The Oxford Handbook of Applied Philosophy of Language 9780192657985, 0192657984

Citation preview

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

FRONT MATTER

p. iv

Copyright Page  https://doi.org/10.1093/oxfordhb/9780192844118.002.0003 Published: May 2024

Page iv

Subject: Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © The several contributors 2024 The moral rights of the authors have been asserted All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the

above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2023948073 ISBN 978–0–19–284411–8 DOI: 10.1093/oxfordhb/9780192844118.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

FRONT MATTER

Contributors  Published: May 2024

Subject: Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

Anita L. Allen, Henry R. Silverman Professor of Law and Professor of Philosophy, University of Pennsylvania Luvell Anderson, Associate Professor of Philosophy, Syracuse University Louise Antony, Professor Emerita, University of Massachusetts Amherst Axel Barceló, Instituto de Investigaciones Filosó cas, Universidad Nacional Autónoma de México Teresa Blankmeyer Burke, Director, School of Arts & Humanities, and Professor of Philosophy, Gallaudet University Herman Cappelen, Chair Professor, Philosophy, University of Hong Kong Noël Carroll, Distinguished Professor, City University of New York Tina Chanter, Lecturer in Philosophy, Newcastle University Archie Crowley, Assistant Professor of Linguistics, English Department, Elon University Matthew J. Cull, Interdisciplinary Research Fellow, Centre for Biomedicine, Self and Society, University of Edinburgh Andrew Cutrofello, Professor of Philosophy, Loyola University Chicago Josh Dever, Professor of Philosophy, University of Texas Austin Cameron Domenico Kirk-Giannini, Assistant Professor, Department of Philosophy, Rutgers University– Newark J. L. Dowell, Professor, Department of Philosophy, Syracuse University

Ray Drainville, Assistant Professor, Stratford School of Interaction Design and Business, University of Waterloo Ángeles Eraña, Instituto de Investigaciones Filosó cas, Universidad Nacional Autónoma de México Daian Flórez, Professor of Philosophy of Science and Technology, Philosophy Department (Caldas University), Universidad de Caldas Michael Glanzberg, Professor of Philosophy, Rutgers University Iz González Vázquez, Department of Philosophy, University of She p. x

eld

Gabriel Greenberg, Associate Professor, Department of Philosophy, University of California, Los Angeles Sally Haslanger, Ford Professor of Philosophy and Women’s & Gender Studies, MIT E. M. Hernandez, independent scholar Jules Holroyd, Senior lecturer, Department of Philosophy, University of She

eld

Timothy B. Jay, Professor Emeritus Psychology, Massachusetts College of Liberal Arts Robin Jeshion, Professor of Philosophy, School of Philosophy, University of Southern California Je rey C. King, Distinguished Professor of Philosophy, Rutgers University Anna Klieber, Lecturer in Philosophy, University of Cardi Quill R Kukla, Professor of Philosophy, Georgetown University John Kulvicki, Professor, Department of Philosophy, Dartmouth College Ernie Lepore, Board of Governors Professor, Rutgers University Matthew McKeever, Postdoctoral research fellow, AI & Humanity Lab, University of Hong Kong José Medina, Walter Dill Scott Professor of Philosophy, Northwestern University Eliot Michaelson, Reader, Department of Philosophy, King’s College London Mari Mikkola, Professor and Chair of Metaphysics, University of Amsterdam Jessica Pepp, Researcher and Burman Fellow, Uppsala University Rebecca Roache, Senior Lecturer in Philosophy, Department of Politics, International Relations, and Philosophy, Royal Holloway, University of London Martina Rosola, Post-doc Fellow, Sustainable Development and Ecological Transition (DISSTE) Department, Università del Piemonte Orientale Esa Saarinen, Professor Emeritus of Applied Philosophy, Department of Industrial Engineering and Management, Aalto University Jennifer Saul, Waterloo Chair in Social and Political Philosophy of Language, Philosophy, University of Waterloo Rachel Sterken, Associate Professor, Department of Philosophy, University of Hong Kong

Una Stojnic, Assistant Professor, Princeton University Stephen Yablo, David W. Skinner Professor of Philosophy, Department of Linguistics and Philosophy, MIT

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

1 Introduction: Applied Philosophy of Language  Luvell Anderson, Ernie Lepore https://doi.org/10.1093/oxfordhb/9780192844118.013.56 Published: 22 May 2024

Pages 1–2

Abstract What is applied philosophy of language? Your answer to this question will likely depend on how you conceive the di erence between applied philosophy and “philosophy proper.” Over the past decade, philosophers of language have increasingly focused on topics that are categorized under the heading “social philosophy of language,” which includes propaganda and ideology, language and gender, and slurs. However, part of our motivation for compiling this handbook is to consider work that draws on more than social and political thought.

Keywords: applied philosophy, social philosophy of language, philosophy, philosophy of language, public philosophy Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

Is there such a thing as Applied Philosophy of Language? Your answer to this question will likely generate thoughts about a broader distinction between applied philosophy and “philosophy proper.” That there is such a distinction seems uncontroversial because applied philosophy is a recognized sub eld consisting of books, journals, handbooks, and even job ads for philosophy positions. Much of the work along these lines has been in applied ethics, but other works in areas like law, education, and art also tend to qualify as applied philosophy. If there is such a thing as applied philosophy of language, you might expect it to derive from a general de nition of applied philosophy. People have interpreted the notion as an enterprise largely concerned with practical matters. According to some conceptions, philosophy proper abstracts away from concrete issues to arrive at universal principles or something along those lines. Alternatively, applied philosophy focuses on the speci c challenges and issues that emerge in a particular practical domain. For example, medical

professionals may need to make decisions about the distribution of care when there is resource scarcity. In these cases, we require the kind of philosophizing aimed at providing explicit norms or rules for speci c situations. The details of concrete situations matter for this sorting out in a way that is taken too narrowly in scope for philosophy proper. But if this is the way to characterize the di erence between applied philosophy and philosophy proper, wouldn’t much of what transpires in the philosophy of language fall under the applied banner? Much of what transpires in pragmatics, for instance, seemingly makes constant reference to details of actual language use. The same can arguably be said of semantics as well. Hence, if we want to admit an applied philosophy of language, the current way of characterizing applied philosophy won’t work. What, then, is Applied Philosophy of Language? The sub eld is quite workable and has yet to assume a de nitive shape. Over the past decade, philosophers of language have increasingly focused on topics that are p. 2

categorized under the heading “social philosophy

of language,” which includes propaganda and

ideology, language and gender, and slurs. This work certainly counts as applied philosophy of language, even if some of it compels theoretical re-imaginings for linguistic analysis. However, part of our motivation for compiling this handbook is to consider work that does not necessarily draw on social and political thought. On this account, we included essays on maps, technology and AI, legal privacy, art, pronouns, and gender, derogatory speech, dogwhistles, and protest. There probably isn’t a clean and uncontroversial way to draw de nitive lines, but there are works that may be categorized under the heading “applied philosophy of language.” These categories could consist of the use of Gricean frameworks to analyze artistic interpretation in Robert Stecker’s Interpretation and Construction: Art, Speech, and the Law; the application of David Lewis’s scoreboard semantics in Rae Langton and Caroline West’s “Scorekeeping in a Pornographic Language Game”; and the use of speech-act theory in Judith Butler’s Excitable Speech. All these studies appear to be instances in which tools from the philosophy of language pertain to other areas, such as aesthetics and the discussion of hate speech in legal contexts. That is, tools from the philosophy of language shuttle through di erent subjects in pursuit of understanding. It also appears that these tools can be utilized in a manner that indicates applied philosophy is underway, in turn, possibly distinguishing some work on slurs as applied philosophy of language. Some studies, for instance, draw from discussions in critical theory about ideology to account for slurs’ semantic and pragmatic function. Thus, both sets of examples intimate two ways of identifying applied philosophy of language: the rst as exporting tools in the philosophy of language to domains outside language, and the second as importing tools outside the philosophy of language to linguistic investigation. The last decade has also witnessed a tidal wave of new philosophical essays with an eye on a public audience. Most philosophers have not been trained in general writing for the public; accordingly, much that has been written has been cast well over the heads of a general audience. But even in the face of these failures to reach and in uence a general audience—considering the urgency with which the public has persistently clamored for attention on key debates surrounding race, gender, and politics, for example—the demands on analytically trained philosophers to turn their expository and critical eyes on these topics and debates have not abated. With these demands in mind, the current volume constitutes one such e ort. There are nearly thirty essays contained in the pages that follow, tackling a spectrum of positions on topics like Conceptual Amelioration, Derogatory Language, Language and Gender, Law, Politics, Bias, Language and Power, Interpreting Maps, and Language in Art. Just how successful these essays are in sharpening our focus on the various debates they address and adopting philosophical positions on these debates is for readers to decide. We hope this handbook provides a fertile stalking ground for those looking for guidance and inspiration in this emergent eld.

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published online: 22 May 2024 Published in print: 02 May 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

2 Amelioration as Course Correction  Sally Haslanger, Stephen Yablo https://doi.org/10.1093/oxfordhb/9780192844118.013.24 Published: 22 May 2024

Pages 5–58

Abstract Ameliorative inferences are normally read as practical: let w henceforth mean N , because N is somehow better than (its earlier meaning) M . So, let “meat” have a meaning that holds of Impossible Burgers. We suggest a theoretical reading: w does mean N, because M (once thought to be the meaning) is problematic. Supporting the theoretical reading is a metasemantic principle of reference electromagnetism, according to which words are repelled by unfortunate meanings and gravitate towards fortunate ones. Metasemantic electromagnetism falls naturally out of Davidsonian/Lewisian principles of charitable interpretation. Speakers are lovers of the good; otherwise, we couldn’t make sense of them. But it may take a while for the good to reveal itself.

Keywords: amelioration, conceptual engineering, semantic externalism, pragmatism, normative externalism, metasemantics, charity Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

She had sacri ced herself to a conception of loyalty that was private and unalterable. —George Orwell, 1984 If we want everything to stay as it is, everything has to change. —Giuseppe di Lampedusa, The Leopard

1. Introduction: Is and Ought in Semantics An ought cannot be derived from an is, according to Hume. He had moral oughts in mind; but we might consider as well oughts from other domains. What a word means has implications, we’re told by Kripke and others, for how it ought to be used: The point is not that, if I meant addition by ‘+’, I will answer ‘125’, but that, if I intend to accord with my past meaning of ‘+’, I should answer ‘125’. Computational error, niteness of my capacity, and other disturbing factors may lead me not to be disposed to respond as I should … The relation of meaning and intention to future action is normative, not descriptive. (Kripke 1984, 37) If meaning facts have normative implications, it seems by Hume’s dictum that meaning cannot be captured p. 6

descriptively. Or perhaps, we should reject the

normative-implication thesis precisely on the ground that

it makes meaning irreducible. Either way, we are thinking like Hume. Whether an is can be derived conversely from an ought was not discussed by Hume. But the prospects appear just as dim. Oughts do not even appear to get much non-deductive—predictive, say, or explanatory— 1

traction on the descriptive facts. Counting P likely on the basis that it ought to be the case is a well-known fallacy, the fallacy of wishful thinking. Explanations of the form ≪P is the case because it ought to be that

P≫ have a famously poor track record. Air does not really rush in because nature abhors a vacuum. It is not because e

cient motion is “better” that objects follow the principle of least action. One might as well try to 2

explain noses, following Dr. Pangloss, as meeting the need for something to hold up our glasses.

Insisting that ought in theoretical contexts be kept at a distance from is seems only reasonable. But there is such a thing as insisting too much. Absurd as Panglossian explanations can be, it goes too far to try to avoid them entirely. Noses may not be for spectacles, but they are in part for getting oxygen into the bloodstream. The desirability of this outcome helps to explain both why noses exist, and why they’re so good at bringing the outcome about. One should not assume out of hand that circumstances are not what they ought to be; or, that oughts are safely ignorable, when considering how matters stand. Philosophers, of all people, should be open to the possibility of a nontrivial connection here. Normative in uence is postulated all over the place in our eld. Think of the link drawn by Davidsonians between the beliefs/desires best rationalizing an agent’s behavior and the beliefs/desires she actually has. Or of Williamson’s contention that the object we perceive is the one that maximizes our perceptual knowledge, knowledge being the best and highest form of belief. Or the proposal by Robbie Williams that I’m to be 3

interpreted so as to come out as responsive as possible to the reasons I’m presented with.

No one should be scandalized, then, by oughts putting pressure on how matters stand descriptively—least of all in the theory of content. And yet, people do seem scandalized by projects in semantics that t precisely this template. What w ought to mean, or what some expert tells us ought to be meant by it, is no guide at all, they say, to the word’s actual semantic properties. Just the opposite, in fact. When a new-sounding meaning N

is recommended to our attention, one assumes that w as it is means something else. Why would N have to

be urged on us, if it was the meaning w already had? ≪w ought to mean N ≫ is accordingly seen as a premise in practical reasoning, not theoretical; the intended conclusion is ≪Let w mean N ≫ rather than ≪w does

mean N ≫. We’ll eventually be pushing back on this, defending the theoretical argument. But let’s explore rst the practical argument for replacing w’s current meaning M with N .

Meaning replacement projects have been criticized on a number of grounds. Some object that the proposed 4

modi cations are not feasible; they cannot as a practical matter be carried out. Others question instead the

advisability of switching M out for N . They’ll be our focus for the next few sections. Advisability worries arise p. 7

when existing meanings

are glori ed or romanticized in some way. M comes with built-in advantages

that its replacement will nd it di

cult to overcome.

Three forms of conceptual romanticism can be distinguished, according to the reasons given for sticking with M . Type-1 (Nature) romanticists point out that M has stood the test of time, meeting our expressive needs over, as Austin puts it, “the lifetimes of many generations.” The idea is sometimes given a naturalselection-y spin. A meaning that’s survived trial by evolutionary re is unlikely to be improved on by a theorist guessing at costs and bene ts. 5

Type-2 (Fidelity) romanticists are impressed by the fact that M is the meaning that got us here. The M

N

questions we care about are framed in terms of w , not w . To abandon these questions in midstream would be counterproductive. Inquiry is supposed to be cumulative, new results building on old. This will not be possible if the words used to state these results are at constant risk of being reinterpreted. The crux for type-3 (Workability) romanticists is that w’s existing meaning “works” for us, which cannot M

always be said for the alternative. We know what we are doing with w ; it can be deployed autonomously, to borrow a term from Kant. Our practice with w would be heteronymous—involving constant checks with an 6

alien-feeling, hard to internalize, rulebook—were we to try meaning N by it instead.

Structure of the paper: Conceptual romanticism is reviewed in the next three sections (secs. 2–4). Section 5 considers how we might cope with meaning change, should amelioration require it. Section 6 raises the possibility that amelioration does not require it. Section 7 introduces the problem of semantic identity over time, on the theory that words are evolving abstract particulars, not unlike songs. Section 8 lays out Lewis’s eligibility theory of meaning-determination. It is suggested in section 9 that eligibility-makers range more widely than Lewis allows. A notion of reference electromagnetism is introduced in section 10 to accommodate eligibility-makers relating to us and our practices. The Frege-inspired “through a glass darkly” model (section 11) takes a rst stab at how reference electromagnetism might work. Section 12 looks at the metaphysics of meaning. Meanings are a special sort of object-kind, like STATUE or SHIP, except that the objects in this case are interpreted words. Disputes about the meaning of terms are like disputes about which of the ship-like objects earlier in the harbor the owner can/should lay claim to. They are resolved not by bringing in further ships (words), but clarifying which of the ships (words) already in play it makes sense to be going on with (section 13). This helps the ameliorist only if w-qua-N (N , recall, is the new meaning) was already in the pool, which might seem ruled out by charity to our former selves (section 14). We respond (in sections 15–17) with a form of charity aimed at minimizing unforgiveable error. Misapplications re ecting limited information, missed opportunities, or the distorting e ects of ideology are not unforgiveable. This leads in section 18 to an alternative model of meaning retention: the “course correction” model. Semantic campaigns are liable to get o

on the wrong foot. They are sometimes put on the wrong foot by

social forces operating in the background. A corrective is needed, and that’s what ameliorists (the kind p. 8

we’re interested in, anyway) try to provide. Some see the

ameliorist as attempting a hostile semantic

takeover. Their hostile takeover is from our perspective the correction of wrong (or suboptimal) turns taken earlier. Mark Wilson speaks in this connection of semantic “detoxi cation” projects, aimed at undoing distortions (Wilson 2006: 545). Important, too, are “desanitization” projects, aimed at exposing awkward 7

details that had been airbrushed out. Debunking accounts can be seen as playing precisely this role (section 19). The nal section attempts to draw these various threads together.

2. Romanticism (1): Nature Start with the idea that existing meanings, forged in the crucible of usage, are unlikely to be improved on by anything reform-minded engineers can devise. This is a common theme in ordinary language philosophy. According to Austin, our common stock of words embodies all the distinctions men have found worth drawing, and the connexions they have found worth making, in the lifetimes of many generations: these surely are likely to be more sound, since they have stood up to the long test of the survival of the ttest, and more subtle, at least in all ordinary and reasonably practical matters, than any that you or I are likely to think up in our armchairs of an afternoon. (1956: 8) 8

Austin’s remarks here t with a larger tradition of treating naturally evolved products as superior. The link he draws between linguistic evolution and biological evolution was there already in Darwin. A struggle for life is constantly going on amongst the words and grammatical forms in each language. The better, the shorter, the easier forms are constantly gaining the upper hand, and they owe their success to their own inherent value. (Darwin 1981: 60, quoting Max Muller) Similar sentiments are expressed in nineteenth-century social theory by Herbert Spencer, to whom the phrase “survival of the ttest” is due (Spencer 1872). His “Over-Legislation” is a paean to the subtlety and complexity of existing social arrangements, described as “utterly beyond human grasp” (Spencer 1865: 62). Attempts at reform run into the law of perverse e ects: In their e orts to cure speci c evils, legislators have continually caused collateral evils they never looked for. (1865: 63) There is no substitute in the end for judgment, the kind that comes from immersion in a culture. Of course, judgment can lead us astray. But the solution is not to install further guardrails; this only exacerbates the problem by impairing judgment further: p. 9

The ultimate result of shielding men from the e ects of their folly, is to ll the world with fools. (Spencer 1865: 349) So, Austin had predecessors. A perhaps surprising successor was Chomsky, who seems to have picked up the 9

10

baton when the two met in the mid-1950s. Chomsky’s target was aspiring social engineer B. F. Skinner : Why should the design of a culture be left so largely to accident? Is it not possible to change the social environment deliberately so that the human product will meet more acceptable speci cations? (Skinner 1965: 426–427) If a science of behavior can discover those conditions of life which make for the ultimate strength of men, it may provide a set of “moral values” which, because they are independent of the history and culture of any one group, may be generally accepted. 11

(Skinner 1965: 445)

This bears on semantic/conceptual engineering insofar as meaning reduces to a kind of socially inculcated behavioral disposition. Behavior which is e ective only through the mediation of other persons has so many distinguishing dynamic and topographical properties that a special treatment is justi ed and, indeed, demanded. Problems raised by this special mode of action are usually assigned to the eld of speech or language… The term “verbal behavior” has much to recommend it … it emphasizes the individual speaker and, whether recognized by the user or not, speci es behavior shaped and maintained by mediated consequences. (Skinner 1957: 2) Supposing with Skinner that verbal dispositions are shaped and maintained by externally organized contingencies of reinforcement, they can be pulled in any direction one likes by rearranging the contingencies. This assumes, however, as Chomsky points out in a scathing review of Verbal Behavior (Skinner 1957), that “the contribution of the speaker is quite trivial and elementary” (Chomsky 1959: 27– 28). Which could not be more wrong, Chomsky thinks; most of the heavy lifting is done by the (brain-based, speaker-internal) language faculty. Skinner’s approach is so caught up in the quest for new solutions that it misses (i) the solutions that have already, over the course of history, been found, and (ii) the way these already found solutions restrict the search space in which reformers necessarily operate.

3. Romanticism (2): Fidelity N

A second objection to replacing M turns on historical resonance. Impressive as N might be in itself, w ’s p. 10

relevance to issues framed with w-as-traditionally-understood

is obscure. Thus Strawson in a famous

discussion of Carnap-style conceptual explication: Philosophical problems about the concepts used in nonscienti c discourse cannot be solved by laying down the rules of exact and fruitful concepts in science. To do this last is not to solve the typical philosophical problem, but to change the subject. (Strawson 1963: 505) Useful work may be done with the spi ed-up concepts; Strawson does not deny it. But this work is liable to bypass the problems originally under discussion. It is loyalty to those original problems that prevents us from replacing M with N . A popular case study here involves formal theories of truth. Tarski believes that “the attempt to set up a structural de nition of the term ‘true sentence’ applicable to colloquial language is confronted with insuperable di

culties” (Tarski 1936/1956: 164). These di

culties stem from the semantic paradoxes,

which on the ordinary, colloquial notion of truth are inescapable. When he says: I now abandon the attempt to solve our problem for the language of everyday life and restrict myself henceforth entirely to formalized languages (165), he is abandoning the traditional Liar problem as handed down from Eubulides. The dialectic is replayed a few decades later in Kripke: Liar sentences are not true in the object language, in the sense that the inductive process never makes them true; but we are precluded from saying this in the object language by our

interpretation of negation and the truth predicate… The necessity to ascend to a metalanguage may be one of the weaknesses of the present theory. The ghost of the Tarski hierarchy is still with us. (1975: 714) But what is to stop us from making Kripke’s interpretation of negation and the truth predicate standard? Children could be raised into the new meanings at logic-themed kibbutzim. Since it would never occur to them to use L is not true to record the fact that L is never in the inductive truth-assigning process made true, L would not cause them the problems it causes us. A population that is not even tempted to call L untrue, despite its acknowledged failure to be true, may have in some sense evaded the paradox; but they have hardly solved it. It is not as if it was just an unforced error on our part to have adopted meanings given which S’s failure to be true entailed the truth of S isn’t true. Various facts become inexpressible (e.g., that The king of France is bald is not true) if this entailment is sacri ced. Kripke-English is expressively impoverished in a way that rightly bothers us. Sweeping trouble under the rug is not the method of true philosophy, especially if it creates an ugly bulge. p. 11

Not much is lost, and a good deal is gained, according to Tarski and Carnap, if for scienti c purposes we adopt spi ed-up meanings not found in ordinary language. Strawson doesn’t necessarily disagree, where scienti c purposes are concerned. But he has his eye also on the unscienti c purposes served by the meanings in use today. One thing we are in danger of losing if these are supplanted is dialectical traction vis à vis existing debates. (How is the new jargon supposed to make contact with the old issues?) Another thing M

we’re in danger of losing is access to the claims we used to make with w . These sorts of worries have taken on a new urgency lately in light of proposals to rework the meanings of words like woman, parent, disability, married, and race. When Cappelen says that “if the revisions succeed, they do not provide us with a better way to talk about what we were talking about,” or Sundell, that “if conceptual engineering takes place, then the representational tools we use to make certain claims will have changed their intensions and extensions,” the focus seems more on expressive access. Dialectical traction comes to the fore when Cappelen adds that answers employing terms with new extensions fail to answer the original questions. These answers concern something new—not what we were originally talking about when we used the [original expressions]. We have the illusion of an answer, but it’s a purely verbal illusion. There’s a lack of continuity of inquiry; the old questions are not being answered. (Cappelen 2018: 100–102) And Sundell, that speakers who express disagreement using the respective meanings will semantically express propositions that are logically consistent. And yet, in many of the cases in which conceptual engineering is said to have taken place, inquiry is continuous, and disagreement is substantive. (Sundell 2020: 588) An author who has been critiqued on this score herself agrees: “Revisionary projects are in danger of providing answers to questions that weren’t being asked” (Haslanger 2000: 34).

4. Romanticism (3): Workability An Adamic ur-language was once postulated in which things went by their true and rightful names (https://en.wikipedia.org/wiki/Adamic language; Losonsky 1992). Spoken originally by everyone, the language fractured at Babel into a confusion of tongues whose word-meaning connections were unprincipled and haphazard. Taken to extremes, this gives us a pure conventionalism according to which any meaning ingenuity can devise is attachable in principle to any old string of phonemes. p. 12

Conventionalism, the idea of meaning assignments as arbitrary, strikes many as just common sense. But it cannot claim a whole lot of support from linguistic theory. Few think any longer of syntactic rules as arbitrary; there are universal constraints on the range of humanly possible grammars and language-speci c constraints depending on how certain parameters are set by French, say, as opposed to Tamil (Baker 2008; Dutilh Novaes 2020). And this is just the beginning. As not any old way of sequencing words makes for a possible sentence, not any old way of stringing phonemes together makes for a possible word. Chomsky and Halle (1968), reviewing phonetic variations on a common English noun, suggest a three-way distinction among such [expressions] as /brik/ (in the lexicon), /blik/ (accidental gap), and /bnik/ (inadmissible). The “accidental gaps” are the items that did not appear in the lexicon but were not ruled out. 12

(470)

The proposed distinction is too coarse, they decide: The matrices /bnik/ and /bnzk/ are both inadmissible in English, but the di erence between them is at least as linguistically signi cant as the di erence between these two matrices and /brik/, /blik/. Hence a real solution to the problem of “admissibility” will not simply de ne a tripartite categorization of occurring, accidental gap, and inadmissible, but will de ne the “degree of admissibility” of each potential lexical matrix. (470–471) Rule-stretching words do exist: tmesis, cwms, phpht, smaragd, chthonic, crwth, subfusc, uhtceare. But their 13

oddity is appreciable, and their pronunciation hard to guess at.

If phonetic/orthographic strings can be more or less admissible as words, ways of assigning them meanings 14

can be more or less admissible, too, a phenomenon known as sound symbolism.

When the tongue is high and at the front of the mouth, it makes a small resonant cavity there that ampli es some higher frequencies, and the resulting vowels like ee and i (as in bit) remind people of little things. When the tongue is low and to the back, it makes a large resonant cavity that ampli es some lower frequencies, and the resulting vowels like a in father and o in core and in cot remind people of large things. Thus mice are teeny and squeak, but elephants are humongous and roar. Audio speakers have small tweeters for the high sounds and large woofers for the low ones. English speakers correctly guess that in Chinese ch’ing means light and ch’ung means heavy. (Pinker 1995: 167) A word’s internal structure (its morphology) imposes further constraints; one doesn’t have to understand “abvolated” to appreciate that it speaks of the past, not the future. Morphological cues are crucial to learnability. How does the child learn to derive Fred was touched by John from John touched Fred, without overgeneralizing and taking John resembled Fred to license Fred was resembled by John? At least part of the

p. 13

answer is that 15

slot into.

how a word is put together bears on what it can mean, and hence the kinds of clauses it can

Even idioms like spill the beans and kick the bucket, traditionally the paradigm of phrases meaning

whatever we want them to, turn out to be constrained in their possible readings by thematic and aspectual 16

features of the words composing them.

The reason for harping on these constraints is that they put the conceptual engineer in a tight spot. How does she know that N is the right kind of meaning for the word w that she wants to attach it to? An example to illustrate the worry is this: Going back to early work on generics, semanticists have distinguished stage17

level predicates (“are making a mess on the lawn”) from individual-level predicates (“are mammals”). The stage/individual distinction can be drawn intuitively, and also on the basis of linguistic patterns ( (Milsark, 2014) (Milsark,

1977) (Carlson, 1977) (Stump, 2012)). Semantically, individual-level predicates express properties that normally are had by items for quite extended periods, often comprising the items’ whole existence. Stage-level predicates, on the other hand, express properties normally had by items for relatively short time intervals. Some examples of both types are as follows: Individual level predicates “is tall”; “is intelligent”; “knows French”; “is a mammal”; “is female”; “is a singer”; “loves Bob” Stage level predicates “is drunk”; “is barking”; “is speaking French”; “is taking an exam”; “is sober”; “is sick”; “is sitting”; “is on the lawn”, “is in the room.” (Leslie and Lerner 2016) A predicate’s level a ects its grammatical possibilities. Stage-level (but not individual-level) predicates are at home, for instance, in locative and perceptual contexts: Sonja is drunk/speaking French/sitting cross-legged by the re. *Sonja is intelligent/knowing French/athletic by the re. Fabrice saw Sonja drunk/speak French/sit cross-legged. *Fabrice saw Sonja intelligent/know French/athletic. Consider now a predicate that has drawn the attention of semantic ameliorators: “is a woman.” It is normally considered individual-level (Ogawa 2001), and the tests bear this out: Fabrice saw Sonja drunk by the re, but he did not see her a woman (or being a woman) by the re. What happens if “woman” is (re)de ned in the way suggested by Haslanger? S is a woman if and only if (i) S is regularly and for the most part observed or imagined to have certain bodily features presumed to be evidence of a female’s biological role in reproduction; (ii) that S has these features marks S within the dominant ideology of S’s society as someone who ought to occupy certain kinds of social position that are in fact subordinate (and so motivates and justi es S occupying such a position); and p. 14

(iii)

the fact that S satis es (i) and (ii) plays a role in S’s systematic subordination, that is, along some dimension, S’s social position is oppressive, and S’s satisfying (i) and (ii) plays a role in that dimension of subordination. (Haslanger 2000)

The de niens here has extrinsic aspects; and extrinsic features are almost by de nition liable to come and 18

go even as the individual remains in relevant intrinsic respects unchanged.

This appears to push us in the

direction of a stage-level interpretation. Sonja is observed and imagined to have various bodily features in 19

Fort Lauderdale over spring break, but not when she is alone on her private island.

A spy keeping tabs on

her over spring break could in principle witness repeated instances of Sonja perceived and imagined to have pregnability-type features. And yet we don’t want to say that Sonja is observed on these occasions (being) a woman. That it legitimates ways of talking that feel borderline ungrammatical seems like a problem for the proposed interpretation. The worry more generally is that some meaning reassignments run so strongly 20

against the grain as to be unworkable—more trouble anyway than they’re worth.

The ameliorist waltzes

through this territory unawares, says the Type-3 (Workability) romanticist, not appreciating how little room there is for maneuver.

5. Coping with Meaning Change The objections of the last few sections take the same form: w’s existing meaning M has crucial advantages given which its proposed replacement N is going to make matters worse. You might expect us to argue that meaning replacement will not make things worse, the objections notwithstanding. That is not the plan; after this section, we will be asking why meaning replacement is thought to be required in the rst place. Not to ignore old school amelioration entirely, let’s consider what might be said if defending N against M  were the plan. Type-1 (Nature) romanticists wonder how invented meanings could hope to improve on meanings “forged in the crucible of usage.” Worries of this general type are discussed by A. O. Hirschman in his critique of reactionary rhetoric, what he calls the “rhetoric of intransigence.” They rest in his view on a wildly implausible thesis: According to the perversity thesis, any purposive action to improve some feature of the political, social, or economic order only serves to exacerbate the condition one wishes to remedy. (Hirschman 1991) It does sometimes happen that e orts at improvement are futile, or wind up aggravating the problem one was trying to solve. The Type-1 romanticist has got to maintain, however, that this is the rule, not the p. 15

exception. How bizarre; it is hard to imagine a less

promising candidate for social-scienti c-law status

than, “Everything back res!” (Hirschman 1991: 11). The trial by re theory is not much help here, if, as Austin admits, superstition and error and fantasy of all kinds do become incorporated in ordinary language and even sometimes stand up to the survival test. (Austin 1956: 185) Must it really back re to try to purge our meanings of superstition, error, and fantasy? Consider 21

Enlightenment-era e orts to purge the meaning of king of its divine mandate aspects.

The usual view is

that these e orts were not pointless at all, but a step on the way to improved social arrangements. The Type-2 (Fidelity) romanticist fears that subbing in a new meaning cuts us o thoughts and questions. This is in line with Hirschman’s

from valuable old

jeopardy thesis, [which] argues that the cost of the proposed change or reform is too high as it endangers some previous, precious accomplishment (1991: 185). But the whole point of putting N in for M may be that the old thoughts and questions are problematic and better abandoned. A related worry is that we are not, if the meanings change, engaged in a continuous line of inquiry, later generations correcting the views of earlier. But does it follow from our meaning di erent things by a word in P, that we are talking past each other when I assert P and you deny it (Cappelen 2018: 97 ; Sundell 2020; Knoll 2020)? This is not obvious. It ought to be enough for disagreement if the 22

proposition that I assert entails the one that you deny. 23

failing in your view because this shared part fails.

Or for my proposition to share a part with yours, P

Or for P to be true for me due to circumstances that 24

con ict with the circumstances that for you make P false.

The ameliorist too quickly assumes that her substitute meaning is workable, says the Type-3 (Workability) romanticist. More attention ought to be paid to the possibility that N will be so impractical that the hoped 25

for bene ts never materialize.

This is in line with Hirschman’s

futility thesis [which] holds that attempts at social transformation will be unavailing, that they will simply fail to “make a dent.” (1991: 185) But we can all agree that various choices of N will be less workable in this or that respect than the existing meaning M . The ameliorist is still in business if she can argue, of speci c proposed reconstruals, that they are so well considered, and speak to such a de nite need, that they bene t us overall. A reconstrual does not have to be cost-free to be worth the trouble, or the nal word on the matter to be worth adopting now. When Tombaugh discovered Pluto in 1930, he thought he’d discovered a planet. He resisted to his death attempts to assign it a lesser status. Assuming that Tombaugh was not simply wrong about Pluto, his disagreement with the International Astronomical Union’s 2006 ruling must be regarded as partly verbal. p. 16

Should it bother us that the ruling

lands us in a verbal dispute with Pluto’s discoverer—a man under no

illusions about its size? Of course it should. One needn’t deny the cost of the ruling to think it a cost still worth paying. The new meaning’s value continues to rise as the celestial bodies pile up that are better quali ed even than Pluto. The caesium-microwave de nition of second is fantastically accurate. Its victory over the traditional de nition in terms of the oscillation of quartz crystals was fully deserved. But relative tness is just that: tness relative to the existing competition. No one would argue on survival-of-the- ttest grounds that we should stick with the caesium de nition forever—even when optical lattices show themselves to outperform caesium on the same tests that had earlier recommended caesium over quartz. One can grant that existing meanings have advantages enough to explain their survival up to now, while reserving judgment on whether they are the best to be going forward with.

6. Who Says the Meaning Must Change? The romanticist’s problems with semantic rebranding were reviewed in sections 2 to 4, and responded to in section 5. Whether you like the responses or not, there is a danger with all of them of conceding too much. This for two reasons.

First, changes in our view of what w means are not always changes in what it does mean. Fish did not acquire a new meaning when it was decided that whales should not be counted into its extension. Nor did atom when the indivisibility requirement was dropped. Jade took it in stride when the kind turned out to be disjunctive. Words can take quite a licking and keep right on ticking—ticking along with the same M . Second, a word’s meaning can change in qualitative respects without changing numerically; the same M 26

continues as w’s meaning even as it evolves (Koslow 2020, 2022).

Mars has had the same enduring planet

as its meaning since before the Explorer arrived. The meaning of water on some views is the scattered object consisting of all the H20 in the universe—an object that takes changing forms over time, and that waxes and wanes and moves about. This is all fair enough, you may say, but sh, water, and the like, are at quite a distance from the words under discussion today: married, artwork, money, meat, labor…. Granted that meaning is retained in the classic cases, the situation with married remains very much in doubt. To get clear on this, we need to consider (a) the meaning-retention mechanisms at work in the classic cases, (b) the extendibility of those mechanisms to married, (c) the availability or not of alternative mechanisms. Questions arise already with (a). It would be one thing if Putnam had gured out the semantics of kind p. 17

terms. The inapplicability of his model to married might then be

cause for concern. But Putnam never

really explains even how water does it. He points at times to indexicality: We have maintained that indexicality extends beyond the obviously indexical words and morphemes (e.g., the tenses of verbs). Our theory can be summarized as saying that words like “water” have an unnoticed indexical component: “water” is stu

that bears a certain similarity

relation to the water around here. (Putnam 1974: 710) This explanation was rejected, though, even at the time (Burge 1979, 1982). The usual view today is that the context-dependence of water, if any, is metasemantic. An outwardly similar word could easily have stood for another substance, just as Mars would have named a di erent planet had Mars-appearances been caused by Venus (Stalnaker 2004). Putnam ran the contingency of reference on the setting in which a word is introduced together with occasion-sensitivity of the sort found with indexicals like here and now. He confused w’s potential to constitute di erent interpreted signs with the semantic phenomenon that is context-sensitivity properly so called (Almog 1984). Nor does it seem plausible that water is, as a matter of xed lexical type, a natural-kind term. Putnam himself gives the example of pencil, which takes on naturalkind-y characteristics when pencils are found to be organisms. Earth, air, and (to a lesser extent) re took 27

the opposite route, losing these characteristics despite being pre-scienti cally on a par with water.

Interestingly Putnam is careful in other passages to keep stipulative-looking features of water out of the meaning, casting them rather as defeasible empirical presumptions: Suppose I point to a glass of water and say “this liquid is called water.” My “ostensive de nition” of water has the following empirical presupposition: that the body of liquid I am pointing to bears a certain sameness relation (say, x is the same liquid as y, or x is the sameL  as y) to most of the stu and other speakers in my linguistic community have on other occasions called “water.” If this presupposition is false because, say, I am—unknown to me—pointing to a glass of gin and not a glass of water, then I do not intend my ostensive de nition to be accepted. Thus the ostensive

I

de nition conveys what might be called a “defeasible” necessary and su necessary and su

cient condition: the

cient condition for being water is bearing the relation sameL to the stu

glass; but this is the necessary and su

in the

cient condition only if the empirical presupposition is

satis ed. If it is not satis ed, then one of a series of, so to speak, “fallback” conditions becomes activated. (Putnam 1974: 702) From this it appears that the arrangements underwriting meaning on a Putnam-like account are not what people often suppose. He is not pulling conditions C out of a hat that qualify any x uniquely satisfying them as the referent “no matter what.” Rather C is believed on empirical grounds to identify the thing, or sort of 28

thing, we wanted to talk about.

This can’t extend to Kripke, you might think, since reference- xing stipulations are for him a source of a p. 18

priori justi cation. But the two are not so far apart if we accept that 29

to empirical presuppositions and, to that extent, defeasible. 30

that has its reference xed by a description.

a priori justi cation, too, is hostage

Anyway it is the rare name, in Kripke’s view,

Even in the meter-stick case, Kripke does not put the referent

entirely at the mercy of the associated description. The stipulator has a length he wants to mark out; the standard meter stick S is chosen with that end in mind. But it is an empirical matter whether stick S really is the intended length. As with Putnam, then, we need to ask what happens if he is wrong and it is a di erent length than intended. It might be, for instance, that the stick is a millionth of an inch long, but emitting magni cation rays that delude us into seeing it as longer. Or maybe the stick is a mile long, but much farther away than anyone had realized. I take it that it is no part of the reference- xer’s understanding of ‘meter’ that it continues to stand for the length of S even if S is much shorter or longer than it appears. (Yablo 2008a: 184) This is part of the reason that Kripke moves quickly on to the model of initial baptism followed by chains of reference-preserving intentions. Some have argued that the initial baptism stands no chance of success unless a sortal comes in to disambiguate; we know a priori that the sortal attaches if the referent exists. Kripke denies, however, that referents must turn out to be of the kind rst supposed: Even if a sortal is used to disambiguate an ostensive reference, surely it need not be held a priori to be true of the object designated. Couldn’t Dobbin tum out to belong to a species other than horses (though super cially he looked like a horse), Hesperus to be a planet rather than a star, or Lot’s guests, even if he names them, to be angels rather than men? (Kripke 1980: 116, n. 58) If Dobbin could turn out to be a zebra, it is hard to see why other animals, thought to be nonexistent, could not surprise us both on the score of existence and biological type. Why should unicorns not be a previously unknown sort of horse? This is one of the likelier scenarios, surely, on which unicorns wind up existing. But then it is not essential to the meaning of unicorn that unicorns are non-horses. Reference- xing as we nd it in nature is not a tool for tying ourselves to the semantic mast. The objections in sections 2–4 were to meaning replacement. To take them on directly, one would argue that it is sometimes advisable, the stated concerns notwithstanding, to replace M with a distinct meaning N . (This was attempted in section 5.) But replacing M with N may not be what the ameliorator was proposing to do in the rst place. Why should this seem like the only option? Part of the answer is that philosophers have

been treating decades old treatments of semantic continuity as the last word. Models that were intended only to open up space for w retaining its meaning are taken now to exhaust the possibilities for such an outcome—as though Putnam had never spoken of cats turning out to be robots, or Kripke of the possibility of discovering that Lot’s guests are angels rather than men.

p. 19

7. Identity over Time Meaning retention is not such a demanding a air as we have been led to believe. Before anyone asks, “How demanding is it, then?” a caveat is in order. Words are various, multitudinous, and (see below) perhaps even plenitudinous. This casts doubt on the prospects for an answer covering every case. How much damage can a word sustain before it gives up its meaning? has some of the same wrongheadedness as, How long is a piece of string? There is no point in wondering about the length in general of a piece of string, because string can be cut to all kinds of lengths. Wondering about the resilience in general of an interpreted word is similarly 31

confused; words can be tted out with all kinds of identity conditions.

Perhaps the question cannot be avoided entirely. Words are abstract continuants persisting through time 32

and across modal space. Something will have to be said about how they do it.

It will not be de nitive or

precise; why would the identity of words be any more reducible to lower-level events than personal identity turned out to be? But we can try at least to map out the territory. One more caveat before we begin, prompted by the analogy with personal identity. An account of identity for Xs appeals sometimes to identity relations among the members of another category Y. Singleton sets, e.g., {Cicero} and {Tully}, are identical i

they have the same object as their sole member (Cicero = Tully). More

generally, x1 = x2 because (or, partly because) y1 = y2, where yi is some important part or aspect of xi. Obviously, there’s a danger here of explaining the obscure in terms of the still more obscure. A (much maligned) approach to personal identity takes persons p to be bodies b amalgamated with souls or minds S ; 33

p1 and p2 are the same person just if b1 = b2 and S 1 = S 2.

The amalgam theory rests on an assumption that

philosophers have come to reject: that whether minds/souls are identical is prior to the identity of the persons pi whose minds/souls they are. Asked to choose between (i)  p1 and p2 owing their identity to some preexisting fact about souls, and (ii)  p1 and p2 sharing a soul on account of being the same person, most of us would go for the latter. (If soul identity were prior, we would have to contend with the possibility of new souls rushing in to replace the old every morning—that was Locke’s objection to the theory.) An analogy suggests itself here between persons and interpreted words. Where metaphysicians speak of bodies, minds, and persons, semioticians distinguish signi ers, meanings, and signs. Words qua signi ers —what Kaplan (1990) calls presemantic or syntactic/lexical words—line up with bodies. Another good term is somatic word, on account both of its bodily connotations and the nice verbal contrast with semantic. Somatic 34

words w will be called words. p. 20

Kaplan’s example is the word spelled c-o-l-o-r in the US and c-o-l-o-u-r in

Canada. Madagascar (which he also

mentions) survived both orthographic changes and referential ones;

it denoted rst the Mogadishu region and then an o shore island. Girl, once spelled g-u-r-l-e, was genderneutral before being restricted in the sixteenth century to female children. About souls S and persons p, we can be briefer. Souls line up with meanings M and persons, with meaninginfused or semanticized words w (henceforth words). There may be just the one Madagascar, but there are two Madagascars—two signs built on the same signi er. Madagascar1 has the mainland referent, while Madagascar2 stands exclusively for the island. Likewise there are two girls, one (obsolete) applying to children of either sex, another, still in use today, applying only to, well, girls.

Words as abstract continuants are individuated genealogically. There would have been a second Madagascar, had Marco Polo invented his word out of whole cloth rather than picking it up from a book. There are two banks, as we know, standing for nancial institutions and riversides. There are also two banks, one deriving from Old High German via French banque, the other from Old Norse banki. Crucially, there would still have been two banks even if they, by some strange uke, meant the same. And there would still have been two banks in that case; for w1 does not count as the same word as w2 unless w1 and w2 are the 35

same word.

8. Embodied Meanings Let’s try now to read these distinctions back into the theory of personal identity. What corresponds on the sign side to the (rejected) picture of persons as composites of minds and bodies? Each w would be an amalgam of presemantic w with a pure meaning M . w1 would be the same sign as w2 just if w1 was identical to w2 and M 1 was identical to M 2. Such an approach seems, as with persons, to get things backwards. Meanings do not have self-standing identities any more than souls do; they are carried along by the words whose meanings they are. Suppose that w1 is meter circa 1800, and w2 is meter as we know it today. Whether w1 = w2 is not obvious, but the answer is plausibly YES. The operational de nition has changed—from one ten-millionth of the distance from the equator to the North Pole then, to the distance light travels in 1/299,792,458 of a second now. But no one thinks, to our knowledge, that meter circa 1800 was knocked out of circulation by a similarsounding word linked constitutively to the motion of light. According to the approach we have just described as backwards, meter1 = meter2 (partly) in virtue of a prior, independently obtaining fact about their meanings M 1 and M 2. It is hard to think what the independently obtaining fact would be. The light-based standard is orders of magnitude more precise, and may be 36

superseded by something more precise yet. SECOND

It is not as if DISTANCE LIGHT TRAVELS IN 1/299,792,458TH OF A

was on the cards all along, the way the de nition of gold as the element with such and such an 37

atomic number was always supposedly on the cards. p. 21

No, the reason meter1 agrees in meaning with meter2 is that they are the same word, and each word has one meaning. Their sameness as words turns in part on how meter has been used over time, and the causal/historical connections, between earlier uses and its uses today. (E.g., did speakers intend to hold later uses hostage to the methods available in 1800?) These connections would not be available as persistencemakers, if words owed their identity over time to independently constituted facts about the identities of free- oating meanings. A more challenging example is adult. It’s agreed all around that adult1, the word constituted by adult in 1800, was restricted to people age 21 and over, while adult2, the word constituted by adult today, applies 38

also to 18- to 20-year-olds.

Does adult retain its meaning over this period? Is adult1 the same word as

adult2? The answer depends as before on the details of adult’s use over 1800 to 2021, and the role adult has played over that period in speakers’ lives. One view, of course, is that the word comes to mean something di erent when the age of majority changes. Alternatively, we could say (with Lampedusa) that the standards for adulthood must change for adulthood itself to remain the same. Intuition goes with Lampedusa here. But we need not insist on the point; it’s enough that a case can be made for adult meaning the same throughout. To observe that 21 > 18 is not enough to settle the question. There are a number of ways of reconciling the drop to 18 with adult possessing a xed—not to say unchanging!— meaning ADULT (A , for short) over the whole period.

Option 1) A does not deliver a single eternally valid cut-o . It determines a rule giving cut-o s as a 39

function of T, the time; A (1800) = 21, A (2021) = 18.

Option 2) A determines a rule that looks to governing practices rather than time. A (Virginia law circa 1800) = 18 ≠ 21 = A (Virginia law circa 2021). Option 3) A always delivers some xed cut-o

or other, but the cut-o

changes as the practice evolves. 40

Compare how games retain their identity through changes in the rules.

Option 4) A always delivers the same eternally valid cut-o . Our forebears were wrong about the cut-o for A ; 18-year-olds were adults even then. Or perhaps the mistake is on our part, or both our parts; or perhaps it’s just hard to tell. These are roughly sketched, and not clearly distinct in some cases. But they’re enough to make the point. Changes in operative standards do not force our hand where adult is concerned. They do not prevent adult from constituting the same word over this period, with (what follows from word identity) a single meaning. For these reasons and others to be mentioned, we will be prioritizing semantic words w—aka signs—over somatic words w and meanings M . The di erences are probably clear enough, but it won’t hurt to lay some of them out explicitly. Words vs. meanings: These are wholly distinct, though w may of course have a meaning—or a bunch of p. 22

them, depending on the metasemantic setting. Somatic

words, though abstract, have

orthographic/phonetic properties, in the way that songs do. Not so meanings; they are not the sort of thing one can write or pronounce. Words vs. words: w is liable to take on a range of meanings; w has just the one meaning M . w has its meaning(s) contingently, but w cannot mean anything other than what it does. w’s meanings lie metaphysically downstream from w; they play no part in making w the word that it is. w’s meaning pertains to the very nature of w. Words vs. meanings: w has orthographic/phonetic properties corresponding more or less to those of w. Like w, w is the kind of thing one can token, by typing out letters (or etc.) in the proper order. Nothing like that is possible for meanings. What one can do is token a w in whose nature it lies to mean such and such. These distinctions bear on the problem of tracking meanings through time. Just as, to use an example of o

Montague’s, we can speak of the water temperature’s rising, but not of the temperature itself (90 F, as it might be) rising, it is M qua meaning of w that changes, not M considered in itself. Imagine, to see why this is so, that w1 and w2 both mean M in 1800. Where should we look for that shared meaning in 2021, to determine whether it has broadened or narrowed? The problem is that it might have broadened qua meaning of w1, while narrowing qua meaning of w2. To insist on knowing whether M “as o

such” has broadened is like asking whether 90 F as such has risen. It’s the same with M ’s continued existence. Perhaps M hangs on as the meaning of w1 while being succeeded as the meaning of w2 by a numerically distinct N . Has M itself survived, or has it been replaced? A bad question, but hard to avoid if one takes meanings to be the primary bearers of semantic persistence properties rather than meaning-infused words. You might think we could dodge the question another way—by junking dynamic meanings and con ning ourselves to the changing relations between presemantic words w and static meanings. Adult meant PERSON 21 OR OLDER

in 1800; what it means now is PERSON 18 OR OLDER. This approach obliterates crucial

distinctions, however. There is all the di erence in the world between w applying more (less) widely because (a) its old meaning M has broadened (narrowed), and (b) it has taken on a broader (narrower) new meaning N . In case (a), w constitutes the same sign w over time, meaning M throughout, as that meaning itself becomes more or less demanding. E.g., adult means ADULT early and late, but one had to be older then to meet its demands. (b) is the case where a signi er w that once constituted the sign w comes later to constitute a di erent sign: w*, meaning M *. Girl ts the second model on the theory that it was true originally of all children. Rather than CHILD narrowing, so that boys were no longer CHILDREN , girl came to mean FEMALE 41 CHILD . How to draw the (a)/(b) distinction in the means-at-T framework is quite unclear.

p. 23

9. Eligibility Theory Back now to the revisionary-sounding de nition of woman as “an individual regularly and for the most part observed or imagined to have certain bodily features presumed to be evidence of a female’s biological role in reproduction, these features playing into the dominant ideology in such a way as to subordinate the relevant individuals.” One may be forgiven for wondering, with Mark Richard (2019), whether this de nition crosses the line between conceptual therapy and stipulative rebranding. (194) One may be forgiven, at any rate, IF the ameliorist is engaged in “a project that seeks to change the meaning of a term” (197)—where changing the meaning is putting a distinct new meaning in place of the meaning woman has had up to now. Is that in fact the project? Amelioration is apt to be portrayed even by its advocates as involving a numerical, not merely qualitative, meaning change. And yet, if one looks at the passages that suggest this, numerical 42

change is more often contemplated as a possibility than laid down as an absolute requirement.

 Haslanger

(2020), in a discussion of “how conceptual amelioration is possible,” asks, Can the content of a concept change, while it remains the same concept? Must we accept content essentialism for concepts? (241) If concepts are meanings and contents are partitions of logical space, then she is asking whether a word can retain its meaning through changes in its intension. This is what occurred it seems with adult, and, to use 43

one of Richard’s examples, rape.

You may be skeptical. Ameliorists speak of associating w with N  because  M  is/was problematic. Isn’t that suggestive, all by itself, of putting one meaning in for another? It is. But suggestions can be implicatures rather than entailments; and that appears to be the case here. w ought to mean  N , therefore it doesn’t mean  N has all the plausibility qua entailment of I tried to make dinner, therefore I didn’t make it. Your having tried to do something is evidence, on the whole, that you succeeded in doing it—granted that it may be a funny thing to say if you succeeded. The probability of succeeding, conditional on trying, exceeds that of succeeding conditional on not trying. The meaning case seems similar. M ’s ill-suitedness to the task could be evidence that w has all along meant something else, granted that N  would be better is a strange

44

choice of words if one thinks N is the meaning w already has.

 I tried to make dinner is the right thing to say,

note, to an audience who suspects you might have made it inadvertently. Similarly N  is what it ought to mean could be the right thing to say to an audience that is shocked to nd a word meaning what it should. This is the possibility that people seem to be overlooking in the amelioration wars. The very p. 24

unfortunateness of w meaning M —the fact that no right-thinking person would

want it to mean that—

is, or could be, reason to think that w doesn’t mean M . Conversely, w lending itself to another interpretation N

counts, or could count, in favor of N being the correct interpretation.

Consider an analogy. Kripke asks why plus should express addition, not quaddition (an operation that agrees with addition until the sum gets too big, and then always outputs 5), assuming that both interpretations can be made to t with actual usage. Lewis replies that addition is the more “eligible” interpretation—more eligible, he thinks, on account of being more natural (Kripke 1984; Putnam 1980; Lewis 1984). That N is the more eligible candidate is not only evidence, according to Lewis, that w means N rather than M ; N ’s greater eligibility is part of what makes it the meaning. The evidential link tracks that underlying constitutive fact about meaning relations. Let a contender for the role of w’s meaning be an interpretation that ts in some sense with usage (Lewis 1974). Charity will be a factor here; an interpretation that convicts speakers of too much error ts worse with usage than one attributing less error and/or (since mistakes are a fact of life) making the errors explicable. However we ll in the details, Lewis grants to Kripkenstein that use-consonance leaves meaning radically underdetermined. (“What we say and think … doesn’t settle what we refer to” [Lewis 1984: 226].) Convinced that meaning is not as underdetermined as all that, he goes looking for a tiebreaker, and nds it in eligibility: N

is the most eligible contender.

∴

 N is what w means

is an epistemically good inference, for Lewis, because that’s how meaning is determined. It lies in the nature of meaning (not that Lewis would put it this way) that the most eligible candidate is the winning candidate.

10. Eligibility-Makers The importance of eligibility as a lter on use-consonant meanings is not in dispute here. But is naturalness 45

the only factor in eligibility?

A case can be made for looking, too, at some of naturalness’s presumed

consequences, among them normative consequences (see below). These consequences could be eligibilitymakers in their own right. Natural properties are supposed to be projectible; it is epistemically more reasonable to see in this green (grue) emerald evidence that the next emerald will be green than that it will be grue. Some relate naturalness to the fact that some true beliefs (or items of knowledge) are more valuable than others. If we care about the truth in the way we should, we will prefer the opportunity to nd out whether neutrinos have mass to the opportunity to memorize the contents of a telephone p. 25

directory. So long as we can make sense of

naturalness for propositions as well as properties, it

is tempting to explain this by claiming that the more natural a true proposition is, the more epistemic value one achieves in believing it (or knowing it, or having high credence in it), at least ceteris paribus. (Dorr and Hawthorne 2013)

Further down the same road is Sider’s (2013) claim that it is better “to think and speak in joint-carving terms” and “worse to employ non-joint-carving concepts” (61). If we agree that our words are better taken to express some concepts (GREEN ) than others (GRUE ), must we agree, too, on the reasons for this, that GREEN is more natural, or joint-carving, than GRUE ? The problem is that from a cosmological perspective, GREEN is quite unnatural. Color impresses man; raven black impresses Hempel; emerald green impresses Goodman. But color is cosmically secondary. Even slight di erences in sensory mechanisms from species to species … can make overwhelming di erences in the grouping of things by color. Color is king in our innate quality space, but undistinguished in cosmic circles. Cosmically, colors would not qualify as kinds. (Quine 1969: 127) 46

Suppose with Quine that GREEN cuts very far from the joints.

What, in that case, makes it a more eligible

meaning than GRUE ? We need, it seems, to nd another eligibility-maker. The fact that GREEN rationalizes inductive inferences better than GRUE seems a point in its favor, even without the supposed foundation in naturalness. Further advantages suggest themselves once the evaluative door is opened. Color talk would be irresponsible if the properties it attributed were not visually recognizable (recognizable, therefore, without knowing the time). Colors are evolutionarily important and play a crucial ecological role. They are far more theoretically fruitful than schmolors. If fruitfulness and the like can enhance eligibility even where naturalness is 47

lacking, then eligibility is not just a matter of naturalness.

Thinker-involving o shoots of naturalness can

play a reference-determining role all on their own. A second reason for doubting that eligibility boils down to joint-carving-ness is this. A candidate referent X is natural “in itself,” not because of its place in the universe. The causal/counterfactual relations X bears to other things do not make it natural; it is thanks, in part, to being natural (e.g., nondisjunctive) that it stands in those relations. And yet relational di erences, according to most philosophers, are what call the shots when it comes to breaking referential ties. An example from Williamson (2007): Emanuel sees a stranger, Celia, standing some distance away. Looking at her face, … he ascribes a character and life-history in considerable detail. In fact, none of it ts Celia. By pure coincidence, all of it ts someone else, Elsie, whom Emanuel has never seen or heard of… A principle of charity that crudely maximizes true belief or minimizes error therefore favors Elsie over Celia as the referent of the pronoun in that context. (262–263) p. 26

48

Many would reach here for causal history:

A causal theorist of reference will point out that Emanuel’s use of “she” in this context is causally related to Celia … In this case, the speci c link is that Emanuel is perceptually attending to Celia and using “she” as a perceptual demonstrative. (263) But what makes that the right sort of causal link to secure reference? A natural idea is this. The perceptual link from Celia to Emanuel matters because it is a channel for knowledge. If “she” refers to Celia, then, in the circumstances, Emanuel expresses knowledge

when he says “She is standing in front of me” … The assignment of Celia wins because it does better with respect to knowledge. (262) This is the knowledge-maximization account of reference. One take-away point is that Williamson agrees with Kripke, Putnam, Burge, and most everyone else that it takes more than naturalness to break referential ties. It takes something relational, with an emphasis on relations to us. A second take-away is that the relation may have a normative character. The normative element is to do with knowledge, on Williamson’s account, but we should keep an open mind on this: X may owe its goodness as an interpretation of w to the 49

goodness in some broader sense of w expressing X .

Terminology going forward: It will be simplest just to give Lewis the phrase reference magnetism; it is by de nition exerted by a would-be referent/meaning X entirely on the basis of X ’s naturalness. Let’s give him eligibility as well, as an umbrella term for factors of the right sort to break ties between use-consonant contenders X and Y . Lewis’s main substantive claim in these terms is that magnetism is all there is to eligibility; it’s the sole tiebreaker. This cannot be right, we have said. Eligibility involves a great deal more, 50

such as facts about w’s causes and e ects (or those of its tokens).

Particularly important are w’s causal 51

relations to us, including forward-looking relations to do with theoretical/practical fruitfulness.

11. Reference Electromagnetism 52

Naturalness as Lewis conceives it is a feature that X has necessarily.

And it seems to be intrinsic to X as 53

well, though how to cash second-order intrinsicness out in his system is not obvious.

This is important

because the non-Lewisian tie-breakers noted above—causal relations to w-talk, projectability of wmeaning-X , epistemic superiority of X- directed theorizing, knowledge-conduciveness of taking w to express X —are contingent and extrinsic. They have in most cases to do with how X relates to us and our winvolving practices. The idea of ranking contents on the basis of how they relate to us and our practices goes back to William p. 27

James. James nds himself wondering (in “The Will to Believe”)

not which X s are eligible to serve as

meanings, but which propositions P are eligible to be believed. A proposition’s doxastic eligibility depends for James on its likely role in our lives—on what we can hope to do with it: Let us give the name of hypothesis to anything that may be proposed to our belief; and just as the electricians speak of live and dead wires, let us speak of any hypothesis as either live or dead. A live hypothesis is one which appeals as a real possibility to him to whom it is proposed (James 1992, 457–8). A dead hypothesis is one that “makes no electric connection with your nature,—it refuses to scintillate with any credibility at all.” His (1992) point about doxastic eligibility-makers, that deadness and liveness in an hypothesis are not intrinsic properties, but relations to the individual thinker, (458) echoes the suggestion above about expressive eligibility-makers (such as, connecting up causally with tokenings of w); some are relational, not intrinsic.

Here is the position we’ve arrived at: a candidate meaning X ’s eligibility is a function not only of its naturalness—a feature X has intrinsically and necessarily—but also its connection with speakers and their activities—a feature X has extrinsically and contingently. A handy name for the connection factor, given that the rst is magnetism, is electricity (taking inspiration from James’s talk of electrical connections with our nature). Putting the two together, we obtain the notion of N outscoring M on electromagnetic grounds. Now we begin to see how speakers and their agendas are worked into an otherwise Lewis-like scheme. The meaning of w is not the most magnetically attractive, but the most electromagnetically attractive, contender.

12. The Through a Glass Darkly Model A model of electromagnetic attractiveness would be nice. Frege takes some rst steps in the Foundations of Arithmetic: What is known as the history of concepts is a history either of our knowledge of concepts or of the meanings of words. Often it is only through great intellectual labour, which can continue over centuries, that a concept is known in its purity, and stripped of foreign covering that hid it from the eye of the intellect. (Frege 1950: xix) He gives as examples “the notions of function, continuity, limit, in nity, negative and irrational numbers.” Where concepts like these are concerned, p. 28

there is no development, no history… If instead … one said ‘history of the attempt to grasp a concept’ or ‘history of the grasp of a concept’, it would seem to me far more appropriate; for the concept is something objective that we do not form and is not formed in us, but that we try to grasp and nally, it is hoped, really grasp—–if we have not mistakenly sought something where there is nothing. (Burge 1984: 7–8) To call the concept “something objective … that we try to grasp” is tantalizingly unclear on the issue that concerns us here: whether M is the meaning on account of features internal to M (like naturalness), or on 54

account of M ’s relations to other things (a linguistic community, as it might be). Let’s dig into this a bit.

A paradigm of what Frege has in mind by “the concept known in its purity” is Weierstrass’s quanti cational de nition of limit (continuity, rates of change): the limit as n approaches in nity of xn is L i xn

for all ε greater than 0, there is a δ greater than 0 such that

is within ε of L when n is greater than N

f is continuous at point a i

for all ε greater than 0, there is a δ greater than 0 such that f(x) is within ε of

f(a) whenever x is within δ of a. f’s rate of change at a is r i

for all ε greater than 0, there is a δ greater than 0 such that |f(x) - f(a)|÷|  x-a|

is within ε of r whenever x is within δ of a. Frege seems to think that these (nineteenth century) elucidations capture what Newton and Leibniz meant by “limit” (etc.) much earlier. Assume for the moment that he is right––and right moreover because these are the most eligible contenders. The question is this: Do they win out purely on grounds of naturalness

(magnetic grounds), or on grounds partly of hooking up in favorable ways with human practices (electromagnetic grounds)? Some of the good-making features Frege stresses are to do with capturing arguments in the net of logic, so that they come out not just analytically but logically valid (Yablo 2010). The second de nition, for instance, allows us to prove by logical means that the product of continuous functions is itself continuous. Is this a magnetic advantage? Perhaps, if one thinks logically articulated notions are especially joint-carving. Or perhaps not. Logical validity comes in handy when it comes to managing huge masses of data with a view to pulling out the information one needs. Manageability and retrievability seem like electrical advantages; M is attractive, not because it traces natural contours, but because of what it lets us do. Applicability is a major theme in Frege’s writing on concepts. He is famously keen on fruitful concepts, ones that facilitate proofs, confer insight, and bring disparate-seeming subject matters together (Tappenden 1995). Kant comes in for criticism on this score. He (Kant) seems to think of concepts as de ned by giving a simple list of characteristics in no special order; but of all ways of forming concepts, that is one of the least fruitful. (Frege 1950: 100) p. 29

Frege would have us take our cue, rather, from fruitful de nitions in mathematics, such as that of the continuity of a function. What we nd in these is not a simple list of characteristics; every element is intimately, I might almost say organically, connected with the others … the more fruitful type of de nition is a matter of drawing boundary lines that were not previously given at all. What we shall be able to infer from it, cannot be inspected in advance; here we are not simply taking out of the box again what we have just put into it. The conclusions we draw from it extend our knowledge. (100) A concept’s fruitfulness may owe something to its t with natural joints. But its t with our cognitive makeup is important too—as is what the concept can do for us inferentially and epistemologically, and 55

(especially) what we need from the concept given the questions that we’re in fact grappling with.

The

utility aspect is stressed as well by Carnap (1962) when he lists fruitfulness as one of four criteria for a 56

successful explication.

Suppose for argument’s sake that seventeenth-century mathematicians were, indeed, confusedly en rapport with Weierstrass’s M , which quanti es over real valued εs and δs. Why was it M they had in mind, rather than Robinson’s M *, which construes rate of change as a ratio of in nitely small hyperreals (Robinson 2016)? Lewisians will say that M is more natural than M *. A more psychologistic answer, developed by Peacocke and others, is that M does better justice to Newton’s “implicit conception” of rates of 57

change.

A third and more interesting suggestion is that M is more fruitful than M * in the sense of facilitating certain 58

kinds of progress. This admittedly puts a word’s meaning at T at the mercy of developments still to come. Not only may we have been in no position at T to appreciate M ’s advantages, M may not even have had the

advantages then. The decision might not yet have been taken to pursue the projects that M facilitates better than M *. It was the ε-favoring road that was eventually traveled, not the ratio-of-in nitesimals-favoring 59

road. But it could at T have been either.

Can the through-a-glass-darkly model accommodate this sort of scenario, where velocity (say) turns out to have meant M all along, but would have all along meant M * had the eld taken later a di erent course? Perhaps it can, with enough fancy footwork. E.g., one could try calling it indeterminate at T whether it was N being seen through the glass darkly or N *. But the model does seem stretched to its limits here. Or consider Mark Wilson’s example of Druidic island dwellers who, on seeing a B-52 crash land on the beach, exclaim, “Lo, a great silver bird has fallen from the sky” (let’s henceforth use “burd,” lest our intuitions be hijacked by the English meaning). The plane does fall, he argues, into the post-crash extension of “is a burd.” But what about the pre-crash extension? Crucially, from Wilson’s (1982) perspective, if the hapless aviators had crashed in the jungle unseen and were discovered by the Druids six p. 30

months later as they camped discontentedly around the bomber’s hulk,

their Druid rescuers

would have proclaimed, “Lo, a great silver house lieth in the jungle.” (550) The bomber cannot be allowed into the pre-crash extension of “burd,” he suggests, since depending on its mode of arrival, the islanders could just as easily have classi ed it as a “house” (or “hause,” by analogy with “burd”). The evidence for assignment of an extension to a predicate should be limited to such linguistic behavior as can be reasonably extrapolated from the community’s contemporaneous practice and should not re ect accidental features of the society’s later history. (552) An alternative model, the course-correction model, will be suggested later that disagrees on this point; it allows for the possibility of assigning “burd,” even pre-crash, a meaning B that holds of the bomber. Equally viable pre-crash was a meaning B * that excluded the bomber. But the pool of viable candidates shrinks over time. And B * loses its viability when the island dwellers brand the bomber a great silver bird, without worrying for a moment that they have modi ed or extended the meaning of “bird” [“burd”] … that classi cation appears spontaneously natural to them in the circumstances. (551) Our di erence with Wilson is that we propose to acquiesce in the Druidical intuition that the meaning hasn’t 60

changed.

When B * loses its viability, nothing is left for “burd” to have meant even pre-crash but B .

Another way to put it is that B turns out to have been the better thing for the word to have meant all along. That the reasons for its betterness do not emerge until after the crash is an important point. But its relevance to the actual meaning facts may be doubted. The Shroud of Turin, taking the legend at its word, came into existence before the events occurred (serving as Christ’s burial shroud) whereby it was the Shroud of Turin that came into existence. Lincoln’s assassination took place, on some views, in the Ford Theater, before the event (his death) occurred that made it an assassination as opposed to just a shooting. What’s so strange then about a word meaning M rather than M *, before the events occur thanks to which M is the meaning?

13. Meanings and Kinds This is a paper in metasemantics, not metaphysics/ontology. But one might worry that an ontology of the kind needed will be on the spooky side. Actually, the framework sketched above (with its words, words, and meanings) can be made to jibe fairly well with standard views of time-extended particulars. The particulars p. 31

in this case are

abstract, not concrete, but they have concrete tokens (concrete articulations, in the

terminology of Hawthorne and Lepore (2011)); that will be enough for our purposes. Start with the idea of somatic words constituting semantic ones. Philosophers have been studying for a while the relations “higher” objects bear to the “lower” objects that compose or constitute them: a statue’s relation to its clay, say, or a person’s to their body. As these examples suggest, the higher object e is typically of a di erent kind K than its constitutor e; e on a useful metaphor is the entity e generates when raised to the power of K . Persons, for instance, are what living 61

human bodies generate when raised to the power of PERSON .

One would like to extend this if possible to the

relation between semantic words w and the somatic words w that constitute them, so that w is w raised to the power M . The parallel is not exact, since words as we’re conceiving them are prior to their meanings, but it will be helpful nevertheless. An obvious point to begin with is that statues look and feel just like the corresponding hunks of matter; there’s agreement in shape, size, weight, and so on. Words being multiply realizable abstract particulars are not of any particular shape or size. But their realizations (“tokens”) have such properties, and when the question is referred to them, we get the expected answer: tokens of adult are by and large indiscernible from the corresponding tokens of adult. Written inscriptions look the same, spoken instances sound the same, and so on. All this ts with the conception of w as the unique w-generated word meaning M . At the risk of multiplying terminologies, where e on a familiar sort of hylomorphic perspective is e-qua-K , w on the 62

picture under construction is w-qua-M .

A refresher on kinds, since they will be serving as a model for meanings. Kinds are properties, more or less. But they are not “regular” properties like Sparky’s property of weighing fteen pounds. That he weighs fteen pounds tells us how Sparky is, where his weight is concerned, but not what he is. Aristotelians 63

sometimes put this by saying that where a thing has its regular properties, it is (of) its kind.

The point of

the is is that kinds are tied up with identity. Regular properties have (i) application-conditions—conditions distinguishing for each world W and time T the φs in W at T from the non-φs, and kinds do, too. But only kinds have (ii) persistence-conditions—conditions endowing each K (at T, in W) with an intraworld past and future, and a transworld career. The kind STATUE , for instance, determines (i) which hunks of matter (at T, in W) compose a statue, and (ii) what the life histories and transworld careers are of the entities thus composed. p. 32

A certain piece s of clay might qualify under (i) to constitute a statue s of Socrates; (ii) then comes in to tell us that (a)  s cannot survive abrupt replacement/rearrangement of s’s clay; (b)  s could not have been made originally of a piece r of clay disjoint from s;

(c)  s goes out of existence if/when s is squashed at, or formed into an ashtray. Are meanings, or the property of meaning thus and such, anything like kinds in these respects? How does a meaning’s role in raising w to the level of an interpreted word w compare to a kind’s role in raising s to the level of a statue s? Take girl. Given that s’s clay at a time is the clay comprising at T the hunk s of matter of which s is then composed, an (a)-counterpart would target the letters comprising the word girl: g, i, r, and l, in that order. One is looking, then, for something like (a*)  w cannot survive sudden replacement/rearrangement of its letters. 64

An abrupt reconstitution as f-e-m-c-h-i-l-d would be too much for girl to bear.

But slower, less drastic

changes might be survivable; the word was once spelled g-u-r-l-e, after all. Corresponding to (b) is the fact, if it is one, that (b*)  w could not have been written originally with entirely di erent letters. Loosely analogous to the limits (c) puts on changes of spatial extension, are those (c*)  w goes out of existence if/when w becomes true of very di erent things puts on changes of semantic extension. The things girl was true of changed in the 1530s when it was restricted to females. The word it had formerly constituted went out of existence then, in accordance with 65

(c*). A new girl took up residence in girl when the old girl left.

14. Many-Thingism and Many-Wordism A classic puzzle case in the theory of material constitution is the Ship of Theseus. A certain ship starts life in the Athens harbor. Over the course of various adventures, its planks are gradually pulled out and replaced. The original planks, stored away in a warehouse, are reassembled at a later date on the same plan. The ship we started with at T thus winds up with two descendants at U, both of which have some claim to be “it.” Does the T-ship survive as its spatiotemporal descendant at U (with planks replaced)? Or does it survive as its material descendant at U, made of the same planks as composed it originally? p. 33

If there is a standard line on this, it’s that the question should be nessed. Talk of ≪The ship in the harbor at 66

T≫ has a schematic element.

Of the various ways this might be modeled, the simplest takes ship-

descriptions (≪The ship …≫) to be referentially indeterminate between items of slightly di erent kinds; the associated persistence conditions are generally in alignment but, occasionally, in philosophical puzzle

cases, pull apart. One candidate referent in the case at hand is a form-based ship s1. s1 is still in the water at U but is made of di erent planks. Another, the matter-based ship s2, sits at U in the town square, where it plays some kind of ceremonial or historical-marker role (Hughes 1997). There were two ship-like items in the harbor initially, the thought is, coinciding at rst but developing over time along di erent tracks. There were probably, in fact, a whole bunch of ship-like items, one for each of 67

the persistence conditions between which the ordinary notion of ship is undecided (Yablo 2024).

No doubt

only some of these are in play in a particular conversation, depending on the participants’ interests and 68

responsibilities.

This is a matter of conversational pragmatics, though, not metaphysics. And so, a modus

vivendi beckons. Which ship-like items are out there is an objective matter. But which of them counts as ≪the ship originally in the harbor≫ depends sometimes on who is asking and why.

The same applies mutatis mutandis to words, especially the meaning-imbued variety. Why would talk of them be any less open-textured than talk of ships? ≪The word spelled a-d-u-l-t≫ is undecided on this view between a bunch of similar sounding words, all carving di erent paths through time and across modal

space. As with ships, certain of these are liable to stand out from the crowd, and not necessarily the same ones to all observers. Now, we did suggest above that the  word spelled a-d-u-l-t in the 1960s took it in stride when the cut-o

age dropped to 18. This should probably be reformulated now that we are contemplating the possibility of a pool of adult-composed words between which community usage was indeterminate. Some 1960s speakers may have been drawn to an in exible adult whose loyalty to the higher cut-o

was absolute; call it adult21.

Legal developments in the 1970s presented these speakers with a challenge, not unlike the challenge presented to marriaget (“t” for traditional) fans by the Supreme Court’s decision in Obergefell v. Hodges. A few diehards stuck, we can imagine, with adult21. They were the ones saying, “Not in my book, you’re not,” to their Vietnam-bound o spring in 1972. Among the non-diehards, some will have switched to adult18, while others, to avoid further humiliation perhaps, adopted the more exible adult that (we assume) everyone uses today. People often remark on the di

culty of deciding whether it changes its meaning to put unaccustomed

objects (18-year-olds, same-sex couples) into a word’s extension. This is to be expected on the present account. The answer to, “Did w su er a meaning-change between T and U?” depends on (i) which word then in use was constituted by w at T, and (ii) where if anywhere we nd that word still in use at U. p. 34

These are tricky issues, especially if there were several words in play at T between which ≪The word

constituted by w≫ is indeterminate. Uncertainty about where ≪The word originally constituted by w≫ is to

be found at a later time is akin to uncertainty about the location later of ≪The ship originally made of those planks≫.

15. The Candidate Pool Ameliorators on a common caricature are people who seek to drag words for political reasons o

their

natural and appointed path. That path is laid down, you say, by “a clear objective kind that is the most explanatorily useful” (Díaz-León 2020a: 185)? It doesn’t matter; they are not deterred. All one needs is to come up with “moral and political considerations that trump” these objective advantages, and then the verdict is clear: We ought to change the meaning of the term, all things considered (Díaz-León 2020a: 185) Changing the meaning corresponds in our system to substituting a new word, with an improved meaning N , for the one that w used to constitute. This is, as Díaz-León says, “the most familiar version of an ameliorative project.” The familiar version needs rethinking. There may well be no such thing as the w best answering to the description ≪Word constituted at T by w≫, as we’ve agreed there is no such thing as the ship-like object

best answering to the description ≪Ship constituted at T by these planks≫. What we have in both cases is a pool of candidate referents whose fortunes are liable to rise or fall over time.

Say, for instance, that s2, the matter-based candidate, is damaged in a re. “The” original ship’s owner makes a claim for repair costs, based on a policy issued at T for his ship. The insurers reply that “his ship” is s1, not s2. They point out that s1 is the ship he has been leasing out and taking payment on, and that the matter-based ship s2 was put together by archaeologists without his knowledge. One can imagine arguments on the other side as well, with one side or the other prevailing. Fruitfulness considerations are going to play a role here; it’s important for instance not to lay down a bad precedent. As with ships, so with words. Reasons may emerge at U for seeing in w1 a happier construal of w than w2. (w2 was, on re ection, an unfortunate choice of word.) The pool of candidate referents for ≪Word constituted at T by w≫ accordingly shrinks, and w’s meaning grows clearer. T-speakers are understood at U to be using

one or another of (or a superposition of) the words still in the pool at U; shrinking the pool therefore makes it less indeterminate what they were saying. The ameliorator is usually portrayed, we said, as wanting to “redirect” w; imposing a new meaning is pulling it o

the course set by the previous meaning. This is not the right way to look at things, if w was

undecided between words not all meaning the same. What the critic calls “pushing w in a new direction” p. 35

might better be seen as opting for words that

were going in the right direction already. There were words

meaning N (or something near enough) in the pool and words whose meanings con icted with N . Our decision is to side with the rst class of words against the second. The words that con icted with N are no longer in the running, as we see it, for the role of ≪Word composed by w at T≫.

16. Pool Rules Hang on, you say. Why think that anything in the pool was going in the right direction already? Especially when it has yet to be explained how words make their way into the pool in the rst place. Consider our shared grasp of adult in 1970, before the age of majority dropped to 18. The guidance it provided in ordinary cases, the sort that speakers might realistically expect to encounter, did not extend to remoter scenarios (cf., Waissman on the open texture of language). Open texture was modeled as adult’s potentially constituting any one of a number of interpreted words all spelled and pronounced the same. Semantic stability would seem to require, on this model, that at least one of these adults was open to applying, under the right conditions, to 18-year-olds. Were there adults of this sort in the pool? One certainly hopes so; it seems more than plausible. But we would like to have more to operate on here than just hunches. 69

Imagine a Bush-era legal theorist

“discovering” in the pre-9/11 pool for torture a word that did not apply

to waterboarding, ngernail extraction, or branding with a red-hot poker. The pool contained in particular a word—torturey—applying only to forms of treatment liable to result in organ failure. Commentators saw this as opportunistic nonsense. Postulate such a word if you like. But don’t pretend it was a reading already in the running when you suddenly developed a need for it. What is the di erence, exactly, with the previous case? Why is the John Yoo line on torture more ridiculous than the standard 1970s line on adult? The ship analogy helps us to answer this. Let it be the form-based ship s1 that catches re this time. The owner seems in for a big loss (insurance was never purchased) until it occurs to him to argue that although “his ship” started out where s1 was, it jumped before s1 caught re to the town square—taking up residence there in the original planks, the ones also by then constituting s2. “His ship” s3 is a gerrymandered object that was coincident with s1 before aligning itself with s2. He would be coming by for it soon … unless the mayor wanted to make a deal.

The truth is that there is no such ship as this scoundrel claims to (still) own. There may be a kooky alwaysship-shaped object with the history he posits. But it does not count as a ship on any sane construal of the word. How we know this is not easy to say, but part of the answer is Davidsonian charity. An interpretation that allowed for discontinuous, uncaused, relocation would be hugely uncharitable to past usage. Only a fool would attempt to track ships perceptually, predict their future whereabouts etc., if ships were the kind of p. 36

thing to jump discontinuously about for no apparent reason. Between

an interpretation that makes us out

to be fools, and one that makes our conduct comprehensible, it is clear enough which to pick. This is the problem with Yoo’s (hence, the Bush administration’s) understanding of torture as the in iction of severe pain—where pain to be severe “must necessarily be associated with “death, organ failure, or 70

serious impairment of body functions.””

A lot of the most obvious-seeming pre-9/11 applications—to

mock execution, branding with a hot poker, and the like—would be incomprehensible if that’s what was meant. People knew that branding did not cause organ failure, yet went ahead and called it torture anyway. 71

Yoo misjudged the contents of the pool.

One more example before we get to our conjecture about how words gain membership in the pool. Girl supposedly came to constitute a new word (meaning FEMALE CHILD ) around 1530. Why do people believe this? Could one not equally say that girls had had to be female all along? One could try, but an explanation would then be needed of why pre-1530 speakers did not withhold the term from male children. Explanations can be imagined, just barely. Maybe they never encountered any male children, because the boys had died or gone into hiding. Maybe they did encounter them, but unwittingly; boys back then went around in disguise. That nothing like this is ever contemplated suggests that pre-1530s speakers would have been, not just mistaken, but inexplicably so, had it been girlf (meaning FEMALE CHILD ) they were using. A pattern is emerging. What keeps torturey out of the pre-9/11 pool for torture? Respect for earlier usage. Reading torturey into pre-9/11 uses of torture makes a mockery of the speakers involved by foisting on them huge numbers of unaccountable mistakes. What keeps girlf out of the pre-1530 pool is similar. The girlf reading makes pre-1530 speakers into fools given to wanton over-application of girl. The idea more generally is that w should be counted at U into the candidate pool for w-at-T if and only if it would not (at U) be uncharitable to T-speakers to read w into their use of w. This reinforces the claim above that normative considerations can bear on “what the actual [meaning] of a certain term is” (Díaz-León, 2020a: 185; emphasis added). A word that people should have been using looks like one that can charitably be ascribed to them. A w that can charitably be ascribed to T-speakers belongs in the candidate pool for w-at-T. A word in that pool is one that can, with no change in its meaning, be the one that w expresses today.

17. Clarifying Charity Interpretation could not get o

the ground, we are told, if there were not a presumption against assigning

false-making meanings to sentences sincerely held true, or normatively perverse meanings to sentences preferred true: p. 37

We could not begin to decode a man’s sayings if we could not make out his attitudes towards his sentences, such as holding, wishing, or wanting them to be true. Beginning from these attitudes, we must work out a theory of what he means, thus simultaneously giving content to his attitudes and to his words. In our need to make him make sense, we will try for a theory that nds him consistent, a believer of truths, and a lover of the good.

(Davidson 2013) This may seem to put charity at odds with the kind of amelioration we’ve been discussing—with “new” meanings read back into past utterances. Mark Richard, drawing on the work of Jill Hasday, gives the example of rape: The nineteenth-century American legal understanding of rape—and surely its dominant social understanding for at least the rst two-thirds of the century—was that it occurred when a man had intercourse with a woman who was not his wife without her consent. The nineteenth-century feminist movement in America began arguing publicly in favor of the idea that a woman “has a right to her own person”—and thus has a right to refuse her husband’s demands for sex—in the mid 1850’s. (Hasday 2000) The legal landscape was, reformers’ e orts notwithstanding, slow to change. The Model Penal Code of 1962 nds a man guilty of rape if he has unwanted sexual intercourse “with a female not his wife.” Not until the 72

1990s was marital rape securely established in the United States as a legal category.

(The dominant social

understanding of rape got there earlier, it seems, but we will not worry too much about the social/legal 73

distinction here. ) Did the countenancing of rape within marriage require rape to take on a new meaning? Charity as it is sometimes represented suggests the meaning did have to change. It was common knowledge, before reformers began their campaign, that in agreeing to marry one gave irrevocable consent to having sex whenever one’s partner requested it—an idea that, as is well known, traces back to the British jurist Matthew Hale, who wrote that marital rape was impossible “for by their mutual matrimonial consent and contract the wife hath given herself up in this kind [i.e., sexually] unto her husband, which she cannot retract.” 74

(Richard 2020: 366)

Think how uncharitable it would be to construe pre-reform speakers as just mistaken about all this. If they were right in rejecting the possibility of “rape within marriage,” and we are right in accepting it (as a category with instances going back centuries), then pre-reform speakers must have meant a di erent thing by rape. More carefully, the word must earlier have constituted a di erent word, not synonymous with the one in use hereabouts today. To assign a false-making meaning would indeed be uncharitable if misstatements—the holding-true of falsehoods—were what interpreters had to minimize. Really though, it is inexplicable, unforgiveable misstatements. We should prefer the interpretation that makes the utterance explainable … it is better to attribute to [the subject] an explicable falsehood than a mysterious truth. (Grandy, 1973: 445) p. 38

Whether marital-rape denialists were inexplicably, unforgivably mistaken is not obvious. Certainly, they were holding-true falsehoods, if their word is the same as ours. But charity was never intended to vindicate everything people say. The Greeks were holding-true falsehoods when they spoke of stars as apertures in the dome of heaven, as 75

were the Babylonians, when they spoke of Hesperus as a star.

Charity is ne with this, provided one can

see why they would in this instance have found the falsehoods plausible. A vindicating interpretation is in fact uncharitable, where the clues speakers have to go on are misleading. For it saddles them with indefensible attitudes. To insist against the evidence on (what turns out to be) the truth is unreasonable. 76

This is not how a charitable interpreter wants to make her subjects look.

What are the lessons for rape? If the case for meaning change turns on the inexplicability of marital-rapeblindness, then it is not very strong. How hard is it really to see why nineteenth-century speakers would consider y un-rape-able by x, when to them, y is x’s property, y has consented for the term of the marriage to sex on demand with x, etc.? Someone with hugely false beliefs (as we assume these are) from which unrape-ability follows would be expected to have a blind spot here. Since their conduct makes sense even if they are using the same  rape  as us, it is not uncharitable so to interpret them. One could argue that charity requires such an interpretation. Otherwise we lose our grip on what precisely was wrong in their day with the behavior that came to be known as marital rape. Imagine someone blurting out, “Stop, this is rape!” during a sexual assault by her husband. She may have been in some sense misusing the word. She would not 77

have been understood, probably even by herself. But it is hard to hear the statement as false.

18. Extending Charity An interpretation that makes T-speakers badly wrong may be charitable nonetheless, provided we can explain the misstatements away as natural in the circumstances. The usual sort of explanation points to a 78

prior “error or oversight”

from which the misstatement ows. People didn’t realize back then that whales

were biologically akin to elephants, or that women were morally akin to men. This is what led them to apply terms like sh and property to things that turned out not to be properly so called. But not all of the statements an interpreter might want to forgive can be traced back in this way to prior miscues. Take, for instance, Newton and Leibniz on instantaneous velocity. A lot of what they said on this topic was ba

79

ing,

and some of it (we are told) was at out wrong. They made mistakes—held-true

falsehoods—that could have been avoided had they been working with epsilons and deltas. But nonadoption of the ε-δ analysis was not an oversight on their part. It didn’t exist in their day; it might not have been comprehensible. Some would argue that it was not determinately the case back then that the ε-δ p. 39

analysis would come to be seen as capturing the meaning. The ratio

analysis might have won out, had the

continuum come to be conceived another way—as a “non-Archimedean” manifold replete with 80

in nitesimals.

Similar remarks apply to Wilson’s Druids. Inability to foresee the arrival of a previously undreamt of winged metallic object hardly counts as a slip or oversight. Still less did the Druids overlook something in failing to realize that the object would rst be encountered in the air rather than on the ground. Their mistakes were innocent, to be sure. An interpretation that represents them as holding-true falsehoods at T (e.g., ≪Burds

are solid rather than hollow≫) is not uncharitable. But what excuses the falsehoods is not that T-speakers had been led astray. Rather, the meaning was still in some sense under construction.

A third example involves measure-words: second, meter, kilogram, etc. These do not change in meaning whenever science comes up with better ways of measuring things. The obvious way to keep the meanings 81

aligned is for the new and improved standard to be read back into earlier uses.

This admittedly runs the

risk of falsifying statements once held true, like ≪This has always weighed a kilogram≫, said by Pierre in 2010, when a kilogram “was” the mass of a metal cylinder at the International Bureau of Weights and 82

Measures in Sèvres.

A charitable interpreter will take pains not to hold such misstatements against Pierre.

But again, charity will be hard to arrange if to receive it, Pierre has to be represented as explicably mistaken.

83

Caesium clocks are highly precise. The de nition they enable of second, adopted in 1967,

represented a

huge advance on the ones employed earlier: Before 1960, the unit of time the second, was de ned as the fraction 1/86400 of the mean solar day. The exact de nition of “mean solar day” was left to astronomers. Measurements showed that irregularities in the rotation of the Earth made this an unsatisfactory de nition. In order to de ne th

the unit of time more precisely, the 11 CGPM (1960) adopted a de nition given by the 84

International Astronomical Union based on the tropical year 1900.

So good is the caesium de nition that we are strongly tempted to read it back into earlier second-talk. One could, it’s true, keep the old de nition(s) around for old-utterance-interpretation purposes. Tokens of ≪The process takes fteen seconds≫ would be interpreted according to the de nition experts were

advocating for when they were produced. This seems a funny policy, though, given that speakers (i) rarely know what measurement experts are saying, and (ii) would be just as happy deferring to experts’ considered judgments as to whatever the best current story was at the moment in question. No doubt some old utterances need to be taken with a grain of salt, since nobody knew back then which way the de nitions were trending. But that is just to say that pre-1967 misstatements should be looked on charitably. The problem, again, is to see under what face-saving “guise” charity is to be extended. It is not as if the 1967 meaning N better captured what second-talk was driving at the whole time. There was an element of happenstance, for one thing, in science hitting on caesium clocks before, say, mercury-ion clocks. Also, the p. 40

talk was driving at N only if it

was NOT driving at the N * obtained by substituting 9,192,631,771 periods

for 9,192,631,770. That second had 9,192,631,770 in its sights pre-1967 rather than 9,192,631,771 would 85

take some arguing.

What primarily gets in the way, however, of looking charitably on pre-1967

misstatements is that we don’t know how to explain them away. All these examples point in the same direction. A distinction is needed between two sorts of non-culpable 86

misstatement.

Some misstatements are explainable away in terms of prior error or oversight. Others

merely re ect that certain semantic opportunities had yet to make themselves available. There are better ways of organizing the use of w—not that anyone at time T would have known of them—adopting which would have saved T-speakers from their mistakes. The term we’ll use is missed-opportunity misstatements, thinking here mainly of opportunities that a) had not yet arisen at T, like that of de ning second in terms of the radiation emitted by microwaved caesium atoms; or b) had not yet at T achieved the status of opportunities—it was not until the late 1960s that an adult in exible as to cut-o

age became untenable.

Falling, arguably, into both these categories is the way of organizing rate-of-change talk proposed by Weierstrass in 1861 (with his ε-δ de nition of the derivative). Weierstrass’s de nition didn’t exist when Newton wrote Principia (T = 1687). And it was opportune (more so than Robinson’s de nition), thanks in part to how mathematicians came to conceive the continuum. Speakers who are pulled o

course by their errors/oversights are not the only ones deserving interpretive

charity. A modicum of charity should be extended to “victims” of missed opportunity too. One probably does not want to speak in the missed-opportunity case of explaining the misstatements away. (≪Newton went wrong on account of not minding his εs and δs≫ has some of the same absurdity as ≪Newton’s head

hurt due to the shortage of Tylenol in his brain≫.) Even if Newton’s misstatements are not explained by the 87

unavailability in 1687 of ε-δ methods,

the possibility remains that they’re excusable on that basis. That

Newton didn’t have εs and δs at his disposal does seem to throw his stumbles into some kind of saving perspective.

Consider now Pierre, who organizes his use of kilogram around that metal cylinder in Sèvres. He is wrong when he endorses ≪The cylinder has always weighed a kilogram≫. It shows him no disrespect to say so,

given that he was working with an inadequate and soon to be superseded de nition. Pierre had no intention, when he gave M (something to do with the cylinder) as the meaning, of swearing absolute allegiance to it. He would presumably have been aware that M had problems, like reliance on a hard to access physical prototype, that would need at some point to be addressed. He could hardly help but be aware as a general matter that M could be improved upon somehow or other. Pierre would likely have jumped at the chance of 88

associating himself with N (something to do with Planck’s constant) had the opportunity been provided. Mistakes that an embrace of N would have prevented are understandable, which is not to say explicable.

p. 41

19. The Course Correction Model Science as Neurath tells it is like a boat that sailors are forced to repair out at sea while remaining a oat in it. A better analogy, since science is an activity, might be the journey the sailors are taking. This preserves the holism while adding an intentional element. The course to be set today depends on the boat’s present location, which is a function of the course(s) set earlier. Also important, of course, is the intended destination. That, too, is up for review, as the travelers get clearer on what they are seeking and where they are likely to nd it. This connects up with amelioration as follows: Meanings are theoretical entities, postulated to make sense, inter alia, of linguistic communication. Hypotheses about what a word means (and has meant) are defeasible, like hypotheses about anything else. This being so, a lot of what gets called “giving the meaning” is better conceived as presenting a hypothesis about the meaning. And a lot of what gets called “changing the meaning” is better conceived as revising that hypothesis. Clearly, the meanings of words are frequently (in some sense) reimagined. The reimagining need not involve the imposition of new meanings. 89

It may be our theory of the meaning that changes.

Why, though, would our theories evolve in an ameliorative direction, with hypotheses attributing productive (fruitful) meanings N winning out over those attributing counterproductive meanings M ? A couple of answers have been suggested to this (both variations on themes sounded long ago by Lewis) to which we’ll soon add one more. First: The role reference magnetism played for Lewis, reference electromagnetism plays for us. That N is the more fruitful candidate makes it electromagnetically more attractive, and hence more eligible for the meaning-of-w role. So much the better for its chances of winning the role. N standing a better chance of winning the meaning role gives ≪w means N ≫ an advantage qua semantic hypothesis over ≪w means M ≫.

The second answer appeals to charity. N is not the meaning unless it’s charitably ascribable to w-users. How could it not be a plus for N, in this respect, that it is a better, more useful, thing for w to mean? Speakers are lovers of the good, remember. The good does not quickly reveal itself, however, which gives the interpreter an opening. She can say: Yes, they accepted a falsehood. But they were doing the best they could with the tools available. They were in thrall at the time to M, being unacquainted with N . What enables this sort of missed-opportunity framing is M ’s unproductiveness relative to N . This is the second reason our words might be expected, other things equal, to gravitate toward productive meanings. Among the hurdles an interpretive hypothesis has to clear are some (to do with charity) that productive N s clear more easily. These “theoretical” advantages cannot be the whole story. If the ameliorist aims only to improve our theory of what w means, why do we nd her so often agitating for a certain meaning? Rather than trying to bring it about that w means N , shouldn’t she limit herself to observations supporting the N hypothesis?

Ameliorists do make observations. They observed in connection with rape that wives are not their husbands’ property. They observed that the length of a certain day in 1900 could not do much to coordinate the use p. 42

decades later of second. They

observed that 18-year-olds risking their lives overseas did not count for

voting purposes as adults. They observed as recently as 1990 that the standard kilogram in Sèvres was hard for physicists to get ahold of, heavier on some occasions than others, etc. So yes, the ameliorist is an observer. Why should this prevent her from trying to in uence events? No one said that she was a disinterested observer. Bringing out awkward facts about M is, or can be, a way of killing its candidacy for the position of what w means. Highlighting N ’s advantages is, or can be, a way of boosting its candidacy. If there’s a contradiction here, then it’s shared by mobilization e orts generally. Social organizers try, by stressing an outcome’s potential fruitfulness, to hasten the day when that potential is realized. Amelioration can be seen as just a special sort of mobilization e ort—where the sought after outcome is that semantic indeterminacy be resolved in favor of these candidates rather than those. This in a nutshell is the course-correction model. Amelioration done right is (on the CC model) selfful lling. N really is, if all goes well, the meaning rather than M . N is the meaning because it’s more eligible; because it exerts a stronger electromagnetic pull; because it is more fruitful/productive, perhaps, in part, because we’ve laid the ground for this outcome by popularizing N and setting out its advantages. The scienti c community arranged at a certain point for second to mean N (something to do with caesium atoms) rather than M (a fraction of the length of a certain day in 1900). They did it by making N more eligible, speci cally, more productive. To be productive, a meaning has to be available for use, which N initially wasn’t. (M had the eligibility advantage to begin with.) N ’s prospects took a turn for the better when 90

atomic clocks were developed in the mid-20th-century. From an early write-up in the New York Times:

Blueprints for the most accurate clock in the universe, tuning in on radio frequencies in the hearts of atoms and thus beating in harmony with the “cosmic pendulum,” were outlined yesterday at … Columbia University by Prof. I. I. Rabi, who delivered the Richtmyer Memorial Lecture under the auspices of the American Association of Physics Teachers. (34) From this it seems that N had only to await the arrival of new technology. But there could have been a number of standards N 1, N 2, … under development, each improving on M but none superior overall to the 91

others. Part of the task in that case would be to build up a constituency for caesium in particular. 92

cognoscenti have rallied around N , it becomes, for the rest of us, an o er we can’t refuse.

Once the

Recall that the

o er includes a pardon for any mistakes we might have been drawn into by that embarrassing early crush on M.

20. Debunking and Charity 93

Back now to the revisionary-sounding de nition of woman oated by Haslanger. p. 43

face of it change the meaning, or (as we prefer to think of it)

Adopting it would on the

the word. Appearances can be deceiving,

though, as with second and adult. What looks, at rst, like proposing a novel meaning is better understood in some cases as proposing a novel hypothesis about the existing meaning. Why would these be so di

cult to distinguish? There’s rst an act-object ambiguity. “A hypothesis” could

be either an act of hypothesizing or the proposition that’s hypothesized. “Proposing” a hypothesis could, second, be either proposing the act, or engaging in it, that is, advancing the proposition as true. To top it all o , proposing N as the meaning has the potential to boost its candidacy for the role, conceivably making it true that N is (and was) the meaning—in which case N was not really new after all. One should be careful

then about calling the proposed meaning novel. It may or may not be novel, depending on whether the proposal succeeds. When second began to be explained in terms of atomic clocks, that was a re nement of the previous de nition. Earlier judgments were not reversed; we were expected only to bear in mind their greater uncertainty. The rethinking of adult-hood in the 1970s did not reverse earlier judgments either. New legislation was passed, and the popular understanding evolved, but the classi catory e ects were mostly 94

forward-looking.

Haslanger’s de nition of woman seems di erent in this respect. (Amazon warrior-

queens are dropped retroactively from the roster.) The transition if it took hold would not be classi catorily conservative … We do not hold constant the classi cation the concept e ects in practice while ipping its emotive valence. (Richard 2020: 374) True, but what happened with rape was not classi catorily conservative either. Assaults that fell initially under “the marriage exemption” came to be seen in retrospect as clear cases of rape. Richard might reply that the switch was conservative in a di erent sense: The purposes the feminists in e ect were assigning to the concept of rape would likely appear to the target audience to be more or less continuous with the cognitive and social purposes the concept already served, even at the beginning of the nineteenth century… A consequence is that it would not feel like false advertising—it would not have been false advertising—for feminists to represent themselves as pointing out that the best way to understand the existing practice of labeling and prosecuting things as rape is as a practice whose rationale … is to condemn sexual violation of women no matter who the agent is. (Richard 2020: 373). This sounds like an appeal to charity. Interpreters are entitled to correct for subjects’ shaky grasp and inconsistent application of the rationale for their own labeling practices. The di erence with woman is that there was an appetite in the target audience for this sort of correction in 1870; that is why it did not feel like 95

false advertising. Whereas there is not much appetite today for the corresponding “correction” re woman. But, should speakers get to decide for themselves what counts as a charitable intervention? Debunking

accounts of ideologically freighted notions are supposed to be uncomfortable. They come packaged, indeed, p. 44

with an account of their own

o -puttingness. The assumptions they target have been implanted so

deeply as to seem obvious, in order that unjust social arrangements take on an appearance of naturalness and inevitability. That patriarchy has done such a number on us looks, in principle, like a wonderful account of why we would rather not see women as ipso facto subordinated. It’s the same sort of account as might once have been given of people’s resistance to the idea of royalty as a social/political status rather than something divinely 96

ordained. Or of the insistence in some quarters that civil “marriage” is not marriage properly so called.

The extension to same-sex marriages seems in this light like no big deal. We have, after all, a story to tell about why earlier generations got it wrong; can obliviousness to the possibility of civil marriage be explained away half as well? Continuing on the theme of giving unjust social arrangements the appearance of naturalness and inevitability, gender status bears comparison with class and caste status. There is a tendency, noted long ago by Russell, to romanticize the downtrodden as if in compensation for their ill treatment:

If virtue is the greatest of goods, and if subjection makes people virtuous, it is kind to refuse them power, since it would destroy their virtue. If it is di

cult for a rich man to enter the kingdom of

heaven, it is a noble act on his part to keep his wealth and so imperil his eternal bliss for the bene t of his poorer brethren. It was a ne self-sacri ce on the part of men to relieve women of the dirty work of politics. And so on. (Russell 2009: 63) Russell’s focus is on subject nations, “noble savages,” the proletariat, women, and children. He might also have mentioned caste. Indian epic poetry tells of worthy and respected outcastes (“Untouchables,” Dalits). The Ramayana is traditionally credited to a Dalit (Valmiki). Veda Vyasa, author of the Mahabharata, supposedly had a Dalit mother. Gandhi wanted Hindus to reconceive this group as Harijans or “Children of 97

God.”

Their occasional earlier veneration notwithstanding, few would insist today on keeping subordination out of the de nition of outcaste, or that of Shudra, the lowest of the macro-castes (Dumont 1980). A charge of conceptual rebranding could be leveled in this case, too. It is not uncommon for caste’s hierarchical 98

character to be denied or downplayed.

The law of Varna, a precursor to the caste system, was defended by 99

Gandhi as (in its heart of hearts) egalitarian:

There is no calling too low and none too high. All are good, lawful and absolutely equal in status. The callings of a Brahmin—spiritual teacher—and a scavenger are equal, and their due performance carries equal merit before God, and at one time seems to have carried identical reward before man. (Ambedkar 2014) Hardly anyone talks this way today, not that no one does. Denialism about caste’s hierarchical character is a persistent strain in Hindu thought. It has not always seemed absurd and continues to have its advocates. p. 45

Denialism about gender’s hierarchical side has its advocates too. This is surprising given how recently it was that women acquired voting rights—to say nothing of rights over their own bodies (which are suddenly again up for discussion). Slaves of a male persuasion won the franchise soon after the Civil War. Women had to wait another ve decades, and then ve more for rights of bodily self-possession. This could be just a string of bad luck. But it is not absurd to think that women are subordinated per se. Anyone who claims to 100

know a priori that such a view distorts the concept has not been paying attention.

The question is not to

be settled from the armchair. Quine once asked, in his attack on analyticity, what the in-principle di erence was supposed to be between reforming logic “as a means of simplifying quantum mechanics,” and “the shift whereby Kepler superseded Ptolemy, or Einstein Newton, or Darwin Aristotle?” (Quine 1951: 40). One might ask in a similar vein what the di erence is supposed to be between Elizabeth Cady Stanton superseding Matthew Hale on the meaning of rape and people like Ambedkar superseding people like Gandhi on the meaning of caste and gender 101

terms.

21. Review and Conclusion Ameliorists by de nition are people who advocate for N as an opportune, fruitful, productive thing for w to mean. Are they advocating for a change of meaning? That is certainly how it looks—or has looked to philosophers who have considered the question. But while meaning change can be the goal, we don’t believe that it must be. Semantic continuity seems out of the question because people are operating with over-rigid assumptions about meaning. To come at it from the other direction, continuity begins to seem less ridiculous if we keep certain semantic possibilities in mind. These are listed below in something like their order of appearance in the chapter. FACTUAL ENTANGLEMENT: Meanings are assumed to be prior to the facts we use them to state. This runs contrary, though, to our actual practices of meaning ascription. We do in practice revise our views of what 102

was meant in light of new empirical ndings,

103

and even non-empirical ndings.

PRESUPPOSITION FAILURE: Meanings are assumed to be presuppositionless. But our ways at least of specifying them are apt to rest on ideas about the properties of the meaning thus identi ed. What to say when these ideas turn out to be incorrect? Maybe, as in the ill-chosen prototype case, that we were wrong about what the meaning was (section 6). HISTORICITY, EMBODIMENT: Meanings are assumed to be free- oating and ahistorical. But our ways of identifying them over time run essentially through the signs w whose meanings they are: a signi er retains its meaning for precisely as long as it continues to constitute the same sign. Anchoring meanings in signs opens the door to variation in what a meaning requires over time, e.g., adult retained its meaning even as p. 46

that meaning became less demanding. The door is opened, too, to retroactively

operating in uences on

words already in use. If the Shroud of Turin could exist before its essential properties (serving as Christ’s burial shroud) were in place, the term rape can predate the events given which it’s the term we’ve been operating with (sections 7, 8). ELIGIBILITY: Meanings are thought to be a function of use. But theorists going back at least to Frege have 104

questioned this.

Contents are arranged hierarchically, for Frege. Some are especially well suited to play 105

the sense role, because they help with (for instance) the codi cation of valid inference patterns.

This

might be seen as a precursor to Lewisian reference magnetism. Use underdetermines meaning on either view. Eligibility comes in to take up the slack (section 9). FRUITFUL CONCEPTS : Fruitfulness considerations may seem to bear only on words’ future semantic properties—on what w shall mean going forward. But they can bear, too, on what w means now, whether or not anyone is in a position to know it. (As seen with measure terms in science, mathematical terms like limit, and social/legal terms  like  married  and  adult.). ELECTROMAGNETISM: Magnetism is a function solely of god’s-eye-view naturalness. Fruitfulness looks as well to “electric” factors such as t with our cognitive makeup and practical projects. Words are drawn to fruitful meanings because fruitfulness makes N more electromagnetically attractive (sections 10–12). SEMANTIC INDETERMINACY: A further resource here is the metaphysical doctrine known as pluralism or many-thingism. Just as it can be indeterminate which of various ship-like items si is “the ship composed at T of thus and such planks, so and so arranged,” it can be indeterminate which of various signs wi is “the word composed at T of thus and such letters, so and so arranged.” An ameliorist not bent on meaning change might be attempting only to reduce the indeterminacy. She might be trying to remove words with the wrong kind of meaning from the candidate pool: the words that w potentially expresses. No new meanings are imposed if a w meaning N was already in the pool (sections 13, 14).

EXPLICABLE MISSTATEMENTS : A word does not belong in the w-pool if it is too far out of sync with people’s use of w at the relevant time T. wi is out of sync with people’s use of w at T if it is not charitably reconcilable with their verbal dispositions at T—if too many sentences that come out false, when w is read as wi, are inexplicably held true at T. Friends of wi are equipped in many cases to explain the misstatements away. The explanation would cite T-speakers’ denial of, or insensitivity to, the facts whatever they are that confer on wi its advantages (e.g., in the case of rape, certain moral equivalency facts). Charity favors the ascription of useful meanings, because their usefulness unlocks resources for explaining misstatements away (section 16). UNDERSTANDABLE MISSTATEMENTS : Explaining them away is not the only way of making sense of misstatements—of why a falsehood would be treated as true. Speakers base their acceptance decisions on heuristics that can lead them astray. There need be nothing irresponsible about this. Heuristics may be all we have to go on at T; there is nothing else for the meaning to be. They take on the appearance of “mere 106

guides” only when something better comes along (section 17). p. 47

This points to a second way charity favors

the ascription of fruitful meanings. The more counterproductive M

is relative to N , the easier it becomes

to cast M as a heuristic issuing sometimes in false conclusions. Insofar as the misstatements are forgivable, proposing N as the meaning does not automatically make T-speakers out to be fools. DEBUNKING ACCOUNTS : Some of the more radical-sounding reimaginings out there, like Haslanger’s de nition of woman, have seemed so at odds with existing usage that the meaning would have to change The thought in Davidsonian terms would be that an interpreter who attempted to read N into present-day usage would saddle us with mistakes not so easy to forgive. This might be right in the end. But it’s not a slam dunk if M has the backing of ideology. You’d expect in that case that speakers would see no real alternative to statements that, when the blinders are o , look rather more debatable. Gandhi thought that Shudras only happened to be subordinated; they were not “lower” by nature. This is just what we say today about women. If ideology could lead him (a moral genius!) o to say it could not lead us o

track about the meaning of caste terms, who’s

track about the meanings of man and woman?

Notes 1.

Special cases aside, such as indirect traction via an agentʼs desire to bring good outcomes about.

2.

“Observe, the nose is formed for spectacles” (Voltaire, Candide).

3.

Williamson (2007); Williams (2018).

4.

Cappelen (2018); Koslow (2022).

5.

You should, as the saying goes, dance with the one that brung you.

6.

Heteronymous reasoning is “reasoning directed from the outside, by an authority that is merely assumed or imposed” (G. Williams 2018).

7.

Linguists have their own notion of amelioration, to do with strengthening positive associations and/or weakening negative ones. Amelioration in the linguistʼs sense is the opposite of pejoration (Traugott et al. 2001: 55). A meaning adjustment that is pejorating to the linguist would be ameliorative in the philosopherʼs sense, if we stand to gain by bringing the unsightly details into view.

8.

Sometimes the superiority has an aesthetic, or anyway, a non-cognitive, aspect. Wittgenstein (1980): “Esperanto. The feeling of disgust we get if we utter an invented word with invented derivative syllables. The word is cold, lacking in associations, and yet it plays at being language” (52). “A language which has not ʻgrown organically,ʼ ” writes Carnap, seemed to Wittgenstein “not only useless but despicable” (Nyíri 1992: 85). (Carnap taught himself Esperanto as a child.)

p. 48

9.

Chomsky was a Junior Fellow at Harvard when Austin gave How to Do Things with Words as his William James Lectures. It would be interesting to know whether Skinner (who was at Harvard too) attended the lectures.

10.

See for more along these lines “The Design of Cultures” (Skinner 1961).

11.

Skinner (1960) attempts to explain in novelistic form how the promotion of survival-conducive values might play out in practice.

12.

Pinker says of the strings ptak thale hlad pla sram mgla vlas flutch dnom rtut toasp nyip that “any native speaker recognizes that thale, pla , and flutch are not English words but could be, whereas the remaining ones are not English words and could not be” (1995: 173).

13.

Similar points apply to proper names, which are on some views not words at all. Kaplan gives the example of someone wanting to call the baby Tkbtbtkbtktb: “No way! It wouldnʼt be English… There are a lot of names, you have a lot of choices, but you canʼt name it that. My point here is that names, like other words, must subscribe to certain regularities. Certain of these regularities have to do with admissible sound and spelling patterns” (1990: 113). Better to say with Chomsky and Halle that certain patterns are dispreferred.

14.

See (Hinton et al. 2006). Austin took an interest in these constraints: “At the time of his death, he was developing a semantic theory based on sound symbolism, using the English gl-words as data” (https://en.wikipedia.org/w/index.php? title=J._L._Austin&oldid=1169193136, accessed September 3, 2023). The gl- words supposedly share a phonestheme, defined as a submorphemic sound pattern lending itself to a certain sort of interpretation. The invited interpretation in this case is to do with light and vision (glare, glass, gleam, glimmer, glint, gloom, and glow). (Austin might have qualified as a Type-3 romanticist too, had he lived to finish his theory.)

15.

Pinker (2013); see also Goddard et al. (1994); Jackendo (2002); Pinker (1995); Johnson (2004); von Fintel and Matthewson (2008); Marchetto and Bonatti (2013); Gardenfors (2014); Machery (2014).

16.

Kovecses and Szabco (1996); Nunberg, Sag, and Wasow (1994); Egan (2008).

17.

“Dogs are mammals” speaks of dogs as a class, while “Dogs are making a mess on the lawn” says only that there are dogs behaving this way.

18.

Ogawa (2001); Haslanger (2011, 2014).

19.

Some may see this as an advantage of the analysis. Maybe woman ought to be understood as stage-level.

20.

Perhaps others are worth the trouble; who knows, when the trouble is rarely even acknowledged? See Machery (2021); Koslow (2022).

21.

Ponet (1556).

22.

The entailment ensures that one of us is wrong, if the other is right.

23.

I accept the part automatically by virtue of accepting P.

24.

Fine (2017); Yablo (2017); Abreu Zavaleta (2021, 2020).

25.

Cappelen (2018); Koslow (2022).

26.

Just as our dogʼs taking on new characteristics doesnʼt mean weʼve got a new dog.

27.

Jade, too, was in the market for a natural-kind-y status that it did not in the end achieve.

28.

Putnam elsewhere gives a more dramatic example where we lack even a false presupposition to point to: the pre2 relativistic definition of an objectʼs energy as ½ mv . (Compare the missed-opportunity cases discussed in section 16.)

29.

Suppose, for instance, that the orbital perturbations Neptune was called on to explain turned out not to be due to a planet at all (as occurred with Vulcan).

30.

“I also think, contrary to most recent theorists, that the reference of names is rarely or almost never fixed by means of description” (Kripke 2011a: 21).

p. 49

31.

Just as statues can; we might have a statue of Socrates coinciding for the moment with a (longer-lived) statue of a generic adult male. (See below for meanings vs. meaning-infused words.)

32.

Compare Thomasson (1999) on the persistence of novels and fictional characters; compare songs, poems, and ceremonies.

33.

Some might think itʼs enough that S 1 = S 2; this possibility is for our purposes best ignored.

34.

Note, we are talking so far about word-types; tokens will come in later.

35.

A real-life example of this might be dog in the Australian language Mbabaram sharing a meaning with our dog (the example is contested; thanks here to Manuel Gustavo Isaac). For more on words in all their flavors and varieties, see Hawthorne and Lepore (2011); Kaplan (2011); and Bromberger (2011).

36.

Especially if, as some speculate, light has been slowing down (Magueijo 2009). Note as well the relativity to our notion of second. See Scharp (2019) for the possibility that meter defined in caesium terms is incommensurable with “the minimal length” (the Planck length, say), which would seem to make the metric system essentially indeterminate.

37.

Similar issues arise, we shall see, for the meaning of second.

38.

Twenty-one was the original age of majority in the United States. It had been 21 (for male commoners, anyway) in England going back to the Magna Carta. It was not until 1970 that the English age of majority dropped to 18 (Family Law Reform Act of 1969). Majority is decided at the state level in the US; most states switched to 18 in the early 1970s. Mississippi seems to be the only remaining holdout.

39.

One can imagine di erent versions of this. T could be the time of utterance (contextualism), the time of assessment (relativism), or the ascribee-stageʼs position in time (subject-sensitive invariantism). The relativistic option is explored in Ball (2020a).

40.

Hacking (1985).

41.

On varieties of meaning change other than broadening/narrowing, see Traugott et al. (2001) and Koslow (2022).

42.

Haslanger (2020a, 2020b) distinguishes epistemic amelioration from semantic amelioration. The first aims at a better understanding of the content; the second aims at a better content. Amelioration of the kind discussed in this chapter belongs in a third category yet. It shares with epistemic amelioration the no-meaning-change feature, and with semantic amelioration the feature of taking a possibly unforeseeable course. This is not the place for a full discussion.

43.

When the content broadened to allow for rape within marriage.

44.

Mutatis mutandis for the probability of w meaning N , conditional on N being a suitable meaning.

45.

Importantly, Lewis has “objective” naturalness in mind rather than something cognitive.

46.

See Byrne (2003) for more on this, especially sec. 8, “Natural Properties and Kripkeʼs Wittgenstein.”

47.

Naturalness-centrism is critiqued from a di erent angle in Williams (2007) and Hawthorne (2007).

48.

As Lewis (1984) says himself, “Many philosophers would suggest at once that the saving constraint has to do with the causal chains that lead into the referrerʼs head from the external things that he refers to” (226). He worries that causal chains are too plentiful to do theoretical work here. (See the main text for Williamsonʼs response.)

49.

See Williams (2018) for an interesting further way of working normativity into metasemantics.

50.

Lewis doesnʼt disagree that causes matter. He thinks he can keep causal factors out of eligibility, however, by (re)filing them under use-consonance. He relies here on a form of descriptivism that reads causation into reference-fixing descriptions. The beliefs a charitable interpreter is called on to vindicate include some about wʼs causal role.

51.

Tappenden (2008b) discusses natural vs. fruitful definitions in mathematics, pointing out that fruitfulness can sometimes favor a disjunctive definition, which appears only in retrospect as natural.

52.

“We should make explicit a claim about naturalness that Lewis presupposes in much of his discussion: Necessity: Facts

about a propertyʼs degree of naturalness are non-contingent. The denial of Necessity would be quite alien to Lewisʼs thought, and would require rethinking many of the other components of the role” (Dorr and Hawthorne 2013).

p. 50

p. 51

53.

Egan (2004); Yablo (1999).

54.

The mathematical examples to come may seem at quite a distance from the amelioration literature as it has developed in the last couple of decades. But see “What Are We Talking About? The Semantics and Politics of Social Kinds” (Haslanger 2005): “In the history of logic and math, inquiry can seem to converge on an idea or concept that we seemed to have in mind all along, even though no one, even the best minds, could have explicated it. (Leibnizʼs early e orts to define the limit of a series is an example.) In such cases, it is plausible to maintain that certain experts were “grasping a definite sense, whilst also failing to grasp it “sharply.”” (50)

55.

See, for instance, Anderson (1995) on the question-driven-ness of inquiry and classification, especially sec. 4 on the Bureau of Labor Statisticsʼ 1994 redefinition of “unemployment.”

56.

“When we compare the explicandum Fish with the explicatum Piscis, we see that they do not even approximately coincide … What was their motive for … artificially constructing the new concept Piscis far remote from any concept in the prescientific language? The reason was that [they] realized the fact that the concept Piscis promised to be much more fruitful than any concept more like Fish. A scientific concept is the more fruitful the more it can be brought into connection with other concepts on the basis of observed facts” (Carnap 1962: 6). For discussion, see Dutilh Novaes and Reck (2017); Dutilh Novaes (2020a, 2020b).

57.

Peacocke (1998a, 1998b); and Higginbotham (1998). Leibniz may be another matter.

58.

Important here are Ebbs (2000); Schroeter and Schroeter (2015, 2020); Jackman (2020); Ball (2018, 2020a, 2020b); Ball and Huvenes (2022).

59.

“In the history of the calculus … there were two rival theories of the continuum. [One] was the Leibnizian theory: the Archimedean continuum extended to a non-Archimedean one by adding infinitesimals and infinitely large numbers. Leibnizʼs theory was the dominant one until the Weierstrassian revolution” (Lakatos 1978).

60.

“Although such Druid contentions may seem perverse or unpalatable, we shall find that we make precisely analogous declarations about the developmental history of our own language all the time!” (ibid., 551). We do make these declarations. Better surely for the declarations to come out correct.

61.

Yablo (1987). To write it mathematically, e is e .

62.

Or perhaps, w-qua-word-meaning-M .

63.

Code (1985).

64.

Likewise for dramatic overnight changes in pronunciation; the Great Vowel Shi played out over several hundred years.

65.

The judgments here are admittedly not crystal clear. But this is to be expected on the proposed analogy; for the persistence conditions of commonsense kinds are not clear either. How much reshaping and/or matter-replacement will a statue put up with? (A perspective on the problem is suggested in the next section.)

66.

Related to what Waismann calls the open texture of language (Makovec and Shapiro 2019).

67.

Compare the neo-Lockean claim that multiple person-type things came into being when you did: a human organism, a biological person (the human being), a forensic person, and so on.

68.

Debt-collectors tend to focus on the forensic person.

69.

John Yoo comes to mind (Goldberg 2005).

70.

“Torture Memos, Wikipedia (https://en.wikipedia.org/w/index.php?title=Torture_Memos&oldid=1160041059).

71.

K

The Yoo definition could have won out; the word spelled t-o-r-t-u-r-e in 2002 would in that case have been torturey. This would have involved a (numerical) change in meaning, we are saying, the original M giving way to a distinct N .

p. 52

72.

Even then, spousal rape was liable to be a sui generis category, not a version of the o ense already on the books.

73.

On Victorian conceptions (outside the law) of marital rape, see Bourke (2008).

74.

The Hale quote is from History of the Pleas of the Crown (1716).

75.

McGinn (1977).

76.

“But it would be more charitable to make allowances for the likelihood that Karlʼs circumstances—his life history of evidence and training, recounted in physical terms in our data base P—may have led him understandably into error. We should at least forbear from ascribing to Karl those of our beliefs and desires which, according to P and our notions of reason, he has been given no reason to share. We should even ascribe to him those errors which we think we would have made, or should have made, if our evidence and training had been like his” (Lewis 1974: 336).

77.

Satan in Society warns against the “first conjugal act” becoming “little else than legalized rape” (Cooke 1871: 146). Genesis and Ethics of Conjugal Love applies the term to spousal sexual violence generally, “notwithstanding its legal recognition by the State and the solemn sanction of the Supervising Church” (Davis 1874: 20). On Conjugal Happiness says of what transpires on the wedding night that it “o en amounts to nothing more or less than rape” (Loewenfeld 1913: 200). Quotes are from Bourke (2008).

78.

Crispin Wrightʼs phrase.

79.

Berkeley (2012).

80.

Archimedean: for all positive x and y, there is a natural number n such that x is less than n times y. The reals are Archimedean, the hyperreals arenʼt. Perhaps the continuum was always going to come out Archimedean; we assume for exampleʼs sake that it could have gone either way (Lakatos 1977). See Builes and Teitel (2020) for considerations potentially favoring the infinitesimal ratio analysis.

81.

With reminders, where necessary, that speakers may not have been party then to all the details.

82.

Not until 2019 did it receive a definition in fundamental physical terms.

83.

One second = 9,192,631,770 periods of the radiation corresponding to the transition between two hyperfine levels of the ground state of the caesium-133 atom.

84.

From the website of the International Committee for Weights and Measures (CIPM) (https://www.bipm.org/en/historysi/second).

85.

Maybe this can be argued; we donʼt know the history. Itʼs enough that the viability of the new definition does not depend on 9,192,631,770 being objectively better.

86.

A misstatement as usual is the holding-true of a sentence S that is, on the going interpretation, false.

87.

Make this stipulative if you like.

88.

Or, he would have jumped at the chance if heʼd had his head screwed on right; he should have jumped.

89.

Van Fraassen on measured quantities: “There is no presuppositionless starting point for coordination. We are, to adapt Neurathʼs metaphor, sailors engaged in a continuing construction, renovation, alteration, and repair of the ways in which we measure our ship while already at sea” (2008, 137). “The parameter that is measured is identified in the historical process by the envisioned eventual stable measuring practice, while it is di erently identified in retrospect by the theory that draws on that history for its credentials” (139).

90.

See also Essen and Perry (1955).

91.

Constituency-building is a job for organizations like the Bureau of Weights and Measures (in the US) and the General International Committee for Weights and Measures (or CIPM, in Europe). The CIPM o ers “Mise en Pratique” guidelines on how to phase new definitions in while maintaining continuity with their cruder predecessors. From the 2019 guidelines for the kilogram: “The mass values of the IPK [International Prototype of the Kilogram] and its six o icial copies are now determined experimentally … Subsequent changes to the mass of the IPK may have historical interest even though the

IPK no longer retains a special status or a dedicated role in this mise en pratique … By following the change in mass of the IPK over time, one may be able to ascertain its mass stability with respect to fundamental constants, which has long been a topic of conjecture.” 92.

The caesium definition had few real competitors. But there are several technologies in the running for “next big thing” a er caesium (Gill 2011).

93.

A woman =Df an individual regularly and for the most part observed or imagined to have certain bodily features presumed to be evidence of a femaleʼs biological role in reproduction, these features playing into the dominant ideology in such a way as to subordinate the relevant individuals.

94.

“A question of interest to older children arose in those states which reduced the age of majority to 18: what e ect does the statutory change have on preexisting support obligations which were defined in terms of majority rather than by stating that support was to continue to a specified age? Jurisdictions which have considered this have generally said that there will be no retroactive e ect. The courts hold that ʻmajorityʼ in those agreements means the age of majority at the time of the agreement.” (Downey 1981: 29).

95.

“One feels that the proposalʼs analysans is pretty much discontinuous with the analysandum on all relevant dimensions” (Richard 2020: 374).

96.

Civil marriage has still not achieved full legal recognition in Egypt, Syria, Jordan, United Arab Emirates, Saudi Arabia, Qatar, Yemen, Iran, Lebanon, Israel, Libya, Mauritania, and Indonesia (Wikipedia).

97.

The term is considered o ensive.

98.

“Hierarchy is at the heart of the ʻunthoughtʼ of modern ideology” (Dumont 1980: xvi).

99.

Gandhi thought Varna more defensible than caste and, at the same time, distanced himself from its main seeming advantage (non-hereditariness): “The law of Varna teaches us that we have each one of us to earn our bread by following the ancestral calling” (Ambedkar 2014; emphasis added).

100. Concepts like “chastity” and “knowing oneʼs place” have managed somehow to survive acknowledgment of their oppressive aspects. 101. Gandhiʼs views on gender are not easily summarized. But a traditionalist, sentimentalizing thread runs through them akin to what weʼve seen with caste. “Gandhiʼs statements supported the sanctity of the family and the home: he reiterated that a womanʼs real place was the household where she was the ʻqueenʼ ” (Bald 2003). For comparison with Ambedkar, see the introduction by Roy to Ambedkar (2014). p. 53

102. E.g., the discovery that the “meter” stick was longer than it appeared, and too long for the purpose. 103. See, for instance, Lakatos (2015); and various papers by Tappenden. “A common pattern in mathematics is the discovery of the proper definition of a word with an already established meaning” (Tappenden 2008a: 267), and “investigations into the structure of numbers discovered what prime numbers really are. The familiar school definition [b | a → (b = 1 or b = a)] only captures an accidental property; the essential property is: a | bc → (a | b or a | c)” (268). 104. Burge (1992, 1998); Yablo (2008b). 105. The transitivity of equinumerosity falls out of the transitivity of identity if ≪The Fs are equinumerous with the Gs≫ is taken to mean that the number of Fs = the number of Gs. 106. Since one never knows if something better is coming, the line between guides and true ultimate meanings is never finally drawn.

References Abreu Zavaleta, Martín. (2020). “Disagreement Lost.” Synthese 199 (1–2): 1899–1932. Google Scholar WorldCat Abreu Zavaleta, Martín. (2021a). “Communication and Indi erence.” Mind & Language 36 (1): 81–107. Google Scholar WorldCat Abreu Zavaleta, Martín. (2021b). “Communication and Variance.” Topoi 40 (1): 147–169. Google Scholar WorldCat Almog, Joseph. (1984). “Semantical Anthropology.” Midwest Studies in Philosophy 9 (1): 478–489. Google Scholar WorldCat Ambedkar, B. R. (2014). Annihilation of Caste: The Annotated Critical Edition. Edited by S. Anand. Introduction by Arundhati Roy. London: Verso Books. Google Scholar Google Preview WorldCat COPAC Anderson, Elizabeth. (1995). “Knowledge, Human Interests, and Objectivity in Feminist Epistemology.” Philosophical Topics 23 (2): 27–58. Google Scholar WorldCat Austin, John Langshaw. (1956). “A Plea for Excuses.” Proceedings of the Aristotelian Society 57:1–30. Google Scholar WorldCat Baker, Mark C. (2008). The Atoms of Language: The Mindʼs Hidden Rules of Grammar. New York: Basic Books. Google Scholar Google Preview WorldCat COPAC Bald, Suresht R. (2003). “The Politics of Gandhiʼs ʻFeminism.ʼ” In Sita Ranchod-Nilsson and Mary Ann Tétreault, eds. Women, States and Nationalism: At Home in the Nation?, 83–100. London: Routledge. Google Scholar Google Preview WorldCat COPAC Ball, Derek. (2018). “Lewisian Scorekeeping and the Future.” Croatian Journal of Philosophy 18 (54): 375–384. Google Scholar WorldCat Ball, Derek. (2020a). “Relativism, Metasemantics, and the Future.” Inquiry 63 (9–10): 1036–1086. Google Scholar WorldCat Ball, Derek. (2020b). “Revisionary Analysis without Meaning Change (or, Could Women Be Analytically Oppressed?” In Burgess, Cappelen, and Plunkett, Conceptual Engineering and Conceptual Ethics, 35–57. Google Scholar Google Preview WorldCat COPAC Ball, Derek, and Torfinn Thomesen Huvenes. (2022). “A Puzzle about Accommodation and Truth.” Philosophical Studies 179 (3): 759–776. Google Scholar WorldCat Berkeley, George. (2012). “De motu” and “The Analyst”: A Modern Edition. Edited and translated by Douglas M. Jesseph. New Synthese Historical Library 41. Dordrecht: Springer Science & Business Media. Google Scholar Google Preview WorldCat COPAC Bourke, Joanna. (2008). “Sexual Violence, Marital Guidance, and Victorian Bodies: An Aesthesiology.” Victorian Studies 50 (3): 419–436. Google Scholar WorldCat

p. 54

Bromberger, Sylvain. (2011). “What Are Words? Comments on Kaplan (1990), on Hawthorne and Lepore, and on the Issue.” Journal of Philosophy 108 (9): 486–503. Google Scholar WorldCat Builes, David, and Trevor Teitel. (2020). “A Puzzle about Rates of Change.” Philosophical Studies 177 (10): 3155–3169. Google Scholar WorldCat Burge, Tyler. (1979). “Individualism and the Mental.” Midwest Studies in Philosophy 4:73–121. Google Scholar WorldCat Burge, Tyler. (1982). “Other Bodies.” in Andrew Woodfield, ed., Thought and Object: Essays on Intentionality, 97–120. Oxford: Clarendon Press. Google Scholar Google Preview WorldCat COPAC Burge, Tyler. (1984). “Frege on Extensions of Concepts, from 1884 to 1903.” Philosophical Review 93: 3–34. Google Scholar WorldCat Burge, Tyler. (1986). “Intellectual Norms and Foundations of Mind.” Journal of Philosophy 83 (12): 697–720. Google Scholar WorldCat Burge, Tyler. (1992). “Frege on Knowing the Third Realm.” Mind 101:633–650. WorldCat Burge, Tyler. (1998). “Frege on Knowing the Foundation.” Mind 107:305–347. WorldCat Burgess, Alexis, Herman Cappelen, and David Plunkett, eds. (2020). Conceptual Engineering and Conceptual Ethics. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Byrne, Alex. (2003). “Color and Similarity.” Philosophy and Phenomenological Research 66: 641–665. Google Scholar WorldCat Cappelen, Herman. (2018). Fixing Language: An Essay on Conceptual Engineering. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Carnap, Rudolf. (1962). Logical Foundations of Probability. Chicago: University of Chicago Press. Google Preview WorldCat COPAC Google Scholar Chomsky, Noam. (1959). “A Review of B. F. Skinnerʼs Verbal Behavior.” Language 35 (1): 48–63.

Chomsky, Noam, and Morris Halle. (1968). The Sound Pattern of English. New York: Harper and Row. Google Scholar Google Preview WorldCat COPAC Code, Alan. (1985). “On the Origins of Some Aristotelian Theses about Predication.” In James Bogen and James E. McGuire, eds., How Things Are, 101–131. Dordrecht: D. Reidel. Google Scholar Google Preview WorldCat COPAC Darwin, Charles. (1981). The Descent of Man, and Selection in Relation to Sex. Princeton, NJ: Princeton University Press. Google Scholar Google Preview WorldCat COPAC Davidson, Donald. (2013). “Mental Events.” In Ned Block, ed., Readings in Philosophy of Psychology, vol. 1, 107–119. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC

Díaz-Léon, Esa. (2020a). “Descriptive vs. Ameliorative Projects: The Role of Normative Considerations.” In Burgess, Cappelen, and Plunkett, Conceptual Engineering and Conceptual Ethics, 170–186. Google Scholar Google Preview WorldCat COPAC Díaz-Léon, Esa. (2020b). “On the Conceptual Mismatch Argument: Descriptions, Disagreement, and Amelioration.” In Teresa Marques and Asa Wikforrs, eds., Shi ing Concepts, 190–212. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Dorr, Cian, and John Hawthorne. (2013). “Naturalness.” In Karen Bennett and Dean W. Zimmerman, eds., Oxford Studies in Metaphysics: Volume 8, 3–77. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Downey et al. (1981). The Legal Status of Adolescents 1980. Report. U.S. Department of Health and Human Services, O ice of the Assistant Secretary for Planning and Evaluation. Dumont, Louis. (1980). Homo Hierarchicus: The Caste System and Its Implications. Chicago: University of Chicago Press. Google Scholar Google Preview WorldCat COPAC Dutilh Novaes, Catarina. (2020a). “Carnap Meets Foucault: Conceptual Engineering and Genealogical Investigations.” Inquiry 1– 27. DOI: 10.1080/0020174X.2020.1860122 Google Scholar WorldCat Dutilh Novaes, Catarina. (2020b). “Carnapian Explication and Ameliorative Analysis: A Systematic Comparison.” Synthese 197 (3): 1011–1034. Google Scholar WorldCat p. 55

Dutilh Novaes, Catarina, and Erich Reck. (2017). “Carnapian Explication, Formalisms as Cognitive Tools, and the Paradox of Adequate Formalization.” Synthese 194 (1): 195–215. Google Scholar WorldCat Ebbs, Gary. (2000). “The Very Idea of Sameness of Extension across Time.” American Philosophical Quarterly 37 (3): 245–268. Google Scholar WorldCat Egan, Andy. (2004). “Second-Order Predication and the Metaphysics of Properties.” Australasian Journal of Philosophy 82 (1): 48– 66. Google Scholar WorldCat Egan, Andy. (2008). “Pretense for the Complete Idiom.” Noûs 42 (3): 381–409. Google Scholar WorldCat Essen, Louis, and Jack V. L. Parry. (1955). “An Atomic Standard of Frequency and Time Interval: A Caesium Resonator.” Nature 176 (4476): 280–282. Google Scholar WorldCat Evans, Gareth. (1973). “The Causal Theory of Names.” Aristotelian Society 47:187–208. Google Scholar WorldCat Fine, Kit. (2017). “A Theory of Truthmaker Content II: Subject-Matter, Common Content, Remainder and Ground.” Journal of Philosophical Logic 46 (6): 675–702. Google Scholar WorldCat Frege, Gottlob. (1950). The Foundations of Arithmetic: A Logico-Mathematical Enquiry into the Concept of Number. Trans. J.L. Austin. Oxford, UK: Basil Blackwell. Google Scholar Google Preview WorldCat COPAC

Frege, Gottlob. (1884). The Basic Laws of Arithmetic: A Logical Mathematical Investigation of the Concept of Number. Breslau: Verlag von Wilhelm Koebner. Google Scholar Google Preview WorldCat COPAC Gardenfors, Peter. (2014). The Geometry of Meaning: Semantics Based on Conceptual Spaces. Cambridge, MA: MIT Press. Google Scholar Google Preview WorldCat COPAC Gill, Patrick. (2011). “When Should We Change the Definition of the Second?” Philosophical Transactions of the Royal Society A:Mathematical. Physical and Engineering Sciences 369 (1953): 4109–4130. Google Scholar WorldCat Goddard, Cli , and Anna Wierzbicka. (1994). Semantic and Lexical Universals: Theory and Empirical Findings. Amsterdam: John Benjamins. Google Scholar Google Preview WorldCat COPAC Goldberg, Karen, and Joshua L. Dratel, eds. (2005). The Torture Papers: The Road to Abu Ghraib. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Grandy, Richard. (1973). “Reference, Meaning, and Belief.” Journal of Philosophy 70 (14): 439–452. Google Scholar WorldCat Hacking, Ian. (1975). Why Does Language Matter to Philosophy? Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Hacking, Ian (ed.). (1985). “Rules, Scepticism, Proof, Wittgenstein.” In Exercises in Analysis: Essays by Students of Casimir Lewy, 113–124. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Hale, Matthew. (1716). Pleas of the Crown. London: Printed by J.N. assignee of Edw. Sayer, Esq. for D. Brown … [and 4 others]. Google Scholar Google Preview WorldCat COPAC Hasday, Jill Elaine. (2000). “Contest and Consent: A Legal History of Marital Rape.” California Law Review 88:1373. Google Scholar WorldCat Haslanger, Sally. (2000). “Gender and Race: (What) Are They? (What) Do We Want Them to Be?” Noûs 34 (1): 31–55. Google Scholar WorldCat Haslanger, Sally. (2005). “What Are We Talking About? The Semantics and Politics of Social Kinds.” Hypatia 20 (4): 10–26. Google Scholar WorldCat Haslanger, Sally. (2011). “Ideology, Generics, and Common Ground.” In Charlotte Witt, ed., Feminist Metaphysics, vol. 6, 179–207. Dordrecht, Heidelberg, London, New York: Springer. Haslanger, Sally. (2014). “The Normal, the Natural and the Good: Generics and Ideology.” Politica & Società 3 (3): 365–392. Google Scholar WorldCat Haslanger, Sally. (2020). “Going on, Not in the Same Way.” In Burgess, Cappelen, and Plunkett, Conceptual Engineering and Conceptual Ethics 230–260. Google Scholar Google Preview WorldCat COPAC Haslanger, Sally. (2020). “How Not to Change the Subject.” In T. Marques and A. Wikforss, eds., Shi ing Concepts: The Philosophy and Psychology of Conceptual Variation, 235–249. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC

p. 56

Hawthorne, John. (2007). “Craziness and Metasemantics.” Philosophical Review 116 (3): 427–440. Google Scholar WorldCat Hawthorne, John, and Ernest Lepore. (2011). “On Words.” Journal of Philosophy 108 (9): 447– 485. Google Scholar WorldCat Higginbotham, James. (1998). “Conceptual Competence.” Philosophical Issues 9:149–162. Google Scholar WorldCat Hinton, Nichols, Leanne Hinton, Johanna Nichols, and John J. Ohala. (2006). Sound Symbolism. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Hirschman, Albert O. (1991). The Rhetoric of Reaction. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Hughes, Christopher. (1997). “Same-Kind Coincidence and the Ship of Theseus.” Mind 106 (421): 53–69. Google Scholar WorldCat Jackendo , Ray. (2002). Foundations of Language: Brain, Meaning, Grammar, Evolution. Oxford University Press. Google Scholar Google Preview WorldCat COPAC Jackman, Henry. (1999). “We Live Forwards but Understand Backwards: Linguistic Practices and Future Behavior.” Pacific Philosophical Quarterly 80 (2): 157–177. Google Scholar WorldCat Jackman, Henry. (2020). “Temporal Externalism, Conceptual Continuity, Meaning, and Us.” Inquiry 63 (9–10): 959–973. Google Scholar WorldCat James, William. (1992). Writings 1878–1899: Psychology, The Will to Believe, and Other Essays in Popular Philosophy. New York: The Library of America. Google Scholar Google Preview WorldCat COPAC Johnson, Kent. (2004). “From Impossible Words to Conceptual Structure: The Role of Structure and Processes in the Lexicon.” Mind and Language 19 (3): 334–358. Google Scholar WorldCat Kaplan, David. (1990). “Words.” Proceedings of the Aristotelian Society supplementary volumes, 64: 93–119. Google Scholar WorldCat Kaplan, David. (2011). “Words on Words.” Journal of Philosophy 108 (9): 504–529. Google Scholar WorldCat Knoll, Viktoria. (2020). “Verbal Disputes and Topic Continuity.” Inquiry 1–22. DOI: 10.1080/0020174X.2020.1850340. Google Scholar WorldCat Koslow, Allison. (2020). “Meaning Change in Theory and Practice.” PhD thesis, MIT. Google Scholar Google Preview WorldCat COPAC Koslow, Allison. (2022). “Meaning Change and Changing Meaning.” Synthese 200 (2): 1–26. Google Scholar WorldCat Kovecses, Zoltan, and Peter Szabco. (1996). “Idioms: A View from Cognitive Semantics.” Applied Linguistics 17 (3): 326–355. Google Scholar WorldCat

Kripke, Saul. (1975). “Outline of a Theory of Truth. Journal of Philosophy 72:690–716. Google Scholar WorldCat Kripke, Saul. (1980). Naming and Necessity. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Kripke, Saul. (1984). Wittgenstein on Rules and Private Language: An Elementary Exposition. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Kripke, Saul. (2011). “Vacuous Names and Fictional Entities.” In Philosophical Troubles: Collected Papers, vol. 1, 52–74. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Lakatos, Imre. (1978). “Cauchy and the Continuum.” Mathematical Intelligencer 1 (3): 151–161. Google Scholar WorldCat Lakatos, Imre. (2015). Proofs and Refutations: The Logic of Mathematical Discovery. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Leslie, Sarah-Jane, and Adam Lerner. (2016). “Generic Generalizations.” In Stanford Encyclopedia of Philosophy. (Winter 2016 edition). Ed. Edward N. Zalta. Google Scholar Google Preview WorldCat COPAC Lewis, David. (1974). “Radical Interpretation.” Synthese 27 (3–4): 331–344. Google Scholar WorldCat Lewis, David. (1984). “Putnamʼs Paradox.” Australasian Journal of Philosophy 62 (3): 221–236. Google Scholar WorldCat Losonsky, Michael. (1992). “Leibnizʼs Adamic Language of Thought.” Journal of the History of Philosophy 30 (4): 523–643. Google Scholar WorldCat Machery, Edouard. (2021). “A New Challenge to Conceptual Engineering.” Inquiry, DOI: 10.1080/0020174X.2021.1967190. Google Scholar WorldCat Makovec, Dejan, and Stewart Shapiro, eds. (2019). Friedrich Waismann: The Open Texture of Analytic Philosophy. Cham, Switzerland: Palgrave Macmillan. Google Scholar Google Preview WorldCat COPAC p. 57

Marchetto, Erika, and Luca L. Bonatti. (2013). “Words and Possible Words in Early Language Acquisition.” Cognitive Psychology 67 (3): 130–150. Google Scholar WorldCat Magueijo, Joao. (2009). “Bimetric Varying Speed of Light Theories and Primordial Fluctuations.” Physical Review D 79 (4): Article 043525. Google Scholar WorldCat McGinn, Colin. (1977). “Charity, Interpretation, and Belief.” Journal of Philosophy 74 (9): 521–535. Google Scholar WorldCat Nunberg, Geo rey, Ivan A. Sag, and Thomas Wasow. (1994). “Idioms.” Language 70 (3): 491–538. Google Scholar WorldCat Nyíri, J. C. (1992). “On Esperanto: Usage and Contrivance in Language.” In Tradition and Individuality, 85–92. Springer.

Google Scholar

Google Preview

WorldCat

COPAC

Ogawa, Yoshiki. (2001). “The Stage-Individual Distinction and (In)Alienable Possession.” Language 77:1–25. Google Scholar WorldCat Peacocke, Christopher. (1998a). “Implicit Conceptions, Understanding and Rationality.” Philosophical Issues 9:43–88. Google Scholar WorldCat Peacocke, Christopher. (1998b). “Implicit Conceptions, the ʻA Priori,ʼ and the Identity of Concepts.” Philosophical Issues 9:121– 148. Google Scholar WorldCat Pinker, Steven. (1995). The Language Instinct. New York: Harper Perennial. Google Scholar Google Preview WorldCat COPAC Pinker, Steven. (2013). Learnability and Cognition: The Acquisition of Argument Structure. Cambridge, MA: MIT Press. Google Scholar Google Preview WorldCat COPAC Ponet, John. (1556). “A Short Treatise of Politike Power, and of the True Obedience with Subjects Owe to Kings… Putnam, Hilary. (1974). “Meaning and Reference.” Journal of Philosophy 70 (19): 699–711. Google Scholar WorldCat Putnam, Hilary. (1980). “Models and Reality.” Journal of Symbolic Logic 45 (3): 464–482. Google Scholar WorldCat Quine, W. V. (1951). “Two Dogmas of Empiricism.” Philosophical Review 60 (1): 20–43. Google Scholar WorldCat Quine, W. V. (1969). “Natural Kinds.” In Ontological Relativity and Other Essays, 114–138. New York: Columbia University Press. Google Scholar Google Preview WorldCat COPAC Ranchod-Nilsson, Sita, and Mary Ann Tétreault, eds. (2003). Women, States and Nationalism: At Home in the Nation? London: Routledge. Google Scholar Google Preview WorldCat COPAC Richard, Mark. (2019). Meanings as Species. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Richard, Mark. (2020). “The A-Project and the B-Project.” In Burgess, Cappelen, and Plunkett, Conceptual Engineering and Conceptual Ethics, 358–378. Google Scholar Google Preview WorldCat COPAC Robinson, Abraham. (2016). Non-standard Analysis. Princeton, NJ: University Press. Google Scholar Google Preview WorldCat COPAC Russell, Bertrand. (2009). “The Superior Virtue of the Oppressed.” In Unpopular Essays, 72–78. Abingdon, UK: Routledge. Google Scholar Google Preview WorldCat COPAC Scharp, Kevin. (2019). “On the Indeterminacy of the Meter.” Synthese 196 (6): 2487–2517. Google Scholar WorldCat Scharp, Kevin. (2020). “Philosophy as the Study of Defective Concepts.” In Burgess, Cappelen, and Plunkett, Conceptual Engineering and Conceptual Ethics, 396–416.

Google Scholar

Google Preview

WorldCat

COPAC

Schroeter Laura, and Francois Schroeter. (2015). “Rationalizing Self-Interpretation.” In Chris Daly, ed., The Palgrave Handbook of Philosophical Methods, 419–447. London: Palgrave Macmillan UK. Google Scholar Google Preview WorldCat COPAC Schroeter, Laura, and Francois Schroeter. (2020). “Inscrutability and Its Discontents.” Canadian Journal of Philosophy 50 (5): 566– 579. Google Scholar WorldCat Sider, Theodore. (2013). Writing the Book of the World. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Skinner, B. F. (1957). Verbal Behavior. New York: Appleton-Century-Cro s. Google Scholar Google Preview WorldCat COPAC Skinner, B.F. (1961). “The Design of Cultures.” Daedalus 90: 534–46. Google Scholar WorldCat Skinner, B. F. (1965). Science and Human Behavior. New York: Simon and Schuster. Google Scholar Google Preview WorldCat COPAC Skinner, B. F. (1969). Walden Two. Indianapolis, IN: Hackett. Google Scholar Google Preview WorldCat COPAC Spencer, Herbert. (1865). Essays: Moral, Political, and Aesthetic. New York: D. Appleton and Company. Google Scholar Google Preview WorldCat COPAC p. 58

Spencer, Herbert. (1872). “Mr. Martineau on Evolution.” The Contemporary Review, 1866–1900 20:141–154. Google Scholar WorldCat Stalnaker, Robert. (2004). “Assertion Revisited: On the Interpretation of Two-Dimensional Modal Semantics.” Philosophical Studies 118 (1): 299–322. Google Scholar WorldCat Strawson, Peter Frederick. (1963). “Carnapʼs Views on Constructed Systems versus Natural Languages in Analytic Philosophy.” In Paul A. Schilpp, ed., The Philosophy of Rudolf Carnap, 503–518. Cambridge, UK: Cambridge University Press. Sundell, Timothy. (2020). “Changing the Subject.” Canadian Journal of Philosophy 50 (5): 580–593. Google Scholar WorldCat Tal, Eran. (2011). “How Accurate Is the Standard Second?” Philosophy of Science 78 (5): 1082–1096. Google Scholar WorldCat Tappenden, Jamie. (1995). “Extending Knowledge and ʻFruitful Conceptsʼ: Fregean Themes in the Foundations of Mathematics.” Noûs 29 (4): 427–467. Google Scholar WorldCat Tappenden, Jamie. (2008a). “Mathematical Concepts and Definitions.” In P. Mancosu, ed., The Philosophy of Mathematical Practice, 256–275. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Tappenden, Jamie. (2008b). “Mathematical Concepts: Fruitfulness and Naturalness.” In P. Mancosu, ed., The Philosophy of Mathematical Practice, 276–301. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC

Tarski, Alfred. (1936/1956). “The Concept of Truth in Formalized Languages.” In Logic, Semantics, Metamathematics, trans. Joseph H. Woodger, 152–278. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Traugott, Elizabeth Closs, and Richard B. Dasher. (2001). Regularity in Semantic Change. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Voltaire. (2013). Candide, or Optimism. London: Penguin UK. Google Scholar Google Preview WorldCat COPAC van Fraassen, Bas C. (2010). Scientific Representation: Paradoxes of Perspective. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC von Fintel, Kai, and Lisa Matthewson. (2008). “Universals in Semantics.” Linguistic Review 25 (1–2): 139–201. Google Scholar WorldCat Williams, Garrath. (2018). “Kantʼs Account of Reason.” In The Stanford Encyclopedia of Philosophy (Summer 2018 edition). Ed. Edward N. Zalta. Google Scholar Google Preview WorldCat COPAC Williams, J. Robert G. (2007). “Eligibility and Inscrutability.” Philosophical Review 116 (3): 361–399. Google Scholar WorldCat Williams, J. Robert G. (2018). “Normative Reference Magnets.” Philosophical Review 127 (1): 41–71. Google Scholar WorldCat Williamson, Timothy. (2007). The Philosophy of Philosophy. Malden, MA, Oxford, UK, Carlton, Australia: Blackwell Publishing. Google Scholar Google Preview WorldCat COPAC Wilson, Mark. (1982). “Predicate Meets Property.” Philosophical Review 91 (4): 549–589. Google Scholar WorldCat Wilson, Mark. (2006). Wandering Significance: An Essay on Conceptual Behavior. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Wittgenstein, Ludwig. (1980). Culture and Value. Trans. Peter Winch. Oxford: Blackwell. Google Scholar Google Preview WorldCat COPAC Yablo, Stephen. (1987). “Identity, Essence, and Indiscernibility.” Journal of Philosophy 84:293–314. Google Scholar WorldCat Yablo, Stephen. (1999). “Intrinsicness.” Philosophical Topics 26 (1/2): 479–505. Google Scholar WorldCat Yablo, Stephen. (2008a). “Beyond Rigidification: The Importance of Being Really Actual.” Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Yablo, Stephen. (2008b). “Carving Content at the Joints.” Canadian Journal of Philosophy 38 (supp1): 145–177. Google Scholar WorldCat Yablo, Stephen. (2017). “Precis of Aboutness.” Philosophical Studies 174 (3): 771–777. Google Scholar WorldCat

Yablo, Stephen. (2024). “Everything Exists: Plenitude, Kinds, and Paradox.” In Javier Cumpa, ed., The Question of Ontology. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

p. 59

3 A World Where Many Worlds Fit  Ángeles Eraña, Axel Barceló https://doi.org/10.1093/oxfordhb/9780192844118.013.2 Published: 22 May 2024

Pages 59–78

Abstract The Zapatista army’s radically pluralist proposal to build a world where many worlds t requires transforming our understanding of many basic ontological dualities, like the relation between the world and our conception of it, as well as our social practices, including our linguistic practices. This emancipatory project requires that we also revise many of the theoretical assumptions of our philosophical conception of language, stressing its genuinely social and material nature. This chapter sketches both the Zapatistas’ pluralist constructivist ontology and the revisions required for developing a philosophy of language that contributes to this emancipatory project.

Keywords: epistemic pluralism, linguistic practices, constructivist ontology, Zapatista movement, ontological dualities Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Disentangling an Ambiguity 1

The Fourth Declaration of the Lacandon Jungle, written and stated by the Zapatistas, starts as follows: Many words walk in the world. Many worlds are made. Many worlds are made for us. There are words and worlds which are lies and injustices. There are words and worlds which are truths and truthful. We make true words. We have been made from true words. In the world of the powerful

there is no space for anyone but themselves and their servants. In the world we want everyone ts. The world we want is one in which many worlds t. (our emphasis)

This latter assertion is in many ways ambiguous. It could be read in many di erent ways. Here we want to explore the following two: i. We want a world in which many societies t. 2

ii. We want a world in which di erent conceptions of the world t.

Our main contention is that (ii) boils down to (i). The way in which we live our lives is the foundation on which we build our conception of the world. The former, in turn, is determined by the social relations that structure our societies. Therefore, as we will argue, (ii) boils down to (i). Thus, we can understand the Zapatistas as saying that our world should be capable of supporting many di erent (though not all) 3

arrangements of social relations. p. 60

Moreover, from our perspective, a proper understanding of the Zapatistas’ initial assertion leads us to revise (and reject) the following two oppositions: a. The world as it is vs. the world as it is conceived (so that, even though the world might be conceived in 4

many ways, there is only one way the world actually is).

b. One vs. many (so that there being one world is preferable or more fundamental than there being 5

many).

The rejection we are proposing will lead us to endorse two very controversial theses. The rst asserts that the world as it actually is is not di erent from the world as it is conceived. The second stems from some ontological commitments that we associate with what we know about Amerindian thought, and leads us to sustain a relational ontology. Our defense of these two theses aims to show how the Zapatistas’ emancipatory proposal challenges some of our more fundamental and well-entrenched theoretical commitments. To reach the conclusion just mentioned, we will proceed as follows. We will rst show that there are good reasons to think that our conceptions of the world are very much determined, or delineated, by the social relations that actually take place in the societies we live in. Then, we will argue that if we adopt a materialist standpoint, it becomes plausible to assert that a conception of the world is neither di erent nor independent from the world as it actually is. Once we have established this, we will explore some of the ontological principles underlying Amerindian thought that help to show that it is not only not contradictory to assert that there is a world made up of many worlds, but rather, it is attuned to the way the world really is. We hope to show further that these are some of the reasons why the Zapatistas’ invitation is an emancipatory one. Finally, we will explore how our conception of language, as an essential site and tool of world-building, must evolve to contribute to this emancipatory project.

2. The Zapatistasʼ Claim As we just mentioned, we think striving for a world that not only allows but also promotes di erent conceptions of itself is nothing over and above desiring a world in which many di erent societies peacefully coexist. To better understand our contention, it is important to develop its terms more carefully. Let us assume, to start, that a society is a structured set of social relations, and that di erent structured sets result in distinct particular social systems, e.g., capitalism, feudalism, etc. From this perspective, societies are displayed as organized systems of social relations. The aforementioned sets are said to have a structure and not just a character because their con guration of elements is not a random collection, but a hierarchically ordered organization where some of the elements are more fundamental and stable than

p. 61

others. These components serve as individuation criteria for social systems: if they are

modi ed, the

system cannot be said to remain the same. Moreover, they delineate the correct or acceptable ways of thinking about (interpreting, perceiving) facts, events, or processes within a particular system. The Zapatistas think, for instance, that a crucial element in capitalism is bellicosity: our ways of thinking about others and ourselves, of interpreting facts and events, and of treating each other are very much modulated by it. From their perspective, war (and the di erent types of belligerences it brings about) is something without which such a system would not persist (Subcomandante Marcos 2017; Eraña 2021). In the Preface and the Postface of The Capital, Marx (1887) claims that (social) facts occur independently of our consciousness or will, and that they determine the latter. He further establishes that humanity “is no abstraction inherent in each single individual. In its reality it is the ensemble of social relations” (Marx 1969). There have been many interpretations of Marx’s ideas. Beyond the question of which is the correct one (if any), the point we want to stress from these citations is that it is plausible to assert that we (humans) are what our societies allow us to be. In other words, the way we live or explain reality is dependent on the particular arrangement of social relations we inhabit. When we speak about social relations, we are referring not only to modes of production, but also to the way we treat (think about and perceive) each other in ordinary circumstances—the ways in which we organize markets, institutionalize academic interactions, use language and produce speech, etc. All these di erent types of interactions are constitutive of us in the sense that they structure the way we actually do things, our cognitive abilities, the way we see and think about the world, etc. (Eraña 2021). It is in virtue of our participation in them that we assign certain meanings (and not others) to our experiences and interpret each other’s behavior. Furthermore, they urge us to perform the roles we execute in society and lead us to attain a notion of who we are. In Horstman’s words, “the totality of reality presents itself as the re ex of the way in which we understand ourselves, that is, of the way in which … we are capable of conceptualizing ourselves” (2014: 70). They also produce an ideal of who we want to be, and a conception of the world. Another way to put these ideas is as follows: the di erent shapes that social relations can take produce di erent kinds of social being. We are instantiations of ways of being social. We are, in other words, a way of life (an ensemble of social practices) which codevelops with social systems: we continually reproduce and transform the social practices that constitute the world we live in (Eraña 2021: 117). If this is the case, and if we accept that our conceptions of ourselves are intrinsically linked to the conceptions we have of the world, then we can accept that our social relations are constitutive of our conceptions of the world. In sum, if what we have so far argued is plausible, then it is reasonable to accept that the (actual) society we live in constitutes our conception of the world. Societies can be individuated through the speci c con gurations of social relations that they bear. Thus, we can accept that our conception of the world is a concept not only of the world, an idea (or a bundle of ideas) of how it works or is organized, but also (and p. 62

more importantly) of the particular social practices that take place on it. A conception of the

world is a

way to relate with others and with the world. We can plausibly claim that the Zapatistas advocate for a world which can support many di erent (though not all) arrangements of social relations. They would not accept that it must uphold all possible or actual arrangements. In particular, they would not consent to any organization of social relations that precludes others from existing (that is why, from their perspective, capitalism would be immediately ruled out).

3. The World and Its Conception When an originary people dies, a world is extinguished. —Juan Chávez, apud Subcomandante Insurgente Galeano

As we have just argued, every particular arrangement (or structured set) of social relations implies a particular organization of social practices i.e., the way we relate with others and the world involves speci c ways of doing things. Moreover, social relations determine correct ways of moving around the world, of understanding it and of interpreting our everyday life experiences. In this sense, they establish normative criteria that allow us to assess our own (and other people’s) practices and behaviors, to judge when some interpretations of facts are faulty (or incorrect or false) and when they are not. In other words, we have ascertained that social reality can achieve “di erent forms based on the ways we orient and shape our cooperative, interdependent, practical activities” (Thompson 2017: 210). We now want to maintain a more controversial thesis, i.e., there is no substantial or sharp di erence between the world as we conceive it and the world as it (actually) is. To support this latter contention, we will rely on Marx’s assertion that “the ideal is nothing else than the material world re ected by the human mind and translated into forms of thought” (1887: 14). It is 6

important to underline the materialist character of this dialectic. In contrast to materialisms of other sorts,

this establishes that reality takes the form of human activity or practice (Marx 1969). Reality is thus a sociohistorical product that stems from human activity, and which cannot be understood or accounted for or transformed if we do not acknowledge its historical and contingent character (Álvaro 2012). We said before that a society is a set of structured social relations and that our conception of the world is constituted by it (and, thus, that (ii) boils down to (i)). Our reality is thus constituted by whatever social practices take place in our (actual) world. This generalizes to any other world (possible or actual): each one of them is constituted by its own set of social practices. The world, from this perspective, is “a kind of social projection” (Haslanger 2012: 98): we can only speak of what takes place in the world from a perspective i.e., from a particular point of view which is determined by the society we live in and which, at the same time, determines how we experience reality, how we perceive and conceive it. p. 63

These ideas are related to a radical social constructionist approach, according to which the world as it is is (nothing over and above) the world as we conceive it. This proposal has some advantages over strong realism: for instance, it allows us to abandon the idea according to which things are the way they need to be, or that there are things which are inevitable (Haslanger 2012: 114, citing Hacking 1999: 12) and they thus cannot be changed. In this sense the perspective we have been putting forward allows us to think that we can 7

transform the world in which we live and thus motivate us to struggle in such direction. However, it also has its caveats: if there is not one way the world is, how can we tell which ways of being (or which worlds) are good or bad (just or unjust). In order to show that it is possible to prevent the sort of relativism that would likely follow from the caveat we just mentioned, it will be useful to distinguish between two ways of conceiving the world: (a) we conceptualize by adjusting our ideas to the world (or at least by trying to do it), by seeking to nd possible misconceptions, by contrasting our world view with others’, by confronting our conceptions with facts or explanations that they cannot accommodate, by bringing our normative structures into question. (b) we project our ideas to the world by attributing them to others and by extending them to reality. This latter mode leads to alienation i.e., it leads us to identify being and thought, and to assume that there is an absolute reality which is embodied in di erent particular forms which stem from our own (sometimes even individual) intuition (Pérez 2017: 40). It masks reality, covers it up and impedes us from revising our practices and our social structures. It rei es the social patterns that constitute our contingent historical moment (Boyd 1991: 147) and prevents us from organizing ourselves and our social relations in ways that are more just.

The rst mode, on the other hand, is a critical way of conceptualizing the world which opens the possibility for genuine emancipation. Whenever our thought is based on e ective reality (i.e., the real material structure underlying and supporting and constraining the arrangement of social relations in which we live) it has the power for real transformation since it leads to understanding and transformation (both of which require each other). Transformation of reality does not (cannot) depend on a voluntary decision of a single person. It requires collective resistance, rebelliousness, organization. With these ideas in mind, we can argue not only that the caveats in the thesis that the world as it is is nothing over and above the world as we conceive it can be avoided by taking a materialist stand. In the following sections we will show that these ideas are consistent with a kind of pluralism that is advantageous for emancipation. Before doing so, we will better explain our constructivist materialism and describe some ontological principles sustained by Amerindian peoples which help substantiate and endorse the idea that there is no one world without many of them. When we talk about creating or constructing worlds, we are talking not just about conceiving the world in p. 64

new imaginative ways. Rather we are talking about actually

building new societies i.e., of engaging in new

ways of relating with others (new ways of treating them and of being treated), of listening and speaking with them, of perceiving and conceiving them, of relating with the world. Thus, creating a world is not a matter of individually (or just theoretically) devising an idea and trying to implement it in the world. Rather it is a question of collectively changing our practices (including our linguistic practices, as we shall see). To achieve this goal, we need to understand reality (i.e., the real structure underlying and supporting the arrangement of social relations in which we live). And we can only do this if we transform the world: we need to understand reality in order to transform it, but without its reconstruction we cannot fully interpret it or understand it. As Marx says: inquiry… has to appropriate the material in detail, to analyze its di erent forms of development, to trace out their inner connection. Only after this work is done, can the actual movement be adequately described. If this is done successfully, if the life of the subject-matter is ideally re ected as in a mirror, then it may appear as if we had before us a mere a priori construction (Marx 1887: 14). Understanding the world requires going beyond individual assessments and delving into the details of sociality. In this sense there is no individual right or wrong consciousness, rather there are defective social organizations (and practices) and non-defective ones. Inquiring about the former can make us aware of the wrongful relations, practices and norms that make up our social reality and, thus, allow us to acknowledge the need for transformation. In other words, we become aware that it is of no necessity that the world is as it appears to be, and that there are many possible ways for it to be. We can stop reifying reality and become capable of transforming it. Usually, falsehood is thought of as a lack of correspondence between the world and how we think about it. But this conception of falsehood is based precisely on the world/worldview dichotomy that we want to question. Instead, we propose to think of false belief as a result of de ciencies in the system of social relations in which the belief occurs. False social consciousness, for example, should not be conceived, then, as a type of belief whose de ciency is that it does not correspond to the real social situation of the agent, but rather as the result of a system of social relations that does not allow the agent to assess the social system she inhabits (and/or to positively contribute to attain justice or fairness). To give another example, which we will detail later, when we mistakenly conceive of negation as a hierarchical relationship that places nonbeing in a subordinate position to being, instead of nding ourselves in an epistemic state that represents the world in a way that does not correspond to the world as it is, that is, independent of how we conceive it, this conception of negation constitutes the epistemic manifestation of a way of thinking about reality, which in turn is a way of relating to each other, in which subordination and hierarchy are the prevailing

schemas of thought. In other words, we mistakenly think of negation as hierarchical, not so much because we fail to notice that it is not hierarchical, as much as we participate in social relations that make it hard not to see hierarchies everywhere. p. 65

Our actions continually produce and reproduce the world. However, and this should not be controversial at all, they can also (and sometimes actually do) transform it. Through our words and deeds, we create new worlds: new societies, new ways of relating among ourselves, new ways of conceiving ourselves and the world. This is not something that merely a ects our so-called social world, leaving aside a so-called natural (or objective) world. From our perspective, there is no aspect of our world that is not a ected by this continuous social creation and transformation. Everything that composes our world is social in this sense. Even material things are not just material, but also social. As Haslanger has stated, “things are graced with the term ‘real’ not by virtue of some intrinsic fact about them but by virtue of some relevant social fact for example, our nding them useful or, perhaps, politically expedient” (2012: 109). Everything in the world is the object it is, at least in part, because of the role it plays in our social practices. Andromeda, for instance, is a real object in our world because of how we have conceived it. We think of it as being material and therefore as something that would exist even if there were no societies or humans. Andromeda—the galaxy also known as Messier 31 which might collide in around 4-5 billion years with the Milky Way; the socially meaningful entity that we use to explain and understand certain facts about the word—however, would not be part of our world were it not for how we are socially organized. Earthquakes are another good example. As physical events they would still take place as the consequences of volcanic activity, the motion of tectonic plates or any other known cause, even if there was nobody to conceive or 8

perceive them. However, since nobody would be there to experience or be a ected by them, they would not belong to any world at all. They are part of our world (at least partly) in virtue of the fact that their occurrence might cause harm to people or changes on Earth’s con guration that might somehow later a ect our lives. Social relations, as any social phenomenon, have a dual nature: they are objective and subjective at the same time. They are constituted by the meanings that we attach to them, but also by brute facts, by a material reality underlying them (Krause 2012: 342). The problem however is that sometimes we mix the meaning we assign to facts with reality, and we are thus incapable of grasping the “ontological structure of layers that dialectically mediate each other” (Thompson 2022: 151) and that end up producing a particular world. We end up blurring the distinction between “objective” and “independent” reality (Haslanger 2012: 111) i.e., the di erence between the social character of reality and the fact that there might be existing things that we do not know and might never get to know. Let us think again of Andromeda or earthquakes. The huge collection of gas, dust, and stars held together by gravity, or the shaking of the surface of the Earth are conceived of as independent facts i.e., as facts that would exist or occur no matter what we know, do, or perceive. At the same time, however, Andromeda and earthquakes are said to be ‘real’ because there are people who perceive and (think they) know them; because there are perspectives that provide particular descriptions of them. Merging these distinctions makes it di p. 66

cult to acknowledge that our conceptions of the world are constrained by material reality, but that it

takes the form it actually does because we signify it in a

particular way, because we organize ourselves in

some speci c way. Once we disentangle these ideas, it becomes plausible to think that our conceptions of the world and the world as it is, are intrinsically tied to each other, and that changing the former implies transforming the latter and vice versa. If what we have been arguing is accepted then it is reasonable to assert that the world can take many shapes, that every structured shape is a society and thus a world on its own. We can further state that there are as many worlds as arrangements of social relations there are (and probably as many as there can be). In this 9

sense there is not one world. It is now unclear how exactly a world can support (the existence) of di erent

worlds. Is this assertion not contradictory? In what follows we will examine what we take to be the Zapatistas’ negative answer to this question which, from our perspective, relies in the idea that there cannot be one world without many nor vice versa.

4. One World, Many Worlds If we assume, as much of our philosophical tradition has done, that in any pair of ontological categories, one is fundamental and the other derived (e.g., presence over absence; individual over collectivity; one over many, etc.), then the sentence “a world where many worlds t” might sound either contradictory or meaningless. Underlying this line of thought is a binary logic according to which the world can be divided or organized into polarities or into discrete categories that are made up of two di erent, opposite and mutually 10

exclusive parts.

As one of us has argued elsewhere, this way of conceiving the world is attached to an

asymmetrical social organization (Eraña 2021). Moreover, as we have here argued, there are some good reasons to assert that the world as it is, is just the way we conceive it. Thus, di erent conceptions of the categories with which we cut the world into pieces may bring about di erent worlds, di erent structured sets of social relations, practices, etc. If the assumption in question is true, then we would be facing a conundrum: a world seems to be a plural entity, i.e., one that is made from many di erent (types of) entities. However, it does not seem possible that a world is one thing because being one is not being many and being many means not being one. What kind of entity is a world? It could be answered that there is one world with di erent parts, or with many di erent explanations or conceptions of it. The former (the totality or the substance), it could be asserted, is fundamental, while the latter (be it the explanation or the parts) are derived. Such answer would imply that the world as we conceive it is derived from and subordinated to the world as it is. But, as we have argued, there are good reasons to think that this is not the way these two notions are related to each other. In what follows we will present an approach to ontology that stems from our reading of the way in which some Amerindian peoples think about how the world is made up and parsed down. From their perspective, relationality is fundamental, and thus oppositions are not exclusive but, rather, p. 67

complementary. By analyzing some of their main ideas, we hope to

show that it is possible to think that

the world is one and many at the same time and that the sentence “a world where many worlds t” is therefore not only not contradictory but substantially meaningful. Recent and old documents, as well as anthropological studies, suggest that Amerindian thought relies on some ontological fundamentals that do not presuppose the aforementioned ontological asymmetries nor accept that there is an opposition within ontological twosomes. There are several reasons to assume that Zapatistas’ political and social thinking hinges on some of those fundamentals. Hence, there are good reasons to assert that the pronouncement motivating our discussion is not a mere metaphor and that it displays a critical and radical proposal regarding how to transform social reality in order for us to live in a better world. Two of these Amerindian ontological principles are the following: 1. The duality principle (of complementary opposites): nothing exists without a complementary pair. 2. The linking principle: everything that exists is “related, or more importantly bonding [with other existences], looking for the proportion, oscillating between a

nities and repulsions” (Favela 2014:

11

49).

The rst principle was used to explain the order of the universe, its movement, and its intrinsic diversity (López Austin 1984: 59). It refers to particulars which are not single unities but pairs (Favela 2014). Earth

and sky, hot and cold, man and woman are conceived as polar pairs: their terms seem to be opposite, but it is impossible to conceive one without the other and thus they complement each other. The second principle, on the other hand, establishes that the relatedness of existence is fundamental and leads us to the idea of totality: “each term on a pair has the properties it has in virtue of its being part of this totality” (Gómez 2009: 127) and of the constitutive relation of the pair. From this perspective, there is no ontological primacy of one term in a pair over the other. Rather, each and every twosome is grounded on its particular relation to each other and to the resulting whole. The following metaphor might help illustrate these ideas: existence is grounded in a warp over and above which a weft is inserted, giving a place to particular ways of tying stu up, on top of this structure di erent weaves, each coupled pair, reveal themselves. In the Story of One and All, the Zapatistas say: Once there was a time when there was no time. It was the time of the beginning. It was about dawn. It was neither night nor day. Time was just there, without going anywhere or coming from anyplace. There was no light but it wasn’t dark either. It was the time when the great gods were alive … The eldest of our elders say that those rst gods were seven, and each was two. The eldest of our elders say that seven is the number the ancients used for the whole, and that one is always two so as to be able to walk together … the very rst gods … talked a lot … they couldn’t understand each other. (Subcomandante Insurgente Marcos 1998) p. 68

This tale tells us what there was before anything existed and how things came to make sense once the world was created: time had to start moving, and thus to move what used to be static. In this cosmogenesis, to create order in the world, i.e., to make it meaningful, means arranging dualities, pairing complementary opposites (Kusch 2007; Favela 2014). Only after this arrangement, all actual (and possible) twos started a dialectical movement from one of its terms to the other and back to the rst (e.g., from day to night to day, from rain to drought and back to rain, etc.). From this perspective, one (entity) cannot exist without an/the other: the world is made from multiple cohabiting complementary pairs (dualities), which constantly unfold and overlap, and which could not exist if not coupled. This is the reason it has been said that the duality principle should better be called the “parity principle” (Marcos 2014; Favela 2014). Only if one is two can it move or change, and one cannot be two if it does not move or change, there is light only if there is dark, and vice versa, and so on. Each of these pairs is not static or xed, time (history) rearranges them continually and thus provides the world with di erent structures. If we put these ideas together with the second aforementioned principle (the linking postulate). It becomes apparent that each pair is not just two singular things put together. Rather, they constitute an integral whole whose foundation is a dialectical relationship. The terms that make up a pair are two faces of the same coin: the rst is both what the second is, and what it is not, and vice versa. According to López Austin (2016), these principles were substantially misunderstood by Spanish interpreters when they arrived to Mesoamerica. In particular, they failed to understand how “the preeminence of one of the opposites supports that in a given a pair only one of the terms becomes apparent, as if it was representing and implying the other” (2016: 68). The main point here is that pairs are not static entities, and thus there are times where one term in each one becomes apparent, but that does not mean that it prevails, it just means that it is its time to play its role. These ideas have a close correlate in what has been called the “ontology of the common” (Álvaro 2012) and the dialectical character underlying what is known as the critical method. The former asserts that if there is something which is metaphysically fundamental, it is relationality: the world is composed of whatever there is between di erent entities, the set of diverse and heterogeneous relations that constitute it. Reality is thus

de ned (constituted) by multiple and active relations (Balibar 2006: 36). As we have already mentioned, social relations constitute us as much as we constitute our (social) reality, thus it is the structure of our relationships that exhibits itself in social practices which determines how the world actually is, and how we conceive it. The latter, on the other hand, establishes that in order for there to exist a dialectical move, existing things have to come as pairs. Each pair in dialectical opposition constitutes a unity-totality and each term on it “has the properties it has by its belonging to such totality and its peculiar relation (basically of ‘absence’) to the other [term in that pair]” (Gómez 2009: 7). Dialectical relations imply a continuous transition from one term in a pair to the other. It therefore seems p. 69

reasonable to assert that all there is exists because it coexists,

that any entity is a complex entity. If this is

the case, and if social reality is irreducible and fundamental, then the network of relations that constitutes reality seems to be its ground, that is populated by a diverse of complex couples. The network of relations in question is conditioned by its historical location. They can and normally do change. Thus, the speci c structure a world takes depends on how di erent networks of relations are accommodated in this, our world. If we take these proposals seriously then the idea that the world is multiple worlds (and that it can be multiple worlds at the same time) makes a lot of sense. We want a world in which many societies t, i.e., where di erent social structures coexist. A world like this would acknowledge diversity and would have room for it. It would, however, preclude the existence of any social organization that requires oppression to subsist, any world in which a social structure aims to have priority over others. Moreover, if these ideas sound acceptable, then it makes sense to assert that it does not seem to be the case that there being one world is (methodologically) preferable or (ontologically) more fundamental than there being many. Neither does it seem to be contradictory, paradoxical, or nonsensical to think that within a world many other worlds might exist.

5. Towards an Emancipatory Philosophy of Language While in other places they are ashamed of what they are, of their skin color, of your language of the blood that runs through your veins, we indigenous peoples wear it with pride, because what others despise means life. (Subcomandante Insurgente Marcos 2007; our translation) Language features prominently in many of our social practices. As such, it is constitutive of many of the world-making practices we have been talking about (Mendoza Sánchez and García Rodríguez 2015). It is not surprising, therefore, that the notion of “language” as an intrinsic aspect of a social order occurs many times in the Zapatista texts—there is talk of a neoliberal language and of the double-speak of the Mexican government, but there is also talk of the languages of the peoples and their central role in the emancipatory project of the Zapatistas. Philosophy of language, as a world making social practice itself, can and must transform itself in order to enable this radical transformation to take place. It might be tempting to think that a philosophy of language that contributes to this emancipatory project should be an appropriate kind of relativism, one where truth is relativized to worlds/societies. If we want a world where many worlds t, we might want to have a notion of proposition and/or truth that allows for

p. 70

propositions to be true in one world or according to one society and not true in other worlds or according to other societies, like the theories proposed by authors like Richard (2004) or MacFarlane (2014). Unfortunately, we do not nd the usual relativist accounts in contemporary analytic philosophy to be genuinely social. To take societies to be mere contextual indexes we must take in account when interpreting speech, while keeping the traditional vision of speech as an individual action, is at most a brief concession that does not recognize the radical social nature of speech. In traditional relativist accounts of speech, one may be able to model how successful communication between individuals from di erent societies is possible, and even account for the way this exchange may be interpreted from a di erent third perspective of eavesdroppers. But this way of modeling pluralism still represents societies as features of individuals and speech as essentially a tool of individuals to interact with other individuals. Instead, we propose a thoroughly social and materialist account of speech, one where speech acts are not individual actions but genuinely social actions. Consider, for example, the very text you are reading right now. Presumably, its goal is to partake in the ongoing philosophical conversation about the relation between language and politics. However, after Lewis’s seminal proposal (Lewis 1979), according to the prevailing way of thinking of conversation in analytic philosophy, we are making a series of assertions through which we are trying to communicate some beliefs we might want you to adopt. This way of approaching what is happening in this text has serious shortcomings. First of all, it places agreement as the ultimate goal of conversation. As such, it conceptualizes disagreement in conversation as something that must be overcome. This is clearly in direct tension with the sort of pluralism advocated in Zapatista thought. A world where many worlds t must be a world of healthy disagreement, even (or maybe specially) in the context of conversation. A world where many worlds t cannot be a mere aggregate of independent worlds, but must be one where interaction and dialogue prevail. Thus, the traditional assertion-centric view of conversation won’t do. Second, we are two persons, with di erent beliefs and perspectives on the issues contained in this text. Similarly, who are you, the reader? The traditional conception of speech has almost completely focused exclusively on speech acts performed by individual agents [i.e., speech acts where there] is a single speaker and a single hearer involved … Though this focus may initially have been due to Searle’s methodological approach of starting his analysis of speech acts with clear-cut examples, it has led in the end to an unnecessary and undesirable bias in speech act theory. [With rare exceptions] collective speech acts have escaped the attention of almost all philosophers of language thus far. (Meijers 2007: 3–4) In these traditional accounts, speech acts performed by and for individuals are presented as the paradigm to which theories of collective speech must conform, as if all that was required to account for collective speech was just to extend the traditional account of individual speech. But extending this account has proved

p. 71

elusive, for it demands thinking

of collectives as individuals—that is, to think of the collective agent

“Ángeles and Axel” as the (singular) speaker of these words. Thus, we are forced to face esoteric metaphysical questions, like: How it is possible for an agent to reside in two persons? Do we both need to believe everything we say for our assertions to be felicitous? Or is it enough for one of us to, and if so, what is the right attitude for the other to take? Should one of us defer to the words written by the other? And so on (Hughes 1984; Meijers 2007; Schmid 2014; Tolle sen 2020). We acknowledge the value of the many proposals to deal with these two postulated entities in the literature —we know there is ample and insightful work on collective action, in general, and collective speech, in particular, and on ideal audiences as well—but we are skeptical of the need for so much work to shoehorn what is essentially a social phenomenon into an individualistic paradigm (Eraña and Barceló Aspeitia 2016).

Instead, the Zapatistas encourage us to explore a di erent approach to this phenomenon, one that does not try to reduce social and collective phenomena to individual action and thought, but instead recognizes the irreducibility of the social. If we go back to the short paragraph at the beginning of the chapter, it is no accident that the subject of those sentences is always “we” and that we introduced the paragraph as being stated by the Zapatistas, in plural, as well as by Subcomandante Insurgente Marcos. What the Zapatistas exhort us to realize is that this sort of dialectic relation between individual and collectivity need no longer be considered an extraordinary aberration that needs to be shoehorned back into an individualistic model, but, instead, should take central stage in our philosophical approach to language. To achieve this goal, we propose to recover the materiality of speech. On this proposal, speech is not fundamentally a sort of interaction between individuals but a genuinely social practice: the coproduction and further circulation of text, a material resource in the pursuit of social value. Under this perspective, what we, Ángeles and Axel, are doing here is not merely a collective act of assertion; it is engagement in a social practice that precedes us and will continue long after we are gone. We are producing a text, this text, under the presumption that it is worth making public, and in doing so, are contributing to the social practice of philosophy and, in particular, the social discussion of the politics of language. And this text we produce is not the mere material wrapping of abstract ideas or of psychological thoughts that were once stored in our minds, but a material resource for world-making, and therefore, for society-making. Its value lies in this, its potential for change. In the end, it does not matter what we believe as individuals, but what we can say to make this a better world. A few lines ago we said that when we talk about creating or constructing worlds, we talk, not just about conceiving the world in new imaginative ways, but mainly about engaging in new ways of relating with others and with the world. Texts are part of the material resources that are recruited in this engagement, and language is a central part of what we want to change and conceive in an innovative and more just way. As philosophers we aim to understand language both to make the most of what it currently is—not in a sel sh or individualistic way but as society, the ultimate language bearers—and to responsibly steer it in new, more just directions. p. 72

Switching the focus to the materiality of text and language has many theoretical and practical advantages to an emancipatory project. As David Bleich has stated: The materiality of language is an axiom, a postulate, a fundamental assumption, a Kuhnian paradigm that leads to new approaches to the study and teaching of language, but is still not a part of our ordinary sense of what language is and does. It suspends the preoccupation with the referential and communicative functions of language and tries to examine all functions of speech and writing relative to the social, intersubjective scenes of their use. Language is material in the sense that it has tangible e ects and that it matters all the time. (Bleich 2003: 469) By conceiving of texts as material ‘ “resources for reasoning’ [that] get created, circulated (or spread through contagion), or tampered down” (Goodwin 2020: 179), we must acknowledge the need to secure its fair distribution and exploitation, for example. By thinking of language as social and material, we are closer to recognizing the value of linguistic pluralism, much more so than if we think of it just as a means of communication. As has been widely defended within linguistics (e.g., Bordelois 2010: 59–60; Guyot 2010; Vázquez Rojas Maldonado 2020), when we think of language primarily as a means of communication between individuals, there is a huge temptation to jump to the conclusion that languages ought to be evaluated by how well they ful ll this function. Consequently, one might think that one ought to eliminate any linguistic barriers to communication and, therefore, that the more people share a common language, the better (Baetens Beardsmore 2010). When we think of language instead as essentially socially situated,

we start to understand languages as means of world sharing and thus think instead that the more languages are spoken, the better. Languages as material resources enable and constrain social action and, consequently, enable and constrain social world-making. What languages we speak partially determines what worlds we create. In a dialectical relationship, communities create languages as much as languages create communities. A world where many worlds t must be a world where many languages t. It is hard to overstate the importance of language, and in particular, of the Mayan language to the emergence of the Zapatista movement. For their speech in front of the Frente de Pueblos en Defensa de la Tierra (People in Defense of the Land/Earth’s Front) at Atenco on July 23, 2006, Delegate Zero related the origins of the Zapatista Movement as a movement that sprang for and from below. He related the story, from the very early days of the movement, about meeting a fellow by the name of Antonio who invited him to talk with other members of his community about Zapata and the ght to return the Land to the People; but, he said, “Do not talk as you do, because you use hard words, we do not understand you; … you have to talk like us, that is, you have to learn to speak Mayan, which is our language” (Delegado Zero 2006a; our translation). In his own account, by learning the Mayan language, Delegate Zero learned that he had not come to teach, but to learn; not to lead, but to obey. By learning their language, he not only earned their p. 73

trust, but literally became one of them—but not of a preexisting them he

just joined, but of a new them

they built together: a new community, a movement, and a new world. A few months earlier, in Veracruz, when talking about the emergence of the Zapatista movement, Delegate Zero also appealed to the importance of learning Maya, insisting that “the EZLN does not use a language di erent from that used by the indigenous Zapatista communities to explain to themselves where they are and to explain to themselves what they want to do. And if it is successful, it is because it is the same simple language that we use and that most people use” (Delegado Zero 2006b), and he contrasted this with the usual custom of politicians of dressing up in the traditional attire of the communities they visit during their political campaigns. Languages are not disguises one can just try on to pass as member of a community, languages are not sites that one can just pass by, that one can just visit, but actual ways of living, ways of inhabiting the world. Finally, proposing such a radical change in our ontology as the one the Zapatistas propose, will also require making adequate adjustments to our way of thinking the relationship between language and world. The traditional models we use when trying to understand di erent features of language tend to reproduce our prevailing ontological prejudices. For example, traditional formal semantic models assume an ontology where individuals are fundamental, and the referents of plural expressions are derived through mereological or set theoretical construction (Winter and Sha 2014). This way, they reproduce a hierarchical view of the duality between individuals and collectivities. In a similar fashion, the classical, extensional treatment of negation assumes a hierarchical view of what is over what is not (Eicher 2018), because, on the one hand, it operationalizes the idea that what is determines also what is not, i.e., that once it is determined whether a proposition is true or not, the truth value of its negation becomes determined as well, and vice versa. This mutual determination becomes hierarchical once one proposition is marked (the negative one), while the other is left unmarked as the default (the positive one). In other words, an asymmetry is introduced into the logical relation of negation by the convention of having a symbol for negation, but not one for a

rmation.

In possible world semantics, this hierarchy manifests itself in the traditional way of characterizing possible worlds by the set of true positive atomic propositions, instead of by the set of true negative atomic propositions, even though both characterizations would be equivalent in a classical framework. In both cases, there is no asymmetry in the logical phenomenon itself. Yet it is introduced by us in our theoretical approach to such phenomena. In truth-maker semantics, we see a similar bias of what is over what is not in the traditional assumption that an account of truth that includes only positive facts is preferable over one

that contains also negative facts (Russell 1919). One might argue that the former is preferable over the later as a matter of parsimony, but this would not explain why it is positive facts which are assumed as unproblematic and it is only the existence of negative facts which is put into question. In other words, even if we agreed on a theoretical preference for only one kind of fact, why would positive facts be preferable over p. 74

negative ones? Similarly, one might argue that a symbol for a

rmation is not

necessary once we have

one for negation, hence having both symbols would make our formal syntax unnecessarily complex. However, the dual claim is also true, i.e., a symbol for negation would not be necessary if we had one for a

rmation. In other words, even though it is true that a single symbol is su

why the single symbol must correspond to negation instead of a facts that impose a hierarchy between negation and a

cient, this does not explain

rmation. Once again, it is not the logical

rmation, but us by the way we study logic.

Fortunately, many recent proposals in logic, metaphysics, and the philosophy of language are at least friendly to this endeavor. Taking negation to be an intensional operator, as is done in some paraconsistent logics, for example, allows for a more ontologically balanced view of the relation between what is and what is not, and the resulting dialetheist logics can also serve as better logical foundations for true dialectic thought (Estrada-González and Olmedo-García 2007). Under an intensional interpretation of negation, to determine the truth value of a negative proposition, it is not enough to ask whether the negated proposition is true or not. We need separate provisions to determine the truth of both positive and negative propositions. This means that fully representing reality requires determining not only what is but also what is not, in a way that is neither subordinate nor fully indi erent to what is (Estrada-González 2014). This has not proved to be an easy task, but it is one that we consider to be in the right direction. It is not that linguistics and traditional analytic philosophy of language have been unaware that their individualistic model was an abstraction that purposely ignored the complex social aspects of linguistic phenomena. It is, instead, that they thought the abstraction was of theoretical value, and that the cost was more than justi ed (e.g., Johnstone 1996; D’Agostino 1979, etc.). After all, analytic philosophy of language has been built on the paradigm of scienti c knowledge construction, where abstraction is king (Barceló 2019) and thus has no problem recognizing that for the longest time it has purposely ignored the social dimensions of speech and language. What has become more di

cult to ignore in recent times is that such

purposeful abstraction of the social has had a political cost greater than any putative theoretical value it might be balanced with, as this very handbook bears witness to. As we have mentioned, the goal of this emancipatory philosophy of language is not to merely t our ideas about language to the reality of language; it is also to project our emancipatory ideas and ideals to language, not only as it is, but also as we want it to be, i.e., to understand language as well as transform it. The point is not so much that the current picture of language being painted by contemporary analytic philosophy of language is inaccurate—indeed, contemporary philosophy of language has made great strides in improving our understanding of language as it currently is—yet we think it can and must do better. Mostly, it must do better for us, the people. Thus, even though our proposal is based on e ective reality (i.e., the real structure underlying and supporting actual linguistic practice), it aims for real transformation. As such, it is an invitation to join in a collective e ort of philosophical resistance, rebelliousness, and organization.

p. 75

6. Some Final Thoughts The way a society is structured depends on the practices it supports and promotes. These practices are conditioned by their historical placement and are open to transformation once a critical perspective is adopted. They can thus take di erent forms and help to create better ways of relating to each other. Any theory that tells us otherwise, that presents our current societal arrangements as perennial must be rejected, for it is incapable of appreciating the potential and bene cial “forms of life and ends that social life is capable of achieving” (Thompson 2017). What the Zapatistas seem to be saying, or what we can take them to say, is not only that di erent ways of organizing ourselves t into our world, but also that it is necessary that they exist for a better world to exist. However, not all possible social structures would t in such a world: worlds in which only one world ts could not survive. This is so, not only because their existence denies the possibility of any other structure to exist, but also because they are not apt to lead to justice. For example, if we accept that capitalism is a social system (a social formation) that requires oppression and exploitation in order to persist, then a capitalist world would be precluded.

Notes

p. 76

1.

When we speak of the Zapatistas, we are talking about the EZLN (Zapatista Army for National Liberation), an army basically made up by indigenous people. They rose in arms against the Mexican government on January 1, 1994. Twelve days a er uprising they decreed a ceasefire, and they have been very active in the pacific (sociopolitical) construction of an alternative world since then. Up to now, they have stated six declarations, each of which has been accompanied by a sociopolitical proposal for civil society.

2.

In this reading we could find another ambiguity: between di erent conceptions of the world and di erent conceivable worlds. We think that this second possibility is not what is meant by the Zapatistas.

3.

To be consistent, they couldnʼt accept an arrangement of social relations which precluded other arrangements from existing (that is why capitalism and many other kinds of authoritative or exclusivist arrangements of social relations would be immediately ruled out).

4.

Asserting that the world is identical to our conceptions of it is worrisome, since it can lead to an undesired relativism. However, as we will argue, if we take a materialist standpoint, this worrisome option is not available.

5.

This idea seems to imply that given that there is but one world (and thus one way the world can be), we are destined to live in the world as it is.

6.

Marx (1969) criticizes, for instance, Feuerbachʼs materialism because it “does not comprehend sensuousness as practical activity” and is thus the “contemplation of single individuals and of civil society,” instead of being “human society, or social humanity.”

7.

We can fight and make it a better world (a world with more justice, with peace, without poverty, etc.)

8.

In fact, earthquakes are a better example, because those that are caused by human activity would not exist at all if there were no societies and probably would not exist if our societies were not as they actually are (e.g., all of those that are caused by nuclear tests would simply not take place if there were no nuclear tests done).

9.

There can exist di erent worlds within one world. The question is whether they can all coexist. We could say that each of them is just a part of a world (as opposed to each being a coexistent totality). However, as we will see in what follows, thinking of them as totalities seems to be more consistent with the ontology that, as fast as our understanding goes, underlies the Zapatistasʼ world view.

10.

There is a great deal of literature which argues that this logic underlies much of our most prevailing theories of gender, society, etc. For a more detailed account of these ideas, cfr. Favela (2014); Segato (2016); Gutiérrez (2014); Mbembe (2020);

Eraña (2021). 11.

Neither of these principles is meant to be self-evident. The evidence that Amerindian people use to support them can be found in original documents and, more accessible to us, in anthropological studies, such as those by López Austin (1984) or León Portilla (1983). More importantly, however, these principles are consistent with their world, with their conception of it and their way of organizing life. These are constitutive (and are constituted by) their social relations, which becomes evident the more one gets to know the Zapatista movement and their way of life.

References Álvaro, Daniel (2012). “Marx y la ontología de lo común.” Nómadas 36 (4): 13–29. Google Scholar WorldCat Balibar, Étienne (2006). La filosofía de Marx. Translated by Horacio Pons. Buenos Aires: Nueva Visión. Google Scholar Google Preview WorldCat COPAC Baetens Beardsmore, Hugo (2010). “Taaldiversiteit—inzet en strategiën.” Bulletin dés Séances 46 (4): 427–437. Google Scholar WorldCat Barceló, Axel (2019). Sobre el Análisis. Mexico City: Instituto de Investigaciones Filosóficas, Universidad Nacional Autónoma de México (UNAM). Google Scholar Google Preview WorldCat COPAC Bleich, David (2003). “Materiality, Genre, and Language Use: Introduction.” College English 65 (5): 469–475. Google Scholar WorldCat Bordelois, Ivonne (2010). Del silencio como porvenir. Buenos Aires: Libros del Zorzal. Google Scholar Google Preview WorldCat COPAC Boyd, Richard (1991). “Realism, Anti-foundationalism and the Enthusiasm for Natural Kinds.” Philosophical Studies 61 (1): 127– 148. Google Scholar WorldCat DʼAgostino, F. B. (1979). “Individualism and Collectivism: The Case of Language.” Philosophy of the Social Sciences 9 (1): 27–47. Google Scholar WorldCat Delegado Zero (2006a). “El Delegado Zero en Atenco.” Enlace Zapatista, July 23, 2006. https://enlacezapatista.ezln.org.mx/2006/07/24/el-delegado-zero-en-atenco-23-de-julio/. Delegado Zero (2006b). “En Huayacocotla Veracruz.” Enlace Zapatista, February 27, 2006. https://enlacezapatista.ezln.org.mx/2006/02/27/en-huayacocotla-veracruz-26-de-febrero/. Eicher, Lauren (2018). “Sacred Truths, Fables, and Falsehoods: Intersections between Feminist and Native American Logics.” American Philosophical Association Newletter: Native American and Indigenous Philosophy 18 (1): 4–11. Google Scholar WorldCat Eraña, Ángeles (2021). De un mundo que hila personas. Mexico, IIFs, UNAM y UAM-C. Google Scholar Google Preview WorldCat COPAC Eraña, Ángeles, and Barceló Aspeitia, Axel Arturo. (2016). “El conocimiento como una actividad colectiva.” Tópicos 51 (51): 9–35. Google Scholar WorldCat p. 77

Estrada-González, Luis (2014). “On the Possibility of Realist Dialetheism” SATS 15 (2): 197–217. Google Scholar WorldCat Estrada-González, Luis, and Claudia Olmedo-García (2007). “Can Paraconsistency Replace Non-Monotonicity?” LA-NMR 2009: 217–224. Favela, Mariana (2014). “Ontologías de la diversidad.” In Márgara Millán (coord.), Más allá del feminismo: Caminos por andar. Mexico City: Red de Feminismos Descoloniales, 35–60. Google Scholar Google Preview WorldCat COPAC

Gómez, Ricardo (2009). “Karl Marx: Una concepción revolucionaria de la economía política como ciencia.” Herramienta 13 (40): 119–141. Google Scholar WorldCat Goodwin, Jean (2020). “Should Climate Scientists Fly? A Case Study of Arguments at the System Level.” Informal Logic 40 (2): 157–203. Google Scholar WorldCat Gutiérrez, Raquel (2014). “Políticas en femenino: Reflexiones acerca de lo femenino moderno y del significado de sus políticas.” In Márgara Millán (coord.), Más allá del feminismo: Caminos por andar. Mexico City: Red de Feminismos Descoloniales, 87–98. Google Scholar Google Preview WorldCat COPAC Guyot, Jacques (2010). “La diversidad lingüística en la era de la mundialización.” Historia y Comunicación Social 15:51–66. Google Scholar WorldCat Hacking, Ian (1999). The Social Construction of What? Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Haslanger, Sally (2012). Resisting Reality. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Horstman, Rolf-Peter (2014). “Substance, Subject and Infinity: A Case Study of the Role of Logic in Hegelʼs System.” In K. Deligiorgi, ed., Hegel: New Directions. London: Routledge, 69–84. Google Scholar Google Preview WorldCat COPAC Hughes, Justin (1984). “Group Speech Acts.” Linguistics and Philosophy 7:379–395. Google Scholar WorldCat Johnstone, Barbara (1996). The Linguistic Individual: Self-Expression in Language and Linguistics. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Krause, Juljan (2012). “Collective Intentionality and the (Re)production of Social Norms: The Scope for a Critical Social Science.” Philosophy of the Social Sciences 42 (3): 323–355. Google Scholar WorldCat Kusch, Rodolfo (2007). América profunda: Obras Completas. Tomo 2. Buenos Aires: Editorial Fundacíon Ross. León Portilla, Miguel (1983). La filosofía Náhuatl. Mexico City: Instituto de Investigaciones Históricas, Universidad Nacional Autónoma de México (UNAM). COPAC Google Scholar Google Preview WorldCat Lewis, David (1979). “Scorekeeping in a Language Game.” In R. Bäuerle, U. Egli, A. von Stechow, eds., Semantics from Di erent Points of View, Springer Series in Language and Communication 6. Berlin: Springer, 172–187. https://doi.org/10.1007/978-3-64267458-7_12. Google Scholar Google Preview WorldCat COPAC López Austin, Alfredo (1984). Cuerpo humano e ideología. 2 tomos. Mexico City: Instituto de Investigaciones Antropológicas, Universidad Nacional Autónoma de México (UNAM). Google Scholar Google Preview WorldCat COPAC López Austin, Alfredo (2016). “La cosmovisión de la tradición mesoamericana.” Arqueología mexicana 69 (2): 23–89. Google Scholar WorldCat MacFarlane, John (2014). Assessment Sensitivity: Relative Truth and Its Applications. Oxford: Oxford University Press.

Google Scholar

Google Preview

WorldCat

COPAC

Marx, Karl. (1969). “Theses on Feuerbach.” In Marx/Engels Selected Works, vol. 1, Translated by W. Lough, 3–15. Moscow: Progress Publishers. Marxist Internet Archive. https://www.marxists.org/archive/marx/works/1845/theses/theses.htm. Google Scholar Google Preview WorldCat COPAC Marx, Karl (1887). “A erword to the Second German Edition,” in Capital, vol. 1, translated by Samuel Moore and Edward Aveling, edited by Frederick Engels, 9. Moscow: Progress Publishers. Marxist Internet Archive. https://www.marxists.org/archive/marx/works/download/pdf/Capital-Volume-I.pdf. Google Scholar Google Preview WorldCat COPAC Marcos, Silvia (2014). “Feminismos en camino descolonial.” In Márgara Millán (coord.), Más allá del feminismo: Caminos por andar. Mexico City: Red de Feminismos Descoloniales, 15–34. Google Scholar Google Preview WorldCat COPAC p. 78

Mbembe, Achille (2020). Necropolítica. España, Melusina. Google Scholar Google Preview WorldCat COPAC Meijers, Anthonie (2007). “Collective Speech Acts.” In S. L. Tsohatzidis ed., Intentional Acts and Institutional Facts. Theory and Decision Library 41. Dordrecht: Springer. Google Scholar Google Preview WorldCat COPAC Mendoza Sánchez, Citlalli, and García Rodríguez, Raúl Ernesto (2015). “Discursos Ético-Políticos y Prácticas Sociales Cotidianas en las Comunidades Zapatistas.” Psicologia & Sociedade 27 (3): 650–659. Google Scholar WorldCat Pérez Varona, Wilder (2017). Marx: Política y enajenación. La Habana: [email protected]. Google Scholar Google Preview WorldCat COPAC Richard, Mark (2004). “Contextualism and Relativism.” Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition 119 (1/2): 215–242. Google Scholar WorldCat Russell, Bertrand (1919). “I—On Propositions: What They Are and How They Mean.” Aristotelian Society Supplementary Volume (Vol. 2, No. 1, pp. 1–43). Oxford, UK: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Schmid, Hans Bernhard (2014). “Expressing Group Attitudes: On First Person Plural Authority.” Erkenntnis 79:1685–1701. Google Scholar WorldCat Segato, Rita (2016). La Guerra contra las mujeres. Madrid: Traficantes de Sueños. Google Scholar Google Preview WorldCat COPAC Subcomandante Insurgente Galeano (2017). “Una historia para tratar de entender.” http://enlacezapatista.ezln.org.mx/2016/11/17/una-historia-para-tratar-de-entender/. WorldCat Subcomandante Insurgente Marcos (1996). Fourth Declaration of the Lacandona Jungle, in https://schoolsforchiapas.org/library/fourth-declaration-lacandona-jungle/. WorldCat Subcomandante Insurgente Marcos (1998). “La historia del uno y los todos.” http://enlacezapatista.ezln.org.mx/1998/12/11/historia-del-uno-y-los-todos/. WorldCat

Subcomandante Insurgente Marcos (2007). “Participación de la Comisión Sexta en la subsede de Rancho El Peñasco, territorio de la nación TOHONO ODʼHAM, Magdalena de Kino, Sonora, 8 y 9 de octubre de 2007.” https://enlacezapatista.ezln.org.mx/2007/10/10/participacion-de-la-comision-sexta-en-la-subsede-rancho-el-penasco-8-y-9-deoctubre/. WorldCat Thompson, Michael J. (2017). “Collective Intentionality, Social Domination and Reification.” Journal of Social Ontology 3 (2): 207– 229. Google Scholar WorldCat Thompson, Michael J. (2022). “The Ontology of the Social as the Ontology of Social Forms.” In J. Halley and H. Dahms, eds., The Centrality of Sociality: Responses to Michael Brownʼs ʻThe Concept of the Social in Uniting the Humanities and Social Sciences.ʼ Emerald Press, 149–173. Google Scholar Google Preview WorldCat COPAC Tolle sen, Deborah (2020). “Can Groups Assert That P?” In Sanford Goldberg, ed., The Oxford Handbook of Assertion. New York: Oxford University Press, 327–344. Google Scholar Google Preview WorldCat COPAC Vázquez Rojas Maldonado, Violeta (2020). “Contar lenguas.” Otros Diálogos de El Colegio de México (online), no. 10. https://otrosdialogos.colmex.mx/contar-lenguas. Winter, Yoad, and Sha, Remko (2014). “Plurals.” In Shalom Lappin and Chris Fox, eds., Handbook of Contemporary Semantics, 2nd edn. Hoboken: Wiley-Blackwell, 77–113. Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

p. 79

4 Silencing and Assertion: An Account of Their Conversational Dynamics  J. L. Dowell https://doi.org/10.1093/oxfordhb/9780192844118.013.3 Published: 22 May 2024

Pages 79–103

Abstract Silencing occurs when a conversational participant is not accorded an ordinary speaker’s standing to update the conversational record. This chapter argues that silencing poses a distinctive problem for assertion in contexts of inquiry. In the standard model, an assertion met with silence is one that has not been rejected by any participant. Like a

rmation, silence adds the asserted content to the

Common Ground. When a conversational participant is silent because she has been silenced, however, her silence does not indicate assent. In such cases, assertions cannot succeed in their characteristic perlocutionary e ect, mutual audience belief. The remedy for this distinctive problem of assertion requires a remedy for the conditions that give rise to silencing. The chapter closes with a proposal for how to revise the standard model to allow it to represent the phenomenon of silencing, as well as the conditions needed for repair.

Keywords: assertion, silencing, Common Ground, discourse goals, discourse dynamics, context, conversational scoreboard Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

On the account of the speech act of assertion that I have been using, an assertion is something like a proposal to change the context by adding the content expressed in an assertion to the common ground. … I am not claiming that one can de ne assertion in terms of [that] context-change rule, since that rule will govern speech acts that fall under a more generic concept [e.g., conjecture] … whether a speech act counts as an assertion or not will be partly a matter of … what [the discourse’s] common purposes are taken to be….

… [that] model of the speech act of assertion is highly idealized and oversimpli ed … One speci c way that model is oversimpli ed is this: the context-change rule is a default rule. Although it is part of the assertion game that the interlocutors have the opportunity to reject the assertion, it is a rule of the game that the content of the assertion becomes part of the common ground if it is not rejected. But of course, in any realistic situation, it cannot be assumed that the addressees accept every assertion that they do not explicitly reject. Consider [for example] a lecture … [it] is not expected that hearers will interrupt at each point where they are unwilling to accept the speaker’s claim.” —Robert Stalnaker (2014: 89–90); last three emphases mine The ability to perform speech acts can be a measure of political power … Conversely, one mark of powerlessness is an inability to perform speech acts that one might otherwise like to perform. … Members of a powerless group may be silent because they are intimidated, or because they believe that no one will listen. —Rae Langton (1993: 314–315) Speakers require audiences to ‘meet’ their e ort ‘halfway’ in a linguistic exchange … in short, to p. 80

communicate we all need an audience willing and

capable of hearing us … An instance of

silencing concerns a single, non-repetitive instance of an audience failing to meet the dependencies of a speaker, whereas a practice of silencing, on my account, concerns a repetitive, reliable occurrence of an audience failing to meet the dependencies of a speaker. —Kristie Dotson (2011: 238–241) Mr. Vice President, I’m speaking. I’m speaking. If you don’t mind letting me nish, we can have a conversation. —Kamala Harris, US vice-presidential debate, October 7, 2020 The free speech of men silences the free speech of women. It is the same social goal, just other people. —Catherine MacKinnon (1987: 156)

1. Introduction Above, Robert Stalnaker suggests that in “realistic scenarios,” we cannot always assume that an assertion 1

met with silence has been assented to by every conversational participant. Kristie Dotson and Rae Langton consider circumstances in which this is so. Potential speakers may choose to remain silent because they have been intimidated or believe that they will not be ‘heard’ or ‘listened to’ by their audience. In the literature on epistemic injustice and feminist philosophy of language, such cases are known as a type of 2

silencing. As a rst pass, these are cases in which a speaker with su

cient epistemic reason to challenge an

assertion is unable to do so because of her failure to be accorded the ordinary conversational standing to reject her interlocutors’ assertions. A more precise characterization of the phenomenon of interest will be given in sections 2 and 3. For now, let’s make do with a concrete illustration. I once heard of an academic who yelled at his colleagues. He didn’t always yell. He yelled only when the issue of diversity came up. And he didn’t yell at all his colleagues. He would only yell at a colleague from a 3

marginalized group when she spoke up in favor of a diversity-promoting proposal in a faculty discussion.

 Sometimes. Other times, he would lie in wait, yelling at his outspoken, marginalized colleague when the meeting was over and his other colleagues had left. “Oh, he has a temper,” the other faculty would tell their marginalized colleagues. “Pay him no mind. We don’t.” It was generally agreed, though, that the yelling was unpleasant. When meetings were held, it became di p. 81

cult for marginalized faculty to speak up or be heard. Over time, meetings on diversity were postponed

to avoid their colleague’s wrath. It was not long

before the department could no longer have open and

honest discussions among the faculty about diversity. Call this case “Yelling Colleague.” Here I argue that representing the type of silencing exempli ed in Yelling Colleague requires an addition to Stalnaker’s idealized model of assertion. As we’ve seen, on that idealized model an assertion successfully updates the set of propositions mutually presupposed for the purposes of conversation (aka the Common Ground) just in case it is not rejected. Rejection here requires an act on the part of a conversational participant. An assertion met with silence is one that has not, in this sense, been rejected by any participant. In a case in which a potential speaker is silent because she has been silenced, though, her silence does not indicate assent. In such cases, assertions cannot be successful. Let’s draw out the implications of this. In much of the existing literature on silencing, the primary focus is (rightly) on the injustice to the speaker of failing to accord her an ordinary speaker’s conversational standing, e.g., the presumption that she means what she says or that her assertions are to be assessed on 4

epistemic grounds. Here, though, we are focused on the consequences of a certain type of silencing for assertion as such. In conversations in which a participant is silent because she has been silenced, silence does not indicate assent. And when it can no longer indicate assent, assertions can no longer do their job of helping inquirers become more jointly opinionated about what the world is like. In other words, silencing is more than an injustice. As illustrated by Yelling Colleague, it is a form of communicative breakdown in contexts of inquiry. Making good on these claims will proceed in ve stages. In section 2, I’ll consider in more detail some general conditions on successful assertion. In section 3, I’ll consider a rather ordinary type of conversational situation that will allow us to isolate the di erence silencing makes to an assertion’s success. Isolating the di erence that makes a di erence to an assertion’s success or failure in our cases will allow us to see what must be true of a conversational context in order for silence to indicate assent, and so for an assertion to be successful. Spoiling the suspense, I’ll argue that the source of conversational breakdown in the cases of interest is silencing itself. For this reason, silencing presents a distinctive form of assertional barrier, one that cannot be explained on other grounds, such as a failure of mutual cooperation. In section 4, I’ll argue that for an assertion’s non-rejection to signal mutual assent, conversational participants must each believe that they have the standing to assert or reject an assertion successfully on 5

epistemic grounds alone. As we’ll see, though, except in somewhat unusual cases, such joint belief will depend upon each participant’s having such standing. In section 5, I’ll o er reasons to think that modeling the conversational dynamics of silence requires adding conversational standing as an element of contexts of utterance. Finally, in section 6, I’ll propose a way to model conversational standing using resources we have independent reasons to posit in an overall model of how communication occurs.

p. 82

2. Assertion and Its Success Conditions On a widely accepted, theoretical notion of contexts of utterance, contexts play a dual role in explaining communication. First, they play a role in determining the content or, more broadly, the signi cance of an utterance. Second, they play a role in representing how the joint acceptance of an utterance impacts a conversational exchange. A speci c account of what contexts need to contain for them to play these two roles is an account of conversational dynamics. Following David Lewis, a standard way of representing contexts is in terms of a conversational scoreboard. As I shall understand them, a conversational scoreboard represents the features of a conversational situation that play this dual role and that participants need to track for successful communication to occur. On the in uential Stalnakerian model, contexts include a Common Ground (CG), i.e., a set of propositions presupposed for the purposes of conversation. In our opening passage, Stalnaker suggests that an assertion is a proposal to update the Common Ground with its content. This is the “context-change rule” for assertions, i.e., the rule governing how they are to update the context of utterance. However, this rule does not yet distinguish assertions from other sorts of speech acts. For example, like assertions, conjectures are proposals to update what is presupposed for conversational purposes. Whether or not such a proposal is an assertion, Stalnaker suggests, depends on what those conversational purposes are. Discourse goals, aka conversational purposes, are also widely recognized as an element of Lewisian 6

scoreboards, as they are understood here. What must a conversation’s discourse goal be in order for an utterance to be an assertion, rather than a conjecture? One widely recognized hallmark of assertions, one that contrasts them with conjectures, is that they express the speaker’s belief. A natural suggestion, then, is that assertions may be distinguished from conjectures in contexts of inquiry (COIs), i.e., contexts in which the primary overarching discourse goal is for participants to become better jointly opinionated about what the 7

actual world is like.

COIs are fairly ubiquitous. They include discussions among members of research groups, activists, and policymakers. More broadly, any conversational situation in which interlocutors are engaged in joint practical decision-making where the stakes are su

ciently high will be one in which facts about what the

actual world is like will be of signi cant conversational interest, e.g., discussions between parents, teachers, mechanics, and medical professionals. These are all contexts of inquiry. On one way of thinking of COIs, the Common Ground represents what is mutually believed. Its status as such is recorded on the conversational scoreboard as a discourse goal of inquiry. On another way of thinking of them, that the status of what is presupposed as mutual belief is itself presupposed (and so part of the Common Ground). In what follows, I’ll treat discourse goals as a separate element of the Lewisian scoreboard, though nothing here hangs on this choice. On the resulting picture, an assertion is a proposal to update a Common Ground that represents a state of mutual belief. p. 83

When is such a proposal successful? There are two di erent ways an utterance might succeed or fail as an assertion. The rst is illocutionarily. For our purposes, to succeed illocutionarily as an assertion, an utterance needs only to have an assertion’s characteristic conversational update rule in a context in which a primary discourse goal is a better informed, mutual belief state. This is why what the lecturer does, in Stalnaker’s example, still counts as asserting, even though she does not expect that everyone will come to believe everything she says. Here our focus is on the conditions under which an assertion is perlocutionarily successful in the respect that gives it its distinctive point. The characteristic perlocutionary aim of an assertion is for the audience to come thereby to believe its content. In this respect, an assertion is successful when its content is added to the common stock of belief. Under what conditions is an assertion’s content added to the set of propositions mutually believed? Above, Stalnaker suggests that the “ideal” situation is one in which the absence of explicit rejection signals assent.

In such circumstances, an assertion’s content becomes part of what is mutually believed if conversational participants explicitly assent or remain silent. It’s the latter phenomenon that is of interest here. To underscore: on Stalnaker’s idealized model, silence is a form of conversational acceptance. This gives silence itself a role to play in conversational dynamics. Ideally, a proposal to update an element of a 8

conversational scoreboard, such as the Common Ground, is successful when it is met with silence. Here we are interested in a particular form of departure from this ideal—namely, cases in which a conversational participant with su

cient epistemic reason to reject an assertion’s proposal fails to do so because she has

been silenced. “Silencing” is a term of art drawn from the literature on this topic. A speaker is silenced, roughly, when her audience fails to accord her the standard speaker’s authority in some respect—for example, to be taken to 9

mean what she says (as in the case of sexual refusals) or to have her assertions accepted or rejected on 10

epistemic grounds alone (as in the case of testimony).

One way this can happen is if a speaker rejects an

assertion with good epistemic reason, but she is ignored for non-epistemic reasons. In contrast, in the cases of primary interest here, the silenced potential speaker is literally silent as a result of her recognition of her failure to be accorded the conversational standing to have her assertions assessed on epistemic grounds 11,12

alone.

What happens to an unrejected assertion made in the course of a conversation in which at least one member is silent because they have been silenced? To address this, I next consider a more careful representation of the phenomenon at issue.

3. Silent from Silencing Consider a series of relatively simple, clear examples of a context of inquiry, each of which is a variation on the same basic case involving a meeting of lab workers who are trying to identify the explanation for the failure of their experiment. They are together reviewing the assumptions on which the experiment was p. 84

based to identify which

is most likely false, given their collective information. The experiment was based

on assumptions A, B, and C. The overarching goal is to use the conclusion of their discussion to develop the best experimental redesign. The lab is organized into the usual hierarchy of principle investigator, sta scientists, postdocs, graduate students, and techs. Deliberation begins with exclusive focus on assumptions A and B. (Call this basic case “Lab Meeting.”) From this basic case, we’re interested in the contrast between two sets of cases. In each case, the Sta Scientist makes an assertion. The rst pair illustrates cases in which the context of utterance is ideal in Stalnaker’s sense, i.e., ideal with respect to an assertion’s ability to be successful. The second set illustrates non-ideal cases of the kind of interest here—cases in which a speaker with epistemic reason to challenge an assertion has been silenced. Individual cases within a set are contrasted by whether or not the Sta Scientist’s assertion is challenged. The rst member (or members) of each set represents a case in which a postdoc has epistemic reason to challenge the Sta

Scientist’s assertion and does so. In the second member

of each set, Postdoc remains silent. In the ideal case, this will be because she lacks epistemic reason to challenge that assertion. In the non-ideal case, this will be because she has been silenced.

3.1. Lab Meeting Consider rst, Ideal 1: Reasons-Successful Challenge

Principle Investigator: Ok, team. What went wrong? Was our mistake in assuming A? Or assuming B? Sta

Scientist: Well, for [reasons] assuming B has to be the mistake.

PI: That sounds right. Thoughts? Anyone else? Rest of group: {Silence as each surveys their epistemic reasons for and against rival explanations.} Postdoc: Maybe not. For [other reasons], B could well be correct. For [further reasons], assuming C may be the culprit. SS: What? Oh, come on. That’s ridiculous—C has to be true. PI: No. Postdoc is right. Though it seems unlikely, [further reasons] are good reasons to think C can’t be ruled out as the faulty assumption. That possibility needs to be investigated. Update: That either B or C is false is added to the CG. The CG represents a mutual belief state. Ideal 2: No reason–No challenge Principle Investigator: Ok, team. What went wrong? Was our mistake in assuming A? Or assuming B? Sta p. 85

Scientist: Well, for [reasons] assuming B has to be the mistake.

PI: That sounds right. Thoughts? Anyone else? Rest of group: {Silence as each surveys their epistemic reasons for and against rival explanations. No one nds any that is more plausible, given their evidence.} Principle Investigator: No further thoughts, then? Ok. Let’s redesign the experiment without assuming B. 13

Update: That B is false is added to the CG. The CG represents a mutual belief state. Non-ideal 1a and 1b: Reasons–Unsuccessful challenge 1a: Ridicule PI: Ok, team. What went wrong? Was our mistake in assuming A? Or assuming B? SS: Well, for [reasons] assuming B has to be the mistake. PI: That sounds right. Thoughts? Anyone else?

Rest of group: {Silence as each surveys their epistemic reasons for and against rival explanations} Postdoc: Maybe not. For [other reasons], B could well be correct. For [further reasons], assuming C may be the culprit. SS: What? Oh, come on. That’s ridiculous--C has to be true.PI: {Pause} No further thoughts, then? Ok. Let’s redesign the experiment without assuming B. Postdoc: {Whaaaat? What about [other reasons] and [further reasons]?} Update: Crash. While B’s falsity may come to be presupposed for the remainder of the conversation, the CG no longer represents a mutual belief state (as Postdoc remains unconvinced). 1b: Ignore PI: Ok, team. What went wrong? Was our mistake in assuming A? Or assuming B? SS: Well, for [reasons] assuming B has to be the mistake. PI: That sounds right. Thoughts? Anyone else? Rest of group: {Silence as each surveys their epistemic reasons for and against rival explanations} Postdoc: Maybe not. For [other reasons], B could well be correct. For [further reasons], assuming C may be the culprit. PI: {Pause} No further thoughts, then? Ok. Let’s redesign the experiment without assuming B. Postdoc: {Whaaaat? What about [other reasons] and [further reasons]?} Update: Crash. While B’s falsity may come to be presupposed for the remainder of the conversation, the CG no longer represents a mutual belief state (as Postdoc remains unconvinced).

Non-ideal 2: Reasons–No challenge Principle Investigator: Ok, team. What went wrong? Was our mistake in assuming A? Or assuming B? Sta

Scientist: Well, for [reasons] assuming B has to be the mistake.

PI: That sounds right. Thoughts? Anyone else? p. 86

Rest of group: {Silence as each surveys their epistemic reasons for and against rival explanations} Postdoc: {Maybe not. For [other reasons], B could well be correct. For [further reasons], assuming C may be the culprit. But Sta

Scientist always ridicules what I say. What’s the point of speaking

up? No one ever sticks up for me.} Principle Investigator: {Pause} No further thoughts, then? Ok. Let’s redesign the experiment without assuming B. Update: Crash. While B’s falsity may come to be presupposed for the remainder of the conversation, the CG no longer represents a mutual belief state (as Postdoc remains unconvinced). 14

All of our non-ideal cases are examples of silencing.

For now, though, set aside Non-ideal 1a and 1b and

focus on the second non-ideal case. We are interested in understanding the conversational dynamics of silence as a result of silencing. In the contrast between Ideal 2 and Non-ideal 2, we can see that silencing can make a di erence to an assertion’s ability to e ectively advance the conversational goal of becoming better, jointly opinionated about what the actual world is like. In Ideal 2, Sta

Scientist’s assertion is successful:

Had anyone believed they had epistemic reason to challenge his assertion, they would have done so. But since they don’t, no one does. The content of his assertion is thereby added to the common stock of belief. In Non-ideal 2, in contrast, although Postdoc believes she has su

cient epistemic reason to challenge Sta

Scientist’s assertion, she does not do so, thinking her challenge will go unnoticed or worse. Since Postdoc does not come to accept what Sta

Scientist says, the latter’s assertion is not able to update the mutual

belief state represented by the conversation’s Common Ground. There are two sides to this conversational failure. First, Sta

Scientist’s assertion has failed to achieve the characteristic perlocutionary aim of

asserting. In this sense, it is a failed assertion. Second, the conversation has become one in which its guiding 15

goal, the goal of inquiry, can no longer be achieved.

This is a form of conversational breakdown. As we’ll

see, these failures are characteristic of silencing. Here we’re interested in better understanding the source of the failure of Sta

Scientist’s assertion in Non-ideal 2. In section 4, we’ll consider what added contextual

features might constitute a repair. However, at this point one might wonder whether the phenomenon of interest here is really one that is not already well understood. Prima facie, it may seem that the di

culty in

Non-ideal 2 is simply that by remaining silent when she has good reason to speak up, Postdoc is uncooperative in the Gricean sense. Grice’s Cooperative Principle: Make your conversational contribution such that it is required, at the stage at which it occurs by the accepted purpose… of the talk exchange in which you are engaged. Grice (1989: 26) On Stalnaker’s account, not merely assertion but communication more broadly is not possible absent 16

common interest and common knowledge of that common interest.

This is plausibly so even on more 17

conventionalized, non-Gricean accounts of how communication via language is best explained. p. 87

Postdoc’s failure to cooperate is the source

of the failure of Sta

If

Scientist’s assertion in Non-ideal 2, then

that case does not illustrate a phenomenon that is not already well understood in terms of Stalnaker’s framework.

A more careful examination of that case, however, shows that the source of the failure cannot be that Postdoc fails to cooperate by not speaking up. To see this, consider the minimal contrast case to Non-ideal 2 in which Postdoc does speak up. Since by stipulation, Non-ideal 2 is a case in which Postdoc is silent because she has been silenced, the minimal contrast case will be one in which she speaks up, while the conditions that give rise to her silencing remain. This is possible because one may be silenced in our sense without being silent. In our sense, recall, a potential speaker is silenced when she has been deprived of a participant’s ordinary standing to update the conversational record. More speci cally, in the cases of interest here, a silenced speaker is one deprived of the standing to have her assertions accepted or rejected on epistemic grounds. The minimal contrast with Non-ideal 2 in which Postdoc does speak up, then, will be a case like Non-ideal 1a or 1b, cases in which her challenge is met with rejection on non-epistemic grounds. But these cases are also cases in which Sta

Scientist’s assertion fails. Because her challenge is rejected on

non-epistemic grounds, Postdoc remains unconvinced and the new Common Ground no longer represents a mutual belief state. This suggests that the source of the failure in our non-ideal cases is not Postdoc’s failure to cooperate. It is that Postdoc has been silenced. While the failure of Sta

Scientist’s assertion in Non-ideal 2 is not plausibly explained by the hypothesis

that Postdoc is uncooperative, perhaps it can still be explained as a failure of the Cooperative Principle. Perhaps the source of the failure is that Sta

Scientist lacks su

cient commitment to shared conversational

goals. This is perhaps suggested by the contrast between Non-ideal 1a (reasons-challenge-ridicule) and Nonideal 2 (reasons-no challenge). In 1a, Sta That may suggest that Sta

Scientist e ectively blocks Postdoc’s challenge through ridicule.

Scientist is more committed to blocking challenges to his claims than he is

committed to the joint goals of inquiry. In such cases, a participant rigs the payo s for mounting a challenge to his assertion in order to shut down even well-grounded dissent. To the extent that a participant’s conversational contributions aim more at advancing personal interests than the joints goals of inquiry, they violate Grice’s principle. If Postdoc were silent in Non-ideal 2 solely because she knows Sta Scientist has rigged the payo s in this way, the explanation for the conversational breakdown in that case might seem to be Postdoc’s knowledge that Sta

Scientist is uncooperative. In light of this, a failure of the

Cooperative Principle might still seem to explain the conversational breakdown in Non-ideal 2. However, this explanation, too, cannot be right because Sta

Scientist’s insu

cient commitment to the

goal of inquiry cannot be the sole reason why Postdoc is silent in Non-ideal 2. To see this, rst consider the contrast between Non-ideal 2 and Ideal 1. In the latter case, Sta

Scientist ridicules Postdoc’s contribution

without engaging it on epistemic grounds. So, the hypothesized respect in which he is uncooperative in Non-ideal 2 is also a respect in which he is uncooperative in Ideal 1. In the latter, however, his ridicule does not prevent that challenge from receiving a hearing. There is no conversational breakdown in that case p. 88

because his ridicule does not have the desired e ect. The

Principal Investigator takes her contribution

seriously, and consequently, her assertion updates the conversational record. This pair of cases illustrates the respect in which silencing is a ‘team sport.’ Setting aside conversational dictatorships, silencing an individual in a group conversation of three or more is not something one can 18

accomplish on one’s own.

Sta

Scientist’s ridicule does not by itself deprive Postdoc of her ordinary

conversational standing to have her assertion assessed on epistemic grounds. The silencing in our Non-ideal cases and more generally is secured by an enabling audience. Since the di erence between Ideal 1 and Nonideal 2 isn’t that Sta

Scientist is cooperative in the former, but not the latter, but rather that his

hypothesized lack of cooperation in the latter but not the former leads to silencing, a better explanation for 19

the conversational breakdown in the latter in contrast to the former is silencing itself.

For the anticipation

of ridicule to play a role in the explanation for why Postdoc remains silent in a case like Non-ideal 2, it must play a role in her expectation that she will be silenced, as in Non-ideal 1a, in contrast to Ideal 1. These considerations suggest that even in cases of explicit bullying, it is the silenced individual’s understanding of her de facto conversational status that explains her silence, not the bullying by itself.

Second, notice that the hypothesis that Sta

Scientist is uncooperative cannot explain the full range of cases

like Non-ideal 2. That Postdoc anticipates Sta

Scientist’s successful ridicule is but one way that silencing

can lead to a case like Non-ideal 2. There are many di erent ways a speaker might be silenced, explicit bullying being only one. Other ways of silencing are compatible with conversational situations in which each participant sincerely accepts the goals of inquiry as their overarching discourse goal and aim to speak so as to further those goals. For example, a would-be challenger may recognize that her fellow conversationalists 20

do not accord her a standing to mount a challenge that is commensurate with her subject matter expertise. Or she may recognize that her interloctors’ behaviors manifest unacknowledged, implicit bias toward individuals who share her group membership, which makes expressing her challenge pointless (or worse). 21

Or she may recognize that the “pernicious ignorance”

22

of or stereotypes held by her audience

renders

them unable to appreciate her challenge. All such cases will lead to conversational breakdown because her challenge goes unaired. Since the challenge goes unaired precisely because the potential challenger is aware she has been silenced, silencing itself best explains the breakdown in such cases. We’ve now seen that silencing itself poses a distinctive problem of assertion, one that leads to a form of 23

conversational breakdown that cannot simply be explained as a failure of cooperation.

Before considering

what the repair for this problem might be, consider an aspect of its signi cance. Prima facie, from the point of view of at least some participants in Non-ideal 2, their situation may appear indistinguishable from Ideal 2. That is, at least some participants may not be able to tell whether Postdoc is silent because she assents to what Sta

Scientist has said or because she has been silenced. This means that the di erence between an

assertion’s perlocutionary success and failure may be indistinguishable to some conversational participants. Call this “the Predicament of Inquiry.” This predicament is of no mere theoretical interest. In p. 89

contexts of inquiry,

the di erence between the two cases may be of great practical signi cance, as in Lab

Meeting, where the di erence may make a di erence to a future experiment’s success. As I’ll now argue, the locus of our problem of assertion suggests that the repair will take the form of a repair to the conditions that give rise to silencing itself.

4. The Repair What must contexts of utterance be like for assertions to avoid failure from silencing? To get the proposal on the table, consider rst a plausible principle of rational discussion in contexts of inquiry. Equal Standing: Interlocutors must each have equal standing to update the conversational record on epistemic grounds alone. To see what this principle entails (and what it does not), note rst that it does not require that each participant’s assertion be assessed on epistemic grounds alone. Participants may recognize other conversational goals that constrain the goal of inquiry. For example, suppose the lab members have a schedule to which they’ve all agreed to adhere. This may constrain the length of time they have in which to deliberate. This may mean that some farfetched, but not conclusively false, hypotheses get set aside for reasons of time, rather than reasons of decisive evidence. What Equal Standing does require is that to whatever extent other goals constrain the goal of inquiry, they constrain the assessment of each participant’s contributions alike. This does not require that each participant’s contributions are scrutinized with equal care. In cases like Lab Meeting in which there are mutually recognized di erences in subjectrelevant expertise (e.g., between the principal investigator and a rst-year grad student), there may be di erences, compatible with Equal Standing, in the quality of the assessment of contributors’ assertions. But assessments of a participant’s contribution based on her level of expertise will be epistemic assessments.

What Equal Standing requires is that no participant’s contributions are assessed less on epistemic grounds than any other’s. Plausibly, the goal of joint inquiry, of a group’s coming together to share information so as to become better, more jointly opinionated about what the world is like, is better pursued in conversations in which Equal Standing obtains than in conversations in which it does not. As I’ll also now suggest, Equal Standing has a role to play in explaining the di erence between our Ideal cases and our Non-ideal cases in which silencing leads to its characteristic conversational breakdown. To see this, rst consider Non-ideal 2. In that case, the barrier to Postdoc’s raising her challenge is her belief that her challenge will not be assessed equally on epistemic grounds. That is to say, the barrier is her failure to believe that she herself has equal conversational standing. What is true of Postdoc, though, is true of any arbitrary conversational participant. Where any p. 90

conversational participant is silent from silencing, this will be because of

their belief that they fail to

have equal conversational standing. For this reason, for silence to indicate assent, each participant must believe of themselves that they have equal standing. Call this hypothesis “Re exive Belief in Equal Standing.” Re exive Belief in Equal Standing: Each conversational participant P believes of themselves that they have equal conversational standing. Because of the Predicament of Inquiry, however, Re exive Belief does not su

ce for silence to indicate assent

in such cases. To see this, focus again on Postdoc. What’s needed in addition is that her audience also believes that she herself believes that she has such standing. Without such belief, an audience could not assume that her silence indicates assent. (It may instead indicate that she does not think her challenge will receive a fair hearing.) Again, though, there is nothing special about Postdoc that makes what is true of her fail to be true of any other conversational participant. In general, for any conversational participant’s silence in response to an assertion to indicate assent, it must be the case that they each believe that each other also believes that they themselves have equal standing. Call this hypothesis “Mutual Re exive Belief in Equal Standing.” Mutual Re exive Belief in Equal Standing: Each conversational participant P believes that each other conversational participant N believes that N has equal standing (where P ≠ N). When those conditions obtain, there is no barrier from silencing to prevent a participant from speaking up. So, when those conditions obtain and a participant remains silent, there is no such barrier to that silence’s 24

indicating assent.

However, while adding these conditions su participant’s speaking up, they do not yet su

ce to ensure that there is no barrier from silencing to a ce to ensure there is no conversational breakdown of the kind

characteristic of silencing. To see this, we need to distinguish between the context as it is before an assertion is challenged and the context as it is afterward. Consider, rst, the context as it is before an assertion is challenged. We are considering now a version of Non-ideal 2 to which Re exive Belief and Mutual Re exive Belief are added. The result is a case in which there is no longer a barrier from silencing to Postdoc’s speaking up. Since Non-ideal 2 is a case in which Postdoc does not speak up because she does not believe she has equal standing, adding those conditions will result in a case in which Postdoc does speak up (and in which other participants expect her to do so). As we saw in section 3, a case like Non-ideal 2 that di ers, at most, in Postdoc’s speaking up will be a non-ideal case of type 1. Adding Re exive Belief and Mutual Re exive Belief does not alter the conditions in Non-ideal 2 that give rise to silencing. A case that is

like Non-ideal 2 with respect to those conditions, in which Postdoc speaks up and her interlocutors expect her to do so, then, is still one in which she has been silenced. Non-ideal 1 cases exhibit the same sort of p. 91

assertional failure and conversational

breakdown the type 2 cases do. So, adding Re exive Belief and

Mutual Re exive Belief to cases like Non-ideal 2 does not su

ce for avoiding breakdown. Whatever else is

needed, it will be what is also needed to avoid conversational crashes in Non-ideal 1 cases. Let’s turn our attention to those cases, then. Our question is: What must be added to Non-ideal 1 cases in order to avoid breakdown from silencing? In those cases, Postdoc speaks up, but is not heard. What needs to be added, I’ll now suggest, is the persistent mutual belief that Equal Standing obtains. To see this, recall that, in those cases, Postdoc’s challenge does not receive a ‘fair hearing’—i.e., it is assessed along some unequal, non-epistemic dimension. She realizes this and is unconvinced by Sta Scientist’s assertion. This results in conversational break down in the characteristic way. Consider next the situation from the point of view of an additional participant: Graduate Student. By Re exive Belief, Graduate Student believes that she, too, has equal standing. By Mutual Re exive Belief, she believes that each other participant believes that they themselves have equal standing. So, she expects that a participant, Postdoc, for example, will speak up if they have an epistemic reason to challenge Sta

Scientist’s assertion. Suppose

in addition, though, that Graduate Student does not believe Equal Standing obtains. For this reason, when Postdoc’s challenge is ridiculed or ignored, she cannot be sure that it has received a ‘fair hearing’—i.e., that it has been assessed along equal, epistemic dimensions. In that case, Graduate Student may not come to believe what Sta

Scientist has said. The result would be conversational breakdown of our characteristic

type. However, were Graduate Student to believe and continue to believe that Equal Standing obtains after the response to Postdoc’s challenge, there would be no such barrier to her coming to believe what Sta Scientist asserts. There is nothing special about Graduate Student’s situation, however. What is true of her is true of each conversational participant. What, in addition, is needed to avoid such breakdown, then, is that each participant believes that Equal Standing obtains and continues to do so throughout the conversation. In other words, mutual belief in Equal Standing must be persistent. I’ll call this thesis “Persistence.” Persistence: To avoid conversational breakdown from silencing, participants must mutually believe that Equal Standing obtains (at least, to a close enough approximation) and that mutual belief must persist throughout their conversational exchange. This is to accord Equal Standing a signi cant role in explaining why conversational breakdown from silencing occurs when it does, and also why it does not when it does not. The role Equal Standing has been shown to play in explaining the di erence between our Ideal and Non-ideal cases thus far is indirect: The explaining proceeds via participants’ mutual belief that it obtains, rather than via its obtaining. Consider now the relationship between Persistence and Equal Standing itself. A conversational situation in which Equal Standing obtains is one in which there is no silencing. That’s because, as we’re understanding it, silencing is the failure to accord a participant an ordinary speaker’s p. 92

standing to update the conversational record. Persistence is strictly consistent

with Equal Standing’s

failure to obtain. This means that removing the barriers to breakdown from silencing is strictly consistent with the conditions that give rise to silencing. Next, we want to investigate how common we should expect these strictly consistent conditions to obtain. Consider rst a hypothesis I’ll call “Responsive.” Responsive: Belief that Equal Standing obtains is responsive to whether or not it does obtain. If Responsive is correct, then it will typically be the case that situations in which there is persistent mutual belief that Equal Standing obtains will be situations in which it does obtain. And if that is correct, then,

General Requirement: Avoiding breakdown from silencing generally requires Equal Standing. Here’s the structure of these considerations: (a)  Responsive means that Persistence requires Equal Standing. (b) Avoiding breakdown requires Persistence. (c)  Responsive. (Supposition.) (d) Therefore, General Requirement. What remains is to show that Responsive is quite plausible. To do that, we consider as a type, the situations in which Persistence obtains, but Equal Standing does not. These will be cases in which Responsive fails— interlocutors’ persistent mutual belief that Equal Standing obtains is not responsive to its failure to obtain. These will be situations in which one or more participants have been silenced, but no conversational participant appreciates this. Let’s consider such a situation, focusing again on the conversational perspective of a target of silencing. Consider again a case in which Postdoc speaks up, but is not heard, as when she is ignored in Non-ideal 1b. Unlike in that case, however, imagine now that she does not realize that she’s been silenced. For example, perhaps she is prone to extreme self-doubt, which causes her to question what she believes partly because she believes it and others don’t. In such a case, she may mistakenly defer to Sta

Scientist and come to

believe what he’s said. Though such deference is suboptimal for the goals of inquiry, so long as her interlocutors likewise don’t realize she’s been silenced, her silencing won’t lead to its characteristic conversational breakdown—the Common Ground will still represent the stock of mutual belief. Since it is her failure to believe that she has been silenced that plausibly prevents conversational breakdown in such a case, it, too, ts with our central hypothesis: Whether or not silencing plays a role in conversational breakdown depends on whether or not Equal Standing is mutually believed to obtain. Our question is whether that mutual belief can be reasonably expected to persist in a conversation in which Equal Standing does not obtain. p. 93

I’ll now suggest that there are reasons to think that in many cases, Equal Standing itself plays an important role in avoiding the conversational breakdown characteristic of silencing. As we’ll see, that role is in signi cant respects analogous to the role Grice’s Cooperative Principle plays in making communication possible. This will be because Responsive is, indeed, plausible. To begin thinking about this issue, consider rst a few di erent roles we might think cooperation plays in making communication possible. First, we might think communication requires cooperation itself. Second, we might think that communication requires mutual belief that there is cooperation. Or nally, we might think that it requires both. Communication with assertion is a speaker’s attempt to get an audience member to believe what she’s said. For communication to be successful, an audience member, M, must have reason 25

to accept what a speaker, S, has said.

To see this, suppose that M does not. In that case, she will not accept

what S has said. But if M does not accept what S has said, then S has failed to achieve the characteristic aim of speaking. To have a reason to accept what a speaker says, an audience member will need to think that the speaker aims, with her utterance, to advance a mutually accepted discourse goal. But that is just to say that she will need to believe that the speaker is cooperative. Equally, for S to accept M’s a

rmation as indicating

acceptance, S will need to believe that M is likewise cooperative. So, for communication to be possible, 26

conversational participants will need to mutually believe that each is cooperative.

So far, we have reasons to think that successful communication requires mutual belief in cooperation. Does it also require that, that belief is true? In short, one-o

conversations, successful communication may be

compatible with a temporary lapse in cooperation, as when a stranger deliberately o ers misleading

directions to someone who is lost, making their situation worse. In longer or iterated conversations, failures to cooperate undermine the mutual belief in cooperation communication requires. Similarly, widespread cooperative lapses in short, one-o

exchanges would also undermine the trust required for communication.

So, although some forms of minimal communication may occasionally be possible absent actual 27

cooperation, it is generally required for communication to occur.

Equal Standing plays a somewhat similar role in contexts of inquiry. As we’ve seen, perlocutionarily successful assertion requires mutual belief that Equal Standing obtains. In a short, one-o

conversation,

successful assertion may be compatible with its failure, as in the case in which Postdoc defers to Sta Scientist, though he has given her no good reason to do so. But, in the way that it does not take long to recognize an uncooperative interlocutor as such, it generally does not take long to realize that one has not been accorded that same standing as others to contribute to a conversation. This is because the deprivation of such standing—manifesting, as it does, in a pattern of being ignored, ridiculed, or otherwise explicitly discouraged from speaking up—is not di

cult to spot. Responsive is plausible, in other words, because, like

failures of cooperation, failures to be accorded equal standing are facts that over time are not di

cult to

appreciate. This means that in longer or iterated conversations, failures of Equal Standing will undermine the mutual belief that it obtains. In other words, Persistence requires Equal Standing. p. 94

So far, I’ve argued that successful joint inquiry requires the mutual belief that Equal Standing obtains. The persistence of that belief in iterated discussions or discussions of any reasonable length generally require that Equal Standing obtains, at least to some good approximation. This suggests that avoiding our characteristic conversational breakdown in contexts of inquiry generally requires that Equal Standing obtains, i.e., General Requirement. However, at this point, even one who grants that the type of conversational breakdown of interest here is not best understood as a failure of cooperation might still wonder whether some other explanation isn’t available. I’ll address this question in the next section by showing that conversational standing gures in explanations both of the conversational signi cance of silence and of its update e ects on contexts in a broader range of cases than those that exhibit silencing. Thus, we have independent reasons to accord conversational standing a role in explaining the dynamics of silence that is quite independent of the need to explain cases of silencing. In this way, according it that explanatory role is not ad hoc.

5. Modeling Conversational Standing I: Reasons to Update the Scoreboard Model “Conversational standing,” in the sense reserved here, is an ability to alter the conversational scoreboard— to o er and have accepted conversational “moves” and to accept or reject the moves of others. As a status 28

that is conferred on participants in a conversation by participants, it is a social power.

In section 6, I’ll

consider how the role standing plays in conversational dynamics might best be modeled. But here, I address the question of whether we have any independent reason to think that conversational standing as such plays a role that needs modeling. To begin exploring this, recall the two roles played by contexts of utterance in explaining communication. First, there is the role contexts play in determining the content or, more broadly, the signi cance of an utterance. For example, we saw in section 2 that whether or not a proposal to update the Common Ground is an assertion or a conjecture depends upon a conversation’s discourse goals. Thus, discourse goals are an element of non-defective contexts of utterance, which earn their place on the scoreboards which represent them in part by guring in explanations of an utterance’s force. Second, the notion of a context also plays a role in explaining an utterance’s e ect on a conversation. In this second role, contexts are that which is updated when an utterance has been accepted. For example, the

acceptance of an assertion updates what is presupposed for the purposes of conversation. For these reasons, we posit a Common Ground as an element of non-defective contexts of utterance. p. 95

These considerations suggest that guring in plausible explanations of how speakers are able to track the signi cance and e ect of an utterance serves as good grounds for positing a feature of conversational situations as an element which speakers track in order to do so. Given that conversational scoreboards as we are understanding them here represent the elements of context so understood, this is just to say that playing such a role su

ces for positing a feature of such situations as an element on conversational

scoreboards. An independent test for whether standing plays the conversational role proposed here, then, will consider whether standing gures in explanations of this kind. To begin exploring this, rst consider again our paradigmatic case of silence due to silencing—Non-ideal 2. Because Postdoc lacks the standing to successfully challenge or reject Sta

Scientist’s assertion, her silence

does not indicate acceptance of what he’s said. Similarly, in Stalnaker’s lecture example, it is because audience members lack (and are known to lack) permission to interrupt the lecturer that their silence does not signal assent to all that she asserts. This suggests that by tracking comparative conversational standing, participants are able to track when silence does and does not express acceptance of another’s assertion. Thus, an interlocutor’s conversational standing gures in explanations of the signi cance of her silence. Similar considerations apply to other sorts of speech act. Consider, for example, requests or imperatives. Ideally, silence signals an addressee’s acceptance of a request or a directive. In the literature on imperatives, 29

acceptance is modeled as an addition to the addressee’s ‘to do’ list.

As I think of them, to-do lists are

elements of a scoreboard that represent each participant’s commitments. A commitment to φ-ing makes one open to criticism for failure to φ. In a non-ideal case, where an addressee lacks the standing to refuse a request, her silence does not indicate acceptance and so does not saddle her with a commitment to comply. In this case, too, then, the signi cance of an interlocutor’s silence is explained in part by her conversational standing. As we’ve already seen in contrasting Ideal 2 with Non-ideal 2, conversational standing also has a role to play in explaining the update e ects of silence. It does as well in Stalnaker’s lecture example. Because audience members lack the standing to reject what the lecturer says during the lecture (and because they are known to), an audience’s silence does not add the content of the lecturer’s assertions to the Common Ground. Putting these considerations together, we see that conversational standing does indeed pass our independent test for positing it as an element of conversational situations interlocutors need to track in order for communication to be successful. Our focus here is on the conversational signi cance of silence. But similar points are plausible for overt acts of assent. When a participant lacks the standing to reject an assertion or request, her overt verbal a

rmation does not signal her assent. In the case of assertion, her verbal a 30

in the content asserted. p. 96

Nor, in the case of an imperative, does her verbal a

rmation does not signal belief rmation signal a commitment

to comply. (She does not thereby become criticizable for failure to comply.) This suggests that the signi cance of overt a

tracking

rmation likewise requires tracking participants’ comparative conversational

standing. Moreover, if we accept that which speech act an individual has performed depends at least in part on her intentions and also, with Stalnaker, that “an intention requires the belief that there is at least a reasonable 31

chance of success,”

then the deprivation of conversational standing will also gure in the explanation of

certain cases of illocutionary silencing. Illocutionary silencing is the deprivation of the ability to perform a speci c speech act (see Langton 1993). Suppose in addition, we accept, as is plausible, that speech acts are characterized in part by their perlocutionary aims: an assertion with audience belief, a directive or request with a commitment to comply, etc. In that case, a speaker who knows that her conversational standing deprives her of any reasonable chance of perlocutionary success will be unable to form the intention

necessary to assert something or request something. When such a speaker is silent for these reasons, her comparative standing will gure in an explanation of her illocutionary silencing. That the notion of conversational standing developed here is able to gure both in explanations of that phenomenon and in the cases of perlocutionary silencing considered here is an advantage of the account. These di erent ways that a speaker may be silenced are united by being cases of silencing. It is an advantage of an account to identify a 32

single source for that commonality.

Together, these considerations give us good reason to think that conversational standing is a feature of conversational situations speakers track in order to track the signi cance and impact of an utterance (or of silence). Second, there is an additional, independent reason to think that relative conversational standing is an aspect of context that can be updated, as when the chair of a meeting opens the oor to discussion or recognizes a contribution from the oor. The e ect of such recognition is to alter the standing of an audience member from one who is not permitted to speak to one who is. More generally, changes in standing to contribute to a conversation are an important aspect of exchanges that we track whenever we adopt a norm of turn-taking in speech. If turn-taking were not a default rule often enough to make speaking worthwhile, communication would not be possible. It’s important to note the independence of several of these reasons to add conversational standing to Lewisian scoreboards. Some of the considerations raised above are quite independent of the claim that standing can gure in plausible explanations of silencing. For example, as we’ve seen, standing gures in a plausible explanation of the impact of the lecturer’s assertions on the scoreboard in Stalnaker’s example. It also gures in a plausible explanation of the signi cance of her audience’s silence. This shows how the explanation of silencing in terms of standing is not ad hoc, but part of a more systematic explanation of a broader set of phenomena of the kind we look to scoreboards to represent and contexts of utterance to explain. These considerations together provide good reason to think that comparative standing earns its keep as an element on Lewisian scoreboards and in an account of conversational dynamics.

p. 97

6. Modeling Conversational Standing II: Updating the Scoreboard Model Recall that here a Lewisian scoreboard represents the information available to conversationalists in nondefective contexts. “Information” here is non-factive. For example, a proposition that is false at the world of evaluation may nonetheless be a part of the Common Ground. In cases in which mutual presuppositions su

ce or approximately su

ce to make it so, e.g., with a conversation’s discourse goals, it can be assumed

that Lewisian scoreboards record accurately. Here I suggest how we might model the role comparative standing plays in conversational dynamics, borrowing aspects of the framework Gillian Russell (2019) has developed to model the dynamics of subordinating speech. Russell adds to the usual scoreboard a ranking of individuals or groups in terms of their comparative social status. Status in her sense confers on status-holders a set of permissions and requirements with respect to other status-holders, both in- and out-group. Individuals or groups ranked higher in the ordering enjoy greater permissions and fewer impermissions than those ranked lower. An individual’s collection of permissions and impermissions de nes her social powers. As she notes, one such 33

power is the power to update the conversational record. “conversational standing.”

This social power is what I am calling

I propose to represent such standing on scoreboards by an ordered pair of elements: the rst, a set of 34

individuals or groups, and the second, a ranking of its elements (⟨Γ,≤⟩).

When there are no comparative

rankings, this second element is represented by the empty set. Otherwise, rankings are preorders— transitive and re exive relations on a set. (“a ≤b” reads “a is at least as well ranked as b,” where “a” and 35

“b” are members of Γ. ) When Equal Standing is presupposed to obtain, that ranking will rank all 36

participants alike.

The addition of this pair of elements permits the modeling of various phenomena discussed here. First, a participant’s comparative conversational standing plays a role in determining the signi cance of her silence. When her standing su

ces to give her the power to reject conversational ‘moves,’ her silence

signals assent. When that standing does not, her silence does not. Stalnaker’s lecture example illustrates. The conversational prohibition on an audience member’s interrupting the lecture partly explains why their silence does not signal agreement with what the lecturer asserts. Second, it allows us to represent changes in such permissions and prohibitions and so changes in the conditions for when a participant’s silence does and does not signal assent, as when the chair opens the oor to discussion. These changes in standing, in turn, can change the signi cance of an audience member’s silence (as when she has been given permission to speak by the chair). The addition of this element to the scoreboard also allows us to represent the di erence between a well-run conversation and conversational breakdown. In the literature on contexts, conversational breakdowns of p. 98

various types are explained in terms

of defective contexts of utterance. Such defects are represented by

defective scoreboards. For example, on a non-defective scoreboard, there is a Common Ground that represents what is mutually presupposed for the purposes of conversation. But it is possible for participants to be mistaken about what each of them jointly presupposes. In cases in which presuppositional di erences 37

make a di erence to interpretability, there will be conversational breakdown.

Similarly, the conversational breakdown found in our non-ideal cases may be explained by a defective ‘score’ for conversational standing. When a participant knows she has been silenced, for example, but at least some of her interlocutors believe she has been accorded equal standing, the signi cance of her silence will not be uniformly interpreted. This is what explains the type of conversational breakdown such cases exhibit. It is also what gives rise to the Predicament of Inquiry. Positing a ‘score’ for comparative conversational standing allows scoreboards to represent the source of the defects in such cases in just the way that a ‘score’ for what is mutually presupposed allows the Common Ground to represent the source of other sorts of failure of interpretability.

7. Conclusion Let’s sum up the lessons here with a return to Yelling Colleague. Here I’ve argued that silencing of the type illustrated in that case is not only a form of epistemic injustice. In addition, such cases exhibit two related sorts of communicative failure, one assertional; the other, conversational. First, such silencing prevents challenges from being raised or properly assessed. This can deprive silenced individuals of su

cient reason to believe the proposition asserted. As such, silencing is a form of practical

irrationality. In silencing his marginalized colleagues, Yelling Colleague undermines a condition necessary for the perlocutionary success of his own assertions. Second, when a conversational participant does not come to believe what has been asserted in a context of inquiry because she has been silenced, the Common Ground no longer represents what is mutually believed. As a result, the context becomes defective in a way that undermines the participants’ ability to achieve their discourse goal of inquiry.

The explanation for conversational breakdown in our cases, I’ve argued, is silencing itself. This means that the repair for such breakdown is a repair to the conditions that give rise to silencing. An individual is silenced relative to a conversation when she has been deprived of the ordinary speaker’s ability to contribute. In Contexts of Inquiry, she is deprived of the ability to add to the common stock of belief or to reject proposals to add to it. In the philosophical literature on testimony, this is characterized as a form of epistemic injustice. In contexts in which demands are made of her, her refusals or her own demands may fail to be taken as such for the purposes of the conversation. In a worst-case scenario, this may leave her vulnerable to predation. p. 99

Silencing is avoided in contexts in which Equal Standing obtains. When it does, participants in a conversation are each equally able to contribute. In COIs, this amounts to an equality in the ability to have one’s assertions (or rejections of assertions) assessed equally on epistemic grounds. The forms of communicative failure that result from silencing are avoided when there is persistent mutual belief that Equal Standing obtains. “Conversational standing,” in the sense reserved here, is an ability to o er and have accepted contributions to a conversation and to accept or reject the contributions of others. As a status that is conferred on participants in a conversation by participants, it is a social power. I’ve argued that in virtue of its explanatory role, such standing should be seen as a part of contexts of utterance in the theoretical sense of elements of a conversational situation that speakers track in order to track the ow of information. Importantly, the argument for the claim that standing does play the needed explanatory role includes considerations that are quite independent of its ability to explain silencing. In this way, conversational standing’s explanatory role is shown to be systematic across a broader range of phenomena, and not ad hoc. In terms of a Lewisian scoreboard, such standing is represented as a set, ranking pair, where the set may be made up of individuals or groups and the ranking is over the elements in that set. The ranking ranks those elements in terms of their ability to update the scoreboard. Those abilities, in turn, are represented as a participant’s bundle of permissions. Given the harm silencing does to those silenced, as well as the signi cant costs of conversational breakdown in many Contexts of Inquiry, a pressing normative question is how conversational participants might promote Equal Standing. That is a di

cult question that will not be answered simply by understanding how

language enables communication. How to make a start, though, is easy enough to see. The rst step will be self-awareness and an awareness of conversational patterns exhibited by one’s interlocutors. Is everyone given an opportunity to speak? Or is the oor dominated by the loudest person? Does each participant equally attend to the words of each of the other participants? Or are the contributions by some systematically ignored? Are contributions assessed equally on the basis of appropriate considerations, such as good evidence in the case of assertions? Or are some contributions assessed through a lens of inappropriate considerations, such as contempt or personal pride?

Notes 1.

Many thanks to Luvell Anderson, Kim Frost, Samia Hesni, and Mike Rieppel for discussion of the ideas contained here. Thanks also to the audience at the presentation of these ideas at the First Annual Meeting of the Philosophy of Language Association, June 10, 2022.

2.

For two good overviews of the literature on silencing, see the entries in Stanford Encyclopedia of Philosophy: “Epistemological Problems of Testimony” by Nick Leonard (2017); and “Feminist Philosophy of Language,” by Diaz Leo, Hesni, and Saul (2018).

3.

“Persons from a marginalized group” in the sense reserved here is anyone who is a likely target of discrimination or bias in

a particular situation on the basis of her membership in a historically marginalized or oppressed group. To illustrate, a Black professor might be a marginalized person at a non-HBUC (historically Black college or university), but not at an HBUC.

p. 100

4.

See, for example, Langton (1993); Maitra (2009); Dotson (2011); and Kukla (2014).

5.

This formulation is somewhat rough. It will be made more precise in section 4.

6.

See Roberts (2004).

7.

This use of “contexts of inquiry” is quasi-stipulative. A conversational situation will not count as a context of inquiry in the reserved sense unless the predominant conversational purpose is to become better jointly opinionated in the relevant way. Exchanges in such contexts may be governed by other mutually recognized constraints, e.g., temporal ones. But it will not be a context of inquiry if the overarching goal isnʼt to become as reasonably and mutually opinionated as interlocutors can, subject to those constraints.

8.

Here I focus primarily on assertion, although parallel observations apply to other forms of speech acts, such as requests or demands. For example, a perlocutionarily successful directive commits the addressee to comply. Under ideal conditions, we might suppose, an addresseeʼs silence in response to a directive su ices to generate such a commitment. (Here I think of such commitments as making one liable to criticism for failures to comply.) For further discussion of these issues, see section 5.

9.

See Langton (1993); Maitra (2009); Kukla (2014); and Hensi (2018).

10.

This gloss is mine. But for a similar gloss, see Adler (2017) and Diaz Leo, Hesni, and Saul (2018).

11.

This formulation is somewhat rough. It will be refined in section 4. The notion of conversational standing, le intuitive in this formulation, will be further explained in section 5.

12.

What Dotson calls “testimonial smothering” is an example of the type of phenomenon I have in mind: Testimonial smothering … is the truncating of oneʼs own testimony in order to insure that the testimony contains only content for which oneʼs audience demonstrates testimonial competence. … [This] kind of testimonial oppression occurs because the speaker perceives oneʼs immediate audience as unwilling or unable to gain the appropriate uptake of pro ered testimony. (Dotson 2011: 244)

p. 101

13.

Thereʼs an additional ideal case in which a challenge is mounted but unsuccessful, because it is itself met with a successful challenge. Since such cases o er no additional insight into the concerns here, they are omitted from the discussion.

14.

Itʼs important to bear in mind here that these cases are designed to illustrate silencing. A reader who believes that some element crucial to their ability to do so has been le out of their descriptions should feel free to add that element. Multiple di erent causes of silencing are recognized in the literature. The conversational e ects of silencing are the issue here, not its causes.

15.

Does it matter here whether or not Postdoc actually does have su icient epistemic reason to pose a successful challenge? No. Suppose she does not. If she did not, but believed that she did and had not been silenced, she would have stated her grounds for dissent. She would thereby have been given an opportunity to subject those reasons to joint scrutiny and an opportunity to become either convinced or to raise additional grounds for dissent. What is important in contexts of inquiry is that dissent is voiced and considered. Even if Postdocʼs dissent had been voiced, considered, and jointly set aside, it would still be possible for participants to track what is and is not jointly believed.

16.

Stalnaker (2014: 42). Plausibly, licensing the Stalnakerian assumption that silence is a form of acceptance requires that mutual cooperation is an element in the Common Ground.

17.

Stone and Lepore (2015). Even on a conventionalized account, common interest and common knowledge will figure in explanations of why speakers say what they do and why hearers accept what they say, even if those common interests donʼt play a role in determining which proposition an utterance expresses. More broadly, any view on which discourse

goals and the Common Ground are conventionally supplied elements of a scoreboard will be a view on which communication requires common presupposition and interests. 18.

Using Lab Meeting as our basic case, hereʼs an example of a conversational dictatorship: The dictator puts the rest of the audience “on notice” that supporting Postdocʼs challenge will be costly. If Sta Scientist were such a dictator, his failure to cooperate might in some sense fully explain the conversational breakdown. But it will do so insofar as it fully explains why she is silenced. Silencing here will be the most proximate cause, and so better explain the conversational breakdown.

19.

Philip Pettit has suggested that where there is freedom of speech, silence communicates assent; for example, he writes, The silent observer gets as close as makes no di erence to the position of meaning or communicating by her silence that she approves of what she observes …. [S]ilence in the presence of freedom of speech is itself capable of becoming a form of meaning and communication. (Pettit 1994: 49) As we can see from the di erence between Non-ideal 2 and Non-ideal 1a and 1b, however, the di iculty Postdoc faces is not that she is not free to say what she thinks. It is, rather, than she has not been accorded the standing to be heard. For this reason, freedom of speech does not su ice for silence to indicate assent or approval. See also Langton (2007) on Pettit (1994).

20.

Such a case would be an example of what Dotson has called “testimonial quieting” (see Dotson 2011: 242).

21.

Dotson (2011) introduces the term “pernicious ignorance” into the literature to denote “any reliable ignorance that, in a given context, harms another person (or set of persons).” Pernicious ignorance leads to epistemic injustice by “lead[ing] … an audience [to] fail to meet speaker dependencies in a linguistic exchange.” Thus, pernicious ignorance is one source of harmful silencing. An example of pernicious ignorance in Dotsonʼs sense would be a White personʼs refusal to consider the literature on implicit bias, insisting instead on her own “color blindness,” in the context of a discussion of fair hiring practices.

22.

For a discussion of stereotypes and their role in social cognition, see Wodak, Leslie, and Rhodes (2015).

23.

There are further cases that are not cases of silencing, but may nonetheless mimic them. These are cases in which the silent participant is put o speaking up because of some cost to herself of speaking up, not because, had she done so, she would not have been given a hearing. But if she has the standing to be given a hearing, she is not silenced. These cases can be explained as failures of cooperation. There is a continuum of cases of this type, though. Consider cases in which it is not only personally costly for a speaker to pose a challenge, but it is also costly for other group members to speak up so as to ensure that her challenge receives a hearing. A case like Ideal 1, but in which it is the PI who ridicules the Postdoc could be such a case. In such cases, at a certain point it is plausible that the costs of speaking up have reached a level at which the assessment of her assertion itself is no longer purely epistemic (in that it is also assessed, for example, on the basis of whether it advances anotherʼs personal conversational goals). At the limit, these are, in e ect, cases of conversational dictatorships, cases in which an individual is able to make the penalties for supporting a challenge so great that no challenge is conversationally assessed on epistemic grounds alone. These are cases of silencing.

24.

This is to say that these conditions su ice for avoiding breakdown from silencing in cases like Non-ideal 2. Thatʼs compatible with there being other reasons not to speak up, as Stalnakerʼs lecture example illustrates. For this reason, the condition discussed is necessary, but not su icient, for silence to signal assent.

25.

Having a reason here might simply amount to having no good reason not to accept what the speaker says.

26.

In a somewhat similar vein, Stalnaker (2014) writes,

p. 102

The cooperative principle is compatible with lots of conflict. It can be satisfied in contentious debates and negotiations as well as in simple exchanges of information. But for communication (trying to get people to believe things by meaning them) to be possible, there must be a recognized common interest in sharing certain information. (41–42)

27.

See also Stalnaker (2014: 41–43).

28.

Hesni (2018) also makes use of a notion she calls “conversational standing.” However, her notion di ers from the one here in that it is normative and conversation-independent.

29.

See Ninan (2005); Portner (2007); and Roberts (2018).

30.

Thanks to my colleague Luvell Anderson for this observation.

31.

Stalnaker (2014: 41).

32.

For compelling reasons to think that such cases should be given a unified treatment, see Hesni (2018). In addition, the explanation of illocutionary silencing provided here does not, in contrast to some accounts (e.g., Kukla 2014), have the consequence that which speech act an agent performs is not in any sense under her control. On Kuklaʼs (2014) account, whether or not an utterance is the speech act a speaker intends depends upon whether it is recognized as such by her addressee. Here, in contrast, whether an utterance is a speech act of the type the speaker intends depends in part upon those speaker intentions. For a nice discussion of why an account of illocutionary silencing that does not depend upon an addresseeʼs uptake is preferable, see Hesni (2018).

33.

Russell (2019).

34.

One way to understand the proposed ʻadditionʼ to Lewisian scoreboards defended here is by thinking of comparative conversational standing as part of the Common Ground. For these reasons, the proposal here is compatible with the overarching Stalnakerian framework for modeling conversational dynamics in terms of a Common Ground.

35.

a is ranked at least as well as b, roughly, when, for every conversational permission b has, a has that permission as well. a is ranked better than b when b is not correspondingly ranked at least as well as a, i.e., when a has permissions b lacks. Stalnakerʼs lecture example illustrates. During her talk, the lecturer is ranked strictly better than her audience members, who lack the permission she has to update the conversational record.

36.

Such egalitarian rankings will be total and symmetric.

37.

For a discussion of the di erences between defective and non-defective contexts and between harmless and nonharmless defective ones, see Stalnaker (2014: 84–86).

p. 103

References Dotson, Kristie. 2011. “Tracking Epistemic Violence, Tracking Practices of Silencing.” Hypatia 26 (2): 236–257. Google Scholar WorldCat Grice, Paul. 1989. “Logic and Conversation.” In Studies in the Way of Words. Cambridge, MA: Harvard University Press, 22–40. Google Scholar Google Preview WorldCat COPAC Hesni, Samia. 2018. “Illocutionary Frustration.” Mind 127 (508): 947–976. Google Scholar WorldCat Kukla, Quill. 2014. “Performative Force, Convention, and Discursive Injustice.” Hypatia 29 (2): 440–457. Google Scholar WorldCat Langton, Rae. 1993. “Speech Acts and Unspeakable Acts.” Philosophy and Public A airs 22 (4): 293–330. Google Scholar WorldCat Langton, Rae. 2007. “Disenfranchised Silence.” In Common Minds: Themes from the Philosophy of Philip Pettit, eds. G. Brennan, F. Jackson, and M. Smith, 199–215. Oxford: Oxford University Press). Google Scholar Google Preview WorldCat COPAC Leonard, Nick. 2017. “The Epistemological Problems of Testimony.” Stanford Encyclopedia of Philosophy (Spring 2023 Edition), Edward N. Zalta and Uri Nodelman (eds.), https://plato.stanford.edu/archives/spr2023/entries/testimony-episprob/. Google Scholar Google Preview WorldCat COPAC MacKinnon, Catherine. 1987. “Not a Moral Issue.” In Feminism Unmodified. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Maitra, Ishani. 2009. “Silencing Speech.” Canadian Journal of Philosophy 39 (2): 309–338. Google Scholar WorldCat Ninan, Dilip. 2005. “Two Puzzles about Deontic Necessity.” In New Work on Modality. MIT Working Papers in Linguistics 51, eds. J. Gajewski, V. Hacquard, B. Nickel and S. Yalcin, 1–26 Cambridge, MA: MIT Press. Google Scholar Google Preview WorldCat COPAC Pettit, Philip. 1994. “Enfranchising Silence: An Argument for Freedom of Speech.” In Freedom of Communication, eds. Tom Campbell and Wojciech Sadurski, 45–55. Aldershot, UK: Dartmouth Publishing Company. Google Scholar Google Preview WorldCat COPAC Portner, Paul. 2007. “Imperatives and Modals.” Natural Language Semantics 15:351–383. Google Scholar WorldCat Roberts, Craige. 2004. “Context in Dynamic Interpretation.” In Handbook of Pragmatics, eds. Laurence Horn and Gregory Ward, 197–220. Oxford: Blackwell. Google Scholar Google Preview WorldCat COPAC Roberts, Craige. 2018. “Speech Acts in Discourse Context.” In New Work on Speech Acts, eds. Daniel Fogal, Daniel W. Harris, and Matt Moss, 317–359. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Russell, Gillian. 2019. “Subordinating Speech and Speaking Up.” In Oxford Studies of Philosophy of Language, vol. 1, eds. E. Lepore and D. Sosa, 178–207. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC

Saul, Jennifer, Esa Diaz-Leon, and Samia Hesni. “Feminist Philosophy of Language.” Stanford Encyclopedia of Philosophy (Fall 2022 Edition), Edward N. Zalta and Uri Nodelman (eds.), https://plato.stanford.edu/archives/fall2022/entries/feminismlanguage/. Google Scholar Google Preview WorldCat COPAC Stalnaker, Robert. 2014. Context. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Stone, Matthew, and Ernie Lepore. 2015. Imagination and Convention. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Wodak, Daniel, Sarah-Jane Leslie, and Marjorie Rhodes. 2015. “What a Loaded Generalization: Generics and Social Cognition.” Philosophy Compass 10 (9): 625–635. Google Scholar WorldCat

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

5 The Pragmatics of Technologically Mediated Online Speech: “Don’t @ Me!”  Quill R. Kukla https://doi.org/10.1093/oxfordhb/9780192844118.013.5 Published: 22 May 2024

Pages 104–122

Abstract This chapter explores the social force and pragmatic structure of an emerging new class of speech acts —including, for instance, ‘liking’, retweeting, and ‘at’-ing—that have only recently become possible to perform. New technology and communication infrastructure is generating the possibility of new pragmatic forms of speaking, which can have no face-to-face equivalent. Standard pragmatic types of speech acts—interrogatives, promises, imperatives, and so forth—are essentially mediumindependent; they can be performed with one’s voice, in writing, over electronic media, in sign language, etc. But as the chapter will try to show, other pragmatic speech-act types are inherently medium-dependent and technologically mediated. It will defend two theses: (1) the technologically enabled online speech acts it explores have distinctive pragmatic structures, which are irreducible to speech-act types performed over other media or in person; and (2) many of these online speech acts are socially potent: they serve to constitute and organize social space in important ways. In particular, they are used to negotiate and solidify the boundaries between group insiders and outsiders and the norms of social engagement within groups

Keywords: online speech, pragmatics, insider speech, social groups, language and technology Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction This chapter explores the social force and pragmatic structure of an emerging new class of speech acts that have recently become possible to perform. These are speech acts that can occur only on social media and in electronic form. I will claim that that new technology and communication infrastructure is generating the possibility of new pragmatic speech-act types, which don’t have any easy face-to-face equivalent. Using the tools of speech act theory, I will identify and analyze several new kinds of speech acts, with an eye to how they shift and shape social space. I will try to show that many of these online speech acts are socially potent: they serve to constitute and organize social space in important ways. In particular, they are used to negotiate and solidify the boundaries between group insiders and outsiders and the norms of social engagement within groups. They also create new kinds of speaker authority, allowing speakers to control social space in new ways. There are all sorts of things that one can do online with speech other than simply convey semantic content or directly mimic standard o

ine speech acts. Gretchen McCulloch (2020), for example, argues that emojis

and some acronyms, like LOL, function more like gestures than like carriers of content; they operate on other bits of speech to x tone, mood, and salience. My interest is in a speci c subclass of the forms of online communication. I am interested in identifying speech acts we can perform online that add expressive p. 105

power to language by allowing us to perform types of speech

acts with a distinctive pragmatic structure,

where the possibility of this pragmatic structure is enabled by the material technology of social media platforms themselves. I am especially interested here in the distinction between speech act types that are medium-independent and those that are medium-dependent. Familiar kinds of speech acts—interrogatives, promises, imperatives, and so forth—are medium-independent; they can be performed with one’s voice, in writing, over electronic media, in sign language, etc. When speech acts are medium-independent in this way, we can often establish that we are performing them (assuming we have the proper authority and felicity conditions) just by using words to announce that we are doing so. We can say, “I am ordering you to close the door,” or “I am asking you whether you are coming tonight,” and thereby determine that the speech act is an order or an interrogative. We can use explicit language to di erentiate between hard to separate speech-act types: For instance, I might say, “Will you talk through this paper with me tonight? I am just asking as a favor, not demanding that you do it.” In these cases, the words used lock down which speech act it is. But other speech acts depend for their performance on the materiality of how they are performed. I cannot prescribe a drug just by saying drug quantities, even if I add, “I am prescribing”—I have to write them on the right kind of o

1

cial paper or form, for example. Online speech-act types of the kind I care about here are materially

di erentiated and medium-dependent, as we will see. No grammar or explicit description will make them the speech act they are—they need to be performed with the right material technology.

2. A Framework for Analyzing the Pragmatics of Speech Acts Beginning in the middle of the twentieth century, Wittgenstein and Austin (among others) transformed the philosophy of language by teaching us that we do more with speech than simply transfer information. Speaking is an action, and through the act of speaking we structure the social world in a variety of concrete ways by shifting norms. We can command, invite, entreat, persuade, arrest, sentence, call to order, warn; these actions are accomplished in the act of speaking. Speech can perform such social transformations only when a speech act is felicitous—that is, it is performed in the right context, following appropriate conventions, and the speaker has the proper authority to perform it. A professor can felicitously command a student to turn in their assigned paper. This is so

because of the norms that govern the context of a university classroom, and the professor’s position as an authoritative speaker of a certain sort, with respect to the student, in that context. It would be infelicitous for the student to command the professor to submit a paper in return; this order would be not merely inappropriate, but nonsensical given the context. All speech acts have felicity conditions of this sort. p. 106

‘Yo!’ and ‘Lo!’: The Pragmatic Topography of the Space of Reasons (Kukla and Lance 2009) built a theoretical framework for analyzing the pragmatic structure of speech acts that can be layered on top of classic speech act theory, and I will draw here upon that framework, as well as on the extended version of that framework developed in Herbert and Kukla (2016). In this framework, speech acts perform normative functions that can be categorized by the personal and contextual conditions that entitle them to be performed, and by how they shift normative statuses in social space. Their entitling conditions are their ‘input,’ and the normative changes they bring about when they are successful are their ‘output’. Both the input and the output of a speech act type can be agent-neutral, agent-relative, or community-speci c. Speech acts with agent-neutral outputs speak impersonally, calling for the same normative uptake from anyone who might happen to hear them. Newspaper headlines and impersonally stated declaratives are paradigmatic examples. Although di erent hearers will, of course, respond di erently to such speech acts, they are not designed to impose di erentiated normative statuses on di erent people. In contrast, imperatives, requests, and promises are examples of speech acts with agent-relative outputs: When I request a favor, for instance, I do so of someone(s) in particular, and in doing so, I call for a speci c transformation in my target’s normative status; they may turn down my request or grant it, but either way, they are now in the distinctive position of having had a request made of them. In contrast, a passerby who overhears my request is under no pressure to turn it down or grant it; indeed, for them to do so would be a kind of 2

infelicity. Speech with a community-speci c output, as a matter of its pragmatic structure, di erentially targets the normative status of members of a community qua community members. For instance, legislative speech speci es a rule that applies to anyone under the jurisdiction of the law, whoever they happen to be. It does not target any individuals in particular, so its output is not agent-relative, nor does it have any force 3

for an overhearer from another jurisdiction, so its output is not agent-neutral.

Likewise, the inputs of speech acts—that is, the normative statuses that enable a speaker to perform a speech act—can be agent-neutral (available to anyone regardless of their particular authoritative position); or they can be agent-relative (indexed to a speci c authoritative position); or they can be communityspeci c (felicitous only when performed by insiders qua members of a community). Anyone can assert that Paris is the capital of France, even though a given speaker may not actually have the warrant to back up that claim. There is nothing about the act of assertion that indexes it to anyone in particular, so it has an agentneutral input. But only the owner of an object can felicitously o er it up for sale; only the ones who will be having sex can consent to that act, and so on; these speech acts have agent-relative inputs. Speech acts with community-speci c inputs are, in e ect, spoken in the plural rst-person, and entitled by the authority of a group: “We have decided that you can join our sorority” would be an example. Speech acts with community-speci c inputs may be performed by individuals speaking as community representatives, who are entitled in virtue of their community membership, or by a collective or plural speaker. p. 107

Some speech acts are community-speci c in their output in that their normative impact is indexed to a particular pre-existing community of insiders. A legislative speech act typically transforms the normative status of people who are in the well-de ned, settled community of those to whom the laws apply. But speech acts can also create or negotiate the boundaries of a community, not just re ect them. For example, when a dissertation director says “Congratulations, Doctor Singh!” to mark a successful dissertation defense, they are speaking as a member of the academic community and simultaneously welcoming a new member into that community, changing its membership. And a declaration annexing a suburb to a city creates a new community with larger boundaries than the original city, with formal and informal repercussions.

Even more strongly, a speech act may help to establish a community in the rst place, working to delineate its boundaries or a character, and thereby helping constitute the possibility of community-speci c speech acts indexed to that community. Consider, for instance, jokes about millennials or the expression “OK Boomer!” Although these lose their constitutive power after being repeated enough, part of the work they do at rst is not just to re ect the character of a generation but to delineate a generation as a community with a speci c character (often by contrast with one being mocked), and to negotiate who counts as an insider and who counts as an outsider. Memes that draw upon a speci c aesthetic sensibility, sense of humor, or set of cultural references may be used to help pull together a community identity and to create insiders by speaking to them. Hence, speech with a community-speci c output may do constitutive work of creating its audience. Later in the chapter, we will see how some kinds of electronic speech can also function to create a community on the input end; they help generate a kind of collective or plural speaker who speaks as a ‘we’. Technology, it turns out, can generate not just new kinds of speech acts but new kinds of speakers. Agent-relative, agent-neutral, and community-speci c inputs can each be combined with any of the three kinds of outputs in any given speech act. So, for example, the 1980s chant, “We’re here, we’re queer, get used to it!” has a community-speci c input, since it was only felicitous as performed by insiders, speaking as members of a meaningful group. But its output was pointedly agent-neutral: it was anyone and everyone, regardless of their normative position, who were being told to get used to it. It certainly wasn’t aimed solely at community members, who were, presumably, already used to it. On the other hand, declaring an o

cial

war, for instance, may have an agent-relative input; only certain people with distinctive social power may have the authority to declare a war. But the declaration has community-speci c outputs, because now everyone in the nation that has declared war (and everyone in the nation against whom war has been declared) is at war. Thus, we have, in e ect, nine possible categories of speech acts, because each speech act may be agentrelative, agent-neutral, or community-speci c in its input, and this can be combined with its being agentrelative, agent-neutral, or community-speci c in its output. (Also, single utterances typically have multiple speech-act functions, so the situation is even more complex in practice.) Going forward, I will analyze p. 108

speech acts partly

by placing them within this space of possibilities, teasing out their inputs and outputs

in order to pin down how they function as social interventions. I’ll use this framework to place and to analyze the pragmatic structure of a series of di erent kinds of technologically mediated speech acts on social media. In what follows, I will examine ve kinds of technologically mediated speech acts: (1) liking ( -ing or ♥ing) a post on Facebook or X/Twitter; (2) tagging or ‘@-ing’ someone on Facebook or X/Twitter; (3) retweeting or sharing someone else’s public post; (4) tagging a tag group on Facebook; and (5) hashtagging on X/Twitter. Before I begin, I need to make clear that I am not giving an exhaustive account of how these technologies can be used in communication. I am analyzing speech-act types that essentially use these technologies, not the technologies themselves. I want to show common types of pragmatic speech act categories that have been generated by the technologies. People can always use the technologies to do something di erent. Indeed, as soon as a new language norm emerges, people tend to be creative about ri

ng on it and

subverting it, and we see this online too. I will point out these nonstandard ri s here and there, but there is always room for new ri s and new uses. Thus, my medium-dependence claim is that the technologies I will be examining are necessary, or at least integral, to the speech act types they support—they make new speech-act types possible—not that they are su

cient to lock down the speech act types that use them.

3.

-ing or ♥-ing

“Liking” a Facebook post with a

4

, or “loving” a tweet with a ♥ (I will refer to both as acts of ‘liking’) are

semantically minimal but pragmatically complex speech acts. On Facebook, a small number of other reactions are possible. It would be worth an entire paper, I think, to analyze the semantic and pragmatic e ects of o ering people a selection between seven preset emotional reactions. There is an interesting capture of our emotional expressive possibilities and interpellation into a attened set of emotional responses that goes on here, and an interesting social grammar for when to use which reaction. That said, I will focus here only on the ‘like’, which especially interests me because of its maximum semantic exibility. Semantically, these speech acts have little stable content other than to communicate a pro-attitude of some sort or other. The kind of pro-attitude expressed varies widely: The approval could be of the content of the original post or tweet, or of the fact that the person posted it. You may ‘like’ something because it makes you feel good, because it provides cutting political critique, because you are happy for the original poster, because you approve of their presumed reasons for posting it; because you hate the content but like that the poster brought it to your attention, and so forth. You might ‘like’ something as a minimal act of etiquette if it expresses good news for the poster. So the semantic content of a ‘like’ is very thin and indeterminate. The ‘like’ functions as a kind of positive reactive attitude tool, but it conveys no more stable semantic content than that. p. 109

The input of the ‘like’ is inherently agent-relative, as a matter of technological necessity and pragmatic force. Unless someone hacks your account (in which case they are fraudulently speaking as you anyhow), you are the only one who can ‘like’ things on your own behalf; the act of ‘liking’ is inherently tethered to the liker. There are some acts that are inherently individuating, in that they can only be performed by their individual owner: Only I can express my own pain; only I can die my own death; only I can ‘like’ social media posts on my own behalf. Someone else might comment, “Quill likes what you said here!,” but pragmatically 5

and technologically, this is a categorically di erent act from my pushing the like button.

The output of the ‘like’ is a complex hybrid of an agent-relative and a community-speci c output. On the one hand, the ‘liking’ has a second-personal, agent-relative output: it speaks to the original poster, reacting approvingly at them; you ‘like’ the post of someone in particular, and your liking is directed to them. On the other hand, this act of reacting is, by nature, visible to everyone in the original poster’s circle, so in ‘liking’ one is also speaking to a curated community of social media users. (In the case of a public post or tweet, the ‘liking’ has an agent-neutral output.) Telling someone privately that you like something they posted is a di erent discursive act from ‘liking’ it for all to see. An interesting pragmatic feature of ‘liking’ is that it is a way of making a move in a conversation that doesn’t invite or even easily technologically allow a response. It can be used as a way of acknowledging that someone has spoken and accepting the legitimacy of their speech act (while not necessarily agreeing with it), without adding anything to the conversation that calls for a response, partly because materially, it modi es the original post instead of adding a new one. It is a new speech act, but one that piggybacks on the materiality of the thing it is responding to. This kind of piggybacking is a distinctive thing we can do online that has no clear equivalent in regular speech. In normal conversation, speech acts call for some response; it is awkward and a bit of a norm violation to leave a speech act completely unacknowledged. For this reason, 6

conversations can actually be kind of hard to end! But because a ‘like’ is as close to a semantically empty acknowledgment as possible, and because it is attached physically to a post or comment and there is no easy technological way of directly responding to it, it serves as a smooth conversation ender if needed. This is useful as a way of wrapping up online conversations, especially since body language is unavailable. We may like a comment even if we disagree with it, simply to second-personally and publicly signal acceptance of it

as legitimate a move in the conversation, and to mark its call to us to respond without, in turn, calling for continuation of the exchange. There are some categories of posts—posts about having a new baby, getting married, getting tenure—that in e ect call for likes. Since no one knows exactly who sees these posts, it is not exactly a transgression not to like them, but social etiquette encourages giving them at least a like. These sorts of rules of etiquette help solidify our community norms for what sorts of lives and identities we are supposed to have. The ‘like’ speech act serves as a quick and minimal positive reactive attitude that tells someone that they are doing things p. 110

right, and tells the rest of the online community that this is how things

are supposed to go. Norms of

liking show everyone in the community what we celebrate, how we are supposed to behave, what the shape of our life narratives is supposed to be. They also show everyone in the community how and when it is appropriate to post about one’s life, and help shape the norms for curating a public identity and narrative. Because of its complex output, the ‘like’ as a reaction does double duty, communicating approval to the original poster but also communicating to the community how one is supposed to act, both online and o .

4. @-ing In this section, I discuss tagging a person on social media by way of the @ function, or @-ing (“at-ing”). This can be done in multiple platforms. When someone is @-ed, their name shows up as clickable, and they typically receive a noti cation that someone has mentioned them. On the surface, what an @ does is allow the person to see that they have been addressed or mentioned. The @ has no semantic content of its own, nor does it act as a semantic operator on the name of the person tagged; its e ect on the larger speech act in which it is embedded is purely pragmatic, and it can be embedded in any semantically contentful online speech act that can contain a name. As we just saw, ‘liking’ a post, because it cannot be directly responded to and because it is relatively free of de nite semantic content, serves as a handy conversation ender when needed. However, @-ing someone typically has the opposite e ect: it calls upon the person tagged to respond, continuing a conversation or drawing them into one. While being @-ed does not automatically obligate someone to respond—and while sometimes we don’t actually expect the person to respond—the structure of the speech act, including its physical structure, calls to the person who is @-ed and places at least some minimal normative pressure on them to respond. This is for three reasons. First and most simply, an @ is a second-personal speech act, and all second-personal speech acts call for 7

some sort of acknowledgment or response by default. This is true even when the @ is embedded as a mention in a third-personal speech act, because the @ part of the speech act is directed at the person named; indeed, this is built into its very name and typography. But second, the technology ensures that the person spoken to is noti ed in real time that they have been addressed, and the speaker’s knowledge that the one addressed knows that they have been addressed strengthens the call to respond. We @ someone when we don’t just want to speak to them, but want to ensure that they know they have been spoken to. Third, when someone is @-ed, because their name shows up as highlighted, everyone else can also see that they have been spoken to and called to respond. This, too, is a product of the technology, and it strengthens the call to respond. If the one @-ed fails to respond, everyone can see that they did not take up the call, and everyone knows that they were noti ed. Thus, people can use the @ function in social media to issue a p. 111

particularly intense second-person call for a response by doing so in

a way that they know will be

received and by displaying to everyone else that they are issuing a call, which everyone else also knows will be received. The most interesting feature of @-ing, in my view, is that it calls to someone, but in the public eye. In e ect, the @ serves as a technological means of pointing a speech act at someone, ensuring that the directed

address is received, and ostending in a public way that this has all happened. This public ostension means that anyone can hold the person @-ed accountable for responding. Just using someone’s name and publicly addressing them does not have the same e ect, because not only does it not guarantee that the address will be received, but it does not demonstrate to others, visually, that a call was issued and received. So @-ing has an agent-relative output, since it is by design directed at the one named. But it also has a communityspeci c or agent-neutral output (depending on the audience for the post), since it displays to everyone who can see the post that someone has been second-personally addressed. This means that @-ing is often an especially powerful form of address. Because of the publicity of the directed address, the @ makes felicitous a distinctive range of responses from the larger community if the one @-ed fails to respond, for example. Failure to respond can be given uptake by anyone in the community as rudeness or as not having anything to say. Because the one addressed has been noti ed that they’ve been addressed and everyone knows this, simply ignoring the call to respond becomes a socially loaded response that cannot conveniently be confused with mere ignorance that they were spoken to. Moreover, by doing so in a publicly displayed way and ostending the address for everyone, it involves the whole community in holding them accountable for responding. Many uses of @-ing are benign, and just re ect an honest desire to call someone’s attention to something or to call them into a conversation. And, despite their function as a second-person address, sometimes they can be safely ignored, such as when many folks are @-ed at once in an informational post. The strength of their pull to respond, and the extent to which they marshal the public eye to enhance this pull, varies greatly. However, quite regularly, @-ing is used to try to force someone to enter into or stay in a conversation; hence, the slang expression “Don’t @ me!” that is used to object to being called out to account for oneself in a public way. A common trend, at least among academics on social media, is to use @-ing to extract labor from members of a nondominant group. Someone will, for example, post an article and then ask what their ‘trans friends’ or their ‘Black friends’ think of it, or will ask them to explain something, tagging people from these groups. This puts the targets of the tagging in the position of either o ering free intellectual labor—which can easily come to include emotional labor if their responses are then attacked or treated with suspicion—or looking, in a highly public way, as though they are unwilling or unable to respond. The @ gives us a material means for performing this public second-person call to respond. In-person equivalents to this sort of public ostensive second-person call are hard to come by, but they exist. There are some ritual settings in which they happen: for instance, an o p. 112

marriage ceremony whether

ciant asking someone in the context of a

they take their partner to be their spouse. But notice that this pragmatic

structure is also medium-dependent. Both the @-ing case and the marriage ceremony case depend on their materiality for their distinctive pragmatic structure.

5. Tagging a Tag Group on Facebook This section is devoted to a much more playful speech act: tagging a tag group on Facebook. A tag group is a group that exists solely or overwhelmingly so that people can tag the name of the group as a comment on someone else’s post. When it is used as a tag, the name of the group is designed to function as an entire speech act. Thus by ‘saying’ the name of the group in the form of a tag, you tie your speech act to everyone else who has tagged the same tag group. To use the tag, one must be a member of the group, but the groups 8

exist almost entirely to support the tagging of itself.

Tag groups are designed to be funny. Some popular examples include the groups “Sounds like another man who needs his penis stolen by a witch”; “Literally nobody is saying that”; and “Marie Condo asks, ‘Does this

man spark joy?’ Alas no, into the pit he goes.” There are even meta-tag groups for performing tagged speech acts about the act of tagging tag groups, such as “u leave my tag group obsession alone”; “tag groups are my love language”; and “my therapist says I am not allowed to join any more tag groups.” The mere existence of these speech acts is fascinating for at least two reasons. First, since the groups they tag exist solely to enable the speech act of tagging them, this tagging is a technologically enabled speech act that truly can have no pragmatic equivalent o

ine. Second, tag groups are decidedly not a use of Facebook

that any of its designers foresaw (Lorenz 2019). They are an example of a discursive practice that is emerging from the bottom up through creative speakers’ uses of technology and developing its own norms 9

organically, in ways that are unpredictable, even from the point of view of the social media platform designers who gave us the technology that enables these speech acts. Moreover, this is a practice that adds expressive power to language that simply didn’t exist before, so it is a lovely example of how technology can enable us to do new things with words. Since one needs to belong to a tag group in order to tag it, the input of the act is community-speci c; quite literally, by tagging a tag group, you are performing an act that displays its own enabling condition of membership in a group. However, these are maximally thin communities; they exist only for the purpose of letting people display their membership in them through the tags. Hence this is a kind of formalistic limit case of a community-speci c input; one must belong to the community to perform the speech act (as a matter of material necessity), but there is nothing more to belonging to such a community other than gaining the ability to perform this speech act. What ties the group together is just that its members wish to have this conversational move available to them, and to make it look like one that is backed by a group. So, the group is constituted in its own membership displays. p. 113

As for the output, it typically has both agent-relative and community-speci c dimensions. People usually direct tags of tag groups at someone else in the conversation, but they do so in a way that displays that we feel this way about that person and their contributions. The whole di erence between simply writing out as a comment in response to an obnoxious, unnecessary comment from a man, for instance, “Marie Condo asks, ‘Does this man spark joy?’ Alas no, into the pit he goes” and tagging this group, is that the former shows up as an idiosyncratic response, whereas the latter directly invokes the existence of a large shadow community of people who are all ready to respond in this way to conversational moves of this sort. Moreover, the tag invites other people who see it to join that community and avail themselves of the tag in the future. The (playful) authority by which the performer of this type of speech act speaks is distinctive. The tag groups not only o er users a repertoire of witty responses, but using one of them performs a kind of authority that comes from speaking in a plural voice; the implication is that there is a mass of folks who have already united around this being the socially appropriate response to whatever it was that was being responded to. (Tag groups are never used, at least as far as I have seen, as conversation initiators.) The implication is that the one being responded to has said or done something of a sort so socially familiar that a whole group has established a shared appropriate response, and that the speaker is an insider in that group.

6. Retweeting and Sharing A tweet or public Facebook post can be retweeted or shared by anyone, with or without added commentary. Here I am interested in the discursive act of retweeting or sharing; I will use “retweeting” to mean both. Robert Brandom (1983, 1998) argued that assertions are structurally public, in the sense that they issue universal reassertion licenses; they o er up to anyone the option to make the same assertion, citing the authority of the original. As Brandom describes it, the reassertion is a re-performance of a type-identical speech act, although its justi cation will be di erent from that of the original. Regardless of whether

Brandom is right in all the details about how assertion works, notice that we have a pragmatic structure in online sharing that is both importantly analogous and importantly disanalogous with the one he describes. Tweeting or posting something publicly also issues a kind of universal public license—a license to retweet. Once a public post is out there, there are no material or normative constraints on who can retweet it; this is built into making it public. But the retweet cannot be assumed to be the same speech act as the original. Someone may retweet in their own voice or at a critical distance from what they are retweeting, and there are no codi ed rules for telling which is which. The retweet might be a straightforward reassertion or reperformance, but it might also be an act of satirization, fomenting protest, mockery, ampli cation, or any number of other discursive moves. And the retweet itself can in turn be retweeted, and it can shift p. 114

pragmatics with

each iteration. Semantically speaking, if no commentary is included, a retweet might be

asserting the content of a tweet, or negating it, or doing something else altogether. The semantic content of a retweet will depend on what we take its pragmatic force to be. So, without context, one cannot lock down the conversational move made by retweeting, either semantically or pragmatically. Interpreting a retweet properly requires knowing a complex combination of facts about who the speaker is and what else they typically post, the circles they travel in and what can likely be taken as shared background assumptions among those circles, and so forth. Because a retweet is not a simple reperformance, and because its semantics and pragmatics are contextually dependent, communication theorists have been especially interested in this kind of speech act, which has received more scholarly attention than the others I have discussed. Linguists have tried to uncover systematic patterns in what a retweet means or accomplishes; for instance, retweets without commentary 10

are more likely to signal agreement, although they do not always.

A retweet always has an agent-neutral input because it is built into the technology that anyone can retweet. But retweets are used for a variety of pragmatic purposes and perform various kinds of social work, and they have various kinds of inputs and outputs. For instance, retweets may be used to build up and curate an individual’s social media ‘brand’ or public identity; part of the point of retweeting may be to communicate that this is the type of content associated with the speaker. Speaking as brand-building is agent-relative in its input and agent-neutral in its output. Alternatively, a retweet may be part of a process of communitybuilding. A tweet or post may get passed around a community, where part of the point is not just to give access to its content, but to establish that this is the kind of thing that we share: this is our kind of humor, our kind of political commentary, the kind of thing we nd salient, the framework that we think is the right one for talking about an issue, or the kind of writer we like to amplify. Through a network of often-repeated retweets, a community can articulate by exemplars its shared identity. Here the output is communityspeci c. Relatedly, combining the two previous functions, someone might share a tweet or post to publicly demonstrate that they are part of or belong in a group—that they are the kind of person who posts the right sorts of things to belong. Sometimes the point of retweeting is not to signal what sort of content “we” approve of or mock, but to solidify internal group norms for how to communicate. Consider, for instance, memes or slang- lled jokes, which may be retweeted often within a community. Becoming familiar with “our” memes and slang can help carve out norms for insider speech and a canon of insider jokes. Memes, in particular, are built to be joke schemas that can be lled in with various jokes, which means that when a group embraces a meme as part of its insider speech, the meme can be a powerful generator of insider speech acts that rea

rm the

group’s cohesiveness, boundaries, and identity. It is interesting that there is any pragmatic point to repeated retweeting or sharing the same content within a subcommunity. Once a group begins to pass around a post, it quickly becomes redundant as far as the communication of its content is concerned. When many people in a group retweet or share the same thing, p. 115

the purpose becomes

decreasingly to communicate the content of the post, and increasingly to

demonstrate that this content expresses our identity, that it is important for people like us. The function of

the speech act in such cases is community-speci c in both input and output. The fact that retweeting can help de ne a group identity, group norms, and particular individuals’ insider status helps to explain why there is a social purpose to repeated retweeting; whereas we may nd this mysterious if we see speech acts as primarily functioning to deliver semantic information. Redundant group retweeting may also help to build echo chambers. Typically, we think of echo chambers as social environments in which a group only has access to a certain kind of semantic content because of builtin selection biases. The idea is that even a continent and curious observer inside an echo chamber will encounter only evidence that supports the kinds of views they already are disposed to endorse. Thi Nguyen (2020) distinguishes echo chambers from mere lter bubbles, adding the requirement that from inside the echo chamber, dissenting voices will be discredited simply in virtue of being dissenting; for example, within the anti-vaccination echo chamber, pro-vaccination voices are automatically seen as aligned with an untrustworthy government conspiracy, hence the information they provide is discounted. Retweeting can potentially contribute to echo chambers in this sense. As sources that t with a group’s sense of reality and identity are repeatedly retweeted, the repetition may bolster their apparent legitimacy and salience, even without their content being supported by other sources. Moreover, although few of us are in a position to know exactly how the algorithms for our social media feeds work, it seems plausible that a repeated tweet will have a larger impact on what other content we receive, so we will end up seeing similar information from other sources. As sheer repetition becomes a felt measure of legitimacy and insider status, any dissenting information that makes its way into the echo chamber automatically looks ‘fringe’ and outsidery. Helping to solidify an echo chamber is not likely to be the explicit goal of retweeters, but I suspect that it is a predictable e ect.

7. Hashtagging on X/Twitter The hashtag on X/Twitter is used to connect whatever symbolic string is hashtagged to all other instances of 11

that same symbolic string being hashtagged in other tweets, whether one’s own or other people’s.

So the

hashtag serves to tie a speech act together with other speech acts with which it shares a particular part. Hashtagging also makes a tweet searchable by way of the hashtag. Hashtagging is a pervasive and multifaceted practice, and people hashtag for a variety of purposes. Whatever the motive for using it, including a hashtag in one’s speech act has the interesting e ect of tying that speech act technologically to every other speech act that contains the same hashtag. Anyone can use a hashtag, so the act of hashtagging is agent-neutral in its input, although a hashtag can be embedded in almost any kind of online speech act. Furthermore, the output of hashtagging is also agentp. 116

neutral, in a distinctively strong

sense. The hashtag renders the text clickable, so anyone who clicks on

any instance of the hashtag will be taken to a place where they can see all other uses of the hashtag; this means that the output is for anyone in a concrete sense that doesn’t apply to regular in-person speech. The hashtag functions to allow anyone to nd the speech act regardless of whether it was directed at them or whether they were part of the original community in which it circulated. We saw that @-ing someone in a tweet or a Facebook post is a way of technologically ensuring that it has an agent-relative output; the @ points the speech act at someone, for all to see, and second-personally calls for recognition from them. In contrast, the # technologically enables the speech act to have an agent-neutral output. Interestingly, linguists have found that this contrast is re ected in the norms of tweeting: tweets with hashtags tend to be written in a more formal style appropriate for address to a broad audience, whereas tweets with @s tend to be written less formally and with more slang, appropriate to a more targeted audience (Pavalanathan and Eistenstein 2015).

There is an important sense in which, by hashtagging, one contributes to a kind of collective speech act consisting of all the tweets with that hashtag. This function can be used strategically, to display and expand the social weight behind a speech act, such as when people hashtag the name of a victim murdered by police, or hashtag a call for a strike. (This same function can also be playfully subverted, for instance, by creating an absurd and overly speci c hashtag that no one else will use.) Part of the power of the hashtag is that one need not announce the collective weight and authority behind one’s speech act, and can instead technologically enable its visibility directly. When a hashtag spreads because of an explicit campaign to spread it, it can be used to create a speci c kind of rst-person plural voice on social media. In using the hashtag, individuals may pointedly speak as part of a ‘we’ who is producing a collective speech act. Indeed, although the use of a hashtag has an agent-neutral input, when many people use a hashtag, it can either re ect or, more interestingly, generate a community, which can then engage in community-speci c speech with a community-speci c input. And though each individual use of a hashtag has an agent-neutral input, the collective speech act made up of all the speech acts sharing the hashtag may itself be one with community-speci c input—one structurally issuing from the community linked by this multipart speech act. The collective speech act can be community-speci c in its input when joining in becomes a display of insider status or when the push to spread a hashtag proceeds by drawing upon or forging a sense of belonging to a larger whole. An excellent example of a socially potent hashtag that has functioned to generate the possibility of speech with a community-speci c input is the #MeToo hashtag. I will spend some time on this example since #MeToo speech acts are especially complex in their pragmatic form and impact, as well as being a 12

watershed example of the social power of technologically enabled online speech acts.

By speaking in a rst-person voice, #MeToo speech acts have an agent-relative input in addition to the agent-neutral input that all hashtagged speech acts generally have. One can witness one’s own sexual violation only for oneself—hence the “me.” One cannot felicitously #MeToo anyone’s experience but one’s p. 117

own. At the same time, the

“too” part of the tag explicitly ties the speaker to a community of fellow

speakers, and serves to give the speech act a community-speci c input. The hashtag form of a #MeToo speech act materially ties the reporting of a sexual violation to other such reports and thereby ties the reported event to other such events. The hashtag makes this collective tie into a technological feature of the speech act, and it gives the speech act an agent-neutral output, as all hashtagged speech acts have. The agent-neutrality of the output is crucial; #MeToo posts are not directed to perpetrators. They do not do anything that has an agent-relative, second-personal output; they do not seek an apology, or demand accountability, or even take the form of an accusation. Rather, they simply make a report on what happened available to anyone. Most often, they don’t even identify a perpetrator (and sometimes don’t even o er any details about the event), but focus instead on two functions: reporting that a sexual violation happened to that person, and tying that report to other such reports by way of the hashtag, participating in a plural and collective act of reporting. It is notoriously di

cult for typical rst-person reports of sexual violations to receive the proper uptake. I

have argued elsewhere (Kukla 2014) that these reports are often treated as expressives; they are given uptake as though they are expressions of someone’s subjective reaction to an event, rather than as observations of empirical facts. When victims report sexual violations, their words are treated as subjective responses in need of interpretation, if not as outright fabrications. This treatment of attempted reports as hermeneutically challenging subjective expressions means that it is di

cult for victims’ attempted

testimony to accrue as evidence that sexual violation is widespread. We get evidence through reports, not through expressions of subjective feelings. For this reason, the case that sexual violation is common does not build in the way that it should. Moreover, even if reports of sexual violations are believed, the victim’s testimony often is treated as exceptional—as describing an event so horri c and peculiar that the perpetrator

must be an aberrant monster or the victim must have somehow done something to encourage or enable their own violation. The assumption then is that sexual violation is neither common nor culturally normalized, nor structurally enabled by our institutions. Of course sexual violation can look exceptional and rare only because the many reports of it are not taken as evidence of its pervasiveness. These reactions feed o

one another, each helping the other to portray sexual violation as rare and unlikely, hence requiring an

especially high burden of proof before we believe it happened. Notice what happens, at the level of discursive pragmatics, to victim reports that use the #MeToo hashtag. The tying of this speech act to many other such speech acts anchors the event in the objective world rather than in subjective space. It directly and technologically combats the tendency to take them as expressives rather than reports. The sheer volume of speakers in the plural discursive speech act that is the #MeToo movement solidi es speaker authority. It also pushes back directly against the idea that the event is rare or to be blamed on the victim’s unusual susceptibility to abuse. The fact that this volume of reports is available at the click of a mouse makes the quotidian, routine character of sexual violation immediately available to anyone. p. 118

Because #MeToo speech acts don’t generally name perpetrators, and because they technologically bind together multiple people making such reports, these speech acts usually center victims. It is normally almost socially impossible for us to talk about a sexual violation without centering the perpetrator: Is he actually guilty? How can he be such a monster? How will we fairly decide on his guilt? What should be done 13

to him? How will his life be destroyed by this event?

But in speaking as a victim, the author of a #MeToo

speech act ties their speech to that of other victims. This both centers victims as a collective group and gives each victim’s attempt to center herself added weight. The emphasis is on what events the victim experienced and on the shared character of the experience. This centering of the victim and creation of a collective, plural, victim-centered discursive act has made a substantial di erence to the discourse around sexual violations more generally. The pragmatics of #MeToo speech acts are especially complex compared to typical hashtagged speech acts, because of the inclusion of the rst-personal “Me,” combined with the “Too” that reaches out to forge community with others and to build a plural voice. But the main point for me here is that this is an example of a speech act type which we have all seen to have intense power to reorganize social space. The power of this speech act depends on its online, technologically enabled form, which allows it to function in the way that it does. It is not incidental that the name of the #MeToo movement includes the hashtag; this piece of technology is essential to the pragmatic work done by the movement. Whether a given reader clicks on the hashtag when reading a #MeToo testimonial or not, the fact that it is clickable, so that any given testimonial is immediately visible as part of a larger chorus of voices, is essential to its expressive function and pragmatic force.

8. Conclusions I have explored ve kinds of online speech acts, of sorts that were not technologically possible until the advent of social media platforms in more or less their current form. In each case, I have tried to show that the speech acts in question reorganize social space, and build our expressive capacities in ways that do not simply mimic what was available to us o

ine. Notice that in all cases, the material form of the speech act

helps constitute its social and expressive power. Also, the innovation has not been at the level of semantic content, but at the level of pragmatic force, where this pragmatic force depends on new material possibilities. #s,

s, and @s don’t change the truth conditions of a sentence; they don’t contribute to its

content. But they are material mechanisms that change what the speech act does as an intervention in social space.

Meanwhile, our grammar is of necessity in the process of evolving to incorporate these new kinds of speech acts. What grammatical part of speech is a ‘like’? What is the grammatical impact on a whole sentence of p. 119

inserting a # or an @? These are not

questions that are answered by traditional grammar, and language

is having to quickly adapt to integrate these new possibilities for discursive moves. I have avoided discussing online discursive innovations whose primary impact is semantic. In the semantic domain as well, however, we are seeing the emergence of discursive moves that add irreducibly to the expressive power of language. Emojis clearly have semantic impact on the speech acts in which they are used, but there is no way of reading this semantic e ect o

of our prior knowledge of the semantic structure

of English. As emojis become richly embedded in elaborate discursive practices and norms, it becomes less and less likely that their semantic impact can be reduced to expressions that could be made in traditional spoken English. Ethan Nowak (2020) has recently argued that some types of speech acts are language-dependent: They cannot be performed in every language, because they depend on the features and the history of the language itself. For example, he discusses poetic expressions that depend on the phonetic, etymological, and rhyming structure of the language for their expressive force. He also talks about performing respect by using a formal form of address, like “Vous” or “Sie”. The action of vouvoyer-ing or of Siezen-ing someone, for example, is simply not performable in English. Indeed, these actions are arguably not even the same as one another, because each has accrued a speci c history of material and normative import within its respective language. Nowak’s point is that speech-act theory must attend to dimensions of a language that cannot be understood in terms of semantic or grammatical abstractions, but only in terms of the actual concrete functioning of a particular language. My point is not the same as Nowak’s; after all, one can perform any of the technologically enabled actions that I have discussed here in any language. But it is a similarly spirited point: Many speech-act types depend for their performance and pragmatic form not just on their semantics and grammar, or even on these things put together with their material circumstances of performance, but on the concrete materiality of the speech act itself. It seems to me that philosophy of language, if it wishes to take seriously the idea that speaking is a material action, needs to start attending in much more detail to the material speci city of speech. I have been speci cally fascinated here by the impact on discursive possibilities of changes to the mechanics of how speech is performed and disseminated. The material changes in communicative possibilities have given rise to new pragmatic speech act types that are dependent on their material and technological form. We cannot perform the communicative actions of liking, @ing, retweeting, or hashtagging just by saying explicitly that this is what we are trying to do or by using a distinctive grammar; the pragmatic force is tethered to the materiality of the speech act. Saying “I like your comment” is not the same act as

-ing a

comment. These are emergent new pragmatic categories of speech acts. The Internet and its attendant and ever-evolving changes in our material possibilities for discursive communication have expanded the pragmatic repertoire of the language. Not only have we seen the emergence of new pragmatic types of speech acts, but we have seen that these can be socially potent: they can build public identities, establish p. 120

group insiders and outsiders, forge plural speaking subjects, pull people into

conversation against their

will, release them from conversation, and negotiate and display group norms of social and discursive behavior, for example. If we wish to understand speaker authority, the formation and maintenance of social groups, or the emergent norms of discourse and conversation, we ignore online speech acts at our peril. The Internet is currently our primary medium for public speech and shared discourse. Especially since the COVID era, when more of our speech than ever is technologically mediated, we need to dramatically expand our understanding of speech-act theory to encompass online speech.

Notes

p. 121

1.

I suspect that, ultimately, this distinction is a continuum, and that all speech acts are to some extent medium-dependent: the materiality of their performance matters to their force. I do not try to make a case for this stronger claim here.

2.

They might leap in and o er to do what I am requesting, but this is a new speech act—an o er—and not itself a response to the normative claim made by the request.

3.

Gretchen McCulloch (2020, 340–341) points out that speakers o en use ʻvaguebookingʼ or inside references on social media to speak to a specific community of insiders despite the fact that a larger audience will see the post. The speech act is meaningless except for those to whom it is speaking. This is also an example of speech with a community-specific output.

4.

I recognize that Elon Musk has now o icially renamed “tweets” as “posts”, “tweeting” as “posting”, and so forth. But this generic language is confusing because one also “posts” on Facebook and elsewhere. In addition, the informal culture of online media has solidified for years around the activity of “tweeting.” So I have elected to keep the technically-nowarchaic language.

5.

A complexity here is that some people, especially celebrities, get others to run their social media accounts, so the “I” that speaks on social media might not correspond neatly to a single fleshy individual; it may be a complex socially constructed speaker. But of course, there are such complex speakers o line as well.

6.

See Kukla and Lance (2009, chap. 7), for a detailed discussion of this structural dynamic and the reasons for it.

7.

See Kukla and Lance (2009) for a detailed defense, but I hope this is intuitively clear.

8.

I originally wrote this section in the spring of 2020. Since then, tag groups have gone somewhat out of fashion, and while they still exist and are used, they are becoming less common, perhaps as younger users have continued to abandon Facebook. This kind of fluidity of online speech is interesting in its own right, and so I decided to leave this section in the chapter.

9.

But see note 7 just above!

10.

A prominent and classic (although likely outdated) discussion of the discursive move of retweeting is Boyd, Golder, and Lotan (2010).

11.

Facebook also employs hashtags, but they are not as heavily used there; and since Facebook posts are much more likely than tweets to be restricted to a friend group, their hashtags donʼt tend to have as much power, and so I focus on hashtag use on X/Twitter.

12.

The next several paragraphs overlap in content with the discussion of #MeToo in Kukla, Herbert, and Watson (forthcoming).

13.

Kate Manne (2017) coined the term “himpathy” to name our strong cultural tendency to center harms to male perpetrators in our discussions of sexual violations and other abuses by men of women.

References Boyd, Danah, Scott Golder, and Gilad Lotan. (2010). “Tweet, Tweet, Retweet: Conversational Aspects of Retweeting on Twitter.” 43rd Hawaii International Conference on System Sciences, Honolulu, HI, pp. 1–10. doi: 10.1109/HICSS.2010.412. Brandom, Robert. (1983). “Asserting.” Noûs 17 (4): 637–650. Google Scholar WorldCat Brandom, Robert. (1998). Making It Explicit: Reasoning, Representing, and Discursive Commitment. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Herbert, Cassie, and Rebecca Kukla. (2016). “Ingrouping, Outgrouping, and the Pragmatics of Peripheral Speech.” Journal of the American Philosophical Association 2 (4): 576–596. Google Scholar WorldCat Kukla, Rebecca. (2014). “Performative Force, Convention, and Discursive Injustice.” Hypatia 29 (2): 440–457. Google Scholar WorldCat Kukla, Rebecca, and Mark Norris Lance. (2009). ʻYo!ʼ and ʻLo!ʼ: The Pragmatic Topography of the Space of Reasons. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Kukla, Quill R, Cassie Herbert, and Ari Watson. (Forthcoming). “Sexual Violation and the Language of Repair.” In The Philosophy of Sexual Violence, edited by Georgi Gardner and Yolanda Wilson. New York: Routledge. Google Scholar Google Preview WorldCat COPAC Lorenz, Taylor. (2019). “The Groups Bringing Forum Culture to Facebook.” The Atlantic, May 17. Manne, Kate. (2017). Down Girl: The Logic of Misogyny. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC McCulloch, Gretchen. (2020). Because Internet: Understanding the New Rules of Language. New York: Riverhead Books. Google Scholar Google Preview WorldCat COPAC Nguyen, C. Thi. (2020). “Echo Chambers and Epistemic Bubbles.” Episteme 17 (2): 141–161. Google Scholar WorldCat Nowak, Ethan. (2020). “Language Loss and Illocutionary Silencing.” Mind 129 (515): 831–865. Google Scholar WorldCat

p. 122

Pavalanathan, Umashanthi, and Jacob Eisenstein. (2015). “Audience-Modulated Variation in Online Social Media.” American Speech 90 (2): 187–213. Google Scholar WorldCat

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

p. 125

6 The Truth about Slurs  Robin Jeshion https://doi.org/10.1093/oxfordhb/9780192844118.013.6 Published: 22 May 2024

Pages 125–148

Abstract Linguistic theories of slurs confront a problem about truth. Identity theories are theories on which slurs and neutral counterparts are co-designating, and sentences containing them are truthconditionally equivalent to sentences where the slur is replaced by its neutral counterpart. Identity theories immediately lead to disturbing consequences about truth. If slur S designates what its neutral counterpart NC designates, then since ‘there are NCs’ is true, so too is ‘there are Ss’. Pre-theoretically, this consequence appears unacceptable. To many, it appears to manifest bigotry and as such must get the facts wrong. Such consequences are one impetus for non-identity theories: Null extension nonidentity theories, like those of Hom and May, on which slurs have an empty extension, and an alternative non-identity theory due to Richard, on which utterances with slurs lack truth value. This chapter o ers novel arguments that non-identity theories are not superior accounts of slurs. It advances four conditions on a theory of slurs and argues that null extension non-identity theories are, in principle, unable to satisfy one of them, the Compositionality Condition. It shows why Richard’s rationale for denying truth to slur-containing utterances is unconvincing. Finally, it defends identity theories, especially expressivist versions, from the charge that they cannot accommodate data about Frege’s puzzle, conceivability, and a version of the Frege-Geach puzzle.

Keywords: slurs, epithets, pejoratives, truth, derogatory words, expressives, projection, compositionality, Richard, Hom and May Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction All linguistic theories of slurs confront a problem about truth. Because slurs are expressions for groups, we must ask, do slurs designate anyone? In particular, do slurs designate the same groups their neutral counterparts designate? Views on which slurs and their neutral counterparts are designationally equivalent immediately lead to disturbing consequences about truth. If Kike designates what Jew designates, and if it is true that

[1NC]

Jake is a Jew

then

[1]

Jake is a Kike

1

is true. More generally, since it is true that there are Jews, so too is it true that there are Kikes. As since it is true there are Hispanics, so too is it true that there are Spics. And so on. These are the perverse and disturbing consequences of what I call identity theories, linguistic theories on which slurs, in their literal uses, are designationally equivalent with their neutral counterparts, and slurcontaining declarative sentences are truth-conditionally equivalent with sentences where the slur is 2

replaced by its neutral counterpart. Numerous linguistic analyses, including prohibitionism, pragmatic a

3

4

5

liationalism, perspectivalism, and expressivism, constitute identity theories.

Pre-theoretically or intuitively, these consequences appear unacceptable. To many, they appear to be themselves manifestations of anti-Semitism, racism, bigotry. It appears that they must get the facts wrong. p. 126

Such consequences are one impetus for non-identity theories, those that deny either the designational equivalence of slurs and their neutral counterparts or deny that utterances of slur-containing sentences can be true. Theories that deny designational equivalence typically advance meaning analyses for slurs that are empty. On such null extension non-identity theories, the null extension is secured by the slurs’ encoding some variety of derogating or dehumanizing linguistic content for being in the group. For instance, Hom [Combinatorial Externalism]: Chink means “ought to be subject to higher college admission standards, and ought to be subject to exclusion from advancement to managerial positions, … because of being slanty-eyes and devious, and good at laundering, and … all because of being 6

Chinese.”

Hom and May [Moral Innocence]: Kike means “ought to be the target of negative moral evaluation 7

because of being Jewish.”

In both views, the semantic content that encodes slurs’ conventional function as a tool of derogation also generates the null extension. Clearly, both views honor the intuition that [1] is not true and that sentences like ‘There are Kikes’ are false.

An alternative non-identity theory due to Mark Richard maintains that, unlike utterances containing neutral counterparts, utterances with slurs lack truth value: they are neither true nor false. They, too, honor the intuition that neither [1] nor ‘There are Kikes’ are truths. Are non-identity theories superior analyses of the semantics of slurs? In this chapter, I o er novel arguments that, despite the initially more attractive consequences, nonidentity theories are not better. In section 2, I o er four conditions on a theory of slurring terms and then argue that null extension non-identity theories are, in principle, unable to satisfy one of them, what I call the Compositionality Condition. Because all the conditions are relatively weak and exceptionally general, the failure to satisfy Compositionality is a serious failing of a theory of slurs. In section 3, I argue that while Richard’s expressivist non-identity theory that denies truth to utterances containing slurs satis es our conditions, his main rationale for why utterances should be denied truth value is unconvincing. In section 4, I defend identity theories, especially expressivist versions, from the charge that they cannot accommodate certain data involving truth conditions. One important quali cation at the outset: in my view, there are numerous varieties of slurring terms. What I call canonical slurs are typically used as weapons to reference and derogate a whole group on the basis of socially signi cant categories like ethnicity, race, religion, sexual orientation, gender identity, ability, nationality, and class. For instance, Kike is a canonical slur for Jews, and Chink is a canonical slur for Chinese persons. The targets of canonical slurs often value the social groups that they are derogated for with slurs. Canonical slurs typically have neutral counterpart terms that non-pejoratively reference the groups that the slurs derogate. By contrast, gendered slurs like slut, bitch, and wuss have a range of basic uses that do not reference a whole gendered group: bitch and slut are not normally used to reference and derogate all girls p. 127

and women. Instead, they

are commonly used to police those whose actions and demeanors break or defy

norms on girls and women. Slut is applied to those who break, or are feared to be breaking, religious and patriarchal norms governing sexuality. Stereotyping slurs like Uncle Tom, Welfare Queen, Karen, and Basic Bitch all evoke a rich stereotype of a social or intersectional group. Gendered slurs and stereotyping slurs do not typically have syntactically simple neutral counterpart expressions. In this chapter, I con ne my discussion of truth in linguistic accounts of slurs exclusively to canonical slurs.

2. Basic Conditions on Theories of Slurs In the following subsections, I articulate four general conditions on a theory of slurs: the Target Condition, the Linguistic Encoding of Derogation Condition, the Projection Condition, and the Compositionality Condition. I then advance an argument resting on those conditions to demonstrate the untenability of nullextension non-identity theories which generate the empty extension by semantically encoding a content that explicitly derogates on account of group membership. While previous discussions of non-identity theories have focused on projection phenomena, this argument identi es their key problem in ful lling the Compositionality Condition.

2.1. The Target Condition No one disputes the fact that particular slurring terms are systematically related to particular groups as devices for designating those groups—the very groups that their neutral counterparts designate. And no one disputes the fact that such regular applications ought to be accounted for as a matter of linguistic convention. Someone who, while aiming to speak literally, regularly applies Kike to African Americans is linguistically incompetent with the term (Richard 2008). A similar point can be framed in terms of linguistic understanding. Someone who understands utterances containing Chink as being about Jews misunderstands that slurring term. To account for such linguistic incompetence and misunderstandings, all linguistic theories need to specify conventions for correct applications. I synthesize this constraint as a Target Condition: Target Condition: a theory of slurs’ conventional linguistic properties ought to encode which group is the target of each slurring term. The Target Condition makes no claim about how the target group is linguistically encoded. Nevertheless, p. 128

one might think it is enough to require that slurs have the same

extension as their neutral counterparts.

Copp and Sennett (2015, sec. 8), for instance, claim that non-identity theories are undermined by the fact that pejoratives are used to refer to the same people as the corresponding neutral counterparts are used to refer to. ‘Kike’ is used by anti-Semites to refer to Jews and everyone who knows how ‘kike’ is used knows this. There is no uncertainty in the minds of those who are competent with the use of ‘kike’ regarding to whom an anti-Semite means to refer when using ‘kike’. This point undermines [Non-Identity] theories. Copp and Sennett o er this passage as a summary of their case against non-identity theories. Their fuller argument incorporates commitment to a Lewisian ‘best- t’ semantics on which agents’ actual use of a term and dispositions to use it determine its extension. On it, because anti-Semites are the only users of the term, their application of kike to Jews engenders that Jews are the best- t referents for that slur. Copp and Sennett’s argument would be a swift case against non-identity theories. But the argument does not go through. The primary reason is that the Target Condition is thin. Theories like Hom and May’s Moral Innocence on which kike is synonymous with Jewish and contemptible on account of being Jewish (or worthy of negative moral evaluation on account of being Jewish) satisfy it by incorporating the target of kike within the slur’s meaning. The same goes for theories like those of Hom (2008) and Neufeld (2019). The descriptive meaning analysis speci es that Jews are the targets of the slur. The problem for Copp and Sennett is this: that anti-Semites use kike to refer to Jews doesn’t itself entail that it in fact refers to Jews. One way to see this is to compare slurs with terms like unicorn horn, a comparison Hom and May themselves develop to bolster their conception of slurs as terms of ctional discourse. Throughout antiquity and through the Renaissance, unicorns were widely believed to exist and to possess magical powers. People believed that unicorn horns could heal the sick, detoxify poisons, and cure demonically induced melancholia. Naturally, unicorn mythology believers used unicorn horn to apply to what they thought were unicorn horns. About this, they were wrong: what they thought were unicorn horns were in fact the tusks of decidedly non-magical sea creatures, the humble narwhal. Although they used unicorn horn to designate narwhal tusks, the term need not—arguably should not—have the same extension as narwhal tusk. A theory on which terms for mythological creatures, like unicorn, are extensionless is not 8

implausible. Thus, unicorn horn may, arguably, be taken to have an empty extension.

Of course, I am not here committing to such a semantics. My claim is just that a meaning analysis on which unicorn horn has an empty extension is not ruled out straightaway by the fact that believers in the unicorn

mythology used unicorn horn, and used it competently, to refer to narwhal tusks. What they lacked was not linguistic knowledge but empirical knowledge—namely, that narwhal tusks are not the horns of unicorns. Similarly, the fact that anti-Semites apply Kike to Jews, and do so competently, does not by itself rule out a 9

semantics on which Kike has null extension. p. 129

The upshot is that even though by our Target Condition a linguistic theory must encode slurs’ targets, who they are about, this alone is insu

cient to rule out Non-Identity theories.

2.2. The Linguistic Encoding of Derogation Condition A theory of slurring terms also needs to explain how slurs di er from their neutral counterparts. Slurs, when used literally, are tools of derogation, terms used systematically to dehumanize their targets. Their neutral counterparts are not. Certainly, they, too, can be used to derogate: one can infuse Jew in [1NC] with contemptuous intonation, therewith making the speech act a derogation. But this is not one of neutral counterparts’ prime functions and they themselves, as words, neither encode nor implicate derogating 10

contents or attitudes.

In maintaining that slurs function as tools of derogation and that this marks a central di erence between them and their neutral counterparts, I have thus far stated nothing about the scope of the derogation, what form it takes, and how it is accomplished. The condition I put forward is this: slurs’ functioning as tools of derogation ought to be explained, at least in part, by appealing to slurs’ conventional linguistic properties. That is, part of what distinguishes slurs from their neutral counterparts ought to be characterized either at the level of what is encoded as semantic descriptive content within slur-containing utterances or by systematic presuppositions, by rules governing slurs’ use, by what uses of them conventionally implicate, or by what they implicate as generalized conversational implicatures. I call this constraint the Linguistic Encoding of Derogation Condition. Linguistic Encoding of Derogation Condition: a theory of slurs’ conventional linguistic properties ought to encode at least part of slurs’ function as tools of derogation. While this condition requires extensive argument, I bypass it here because all theorists who dispute slurs’ and neutral counterparts’ designational and truth-conditional equivalence or that slur-containing utterances cannot express truths or falsehoods agree that slurs’ distinctive standing as tools of derogation 11

are, at least in part, conventional linguistic properties.

Note that our condition rests on understanding slurs as tools of derogation, not on their causing someone or a group to be o ended, to ‘take o ense.’ One who takes o ense to an utterance feels indignant, outraged, or a ronted by it, and may display that feeling publicly by taking a stand in opposition to the utterance. Taking o ense is a reaction to something. Whether and the degree to which a slur-containing utterance inclines people to take o ense is a function of who hears the utterance, those agents’ psychological sensitivity, moral sentiments, political beliefs, dispositions in that moment toward indignance and outrage, and so on. p. 130

Yet a term

can be a tool of derogation without its engendering o ense in those who hear it used in a

particular context.

2.3. The Projection Condition Slurs function to derogate not only in utterances of simple declaratives involving a predication of the slur, as in [1], but also in utterances of complex slur-containing constructions involving negation [2], disjunction [3], conditional and modal statements [4]–[6]; and imperatives [7], interrogatives [8], and vocatives [9].

[1]

Jake is a Kike.

[2]

Jake is not a Kike. He just looks like one.

[3]

Yao is either a Chink or a Jap.

[4]

If there are faggots in the bar, my father wonʼt go in.

[5]

If there are no Chinks in the orchestra, I will be surprised.

[6]

Thereʼs no way a Kike would work for that company.

[7]

Stay away from that group of Spics.

[8]

Is there a Chink cooking in the kitchen?

[9]

Wherever you go, Chink, Iʼll be taunting you.

Here we assume that the slurs in [1]–[9] are not functioning metalinguistically: they are not implicitly about language. Additionally, the slurs in the utterances are not issued as echoing or as denials of prior racist 12

content.

In [2]–[9], the slurs derogate their targets just as they do in [1], when the slur occurs as a simple predication. Despite the slur’s occurrence in the antecedent of the conditional, [4] derogates gay or queer persons. One couldn’t cancel the derogation by pleading “I only said if there are faggots in the bar.” Similarly, in [8], that the speaker is just asking whether (as opposed to stating that) a Chink is in the kitchen does nothing to excise the utterance’s derogation. And in [2], though with the rst sentence, the speaker negates the predication in [1], the slur still derogates Jews. Negation does not mute slur-containing utterances’ derogation. This non-interaction of logical connectives with slurs’ derogation is a familiar fundamental property of slurs: slurs’ derogation projects through negations, conditionals, interrogatives, imperatives, and vocatives. The utterances of [2]–[9] are all derogatory. Additionally, [2]–[9] appear to be derogatory in the same way as when they occur in simple predications like [1]. What explains the derogating element for a single slur does not shift with embeddings in the di erent logical operators and grammatical forms. Thus, a linguistic theory ought to provide a common analysis of the derogation inherent in all these slur-containing constructions. Call this the Projection Condition. Projection Condition: a theory of slurs’ conventional linguistic properties ought to provide a common analysis of the derogation inherent in all non-metalinguistic non-echoic slur-containing constructions. p. 131

Note that this condition does not preclude there being multiple forms of derogation within the individual constructions in [1]–[9], say. It requires only speci cation of a source of derogation common to all.

2.4. The Compositionality Condition When bigots use slurring terms, they aim to be speaking about the members in the group the slur’s neutral counterpart designates. We drew attention to this feature of slurs in the Target Condition, illustrated in [1]. The same obtains for the complex constructions in [2]–[9].

In non-metalinguistic, non-echoic constructions involving the interaction of a logical connective and a slurring term, used literally, the connective always operates on whatever represents the target group. Take, for instance, constructions involving negation interacting with a slurring term, as in [2] and [5]. What [2] encodes is that Jake is not Jewish—coupled with a derogation of Jews. One who utters [2] aims to say that Jake is not Jewish, as well as to derogate Jews. In [5], negation similarly operates on that aspect of the slur that represents the target group, not the derogation itself. The speaker who utters it aims to say he will be surprised if there are no Chinese persons in the orchestra. Similarly for modal operators. [6] encodes that it is impossible that a Jewish person would work for the company. A theory of slurs must reveal how slurs interact with logical operators. Compositionality Condition: a theory of slurs’ conventional linguistic properties must ensure that in all non-metalinguistic non-echoic slur-containing constructions, logical operators’ interaction 13

with slurs is independent of their derogation and act only on the group designation.

2.5. The Argument against Null Extension Non-Identity Theories My main argument for why null extension non-identity theories are untenable is that there is no viable way for them to simultaneously satisfy the Target Condition, Derogation Condition, Projection Condition, and Compositionality Condition. While I illustrate the main argument using Hom and May’s (2018) semantic analysis, the argument provides a general problem confronting all null extension non-identity theories on which the null extension is generated from slurs’ encoding a content that derogates the target group on account of being in the target group. Hom and May (2013, 2018) advance a general semantic analysis of pejoratives for groups. For them, pejoratives are complex expressions, composed of a general lexical marker PEJ that functionally combines with a neutral counterpart term for a social group to yield a pejorative expression. Kike is analyzed as p. 132

PEJ(Jew). Generally, PEJ(ξ)

combines with a counterpart term, t, denoting a social group (racial, religious,

ethnic, sexual orientation…) to form a pejorative expression, PEJ(t). PEJ (ξ) expresses the concept ought to be the target of negative moral evaluation because of being ξ (2013, 298). For them, all slurs have an empty extension for no one ought to be the subject of negative moral evaluation on account of being in a particular religious, ethnic, etc. group. Because slurs are formed from their neutral counterpart expressions, PEJ(t), their linguistic meaning encodes the target group, so Hom and May satisfy the Target Condition. They satisfy the Derogation Condition because, for them, slurs’ linguistic meaning attributes the derogating content ought to be the target of negative moral evaluation because of being ξ. What about the Projection Condition? For Hom and May, the semantics of slurs encode a negative 14

prescriptive predicate ought to be the target of negative moral evaluation because of being ξ.

However, while

the semantics generates a derogation of an individual for simple predications like [1], it does not for [2]– [9]. For instance, [8] encodes a question of whether there is someone who ought to be the target of negative moral evaluation because of being Chinese, which is not itself an attributed negative moral evaluation. Hom and May do, however, claim that while complex constructions do not semantically encode a negative moral evaluation, they systematically conversationally implicate or presuppose a derogation. Simply by using a slur, a speaker presupposes that the slur’s extension is not empty—and this existential presupposition 15

constitutes a derogation of the target group.

For them, such a presupposition would count as a constant

and common explanation of the slurs’ derogation across all non-metalinguistic, non-echoic constructions. 16

Although the presence of the presupposition has been questioned, I grant it here. an existential presupposition, Hom and May satisfy the Projection Condition.

I allow that by appeal to

What I challenge is their ability to satisfy the Compositionality Condition. Their semantics cannot capture how logical operators interact with slurs. Consider [2] and suppose that Jake is in fact not Jewish. Suppose also that the speaker of [2] knows this fact and speaks literally and sincerely. On Hom and May’s analysis, [2] means: Jake is not one who ought to be the target of negative moral evaluation because of being Jewish. But this meaning analysis does not capture the meaning constrained by our Compositionality Condition, which matches the speaker’s intended meaning. With [2], the bigot aims to say just that Jake is not Jewish—while also derogating all Jews. He is therefore not expressing a content which allows that Jake is one of those exceptional Jews who doesn’t deserve negative moral evaluation on account of being Jewish. We can even stipulate that our speaker believes and says that all Jews are Kikes. When he utters [2], therefore, he only denies that Jake is Jewish while expressing a wholesale derogation of Jews. Yet on the PEJ(ξ) analysis, negation functions to deliver a reading allowing that Jake is an exceptional Jew who ought not be derogated for being Jewish. The semantics does not satisfy the Compositionality Condition. More, it cannot satisfy the condition because the primary problem stems from the general form of the semantic content of slurs on all p. 133

null extension non-identity theories: because the contents derogate on the basis of the group membership, negation of that content will leave open a reading that allows for the ‘exceptional’ group members. The same point applies to [6], a statement the bigot uses to express a relationship between their surprise and the presence of Chinese persons in the orchestra. Yet on the PEJ(ξ) analysis, [6] means: If there are no persons who ought to be the target of negative moral evaluation because of being Chinese, I will be surprised. This is decidedly not what the racist intends or means in his competent use of negation in the antecedent of [6]. Here again, the failure of the Compositionality Condition is general, cutting across null extension nonidentity theories, and, because of the form of their semantically encoded derogating content, reveals they are in principle unable to satisfy all our conditions. Although, strictly speaking, this argument just presents a serious problem for null extension non-identity theories, it also illuminates why we ought to embrace identity theories: slurring terms’ projection and compositionality phenomena force a separation of the slurs’ linguistically encoded derogation and their group classi cation.

3. Richardʼs Reductio 3.1. Richardʼs Reductio and Analysis of Misrepresentation Is there an alternative way to satisfy our four conditions while avoiding identity theories? I turn now to consider Mark Richard [2008] novel non-identity theory that is distinctive because to account for the lack of truth value in all slur-containing utterances, he does not propose that slurs have null extension. Instead, he proposes an expressivist analysis on which the Target, Derogation, Projection, and Compositionality Conditions are all satis ed because slurs derogate by expressing contempt for target groups on account of their group 17

membership, yet no sentences containing slurs express truths.

For him, utterances of all slur-containing

sentences are neither true nor false. Richard o ered an alternative rationale for why all utterances of sentences containing slurs cannot be true. To underwrite his view, Richard advances a reductio of any view on which slur-containing utterances encode truths: Let S be some odious racial slur. Imagine standing next to someone who uses S as a slur. Perhaps you are in front of a building where targets of the slur live and work; the racist mutters That building is full of Ss. … if we admit that [what the racist said was true], we must believe that it is true

that the building is full of Ss. And if we think that, we think that the building is full of Ss. We think … what and as the racist thinks. This certainly seems to make us complicit in the racist’s racist 18

attitude, and thus to some extent racists ourselves.

In this reductio, Richard locates a practical, widespread, and serious moral cost to counting the racist’s p. 134

utterance to be true. The cost is practical and widespread because

it makes all who believe what the racist

says was true complicit in the racist’s racist attitude. It is serious because, as Richard notes, such complicity makes all such believers, to an extent, racist themselves. The argument is undeniably seductive insofar as it transfers to complex lexical constructions containing negation and conditionals through which the derogatory quality of the slurs project. If Richard’s racist muttered There are no Ss in that building (or: If there are Ss in that building, the police won’t come), and we believe what they said is true, then we believe that there are no Ss in that building (or: that if there are Ss in that building, the police won’t come). And thus, we think that there are no Ss in that building (or: that if there are Ss in that building, the police won’t come) and thereby think what and as the racist thinks. In this way, Richard’s argument applies broadly and avoids problems associated with projection and compositionality. What engenders the conclusion that we must be racists ourselves by accepting that racists utter truths? The key resides in Richard’s conception of the relationship between thought, expressions of and demonstrations of a ect, representational content, and truth value. For him, a contemptuous attitude toward an individual or group is part of what one thinks when one thinks with a slurring term, and, as such, contributes to one’s representation of them. “To have an attitude of contempt towards someone because of their race or 19

ethnicity is … to represent one’s target … as contemptible because of his race or ethnicity.”

Yet, for all

properties like race, religion, ethnicity, and sexual orientation for which such contempt is always misplaced, all slurring thought and talk constitutes or involves a misrepresentation of its subject. Thus, a thought that

[1]

Jake is a Kike

is a misrepresentation in precisely one same way as is a thought that

[2]

Jake is not a Kike

for both incorporate misplaced attitudes of contempt toward Jews. Since utterances of such thoughts involve the expression of contempt and thus the display of such misrepresentations, and because, according to Richard, misrepresentations are incorrect representations, all slurring thoughts and their expression in language lack truth-value. As Richard says, “[T]hat what is said in slurring speech is neither true nor false comes rather directly from the everyday observation that to think [like the bigot] … is to think wrongly 20

because it involves misplaced contempt and hostility.”

To appreciate the reductio, distinguish, as Richard does, classi cations from representations. One who asserts [1] is engaged in classi cation. In thinking and speaking with Kike, they classify individuals in the same way as one who uses Jew, as in [1NC]. They are engaged in an activity that is answerable to the world, and as such

can succeed or fail. Yet for Richard, there’s no direct entailment from someone engaging in classi cation to their having said something true. He denies that classi cation is straightaway, without additional argument, 21

su cient for truth or falsity. p. 135

Instead, Richard closely connects the notions of a

rming a thought as true, approving that thought, and

endorsing it as a right representation. He writes: [T]o say that a thought is true is, inter alia, to approve of it, to endorse it: if I say your thought is true, I am saying that you are representing the world aright. To say that your thought is true is to endorse it as representing the world correctly. But when the anti-Semite thinks slurringly of Jews, he is not representing the world aright. We cannot approve of representing or thinking of people in the way in which he does, for we would approve of thinking of people as inferior merely because of 22

their religion. It is wrong to represent anyone in that way.

He summarizes by claiming that “we cannot endorse the anti-Semite’s slurring performance as true, as it 23

involves misrepresentation.”

3.2. Misrepresentations: Misplaced A ect and Its Impact on Truth As I understand him, Richard’s primary rationale is that because contemptuous thought about the target is intrinsic to slurring thought and speech, it misrepresents, and misrepresentations being misrepresentations deprive the thought and communicated content with truth and falsity. That is, quite generally, 24

misrepresentations in thought and talk undermine possession of truth value.

I am unsure whether Richard is correct that in thinking with a slur, a bigot’s contemptuous attitude is ‘part’ of his thought. Does this mean that the attitude is so intrinsic to the representation of the group that it is inseparable from the referential aspects of the thought? Or only that the attitude is inextricably tied to the thought, and always arises with the thought? I will not attempt to settle these matters. My critical analysis will not hinge on how an a ective attitude integrates with thought. For the sake of argument, I grant Richard’s position. What I question is, rst, whether misrepresentations incorporating misplaced contempt must deprive a thought and its utterance of truth-value. To bolster the point, consider a distinct instance of the more general phenomena—misrepresentations due to misplaced, unjusti ed a ect. Consider an example involving 25

misplaced fear.

Imagine someone so pathologically afraid of all spiders, their fear is indiscriminate.

Though they know most spiders are harmless, they fear the entire species. Their spider concept, we suppose, incorporates this fear, and does so in whatever way, according to Richard, the bigot’s concept of a group incorporates contempt and hostility. Imagine that a harmless spider crawls on our subject, and, though they recognize the spider to be harmless, they scream violently in panic:

[10]

Thereʼs a spider on my arm!

deploying their fear-laden concept. Because we are assuming that the speaker’s a ective state is part of p. 136

their representational thought content, their panic permeates their

thought. Nevertheless, what the

speaker says is true. Though, we are supposing, their a ective attitude pervades the representation, the panic itself does not compromise the truth-evaluability of what they express in [10], which turns only on 26

whether the spider is on the speaker’s arm.

The point should transfer readily to slurs.

Consider another example employing concepts of racial or religious groups incorporating misplaced esteem based on in-group membership. We imagine a group ideology that incorporates norms of respect for all persons, so it does not denigrate other groups as lesser persons. However, within the ideology, the concept of their own group accords esteem to in-group members simply for being in the group. To make it concrete, image that the speakers of [11] and [12] accord esteem to Jewish and Japanese persons simply on account of their group membership.

[11]

Albert Einstein was Jewish.

[12]

Kimi Nakamura, that girl in your science class, sheʼs Japanese.

Such high regard saturates the speakers’ individual Jewish and Japanese concepts and their utterances both encode veneration. Such misplaced esteem constitutes a misrepresentation. Our examples manifest the same misrepresentation for why, according to Richard, slurring thought and talk is truth-valuelessness. They are inaccurate along a moral dimension: no one is an inferior person on account of group membership. So too, Jews qua Jews, and Japanese qua Japanese are not esteem worthy. Yet the thoughts and utterances containing these misrepresentations don’t for that lack truth-value. We should retain the same standard on misrepresentations with slurs, and reject their alleged truth-valuelessness. Finally, take slurring speech acts conveyed with bare pejoratives fronting neutral counterparts or with contemptuous intonation on the neutral counterpart:

[13]

That building is full of fucking Jews/goddamn homosexuals.

[14]

That building is full of Jews

C

Though the speakers misrepresent, what they’ve said is not thereby bereft of truth value. Where does Richard’s reductio go wrong? Recall: he claims that if we allow that what the racist says is true when he says That building is full of Ss, then we must believe that it is true that the building is full of Ss. Thus, he concludes, we must think that the building is full of Ss, which is to think what and as the racist thinks. The misstep resides in assuming that if we allow that what the racist says is true, we must thereby think as the racist thinks. We can allow what the racist says is true, without believing as the racist thinks. We need only believe that it is true that that building is full of NCs. Before concluding this section, it’s worth advancing a secondary objection pinpointing why the core p. 137

problem with slurs is independent of truth values. Richard’s

argument focusing on depriving utterances

of truth value is inapt, for, as framed, it depends upon believing what the racist asserts. Yet notice: although his argument generalizes to lexically complex declaratives, as framed, it does not transfer to nondeclarative slur-containing utterances, as when the bigot asks, as in Are there Ss in that building? or uses a vocative, as in Get lost, Ss! After all, there’s nothing here to believe true. This shows that the role of truth in the communicative transfer of racist thought isn’t what generates the power of Richard’s thought experiment. One can see this by reformulating the argument with Are there Ss in that building? or Get lost, Ss! We still confront the essence of the problem upon hearing the bigot derogate with both the question and

vocative. To understand the bigot, we must grasp what is said. The main matter is then just ‘thinking … as the racist think[s]’. Thus, the crux of the matter is not truth value at all. What is at stake, at bottom, concerns instead the pernicious power of the a ective stance the racist expresses in all slur utterances—declaratives, imperatives, vocatives—and it, not truth, ought to be the locus of our moral disapproval. I’ve argued that Richard’s reductio collapses because accepting a slur-containing utterance as true does not require that we represent as the racist represents. We can accept utterances as true that contain morally misguided misrepresentations and without thinking as the racist thinks. I’ve argued, further, that replacing Richard’s reductio with a version involving slur-containing non-declaratives reveals that the essence of the worry about slurs involves co-thinking with the racist, not truth.

4. Three (Apparent) Problems about Truth for Expressivist Identity Theorists In section 2, I advanced conditions on a linguistic theory of slurs to explicate fundamental problems for null extension non-identity theories. In section 3, I argued against Richard’s rationale for why slur-containing utterances are deprived of truth value. In this section, I change course to defend identity theories and, especially, expressivist identity theories from the charge that they cannot accommodate certain data involving truth conditions. Hom and May have advanced three arguments aimed at establishing that expressivist identity theories are riddled with distinctive problems resulting from taking utterances of slurcontaining sentences to be truth-conditionally equivalent to utterances of sentences that swap the slur for its neutral counterpart. I defend expressivist identity theories against the charges that they cannot successfully explain versions of Frege’s Puzzle (section 4.1); that it confronts a problem due to modal conceivability (section 4.2) as well as one analogous to the Frege-Geach puzzle that plagues expressivist 27

analyses of moral language (section 4.3).

p. 138

4.1. The Fregeʼs Puzzle Argument Hom and May (2013, 2014) advance what they describe as a “version of Frege’s Puzzle” as engendering a serious problem for expressivist identity theories. Recall Frege’s puzzle: assuming a and b have the same referents, sentences having the form ‘a=a’ are trivially true while sentences of the form ‘a=b’ are nontrivial. According to Hom and May, because we “posit that Jew and Kike have identical semantic values,” we face a version of Frege’s puzzle in sentence pairs like:

[15]

Jews are Jews

[16]

Jews are Kikes

for we cannot account for why [15] is trivial and cognitively uninformative, while [16] is non-trivial and cognitively informative. Hom and May o er a solution on our behalf, invoking standard Millian resources, distinct modes of presentations, alternately associated with Jew and Kike: When an agent A thinks of Jews in an anti-semitic way, she employs one mode. Call it mK. When A thinks of Jews in a non-anti-semitic way, she employs another mode. Call it mJ. The words “kike”

and “Jew” are the outward re ections of these ways of thinking, even though they do not di er in 28

the propositional content with which they are associated.

Yet, they maintain, this, too, engenders a problem, for it forces the non-racist identity theorist who holds that Jews are one and the same as Kikes, to think of Jews via racist modes of presentation, mK. Echoing 29

Richard, this is untenable for it forces the non-racist “to think of Jews as the anti-semite does.”

The solution on o er should not tempt an identity expressivist. There is no need to appeal to di ering modes of presentation associated with the slur and its neutral counterpart, and no advantage gained by doing so. The Frege’s Puzzle Argument is best dealt with in one of two related ways. The rst grants that this is a bona de version of Frege’s puzzle and explains away the appearance of non-triviality in [16]. To do so, rst note that this version of Frege’s puzzle is supposed to be rooted in the fact that one term is a pejorative and the other is not, not simply in the fact that we have two distinct co-designating terms anking the identity sign—as in, say,

[17]

Jews are Hebrews,

where Hebrews is an alternative non-pejorative term for Jews. The sentences [15] and [17] could generate a run-of-the-mill version of Frege’s puzzle. The one presented by Hom and May is, however, supposed to be atypical, resulting from something other than there being two distinct yet co-designating terms. So, bracketing the non-triviality due to that, the expressivist should simply deny that with respect to the p. 139

proposition expressed,

[16] is no less trivial than [15]. They both encode propositions that are recognizably

uninformative. What gives [16] its veneer of non-triviality (and it is non-trivial—just not at the level of propositions) is that the negative a ective stance encoded in Kike, when contrasted with Jew, is inevitably raised to salience. The e ect is compounded by the fact that in [16] Kike occurs on the right side of the identity sign. Just by transposing the order of the terms

[18]

Kikes are Jews

we can partly excise this e ect, resulting in a sentence that appears far more trivial. We may, similarly, excise the e ect of the di ering terms by substituting in what is, even according to Hom and May, a semantically equivalent expression to Kike:

[19]

C

Jews are Jews

and this does clearly ring of triviality with respect to the proposition-expressed. The second response e ectively starts the dialectic where we just ended up. Consider [19] or its companion sentence [20] employing the bare expressive fucking to generate pejoration:

[20]

Fucking Jews are Jews

Must we accept this as a well-motivated version of Frege’s Puzzle? Analogous versions would have to obtain for

[21]

Fucking dogs are dogs

[22]

Goddamn cats are cats

But these do not seem to be well-motivated instances of Frege’s puzzle. A perfectly acceptable—and plausible—expressivist response is just to deny that there is a genuine version of Frege’s puzzle here.

4.2. The Modal Conceivability Argument With their modal conceivability argument, Hom and May aim to establish that slurs and their neutral counterparts cannot have the same extension. Following the structure of familiar conceivability arguments concerning the extension of natural kind terms, their argument runs as follows. It’s conceivable for there to be Jews without being Kikes. Whatever is conceivable is possible. Therefore, it is possible for there to be Jews without being Kikes. p. 140

Yet if Jew and Kike have the same extension (and do so necessarily), as I and other identity expressivists maintain, it is not possible for there to be Jews without being Kikes. Thus, claim Hom and May, truthconditional equivalence of slur and neutral counterpart is incompatible with the modal conceivability argument’s conclusion. The way to respond to this argument is by challenging the conceivability claim. Why think it is conceivable for there to be Jews without being Kikes? Prima facie, identity expressivists and all identity theorists have no impetus to sign on to its conceivability for this would mean conceiving of a world in which there are, as they would put it, Jews but … no Jews. Hom and May motivate their premise by suggesting that we “consider a world that is mostly like ours, except that it is morally perfect, and so devoid of racism of any kind” and maintain that in that world, “it is 30

clearly conceivable for there to be Jews without Kikes.”

The relevance of such a world to the issue at hand

is doubtful. True, in this world devoid of racism, there will be Jews but no Jews subject to anti-Semitism, and thus no Jews who are called Kike. But that doesn’t contribute to establishing that it’s conceivable for there to be Jews but no Kikes. Return to our comparison between slurring terms and bare expressives. Assuming that semantically, bare expressives do not contribute any descriptive content, we would and should be unmoved by the idea that it is conceivable for there to be dogs without there being goddamn dogs. Imagining a world full of eternally patient dog lovers does nothing to establish otherwise. It only exhibits that there’s a world in which people wouldn’t say goddamn dog to vent their dislike of or frustration with dogs.

The deeper problem with the modal conceivability argument is that it encourages us to re ect upon the semantic properties of slurs by re ecting on the non-semantic question “what’s the di erence between 31

being a Jew and being a Kike?”

In my view, doing so prejudicially frames and warps the semantic

discussion. Resorting again to our comparison, it is in many ways analogous to what has gone wrong when, in investigating the semantics of intensi ers, to discern the di erence between

[23]

Get the dog out of the yard.

[24]

Get the fucking dog out of the yard.

we kick o

discussion by asking the non-semantic question, “What is the di erence between being a dog

and being a fucking dog?” This is not the right question to ask. In the next section on the Frege-Geach problem, I deepen our understanding of bare expressives to further underscore the rhetorical point here.

4.3. The Frege-Geach Problem The Projection Condition pro les how the derogation of slurs project out of various complex constructions p. 141

and do not interact with logical operators. I sketched above

why separating the derogating component of

slurs from their designational and truth-conditional contributions helps explain projection. Interestingly, some believe that identity expressivists confront a problem concerning projection. According to Hom and May (2013), identity expressivists run into a version of the Frege-Geach problem by treating slurs’ derogating semantic features in a way that parallels those of “non-racist pejoratives” and the derogating element of slurs as independent of truth conditional content. The Frege-Geach problem is canonically directed toward non-cognitivist treatments of moral terms like good and wrong. Non-cognitivists are united in regarding moral terms as lacking ordinary semantic descriptive content—in contrast with terms like ripe, round, and red. For them, moral terms possess only expressive meaning. So, very roughly, the meaning of “X is good” is something like the expression of an approving-attitude toward X; the meaning of “X is wrong” is the expression of a disapproving attitude toward X. The problem is that linguistic evidence suggests that moral terms at least sometimes appear to function like ordinary descriptive terms. To exhibit this, Geach advanced a simple argument form. Consider: Lying is wrong. If lying is wrong, getting your little brother to tell lies is wrong. Therefore, getting your little brother to tell lies is wrong. On the basis of logical and semantic grounds, the content of “lying is wrong” must have exactly the same meaning regardless of whether it occurs alone, as in the rst premise, or whether it occurs as the antecedent of the conditional, as in the second. Otherwise, the conclusion could not be derived. Yet, intuitively, there is nothing expressed—no disapproving attitude encoded—when “lying is wrong” occurs as the antecedent of 32

the conditional, as in the second premise.

Plainly, the linguistic phenomena here about the semantics of good and wrong di ers from expressivist accounts of slurs—for their derogating qualities are not blocked when they interact with logical operators: they project. So identity expressivism is not ruled out by projection considerations. But it is precisely this point that Hom (2010) and Hom and May (2013) challenge. To exhibit their Frege-Geach problem for slurs,

Hom and May o er a series of varied examples involving “non-racist pejoratives” in which “the derogatory element does not scope away, as predicted if the derogatory element were non-truth-conditional” (2013, 306).

Negation [25]

John didnʼt fuck the managing partnerʼs daughter. (He hasnʼt even met her.)

Conditionalization [26]

If John fucks up another case, then he will be fired. (But I donʼt think he will because heʼs working much harder now.)

p. 142

Event Quantification [27]

Every time someone fucks up a case, the senior partner blames John. (But as long as Iʼm not blamed for it, I donʼt care.)

Tense [28]

John was a fucker in law school, but he has improved since then. (I like him quite a lot now.)

They conclude that evidence for the interaction between the general pejorative and operator is evidence for regarding “expressive content as super uous” (2013, 304)—apparently for both racist and non-racist pejoratives alike. Their idea is that because fucks interacts with, say, negation in [25], it contributes to the sentence’s truth conditions. The same obtains for fucks up in [26] and [27] and fucker in [33]: each interacts with the logical operators and contribute therewith to the truth conditions. In reply, the rst thing to say is: there are f-bombs and there are f-bombs, and you have to be careful about which ones inspire expressivists about slurs. I and many expressivists about slurs are moved by Chris Potts’s analyses of bare expressives. We regard the expressive component of slurs as in some respects functioning semantically in the way that the f-bomb functions precisely when it is an a ective intensi er. For instance, the f-bombs in [29]–[32] all function as a ective intensi ers:

[29]

If Laura insists on getting a fucking dog, Iʼm moving out.

[30]

Every time someone fucking sneezes, I get sick.

[31]

John was a fucking lawyer, but now heʼs a philosopher.

[32]

That is really, really fucking brilliant! [Bono]

In [29]–[32], fucking contributes nothing to truth-conditional content; it contributes only, typically, something about the speaker’s attitudes. The negative a ect projects and it does not interact with the sentential operators. However, identity expressivists do not and should not align the expressive component of slurs with all other 33

lexical items having the same stem, root, or etymological origin as fuck.

In particular, we do not maintain

that slurs’ expressive component is akin to the f-bomb, qua verb to fuck, to indicate, coarsely, the activity of sexual intercourse, as in [25]. We do not maintain that slurs’ expressive component is akin to the f-bomb, qua verb to fuck-up, to indicate, coarsely, messing something up, an in [26] and [27]. And we do not maintain that slurs’ expressive component is akin to the f-bomb, qua noun fucker, to indicate, coarsely, well, whatever it is that goes into being a fucker … something like being an asshole, I presume, as in [28]. So, the fact that these varieties of f-bombs do interact with logical operators doesn’t establish or provide evidence for a Frege-Geach type problem. It’s worth pausing for a quick historical note corroborating our distinguishing fucking as an a ective p. 143

intensi er from other forms: In 1967, James McCawley

commenced investigation into the syntactic and

semantic properties of fuck. He distinguished two verbs, fuck 1 and fuck 2, the rst the copulating fuck exempli ed by Hom and May’s [30], the second the angry fuck exempli ed by fuck you, and demonstrated that they had radically di erent syntactic pro les. A few years later, a group of his students argued in a similar vein that fucking, as in our [29]–[32], is not a standard adjective, having entirely di erent syntactic 34

properties from regular adjectives.

Their work provides some early support for Potts’s general notion of a

bare expressive and our construing fucking in [29]–[32] as an a ective intensi er, to be sharply distinguished from the copulating fuck in [25] and meaning-extended ‘messes up’ fuck in [26] and ‘asshole’ fuck in [28]. This brings us to the second thing: Although I deny that Hom and May’s examples of f-bombs exemplify the expressive features of slurring terms, they do usefully illustrate how only one aspect of slurring terms’ meaning interacts with their truth conditional content. Take [25]. The only part of the meaning of fuck that interacts with negation is that which indicates having sex. [25] is truth conditionally equivalent with

[33]

John didnʼt have sex with the managing partnerʼs daughter.

In [26] and [27], the only part of the meaning of fuck-up that interacts with the conditional is that which indicates messing up. Both are truth-conditionally equivalent to sentences substituting messes up for fucks up. The coarseness inherent in all is separate, and just as present in “John didn’t fuck the partner’s daughter” as it is in “John did fuck the partner’s daughter”. The coarseness projecting out would therefore

need to be explained separately—either as a matter of the expressions’ semantics or more probably in these examples, at the pragmatic level, as a matter of their register. This in fact aligns with identity expressivism’s proposal: that slurs possess truth conditionally relevant content and, along a separate dimension, a derogating expression of contempt, which projects. In sum: Hom and May’s Frege-Geach problem rests on con ating syntactically distinct lexical items, a ective intensi ers and verbs and nouns that contribute to truth conditional content. They thereby fail to carve out a genuine Frege-Geach problem for identity theorists of slurs. However, because the syntactically distinct lexical items in their series of example all possess a similar coarseness, they e ectively underwrite our view by displaying another variety of non-truth-conditionally relevant meaning to be accounted for, most likely in terms of register or coloring.

5. Conclusion In this paper, I have argued that null extension non-identity theories are linguistically inadequate and Richard’s rationale for denying slur-containing utterances truth value is suspect. I have defended p. 144

expressivist identity theories from the charges that they have

special problems rooted in Frege’s

problem, Modal Conceivability, and the Frege-Geach problem. Identity theories do indeed entail that because it is true that there are Jews, so, too, is it true that there are Kikes—an awful thing to write and say. But the right response to this observation is not to reject identity theories. What’s far better, we’ve learned, is to say that I, we, would never put these truths this way. We would instead say the boring: since it is true that there are Jews, so, too, is it true that there are Jews. Moreover, the entailment is not as noxious as it may have initially appeared. For we’ve seen from slurs’ derogation in questions, imperatives, and vocatives, denying truth to simple predications with slurs does not ultimately confront the general perniciousness of slurs, so there is nothing ultimately to be gained by doing so.

Acknowledgments I presented some material from this paper at Harvard 2010, at the 2010 Southern California Philosophy Conference, and at a 2011 Paci c APA symposium, and presented earlier versions of most of the chapter in my second Wedberg Lecture, 2015, Stockholm, and at LOGOS, University of Barcelona, 2015. Many thanks to all who attended. For extensive feedback and challenges, special thanks to Jonas Åkerman, Luvell Anderson, Esa Diaz-Leon, Manolo García-Carpintero, Kathrin Gluer-Pagin, Ned Hall, Anandi Hattiangadi, Teresa Marques, Michael Nelson, Peter Pagin, Mark Richard, Ken Taylor, and Åsa Wikforss. Deep gratitude to Luvell Anderson for his superhuman patience and warm encouragement.

Notes 1.

Richard (2008, 13): “ ʻFrogʼ is a derogatory term for the French, and a derogatory term for the French is a term for the French. But if S is a term for the French and f is in fact French, then if I point at f and utter He is an S, I speak truly.”

2.

Anderson and Lepore (2013).

3.

Bolinger Jorgensen (2017); Nunberg (2018); Pullum (2018); Swanson (2022).

4.

Camp (2013, 2018); Rappaport (2019).

5.

Potts (2005, 2007); Jeshion (2013, 2016, 2018), Marques and Garcia-Carpintero (2020).

6.

Hom (2008, 431).

7.

Hom and May (2013, 298; 2018). Additional null extension Identity Theories are in Bach (2018) and Diaz-Leon (2020). I discuss Bachʼs view in Jeshion (2017). Another null extension view is due to Neufeld (2019), who o ers an essentialist model whereby slurs represent that thereʼs an essence of a group that is causally responsible for negative features stereotypically associated with the group. For her, slurs are extensionless because there is no essence that is causally responsible for such stereotypical features.

8.

Hom and May (2013, 299) advance extensive reasons for thinking that slurs, like terms of myth, have empty extensions. Their chief claim is that these classes of terms are lexically marked at an abstract level, the slurs as pejoratives, the terms like unicorn, unicorn horn, and witch as expressions within magical discourse. Terms of magical discourse are so loaded with ontologically and supernaturally false assumptions about the worldʼs causal structure that they necessitate null extensionality. Pejoratives are similarly so “loaded with false assumptions about the social and ethical structure of the world” that they “drain extensionality.” While I find the analogy between the two types of expressions uncompelling, it would take extensive argument to demonstrate why. For discussion, see Marques (2017).

9.

Hom and May work the unicorn horn analogy to make this point against Sennet and Copp: that best-fit semantics plus anti-Semitesʼ actual and dispositional patterns of use of Kike do not entail that Kike refers to Jews, just as it does not guarantee that unicorn horn refers to (the non-existent) unicorn horn.

10.

Cf., Jeshion (2021) for a more extensive discussion on why itʼs necessary to distinguish slurring words from slurring speech acts that function to derogate on the basis of groups.

11.

Linguistic Encoding of Derogation Condition is controversial for semantic minimalists like Anderson and Lepore (2013), who account for all that di erentiates slurs and neutral counterparts extra-linguistically, by reference to prohibitions instituted on their use, and for pragmatic a iliationists like Bolinger Jorgensen (2017), Nunberg (2018), and Swanson (2022), who regard slurs as ideology-cues or ways of signaling ideological a iliation.

12.

That is, they are not functioning akin to Homʼs “There are no Chinks at the University of California. There are only Chinese persons,” which would be made to echo or respond to prior salient events in which Chinese persons are called Chinks or experience other forms of racism. Here, I appeal only to intuitive notions of metalinguistic content, discourse denials, and echo. A rigorous analysis requires evaluating specific accounts of metalinguistic negation and denial, which are controversial. Cf., Horn (1985); van der Sandt and Maeir (2003); and Davis (2011). For classical views on echo, see Wilson (2006); Sperber and Wilson (1986). I discuss some of the Hom cases briefly in my (2013).

13.

The Compositionality Condition of course allows that logical operators can be used in metalinguistic denials like ʻThere are no Chinks, only Chinese peopleʼ, group-extending uses like ʻThat lawyer is such a kike – but no, heʼs not a Jewʼ and groupcontracting uses, as in ʻMy best friend is a Jew, but heʼs not a Kikeʼ. The condition itself only requires that in sentences like [2]–[9], the logical operators interact with the slurʼs representation of the target group. Cf. Jeshion (2013) for more on Gcontracting and G-extending uses of slurs.

14.

Hom and May (2013, 310).

15.

Hom (2008) and Hom and May (2013) claim such conversational implications and presuppositions o end, but do not derogate. Here, I am using ʻo enseʼ di erently, to indicate a response to a slur, not what the slur means or how it functions to derogate. Nothing hinges on this. I am allowing that such presuppositions constitute derogations and satisfy our Projection Condition.

16.

Cepollaro and Thommen (2019).

17.

Cf., also Kaplan (2005) for expressivist fulfillment of the Derogation Condition. Kaplan grapples with the issues about truth conditions, without a conclusive semantic theory.

18.

Richard (2008, 13).

19.

Richard (2008, 14).

20.

Richard (2008, 37).

p. 145

p. 146

21.

Richard (2008, 24) is here criticizing Gibbard (2003, 298–302) for maintaining that slurs di er from thick concepts, and that the classification that slur users share with those that use their neutral counterparts is su icient for their thoughts and utterances of them being truth evaluable.

22.

Richard (2008, 24).

23.

Richard (2008, 25; my emphasis).

24.

An alternative interpretation—that what deprives the slurring thought and speech of truth value is in part the moral wrong in thinking as the bigot thinks—attributes to Richard a kind of moral encroachment into semantics. This alternative reading seems implausible insofar as it neglects Richardʼs emphasis on misrepresentation.

25.

Thanks here to Ned Hall, who o ered a similar example during the question period in my 2010 Harvard talk.

26.

In this example, thereʼs no particular word or concept that incorporates or necessarily triggers the subjectʼs a ective attitude. While this may be a relevant di erence, Richard himself does not appeal to either, only to having a thought involving misplaced a ect.

27.

Hom and May also argue that expressivist identity theories are saddled with problems explaining indirect quotation and attitude attributions. I leave these to the side because I discuss similar objections due to Kent Bach in my (2017, sec. 4). There I advocate an analysis on which indirect quotation of utterances containing slurs are ambiguous or undetermined from the reported sentence alone whether the speaker is expressing her own attitudes or saying something about the attitudes of the one whose words she is reporting. I argue that positing such an ambiguity is not ad hoc, for it is necessary to handle indirect quotation of bare expressives as well, as in ʻJason said that the cat hissed at the goddam dog.ʼ On one reading, it encodes Jasonʼs negative attitudes about the dog, on another, the speakerʼs. I also argue there that attitude attributions involving slurs are fruitfully seen as akin to free indirect discourse.

28.

Hom and May (2014, sec. 2).

29.

Hom and May (2014, 117).

30.

Hom and May (2013, 305).

31.

Bach (2015) o ers just this question to motivate his Loaded Descriptivism.

32.

For extensive discussion of the development of the problem and non-cognitivistsʼ attempted solutions, cf., Schroeder (2008).

33.

Hom and May also o er sentences containing bastard: John is not a bastard. (Heʼs extremely nice.) Bastard is even less plausible as a linguistic item inspiring an expressivist view of slurs.

34.

For this content, McCawley wrote under the pseudonym Quang Phuc Dong, in his epic Dong (1967). Zwickey et al. (1992) contains reprints of McCawleyʼs 1967 essays, as well as his studentsʼ responses to and developments of them.

References Anderson, Luvell, and Lepore, Ernie (2013), “Slurring Words,” Noûs, 47, 25–48. Google Scholar WorldCat Bach, Kent (2018), “Loaded Words: On the Semantics and Pragmatics of Slurs,” in David Sosa, ed., Bad Words. Oxford: Oxford University Press, 60–76. Google Scholar Google Preview WorldCat COPAC Bolinger, R. Jorgensen (2017), “The Pragmatics of Slurs,” Noûs, 51(3), 439–462. Google Scholar WorldCat Camp, Elisabeth (2013), “Slurring Perspectives,” Analytic Philosophy, 54(3), 330–349. Google Scholar WorldCat Camp, Elisabeth (2018), “A Dual Act Analysis of Slurs,” in David Sosa, ed., Bad Words. Oxford: Oxford University Press, 29–59. Google Scholar Google Preview WorldCat COPAC Cepollaro, Bianca, and Thommen, Tristan (2019), “Whatʼs Wrong with Truth Conditional Accounts of Slurs?” Linguistics and Philosophy, 42, 333–337. Google Scholar WorldCat p. 147

Copp, David, and Sennett, Adam (2015), “What Kind of a Mistake Is It To Use a Slur?” Philosophical Studies, 172 (4): 1079–1104. Google Scholar WorldCat Davis, Wayne (2011), “ʻMetalinguisticʼ Negation, Denial, and Idioms.” Journal of Pragmatics, 43(10), 2548–2577. Google Scholar WorldCat Dong, Quang Phuc (1967), “English Sentences without Overt Grammatical Subject.” Conneries Linguistiques, 19: 23–31. Google Scholar WorldCat Hom, Christopher (2008), “The Semantics of Racial Epithets,” Journal of Philosophy, 105 (8): 416–440. Google Scholar WorldCat Hom, Christopher (2010), “Pejoratives,” Philosophy Compass, 5(2), 164–185. Google Scholar WorldCat Hom, Christopher, and May, Robert (2013), “Moral and Semantic Innocence,” Analytic Philosophy, 54(3), 293–313. Google Scholar WorldCat Hom, Christopher, and May, Robert (2014), “The Inconsistency of the Identity Thesis,” ProtoSociology, 31, 113–120. Google Scholar WorldCat Hom, Christopher, and May, Robert (2018), “Pejoratives as Fiction,” in David Sosa, ed., Bad Words. Oxford: Oxford University Press, 108–131. Google Scholar Google Preview WorldCat COPAC Jeshion, Robin (2013), “Expressivism and the O ensiveness of Slurs,” Philosophical Perspectives, 27, 231–259. Google Scholar WorldCat Jeshion, Robin (2017), “Loaded Words, Expressive Words: Assessing Two Semantic Frameworks for Slurs,” Croatian Journal of Philosophy, 18(50), 111–130. Google Scholar WorldCat

Jeshion, Robin (2021), “Varieties of Pejoratives,” in Justin Khoo and Rachel Katharine Sterkin, eds., The Routledge Handbook of Social and Political Philosophy of Language. New York: Routledge, 211–231. Google Scholar Google Preview WorldCat COPAC Kaplan, David (2005), “The Meaning of ʻOuchʼ and ʻOopsʼ”, unpublished manuscript. Marques, Teresa (2017), “Pejorative Discourse Is Not Fictional,” Thought: A Journal of Philosophy, 6(4), 250–260. Google Scholar WorldCat Marques, Teresa, and Garcia-Carpintero, Manuel (2020), “Really Expressive Presuppositions and How to Block Them,” Grazer Philosophische Studien, 97(1), 138–158. Google Scholar WorldCat Neufeld, Eleanor (2019), “An Essentialist Theory of the Meaning of Slurs,” Philosophersʼ Imprint, 19(35), 1–29. Google Scholar WorldCat Nunberg, Geo rey (2018), “The Social Life of Slurs,” in Daniel Fogel, Daniel W. Harris, and Matt Moss, eds., New Work on Speech Acts. Oxford: Oxford University Press, 237–295. Google Scholar Google Preview WorldCat COPAC Potts, Christopher (2005), The Logic of Conventional Implicatures. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Potts, Christopher (2007), “The Expressive Dimension.” Theoretical Linguistics, 33(2), 165–198. Google Scholar WorldCat Pullum, Geo rey (2018), “Slurs and Obscenities,” in David Sosa, ed., Bad Words. Oxford: Oxford University Press, 168–192. Google Scholar Google Preview WorldCat COPAC Rappaport, Jesse (2019), “Communicating with Slurs.” The Philosophical Quarterly, 69, 795–816. Google Scholar WorldCat Richard. Mark (2008), When Truth Gives Out. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Schroeder, Mark (2008), “What Is the Frege-Geach Problem?” Philosophy Compass 3(4), 703–720. Google Scholar WorldCat Sperber, Daniel, and Wilson, Deirdre (1986), Relevance: Communication and Cognition. Oxford: Blackwell. Google Scholar Google Preview WorldCat COPAC Swanson, Eric (Forthcoming.), “Slurs and Ideologies,” in Robin Celikates, Sally Haslanger, and Jason Stanley, eds., Rethinking Ideology. Oxford University Press. Google Scholar Google Preview WorldCat COPAC van der Sandt, Rob, and Maeir, Emar (2003), “Denials in Discourse,” unpublished manuscript. p. 148

Wilson, Deirdre (2006), “The Pragmatics of Verbal Irony: Echo or Pretence?” Lingua, 116, 1722–1743. Google Scholar WorldCat Zwickey, Arnold, Salus, Peter, Binnick, Robert, and Vanek, Anthony, eds. (1992), Studies out in Le Field: Defamatory Essays Presented to James McCawley on the Occasion of His 33rd or 34th Birthday. Philadelphia: John Benjamins. Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

p. 149

7 Slurring Words and Slurring Articulations  Una Stojnić, Ernie Lepore https://doi.org/10.1093/oxfordhb/9780192844118.013.7 Published: 22 May 2024

Pages 149–174

Abstract Slurs are epithets that denigrate a group simply on the basis of membership, e.g., on the basis of race, ethnicity, origin, religion, gender, or ideology. They provide powerful linguistic weapons, carrying a characteristic pejorative sting or punch, prone to cause o ense, outrage, and even injury. So much so, that they can be subject to media censorship, and sometimes even legislation. As to the nature and source of their characteristic sting, the predominant position is to invoke some aspect of meaning— either semantically encoded or pragmatically conveyed. It is because of what a particular slur term means or conveys that it is suitable to be thus weaponized. Consequently, most e orts at understanding pejorative language have been attempts to characterize the meanings and how these meanings compose with the meanings of other sorts of expressions. The few who reject this assumption locate the source of o ense in the taboo status of pejorative language. In other words, the slurs themselves and/or their associations are the source of their o ensive sting, not their meanings. The chapter challenges both sorts of approaches and defends a novel alternative according to which the source of a pejorative e ect is negative associations triggered, not by slurs, but rather, by certain articulations of these expressions—phonological or orthographic. We need to distinguish slurs from their articulations because, surprisingly, the latter can trigger an e ect even when the former is not tokened, and even when articulated, a slur can lose its o ensive potency if its articulation is standard, or so we will argue.

Keywords: slurs, predication, presupposition, conventional implicatures, expressives, tone, perspective, prohibition, articulations Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction Slurs are particularly in ammatory expressions with a distinctive o ensive potential; they target groups on the basis of race, nationality, religion, gender, sexual orientation, origin, immigrant status, or other *

demographics. A central goal in the study of slurs is to explain why their uses tend to o end. The consensus 1

has been that it is because they convey or otherwise express a semantically encoded pejorative content.

Contemporary semantic theories teach us that there are various ways an expression can encode semantic content; and so, the investigation begins by asking whether slurs carry a potential to o end in any of the known ways for encoding contents. Can slurs o end because of what they predicate of, or presuppose, conventionally implicate, or otherwise semantically express (as opposed to describe) about, their targets? Most theorists answer a

rmatively and, thus, are committed to slurs having the potential to o end because of

the contents they encode; they diverge only over the mechanisms of implementation. Our aim in section 2 is to survey this ever-growing body of literature with a critical eye on this consensus: can any of these content strategies explain why uses of slurs have the potential to o end? In section 3, we will present a relatively recent alternative proposal, Prohibitionism, according to which slurs harbor a potential to o end because their occurrences are prohibited (Anderson and Lepore 2013a, 2013b). Although this proposal, we believe, is on the right track, it comes up short as is. In section 4, we will critically turn to the Pejorative Tone Account (Lepore and Stone 2018), according to which slurs are potentially o ensive because they are carriers of “pejorative tone.” Roughly, whenever a slur is tokened, its pejorative tone is triggered, thereby potentially provoking an audience. This account, we believe, is also on the right track, though in a di erent respect than Prohibitionism; but, we will argue, it shares its shortcomings. Indeed, in p. 150

section 5, we will look at cases where a non-slur term, solely in virtue of sharing an

articulation of a slur,

or having a resembling articulation to that of a slur, inherits a potential to o end, as re ected, for instance, in the heated exchange over the o ensiveness of an adverb which bears phonetic (and orthographic) similarity to a highly charged racial slur targeting African Americans, the N-word, but which is well-known to be etymologically and semantically unrelated (Kennedy 2002, 94–95). This, we argue, is indicative of the key weaknesses in all familiar accounts of slurs, save for our preferred one (Stojnić and Lepore 2022, forthcoming). In short, we argue, neither the content of slurs, nor their prohibition, nor any concomitant negative tone they might carry, alone can account for their o ensive potential. In their place, we will present an account that captures both the o ensive potential of slurs and the inherited one of unrelated, but resembling, expressions, by arguing that the trigger of the o ensive potential is an articulation of a slur, which can be 2

present even when the slur is not (as well as absent when the slur is tokened).

2. Encoded Meaning In this section, we will brie y canvass several familiar ways in which slurs have been thought to carry the potential to o end in virtue of semantically encoding certain contents—namely, via predicative, presuppositional, conventionally implicated, or expressive content.

2.1. Predicative Accounts Words like ‘African American,’ ‘gay,’ ‘Asian,’ ‘Mexican,’ ‘Italian,’ ‘Irish’ are all neutral counterparts to 3

familiar slur terms; they lack a distinct potential to cause o ense. Those who want to trace the o ensive potential of slurs to the meanings they encode qua predicates—i.e., as a matter of truth-conditional content 4

—must hold that, though slurs and their neutral counterparts are coextensive, they are not synonymous. A slur and its corresponding neutral counterparts pick out the same target group, but not in the same way, much like ‘equilateral triangle’ and ‘equiangular triangle.’ So, according to the Predicative Content Hypothesis, the reason why “John is African American” is not o ensive in the way that the sentence with ‘African American’ replaced by the N-word is, must be because these two expressions encode di erent predicative contents; predicative uses of slur terms supplement the meanings of their neutral counterparts with additional, pejorative contents over and above that of the neutral counterpart. One challenge for the simple view just sketched is that if a slur and its neutral counterpart are coextensive, then assuming the neutral counterpart just picks out a target group, the supposedly o ensive content of the 5

slur would have to be true of that group as well. But this seems problematic; indeed, many contributors p. 151

report the intuition that

predications of slurs, even when applied to target-group members, are false (or,

6

at least, not true). Accordingly, they insist, (1) is either truth-valueless or false:

(1)

7

Hermione is a mudblood.

In response, some Predicativists have proposed a variant of the account, according to which slurs have null extension, much like, say, ‘unicorn’ (Hom 2008; Hom and May 2013; Neu eld 2019). For Hom (2008), the semantic content of a slur is speci ed by a schema “ought to be subject to p*1 + … + p*n because of being d*1 + … + d*n all because of being NPC*,” where p*1 + … + p*n specify some discriminatory treatment, d*1 + … + 8

d*n some derogative properties, and NPC* is the neutral counterpart of the slur.” Slurs with such meanings are taken to determine a null extension. However, this version of Predicativism faces the ipside problem, for there is also an opposing intuition that some utterances of slurs can be true, as well as non-trivially false. Camp (2013) points out that it is clear what it takes to win a (bigoted) bet of the following kind:

(2)

9

I bet you the next hire will be a mudblood.

The view that would hold that slurs have empty extensions would predict that accepting any such bet is guaranteed to result in a win. But that’s wrong: if the next hire is a muggleborn, one would lose the bet to the speaker of (2) if one were to accept it. You win i the next hire is a non-muggleborn or a non-wizard. This points to a dilemma for a Predicativist: in order to capture that utterances containing slurs can have non-trivial truth-conditions (as required for non-triviality of (2)), we cannot posit that slurs in general have null extensions; but if we are to capture that slurs get something wrong, that they misrepresent the target group, we cannot maintain that a slur is coextensive with its neutral counterpart, but also predicates some o ensive content of the target group. Worse, as it turns out, Predicativism doesn’t fare well against a number of well-known tests for determining whether a certain content is predicatively encoded. One such test (among many) examines the

10

behavior of a predicate within the scope of linguistic negation, and licensing of direct denial.

So, for

example, anyone who wants to disagree with what (3) ascribes to Hermione can do so by negating it (as in 3a) or with a simple denial (as in 3b):

(3) Hermione is a muggle-born wizard. a.

Hermione is not a muggle-born wizard.

b.

Thatʼs false.

On the basis of this test, we surmise that ‘is a muggle-born wizard’ is predicating being a muggle-born wizard to Hermione. But if replacing ‘muggle-born wizard’ with the slur ‘mudblood’ results also in predicating some pejorative content to her (or to the target group she is alleged to be a member of) because p. 152

of what the occurrence of the slur, but

not its neutral counterpart, predicates of her, then we should be

able to discharge this potential o ensive content by negating or denying the resulting slurring sentence, as in (4a)–(4b). Yet the widespread intuition is that any such negation and denial merely denies the group membership, but remains as in ammatory as the corresponding a

11

rmation.”

(4)   Hermione is a mudblood. a.

Hermione is not a mudblood.

b.

Thatʼs false.

These considerations lead us to conclude that the o ensive potential of a slur cannot lie in its predicative content.

2.2. Presuppositional Accounts An alternative proposal for how slurs encode potentially o ensive content is through a mechanism of presupposition (Schlenker 2007; Cepollaro 2015; Cepollaro and Stojanovic 2016). Utterances that feature presupposition triggers are only appropriate if the presupposition triggered is accepted in the context—if it is treated as common ground by the interlocutors (Stalnaker 1973, 1974). For instance, utterances of (5a)– 12

(5c) are all appropriate only if it is mutually accepted among interlocutors that John used to smoke.

(5) a.

John has stopped smoking.

b.

Has John stopped smoking?

c.

John has not stopped smoking.

We say that the speaker in using ‘stopped smoking’ presupposes John used to smoke. And this presupposition is said to “project” out of the questions or negation: notice that even in (5b), while the speaker is not making an assertion, they nevertheless incur a commitment to the presupposed content. This might seem promising for the Presuppositional Accounts of slurs: since, in normal circumstances, it is not possible to ask, “Is John an X?” or assert either “John is an X” or “John is not an X”, where X is a slur, without potentially o ending the target group, perhaps, the o ensive e ect is encoded as matter of presuppositional content. There are various ways one can spell out a presuppositional account, contingent on how one chooses to 13

characterize presupposed content.

For instance, Schlenker (2007) argues that slurs presuppose, roughly,

that the speaker has a negative attitude toward the target group, e.g., that they believe the target group is despicable. Stojanovic and Cepollaro (2016) suggest that the presupposition should be characterized in p. 153

14

“objective” terms: the slur presupposes that the target group is despicable.

In either

case, an utterance

containing a slur will signal that the speaker is committed to the presuppositional content. Whatever its attractions, the thesis that slurs presuppose, as common ground among interlocutors, a potentially o ensive content, is vulnerable to a number of objections. For one, the speaker needn’t have any negative attitudes towards the target group in uttering the slur—and they might coherently deny that they do:

(6) a.

Mudbloods are my favorite wizards!

b.

Mudbloods are actually very respectable wizards and not despicable at all.

c.

I donʼt think mudbloods are despicable.

Such utterances, if o ensive, are not self-undermining, as would be an utterance of ‘John stopped smoking, 15

but he never smoked.’

Further, the Presuppositional Accounts fail to predict that the o ensive e ect of slurs projects from a wider range of embeddings than presuppositions do. It is well known that certain linguistic environments behave as “presupposition lters,” or “plugs,” blocking or ltering out the presupposition from projecting to the 16

whole sentence.

For instance, when a trigger is embedded in the consequent of a conditional, the

presupposition fails to project if the antecedent entails it: (7) does not presuppose that John used to smoke, despite the presupposition trigger ‘stopped.’

(7) If John used to smoke, then he has stopped smoking.

It is similarly well-known that indirect reports plug the presuppositions of their complement clauses: (8a) does not presuppose that John used to smoke, which is why one can consistently utter it while denying that he did (as (8b) illustrates). Whoever utters (8b) is not trying to get his audience to go along with the proposition that John used to smoke.

(8) a.

Frank said that John stopped smoking.

b.

Frank said that John stopped smoking, but John has never smoked.

But then, if slurs trigger their o ensive e ect as a matter of presupposition, this e ect should be canceled, or ltered out, under presupposition plugs, or lters. Yet, this prediction is not borne out in practice. (9a)– 17

(9b) inherit any o ensive e ect that ‘mudblood’ normally triggers in simple predications like (1).

(9)

p. 154

a.

If muggleborn wizards are despicable on account of being muggleborn, then Hermione is a mudblood.

b.

Draco said Hermione is a mudblood, but muggleborn wizards are not despicable on account of being muggleborn.

The moral is that slurs involve a distinctively robust projective behavior, much stronger than ordinary presupposition triggers. And so, slur terms cannot carry a potential to o end through presuppositional content.

2.3. Conventional Implicature Accounts Some contributors propose that pejorative content is encoded, not as a matter of presupposition, but rather, 18

as a matter of conventionally implicated content.

The notion of conventional implicature (CI) goes back to

Grice (1961, 1989). Grice observed that sentences like (10a)–(10b) di er in meaning, but are nevertheless truth-conditionally equivalent: they are true in exactly the same circumstances.

(10) a.

John is tall but handsome.

b.

John is tall and handsome.

The di erence in meaning comes down to the fact that (10a) draws a contrast between being tall and being handsome, whereas (10b) does not. And while it is not a part of the truth-conditions, this contrast is still conventionally encoded in the lexical meaning of ‘but.’ For one, anyone who understands the meaning of 19

‘but’ understands that it signals this contrast.

But further, the contrast is detachable—one can express the

same truth-conditions in di erent wording without conveying contrast, say, by replacing (10a) with (10b)— and it is not cancelable—if one uses ‘but,’ one can’t avoid the commitment to the implicated contrast by explicit denial; such attempts, as in (11), sound incoherent:

(11) John is tall but handsome, and thereʼs no contrast between being tall and being handsome.

Williamson (2009) argues that the o ensive content of a slur is encoded in the same way the contrastive content of ‘but’ is. This would explain why the o ensiveness is hard to cancel; one cannot cancel the o ensiveness of a slur by following up with an explicit denial of the supposed conventionally implicated 20

content:

(12) Hermione is a mudblood, but muggleborn wizards arenʼt inferior on account of being muggleborn.

Further, this would explain why the o ensiveness persists even in some complex embeddings, for instance, in questions or antecedents of conditionals; for the same holds for the contrast encoded in ‘but’:

(13)

p. 155

a.

Is John tall but handsome?

b.

If John is tall but handsome, then heʼd be a good cast for the role.

Further, if (10a) and (10b) are truth-conditionally equivalent, the falsity of the implicature would not a ect the potential truth of the asserted content. This would capture the sense in which the pejorative both gets

something wrong and something right: (1) would be true just in case Hermione is a muggleborn wizard, and (given the facts of the ction in question) it would be true regardless of the falsity of the claim that 21

muggleborn wizards are inferior or despicable on account of being muggleborn (Williamson 2009).

Moreover, conventional implicatures cannot be targeted with direct denial. Saying “That is false” in response to (10a) cannot be understood as a denial of the contrast between handsomeness and height. The analogy with slurring should be obvious: as noted before, saying “That’s false” in response to a slurring predication is not an adequate denial of its o ensive potential. Further, conventional implicatures, unlike presuppositions, are standardly taken to project out of 22

presupposition plugs; indeed, they have been claimed to always have the widest scope (Potts 2005).

We

have seen that slurs also share this feature, projecting even out of ‘says that’ reports. These commonalities between traditional CI items and slur terms certainly seem to support a Conventional Implicature proposal for how slur terms encode pejorative content. But there remain problems with the account. For one, the explicit denial of a CI is supposed to reveal a kind of linguistic incoherence—(11) not only fails to cancel the commitment to the implicated content but sounds incoherent, as if the speaker doesn’t understand the word ‘but.’ By contrast, while (14) as a whole still triggers the o ensive e ect of the slur, it’s not incoherent in that same way:

(14) Hermione is a mudblood, but I donʼt mean she is inferior/despicable on account of being muggleborn.

Indeed, someone reared in a bigoted community that only ever uses a slur term for muggleborn wizards, but who denounces this world view, could coherently say something like (15):

(15)   Mudbloods arenʼt inferior on account of being muggleborn! Itʼs a disgrace we treat them that way!

(14) and (15) should be not only o ensive, but linguistically confused in the way (11) is. But they don’t 23

appear to be.

2.4. Expressivist Accounts A related sort of content strategy—exempli ed by Expressivist Accounts—identi es the o ensive potential of a slur with a conventionally encoded expression of the speaker’s negative attitude or sentiment toward a target group (beyond content predicating group membership (Potts 2007; McCready 2010, inter alia)). So construed, this sort of account maintains that the expressive content conventionally conveyed by a slur is p. 156

like a regular

CI, inasmuch as its o ensive e ect projects out of complex embeddings, including

presupposition plugs. Further, since the Expressivist Accounts identify the o ensive content, not with negative propositional content, but rather with an expression of a negative attitude or sentiment, they avoid the challenge of trying to fully specify o ensive content in descriptive, propositional terms. This feature helps explain why a use of a slur can coherently occur in conjunction with the denial of any particular descriptive content. However, a version of the problem that faces CI accounts extends to the Expressivist Accounts as well. The problem, simply put, is that a speaker can utter a slur without having any (occurrent or standing) negative attitude or sentiment toward the target group; such utterances, while bigoted, are surely not linguistically awed. Recall (6a)–(6c).

(6) a.

Mudbloods are my favorite wizards!

b.

Mudbloods are actually very respectable wizards and not despicable at all.

c.

I donʼt think mudbloods are despicable.

Whatever is o

with (6a)–(6c), they are not linguistically deviant. Perhaps, a more sophisticated version of

Expressivism would fare better, one that maintains that what is conventionally signaled is not an expression of a sentiment or attitude, but rather an expression of a perspective about the target group (Camp 2013). According to Camp, a perspective is an intuitive way of organizing thoughts and feelings. Her proposal is that a slur conventionally encodes (beyond predicating group membership) an expression of a perspective, by which she means an expression of a commitment to “an integrated, intuitive way of cognizing members of the targeted group” (Camp 2013, 335). A perspective itself need not involve any speci c truth-conditional content, but, rather, serves as “a lens for integrating and explaining truth-conditional contents,” and it need not involve any particular feeling, but may motivate “certain feelings as appropriate” toward the target group. Camp further writes, “Whether sincerely or disingenuously, the speaker of a slur who invokes cancellation exploits the indeterminacy, open-endedness, and abstractness of perspectives to disavow any particular factual or a ective commitment” (2013, 343). This, Camp argues, is why utterances of (6a)–(6c) can be coherent, albeit still o ensive. Her idea is that since perspectives are not characterized by any speci c attitude or emotion, it is possible to token a slur while denying any particular attitude or emotion toward a target group. Just the same, it is not entirely clear to us that Camp’s view fully holds up under scrutiny. If slurs conventionally signal perspectives, then, in some sense, the perspective associated with a slur is (or, at least, partially determines) part of its meaning; so, it would seem that insofar as someone uses a slur in a discourse that conveys content radically un tting of that perspective, they should be guilty of a linguistic error. But it is easy enough to imagine someone raised in a bigoted community who only has a slur in their vocabulary for making reference to a particular target group. We can further imagine that this speaker has

p. 157

also come to denounce the complete standard perspective of their

community as a way of thinking about

that particular group. This speaker might then go on to say something along the following lines:

(16) The perspective with which my community thinks about/sees mudbloods is bigoted and wrong.

While this choice of words still carries o ensive potential in (16), the utterance is not linguistically awed. But presumably, the speaker doesn’t subscribe to the negative perspective toward muggleborn wizards, allegedly conventionally signaled by the slur, nor do they think of it as appropriate: they explicitly deny this. Even once we grant that one can deny any speci c content or emotion yet still signal commitment to a perspective, presumably one can signal one does not endorse a perspective as a whole if one expresses sentiments and thoughts su

ciently radically opposed to, or un tting of, the perspective in question.

Be that as it may, there are further signi cant problems that face all Expressivist Accounts, even those of a perspectival kind; indeed, ones that extend to all content-based strategies for accommodating the o ensive potential of slurs. We turn to them directly.

2.5. Innocent Uses of Slurs? So far, we have seen that slurs exhibit robust projective behavior, with their o ensive e ect projecting even out of presupposition plugs and also resisting cancelation. But the full extent of their projection is even more drastic. Slurs tend to retain their o ensive potential even in such linguistic environments in which any type of content—truth-conditional, presuppositional, conventionally implicated, or expressive—is rendered inert (and, as we will show later, even when no slur and, therefore, no meaning of a slur, can be in play). It has been observed in the literature that the o ensive potential of slurs persists even in meaning attributions (Anderson and Lepore 2013a, 36; Hornsby 2001, 130; Saka 2007, 122; Williamson 2009, 142). For instance, replacing either or both occurrences of “African-American” with the N-word in meaning attributions like “ ‘African American’ means African American,” according to many, can be o ensive. But content-based accounts have no way of explaining this. For, as Anderson and Lepore (2013a) point out, encoded contents are, in a relevant sense, semantically inert inside meaning attributions:

(17) a.

“ ʻBinyamin is Jewishʼ means Binyamin is Jewish” predicates nothing of Binyamin.

b.

“ ʻJohn is tall but handsomeʼ means John is tall but handsome” draws no contrast between height and handsomeness.

c.

“ ʻJoe stopped smokingʼ means Joe stopped smoking” doesnʼt presuppose Joe once smoked.

d.

“ ʻouchʼ means ouch” does not express a state of mind (e.g., pain).

e.

“ ʻvousʼ means vous” does not commit the speaker to addressing anyone in a formal/polite manner nor does it signal 24 politeness.

p. 158

The at-issue truth-conditional content (viz., 17a), conventionally implicated content (viz., 17b), presuppositional content (viz., 17c), and expressive content (viz., 17d and 17e) all are rendered inert in meaning attributions. So, if the o ensive potential of slurs is encoded as a part of content, then the potential o ensiveness of meaning attributions remains mysterious. Nor can it be explained by saying the words on the right-hand 25

side are (implicitly) mentioned (not used);

as Anderson and Lepore (2013a) point out, we can substitute

synonyms inside meaning attributions salva veritate, as in “ ‘bachelor’ means unmarried man”— establishing they’re not mentioned. So, how could a content strategist explain why slurs inside these sorts 26

of attribution (ever) have the power to o end?

This is probably a good juncture at which to reinforce a distinction that is often con ated in the literature, but which must be respected in discussions of the o ensive potential of slurs—namely, the intent behind a tokening of a slur versus its pejorative e ect, that is, the kind of “viscerally palpable” (Camp 2013), cringeinducing, sting one typically experiences when confronted with a slur. As we have seen, some authors dismiss or play down the o ensive potential of slurs in meaning attributions or quotative environments, arguing that any o ense taken in these contexts must be a result of confusion. Indeed, some argue that some uses of slurs are similarly insulated from o ensive potential; for some uses are indispensable for making a particular—e.g., pedagogical, or artistic—point, and due to this indispensability, combined with 27

the lack of malicious intent to denigrate, they are insulated from o ensiveness (Hom 2008).

But any

arguments for the aptness of tokening a slur in certain contexts, for example, for pedagogical purposes or in reports of others’ speech, or within a ctional or artistic context, obviously rely on the reactive sting these expressions are prone to e ect—one that a neutral counterpart wouldn’t; otherwise, why bother using one term over the other? Presumably, tokening a slur in such contexts is apt or indispensable only because it achieves something its neutral counterpart could not: it carries a particular e ect—a distinctive sting. This is not to say that some tokenings cannot be blameless or excusable, but the question of blameworthiness is downstream from the fact that the sting transpired. In short, we have to distinguish the blameworthiness of tokening a slur from the distinctive sting that confrontations with it might evoke, and which prompt questions of blameworthiness in the rst place. We might excuse an ignorant child or non-native speaker even if we nd their tokenings painful. And that the sting is present even in excusable tokenings explains why such tokenings still carry the risk of o ending (as documented by numerous examples of o ense even 28

in innocuous contexts).

2.6. Interim Summary In summary of what has been established thus far, our brief investigations of di erent content-based p. 159

accounts of slurs inspires pessimism toward content-based explanations.

The o ensive e ect of slurs

persists even in environments in which content—of any familiar kind—is rendered inert. This e ect is mysteriously ine able, and resists being re-expressed or paraphrased in other words. In light of these observations, we look for an explanation elsewhere. The conclusion to extract from our sojourn into content-strategies is that no special contents must attach to a slur in order to accommodate its potential to o end. We thus turn to non-content explanations in sections 3 to 5. Our own account (spelled out in section 5) will draw on the idea that typical tokenings of slurs give rise to negative associations, where these are, in certain respects, akin to the open-ended set of cognitive and emotive contents that constitute Camp’s perspectives. However, we think such associations must not be con ated with (an aspect of) the meaning of a slur nor attach to slurs themselves in the rst place (conventionally or otherwise). Rather, they are evoked through a display of certain articulations of slurs—orthographic or phonetic forms via which they are articulated.

Before we proceed to esh out our preferred proposal, we rst, in sections 3 and 4, turn to other noncontent strategies that fare better than content-based accounts in empirical coverage, but ultimately fall short.

3. Prohibitionism The rst non-content proposal of slurs we will consider is Prohibitionism: this is the view that slurs are prohibited words and the o ensive e ect of tokening one results from violating its prohibition (Anderson and Lepore 2013a, 2013b). So, on this view, it’s in virtue of a slur being banned that it acquires the potential to o end when tokened.

3.1. Prohibitionism Elaborated Unlike content theories that tie the o ensive potential of a slur to a layer of content, Prohibitionism ties it to a taboo status of the word itself. This, in turn, is why it can explain the robust projective behavior of the o ensive e ect of a slur: regardless of how deeply embedded its occurrence, it is, nevertheless, an occurrence of a slur; since the taboo applies to all occurrences of slur words, its ban is violated, regardless of 29

the depth of the embedding.

Indeed, since prohibited words are normally prohibited everywhere they

occur, Prohibitionism explains why the o ensive e ect tends to remain even in meaning attributions and under quotation, as discussed in section 2.5. It also explains why the o ense transpires regardless of the intentions and attitudes of the speaker: an inadvertent violation of a taboo, even in the absence of negative attitudes or malicious intentions, is still a violation of a taboo, and as such, it risks o ending. For the same p. 160

reason, Prohibitionism explains why one can express something linguistically coherent

(even if

insensitive and potentially o ensive) with utterances like (6a)–(6c): the speaker can violate a taboo, regardless of background attitudes, intentions, or perspectives. In short, if violations of taboos are what triggers the o ensive e ect of slurs, then Prohibitionism explains why any occurrence of a slur, whether mentioned or used, with or without malicious intent or attitude, can induce a palpable, o ensive sting. The Prohibitionist tells us a word can become taboo for all sorts of reasons, including a complex nexus of sociological, historical, and psychological reasons (Anderson and Lepore 2013a). It can become prohibited 30

by a decision from the relevant authority:

for instance, a group might denounce a certain term, and regard

its tokenings as violations. There can be an explicit rule, norm, or law prohibiting a particular term (e.g., an explicit editorial rule prohibiting occurrences of a term in mass media). A word might be banned because of what it means: it might be coined to signify something abusive or o ensive, and so, consequently become prohibited. Or it might become prohibited because of a history of usage and negative associations that have arisen through this history, and are evoked in its tokenings. Anderson and Lepore (2013a, 40) cite examples illustrating how the history of use of a term along with rights to self-determination might lead to a term becoming banned, even when it was once used self-referentially by the target group (they cite Baugh 1991), or how facts about who coined a word might become relevant to whether a term is prohibited (they cite the exchange in Du Bois 1928/1996). But, importantly, regardless of the possibly complex factors that lead to a prohibition, according to Prohibitionism, the word itself is o ensive because it is taboo. As Lepore and Anderson (2013a, 39), paraphrasing Feinberg (1985), summarize their own position, “[O] ensive words are generated by wordtaboos.” It is the prohibition that imbues the word with its o ensive e ect, because violations of a taboo o end. While many complex factors can lead to a word’s becoming taboo, it is only because it is taboo that its violations become o ensive. This brings us to criticisms of Prohibitionism.

3.2. Criticism Prohibitionism, whatever its advantages, has received its fair share of criticism. For one, we might wonder why banishing a random word renders it o ensive? In the nineteenth century, the Russian Empire banished Ukrainian, Polish, Lithuanian, and Belarusian in areas that it had annexed. These prohibitions did not engender feelings of o ense in Russians, or anyone else, when violated. Secondly, and obviously relatedly, Prohibitionism by itself, critics complain, fails to capture that slurs are banned because they’re o ensive, 31

and not that they are o ensive because they are banned.

There is a third aspect of the account that makes it inadequate, and which we will argue against in Part 4. But before we do so, it will be useful to introduce another non-content strategy: the Pejorative Tone Account.

p. 161

4. Pejorative Tone Frege thought that ‘dog’ and ‘cur’ agree in meaning, but that each “puts us … in mind” of di erent associations. One who uses ‘cur’ “speaks pejoratively, but this is not part of the thought expressed” (Frege 32

1897/1979, 140).

If Frege is correct, it follows that there are interpretive di erences that are not

di erences in meaning (i.e., within Frege’s nomenclature, neither in reference nor sense). Lepore and Stone (2018) invoke a notion of tone as a catchall label for all such interpretive e ects—by which they mean “e ects that are heterogeneous in origin, open-ended, and often non-propositional” (Lepore and Stone, 2018, 133). As they write, so construed, “tone is a messy, heterogeneous, psychological, historical, and social construct, but not a linguistic one” (Lepore and Stone 2018, 139). Crucially, Lepore and Stone argue that the tone of an expression does not correspond to a layer of its meaning, and it is not conventionally encoded. To see their argument for why tone cannot be characterized as meaning, consider their discussion of di erences between the words ‘bloom,’ ‘ ower,’ and ‘blossom,’ as 33

de ned by the Oxford English Dictionary (OED).

These three words pick out the same things. As Lepore and

Stone argue, di erences among them are not semantic, but rather di erences in associations: while all three denote the same thing, according to the OED, they tend to evoke di erent ower-related associations; 34

they invite you to think of—or rather, “picture”— owers in di erent ways.

So, ‘bloom’ (according to the

OED) is more delicate than ‘blossom,’ evoking the culminating beauty of a plant, whereas the former evokes the promise of fruit; yet these associations do not constitute part of the meaning of either term. Someone who only knew the word ‘bloom’ and used it indiscriminately to refer to owers, while stylistically peculiar, would not, Lepore and Stone a

rm, exhibit linguistic ignorance. (Note the contrasts with expressive terms

like ‘ouch,’ ‘oops,’ or ‘damn’: someone who consistently uses these terms in inappropriate circumstances, or combines them with a denial of the expressed sentiment, would be doing something linguistically incoherent, and not merely insincere.) Di erences attributed to tone are thus not a matter of predicative, presuppositional, conventional implicated, or expressive (à la Camp or Potts) content. But then how do these di erences arise?

4.1. Pejorative Tone and Negative Associations Obviously, we can have di erent perspectives on the same topic. We might think of a ower in a way that anticipates the plant’s fruit, or simply as a plant, or through re ecting on the beauty of the culmination of the owering process. But as Camp (2006, 2007, 2009) argues, these di erences in perspective don’t p. 162

amount a airs:

to di erences in representational content; (18a) and (18b) describe exactly the same state of

(18) a.

That is a rose blossom.

b.

That is a rose bloom.

As Lepore and Stone (2018) point out, all owers are parts of plants that bear fruit. Nevertheless, di erent expressions in (18a)–(18b) give rise to di erent associations: without altering the content expressed, they draw attention to, and thereby, invite us to think about, owers in di erent ways. Perspectival di erences, Lepore and Stone maintain, are di erences in tone. They further propose that, similarly, each slur carries a pejorative tone: while slurs do not di er from neutral counterparts in meaning, they di er in tone. A slur, unlike its neutral counterpart, evokes a set of negative associations and invites a negative perspective toward a targeted group. Speci cally, it invites the audience to organize their thoughts about a targeted group through negative stereotypes, and it places disproportionate attention on problematic conceptualizations of its members. Lepore and Stone partially concur with Camp (2013), when she writes that “slurs conventionally signal a speaker’s allegiance to a derogating perspective on the group identi ed by the slur’s extension determining core” (331). However, unlike Camp, who argues that slurs signal perspectives as a matter of linguistic convention—thus turning perspectives into a kind of semantically encoded content—Lepore and Stone (2018) argue that interpretive e ects that evoke perspectives must be understood in a causal framework. Calling a ower a ‘bloom’ doesn’t conventionally communicate that it’s delicate and beautiful—“this bloom isn’t delicate or beautiful” isn’t linguistically incoherent—but it tends to prompt the audience to re ect on it this way. Words can prompt all sorts of associations. Some evoke their most notable users; their present uses thus summon the attitudes or cultural context of those prominent users. Lepore and Stone invite us to consider a hipster who adopts the vivid but outdated slang of the 1930s and 1940s, calling error-prone workers “numbskulls,” addressing strangers in casual encounters as “mister,” or using “clams” as slang for money. These words, because of their ties to an extinct culture, invite an ironic engagement. As Lepore and Stone (2018, 147) write, “[T]o catch the hipster’s tone is to treat the hard-boiled Los Angeles of lm noir or the slapstick New York City of Vaudeville as a model or foil for current circumstances.” The same goes for slurs: they, too, have histories, albeit less innocent ones. They evoke uses embedded within the history, culture, practices, and ideology of oppression and discrimination. But, again, they do so not as a matter of semantics, but rather via causal links established through the history of usage. Just as in the case of the hipster, Lepore and Stone propose, speakers who use these words invite us to see our current circumstances through the lens of prominent speakers who’ve used the word and their cultural and historical context. This is why the uses have the o ensive potential. p. 163

Note further that, since the account of associations is causal, and not semantic, the Pejorative Tone Account —like Prohibitionism—can successfully explain why the o ensive e ect of slurs persists in linguistic environments that otherwise render content inert. The associations causally connected with the term through a history of usage are triggered whenever the term is tokened—even if it occurs in meaning attributions or in quotative environments. Further, since these associations are automatically evoked by the tokening of a slur, we can explain why the o ensive e ect persists regardless of the presence or absence of negative attitudes or malicious intentions. In sum, by agreeing with Prohibitionism not to seek a contentbased account of the o ensive e ect of slurs, the Pejorative Tone Account preserves its good predictions; but in locating the e ect in pejorative tone—the pernicious associations the term triggers—it can explain why a slur is o ensive and why it became taboo in the rst place.

Despite these merits, shortcomings remain for the Pejorative Tone Account and for Prohibitionism as well. Namely, both assume—naturally, but, we argue, wrongly—that the trigger of o ense (and the carrier of negative associations) is the slurring word. As we explain next (and as we have argued in more detail elsewhere (Stojnić and Lepore 2022, forthcoming)), this is a mistake: it is not words that trigger o ensive e ects, but rather, it’s their articulations.

5. The Articulation Account Prohibitionists maintain that it is slur words that are subject to taboo. The Pejorative Tone Account maintains that negative associations attach to slur words. Similarly, the content theories we examined locate the o ensive potential in a slurring word’s meaning or use. Indeed, all of the accounts discussed so far agree that it is, ultimately, the occurrences of slur words that trigger o ensive e ects. To cite an old adage, it is words that have consequences. However, to the contrary, we have been arguing that the o ensive e ect of a slur is triggered not by the word or its meaning but rather by some (though not all) of its articulations (Stojnić and Lepore 2022, forthcoming). To make our case, we rst need to say a few words about this important but remarkably ignored distinction between words and their articulations.

5.1. Words and Articulations The tokenings of any given word can vary in articulation: a word can be spoken or spelled, mispronounced or misspelled; it allows for variation in spelling and pronunciation across time and place. That is to say, words’ various tokens can take on di erent shapes (Lepore 2009; Hawthorne and Lepore 2011). It is thus critical to distinguish words (and their tokenings) from their articulations: words cannot be (obviously) 35

identi ed with shapes, and so, must be distinguished from their various articulations. p. 164

(2022)

 Stojnić and Lepore

argue this distinction is crucial for the analysis of the o ensive e ect of slurs. They o er a

di erent non-content proposal—the Articulation Account—according to which the primary triggers of the o ensive e ects of tokens of slurs are neither the slurs themselves nor their meanings but, rather (some of) their articulations. The articulations trigger o ense because they are the carriers of pejorative tone—those negative associations that arise through a history of usage in a way described by the Pejorative Tone Account. And because articulations evoke negative associations, contra Prohibitionism, it is (some of) their articulations, and not the words themselves, that are subject to taboos. If the Articulation Account is correct, then all of the accounts surveyed so far, including the content ones, Prohibitionism, and the Pejorative Tone Account, have been looking for the cause of o ensive e ect in the wrong place. And to do so con ates words with articulations, which is easier to do than one might think. After all, it’s impossible to token a word without articulating it, and so they are inextricably intertwined, and therefore, one might mistake a property of an articulation for one of a word. Articulations are, for example, graphemic or acoustic sorts of objects; and words are not. But, since we agree that words and articulations are intertwined, why think articulations rather than words trigger o ensive e ects? The short answer is that the e ect tends to persist whenever the standard articulation of a slur is present, even if no slur—indeed, even if no word at all—is tokened; and conversely, the slur loses its potent punch when its articulation is su

ciently far o

36

from a standard articulation.

To illustrate the rst point, consider the potential o ensive e ect triggered by tokening a homophone of a slur. Stojnić and Lepore (2022) illustrate with a recent real-life case where an o ense was triggered by the tokening of a Chinese demonstrative (‘那个’), which happens to acoustically resemble the N-word, in a university classroom setting, in the context of a discussion of ller words in di erent languages. That the o ense was triggered cannot be explained by semantic or etymological confusion, since the speaker announced in advance exactly which word was to be tokened and what the word means in Mandarin. So, no

slur was tokened in this case: the word tokened only accidentally matches in articulation a slur. It does not matter for our purposes whether the speaker has done anything wrong—whether they are blameworthy or to what extent—or whether any o ense taken was warranted. That we are asking whether the speaker is blameworthy or whether the o ense taken is warranted is downstream from the fact that the o ensive e ect was triggered; and since only the articulation was present, and not the slur, it must be the former that was the trigger. In a similar fashion, an o ense can be triggered even when no term—and so, no slur—is tokened. An accidental sneeze that happens to match the articulation of a slur can likewise trigger an o ensive e ect: one would be compelled to apologize and explain the accident. But conversely, if one’s articulation of a slur is so far o

it is not recognizable that the slur was uttered, the o ensive e ect will not be triggered. While

we might be upset or outraged if we know—or learn—what the speaker was tokening, we won’t experience 37

the same sort of palpable o ensive e ect we would have had the articulation been more successful.

Further, as Stojnić and Lepore (2022) argue, distinct (canonical) articulations of a word can di er in p. 165

whether they trigger negative associations (and even to what degree).

For instance, consider the

phenomenon of graphic pejoratives in logographic languages like Mandarin, where a word can be o ensive when written, but not when pronounced, and, further, only when written with a particular choice of a character but not with a di erent (phonetically indistinguishable) one. Since presumably the word means what it means however it’s written, this phenomenon is mysterious on any content account; further, since presumably it is the same word, whether it’s written or spoken, the phenomenon also remains inexplicable by non-content accounts that tie the o ensive e ect to the word itself—as Prohibitionism and the Pejorative Tone Account do. But on the Articulation Account, the phenomenon is not only explicable but expected, as some but not other articulations of a word can vary in the associations they trigger. Finally, the Articulation Account can easily explain why the o ensive e ect tends to be inherited by words whose articulations resemble the standard articulations of a slur. We have already seen this in the case of a Mandarin demonstrative. The same point is illustrated by a well-documented controversy over the incidents of tokening an English adverb that orthographically and phonetically resembles, but is etymologically and semantically unrelated to, the N-word (Kennedy 2002, 94–95). Its inherited o ensiveness has persisted despite the recognized etymological and semantic independence. That the o ensive e ect can be inherited by unrelated words is mysterious on any account that ties the o ensive e ect of a slur to the slur itself or its meaning; but if articulation is what triggers the o ensive e ect, such inheritance is just what we would expect. While the Articulation Account maintains that articulations trigger o ensive e ects—the distinctive o ensive sting that tokening a slur typically provokes—we are not claiming that the articulation alone causes the sting. More precisely, we are not claiming that slurs are o ensive because of how they sound or look. Rather, an articulation triggers an open-ended array of negative associations, which become associated with it through a history of usage. When an articulation is paired with negative associations, much akin to o ensive symbols (e.g., a swastika or a burning cross) or gestures (e.g., the racist appropriation of the OK-gesture, or the Nazi salute), or imagery (e.g., the lawn jockey or the mammy caricature), it gives rise to an o ensive e ect. In this way, the proposal becomes: certain articulations are 38

like o ensive imagery, symbols, and gestures in that they evoke negative associations.

The key insight

here is that a pejorative o ense is triggered by certain vehicles for presenting slur terms to which negative 39

associations attach, and not occurrences of those terms themselves.

Associations, of course, can come and go. They are created through and, thus, re ect the history of usage; but this process is dynamic: negative associations can fade over time and be replaced by di erent neutral or even positive ones (and the other way around). When negative associations fade, the pronunciation of the word in question loses its negative e ects. When they arise, a harmless word can take on a sting it never had

before. The mechanisms by which associations can come and go are beyond the scope of this paper and perhaps philosophy itself. But that this is a real phenomenon is beyond doubt. In short, like Prohibitionism and the Pejorative Tone Account, the Articulation Account o ers a nonp. 166

content explanation of the o ensive e ect of a slur. Like the Pejorative Tone

Account, and contra

Prohibitionism, it maintains that slurs are o ensive because their tokenings typically involve triggering negative associations; but unlike both of these other non-content accounts, the Articulation Account isolates (certain) articulations, rather than words, as triggers of negative associations. As explained, it is this feature that endows the account with an explanatory edge over these alternative accounts.

6. Conclusion We have surveyed a number of accounts of the o ensive e ect of slurring terms. The dominant approaches trace this o ensive e ect to some level of meaning of slurring words: either their truth-conditional, or presupposed, or conventionally implicated content or some further, sui generis type of expressive content. But we have seen there are serious problems with content accounts, indeed, with any account that traces the o ensive e ect to either some semantically encoded or pragmatically conveyed layer of meaning. In response, non-content approaches—Prohibitionism and the Pejorative Tone Account—have proposed that it is not pejorative meanings that trigger o ensive e ect, but rather that slurring words themselves trigger an o ensive e ect, either because they violate a taboo, or because they carry pejorative tone. However, we have argued, even these accounts have misidenti ed the trigger of the o ensive e ect: if the Articulation Account is on the right track, it is slurring articulations, not slurring meanings or words, that trigger the pejorative e ect.

Acknowledgments For helpful discussion, we would like to thank the participants in our joint Princeton-Rutgers graduate seminar on pejorative language in the fall of 2021. We are also grateful to Elizabeth Camp, Michael Glanzberg, Harvey Lederman, Matthew Stone, and Rob Stainton for helpful discussions of the related material.

Notes *

This chapter can be seen as a sort of précis to our manuscript “Inflammatory Language: The Linguistics and Philosophy for Pejoratives” (Stojnić and Lepore forthcoming).

1.

See, in particular, Potts (2007); Schlenker (2007); Williamson (2009); McCready (2010); Camp (2013, 2018); Jeshion (2013a, 2013b, 2016, 2018); Cepollaro (2015); Cepollaro and Stojanovic (2016); Kirk-Giannini (2019), inter alia. This pejorative content is o en characterized as an expression or a declaration of a negative attitude toward the target group. But even those who wouldnʼt identify the content with an expression of, or commitment to, a negative attitude, still typically maintain that slurs have some kind of o ensive content, and that their typical uses signal a speakerʼs commitment to, or approval of, such content. So, for instance, theorists thinking that slurs encode descriptive content that characterizes entrenched negative stereotypical generalizations and corresponding negative valuations based on such generalizations (e.g., Hom 2008) would likewise maintain that (i) attributions of such content reflects negative attitudes towards those targeted, and (ii) that uses of slurs typically signal speakerʼs approval or endorsement of such attitudes. More on this below.

2.

We defend this account in much more detail in Stojnić and Lepore (2022, forthcoming).

p. 167

p. 168

3.

Of course, one can insult with a neutral counterpart, say, if one utters it in the course of saying something insulting. Itʼs not necessary to slur someone to o end them. (Indeed, given customary usage, itʼs not even necessary to use a slur to slur someone.) The point is just that neutral counterparts arenʼt primarily exploited to cause o ense, nor do they in and of themselves do so; with slurs, by contrast, one would be hard-pressed to escape committing the o ense even when not (otherwise) trying to say something o ensive (but for the tokening of the slur). Slurs are prone to cause o ense in a way that neutral counterparts arenʼt. More on this below. It also bears emphasis that the status of a term as “neutral” as opposed to a slur is itself slippery, and subject to historical change. (On this, see in particular, Nunberg 2018; Stojnić and Lepore forthcoming) This is prima facie problematic for those who seek a semantic explanation of the o ensive potential, as this would require positing a change in meaning to accompany any such change. (More below.)

4.

While most authors agree that slur terms and neutral counterparts are (at least) coextensive, the agreement is not universal. We discuss an alternative proposal which denies this later in this section.

5.

Note that this means that utterances predicating a slur of an individual while denying the additional pejorative content should be self-contradictory, like “John is a bachelor but not unmarried” is. That is prima facie an uncomfortable result, as one can in principle use a slur in conjunction with a denial of any specific negative content, without thereby expressing something that sounds self-contradictory. This result also makes it particularly di icult to explain how slurs can be reclaimed, i.e., used by targeted group members in neutral or positive ways, as a signal of camaraderie or a term of pride, and without any derogation.

6.

See, e.g., Dummett (1981); Saka (2007); Hom (2008); Richard (2008); Camp (2013); Neufield (2019), inter alia.

7.

While we acknowledge that in order to appreciate the o ensive e ect of a slur, one has to be confronted with it (or rather, as we argue below and elsewhere (see Stojnić and Lepore 2022, forthcoming), with certain of its articulations), since we maintain that such an e ect is present—and so, can cause o ense—even when slurs are quoted and not used, we refrain from tokening actual slurs in this chapter (but we will, with some trepidation, use canonical descriptions of certain slur terms), and instead, opt to illustrate with a fictional one. We invite readers to test their intuitions against real-life examples or to consult those from the literature. The term ʻmudbloodʼ is a fictional slur from the Harry Potter novels. The word targets wizards who have non-wizardly ancestry, and is understood to be extremely o ensive, akin to nonfictional racial slur terms; its neutral counterpart is explained to be ʻmuggleborn,ʼ standing for ʻborn to non-magic parents.ʼ The term ʻmuggle,ʼ from which ʻmugglebornʼ is derived, is itself an o icial term for non-magic people, and while it sometimes appears to have negative overtones, no neutral counterpart exists. We will use ʻmugglebornʼ as a neutral counterpart of ʻmudbloodʼ. We also disregard the orthogonal complication that ʻmudbloodʼ and ʻmugglebornʼ are fictional terms and will talk throughout as if they had an actual non-empty extension. Since we use it as a stand-in for actual slurs that target actual groups of people, this idealization is innocuous.

8.

Similarly, Hom and May (2013), characterize the content as “ought to be a target of negative moral evaluation, because of being a member of G”, where ʻGʼ denotes the relevant target group. More recently, Neufield (2019) proposed a truthconditional account according to which the meaning of a slur is captured by the schema “S is true of P i P bears the ʻessenceʼ of G—whatever that essence is—which is causally responsible for stereotypical negative features associated with G and predicated of P (where S is a slur, P an individual, and G a social group).”

9.

Again, we invite you to consider a non-fictional counterpart. See Camp (2013) for examples.

10.

The challenge is raised by a number of authors in the literature. See, e.g., Anderson and Lepore (2013a, 2013b); Jeshion (2013a, 2013b); Camp (2013); Whiting (2013), inter alia. For responses o ered on behalf of predicativism, see Hom and May (2013); Neufield (2019). For a critical discussion of these responses, see Stojnić and Lepore (forthcoming).

11.

Again, one can test oneʼs intuition against a non-fictional example, replacing ʻmuggle-bornʼ with a neutral word: ʻgayʼ, ʻJewishʼ, ʻwomanʼ, etc. and ʻmudbloodʼ with a corresponding slur. For instance, saying “Mary is not Jewish” denies that being Jewish is true of Mary; replacing ʻJewishʼ with a corresponding slur under negation still denies that Mary is Jewish but does not neutralize the pejorative e ect: it projects outside negation. Similarly, saying “That is false” in response to “Mary is Jewish” denes that Maryʼs Jewish; saying the same thing in response to a predication with a corresponding slur denies Maryʼs Jewish, but does not deny the o ense.

12.

Even if the presupposition is not in fact a part of the common ground, it can o en be accommodated—the audience can go along and treat the presupposition as mutually accepted, even if it in fact wasnʼt prior to the utterance. For more on

presupposition accommodation, see, e.g., Karttunen (1974); Stalnaker (1974); Lewis (1979); Simons (2006); von Fintel (2008). 13.

In addition, there are various theoretical options depending on whether we think of the presupposition as semantically encoded in the slur term or, rather, as a pragmatic phenomenon. The presuppositional accounts of slurs on o er in the literature tend to posit lexicalized pejorative presuppositions (Schlenker 2007; Cepollaro 2015; Cepollaro and Stojanovic 2016). But since both pragmatic and semantic accounts would invoke pejorative content, our criticisms of positing such content would stick whether we think of presuppositions as semantic or pragmatic. For more on presuppositions, see e.g., Karttunen (1974); Stalnaker (1973, 1974); Heim (1983, 1990); Beaver (2001); Tonhauser et al. (2013), inter alia.

14.

The motivation for this “objective” approach has to do with the observation that one risks complicity if one does not protest othersʼ uses of slurs. This is somewhat di icult to explain if all that the slur presupposes is that the speaker believes some o ensive content. If, by contrast, the presupposition is “objective” (i.e., not describing the speakerʼs mental states), then by failing to challenge it, one tacitly comes to accept the presupposed content (see Cepollaro and Stojanovic 2016; Cepollaro 2015). However, the objective approach still faces a related challenge: simply challenging the presupposed content (that the targeted group is despicable) seems inadequate as a way of rejecting the o ensiveness of the slur and avoiding complicity. “Hey wait a minute! Muggleborn wizards are not despicable!” doesnʼt appear to be an adequate way to challenge the o ensiveness of “Hermione is a mudblood.” To avoid complicity, one would rather protest the very practice of using a slur (e.g., with, “That language is o ensive!”).

15.

Similar observations are made in Camp (2013) and Kirk-Giannini (2019).

16.

This observation goes back to Karttunen (1973). The literature on presupposition projection is vast, but this brief description su ices for our purposes. See, e.g., Karttunen (1974); Stalnaker (1973, 1974); Heim (1983, 1990); Beaver (2001); Tonhauser et al. (2013), among many others, for further discussion of the relevant theoretical issues.

17.

As always, we invite readers to test their intuitions with non-fictional examples.

18.

See, in particular, McCready (2010); Potts (2007); Williamson (2009); Whiting (2013). However, as shall become apparent, these authors di er in how they understand the nature and content of the implicature they posit.

19.

In this vein, Williamson writes: “An Englishman in Italy who thinks that ʻmaʼ is synonymous with ʻandʼ and ʻeʼ with ʻbutʼ is mistaken, for ʻeʼ is synonymous with ʻandʼ, not with ʻbutʼ, and ʻmaʼ is synonymous with ʻbutʼ, not with ʻandʼ. Fully to understand a word, one must have some awareness, however explicit, of the conventional implicatures that it generates” (2009, 24). That is, it is a part of the meaning—albeit not the truth-conditions—of one but not the other, that it implies a contrast between its conjuncts.

20.

It doesnʼt matter whether the supposedly implicated o ensive content is that muggleborn wizards are inferior on account of being muggleborn, or that they are despicable, or what have you: denying any such content in conjunction with the use of the slur wonʼt defuse its potential o ensive e ect.

21.

This prediction has been claimed as an advantage of the Conventional Implicature Account over the Presuppositional Account, since by contrast, utterances with false presupposition lack a truth value altogether (Williamson 2009). However, the argument is a bit quick, since the Presuppositional Account could maintain that an utterance of a slur triggers presupposition accommodation, and that the audience must challenge the presupposition directly in order to block accommodation of the (false) o ensive presupposition, which if unchallenged, will become accepted as true for the purposes of the conversation.

22.

Itʼs not clear that ʻbutʼ has this feature: “Mary said that John is tall but handsome, but thereʼs no contrast between handsomeness and height” might be read as discharging the speakerʼs commitment to the implicated content. At the same time, some have argued that ʻbutʼ isnʼt in fact a conventional implicature precisely on the grounds that the contrastive content of ʻbutʼ inside indirect speech reports is attributed to the attitude holder, not the speaker (Bach 1999). Potts (2005) likewise argued that ʻbutʼ does not contribute any CIs. As many have pointed out, itʼs not entirely clear that there is a uniform kind corresponding to Gricean CIs.

23.

Again, it doesnʼt matter how we specify the pejorative content a CI supposedly conveys: slurring predications can be coherently (if o ensively) combined with the denial of just about any negative content regarding a target group. Indeed, this holds even if a CI is expressive rather than propositional (Potts 2007; McCready 2010). More below.

p. 169

24.

Camp argues that informal versus formal ways of address (ʻtuʼ/ʻvousʼ) should be seen as perspectival expressions— expressions expressing perspectives—along with slurs, slang terms, and thick terms. Others have argued that politeness (formal/informal address) should be theorized in terms of sociolinguistic register, a conventionalized marking of lexical expressions along a certain set of features (e.g., [+/− politeness] or [+/− slang]) signaling the “appropriateness conditions” for the use of an expression. Diaz-Legaspe, Liu, and Stainton (2020) have proposed extending this idea to explain the o ensiveness of slurs in terms of the sociolinguistic register, by arguing that we should posit a feature [+/− derogatory], so that slur terms are o ensive because they are marked as derogatory. Insofar as [+/− polite] appears “inert” in (17e), and insofar as [+/− derogatory] should behave in the same way, then this view still faces a problem of explaining why slurs retain o ensive potential in meaning attributions. Apart from this problem, we note that it is not entirely clear why simply being told that a term is marked as [+derogatory] would su ice to explain its o ensive potential: suppose we are told that “ ʻboopʼ is a term for a group G and itʼs [+derogatory].” This still wonʼt in itself make us feel the palpable sting we typically experience when encountering slurs when we hear ʻboop.ʼ (The same holds if we simply stipulate: “let ʻboopʼ be a [+derogatory] term for Gs.”) Indeed, the o en-felt lack of potency of slurs when we first encounter them in a foreign language, even when we are told they are [+derogatory], suggests that merely knowing that slurs are [+derogatory] doesnʼt in and of itself explain their potency.

25.

See Lepore and Ludwig (2005, 40–41) for further discussion.

26.

This critical point has been extended to quotative uses of slurs. Many (most?) believe replacing ʻAfrican Americanʼ with the N-word in ʻ“African Americanʼ is a term for Black peopleʼ” can easily potentially o end. See Anderson and Lepore (2013a, 2013b), but also Jorgensen Bolinger (2017); Kirk-Giannini (2019); Camp (2013); and Stojnić and Lepore (2022), inter alia. But not all authors concur with this assessment (Potts 2005; 161–162; Hornsby 2001, 129–130; Williamson 2009; Hom 2008, 16). These authors maintain that one can insulate oneself from o ending by utilizing quotes. Still there are many documented examples of o ense taken (even without a risk of use-mention confusion) strongly suggesting that tokenings of quoted slurs carry a risk of o ending that tokenings of their neutral counterparts (quoted or otherwise) do not. (For a small sample, see e.g., Butler, C. (July 29, 2020), “Use of Racial Slur in Class Got Lukewarm Response from Kingʼs University College, Student Says,” CBC News, https://www.cbc.ca/news/canada/london/western-university-kingscollegeracism-1.5664828; Flaherty, C. (August 06, 2019), “The N-Word at the New School,” Inside Higher Education, https://www.insidehighered.com/news/2019/08/07/another-professor-under-fire-using-n-word-class-while-discussingjames-baldwin.) Similarly, there is the familiar practice of avoiding quotative tokenings of slurs given our preference for ʻthe N-wordʼ (and its ilk), which describes a slur without mentioning it. See Stojnić and Lepore (2022) for discussion.

27.

Even Hom (2008) acknowledges that occurrences of a slur retain the potential to denigrate, and can provoke “feelings of squeamishness … not only [when] embedded under negation, conditional antecedents, questions, intensional and fictional contexts, but also…under quotation, in contexts of appropriation, and even [in tokening] mere phonological variants.” However, he argues these intuitions should not be given much weight against content-based accounts, but rather should be dismissed or explained away: “These words are so highly charged, our intuitions have limited value from the outset, and it would be hardly surprising if, at least in some cases, our intuitions were even misleading” (27). But if we are interested in explaining the o ensive potential of slurs in real-life discourse, such data must not be brushed aside. The theoretical question is why such words are highly charged, and in trying to explain this, we canʼt set aside a set of data in which a potent charge is manifested. Homʼs dismissive reaction partly stems from considering only content-based— semantic and pragmatic—accounts of slurs. But as we will see below, content-based accounts donʼt exhaust the realm of possibilities: viable non-content-based accounts—in particular, our own—capture all the data uniformly without dismissing any as a result of noise or confusion.

28.

Compare note 27, above. See Stojnić and Lepore (forthcoming); Anderson and Lepore (2013a, 2013b) for discussion and further examples.

29.

This is an exaggeration: Prohibitionism allows that certain tokenings can be exempt from the ban, which is to explain the possibility of reclaimed uses of slurs (Anderson and Lepore, 2013a).

30.

Although, naturally, the relevant authority is constituted by members of the target group, this neednʼt be the case. Anderson and Lepore (2013a) point out that there could be slurs for groups that cannot themselves denounce the term: e.g., for infants, or people with severe mental disability. This is further complicated by the fact that there can be variation and disagreement within the targeted group as well. As Anderson and Lepore (2013a) point out, for a proposed prohibition to “stick” there should be su icient uptakes; as an illustration, they point to the Reverend Jesse Jacksonʼs proclamation at the 1988 Democratic National Convention that the word ʻblackʼ should not be used. His e ort failed, as

p. 170

p. 171

many members of the targeted community refused to regard the word as taboo, and so as a slur, due to its many positive associations.

p. 172

31.

See Camp (2013) for a version of this objection; see Anderson and Lepore (2013a) for an attempt to respond on behalf of prohibitionism.

32.

For more on interpretation of Frege on tone, see, e.g., Dummett (1981) and Picardi (2007).

33.

Their point could equally well be illustrated with Fregeʼs example of ʻdogʼ versus ʻcurʼ: both terms apply to all and only dogs. Yet they invite you to think of dogs in di erent ways: as Frege says, describing a dog as a ʻcurʼ evokes negative associations that do not arise with the same description featuring ʻdogʼ instead. But the associations are likewise perspectival, and open-ended: thereʼs no specific emotion or content one has to think of when one applies ʻcurʼ as opposed to ʻdog.ʼ

34.

Lepore and Stone do acknowledge that ʻblossomʼ and ʻbloomʼ have several senses, not all overlapping. But, they point out that in their basic sense, ʻflower,ʼ ʻblossom,ʼ and ʻbloomʼ all denote the same thing.

35.

Coincidentally, the so-called Shape Account of words, according to which two utterances are of the same word just in case they exhibit the same shape, has had quite a few defenders (Davidson 1979, inter alia). A common, though not the only, way to maintain this idea is to treat words as abstract types, the identity of which is reflected by a common shape, exhibited by its tokens, where utterances are of the same word just in case they exhibit the same shape. But, as has been forcefully argued by a number of theorists, such accounts cannot adequately explain how words can be spoken or spelled, mispronounced or misspelled, and vary in their canonical spellings and pronunciation over time and place (viz., Hawthorne and Lepore 2011; Kaplan 2011; and Stojnić 2022, inter alia).

36.

For a more detailed defense, see Stojnić and Lepore (2022, forthcoming). Here we just briefly summarize some main arguments.

37.

We should resist the temptation to conclude that a bad articulation of a word fails to be an articulation of that word: which word the speaker tokens is a matter of which item they select from their mental lexicon during the process of word production, not how well they articulate the word they are tokening. See Stojnić (2022) for more on this point.

38.

The proposal is not that slurs are (accompanied by) tacit o ensive gestures, nor that an aspect of their meaning is best characterized as the meaning of an accompanying (tacit) o ensive gesture (as is the case on the gestural account of slurs (Hornsby 2001)). Rather, certain articulations of slurs evoke o ensive associations in the same way an o ensive gesture evokes such associations.

39.

There can be variation in severity of the o ensive e ect: not all slurs (or all articulations of a particular slur) need have equally severe o ensive e ects, even when pertaining to the same subject matter, just as not all symbols and gestures —even when pertaining to the same subject matter—need be equally severe in o ensive e ect.

References Anderson, Luvell, and Lepore, Ernie. (2013a). “Slurring Words.” Noûs 47 (1): 25–48. Google Scholar WorldCat Anderson, Luvell, and Lepore, Ernie. (2013b). “What Did You Call Me? Slurs as Prohibited Words.” Analytic Philosophy 54 (3): 350– 363. Google Scholar WorldCat Bach, Kent. (1999). “The Myth of Conventional Implicature.” Linguistics and Philosophy 22 (4): 327–366. Google Scholar WorldCat Baugh, John. (1991). “The Politicization of Changing Terms of Self-Reference among American Slave Descendants.” American Speech 66 (2): 133–146. Google Scholar WorldCat Beaver, David I. (2001). Presupposition and Assertion in Dynamic Semantics. Stanford, CA: CSLI. Google Scholar Google Preview WorldCat COPAC Camp, Elisabeth. (2006). “Metaphor and That Certain ʻJe Ne Sais Quoiʼ.” Philosophical Studies 129 (1): 1–25. Google Scholar WorldCat Camp, Elisabeth. (2007). “Showing, Telling, and Seeing: Metaphor and ʻPoeticʼ Language.” Baltic International Yearbook of Cognition, Logic, and Communication. Vol. 3, 1–24. Google Scholar WorldCat Camp, Elisabeth. (2009). “Two Varieties of Literary Imagination: Metaphor, Fiction, and Thought Experiments.” Midwest Studies in Philosophy: Poetry and Philosophy 33 (1): 107–130. Google Scholar WorldCat Camp, Elisabeth. (2013). “Slurring Perspectives.” Analytic Philosophy 54 (3): 330–349. Google Scholar WorldCat Camp, Elisabeth. (2018). “A Dual Act Analysis of Slurs.” In Bad Words, edited by David Sosa, 29–59. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Cepollaro, Bianca. (2015). “In Defense of a Presuppositional Account of Slurs.” Language Sciences 52:36–45. Google Scholar WorldCat Cepollaro, Bianca, and Stojanovic, Isidora. (2016). “Hybrid Evaluatives.” Grazer Philosophische Studien 93 (3): 458–488. Google Scholar WorldCat Davidson, Donald. (1979). “Quotation.” In Inquiries into Truth and Interpretation, 79–92. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Diaz-Legaspe, Justina, Liu, Chang, and Staiton, Robert J. (2020) “Slurs and Register: A Case Study in Meaning Pluralism.” Mind and Language 35 (2): 156–182. Google Scholar WorldCat Du Bois, W. E. B. (1928/1996). “The Name ʻNegro.ʼ” In The Oxford W. E. B. Du Bois Reader, edited by Eric J. Sundquist, 70–72. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC

Dummett, Michael. (1981). Frege: Philosophy of Language. 2nd ed. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Feinberg, Joel. (1985). O ense to Others: The Moral Limits of the Criminal Law. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC von Fintel, Kai. (2008). “What Is Presupposition Accommodation, Again?” Philosophical Perspectives 22 (1): 137–170. Google Scholar WorldCat Frege, Gottlob. (1897/1979). “Logic.” In Posthumous Writings: Gottlob Frege, edited by Hans Hermes, Friedrich Kambartel, and Friedrich Kaulbach; translated by Peter Long and Roger White. Oxford: Blackwell. Google Scholar Google Preview WorldCat COPAC Grice, H. P. (1961). “The Causal Theory of Perception.” Proceedings of the Aristotelian Society 35:121–152. Google Scholar WorldCat Grice, H. P. (1989). Studies in the Way of Words. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC p. 173

Heim, Irene. (1983). “On the Projection Problem for Presuppositions.” In WCCFL 2: Second Annual West Coast Conference on Formal Linguistics, edited by Michael Barlow, Daniel Flickinger, and Michael Wescoat, 114–125. Stanford, CA: Stanford University. Google Scholar Google Preview WorldCat COPAC Heim, Irene. (1990). “Presupposition Projection.” In Presupposition, Lexical Meaning and Discourse Processes: Workshop Reader, edited by Rob van der Sandt. Nijmegen, Netherlands: University of Nijmegen. Available at https://semanticsarchive.net/Archive/GFiMGNjN/Presupp%20projection%2090.pdf. Google Scholar Google Preview WorldCat COPAC Hawthorne, John, and Lepore, Ernie. (2011). “On Words.” Journal of Philosophy 108 (9): 447–485. Google Scholar WorldCat Hom, Christopher. (2008). “The Semantics of Racial Epithets.” Journal of Philosophy 105 (8):416–440. Google Scholar WorldCat Hom, Christopher, and May, Robert. (2013). “Moral and Semantic Innocence.” Analytic Philosophy 54 (3): 293–313. Google Scholar WorldCat Hom, Christopher, and May, Robert. (2018). “Pejoratives as Fiction.” In Bad Words, edited by David Sosa, Oxford: Oxford University Press, 108–131. Google Scholar Google Preview WorldCat COPAC Hornsby, Jennifer. (2001). “Meaning and Uselessness: How to Think about Derogatory Words.” Midwest Studies in Philosophy 25 (1):128–141. Google Scholar WorldCat Jeshion, Robin. (2013a). “Expressivism and the O ensiveness of Slurs.” Philosophical Perspectives 27 (1): 231–259. Google Scholar WorldCat Jeshion, Robin. (2013b). “Slurs and Stereotypes.” Analytic Philosophy 54 (3): 314–329. Google Scholar WorldCat Jeshion, Robin. (2016). “Slur Creation, Bigotry Formation: The Power of Expressivism.” Phenomenology and Mind 11: 130–139. Google Scholar WorldCat Jeshion, Robin. (2018). “Slurs, Dehumanization and the Expression of Contempt.” In Bad Words, edited by David Sosa, 77–107.

Oxford: Oxford University Press. Google Scholar Google Preview

WorldCat

COPAC

Jorgensen Bolinger, Renée. (2017). “The Pragmatics of Slurs.” Noûs 51 (3): 439–462. Google Scholar WorldCat Kaplan, David. (1999). “The Meaning of ʻOuchʼ and ʻOopsʼ: Exploration in the Theory of Meaning as Use.” Unpublished manuscript, UCLA. Google Scholar Google Preview WorldCat COPAC Kaplan, David. (2011). “Words on Words.” Journal of Philosophy 108 (9): 504–529. Google Scholar WorldCat Karttunen, Lauri. (1973). “Presuppositions of Compound Sentences.” Linguistic Inquiry 4:168–193. Google Scholar WorldCat Karttunen, Lauri. (1974 ). “Presuppositions and Linguistic Context.” Theoretical Linguistics 1:181–194. Google Scholar WorldCat Kennedy, Randall. (2002). Nigger: The Strange Career of a Troublesome Word. New York: Pantheon. Google Scholar Google Preview WorldCat COPAC Kirk-Giannini, Cameron Domenico. (2019). “Slurs Are Directives.” Philosophersʼ Imprint 19 (48): 1–28. Google Scholar WorldCat Lepore, Ernie. (2009). “Heresy of Paraphrase: When the Medium Really Is the Message.” Midwest Studies in Philosophy 33 (1): 177– 197. Google Scholar WorldCat Lepore, Ernie, and Ludwig, Kirk. (2005). Donald Davidson: Meaning, Truth, Language, and Reality. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Lepore, Ernie, and Stone, Matthew. (2018). “Pejorative Tone.” In Bad Words, edited by David Sosa, 132–154. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Lewis, David. (1979). “Scorekeeping in a Language Game.” Journal of Philosophical Logic 8 (1): 330–359. WorldCat Google Scholar McCready, Elin. (2010). “Varieties of Conventional Implicature.” Semantics & Pragmatics 8:1–57. Google Scholar WorldCat Neufeld, Eleonore. (2019). “An Essentialist Theory of the Meaning of Slurs.” Philosophersʼ Imprint 19 (35): 1–29. Google Scholar WorldCat p. 174

Nunberg, Geo rey. (2018). “The Social Life of Slurs.” In The New Work on Speech Acts, edited by Daniel Fogal, Daniel W. Harris and Matt Moss, 237–295. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Picardi, Eva. (2007). “On Sense, Tone and Accompanying Thoughts.” In The Philosophy of Michael Dummett, edited by Randall E. Auxier and Lewis Edwin Hahn, 491–520. Chicago: Open Court. Google Scholar Google Preview WorldCat COPAC Potts, Christopher. (2005). The Logic of Conventional Implicatures. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC

Potts, Christopher. (2007). “The Expressive Dimension.” Theoretical Linguistics 33 (2): 255–268. Google Scholar WorldCat Richard, Mark. (2008). When Truth Gives Out. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Saka, Paul. (2007). How to Think about Meaning. New York: Springer. Google Scholar Google Preview WorldCat COPAC Simons, Mandy. (2006). “Presupposition Accommodation: Understanding the Stalnakerian Picture.” Philosophical Studies 112 (3): 251–278. Google Scholar WorldCat Stalnaker, Robert. (1973). “Presuppositions.” Journal of Philosophical Logic 2: 447–457. Google Scholar WorldCat Stalnaker, Robert. (1974). “Pragmatic Presuppositions.” In Semantics and Philosophy, edited by Milton K. Munitz and Peter K. Unger, 197–213. New York: New York University Press. Google Scholar Google Preview WorldCat COPAC Schlenker, Philippe. (2007). “Expressive Presuppositions.” Theoretical Linguistics 33: 237–245. Google Scholar WorldCat Stojnić, Una. (2022). “Just Words: Intention, Tolerance and Lexical Selection.” Philosophy and Phenomenological Research 105 (1): 3–17. Google Scholar WorldCat Stojnić, Una, and Lepore, Ernie. (2022) “Inescapable Articulations: Vessels of Lexical E ect.” Noûs 56 (3): 742–760. Google Scholar WorldCat Stojnić, Una, and Lepore, Ernie. (forthcoming). Inflammatory Language: The Linguistics and Philosophy for Pejoratives. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Tonhauser, Judith, Beaver David, Roberts Craie, and Simons Mandy. (2013). “Toward a Taxonomy of Projective Content.” Language 89 (1): 66–109. Google Scholar WorldCat Whiting, Daniel. (2013). “Itʼs Not What You Said, Itʼs the Way You Said It: Slurs and Conventional Implicatures.” Analytic Philosophy 54 (3): 364–377. Google Scholar WorldCat Williamson, Timothy. (2009). “Reference, Inference, and the Semantics of Pejoratives.” In Feshschri for David Kaplan, edited by Joseph Almog and Paolo Leonardi, 137–158. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

8 On Passive Aggression  Rebecca Roache https://doi.org/10.1093/oxfordhb/9780192844118.013.8 Published: 22 May 2024

Pages 175–191

Abstract This chapter de nes passive-aggressive behaviour and explains why we nd it objectionable. It argues that passive aggression essentially involves a hostile attack on a target, which is ambiguous in that, in addition to being correctly describable as such, from the point of view of the target and any onlookers, there is a somewhat plausible (yet false) interpretation according to which the behaviour is innocuous. Passive aggression is objectionable primarily because behaving passive aggressively impedes the agent’s social and professional functioning, and because it involves attacking the target in a way that makes it di

cult for the target to challenge the behaviour. Further, behaving passive aggressively is

bad for passive aggressors because it makes it more di

cult for them to address their feelings of

hostility, and it may exacerbate feelings of powerlessness.

Keywords: passive aggression, ambiguity, ethics, etiquette, interpretation Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction The term ‘passive aggression’ was coined by the US army psychiatrist Colonel William Menninger during the Second World War, in response to a need to update the system for diagnosing mental illness in the wake of the mental health issues experienced by military personnel. Menninger used the term to characterize soldiers who resisted complying with orders—not openly, but instead ‘by passive measures, such as 1

pouting, stubbornness, procrastination, ine

ciency, and passive obstructionism’. Early editions of the

American Psychiatric Association’s Diagnostic and Statistical Manual for Mental Disorders (DSM) included a diagnosis of ‘passive-aggressive personality disorder’, characterized in DSM-I (APA 1952) by ‘passive resistance to ful lling routine demands’, and in DSM-III-R (APA 1987) by a ‘pervasive pattern of passive resistance to demands for adequate social and occupational performance’. As a disorder, however, it was never satisfactorily de ned as a pattern of behaviour that a person manifests generally rather than in response to speci c stimuli, such as the demands of a professional superior. As a result, it was relegated to the appendix of DSM-IV (APA 1994), and it does not appear at all in the current volume, DSM-5 (APA 2013). Despite this, the concept of passive aggression remains useful and in uential. Clinicians and counsellors continue to address problems associated with passive aggression in their patients and clients, and articles o ering advice about dealing with passive aggression regularly appear in popular venues like Psychology Today. The term has even entered our everyday vocabulary and popular culture: it has an entry in the online Urban Dictionary; it is the title of a song by the British rock band Placebo; and a ‘passive-aggressive cats’ internet trope is the focus of YouTube videos with many millions of views. ‘Passive aggressive’ is, of course, a pejorative term. We no more welcome being thought of by others as passive aggressive than we welcome being thought of as lazy or boring. Despite this, exactly what passive aggression is and what is wrong with it are issues that have not received much attention from philosophers. p. 176

My aim here is to o er

such an analysis: to identify the de ning characteristics of passive aggressive

behaviour and assess what is wrong with it.

2. What Is Passive Aggression? Let’s start with some examples of behaviour that, in everyday settings and with the right contextual information, it would be reasonable to identify as passive aggressive. The Wedding Guest: It is Parvati’s wedding day. As she mingles with her guests at the reception, Bruna, a newly divorced family friend, coolly appraises her and remarks, ‘How brave of you to wear that out t, Parvati’. The Colleague: Miho and Cassius both recently applied for the same promotion. Cassius was successful and now leads the team of which Miho is a member. Since Cassius’s promotion, Miho has acted sullenly towards Cassius. Miho committed to writing a report on which the team is relying to meet an important project deadline. On the day of the deadline, Miho calls in sick, and the report is late. The Boyfriend: Erno and Avril, who are in a long-term relationship, are taking a ight together. Erno takes his seat on the plane, as Avril moves along the aisle looking for a space to store her luggage. When she returns to Erno, he has given away her seat to someone else. ‘I thought you’d 2

decided to sit somewhere else,’ he says.

There are features that all these cases have in common, and they tell us something about what passive aggression is. I will argue that passive aggression has two essential features. First, the agent’s behaviour

must constitute a hostile attack on the target. Second, the behaviour must be ambiguous in that it is objectionable under one (correct) interpretation, but innocuous under another (incorrect but plausible) interpretation. Let’s consider each of these features in more detail.

2.1. A Hostile Attack In each of the three cases, and in all instances of passive aggression, the behaviour expresses hostility of the passive aggressor towards the target. The behaviour expresses the agent’s hostility in the sense of being 3

motivated and explained by it. The hostility involves a negative attitude like resentment, jealousy, or a feeling of having been unfairly treated, and is expressed without the agent explicitly stating, or otherwise clearly indicating, that he holds the attitude in question. This, combined with the agent’s e orts to avoid having to discuss his hostility—something we will consider in more detail shortly—means that passive aggression can be an appealing way to behave for an agent who does not want to admit to his hostility, who does not feel entitled to it, or who does not want to su er the negative consequences of a more overt attack. p. 177

In some cases, the

passive aggressor may not even know exactly what sort of hostile attitude he holds, or

why. Sometimes our negative attitudes are deeply buried and can be unearthed and processed only following the sort of deep self-re ection that results from psychotherapy. The passive-aggressive agent expresses hostility towards the target by attacking the target in some way. In The Wedding Guest, Bruna attacks Parvati by insulting her; in The Colleague, Miho attacks Cassius by sabotaging the project; in The Boyfriend, Erno attacks Avril by inconveniencing her. E orts to insult, o end, punish, belittle, embarrass, or otherwise inconvenience or disadvantage the target are typical—but any form of attack is possible, provided that it is compatible with the sort of ambiguity that we will consider in the next sub-section. An agent’s behaving passive aggressively does not entail anything about whether the agent’s attack on the target is warranted: it may be morally or rationally warranted (in The Colleague, perhaps Cassius beat Miho to the promotion by underhand means, which we might think warrants retaliation by Miho) or unwarranted (perhaps Cassius has not wronged Miho, making Miho’s treatment of him unfair). We can draw a distinction between the target of the agent’s hostile attack and the agent’s intended audience. Often, the target is the audience; for example, in The Wedding Guest, Parvati is both the target of Bruna’s attack and her audience, i.e. the person whom she addresses in making the remark that constitutes her attack. Sometimes, the target and the audience are di erent people (or groups of people). Imagine a variation of The Wedding Guest in which, instead of addressing Parvati directly, Bruna remarks to a third person in the company of Parvati, ‘Isn’t Parvati brave to wear that out t?’ In this case, whilst Parvati is the target of Bruna’s attack, her audience is the person to whom she addresses her remark. In other cases, a passive aggressive attack may have no audience at all. In The Colleague, Cassius is the target of Miho’s attack, but the behaviour that constitutes his attack—i.e. taking a day o

work—is not the sort of behaviour

that is addressed to an audience. Whilst passive aggression always has a target, then, it may or may not have an audience; but if it does, that audience may or may not be identical with the target. An important characteristic of passive-aggressive attacks is that they aim to make it di

cult for the target

to respond by challenging their attacker. That the passive-aggressive agent aims at this e ect does not entail that the agent consciously intends it, but rather that this aim—conscious or otherwise—is part of the explanation for why the agent behaves passive aggressively. The agent achieves this aim via the second essential feature of passive aggression: ambiguity.

2.2. Ambiguity In each of the three cases described above, the agent’s behaviour is in some sense objectionable, but it is not clearly objectionable—as would be the case if, for example, the agent had sworn at the target aggressively. From the point of view of the target, the audience (if there is one), and any onlookers, there is at least one somewhat plausible, yet incorrect, possible interpretation of the behaviour according to which it is p. 178

innocuous.

Bruna might be complimenting Parvati on her unique and bold sense of style; Miho might

really be sick; Erno might sincerely think that Avril has decided to sit elsewhere. But in each case, there is another, correct, interpretation according to which the behaviour is objectionable. On these interpretations, Bruna intends to insult Parvati, Miho intends to sabotage the project, and Erno intends to punish Avril for leaving him alone. For the behaviour in question to count as passive aggressive, the objectionable interpretation must be correct. If the innocuous interpretation is correct, the behaviour is not passive aggressive, although it may appear to be so. It is not possible for both the innocuous interpretation and the objectionable interpretation to be correct: the agent’s behaving objectionably entails that her behaviour cannot correctly be interpreted as innocuous. This sort of ambiguity, in which the behaviour is innocuous according to one interpretation but objectionable according to another, is characteristic of passive aggression. Let’s dig a bit deeper into these claims. What does it mean to say that, on one interpretation, the behaviour is ‘innocuous’? There is, after all, a clear sense in which it is not innocuous, regardless of the interpretation. In The Colleague, an important deadline is missed. In all three cases, the target might end up feeling insulted, punished, undermined, or otherwise o ended (for brevity, I’ll use ‘o ended’ to include all the various sorts of negative responses that passive aggression can be expected to provoke in its targets). By claiming that the behaviour has an innocuous interpretation, I mean that there is an interpretation of the behaviour according to which the agent did not intend to o end the target or to bring about any other negative consequences, nor could the agent reasonably have been expected to foresee that any o ence or negative consequences would result from her behaviour; as a result, it is inappropriate to blame the agent for the target’s feeling o ended or for any other negative consequences of her behaviour. Instead, according to the innocuous interpretation, the target’s o ence and any other negative consequences can be attributed either to unfortunate events for which nobody is to blame, or—as is sometimes suggested by passive aggressors when challenged—to the target’s own oversensitivity. What does it mean to say that, on one interpretation, the behaviour is ‘objectionable’? This could mean that the behaviour is ethically objectionable, but passive aggression generally falls short of being unethical. More often, it is simply rude or inconsiderate—that is, contrary to the demands of etiquette, or contrary to the 4

sort of norms and expectations that govern relationships between friends or colleagues. The di erence between ethics and etiquette is explored by Judith Martin—who, under the name Miss Manners, pens a regular etiquette advice column that is published in hundreds of newspapers worldwide—and Gunther S. Stent in their 1990 paper, ‘I Think; Therefore I Thank: A Philosophy of Etiquette’. Martin and Stent view the norms of etiquette and those of morality as forming a spectrum. At one end of the spectrum are norms where the consequences of violation are grave; these are the rules of morality. At the other end are the norms of etiquette, whose violation tends not to result in grave consequences. Martin and Stent tell us that morality is concerned with ‘aspects of the human condition that involve matters of potentially grave p. 179

consequence for life, limb, and property’, whilst

etiquette focuses on ‘matters of potentially less grave

consequences, such as o enses against personal dignity, sacred and profane ritual, and the aesthetic sense’ (1990: 253). This view of etiquette as related to morality, concerned with everyday matters of less grave consequences, and having an important role in conveying respect for the dignity of others, is echoed by philosophers. Sarah Buss (1999) argues that in many cases to be bad mannered is to be immoral, and that an important function of manners is to express respect and to acknowledge the moral status of others (indeed, she observes that saying ‘please’ is equivalent to saying ‘you are worthy of respect’). Karen Stohr, who

discusses Martin throughout her 2012 book, On Manners, also holds the view that etiquette and morality form a spectrum, and concurs with both Martin and Buss that good manners are a way of conveying respect (Stohr 2012). An important feature of etiquette—and especially of the more particular norms that arise between friends— is that many of the norms are not clearly de ned and understood. Its norms vary with context and over time (consider that it was once a mark of good manners for a man to expect a woman to remain seated in a car until he had opened the door for her; now such an attitude is as likely to convey sexism, and therefore disrespect). It can be di

cult to work out which norms are required in a given context, whether certain

norms are now outdated, and which norm should take precedence when there is a con ict. This is why there remains a demand for etiquette columnists like Miss Manners. Moreover, often we are not even conscious of widely followed norms of etiquette and would nd it di

cult to articulate them with any precision (consider

the norms, within a culture or a particular social setting, around what counts as too much eye contact with, or standing too close to, a person with whom one is conversing). All this uncertainty suits the passive aggressor well: the more di

cult it is to identify the appropriate norms for the situation in question, the

less con dent the target is likely to be that the agent is wilfully subverting the relevant norms, and the more plausible is the innocuous interpretation. Note that the behaviour’s having an objectionable interpretation is compatible with the behaviour being deserved by the target. It is possible for an agent to out etiquette or other social or professional norms— which is all that is required for one’s behaviour to count as objectionable in the relevant sense—even if the target deserves it. Sometimes we are justi ed in being rude to people; even so, rudeness is objectionable in the sense of being contrary to etiquette. Passive-aggressive behaviour is necessarily ambiguous in the way just described. To count as being passive aggressive, an agent must make a hostile attack on the target; accordingly, on the correct interpretation of her behaviour, it is objectionable. But there must also be at least one false, but somewhat plausible, interpretation according to which her behaviour is innocuous. What does it mean to say that the innocuous interpretation must be somewhat plausible? The key consideration here is that the agent’s hostile attack must be covert, and the agent must be motivated to keep it so. The agent does not necessarily want to avoid communicating hostility to her target—she may positively aim at doing so—but the target should be discouraged from challenging the agent about her p. 180

behaviour.

The innocuous interpretation helps the agent discourage the target from challenging her

behaviour in two ways. First, the target’s recognition of the innocuous interpretation alongside the objectionable one may confuse the target, leaving her uncertain as to whether she really has been attacked or whether she has incorrectly read an attack into the agent’s innocuous behaviour. Second, even if the target recognizes that she has been attacked, she might suspect that if challenged the agent will insist that the innocuous interpretation is the correct one, which could leave the target looking foolish and aggressive. And of course, even if the innocuous interpretation does not discourage the target from challenging the agent about her behaviour, the agent can respond to such a challenge by appealing to the innocuous interpretation to argue that her behaviour was not objectionable. The innocuous interpretation, then, helps the agent keep her attack on the target covert. We can understand the requirement that the innocuous interpretation be somewhat plausible to mean that it should be su

ciently plausible to serve this purpose. How water-tight the innocuous interpretation needs to be will

depend on the situation, since challenges to the agent are more likely in some circumstances than in others, and the likely aggressiveness of such challenges will vary between situations. Bruna may reasonably judge that she is unlikely to be challenged about her remark to Parvati by onlookers or by Parvati herself, because the norms of politeness at a wedding reception make it likely that any such challenge will appear rude; consequently, the innocuous explanation of Bruna’s behaviour does not need to be robust enough to

withstand sustained interrogation. By contrast, in The Colleague, Miho’s employers may ask for details of his illness and may also ask him to produce evidence in the form of a note from his doctor. As a result, the innocuous interpretation of Miho’s behaviour had better be robust enough to withstand some investigation.

2.3. Test Cases Can a person’s attempt to be passive aggressive fail because the innocuous interpretation of their behaviour is not su

ciently plausible? Yes. Passive aggressors rely on the plausibility of the innocuous interpretation

to keep their aggression covert. If this interpretation turns out not to be plausible then the agent’s cover is blown, his aggression fails to be covert, and his behaviour is consequently not passive aggressive. Even so, whether or not we want to describe an agent whose innocuous interpretation is rejected in this way as having failed to be passive aggressive depends on the context. We’re likely to want to claim that an aspiring passive aggressor whose innocuous explanation is immediately exposed as false by the target has 5

failed to be passive aggressive. But an agent whose attack is exposed by the target only years later succeeded in behaving passive aggressively. Deciding exactly what must hold true of the innocuous interpretation in order for the agent’s behaviour to count as passive aggressive involves deciding what is required in order for the agent’s attack to remain covert, and deciding the latter is largely a matter of setting p. 181

out the necessary and su

cient conditions for covert

behaviour in general (including deciding what sorts

of events that follow the behaviour may undermine its covertness). Rather than attempt to resolve this tangential issue, let us simply content ourselves with the observation that a passive aggressive attack must succeed in being covert, whatever that might involve. We just saw that an agent may fail to act passive aggressively by failing to be passive (that is, covert). An agent can also fail to be passive aggressive by failing to be aggressive (that is, by failing to attack). An agent who attempts a passive-aggressive attack by covertly doing favours for his target fails to be passive aggressive by failing to attack, since covertly doing favours for a person is not a way of attacking them. Can an agent succeed in making a covert attack on his target, yet fail to act passive aggressively by failing to convey aggression to the target? This might happen if the innocuous interpretation is so plausible that the target is left con dently believing that the passive aggressor’s behaviour is benign; consequently, the target does not feel at all o ended. Suppose that in The Colleague, Cassius knows Miho to be an introvert who sometimes does not interact with others. In that case, his sullenness might easily be overlooked. If Cassius also knows Miho to su er from health problems, then Cassius—whilst disappointed and upset that Miho’s apparent illness has resulted in a missed deadline—may not realize that Miho is attacking him. In this case, has Miho tried but failed to be passive aggressive? I think that in this case Miho succeeds in behaving passive aggressively. Just as an anaesthetized person can be attacked by a wasp without noticing, and just as one can be attacked in print by a critic without ever reading or hearing of it, one can be the target of a passive aggressive attack without realizing it, and without the behaviour being any less of an attack. Whilst the target’s obliviousness to the agent’s attack does not disqualify the behaviour from counting as passive aggressive, though, it is likely to be a suboptimal outcome for the agent; after all, when we attack someone we very often (though not always) aim to provoke some response or recognition in them. Having an attack pass unnoticed by the target might be especially wounding for some passive aggressors: many people formerly diagnosed with passive-aggressive personality disorder were observed to su er from a sense of powerlessness, lack of assertiveness, and a feeling that ‘nothing they do seems to make any di erence’ (Wetzler and Morey 1999: 51)—attitudes likely to be reinforced should such a person make an attack that fails to register with the target.

2.4. Similar Behaviours Passive aggression, we have seen, essentially involves both a hostile attack and a certain sort of ambiguity. Behaviours that lack either or both of these features are not passive aggressive. We just considered some such behaviours in our discussion of test cases. Before moving on, let’s consider a non-exhaustive list of few widely recognized types of behaviour that resemble passive aggression, but which fall short of being passive aggressive. p. 182

The rst is double entendre: a type of joke that, interpreted innocuously, is an appropriate comment to make in the context in question, but which bears another interpretation according to which it is a sexually explicit remark. This form is strongly associated with British comedy, and it takes centre stage in the longrunning Viz comic strip, ‘Finbarr Saunders and His Double Entendres’. In one episode, the character Mr Gimlet is inspired by a visit to a pet shop to innuendo-laden reminiscing about his interactions with various animals, including the remark, ‘On a recent holiday to Morocco I paid £10 for the privilege of being tossed o

by a frisky young Arab’ (Thorp 2015). Like passive aggression, double entendre is ambiguous behaviour

that has a correct, norm- outing interpretation (Mr Gimlet is describing in explicit detail a past sexual encounter) alongside an incorrect but plausible innocuous interpretation (Mr Gimlet is merely describing being thrown o

a horse). Unlike passive aggression, it is not a success condition of double entendre that

the objectionable interpretation remains unexposed, and it does not essentially involve a hostile attack. In cases where double entendre does constitute a hostile attack and where the agent is motivated to insist that the innocuous interpretation is correct and the objectionable interpretation incorrect, it counts as passive aggressive. For example, a man who uses double entendre to sexually harass a woman and, when challenged, insists that his behaviour was innocuous, behaves passive aggressively. Also similar to passive aggression is the dogwhistle. Kimberly Witten writes: A dogwhistle is a speech act designed, with intent, to allow two plausible interpretations, with one interpretation being a private, coded message targeted for a subset of the general audience, and concealed in such a way that this general audience is unaware of the existence of the second, coded interpretation. (Witten, n.d.: 2) Dogwhistles can be benign—both Witten and Jennifer Saul discuss the example of the coded messages for parents that often appear in children’s TV shows, which aim only at entertaining cartoon-fatigued parents watching with their children—but they are also used for sinister ends, especially in politics. For example, Saul observes that In the United States, ‘inner city’ has come to function as a dogwhistle for Black. Thus, politicians who would be rebuked if they called for harsher measures against Black criminals can safely call for cracking down on inner city crime. (Saul 2018: 367) This device allows politicians to align their values with those of racist voters without saying anything explicitly racist. Reference to ‘inner city’ is intended to reassure racist voters who want lawmakers to take an aggressive approach to Black criminals while being subtle enough that the coded racist message will remain unheard by voters who would not support a politician whom they believe is racist. Like passive aggression, then, the behaviour of politicians who make use of the ‘inner city’ dogwhistle has an objectionable interpretation (according to which the agent holds racist values) and an innocuous interpretation (according to which the agent is simply concerned about crime). The innocuous interpretation allows the agent to deny the objectionable interpretation.

p. 183

Another example discussed by Saul is former US president George W. Bush’s use of the term ‘wonderworking power’ during his election campaign. This term is recognized by fundamentalist Christians as a reference to the power of Christ. Fundamentalist Christian voters, hearing Bush’s use of the term, see Bush’s values as aligned with theirs. But for voters who are likely to be alienated by fundamentalist Christianity, the term passes unnoticed. Again, there is an objectionable interpretation of Bush’s use of the term—at least, an interpretation that would dissuade many voters from supporting Bush—according to which he is signalling his fundamentalist Christian values, and an innocuous interpretation according to which Bush’s use of the term does not signal anything in particular about his faith. To count as passive aggressive, it must not only be the case that a dogwhistle is ambiguous in the right sort of way and that the agent is motivated to deny the objectionable interpretation; it must also constitute a hostile attack on a target. Bush’s use of ‘wonder-working power’ can hardly be thought of as a hostile attack; as a result, it is not passive aggressive. By contrast, use of the ‘inner city’ dogwhistle plausibly constitutes a hostile attack on Black people. Given that politicians who use this dogwhistle are motivated to keep their racism covert, use of the ‘inner city’ dogwhistle meets the criteria for passive aggression. From these re ections, we can see that dogwhistles are not necessarily passive aggressive, although some— including the ‘inner city’ dogwhistle—meet the criteria for passive aggression. Another behaviour related to passive aggression is sulking. Passive aggression can sometimes look like 6

sulking: it can involve the agent being sullen towards a target, which is also characteristic of sulking. Both aim to express covertly what the agent is unable or unwilling to communicate explicitly. Both, too, express negative attitudes towards the target; with sulking, but not always with passive aggression, these mainly involve a sense of having been wronged by the target. But whilst passive aggressors want to avoid discussing their negative feelings with their targets, sulkers solicit it. Indeed, the ultimate aim of a sulk is sympathetic discussion with the target of the sulker’s negative feelings (plausibly, there are conditions attached to this: the target should correctly guess the negative feelings of the sulker without the aid of clues from the sulker, and the target should make no mention of the sulker’s employment of sulking to achieve his aim). Sulking, 7

then, is not passive aggressive.

3. Whatʼs Wrong with Passive Aggression? ‘Passive aggressive’ is, as we have noted, a pejorative term. Its undesirability is re ected in the fact that it captured the interest of psychiatrists—even nding its way into o

cial nosology—and continues to be

viewed by clinicians and counsellors as a problem behaviour. Let’s explore exactly why we view it as undesirable.

p. 184

3.1. Passive Aggression Is Bad for the Target It seems clear that for Menninger, passive-aggressive behaviour was undesirable because it led to soldiers failing to obey orders, which jeopardized the successful operation of the army as a whole. Menninger explicitly stated that his concern for the wellbeing of the individual soldier was secondary to his concern for 8

the smooth functioning of the military unit of which the individual is a member. Passive aggression, then, was undesirable primarily because of its e ects on the individual’s social and professional functioning. Our everyday reactions to passive aggression re ect this. Passive-aggressive people can be di

cult to deal

with. Just as Menninger’s passive-aggressive soldiers fail to ful l their military duties, our passiveaggressive friends and colleagues fail to act in accordance with social and professional norms. These norms may sometimes be relatively formal and well de ned, like the professional norm that requires Miho to complete his work in advance of the deadline in The Colleague. Often, though, these norms are less well

de ned and sometimes so ill-de ned that their existence comes to light only when they are not met: consider the norm that requires Erno to reserve Avril’s seat in The Boyfriend. People who are intent on outing social and professional norms can cause great inconvenience to others. In addition to the negative impact of passive aggression on the agent’s social and professional functioning, we dislike being the target of other people’s passive aggression because it is a form of attack, and we dislike being attacked. On balance, we prefer that the people with whom we interact are respectful and cordial, and that they are able to resolve disagreements without resorting to attacking behaviour. Whilst we recognize that this ideal is not always realized and that sometimes people attack each other, it often seems that passive aggression is an especially objectionable form of attack. Targets of passive aggression commonly remark that they would have preferred to have been overtly attacked. This preference is expressed in oftheard remarks like, ‘If she has a problem with me, why doesn’t she just come right out and say so?’, ‘I wish he had refused to do it instead of agreeing and then ignoring my emails,’ and ‘It would have been more honest of her to tell me to sod o

than to keep quiet and then unfriend me on Facebook.’ This suggests that

we dislike passive aggression not merely because it is a form of attack, but because of the particular form of attack that it is. What makes the di erence here? The ambiguity of passive-aggressive attacks makes them especially unpleasant for the target, for two main reasons. First, the target may suspect that she is the victim of a hostile attack without being con dent that this is the case, and the resulting uncertainty makes for a more unpleasant experience than the target would 9

have were she certain that she was under attack. Second, the ambiguity makes it di

cult for the target to

challenge the attacker’s behaviour or to address the attacker’s feelings towards her. One reason for this is that the target may suspect that the passive aggressor will, if challenged, appeal to the innocuous interpretation in order to deny having behaved objectionably, which risks leaving the target looking foolish. Another reason is that passive aggression makes calm and sympathetic discussion of the passive p. 185

aggressor’s feelings di

cult. If

the target wishes to initiate a discussion of the passive aggressor’s

hostility towards her, then she must do so by challenging the passive aggressor’s objectionable behaviour, or at least by observing that the latter is attempting to express hostility covertly. This type of approach is hard for the target to take without being somewhat confrontational and accusatory; as a result, the target may be dissuaded from attempting it. Challenging a passive aggressor, then, risks appearing hostile oneself; a risk which the target may reasonably resent, especially given the passive aggressor’s evident wish to avoid appearing hostile. The target of passive aggression must choose between, on the one hand, su ering the passive aggressor’s hostility without challenging it, and on the other, challenging it and running the risk of being accused of oversensitivity or appearing hostile. Because of all this, being the target of passive aggression not only involves being attacked; it also involves being silenced, in an important way. What do these re ections reveal about in what sense, if any, passive aggression is wrong? Well, attacking another person is very often inconsiderate and rude, especially when one has the option of engaging with them in a way that does not involve attacking them. Moreover, attacking a person in a way that makes it di

cult for them to respond is also inconsiderate and rude, in approximately the way that hanging up on

someone to whom one has been talking over the phone, before they have had a chance to respond to what one has said to them, is inconsiderate and rude. For these reasons, passive aggression is bad mannered, or contrary to etiquette. This makes passive aggression doubly contrary to etiquette: in addition to the fact that behaving passive aggressively is itself contrary to etiquette, we saw in section 2 that for behaviour to count as passive aggressive, it must covertly out norms of etiquette or other social or professional norms. In addition to being contrary to etiquette, is passive aggression contrary to morality? No doubt consequentialists and virtue ethicists could make promising cases for the moral badness of passive aggression, but choosing to attack someone passive aggressively rather than overtly is not unethical in the deontological sense of violating the rights of the target. Passive aggression might, however, facilitate and exacerbate immoral behaviour. In many cases, it is immoral to intimidate, harass, or otherwise attack

another person, and doing so passive aggressively rather than overtly can make it less likely that the agent will be challenged about his behaviour and brought to account. Agents who are reluctant to commit such wrongs overtly may, as a result, be willing to commit them passive aggressively.

3.2. Mitigating Factors As with other objectionable behaviours, there may be mitigating factors. It is inappropriate to blame a person for behaving passive aggressively in cases where she has little option to behave in a less objectionable way. A striking example of this is the very rst documented case of passive aggression: Menninger described as ‘passive aggressive’ US soldiers during the Second World War who resisted p. 186

complying with orders ‘by passive ine

measures, such as pouting, stubbornness, procrastination,

ciency, and passive obstructionism’. It is not di

cult to sympathize with Menninger’s passive-

aggressive soldiers, given their circumstances. Complying with orders from their superiors during a war likely required these soldiers to take great risks, including the risk of death. On the other hand, by refusing to comply they risked being punished. Their behaviour can be explained by their desires both to avoid risking death and to avoid being punished for failing to comply with orders. The signi cant negative consequences attaching to non-passive-aggressive courses of action in this case makes the soldiers’ passive aggression less blameworthy than it would otherwise be. Their plight resembles that of a victim of blackmail: just as a victim of blackmail is dissuaded by the prospect of signi cant negative consequences (imposed by the blackmailer) from pursuing any course of action other than that demanded by the blackmailer, Menninger’s soldiers are dissuaded by the prospect of signi cant negative consequences (either imposed by their superiors or arising from their following orders) from pursuing any course of action other than covert resistance. Whilst Menninger, with his focus on ensuring the smooth functioning of army units, viewed this behaviour as pathological, from the point of view of the individual soldier it makes sense as the only available strategy aimed at avoiding the consequences of both compliance and explicit non-compliance. As the blogger Cecil Adams has remarked on The Straight Dope website, ‘If you’ve ever served in the military during wartime … or for that matter read Catch-22, you realize that what the brass calls a personality disorder a grunt might call a rational strategy to avoid getting killed.’ (Adams 10

2003, cited in Lane 2009: 58).

Cases of passive aggression in which the agent is as powerless as Menninger’s soldiers are relatively rare. Most cases of passive aggression that we encounter day-to-day do not involve such severe consequences of non-passive-aggressive action. Even so, sometimes there are factors that might encourage a person to express their hostility passive aggressively rather than non-passive aggressively. In The Colleague, Miho feels jealous and resentful of Cassius’s success. His knowledge that these attitudes are not generally viewed sympathetically gives him a motive to express them passive aggressively, rather than overtly. This motive would be absent were his negative attitudes towards Cassius the sort that are viewed more sympathetically —as would be the case, say, if Miho felt angry and outraged at Cassius for committing a widely-recognised moral wrong. That some negative feelings are not viewed sympathetically should probably not lead us to view their passive-aggressive expression as blameless, but it does go some way towards increasing our understanding of why people behave passive aggressively. It’s not only the agent who might stand to gain by attacking passive aggressively rather than overtly. Sometimes, it bene ts the target (and others) too. Were Miho to express his hostility overtly to Cassius, Cassius might feel under pressure to respond in some way, such as by disciplining Miho or by asserting dominance and authority in some other way. Cassius’s response would not only enable him to object to Miho’s treatment of him, but also to signal to Miho and any witnesses that Cassius will not tolerate Miho’s behaviour. Especially if Miho’s attack on Cassius is witnessed by others, Cassius’s failing to respond p. 187

appropriately might result in his losing the respect of his co-workers. By

contrast, it is easier to pretend

publicly that Miho’s passive-aggressive behaviour is not an attack at all—even if everyone in the workplace is privately certain that it is. This is because a passive-aggressive attack can result in what Steven Pinker and colleagues have called shared individual knowledge yet fail to result in common knowledge. Pinker and colleagues explain that, ‘[i]n common knowledge, not only does A know x and B know x, but A knows that B knows x, and B knows that A knows x, and A knows that B knows that A knows x, ad in nitum’ (Pinker et al. 2007: 837). The transmission of common knowledge can have important social consequences, which can be avoided when the knowledge transmitted is merely shared. This is illustrated vividly by the story of the Emperor’s New Clothes: When the boy called out, ‘The emperor is naked!’ he was not telling the onlookers anything they didn’t already know. Yet he was conveying knowledge nonetheless: Now everyone knew that everyone else knew, and that everyone else knew that they knew, and so on, and that common knowledge licensed the people to challenge the dominance relationship commanded by the emperor. (Pinker et al. 2007: 837) Communicating indirectly, even when there is little doubt as to what is being communicated, can enable one to share information whilst avoiding the sort of relationship renegotiation that would be prompted by communicating overtly. In cases where an agent’s motivation to attack passive aggressively rather than overtly includes a desire to spare the target the signi cant uncomfortable social consequences that might result from an overt attack, we can take the agent’s judgments about the likelihood of these to be a mitigating factor in passive aggression.

3.3. Passive Aggression Is Bad for the Passive Aggressor We’ve considered why passive aggression is bad for its targets and for others who must interact with passive aggressive people. Let’s end by considering how passive aggression is bad for passive aggressors. An obvious way in which passive aggression is bad for passive aggressors is that it does nothing to address and resolve the hostility that it expresses, as would happen were the agent instead to approach the target to request a calm discussion of the agent’s grievances, or were the agent to take some other steps to address the issue causing her to have the negative attitudes in question. But, in fact, the situation for the passive aggressor is likely to be even worse than this. Not only does behaving passive aggressively do nothing to address and resolve the agent’s hostility, but plausibly it makes addressing and resolving her hostility more di

cult than it would have been had she not expressed her hostility at all. This is partly because, as we have

seen, passive aggressive behaviour discourages the target from initiating a calm discussion with the agent of the agent’s hostility. It is also because, having expressed her hostility passive aggressively, the agent herself will likely nd it di p. 188

her earlier behaviour

cult to approach the target for a discussion about her hostility without exposing

as passive aggressive, and therefore objectionable. Anticipating this, the agent may

be dissuaded from attempting such a discussion. We have already seen that there is a sense in which passive aggression silences the target; these current re ections reveal that there is also a sense in which it silences the passive aggressor. That behaving passive aggressively makes it more di

cult for an agent to address and resolve her negative

attitudes is especially concerning given that—as Wezler and Morey observe (1999: 57)—people who regularly behave passive aggressively commonly view themselves as powerless. We can speculate that if viewing oneself as powerless makes one more likely to behave passive aggressively, and if behaving passive aggressively makes it more di

cult to address one’s negative attitudes, then the resulting di

culty in

resolving one’s grievances is likely to exacerbate one’s view of oneself as powerless and strengthen one’s belief that ‘nothing [one does] seems to make any di erence’ (Wetzler and Morey 1999: 51). If this is right,

then feeling powerless and a tendency to favour passive aggressive behaviour are mutually reinforcing states.

4. Conclusion I have attempted here to characterize passive aggression, and to make some remarks about why passiveaggressive behaviour is objectionable. Passive aggression, I have argued, is behaviour that constitutes a hostile attack on a target. This attack takes the form of norm- outing behaviour that is ambiguous in the sense that, from the point of view of the target and any onlookers, it has a plausible (though false) innocuous interpretation. Passive aggression is objectionable to those who encounter it in others because the passive aggressor’s norm- outing behaviour impedes his social and professional functioning; because being the target of passive aggression involves being attacked, and people generally object to being attacked; and because the nature of a passive-aggressive attack is especially objectionable in that its ambiguity both makes for an unpleasant experience for the target, and makes it di

cult for the target to

challenge the passive aggressor’s behaviour. Finally, passive aggression is bad for passive aggressors because it does nothing to resolve the negative attitudes that it expresses, and it may also make those negative attitudes more di

cult to address—which, in turn, might exacerbate the agent’s sense of

powerlessness.

Acknowledgements Several people have helped and encouraged me as I developed my arguments on this topic. These include David Edmonds, who invited me to discuss passive aggression for his podcast, Philosophy 24/7, and chose his interview with me as one of two episodes with which to launch the podcast; Nigel Warburton, who commissioned me to write an article on passive aggression for Aeon; and Katrien Devolder, who interviewed me about passive aggression for the University of Oxford’s Practical Ethics YouTube channel. Along the way, p. 189

each of them o ered

incisive questions, helpful feedback, and enthusiastic support of the project. I have

delivered talks on the topic at the University of Oxford and at Royal Holloway, and I’m grateful to those who attended for their thoughtful contributions. I’m also thankful to Clare Moriarty and Irina Manta, who generously read and provided comments on an early draft of the chapter.

Notes 1.

Cited in Lane (2009: 58).

2.

These examples are an attempt to reflect widespread contemporary intuitions about what constitutes passive aggression based on descriptions of passive aggression by psychologists, psychotherapists, counsellors, and other members of the public. The Wedding Guest is an example of a backhanded compliment, a form of passive aggression discussed by Long et al. (2008) that is the focus of a long thread on Reddit (Reddit 2017), and which is mentioned in three of the top five Google results of a search on the term ʻpassive aggressiveʼ (Whitson 2010; Brogaard 2016; and Wikipedia (n.d.), ʻPassiveAggressive Behaviorʼ, https://en.wikipedia.org/wiki/Passive-aggressive_behavior, accessed 25 January 2020). The Colleague depicts distinct properties of passive-aggressive behaviour (specifically, ʻtemporary complianceʼ and ʻintentional ine iciencyʼ) described in Long et al. (2008). The Boyfriend is based on a real-life case of passive aggression identified by the psychologists E. Tory Higgins and Ozlem Ayduk and reported in Carey (2004). See also Ayduk et al. (2003).

3.

There is more to say, which I will not say here, about what is involved in expressing an emotion or attitude. This issue has been discussed, not only in relation to the philosophy of action, but also in relation to issues as diverse as art and punishment. See, for example, Cockburn (2009); Levy (2011); Levinson (2006); and the essays in Abell and Smith (2016).

p. 190

4.

There is scope to distinguish between etiquette, politeness, manners, and the particular norms and expectations that arise between friends and colleagues, and the writers quoted in this paragraph all make at least some such distinctions. Since the details of these distinctions are not important for our purposes here, I will not discuss them, and in what follows I will occasionally use these terms interchangeably.

5.

We may, in such circumstances, recognize that the agent was attempting to be passive aggressive; as a result, we may be justified in ascribing to the agent the character trait of passive aggression, in that we view the agent as someone who favours a passive-aggressive approach to dealing with others. Given that the agentʼs current attempt at passive aggression has failed, plausibly ascribing to such a person the character trait of passive aggression likely involves assuming that, despite the failure of the current attack to be passive aggressive, her other attempts usually succeed.

6.

Sulking is mentioned as a behaviour associated with passive aggression in DSM-III (APA 1980). However, no analysis of sulking behaviour appears there, and it is not clear whether (and if so, how) it is to be distinguished from mere sullenness.

7.

For more on sulking, see Roache (2023).

8.

ʻThe unit becomes the important element rather than the individual since it is our aim to help that particular unit achieve its goal. In our civilian medical practice we were interested only in the individual. In the Army, we are interested in the company or the platoon and how the man fits into the teamwork of this company or platoonʼ (Menninger 2004: 289).

9.

There is empirical evidence for this. See Bar-Anan et al. (2009).

10.

Given Menningerʼs account of the soldiers in question, it is puzzling that he chose to ascribe to them a form of aggression at all. At best, his evaluation of their behaviour is unsympathetic. At worst, his description of them as aggressive might itself be viewed as a hostile attack, which—given that it is made under the innocuous guise of medical diagnosis—makes Menningerʼs original diagnosis of passive aggression itself an act of passive aggression.

References Abell, Catharine, and Smith, Joel, eds. (2016), The Expression of Emotion (Cambridge, UK: Cambridge University Press). Google Scholar Google Preview WorldCat COPAC Adams, Cecil (2003), ʻWhat Is “Passive-Aggressive”?ʼ, The Straight Dope, 30 May. https://www.straightdope.com/columns/read/2453/what-is-passive-aggressive/. APA (American Psychiatric Association) (1952), Diagnostic and Statistical Manual of Mental Disorders, 1st edn [DSM-I] (Washington, DC: Mental Hospitals Service). Google Scholar Google Preview WorldCat COPAC APA (American Psychiatric Association) (1980), Diagnostic and Statistical Manual of Mental Disorders, 3rd edn [DSM-III] (Washington, DC: Mental Hospitals Service). Google Scholar Google Preview WorldCat COPAC APA (American Psychiatric Association) (1987), Diagnostic and Statistical Manual of Mental Disorders, rev. 3rd edn [DSM-III-R] (Washington, DC: Mental Hospitals Service). Google Scholar Google Preview WorldCat COPAC APA (American Psychiatric Association) (1994), Diagnostic and Statistical Manual of Mental Disorders, 4th edn [DSM-IV] (Washington, DC: Mental Hospitals Service). Google Scholar Google Preview WorldCat COPAC APA (American Psychiatric Association) (2013), Diagnostic and Statistical Manual of Mental Disorders, 5th edn [DSM-5] (Washington, DC: Mental Hospitals Service). Google Scholar Google Preview WorldCat COPAC Ayduk, O., May, D., Downey, G., and Higgins, E. T. (2003), ʻTactical Di erences in Coping with Rejection Sensitivity: The Role of Prevention Prideʼ, Personality and Social Psychology Bulletin, 29/4, 435–448. Google Scholar WorldCat Bar-Anan, Y., Wilson, T. D., and Gilbert, D. T. (2009), ʻThe Feeling of Uncertainty Intensifies A ective Reactionsʼ, Emotion, 9/1, 123– 127. Google Scholar WorldCat Brogaard, Berit. (2016), ʻ5 Signs That Youʼre Dealing with a Passive-Aggressive Personʼ, Psychology Today, 13 November, https://www.psychologytoday.com/gb/blog/the-superhuman-mind/201611/5-signs-youre-dealing-passive-aggressive-person. Buss, Sarah (1999), ʻAppearing Respectful: The Moral Significance of Mannersʼ, Ethics, 109/4: 795–826. Google Scholar WorldCat Carey, B. (2004), ʻOh, Fine, Youʼre Right. Iʼm Passive Aggressiveʼ, New York Times, 16 November, https://www.nytimes.com/2004/11/16/health/psychology/oh-fine-youre-right-im-passiveaggressive.html. Cockburn, D. (2009), ʻEmotion, Expression and Conversationʼ, in Y. Gustafsson, C. Kronqvist, and M. McEachrane, eds., Emotions and Understanding (Basingstoke, UK: Palgrave Macmillan), 126–144. Google Scholar Google Preview WorldCat COPAC Lane, Christopher. (2009), ʻThe Surprising History of Passive-Aggressive Personality Disorderʼ, Theory & Psychology, 19/1, 55–70. Google Scholar WorldCat Levinson, Jerrold. (2006), Contemplating Art (Oxford: Clarendon Press). Google Scholar Google Preview WorldCat COPAC

Levy, N. (2011), ʻExpressing Who We Are: Moral Responsibility and Awareness of Our Reasons for Actionʼ, Analytic Philosophy, 52/4, 243–261. Google Scholar WorldCat Long, J. E., Long, N. J., and Whitson, S. (2008), The Angry Smile (Austin, TX: Pro Ed). Google Scholar Google Preview WorldCat COPAC p. 191

Martin, J. and Stent, G. S. (1990), ʻI Think Therefore I Thank: A Philosophy of Etiquetteʼ, The American Scholar, 59/2, 237–254. Google Scholar WorldCat Menninger, W. C. (2004), ʻContributions of Dr. William C. Menninger to Military Psychiatryʼ, Bulletin of the Menninger Clinic, 68/4, 277–296. Google Scholar WorldCat Pinker, Steven, Nowak, Martin A., and Lee, James J. (2007), ʻThe Logic of Indirect Speechʼ, Proceedings of the National Academy of Sciences of the United States of America, 105/3, 833–838. Google Scholar WorldCat Reddit (2017), ʻWhatʼs Your Best Passive Aggressive “Compliment”?ʼ Discussion Thread. https://www.reddit.com/r/AskReddit/comments/7mnaot/whats_your_best_passive_aggressive_compliment/. Roache, Rebecca (2023), ʻThe Joy of Sulkʼ, Aeon, 12 May. https://aeon.co/essays/sulking-is-a-fascinating-form-of-indirectcommunication. Saul, Jennifer (2018), ʻDogwhistles, Political Manipulation, and Philosophy of Languageʼ, in D. Fogal, D. W. Harris, and M. Moss, eds., New Work on Speech Acts (Oxford: Oxford University Press), 360–383. Google Scholar Google Preview WorldCat COPAC Stohr, Karen (2012), On Manners (New York: Routledge). Google Scholar Google Preview WorldCat COPAC Thorp, S. (2015), ʻFinbarr Saunders and His Double Entendresʼ, Viz, 23 January. https://web.archive.org/web/20210307113250/https://viz.co.uk/2015/01/23/decision-buy-pet-not-taken-lightly/ Wetzler, S., and Morey, L. C. (1999), ʻPassive-Aggressive Personality Disorder: The Demise of a Syndromeʼ, Psychiatry, 62/1, 49–59. Google Scholar WorldCat Whitson, S. (2010), ʻ10 Things Passive-Aggressive People Sayʼ, Psychology Today, 23 November, https://www.psychologytoday.com/gb/blog/passive-aggressive-diaries/201011/10-things-passive-aggressive-people-say. Witten, K. (2008), ʻDogwhistle Politics: The New Pitch of an Old Narrativeʼ, coursework essay, San Francisco State University, https://www.academia.edu/42929858/Dogwhistle_Politics_the_New_Pitch_of_an_Old_Narrative. WorldCat

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

p. 192

9 Taboo Word Research  Timothy B. Jay https://doi.org/10.1093/oxfordhb/9780192844118.013.32 Published: 22 May 2024

Pages 192–208

Abstract This chapter addresses speakers’ use of emotionally hurtful words and their victims’ emotional responses to hurtful words in American English. Hurtful words are de ned here as negatively valenced or bad words used to express one’s feelings and convey those feelings to others. Topics include anger and frustration, hate speech, obscene phone calls, workplace sexual harassment, verbal abuse, internet and online o enses, and research involving children as speaker-victims. The research cited here is drawn primarily from psychological and social science empirical studies on the topic of how people produce and comprehend o ensive speech. The chapter includes an examination of ethical and methodological problems associated with hurtful word research and concludes with suggestions for conducting meaningful research on hurtful words.

Keywords: hate speech, hurtful words, obscene phone calls, sexual harassment, verbal abuse, online o ensives Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction This chapter addresses two issues scholars face when conducting research on taboo words (de ned in section 1.2): awareness of the past problems with research and planning meaningful future research. Due to my training in psychological science, conducting meaningful research involving taboo words has been my primary scholarly endeavor since 1972. The overarching question hang been, What are the appropriate ways to conduct and publish research on taboo words? To improve the way we conduct research, my mission is to outline the major problems we scholars face. These are the issues with taboo word research that we all need to address at some point in our research programs, including those associated with publishing our work.

1.1. Problems from the Past Years ago, I addressed some of the methodological problems with taboo word research (Jay 1977), for example, how to de ne the category of taboo words and how to obtain accurate word frequency and word o ensiveness data. At the time, I noted that standard procedures needed to be established. Unfortunately, in the 1970s, there was scant research on the topic of how to conduct such research. Since then, a signi cant body of scholarly research has been accumulating (e.g., Bergen 2016; Beers Fägersten and Stapleton 2017; Jay 2009). However, the methodological problems that were identi ed years ago remain (e.g., de ning the category of taboo words), and a set of new problems has arisen concerning describing the brain mechanisms underlying swearing or determining how taboo words are used in computer-mediated communication (CMC).

p. 193

1.2. The Problem of Categorizing Taboo Words The initial goal of any scienti c research is to describe the subject of interest. Once a subject matter has been properly de ned, a research plan can be built. The goal for scholars is to be as clear and accurate as possible when de ning what they are studying. Here, we start with the term taboo, which refers to prohibited or restricted social customs. Taboo words are those restricted in usage in any language. The term taboo words is used collectively, or distributively, to refer to several semantic subcategories that make up the corpus of words in many languages that are restricted in usage. Taboo words are a category of words that includes o ensive words and references such as profanity/blasphemy, obscenity/indecency, slang, words for sexualized body parts and behaviors, ancestral allusions, name calling, scatology, racist/gender/ethnic slurs, and vulgarity. Each subcategory contains di erent but overlapping sets of words. In the past, problems have appeared because scholars have not been clear when de ning what they were studying. A common example is using a more speci c term, such as profanity (religious word taboos) or obscenity (legally de ned sexual and excretory word taboos), too broadly. Detailed discussions of de nitional problems appear in the work of Adams (2016), Croom (2014), and Jay (1992, 2000, 2003, 2009, 2017). Taboo words can be single words or phrases or conventionalized, insulting emotional expressions and idioms, for example, you think your shit don’t stink, you don’t know your ass from your elbow, or, as translated from Farsi, what did you do eat donkey brain? One can see what kinds of word taboos exist in a culture by looking for the presence of non-taboo euphemisms, which are acceptable words used to replace taboo words. Examples of the use of euphemisms in American English are saying sugar instead of shit, darn instead of damn, or jeezum crow instead of Jesus Christ! Euphemistic expressions stand in as polite substitutions for impolite taboo counterparts which many speakers are reluctant to utter publicly. Perceiving a taboo word as a member of a category depends on one’s knowledge of the language in question. One needs to have a thorough knowledge of that language to appreciate the various shades of meaning and levels of o ensiveness of words and phrases within a subcategory. Each subcategory of taboo words is

complex, and one needs an understanding of the nuances of taboo word meanings. For example, Leo Rosten (1968) examined the di erent shades of insults in Yiddish, such as di erent ways to refer to a person who is inept: shlemiel, shlepper, shlimazl, shmegegge, or shmo. If you understand Yiddish, you appreciate the di erent shades of meaning each of these insults implies; if you do not understand Yiddish, you might categorize the insults as having the same meaning. Scholarship on taboo words is now global in scope, but not all scholars have native knowledge of the language they research. When we study the meaning of taboo words in a native language or a nonnative language, we need knowledge of what we are trying to analyze. Without a native speaker’s knowledge, we p. 194

might understand only that a word is a taboo word but not understand its ner contextual restrictions and inappropriateness. Native knowledge helps speakers judge how to employ pragmatic strategies involving slurs and insults. Nonnatives and language learners, due to lack of exposure or experience, have not yet acquired these kinds of pragmatic strategies. Consider how sexuality can be described in clinical terms, e.g., intercourse, as well as in vulgar terms, e.g., fucking. The referent is the same, but the expression of it is quite di erent. Native knowledge is used to judge when and where to use intercourse and when to use fucking. We must de ne the taboo words we are studying clearly. Are they sexual words, religious words, slurs, or something else? But once we have described the category of interest, we must not lose sight of the relationship between taboo words and language in general. We should not imply that taboo words are independent of our general semantic lexicon because taboo and non-taboo cognates coexist in human semantic memory. Poop, shit, and feces have essentially the same referent. Vulgar terms (e.g., shit) have a direct semantic connection to the clinical terms (e.g., feces) for the same referent. Speakers will choose the appropriate term to use in a dialogue based on the situation, choosing perhaps feces when communicating with a physician but using shit with a close friend. Although the words di er in emotional force and pragmatic use, they are not isolated from each other in semantic space. Taboo words, although they are studied and reported as if they are independent of non-taboo words, are part of our general lexicon and cannot be completely divorced from it.

1.3. Structural and Functional Di erences Consider how psychologists and neuroscientists execute their taboo word research versus how linguists, philosophers, and sociolinguists execute their research. An approach to taboo word research based on delineating the semantic categories, grammar/syntax, or meaning of utterances involved is referred to as a structural approach. Good examples of a structural approach to language are any of Noam Chomsky’s early works on linguistics, which outlined the grammatical structure of language and, much less so, semantics. Linguists like Chomsky paid little attention to the function of language, why people speak to each other. A structuralist like Chomsky wants to know if sentence X is a well-formed one, but not why a speaker said sentence X. In contrast, a functionalist wants to study why the speaker said sentence X. The functional approach insists on paying attention to the interpersonal social and emotional dynamics that give rise to taboo word use— that is, the basic pragmatic aspects of taboo word use, the how, who, what, when, and where of taboo speech (see Jay 1992, 2000). As valuable and appealing as the structural approach has been for many language scholars, it is insu

cient to capture the rich interpersonal dynamics that give rise to the use of taboo words.

Importantly, we need to know the pragmatic functions of taboo word use in social situations, such as those p. 195

involving humor, self-deprecation, anger, retaliation,

name-calling, or social bonding. Whether a taboo

phrase is grammatical and structurally well-formed, or not, is not so important. What is needed for a more complete picture of taboo word use is an elaboration of various purposes or pragmatic functions that taboo words provide for speakers and listeners.

1.4. Problems Defining Swearing and Accounting for Subjectivity A structuralist approach addresses the nature of the words under study, but it does not address how and why people use taboo words in public—that is, why people swear. Swearing is the use of o ensive emotional speech to express one’s feelings and convey them to others. To answer the question about behaviors associated with taboo word use, we need to look at what people are doing with taboo words. People swear because swearing achieves personal and social goals. The functionalist focus is on what a speaker is trying to achieve with taboo words and how the listener is reacting to them—that is, on nding the pragmatics underlying the swearing. The functional approach determines what people are doing with words, the kind of issues and analysis addressed by Herb Clark (1996) in his book Using Language. A functionalist asks, “Why do people swear?” The primary purpose for swearing is emotional expression; venting anger and frustration are good examples of this. For example, Jay (1992), in a study of emotional speech at a summer camp, found that the majority of the episodes of taboo word use involved the speakers’ expressions of anger or frustration. A related purpose is to communicate our emotions to others so that they know what we are feeling and will be a ected by the words we use. One of the socially negative aspects of taboo words is that they can be used to slur, marginalize, or dehumanize others (Croom 2014). Swearing can achieve social goals such as receiving an apology, starting an argument, dehumanizing others, or gaining sympathy from bystanders. To address the personal or social goals of swearing, we need to address why and how people use taboo words—that is, their pragmatic function. Understanding the function of swearing requires a cultural knowledge (understanding) of what we think people are trying to do. Our awareness of pragmatic functions is based on our (biased) cultural perceptions of speech acts underlying expressions of anger, frustration, self-deprecation, or humor in a given cultural setting. All competent speakers learn now to swear in their native language; whether they do end up swearing is another matter. Native speakers, to avoid breaking conventions or laws, carry with them a cultural knowledge of when and how to use taboo words. We also used this sociopragmatic knowledge when we try to comprehend what others are saying. The notion that there are normative functions for taboo word use implies that there are cultural practices that native speakers learn to help them decide what are the proper and improper uses of taboo words—for example, when to tell an o ensive joke. If native speakers learn how p. 196

to use taboo words through experience, this means that young

children and non-native speakers,

depending on their second-language pro ciency, will be de cient or naive with respect to adult taboo word competency, depending on how extensive the requisite experience is (see Dewaele 2004). The notion of variable speaker knowledge of taboo words raises the problem of subjectivity. How do we decide what a child or non-native speaker intends when saying taboo words? Naive speakers might understand the meaning of a given taboo word but use the term inappropriately. Non-native speakers and children, by means of their upbringing and education, may not have learned the pragmatics of swear words, producing a naivety toward them. We observers (subjectively) can misinterpret what children or non-native speakers are doing with words: did they really intend to be so insulting? Consider the complex social problems of sexual harassment and hate speech; these concepts require a cultural knowledge and a sociopragmatic knowledge of the “who, what, where, and when” regarding the use of o ensive speech. Changing the context of the speech, any of the cultural elements (who and where) may place an incident outside the de nition of harassment. Saying “We should fuck tonight” to a colleague at work could be considered sexual harassment and is quite di erent than saying the same thing to your spouse in your bedroom. Some sociopragmatic experience is necessary to understand this. The legal concept of sexual harassment is premised in part on the nature of “unwanted” comments of a sexual nature, and what is unwanted in a given context (motorcycle factory versus day-care setting) depends on the sensitivity of the speech community therein. What is harassment is one setting may not reach that threshold in another.

Dewaele (2004) demonstrated that native speakers of a language perceive the emotional strength of its swearwords and taboo words to be higher than non-native speakers of that language do. Further, the emotional strength of words is higher for non-natives who learned o ensive words in natural contexts as compared to those who learned them in a classroom setting. This nding raises a more general question regarding the relation between emotion and speech. One of the more mysterious and unpredictable qualities of human speech behavior is understanding how an emotion is expressed through speech: How does anger or depression a ect the way we talk and what we say? Can we control these kinds of problematic emotions that o end people? At what point do deep emotions get translated into speech production? Answering these questions leads to something of an enigma: language scholars use the rational, objective, analytical tools of science in order to understand the nature of pre-verbal, pre-linguistic emotional expressions that are not penetrable by speech because the emotions have yet to be converted to speech (see Jay 2003). The problem is that the pre-verbal or pre-linguistic stages prior to an act of swearing are impenetrable to linguistic analysis (semantics or grammar)—because the emotion stage and the language-production stages exist in di erent codes or formats. Swearing is in language, but the origin of emotion is not in language; emotion is pre-verbal; it is not in language, yet. An emerging problem for taboo word researchers is how to compare public swearing with taboo word use in p. 197

computer-mediated communication (Jay 2018). With increasing

research on taboo word use in CMC

contexts, it is important to note that the pragmatics of swearing in face-to-face encounters is di erent than the pragmatics in CMC. Face-to-face encounters are predicated on a shared physical presence where the characteristics of the speaker and listener are available through eye contact, posture, authority/dominance, physical size, cues as to social status (e.g., clothing and vocabulary), and emotional reactions, as well as voice tone and volume. Online communication lacks these critical physical and social cues, and in many cases a user’s identity is anonymous or misleading. Anonymity and the lack of physical presence allow for more liberal use of taboo words since the consequences for swearing in CMC are not at all clear, not as clear as the consequences in face-to-face violations. Forty-eight years of my research (see Jay 1977, 2018) have driven home the point that swearing is a normal activity, a predictable part of human communication in North America. It is a normal human behavior, not a rare, deviant form of communication. Native speakers know that taboo words are used in the context of general communication between people where one person is usually (but not always) expressing some form of emotion. Emotional o ensive speech is not an oddity; it is how people communicate their emotions to others. It is normal for children and non-native speakers to be eager to learn the meaning of the o ensive words they encounter, whether they will ultimately use them or not.

1.5. Problems with Cross-Cultural Di erences and Accurate Translations A primary goal has been to document research on swearing obtained from a variety of social and cultural situations. The emergence of cross-cultural similarities reveals commonalities in the way humans express their emotions with taboo words. These e orts are bound to o er new information about similarities and di erences in how cultures use taboo words. Research on cross-cultural di erences heightens the need to document the unique ways cultures express themselves. Cross-cultural di erences in taboo word use oppose the tendency to think in terms of a universal “average” speaker, because these di erences cannot be averaged over. To date, cross-cultural research on swearing has failed to address the language practices of marginalized and understudied communities on a global basis. Unless we reach these communities, we will develop an understanding of swearing that is limited in application. Studies of o ensive words in American English are generally restricted in focus to information obtained from college students or from surveys conducted over the internet. As a result, we have little information from English speakers in marginalized communities who

do not attend colleges or do not own or use internet communication architecture (see Jay and Jay 2013). Until scholars address these kinds of sampling limitations, we will continue to perpetuate an incomplete account of how speakers from all social strata use swearwords. p. 198

Another problem for those of us who study foreign languages is that we tend to interpret taboo words in other languages through the lens of our native languages (the subjectivity problem). Native English speakers assess what foreign speakers mean when using sexual terms by using what we know about English sexual terms. As an example of the interpretation problem, Jay (2000) has reported research on the coprolalia (uncontrollable swearing) that non-English-speaking patients with Tourette syndrome utter in their native tongues. English-speaking Touretters usually utter o ensive words such as shit or pussy or fuck. However, Brazilian researchers (Cardoso, Veado, and de Oliveira 1996) translated Touretter’s outbursts merda and buceta as meaning feces and vagina instead of the more o ensive words shit and pussy. Touretters usually do not utter ino ensive clinical terms when they experience their tics; instead, they say the most socially unacceptable words in their lexicon. These mistranslations misrepresent what the Touretters really said, which is more o ensive than the translations. Problems in translating tabooed words from the source language to the target language were recently outlined by Dominguez (2019). It also has been noted that foreign translation can become more accurate when we have considerable experience with the culture (not just the words) we are studying (see Dewaele 2004, 2010). In such cases, knowledge from a subdiscipline such as anthropological linguistics may help overcome the translation challenge. The subjectivity problem is also related to the problem of using our adult lexica to interpret children’s swearing—words that adults might hear as ino ensive (nerd or baby) are usually deemed to be more o ensive to young children (Jay and Jay 2013). Here again, there is little research to guide us on what children know about taboo words (Jay and Jay 2013). Researchers need to be aware of historical, developmental, and cross-cultural di erences that lead to misinterpretations of what other speakers mean when they swear. The subjectivity problem calls for a deeper understanding of the people we study, as well as using informants and translators that can accurately transcribe/interpret what we hear. Here is one nal example of how interpretations have gone wrong. I was enrolled in a graduate-level course on the topic of “metaphors for love” at the University of Massachusetts. Most of the graduate students and the instructor were non-native English speakers. The instructor was collecting love metaphors in English for a book on the topic that was published several months later. As I studied his metaphor book, I noticed that the analysis did not cover what I call the more “lusty” aspects of how native American English speakers talk about love, sex, and interpersonal relationships—the kind of expressions that in the wrong context might constitute sexual harassment or unwanted sexual comments, such as she liked to sleep around, you are a tramp, we had a quickie, we humped each other, or s/he’s a nice piece of ass (see Jay 2000). The author had reported the a ectionate and ino ensive metaphors for love without digging down to see how native American English speakers talk about sexual relationships using o ensive gurative language. Again, a deeper understanding based on native expertise is needed for appropriate analysis.

p. 199

2. Plans for the Future Can we tie together, theoretically, all the diverse research on taboo words published so far? I have speculated on what the nature of a uni ed theory within the eld of psychology could be (Jay 2000, 2003, 2009), but I am not hopeful. Scholarship and methodology standards vary too much across the many disciplines studying taboo word use and swearing. A uni ed theory will not accrue due to cross-disciplinary di erences in research goals; nonetheless we still need to be aware of our di erences. Most researchers need to develop a better understanding of cross-disciplinary approaches to o ensive speech, even though this knowledge will not produce a comprehensive, universal theory of swearing. Although we will gain a better understanding of how others conduct their research, a grand theory tying together diverse research is not probable. What we gain reading cross-disciplinary swearing research is present in a book by Beers Fägersten and Stapleton (2017), which presents new avenues of research for non-English languages that employ a variety of research methods. Awareness of wide-ranging approaches stimulates more interest and work across disciplines. This avenue for broad interests did not exist fty years ago. When I started studying taboo word literature in the 1960s, I had to scour the research in a range of disciplines: psychology, sociology, criminal justice, communication studies, mass media, religion, law, medicine, linguistics, and neurology. Now, because there is much more research and scholarship within each discipline, a scholar can comfortably nestle within his or her discipline without venturing to read what is published elsewhere. This is a mistake; even though it is di

cult to understand the methods and jargon emerging from another area of study, we

should not be reluctant to examine it.

2.1. Planning to Use the Appropriate Research Methods What are the appropriate methods to study swearing? Disciplines that study swearing, neuroscience, sociolinguistics, anthropology, philosophy, law, medicine, and psychology each champion di erent approaches and methods. Each eld has its tried-and-true methods. Due to di erences in practices across language disciplines, not all research methods are of equal value to a given scholar. Each scholar experiences a form of disciplinary separation in language studies because each scholar’s discipline has its own preferred methods. Ultimately, each scholar must make decisions based on where his or her research will be published, such as selecting the right journal and its readers as a target audience. Relatedly, each journal has its unique scope, purpose, or aim. Scholars must seek journals that best suit their long-term goals. Within any discipline a good working knowledge of past research methods is necessary in order to avoid studying a question that has already been answered, one that may be irrelevant, or try to answer a question p. 200

that cannot be answered (e.g., What was the

rst swearword?). Awareness of previous research reveals

which research methods have been most fruitful. More importantly, each researcher needs to address the question of what she or he is trying to achieve—that is, the ultimate purpose of his or her research.

2.2. Planning to Be Aware: Research in Other Disciplines Scholars who are planning research need to be open-minded, inclusive, and multidisciplinary in scope when they are able to be. Recognize that each language discipline indoctrinates scholars to adhere to its particular conventions, dictating how research is conducted and where it is published. Each researcher, because of these cross-disciplinary di erences in training approaches swearing from a di erent angle. A communication studies scholar might construct imaginary scenarios that use taboo words. These constructed scenarios might not include actual public utterances that a psychologist wants to report.

There is danger in being ignorant of the methods of other disciplines. We might venture into troubled waters if we try to use methods that are outside our own training. When initially starting to conduct research, it is best to stick with the most trusted methods, and not try those outside our realm of expertise. Ultimately, we will accumulate a broader expertise as we gain more experience, and then we can experiment with novel methodology. A thorough knowledge of past research is helpful for planning new research. Before collecting new data, we need to be aware of the swearing data that have already been collected in databases of speech, spoken dialogues, or corpus linguistics data. We need to decide if previous taboo word research is relevant for our research. Some data might be too old or too restricted; it might sample speakers who are older or younger, more religious or less religious, or more educated or less educated than our target population. Researchers who study the frequency or the o ensiveness of language samples should plan to collect their own normative data from their own participant pool, because data from a di erent population will not accurately re ect their participants’ perceptions and attitudes. In other words, do not use normative data from a sample that is not representative of your own speakers. Frequency researchers should report the percentage of a sample (e.g., .5%) that includes taboo words or the percentage of an informant’s daily speech that includes taboo words. Reporting raw frequencies (e.g., “there were 6 taboo words”) is not as meaningful as reporting what portion of the sample they represent (e.g., 6/1200). We need to be careful in framing the scope and meaning of our work. Swearing as de ned earlier in face-toface contexts is a spoken behavior that occurs when people become emotional. It would not be accurate to describe ctional, invented dialogue in motion pictures, novels, or Shakespeare’s plays as being the same as natural (real world) examples of swearing. Although these kinds of ctional narratives do employ taboo words, such ctional data are not equivalent to data derived from natural language corpora.

p. 201

2.3. Planning with Corpus Linguistics and Social Network Research Corpus linguistics is the study of language using real-world language texts or corpora, which can be searched for words or phrases of interest via computer. An example of the value of corpus linguistics to taboo word scholars is the work of Tony McEnery. McEnery (2006, chap. 2) examined the use of bad language words (BLWs) derived from the Lancaster Corpus of Abuse (LCA) based on the British National Corpus (BNC), a 100,000,000-word corpus of present-day British English. Importantly for taboo word scholars, McEnery, with an awareness of the criticality of reporting contextual variables, reported BLW use as a function of the speaker’s age, gender, and social class, as well as the target (listener) of the insult and the o ensiveness of the word used. These are the kinds of sociopragmatic variables that engender taboo word use and should be reported in future research. A recent examination of corpus linguistics research in non-English-language samples can be found in Beers Fägersten and Stapleton (2017). Corpus linguistics research can be related to studies using samples from social network speech, both deriving data from social communications. Research on taboo word use in online communities has been evolving over the last decade or so (Jay 2018). For example, Subrahmanyam, Smahel, and Green eld (2006) reported that swearing in online chat rooms occurred in 3% of online conversations, a rate of about one swearword every two minutes. By comparison, studies of spoken English indicated that swearwords occur at a rate of .3% to .7% of all the words spoken in the samples studied (Jay 2009a). In a study of swearing rates in online communication in the United States and the United Kingdom, Thelwall (2008) reported a .2% swearword rate in his UK MySpace study. Most of the young UK users’ (16- to 19-year-olds) MySpace communication contained swearing compared to roughly only one-half of middle-aged UK users’ communications. Thelwall found no gender di erences in swearing in his UK sample but noted that in a US

sample, US males’ MySpace communications contained signi cantly more strong swearing than did females’ MySpace communications. To delineate gendered uses of swearwords on Twitter, Gauthier et al. (2015/2016) examined data gathered from a corpus of one million tweets. They reduced an original set of 788 possible swearwords to a set of 26 swearwords used in their Twitter corpus; this smaller set accounted for 90% of all swearwords in the corpus. Most of the tweets, based on recorded user data characteristics, were from 12- to 30-year-old users. In their corpus, 5.8% of the men’s tweets contained at least one swearword compared to 4.8% of the women’s tweets. Women used the words bitch, bloody, and hell signi cantly more frequently than men; men used cunt, tit, and fuck signi cantly more frequently than women. Gauthier et al. noted that time of day was an important variable for swearword use on Twitter. People started swearing on Twitter between 6 a.m. and 7 a.m., and the rate gradually increased through the day; there was a lull around 7 p.m. to 8 p.m. after which the rate dramatically increased. These types of corpus linguistics and social network studies will be increasingly explored in the future.

p. 202

2.4. Planning to Account for the Role of Context The contextual nature of swearing becomes obvious to the experienced researcher: the comprehension and production of taboo expressions is entirely context dependent. Change the location, social function, or relationship between the speaker and listener, and it will change the meaning of what is said. The task going forward is to account for the contextual variables that a ect how taboo words are used, why they are spoken and how they are understood, as well as, in pragmatic terms, also what is presupposed or implied. We have demonstrated people’s awareness of contextual variables by asking participants to rate the o ensiveness or the likelihood of di erent combinations of speakers, words, and locations, for example, the janitor said hell in the parking lot versus the student said shit in the dean’s o ce (Jay 1992). We (Jay and Janschewitz 2008) also demonstrated that native speakers of English are more sensitive to contextual changes in swearing than non-natives and that it takes some time to become familiar with the sociocontextual variables that a ect swearing. A non-native speaker or language learner might know that shit is a taboo word but not be sensitive to where it can (locker room) or cannot be used (dinner with grandparents). Dewaele (2010) has published extensively on L1 versus LX language di erences with respect to swearing and the use of taboo words. He has demonstrated, for instance, that non-natives acquire the nuances of swearing better through direct experience in the new language community than through classroom-only exercises. The concept of “native speakers” here is based on monolingual speakers or speakers whose mother tongue is English and who, at maximum, acquired some knowledge of other languages in school or their professional career. Thus, it is a Euro-American perspective, not necessarily one that focuses on a multilingual Nigerian or Singaporean who may also be considered a “native” speaker of English elsewhere. Researchers should expect to nd di erences based on age, gender, race, and occupation as a function of a language community’s values: some communities will be more liberal than others, and one will hear swearing there. Swearing is a function of community values. Swearing is also part of masculinity or the ritual performance of power associated not so much with liberality but with aspects of gender roles. Di erent dialects of the same language will also di er according to geographic location and socialeconomic statuses. For those interested in geographical di erences in spoken American English a valuable 1

resource is the Dictionary of American Regional English ([DARE] 1985). As a native of the state of Ohio, I found that insults I had heard growing up such as brier hopper (similar to hillbilly referring to a rural mountain dweller) were only used and known in southwestern Ohio and not elsewhere in the United States. DARE documents these kinds of regional di erences in English language word use. One of the important lessons about context is that these geographical and intrapersonal di erences are critical to understanding how and

p. 203

why people say taboo words (with friends) or do not say taboo words (with strangers) in di erent contexts. I would not call someone a brier hopper where I live now in Massachusetts because people would not understand what it meant. The plan is to be sensitive to and to report contextual nuances that a ect swearing. Scholars, therefore, must not collapse or gloss over these important contextual di erences in order to report “averages.” Answering questions such as “What does the average person say?” is not informative from a contextual point of view because the answer to that kind of question glosses over the signi cance of contextual and individual di erences.

2.5. Planning to Demonstrate How the Taboo Lexicon Evolves from Childhood to Adulthood Although those who study language understand that vocabulary changes dramatically from infancy to adulthood, little research has been dedicated to showing how the swearing lexicon is acquired in the context of a child’s evolving social, emotional, and psychological awareness. Jay Jay and Jay’s (2013) research has examined gender, age, and historical di erences in taboo word use, outlining how boys’ and girls’ lexica changes as they mature. Children start swearing as soon as they speak. The extent of the taboo lexicon of the three- to four-year olds is impressive; boys and girls at these ages are learning name-calling and psychosocial insults, abusive language, common profanities, scatological language, and gender-related insults. The growth in the taboo lexicon starts to level o

at about the time children enter elementary school. Our analysis of child swearing

data by age indicated that, with age, the child swearing lexicon evolves to become more adultlike. The youngest age groups tend to share few taboo words with adults and do not show signi cant correlations with adults based on word frequency. Overall, children showed gender di erences in word frequency like those found in adults (men swear more than women in public), but the lexica of boys and girls were more di erent than those of adult men and women, as evidenced by the many points of divergence in boys’ and girls’ vocabularies over time. Gender di erences in word frequency became most obvious at older ages (generally over ve years), suggesting that the time of the transition to school is when adultlike gendered habits underlying emotional expression start to become salient. Historically speaking, the most frequently used words by adults and children have remained stable with some minor changes emerging over the last forty years. Our (Jay and Jay 2013) study reported similarities and di erences in the data that we collected during the 1980s compared to the taboo word use data we recorded thirty years later. There was some di erentiation over time in the content of children’s taboo lexica. For taboo words that were recorded during both time periods, the frequency relationship can be p. 204

described by correlation which was signi cant and positive. The high correlation

indicated that

children’s high-frequency taboo utterances (e.g., shit) in the 1980s also tend to be high-frequency taboo utterances in the 2013 sample. Although people believe that taboo vocabularies shift dramatically over time, stability is more the rule according to these data. There are very few, if any, developmental studies like this one. Obviously, more work needs to be done to show how children acquire and expand the taboo lexicon, especially in non-English societies and from di erent income and education groups. One nal suggestion for future research: if scholars are working with non-English participants/speakers, the di erences between boys’ and girls’ taboo word use needs to be addressed, as well as how children’s taboo word use changes as they mature.

2.6. Planning to Demonstrate How the Taboo Lexicon Evolves Historically The words in our taboo lexica change over time. The potency of some o ensive words also declines over time. Profane epithets such as Jesus Christ! and Goddamn! once forbidden, are now common. The epithet sucks, once referring to oral sex in the 1960s in the United States, now has lost its original taboo association. The spoken lexicon is organic; it changes and evolves over time. Some words become obsolete and are dropped from usage; other words, especially slang terms, slurs, and insults are coined anew and enter the spoken lexicon. By the time we reach adulthood, it becomes obvious to us that words have changed their meaning or emotional potency over time. Older insults can lose their power to o end. This means that scholars need to be aware of the evolution of the taboo lexicon over time—the birth and death of taboo words. Some frequent words (fuck, shit) have appeared consistently over centuries; others (lickspittle, milksop) have not. Some changes in taboo language re ect general cultural value trends, for example, cultural shifts from liberal to more conservative values, and vice versa. Cultural shifts occurred amid repressive trends during the Victorian era in Europe and in the 1920s during Prohibition in the United States. Language restrictions align with such shifts such that words that were once acceptable become tabooed. These forms of changes in o ensive and inappropriate language have taken in American history have been documented (Jay 2017). Obscene, indecent, and insulting epithets that were banned in the early 1900s are now commonplace in popular media (Jay 1992, 2017). One can expect to see the shifts in speech etiquette like these over time by looking at media standards, for example, by looking at what o ensive words were acceptable in motion pictures or television content over time. Because of changes in a culture’s sociopolitical atmosphere and media standards, our sensitivity to and use of taboo words and insults align with trends that occurred during a conservative trend or liberal trend when each of us grew up (e.g., 1960s in the United States). For p. 205

example, we (Jay 2000) have recorded how elderly men

and women in nursing homes bring to that

setting the o ensive speech learned during their childhood. We reported 90-year-olds frequently using words like, hell, damn, or Jesus Christ. These, however, are not the words that millennials will use when they turn 90. Millennials are more likely to say fuck and motherfucker than their grandparents. The millennials in their old age will use words that were o ensive (but frequent) in the twenty- rst century.

2.7. Planning to Account for the Physiological and Emotional Correlates of Swearing How the human brain is involved in swearing is a primary interest of psychologists and neuroscientists. Neuroscience issues have been of lesser concern for those conducting research in the humanities, but even so, accounting for the biological genesis of taboo speech is important for understanding the entire person. To completely address emotional expressions requires knowledge of the physiological correlates underlying emotion—to understand how a person generates an expression using taboo words from a pre-verbal verbal feeling (usually anger or frustration). Physiological knowledge is also important to understand neurological speech disorders, for example, people who swear uncontrollably or cannot swear at all. Whether you can use neuroscience in your scholarship or not, you will be enlightened by a study of the subcortical brain areas involved in swearing (limbic system, amygdala, basal ganglia) in order to understand how swearing is a ected by brain damage, stroke, Tourette syndrome or prescription medications. Neuroscience literature can provide a better understanding of personality di erences in swearing that are the product of factors such as religiosity, sexual anxiety/guilt, Type A personality traits (characterized by hostility, urgency, and restlessness), extroversion, and agreeableness. These signi cant psychological aspects of personality are related to taboo word use, because when looking at taboo language in public or in written accounts, it is always important to ask, What emotion is being expressed?

We can witness emotion at work in public and private places because we have learned how to perceive emotion in others from an early age. However, we perform poorly at accurately perceiving emotions in CMC and online communications. Our inability to detect what a texter or Facebook friend means to express emotionally when writing has led to serious misunderstandings in online discourse (Jay 2018). One nal neuroscience issue involves how multilingual speakers experience emotions in di erent languages. As a native English speaker, I do have an emotion reaction when I say “shit!” but I do not feel the same emotion when I say “merde” or “scheiße.” I do not feel much of anything. Dewaele (2004, 2010) has addressed these kinds of multilingual feelings to indicate how native speakers feel more emotion in their native languages than they do in non-native languages. Scholars who swear in more than one language know what this means.

p. 206

2.8. Planning for the Campus “Politics” Surrounding Taboo Word Research This section addresses community perceptions and attitudes that arise because of the words we study. Academics work in communities of scholars and administrators. Some communities do not tolerate the use of taboo speech. I work in a psychology department with a handful of colleagues who are aware of my research. I doubt any of them read what I publish in psychology journals. Colleagues outside my department have limited exposure to what I publish. Academics who study swearwords are subject to scrutiny and evaluation by administrators, who, like our colleagues, may have little understanding of the nature of taboo word research; they may have little admiration for what we do. Taboo word research merits department advancement and tenure, although some colleagues and administrators cannot accept it as a legitimate eld of study. We study the language that many community members explicitly want to prohibit; yet those kinds of attitudes should not restrict our work on campus. Consider that we study taboo topics but not the deepest, darkest forms of psychological secrets such as incest, murder, suicide, rape, depression, or addiction that lead victims to seek the help of mental health workers. Psychiatrists work with patients who have experienced traumas that are di

cult to put into words; the patients’ problems are in a sense unspeakable,

maybe even too dark to bring into consciousness. We swearword scholars study the speakable taboos; words that are not supposed to be said in public but are said in public. Taboo words are not unspeakable; they are spoken publicly whether our overly sensitive colleagues or those with conservative moral perspectives like it or not. With this apparent academic dilemma, studying what should not be said, social and political consequences will arise. Studying taboos will always be controversial because the subject of the research is inherently taboo too. We in a sense become contaminated by the words we study. Having faced these issues for four decades, I urge scholars to persist despite the negative reactions of colleagues, administrators, and editors. Whether we like it or not, taboo word research involves political issues (see Jay 2000). What we study is important, and it is associated with the social and political problems of our times, such as sexual harassment, discrimination, hate speech, workplace standards, obscenity, and more. These are issues that bother a lot of people—our colleagues included. Moreover, it is also our responsibility to debunk myths and misrepresentations of taboo speech and swearing behaviors—that fuck is an ancient acronym or that swearing is abnormal, the sign of an impoverished vocabulary, or only a problem for undereducated children and teenagers. Our research can provide for all who read it a better understanding of how humans communicate their emotions in di erent cultures and social situations. We will escape the fact that what we study is somewhat dangerous or at least o ensive to other people. In the end, what we are studying is now a legitimate area of scholarly interest in the context of the important social problems just mentioned. Swearing is a damned fascinating behavior, but we scholars should not be condemned for studying it.

p. 207

Note 1.

There are 6 volumes of DARE published from 1985 to 2013. There is also a DARE Newsletter and Digital DARE.

References Adams, Michael. 2016. In Praise of Profanity. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Bergen, Benjamin. 2016. What the F: What Swearing Reveals about Our Language, Our Brains, and Ourselves. New York: Basic Books. Google Scholar Google Preview WorldCat COPAC Cardoso, Francisco, Claudio Veado, and Jose Teotonio de Oliveira. 1996. “A Brazilian Cohort of Patients with Touretteʼs Syndrome.” Journal of Neurology, Neurosurgery, and Psychiatry 60:209–212. Google Scholar WorldCat Chomsky, Noam. 1957. Syntactic Structures. The Hague, Netherlands: Mouton & Co. Google Scholar Google Preview WorldCat COPAC Clark, Herbert. 1996. Using Language. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Croom, Adam. 2014. “The Semantics of Slurs: A Refutation of Pure Expressivism.” Language Sciences 41:227–242. Google Scholar WorldCat Dewaele, Jean-Marc. 2004. “The Emotional Force of Swearwords and Taboo Words in the Speech of Multilinguals.” Journal of Multilingual and Multicultural Development 25:204–222. Google Scholar WorldCat Dewaele, Jean-Marc. 2010. Emotions in Multiple Languages. Basingstoke, UK: Palgrave Macmillan. Google Scholar Google Preview WorldCat COPAC Dictionary of American Regional English. 1985. Vol. 1: A–C, 380. Edited by Frederic G. Cassidy. Cambridge, MA: The Belknap Press of Harvard University Press. Dominquez, Pedro J. Chamizo. 2019. “Problems Translating Tabooed Words from Source to Target Language.” In The Oxford Handbook of Taboo Words and Language, edited by Keith Allan, 199–217. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Beers Fägersten, Kristy, and Karyn Stapleton, eds. 2017. Advances in Swearing Research: New Contexts and New Languages. Amsterdam: John Benjamins. Google Scholar Google Preview WorldCat COPAC Gauthier, Michael, Adrien Guille, Fabien Rico, and Anthony Deseille. 2015/2016. “Text Mining and Twitter to Analyze British Swearing Habits.” In Twitter for Research Handbook, edited by Clement Levallois, Morgane Marchand, Tiago Mata and Aandre Panisson, 27–43. Paris: Emlyon Press. Google Scholar Google Preview WorldCat COPAC Jay, Kristin, and Timothy B. Jay. 2013. “A Childʼs Garden of Curses: A Gender, Historical, and Age-Related Evaluation of the Taboo Lexicon.” American Journal of Psychology 126:459–475. Google Scholar WorldCat Jay, Timothy B. 1977. “Doing Research with Dirty Words.” Maledicta: The International Journal of Verbal Aggression 1 (2): 234– 256. Google Scholar WorldCat Jay, Timothy B. 1992. Cursing in America. Philadelphia, PA: John Benjamins.

Google Scholar

Google Preview

WorldCat

COPAC

Jay, Timothy B. 2000. Why We Curse: A Neuro-Psycho-Social Theory of Speech. Philadelphia, PA: John Benjamins. Google Scholar Google Preview WorldCat COPAC Jay, Timothy B. 2003. The Psychology of Language. Upper Saddle River, NJ: Pearson. Google Scholar Google Preview WorldCat COPAC Jay, Timothy B. 2009. “The Utility and Ubiquity of Taboo Words.” Perspectives in Psychological Science 4:153–161. Google Scholar WorldCat Jay, Timothy B. 2017. We Did What? O ensive and Inappropriate Behavior in American History. Santa Barbara, CA: ABC-Clio. Google Scholar Google Preview WorldCat COPAC Jay, Timothy B. 2018. “Swearing, Moral Order, and Online Communication.” Journal of Language Aggression and Conflict 6 (1): 107–126. https://doi.org/10.1075/jlac.00005.jay. Google Scholar WorldCat p. 208

Jay, Timothy B., and Kristin Janschewitz. 2008. “The Pragmatics of Swearing.” Journal of Politeness Research: Language, Behavior, Culture 4 (2): 267–288. Google Scholar WorldCat McEnery, Tony. 2006. Swearing in English: Bad Language, Purity and Power from 1586 to the Present. New York: Routledge. Google Scholar Google Preview WorldCat COPAC Rosten, Leo. 1968. The Joys of Yiddish. New York: McGraw-Hill. Google Scholar Google Preview WorldCat COPAC Subrahmanyam, Kaveri, David Smahel, and Patricia Greenfield. 2006. “Connecting Developmental Constructions to the Internet: Identity Presentation and Sexual Exploration in Online Teen Chat Rooms.” Developmental Psychology 43: 395–406. https://doi.org/10.1037/0012-1649.42.3.395 Google Scholar WorldCat Thelwall, Mike. 2008. “Fk Yea I Swear: Cursing and Gender in a Corpus of MySpace Pages.” Corpora 3(1): 83–107. https://doi.org/10.3366/E1749503208000087. Google Scholar WorldCat

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published online: 22 May 2024 Published in print: 02 May 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

p. 467

21 The Contents of Maps  Je rey C. King https://doi.org/10.1093/oxfordhb/9780192844118.013.45 Published: 22 May 2024

Pages 467–488

Abstract It seems clear that maps have contents. Maps represent the world as being a certain way, and do so accurately or inaccurately. Presumably, they do this in virtue of having contents. Further, it seems that we can have attitudes toward the contents of maps. Surely, we can believe or doubt a map. A quick internet search revealed the following examples of people talking about believing maps: “Get out the compass, believe the map and work out where you really are”; “Corsica: Don’t Believe the Maps!”; “Episode 164: Don’t Believe Your Map” (History of the Roman empire podcast); “Turkey vs Russia: Which Map Do You Believe?”; “Which do you believe, the map or the GPS?”; and “Our inclination is to always believe the map.” Presumably, to believe or doubt a map is to have an attitude toward the content of the map. Further, people often talk about maps saying things or telling us things. Again, an internet search yielded the following examples: “This map is a total lie” (said about a Mercator projection); “The map says we’re fucked” (from Freddie’s Dead: The Final Nightmare); “We expect maps to tell us the truth”; and “We are inclined to trust maps and feel they are telling us the truth.” Again, presumably maps say things in virtue of having contents. Given that maps have contents, we immediately confront the question: What kinds of things are the contents of maps? The chapter attempts to answer that question.

Keywords: content, proposition, representation, map, structured proposition, structured content, unstructured content, truth condition Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

It seems clear that maps have contents. Maps represent the world as being a certain way, and do so accurately or inaccurately. Presumably, they do this in virtue of having contents. Further, it seems as though we can have attitudes toward the contents of maps. Surely, we can believe or doubt a map. A quick internet

search revealed the following examples of people talking about believing maps: “Get out the compass, believe the map and work out where you really are”; “Corsica: Don’t Believe the Maps!”; “Episode 164: Don’t Believe Your Map” (History of the Roman empire podcast); “Turkey vs Russia: Which Map Do You Believe?”; “Which do you believe, the map or the GPS?”; and “Our inclination is to always believe the map.” 1

Presumably, to believe or doubt a map is to have an attitude toward the content of the map. Further, people often talk of maps saying things or telling us things. Again, an internet search yielded the following examples: “This map is a total lie” (said about a Mercator projection); “The map says we’re fucked” (from Freddie’s Dead: The Final Nightmare); “We expect maps to tell us the truth”; and “We are inclined to trust maps and feel they are telling us the truth.” Again, presumably maps say things in virtue of having contents. Given that maps have contents, we immediately confront the question: What kinds of things are the contents of maps? The present chapter attempts to answer that question. As a starting point, let’s consider the well-known semantics for maps that was formulated by Casati and Varzi (1999; henceforth, C&V). C&V formulate a semantics for simple maps that they call formal maps. They assume that formal maps are self2

connected square regions of space. Formal maps have no names or symbols, etc. They only have colored regions. Map subregions can be any shape or size and need not be self-connected. Color patches are taken to designate properties, and that a certain region of a map is a certain color represents the region of the world corresponding to that map region as having the property corresponding to the color. A uniformly colored map region is correctly colored when the corresponding world region has the property corresponding to the color. A non-uniformly colored map region is correctly colored when the uniformly colored regions p. 468

constituting it are correctly colored. When a map

is correctly colored, C&V say it is true. A stage of a given

colored map μ is any coloring of some regions of μ, where a region is colored F in that stage only if it is so colored in μ. A well-formed stage of a map μ is a stage σ of a map μ satisfying the following conditions: 1. All colored regions of σ are the same color. 2. If a region is that color in μ, it is that color in σ. A map is built out of map stages as follows: We begin with the empty (blank) map. Then we have the set of atomic, well-formed map stages of one color. Then we have the set of map stages with two colors, and so on, until we reach a nal stage with all the colors of the map. The map is this nal stage. C&V introduce some predicates that play a role in their semantics. First, there is the mereological notion of part for which they introduce the binary predicate ‘P’. ‘Pxy’ is to be read ‘x is part of y’. C&V explicitly allow that the relata here can be things that are as di erent as material bodies, events, geographical regions, and 3

sets of numbers. Indeed, they write that “there is no ontological restriction on the eld of ‘part’.” For our purposes we will be concerned with cases where the relata are spatial regions either of maps or of the world. C&V assume that ‘P’ will be governed by at least the following axioms: P1. Pxx (Re exivity) P2. Pxy & Pyx→ x=y (Antisymmetry) P3. Pxy & Pyz→Pxz (Transitivity) They call the resulting theory Ground Mereology, or M for short. They assume that any part/whole theory will have M as its basis. They consider various extensions of M, but for our purposes, we can consider the following. Add to P1-P3 the following: P5. ~Pyx→∃z(Pzy & ~Ozx) (Strong Supplementation), where ‘O’ is overlap de ned from ‘P’ as follows: Oxy = def ∃z(Pzx ^ Pzy)

The theory consisting of the axioms P1-P3 and P5, C&V call Extensional Mereology (EM). Finally, C&V want to add some sort of fusion principle to EM. They tend to favor the following: P8. ∃xϕ → ∃z∀y(Oyz ↔ ∃x(ϕ ^ Oyx)) (Fusion),

where ϕ is any formula of the language. The theory consisting of adding P8 to EM, C&V call General Extensional Mereology (GEM) and it is one of the mereological theories they clearly are fond of. p. 469

C&V wish to consider some topological extension of GEM by adding a predicate ‘C’ expressing the binary connection relation. ‘Cxy’ is read ‘x is connected to y’. C&V consider the result of adding the following axioms for ‘C’ to GEM: C1. Cxx (Re exivity) C2. Cxy→Cyx (Symmetry) C3. Pxy→Exy (Monotonicity) where ‘E’ (“enclosed”) is de ned as follows: Exy = def ∀z (Czx→Czy)

C&V call the theory resulting from adding to GEM C1-C3 GEMT. This is one of the mereotopological (mereology + topology) theories of interest to C&V in their chapter 4. They actually seem to favor a theory they call GEMTC that features an additional closure condition not belonging to GEMT. But they have no problems with the simpler GEMT, and so we’ll stick with it. Now, what does all this have to do with the semantics of maps? For C&V, a model for a formal map μ is a structure , where R is a set of world regions and f is an interpretation function such that: 1. For any map region μi of μ, f(μi) is a world region in R; and 2. For any map regions μi and μj of μ: a. Pμiμj i

Pf(μi)f(μj); and

b. Cμiμj i

Cf(μi)f(μj)

Finally, in an intended model, for any color F, f(F) is the set of world regions in R having the property corresponding to F. In saying all this, C&V assume that ‘P’ and ‘C’ are governed by any one of the 4

mereotopological theories they discuss in their chapter 4, including GEMT. Note that clause 1 insures that when a map region is assigned a world region, subregions of the map region are assigned to subregions of the world region. And clause 2 insures that the parthood facts and connection relations of map regions of a map mirror those of the world regions they are assigned to. Now let σ be a well-formed atomic map stage of map μ. That means σ is a map stage of μ whose colored regions are all the same color, say F. Let σ1 be the mereological fusion of all colored regions of σ and let σ2 be σ-σ1. Then σ is true relative to i i

f(σ1) ∈ f(F) and f(σ2) ∉f(F). Finally, we can say that a map stage is true

all its well-formed atomic maps stages are true. Note that this clause applies to the map itself, since it is

just the maximal map stage. For future reference, also note that this says that a map stage is true i

its well-

formed atomic map stages are conjointly true.

p. 470

The condition that f(σ2) ∉f(F) in C&V’s de nition of truth for a well-formed atomic map stage σ is intended to capture what Rescorla (2009) dubs the absence

intuition: roughly, that a feature (in our case, a color)

representing a property appears on a map but is absent from a region r of the map represents that the region

of the world corresponding to r fails to possess the property corresponding to the feature. In the present case, the region σ2 of σ fails to be the color F, and so the world region corresponding to σ2 (f(σ2)) must lack 5

the property corresponding to F for σ to be true relative to . Since f maps F to world regions that possess the property corresponding to F, for σ to be true relative to it had better be that f(σ2) ∉f(F).

Of course, to this point all we have done is assigned truth conditions to maps. We can turn this semantics into a theory of unstructured map contents as follows: We are going to de ne what it is for a map stage μ to be true relative to a world w and interpretation function f. First, let’s assume our interpretation functions map regions in formal maps to world regions common to all worlds under consideration. This means that for a region r of map m, we can talk of the world region f(r) without specifying which world it is in. We assume that a given world region f(r) can have di erent properties in di erent worlds. Finally, assume that interpretation functions map colors to the properties they represent rather than the world regions possessing those properties, as on C&V’s approach. A map stage μ is true relative to an interpretation function f at a world w i

(i) f is an interpretation function such that:

1. For any map region μi of μ, f(μi) is a world region; and 2. For any map regions μi and μj of μ: a. Pμiμj i

Pf(μi)f(μj) in w; and

b. Cμiμj i

Cf(μi)f(μj) in w;

and (ii) for every atomic map stage σ of μ, f(σ1) possesses f(F) in w and f(σ2) fails to possess f(F) in w, where 6

σ1 is the mereological fusion of all colored regions of σ, σ2 is σ-σ1 and F is the color of σ’s colored regions. Finally, the unstructured content of a map stage μ relative to an interpretation function f={w| μ is true

relative to f at w}. Of course, in usual cases, a given map will have an intended interpretation function and it is natural to think of the map’s unstructured content as its content relative to that function. When I talk of a map’s content simpliciter, I am talking of its content relative to the intended interpretation function. We have now managed to assign unstructured contents to formal maps. However, as an advocate of the view that the contents of natural language sentences are structured, I believe that many of the considerations that favor that view also favor the view that the contents of ordinary maps are structured. But then, insofar as we intend our assignment of contents to formal maps to be a simple model of an assignment of contents to ordinary maps, we should assign structured contents to formal maps. Let’s look at reasons for thinking the contents of ordinary maps are structured. First, it seems possible for one to believe a map A and fail to believe a map B, where an unstructured account of content is bound to assign A and B the same content relative to the intended interpretation functions. Perhaps the clearest case of this would be the case in which A and B are both accurate maps of the world employing di erent projections. Take the following Mercator and Robinson projections: p. 471

Presumably, an unstructured account of content would assign both maps the same contents relative to their intended interpretation functions. Every world accurately described by one map is accurately described by the other relative to their intended interpretations, and vice versa. Yet it seems that a person who understood the idea that there are di erent map projections could nonetheless think that the rst map accurately represents the world given its interpretation, and that the second doesn’t. Suppose, for example, p. 472

that subjects are told they will be shown maps of the world, some of which are

di erent projections. They

are told that some maps accurately represent the world, and some don’t. They are told to write down which maps they believe and which they don’t. It seems clearly possible that a person could believe the rst map projection and not believe the second one, or vice versa, even though both maps have the same unstructured content. But, assuming that believing a map or not is bearing a relation to its content, these maps had better have di erent contents. This looks to be an argument that map content is structured assuming that the structured account does indeed assign di erent contents to these maps. Of course, some advocates of the view that natural language sentences express propositions that are sets of worlds, most notably, Stalnaker (1984, 1987), have tried to respond to this kind of argument against unstructured content. In King (forthcoming) I argue that Stalnaker’s response fails. The second problem with assigning maps unstructured contents concerns the fact that map contents have truth/accuracy conditions in virtue of representing the world as being a certain way. To put this another way, map contents have truth values at possible worlds. Having truth values at possible worlds— representing the world as being a certain way—is a very unusual property, and we should expect there to be

some explanation of how map contents manage to do this. Sentences, minds, and perceptual experiences all represent the world as being a certain way, and we expect there to be an explanation forthcoming as to how they manage to do this. I think we should say the same thing about map contents. We need an explanation of how they manage to represent the world as being a certain way and so have truth conditions. The problem is that taking map contents to be sets of possible worlds leaves us without any explanation of how map contents manage to be true or false at possible worlds and so represent the worlds as being a certain way. Why would a set of worlds be true or false at a possible world? What is it about the set that would make it do this? Sets of cities, states, countries, planets, solar systems, galaxies, and galaxy superclusters don’t represent anything. Why do sets of worlds? There just is nothing about a set of worlds that would explain why it takes truth values at worlds. So the unstructured-content account of the contents of maps cannot explain why such contents have truth conditions and so represent the world as being a certain way. But as we’ve seen, we need such an explanation. Further, the unstructured account of map contents cannot explain why map contents have the speci c truth conditions they do. Consider some map’s content C qua set of worlds and suppose it contains w1. Is C true or false at w1 and what makes it so? Why would w1 being in C make C true at w1 as opposed to making it false at w1? There just doesn’t seem to be any answer to this question. So the unstructured-content account of maps’ contents cannot explain why such contents take truth values at worlds and so represent the world as being a certain way; nor can it explain why an unstructured map content is true at the worlds it is true at and false at the worlds it is false at. The nal problem I’ll discuss with taking maps’ contents to be unstructured is that it makes belief closed under logical consequence when it doesn’t appear to be. To see why, following Stalnaker (1984), let’s understand what it is to believe a proposition in terms of a belief state: the set of worlds compatible with what one believes; the set of worlds that for all one believes might be actual. To believe that P is for P to be p. 473

true in all the worlds in one’s belief state. Now, suppose one believes that P, one believes that Q, and P and Q entail R. Then P and Q are true in all worlds in one’s belief state. But then, so is R. Hence, one believes R. Another way into the problem is discussed by Stalnaker (1999): From their beginnings, epistemic and doxastic logics—the logics of knowledge and belief have been based on modal logic—the logic of necessity and possibility. Knowledge and belief in such logics are analogous to necessity. There is a wide variety of modal logics but all of the normal ones contain certain distribution or deductive closure principles; for example if ‘ϕ→ψ’ is valid so is ‘☐ ϕ →☐ ψ’. Most versions of epistemic logic are normal in this sense, accepting analogous principles for knowledge and belief. Developers of such logics invariably remark that the principles of deductive closure are unrealistic since it is obviously false that knowers in general know all the 7

deductive consequences of anything they know.

Now, given that belief is closed under logical consequence on an unstructured account of content and that map contents are unstructured, suppose that one is presented with the following map of the United States:

United States Map One is asked to examine the map and decide whether it is accurate. One is then asked whether one believes the map to be accurate, whether one believes the map. Suppose one does. So now, one bears the belief relation to the content of the map. But the content of the map entails that Reno is west of Los Angeles. So if one believes the map, one must believe that Reno is west of Los Angeles. But it seems clear that one might p. 474

believe that the map is accurate—believe the map’s content—without believing that

Reno is west of Los

Angeles. Indeed, this seems to happen all the time: many people believe that an accurate map of the United States is accurate and so believe the map without believing that Reno is west of Los Angeles, or many other things entailed by the map’s content. It is worth noting that one prominent attempt to lessen the bite of the fact that belief is closed under logical consequence given a theory of unstructured content won’t apply here. Stalnaker (1984) thinks that an agent can be in multiple belief states that are compartmentalized or fragmented and so not integrated with each other. Any single belief state will be closed under logical consequence for the reason given above. However, if an agent A is in belief state B1 and believes R relative to it, and A is also in belief state B2 and believes S relative to it, and B1 and B2 are not integrated, there may be no belief state such that A believes R&S relative to it. In such a case, A believes R and A believes S, but A does not believe R&S. In the present case, the agent believes the map’s content M relative to some belief state B, and since M entails that Reno is west of Los Angeles, that agent believes that Reno is west of Los Angeles relative to B as well. So appeals to fragmentation of belief states doesn’t help with the problem here. Here again, then, we have a motivation for assigning structured contents to maps. To be clear, I am not saying that advocates of unstructured content don’t have various things to say in 8

response to the concerns I’ve raised. I am simply pointing out that many of the considerations that motivate structured accounts of contents for sentences of natural language also motivate a structured account of map contents. So let’s try to formulate such an account. It won’t be surprising to hear that the account I am going formulate is in the spirit of the account of the structured contents of natural language 9

sentences defended in King (2007, 2019). The resulting account of structured map contents shares important virtues of the account of structured contents of natural language sentences defended in the latter works. In particular, it has the following four virtues: 1. It says what the constituents of map contents are.

2 It says what it is that structures constituents to yield the structured complexes that are map contents. 3 It shows how/why map contents so construed have truth conditions/represent the world as being a certain way. 4. It yields a notion of map contents that are individuated in the proper way. I take it that by doing 2 and 3, the account of map contents addresses the problem that might be called the 10

problem of the unity of map contents on analogy with the problem of the unity of the proposition.

For in doing

2, the account explains how the constituents are combined in a map content such that the resulting combination is something distinct from a mere collection of its constituents. And in doing 3, we explain how a complex structured entity manages to represent the world as being a certain way. Though I won’t argue the matter here, it seems to me that Russell had both these questions before his mind, perhaps not clearly distinguishing them, in his famous discussion of the unity of the proposition in The Principles of 11

Mathematics. p. 475

As with my account of the structured contents of natural language sentences, a radical

feature of my account of

map contents is that they have truth conditions because of something map users

do to endow them with truth conditions. This will come out in our discussion of 3. In giving an account of structured contents for maps, we’ll use some of the resources of C&V’s semantics for maps discussed above, as we did in assigning unstructured contents to maps. Further, again, as before, we’ll follow C&V in giving our account for their formal maps. It is natural to think of C&V’s formal maps as belonging to a map system M that speci es the allowed construction of formal maps. Consider a formal map m that belongs to the system of maps M. Let f be an interpretation function from map regions in m to world regions and from colors to properties. As before, we assume that the world regions that f maps map regions to are common to all worlds under consideration, and that world regions can have di erent properties at di erent worlds. Let r be an atomic colored region of map m i

r is a region of m that is uniform in color,

where no region of m not in r is that color. If F is the color of r, this means that no region of m other than r is F: r exhausts the F region of m. We are generally going to be interested in atomic colored regions of m that are subregions of m. Recall that map subregions, unlike maps themselves, can be any shape and need not be self-connected. As indicated, the atomic colored region of a given map m that is the color F will be exactly the region of m that is F. If that region is a subregion s of m, it can be any shape and may fail to be selfconnected. The semantic value relative to f of an atomic colored region r of m whose color is F is , where f(r) is a world region and f(F) is a property. A semantic value relative to f of an atomic colored region r of m whose color is F, , is true relative to world w i

f(r) possesses f(F) at w.

I am going to call the atomic colored regions of m the components of m. Now, I take it that in m, its components stand in some spatial relation to each other, thereby making up the map m. For example, consider the following formal map m on whose intended interpretation represents a stretch of Lost Winds beach (light gray) and an adjoining stretch of ocean (dark gray):

p. 476

Here the dark gray and light gray regions are the atomic colored regions of the map m. By standing in a spatial relation R in the map, they constitute the map m. I want to understand spatial relations between atomic colored regions in a map in such a way that a map m’ that is “identical” to m except for belonging to a map system whose maps are a di erent size than m has its atomic colored regions standing in the same 12

spatial relation in it that m’s atomic colored regions stand in, in m.

Here is such an m’:

So we’ll say that two maps m and m’ are such that their atomic colored regions stand in the same spatial relation i

m and m’ have the same number of atomic colored regions, corresponding regions are the same

shape, and the regions t together in the same way in m and m’. This means that if two maps have the same number of atomic colored regions, but the shape of some atomic colored region in one map is not the shape of any atomic colored region in the other map, the atomic colored regions stand in di erent spatial relations in the two maps. The idea is that if one of the two maps has at least one atomic colored region that is shaped di erently from every atomic colored region in the other map, the regions in the two maps cannot t together in the same way, and so must stand in di erent spatial relations in the two maps. I’ll return to this point below.

Just as words are combined in a sentence by standing in a syntactic relation, atomic colored regions are combined in a map by standing in the relevant spatial relation. On my theory of structured propositions, the syntactic relation the words stand in a sentence expressing a proposition provides all the signi cant structure to the relation holding the constituents together in the proposition. On my theory of structured map contents, the spatial relation that the atomic colored regions stand in in a map will provide all the signi cant structure to the content of the map. p. 477

To see the similarity between my accounts of structured propositions and structured map contents, let’s consider the proposition I claim is expressed by the following sentence:

(1) 13

Let’s call the syntactic relation that obtains between ‘Katie’ and ‘swims’ in the sentence here S.

I call

relations like S that lexical items stand in to form sentences sentential relations. In virtue of the existence of the English sentence 1, there is a two-place relation that Katie stands in to the property of swimming. The relation is this: there is a context c and assignment f such that ___is the semantic value relative to c and f of a lexical item e of some language L, and ___ is the semantic value relative to c and f of a lexical item e’ of L such that e occurs at the left terminal node of the sentential relation S that in L encodes ascription and e’ occurs at S’s right terminal node. This relation, I claimed, is the relation that holds Katie and the property of swimming together in the proposition that Katie swims. As a result, I’ll call it the propositional relation of the proposition that Katie swims. As I did in King (2007, 2009, 2013a), I’ll call an object possessing a property, or n objects standing in an nplace relation, or n properties standing in an n-place relation, etc., a fact. Then the proposition that Katie swims is the fact consisting of Katie and the property of swimming standing in the two-place relation mentioned above: there is a context c and assignment f such that Katie is the semantic value relative to c and f of a lexical item e of some language L, and the property of swimming is the semantic value relative to c and f of a lexical item e’ of L such that e occurs at the left terminal node of the sentential relation S that in L encodes ascription and e’ occurs at S’s right terminal node. Note that, as promised, the sentential relation S provides all the signi cant structure to the propositional relation binding together the constituents of the proposition expressed by 1 (Katie and the property of swimming). Now the maps m and m’ above have two atomic colored regions: the dark gray region and the light gray region. Recall that the semantic value relative to f of an atomic colored region r whose color is F is . Call the dark gray and light gray atomic colored regions of m D and L, respectively. For the intended interpretation function fim, fim(D)= a certain world region that in the actual world is stretch of ocean, call it O; fim(L) = a certain world region adjoining O that in the actual world is Lost Winds beach, call it LW; 14

fim(dark gray)= (the property of) being ocean and fim(light gray)=(the property of) being beach.

Then the

semantic value of D relative to fim== and the semantic value of L relative to fim==. The semantic values relative to fim of D and L are the constituents of the map m’s content on analogy with Katie and the property of swimming being the 15

constituents of the proposition expressed by 1. p. 478

interpretation

Exactly similar remarks apply to m’, its intended

function fim’ and its atomic colored regions D’ (its dark gray region) and L’ (its light gray

region). In particular, the semantic values of D and L relative to fim are identical to the semantic values of D’ and L’ relative to fim’. Now in virtue of the existence of m and m’ together with their intended interpretation functions, the semantic values of D and L relative to fim/the semantic values of D’ and L’ relative to fim’ stand in the following two-place relation: There are atomic colored regions r1 and r2 of some map m of some map system M such that m consists of r1 and r2 standing in the spatial relation R (Rr1r2) that encodes conjunction

where the semantic value of r1 relative to fim is ____ and the semantic value of r2 relative to fim is ___. We’ll discuss what it means to say a spatial relation, speci cally R, encodes conjunction subsequently. The content of the maps m and m’ relative to their interpretation functions is the fact consisting of the semantic values of D and L relative to fim/the semantic values of D’ and L’ relative to fim’ standing in this two-place relation: There are atomic colored regions r1 and r2 of some map m of some map system M such that m consists of r1 and r2 standing in the spatial relation R (Rr1r2) that encodes conjunction where the semantic value of r1 relative to fim is and the semantic value of r2 relative to fim is . Call this content BEACH. This content is true/accurate at w i

its constituents are true at w (i.e. O possesses the property of

being ocean at w and LW possesses the property of being beach at w); and for all regions of m uj, uk, Puj uk i Pfim(uj) fim(uk) at w and Cuj uk i

Cfim(uj) fim(uk) at w. Note that as promised the spatial relation R provides

all the signi cant structure to the relation binding together the constituents of BEACH. Now I want to say that any map together with an interpretation function f that su

ces for the existence of

BEACH has BEACH as its content relative to f. Since the existence of each of m and m’ together with their intended interpretation functions su

ce for the existence of BEACH, they both have BEACH as their

content. More generally, a map m constituted by n atomic colored regions standing in the spatial relation R that encodes conjunction has as its content relative to f the following fact: There are atomic colored regions r1, …rn  of some map m of some map system M such that m consists of r1, …, rn  standing in the spatial relation R (Rr1  … rn) that encodes conjunction where the semantic value of r1  relative to f is and …and the semantic value of rn  relative to f is , where Fi is the color of ri. This content is true at w i f(F1) at w and …and f(rn) possesses f(Fn) at w; and for all regions of m uj, uk, Puj uk i Cuj uk i

f(r1) possesses

Pf(uj) f(uk) at w and

Cf(uj) f(uk) at w. Note that since the spatial relation R between atomic colored regions of a map

gure in its content, two maps whose atomic colored regions stand in di erent spatial relations in those maps must have di erent contents. Further, since I’ve claimed that if a map has an atomic colored region that is shaped di erently from any atomic colored region in another map, the atomic colored regions in the maps stand in di erent spatial relations, any two such maps must have di erent contents. How on the present approach to maps’ contents do we capture the absence intuition? Here our only map features are colors, and so we want to capture the fact that if color F occurs on a map and a region r of that map is not F, the map represents the world region corresponding to r as lacking the property corresponding p. 479

to F. There are

at least two ways to capture the absence intuition. The rst follows C&V’s strategy of just

building it into the truth conditions of an atomic map stage. Here we would build it into the truth conditions of the semantic value relative to f an atomic colored region of map m. We previously said that semantic value relative to f of an atomic colored region r of m whose color is F, , is true relative to world w i

f(r) possesses f(F) at w. Instead, we could say that a semantic value relative to f of an atomic colored

region r of m whose color is F, , is true relative to world w i

f(r) possesses f(F) at w and no

region r’ of m not in r is such that f(r’) possesses f(F) at w. This will have the result that BEACH is true/accurate at w i

O possesses the property of being ocean at w and no region r of m not in r1 is such that

fim(r) possesses the property of being ocean at w; and LW possesses the property of being beach at w, and no region r of m not in r2 is such that fim(r) possesses the property of being beach at w; and for all regions of m uj, uk, Puj uk i

Pfim(uj) fim(uk) at w and Cuj uk i

Cfim(uj) fim(uk) at w. The second way of capturing the

absence intuition follows the strategy of Kulvicki (2015). Again, the only representational features of our formal maps are colors. These colors are incompatible in the following sense: no map region can be more than one color. Now suppose that two di erent map colors represent properties P and P’ that are incompatible in the following sense: nothing can possess both P and P’. In fact, the dark gray and light gray colors of m represent incompatible properties relative to the intended interpretation function: nothing can be both ocean and beach. If we were to suppose that, in general, di erent colors in our formal maps always 16

represent incompatible properties, we would account for the absence intuition.

For consider a color

occurring on a map—say the light gray region of m. Now consider every region of m that isn’t light gray. Since every such region is some color other than light gray, every such region gets mapped to a region of the

world that is represented as possessing a property incompatible with the property corresponding to light gray. Hence, we capture the absence intuition: if a color F occurs on a map, any region of the map that is not F represents a world region possessing a property that is incompatible with the property F represents and so not possessing the property F represents. Both strategies for capturing the absence intuition seem viable, and I won’t worry about which we should actually pursue. To this point, we have addressed the rst two points of the four points about map contents that we have claimed that our structured account of map contents would address. We have said what the constituents of map contents are: the semantic values relative to an interpretation function f of the atomic colored regions of the map. In the case of m/m’ and fim/fim’, these are: and . We have also said what it is that structures constituents to yield the structured complexes that are map contents: it is the two-place relation discussed three paragraphs back that binds together and in the fact that we claim is the structured content of m/m’ relative to fim/fim’. We now have to answer question 3: Why does this fact that we claim is the map content (relative to an interpretation function) have truth conditions/have truth values at worlds/represent the world as being a certain way? Before getting to that, there is a feature of the truth conditions of maps generally that needs to be addressed. p. 480

Note that in the case of both our unstructured map contents and our structured map contents, for a map’s content relative to an interpretation function f to be true at a world w, the following condition must be met: for all regions of m ui, uj, Pui uj i

Pf(ui) f(uj) at w and Cui uj i

Cf(ui) f(uj) at w. Where does this constraint

on the truth of a map content at a world w come from? Well, what it says is that the part/whole relations and connectedness relations between map regions mirror those of the world regions the interpretation function maps the map regions to at w. It seems to me that we understand maps and their contents this way because of a general convention about how to understand normal maps. We just conventionally take maps to be such that the part/whole relations and connectedness relations between map regions mirror those of the world regions the interpretation function maps the map regions to at a given world. This is a very natural convention in the sense that it would be hard not to interpret maps this way. In any case, I claim that the requirement that for a map to be true relative to an interpretation function f at a world w, it must be that for all regions of m ui, uj, Pui uj i

Pf(ui) f(uj) at w and Cui uj i

Cf(ui) f(uj) at w is one that conventions about

how to understand normal maps impose on any candidate for being a map’s content. Hence, it is imposed both on the unstructured and structured accounts of map contents considered here. Call this requirement the mirroring condition. Now, in describing the two-place relation that I claim holds the map content’s constituents together in the map content, and in describing the fact that I claim is the map’s structured content, I talked about a spatial relation R between the atomic colored map regions in m/m’ encoding conjunction. Why did I say that, and what does it mean? Each of a map’s atomic colored regions provides information (possibly false, of course) about a world region relative to an interpretation function. If I take the information provided by each of the atomic colored regions about world regions relative to f, how do I arrive at the information the map provides about the entire region the map represents? This is really a question about how to compose semantic values relative to an interpretation function f of atomic colored map regions to yield the information the map as a whole provides about the whole region it represents relative to f. And the answer is that we conjoin the semantic values relative to f of atomic colored map regions. I think that we do this because we interpret the spatial relation R that atomic colored map regions stand in in the map as instructing us to compose their semantic values relative to f by conjoining them. That the spatial relation R instructs us to compose semantic values relative to f by conjoining them is what I mean when I say that R encodes conjunction. That we interpret spatial relations that atomic colored map regions stand in as instructing us to conjoin semantic values relative to f of atomic colored map regions is a matter of a convention as to how we interpret maps. It would be possible, if odd, to have a map system in which we interpreted these spatial relations di erently. For example, the map m/m’ could belong to a map system where we interpreted spatial relations between

atomic colored map regions as encoding disjunction. In that case, m/m’ on its intended interpretation would carry the information that either the world region the light gray map region is mapped to is beach, or the world region the dark gray map region is mapped to is ocean. Recall that in giving C&V’s semantics for p. 481

maps, I highlighted the

fact that on their semantics, a map stage is true i

its well-formed atomic map

stages are conjointly true. This, again, is just a re ex of the fact that we interpret spatial relations between atomic colored map regions in a map as instructing us to conjoin the semantic values relative to f of the atomic colored map regions of the map. Let’s now turn to the question of why the fact that I claim is the content of the maps m, m’ has truth conditions, and why it has the speci c truth conditions it has. Though the two-place relation that I claim holds the constituents together in m’/m’s content is highly complex, let’s suppress that and view it as simply a two-place relation between those constituents. Then we can represent the fact that I claim is the structured content of m/m’ as follows:

(2) Call the relation that binds together the constituents of a map’s content, the content relation. Now, we mentioned that the spatial relation R between m’/m’s atomic colored regions is interpreted as instructing us to compose the semantic values relative to fim/fim’ of these atomic colored regions—the constituents of m’/m’s content—by conjoining them. If the content relation in 2 were so interpreted, m’s content 2 would be true at w i

the conjunction of the semantic value relative to fim of D and the semantic value relative to

fim of L is true at w and the mirroring condition obtains i

the conjunction of and is true at w and the mirroring condition obtains i

is true at w and is true at w and the mirroring condition obtains i

O possesses the property of being ocean at 17

w and LW possesses the property of being beach at w and the mirroring condition obtains.

And these are

BEACH’s truth conditions. So we can explain why 2—the content of m/m’, that is, BEACH—has the truth conditions it does if we hold that its content relation is interpreted as R is: as encoding conjunction/as instructing us to compose semantic values by conjoining them. So we should hold that the content relation is interpreted as encoding conjunction. But it is not enough to just say this. We need to give an account of what it is that constitutes our interpreting the content relation as encoding conjunction. We need to say what it is that we do that amounts to our interpreting the content relation in this way. We want to know what it is that constitutes an agent interpreting BEACH’s content relation as encoding conjunction and thereby endowing BEACH with truth conditions and grasping it. As a rst step in that direction, we should note that in order for us to interpret BEACH’s content relation, we must have cognitive access to it. The idea that we could interpret BEACH’s content relation in a certain way without having any cognitive access to it seems incoherent. To interpret a part of a thing, we have to have cognitive access to the thing, or at least the part of it we interpret. But notice that having this cognitive access to BEACH in order to interpret its content relation cannot be a matter of having some attitude to BEACH qua content. For we are p. 482

trying explain how/

why BEACH is a content—something with truth conditions—by explaining how we

interpret its content relation thereby endowing it with truth conditions. If this explanation presupposes that we have some attitude towards BEACH qua content, we presuppose that BEACH already is a content, thereby undermining our explanation of how/why it is a content. So the cognitive access we have to BEACH that allows us to interpret its content relation must be a kind of having the fact BEACH in mind without having any attitude toward it qua content. The idea is going to be that in grasping BEACH qua content, we rst have cognitive access to it, and then in actually grasping it qua content, we interpret its content relation thereby 18

both endowing it with truth conditions and entertaining it qua content.

Let’s see how we might do that.

First of all, note that map m is a fact in my sense. It consists of objects (paper parts) possessing properties (being colored) and standing in relations (the spatial relation R). Now suppose one is presented with map m and knows about the map system (formal maps) of which it is a member. One now has cognitive access to m in the sense of having m in mind and being able to think about it etc. One thereby has cognitive access to what I’ll call the interpreted map mi, which is simply the map m together with the semantic relations relative to fim the atomic colored regions bear to their semantic values relative to fim. We can represent mi as follows:

where the lines are the semantic relations relative to fim between the atomic colored regions of m and their semantic values relative to fim. Given that one has cognitive access to m initially, one has cognitive access to p. 483

the interpreted map mi in virtue of knowing

what the semantic values relative to fim of the atomic colored

regions of m are. This is just a matter of knowing what world region fim maps each such map region to and what property fim maps each color to. With m in mind, one has cognitive access to mi simply by exercising this knowledge and retrieving the semantic values relative to fim of the atomic colored regions of m. Let’s say that the fact that o is P is a witness for the fact of there being P’s. More generally, say that the fact that o bears R to o’ and the fact that o bears R to something are both witnesses to the fact of there being an x and a y such that xRy, and so on. The former is a witness to the fact of there being an x and y such that xRy with respect to o and o’, and the latter is a witness to the fact of there being an x and y such that xRy with respect to o. Now the interpreted map mi is a witness to BEACH—the fact that I claim is the content of m— with respect to the atomic colored regions of m, m itself, and the map system M. BEACH is the result of taking mi and existentially generalizing on the atomic colored regions of mi, m, and the map system m belongs to. That makes mi a witness for BEACH with respect to the atomic colored regions of mi, m, and the map system M. Now the crucial point: having cognitive access to a witness to a fact allows one to have cognitive access to the fact witnessed. This seems to me to be true in general: having access to the fact that o is P allows one to have cognitive access to the fact of there being P’s. The crucial idea here is that when one has cognitive access to a given fact, one is able to abstract from certain features f1,…,fn of the fact one has in mind, thereby having in mind the fact witnessed with respect to f1,…,fn by the original fact. That is, in abstracting from a

feature fj of fact F one has in mind, one thereby has in mind the fact that F witnesses with respect to fj. So this abstracting from a feature fj of a fact F is existentially generalizing over it, yielding the fact F witnesses with respect to fj. If that is right, then having cognitive access to mi allows one to have cognitive access to BEACH. Let’s now go through the entire process of endowing BEACH with truth conditions and thereby grasping it by interpreting its content relation. Initially, one is presented with the map m and so has it in mind. Recall that one does not yet grasp its content. One simply has the map in mind as an object in the world one is aware of. Next, one exploits one’s knowledge of the semantic values relative to fim of the atomic colored regions of m in retrieving those semantic values relative to fim. One thereby gains cognitive access to the interpreted map mi and thereby has it in mind. At this point, one abstracts from now irrelevant features of mi: the atomic colored regions of m, m itself, and the map system M. These features of mi were just a means for getting from m to the semantic values relative to fim of m’s atomic colored regions. Since that has been done in moving from having m in mind to having mi in mind, there is no reason to retain these features. One now has BEACH in mind, which mi witnessed with respect to the atomic colored regions of m, m itself, and the map system M. Hence, one is now in a position to interpret BEACH’s content relation and thereby grasp m’s content. Note that so far, there has been no composition of the semantic values relative to fim of the atomic colored regions of m, and so the signi cance of the spatial relation R between atomic colored regions in m as instructing us to compose semantic values by conjoining them has not yet been processed. p. 484

Given that one now has BEACH in mind in such a way as to be in a position to interpret its content relation as encoding conjunction, we now need to say what it is that constitutes our so interpreting it. That is, what is it that we do that amounts our so interpreting it? It is simply that we compose the semantic values relative to fim of the atomic colored regions of m by conjoining them. In the end, this is just a matter of the spatial relation R between the atomic colored regions of m having the semantic signi cance it does. That is, it is just a matter of R encoding conjunction. One begins by having the map m in mind. One then accesses the semantic values relative to fim of the atomic colored regions of m. One thereby has the interpreted map mi in mind. One then abstracts from the atomic colored regions of m, m itself, and the map system M that m belongs to. One thereby has BEACH in mind. Only now in grasping BEACH are semantic values composed; as was mentioned earlier, no composition of semantic values occurs in going from having m in mind to having the interpreted map mi in mind to having BEACH in mind. Hence, when the composition of semantic values occurs, it is BEACH that we have in mind/that we have cognitive access to. In now composing semantic values by conjoining them, the semantic signi cance of the spatial relation R that atomic colored map regions stand in in m—encoding conjunction—is nally processed. Since it is BEACH that one has in mind in conjunctively composing semantic values, one counts as interpreting BEACH’s content relation as encoding conjunction, thereby endowing BEACH with truth conditions and grasping it qua content. So interpreting the spatial relation R between atomic colored map regions in m in the way we do, by composing semantic values at a certain point in the process just described, just is interpreting BEACH’s content relation in the relevant way. Note that if interpreting BEACH’s content relation as encoding conjunction just is interpreting the spatial relation R between atomic colored map regions in m as encoding conjunction, then it is metaphysically necessary that R is interpreted as encoding conjunction i

BEACH’s content relation in

interpreted as encoding conjunction. We have now addressed three out of the four points we set out to address regarding structured map contents. We have said what the constituents of map contents are. We have said what binds these constituents together in the map’s structured content. We have explained why/how map contents have truth conditions and so represent the world as being a certain way. We now turn to the nal point: how the current theory individuates map contents. According to the current account, any two maps m1 and m2 have the same content relative to their respective interpretation functions fm1 and fm2 i

m1 and m2 have the

same number of atomic colored regions with the same semantic values relative to fm1 and fm2, respectively,

and the atomic colored regions of m1 and m2 with the same semantic values relative to fm1 and fm2 are the same shape, and the atomic colored regions of both maps stand in the same spatial relation that is interpreted in the same way in each map. The condition that the atomic colored regions of m1 and m2 stand in the same spatial relation that is interpreted in the same way is going to be satis ed by virtually any two maps whose atomic colored regions stand in the same spatial relation, since in general, we interpret the spatial relations between atomic colored map regions as encoding conjunction. p. 485

These identity conditions for maps contents are fairly ne-grained. They allow m and m’ above to have the same content relative to their intended interpretation functions. Further, it allows m to have the same content as any other map m’’ that di ers from m only in what colors its atomic colored regions are, as long as the intended interpretation function of m’’ maps the colors of m’’ to the same properties as the corresponding colors of m get mapped to and maps the atomic colored regions of m’’ to the same world regions as the corresponding atomic colored regions of m get mapped to. But, to repeat, no map can have m’s content whose atomic colored regions di er in number from m’s or di er in shape from m’s or stand in a di erent spatial relation than m’s. This may seem to have some counterintuitive consequences. For example, consider the following map m’’ and assume that relative to its intended interpretation function fim’’, its light gray and dark gray atomic colored regions have the same semantic values as do the dark gray and light gray regions of m relative to its intended interpretation function:

Nonetheless, m’’ has a di erent content than does m, since its atomic colored regions with the same semantic values as m’s relative to their respective interpretation functions are slightly di erently shaped than m’s and so must stand in a di erent spatial relation than those in m do. But, it might be thought, m’’ should have the same content as m. That it doesn’t is a strike against the current account of content, it might be thought. There is a formulation of the current account that would yield the result that m and m’’ have the same content. Just say that the following fact is the content of both: There are atomic colored regions r1  and r2  of some map m of some map system M such that m consists of r1  and r2  standing in a spatial relation (in that order) that encodes conjunction where the semantic value of r1  relative to fim  is and the semantic value of r2  relative to fim  is . On this view, m and m’’ have the same content since p. 486

we have here existentially generalized over the spatial relation R

that atomic colored regions stand in in

constituting a map. Hence, maps may have the same content even though they have di erently shaped atomic colored regions, whose semantic values are the same relative to their respective interpretation functions, and which therefore must stand in di erent spatial relations in constituting their respective maps. The problem with this account goes back to the idea that the Mercator and Robinson projections of

the world map are such that one could believe one and not the other, and so they must be assigned di erent contents. As far as I can see, the best shot at getting that result by scaling up the current account of formal maps’ structured contents to yield structured contents of ordinary maps is to say that the reason that these projections have di erent contents is that their regions with the same semantic values relative to their intended interpretation functions are di erently shaped and so stand in di erent spatial relations in the two maps. But then this forces us back to our original account and requires us to say that m and m’’ have di erent contents for the same reason. Looked at in this way, though, I think the current account is the right way to go. If the way to get the Mercator and Robinson projections having di erent contents is to say that di erences in shapes of map regions and so those regions standing in di erent spatial relations in two maps results in the maps having di erent contents, as seems plausible, then we have to say that two maps that di er only very slightly in this way, as do m and m’’, have di erent contents. Perhaps the following thought experiment will assuage the worry that assigning di erent contents to m and m’’ is a bad consequence of the current account. Because the corresponding atomic colored regions of m and m’’ are slightly di erently shaped, there will be points on m that appear to correspond to ocean, where those same points on m’’ 19

appear to correspond to sand, and vice versa.

Suppose we impose very ne-grained coordinates on both

maps and pick out such a point, call it . Finally, suppose you are asked to bet $100 on whether if you stand at , your feet will be wet. Depending on which map you consult, you may make di erent bets. This is at least some reason to think the maps have di erent contents. So in the end, I think our account of structured map contents individuates them in the right way.

Notes

p. 487

1.

Camp (2018) argues that map contents are not propositional. However, so far as I can see, she takes no position on the question of whether one can believe a mapʼs content. I reject Campʼs argument that map contents are non-propositional, because I reject the account of propositionality she employs. I remain neutral on the question of whether map contents are propositional here. But note that if they arenʼt, I am committed to the claim that some objects of belief are nonpropositional.

2.

x is self-connected i any two parts of x that make up the whole of x are connected. Iʼll discuss connection further below. Since maps are self-connected, they will be associated with self-connected regions of the world.

3.

Casati and Varzi (1999, 32).

4.

Ibid., 194. Though, as I said, they actually favor the theory they call GEMTC.

5.

As far as I can see, the idea that the absence of a map feature at a location l on a map where that feature occurs at other places in the map indicates that what the feature stands for isnʼt present at the world region corresponding to l first occurs in Pratt (1993). It is mentioned in Rescorla (2009), Kulvicki (2015), and Camp (2018), among others.

6.

Note that Iʼve built conditions 1 and 2 into the conditions under which a map is true relative to f at w, whereas C&V make it a restriction on interpretation functions. I think my approach is superior because we should say that a map that violates these constraints for a choice of f and w misrepresents w by not mirroring its mereotopological structure, and so is inaccurate/false instead of simply ruling out such a choice of w and f. A er all, there have historically been such maps and we want to say that they are inaccurate/false, and not simply that they are to be ruled out as maps so interpreted.

7.

Stalnaker (1999, 241).

8.

Some of these responses are discussed in King (forthcoming).

9.

The account of structured contents of natural language sentences defended in King (2019) is slightly di erent from the account in King (2007) and is my preferred account. However, the former is more di icult to explain than the latter, and since the di erences between the accounts donʼt matter here, I have gone with the latter account.

10.

The latter problem(s) is (are) raised and, I claim, solved in King (2009, 2013b).

11.

Section 54.

12.

I assume a map system M specifies the size of maps in it and that all maps in M are the same size.

13.

The syntactic relation obtaining between ʻKatieʼ and ʻswimsʼ in the sentence ʻKatie swimsʼ is likely much more complicated than it is represented to be in 1. This doesnʼt e ect my view. See King (2013a) for discussion.

14.

In general, Iʼll use ʻfimʼ to designate the intended interpretation function of a given map m.

15.

One disanalogy here is that the constituents of the mapʼs content are true and false at worlds; whereas the constituents of the proposition that Katie swims are not. If we want to analogize maps to sentences, we should think of maps as the analogues of conjunctive sentences, where the mapʼs components (its atomic colored regions) are the conjuncts. The semantic values of these components relative to an interpretation function are true or false at worlds and are the constituents of the mapʼs structured content. As weʼll see below, if the mapʼs structured content is true at a world w, all of its constituents are true at w.

16.

I am suppressing reference to interpretation functions here. To be more explicit, the idea is that for all distinct map colors F and Fʼ, every interpretation function maps them to incompatible properties.

17.

Recall that the mirroring condition is imposed on any candidate for being the content of a map in virtue of conventions about how to understand normal maps. I ignore the absence intuition here.

18.

I talk of grasping BEACH qua content to distinguish it from merely having the fact BEACH in mind without grasping it qua content. The latter, I have been calling having cognitive access to BEACH in such a way as to be in a position to interpret its content relation and thereby grasp it qua content.

19.

I am being careful not to say that a point on m represents a world region as being ocean whereas mʼʼ represents that same point as being sand. The dark gray and light gray regions of m and mʼʼ, respectively, are mapped to the exact same world regions, and so they both represent the same world region as being ocean, given the intended interpretation of light gray and dark gray. Similarly for the light gray and dark gray regions of m and mʼʼ respectively.

p. 488

References Camp, Elisabeth (2018), ʻWhy Maps Are Not Propositionalʼ, in A. Grzankowski and M. Montague, eds., Non-Propositional Intentionality (New York: Oxford University Press), 19–45. Google Scholar Google Preview WorldCat COPAC Casati, Roberto, and Achille Varzi (1999), Parts and Places: The Structures of Spatial Representation (Cambridge, MA: MIT Press). Google Scholar Google Preview WorldCat COPAC King, Je rey C. (2007), The Nature and Structure of Content (New York: Oxford University Press). Google Scholar Google Preview WorldCat COPAC King, Je rey C. (2009), ʻQuestions of Unityʼ, Proceedings of the Aristotelian Society, 109, pt. 3, 257–277. Google Scholar WorldCat King, Je rey C. (2013a), ʻOn Fineness of Grainʼ, Philosophical Studies, 163/3, 763–781. Google Scholar WorldCat King, Je rey C. (2013b), ʻPropositional Unity: Whatʼs the Problem, Who Has It and Who Solves It?ʼ, Philosophical Studies, 165/1, 71–93. Google Scholar WorldCat King, Je rey C. (2019), ʻOn Propositions and Fineness of Grain (Again!)ʼ, Synthese, 196/4, 1343–1376. Google Scholar WorldCat King, Je rey C. (forthcoming), ʻUnstructured Contentʼ. Google Scholar Google Preview WorldCat COPAC Kulvicki, John (2015), ʻMaps, Pictures and Predicatiomʼ, Ergo 2/7, 149–173. Google Scholar WorldCat Pratt, Ian (1993), ʻMap Semanticsʼ, in Andrew U. Frank and Werner Kuhn, eds., Spatial Information Theory: A Theoretical Basis for GIS, conference proceedings, A Theoretical Basis for GIS, International Conference, COSIT ʼ95, Semmering, Austria, September 21– 23, 1995, Lecture Notes in Computer Science, no. 716 (Berlin: Springer-Verlag), 77–91. Google Scholar Google Preview WorldCat COPAC Rescorla, Michael (2009), ʻPredication and Cartographic Representationʼ, Synthese, 169, 175–200 Google Scholar WorldCat Stalnaker, Robert (1984), Inquiry (Cambridge, MA: MIT Press). Google Scholar Google Preview WorldCat COPAC Stalnaker, Robert (1987), ʻSemantics for Beliefʼ, Philosophical Topics 15. Reprinted in Stalnaker (1999). I use the pagination of the latter. Google Scholar WorldCat Stalnaker, Robert (1999), Context and Content (New York: Oxford University Press). Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

p. 489

22 Map Semantics and the Geography of Meaning  Gabriel Greenberg https://doi.org/10.1093/oxfordhb/9780192844118.013.23 Published: 22 May 2024

Pages 489–522

Abstract This chapter develops a semantic theory for maps and situates it within the broader geography of meaning and semiotic signi cance. The discussion focuses on three central aspects of map semantics: the use of space, line marking, and linguistic tags. It is argued that the treatment of space in maps must be based on geometrical projection from a viewpoint rather than the traditional analysis in terms of spatial isomorphism. The chapter then shows how to integrate the projection-based semantics of maps and the symbolic semantics of markers and tags, within the framework of possible-world semantics. The resulting account addresses issues of accuracy, viewpoint, orientation, and exhaustivity.

Keywords: maps, cartography, semantics, absence intuition, viewpoint, semiotics, depiction, projection, space, diagram Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Maps and Meaning Maps are rich with potential for meaning. Their geographic contents are what enable simple acts of way nding, say, from the grocery store to the gas station. But they can also reinforce the tectonic outlines 1

of global ideologies, such as the dominance of First World nations over the countries of the Third World.

The many meanings of a map can be visualized as a loosely organized sequence of concentric spheres. At the center is the strictest, most conventionalized interpretation of a map as a representation of objects and features distributed in space. The spheres of meaning radiate out from that core in degrees of social 2

dependence and political signi cance (Figure 22.1).

Figure 22.1. Spheres of meaning for maps. The innermost sphere of meaning is the literal meaning of the map, what I will call its cartographic content. It describes the way that objects and properties are distributed over some geography, and it is determined p. 490

from the map itself by more and less natural

semantic convention. The cartographic content of the map

in Figure 22.2, for example, xes the locations of concretia like roads, rivers, and population centers, but also relatively abstract concepts like political boundaries, zip codes, and types of land usage. There are few limits on the kinds of properties that a map can encode at this level, as long as they are the sorts of things that can be sensibly understood as having geographic locations.

Figure 22.2. Excerpt from Middlesex County Atlas (2002, 32–33), published by Hagstrom Maps. Next, the discourse e ect of a map, on a given occasion of use, corresponds to how the map’s cartographic content is leveraged to contribute to the broader exchange of information in discourse. This, in e ect, is the 3

speech-act level of representation for maps. Here, we can distinguish between assertoric uses of a map to describe how things are (like a road map) and imperatival uses to describe how things should be (like an urban planner’s map for a city). Maps, of course, can also be counterfactual, cautionary, or fantastical. In most of these cases, the map maker will choose a target for the map: the time and place (actual or counterfactual) which the map functions to accurately describe. A map’s discourse e ect also includes the way a map contributes to the broader multimodal discourse in which it may be embedded. Consider an actual

road map: it includes not just a map narrowly de ned, but also an index, legend, distance table, scale, compass rose, descriptive text, and a title. All these aspects combine further to form a more complex whole. Whereas cartographic content and discourse e ect have to do with the articulation and management of p. 491

information, the social semiotics of a map describes the social

forces which lead to a map’s production

and display, and the characteristic e ects it has on the psychologies and social- identities of its audiences. 4

Some of the traditional interests of pragmatics may be found here. We may ask, for example, for what purpose was the map created, or what problem it was designed to solve: navigation, building, warfare, play, 5

boundary-making, colonization, or, more commonly, some mixture of uses. Social semiotics also includes a map’s connotation: the associations, assumptions, and expectations it reliably elicits from audiences in a 6

given cultural setting. For example, how does the design of the map signal its authoritativeness (or tentativeness), its political a

liation, or its basis in scienti c evidence? Here the manner of display is also

relevant: is the map printed on one page of many in an atlas, or is it ostentatiously hung on the wall of a 7

study, or is it part of a company’s logo? What is the displayer trying to say about themselves?

At the outermost sphere is a map’s political meaning: the way it shapes and is shaped by ideology and the societal distribution of power, often independently of the conscious awareness of those who create and use the map. Because maps are so intimately tied to land, and land to sovereignty and nationality, maps are an 8

inevitable source of rich political meaning.

Consider the way a map can make geography “available” to those with access to the map, automatically conferring a kind of epistemic power to the map’s possessors. This might play out at a local level: having a map of a nearby neighborhood allows you to explore, shop, or dine there, even though you may feel yourself an outsider. At a larger scale, cities that are featured in the map of a popular guidebook may make those cities welcoming to tourists. Of course, maps of the globe are the ultimate site of colonial ambition, rendering “undiscovered” lands available for conquest in the minds of the map holders: As much as guns and warships, maps have been the weapons of imperialism. Insofar as maps were used in colonial promotion, and lands claimed on paper before they were e ectively occupied, maps anticipated empire. (Harley 1988, 132) Likewise, the composition of a map can have irresistible e ects on the ways that map consumers understand the distribution of status and power in the world. Most modern maps follow the convention of putting North at the top of the map. Applied to globe maps, this has the e ect of placing most First World nations on top, and most Third World nations below. The stark visual re-centering that occurs when the map is inverted reveals the subliminal e ects of the standard orientation. The inverted world map has the same cartographic content as the original, but at the level of political meaning, it clearly shifts the global 9

south into the center of attention (Figure 22.3).

10

Or consider the late twentieth-century debate over the rise of the Gall-Peters projection.

The classical

Mercator projection, still widely used at the time, preserves local shape, and has functional bene ts for nautical navigation. It also has the e ect of in ating the size of landmasses that are farther from the equator, so that the areas of First World nations in North America and Europe, as well as Russia and p. 492

Greenland are all expanded relative to the map’s total size, while Third World nations in Africa, South Asia, Central America, and South America, along with Australia, are all compressed. The Gall-Peters projection, widely criticized for its supposed aesthetic defects, represents all areas proportionally (Figure 11

22.4).

The inevitable e ects of the two maps on audiences’ assessment of power and importance are

obvious.

Figure 22.3. North-up and south-up Mercator projections. Source: Image from Wikipedia. Public domain: Creative Commons Attribution-Share Alike 3.0 Unported license. Documentation: https://en.wikipedia.org/wiki/File:Mercator_projection_SW.jpg.

Figure 22.4. The Mercator projection (le ) preserves local shape. The Gall-Peters projection (right) preserves area. Source: Public domain: Creative Commons Attribution-Share Alike 3.0 Unported license. Documentation image 1: Documentation image 2: https://en.wikipedia.org/wiki/File:Gall-Peters_projection_SW.jpg. As map consumers, our interests’ range over all forms of meaning, from spatial display to political potency. Still, as the model of concentric spheres suggests, all these meanings depend at least to some extent on the innermost sphere of cartographic content. Understanding how maps express this core is the task of this chapter. Part of what makes maps special is the diversity of expressive resources they marshal to x their core cartographic contents. They are “picture-language hybrids” (Kulvicki 2015, 149), combining pictorial elements to construct space, and a host of symbolic and linguistic elements to enrich this space with features, landmarks, and terrain. In this respect, the study of map semantics joins the broader investigation within cognitive science into iconic and multimodal representations, one that has already seen 12

breakthroughs in our understanding of diagrams, gesture, sign language, and more. p. 493

13

A central theme of this chapter is that we ought to think of maps as a special kind of picture.

Like any other

picture, maps consist of marked 2D planes that represent 3D environments. And tellingly, though this is 14

sometimes overlooked,

maps always represent their subject matter from a point of view. This is usually a

spread out “god’s eye view,” but it is, nonetheless, a point of view showing a particular spatial and temporal part of the world from an overhead position. Just as pictorial perspective is the result of geometrical projection, maps, too, are governed by a range of projective methods. In what follows, I argue that projection from a viewpoint is the central principle of map semantics, just as it is for pictures. Notwithstanding these analogies, maps di er from ordinary pictures by relying more on explicit symbolic signs, and less on implicit visual cues. Thus, they standardly include a large share of symbolic elements, including a library of essentially arbitrary color and line-codes (the sort of thing speci ed in a legend), along with textual labels, and graphical symbols. And whereas pictures typically rely on a range of depth and shape cues to convey volumetric space, maps typically do not engage depth perception, 3D shape 15

perception, or object recognition.

Wollheim put it this way:

To [extract information from maps] we do not rely on a natural perceptual capacity, such as I hold seeing-in to be. We rely on a skill we learn. It is called, signi cantly, “map-reading”: “mapreading.” (1987, 60–61) Crudely speaking, the interpretation of ordinary pictures relies more heavily on normal perceptual processes, whereas that of maps relies more heavily on symbolic conventions. As a result, an explicit semantics of maps is much more nearly within grasp, while a detailed understanding of the visual system is still a monumental work in progress. In the remainder of this chapter, I rst discuss a variety of facts about the expression of cartographic content that should inform any semantics for maps, in section 2. In section 3, I outline a semantic theory which aims to explain these facts and illuminate the general mechanisms of map representation. Section 4 situates maps semantics within the broader geography of meaning.

2. Space, Lines, and Labels Cartographic content is constructed by bringing together two aspects of a map’s meaning: the space it expresses, on one hand, and the objects and properties, denoted by markings and labels, that it distributes throughout this space, on the other. In this section, I highlight a handful of simple but fundamental facts of map representation, organized around the key representational elements of cartography: space, markers (including lines, colors, and textures), and tags (including labels and graphical symbols). Lessons are drawn for the eventual articulation of a semantics for maps.

p. 494

2.1. Space A metric space is one where every point in that space stands in a determinate distance to every other point. The familiar 3D physical space that surrounds us is metric in this sense. Such a space is constituted by an underlying network of metric relations which connect every point to every other point. Of course, a map itself de nes a 2D metric space, while the world it represents is a 3D metric space. How should we understand the relationship between map space and world space? To approach this question, I restrict our attention to maps which express metric content, as opposed to quasi-topological subway maps, and to maps that are planar projections, as opposed to the spherical projections displayed by globe maps. Henceforth, this is what I’ll mean by “map.” To make things concrete, let us introduce a set of axes in the represented world, where the XY plane runs perpendicular to the direction of gravity (or parallel to the ground, were the ground at), and the Z dimension runs parallel to the direction of gravity (or perpendicular to the at ground). We can now state two basic facts about the spatial content of maps. Fact 1. Maps attribute complete metric relations in the XY dimensions. Fact 2. Maps (generally) do not attribute metric relations in the Z dimension. Fact 1 says that a map will determine the X and Y position, akin to longitude and latitude, of every location depicted, relative to every other location depicted. This is why, given a map and a landmark on the map, you can determine the direction and distance “as the crow ies” of every other landmark. The restriction to the XY dimension means that maps attribute complete metric relations to the landmarks they depict—but only when these are projected onto the XY plane.

Fact 2 says that maps do not attribute metric relations in the Z dimension (unless they include special markings, such as contour lines). Maps, in other words, do not attribute relations of depth. This is not the same as attributing zero depth; maps do not require for their accuracy that the geographies they represent be at planes. Rather, they are silent on matters of depth; their spatial content is indeterminate in the Z dimension. An important corollary of Fact 2 is that maps do not express fully metric relations between all the locations or landmarks they depict. Put another way, you cannot, just by reading a map, tell the distance between every pair of landmarks it depicts. This fact is familiar to hikers. Suppose you are located at landmark B, and are trying to decide, from the map in Figure 22.5, whether it will be a faster walk to A or to C. Super cially, it looks on the map like the distance to B is shorter than the distance to A. But such looks can be deceiving. In fact, there is a sharp downward slope that separates A from B, revealed in Figure 22.6, while B and C lie along the same plane. As a consequence, the AB distance is greater than the BC distance (not to mention a more arduous hike), even though, in Figure 22.5, the distance on the page between “A” and “B” is shorter than that between “B” and “C”.

p. 495

Figure 22.5. A map of the terrain.

Figure 22.6. An elevation drawing of the terrain. These elementary observations tend to be overlooked in contemporary semantic accounts of maps. As a consequence, we still have not got a fully satisfactory treatment of what is arguably the most basic feature 16

of cartographic meaning: the expression of spatial relations.

A number of authors adopt a quasi-linguistic view of maps, in which the locations of the map are treated as 17

referring terms, and their denotations are locations in the world.

Markings on maps denote properties,

and these properties are attributed to the world-locations denoted by the map-locations occupied by the markings. At the extreme, views of this kind hold that the spatial content of a map is exhausted by attributing properties to locations in this way. The limitations of such a view have been argued forcefully by Camp (2018, 38), who concludes that “a map makes the spatial relations among denoted object/properties directly available in a way that a list of symbol-pairs plus [an assignment function] does not.” In other words, a purely referential approach to cartographic space fails to explain Fact 1. Pratt (1993, 79–80), for example, begins his exploration of map semantics with an analogical extension of predicate calculus. Just as names in a sentence denote individuals in the world, coordinates in a map denote locations in the world. Locations, in turn, are taken to be geographic individuals. Since Pratt models map coordinates as pairs of numbers, he preserves a measure of metric relations in the 2D space of the map; unfortunately, the crucial attribution of metric relations in the 3D space of the world drops away. Pratt simply stipulates an assignment from map coordinates to locations; not only is the assignment unresponsive to the spatial properties of the map coordinates, it returns no spatial attributes to be associated with the locations that are its values. This is a consequence, Camp (2018, 27) argues, of treating maps too much like an exotic form of symbolic representation, and not taking seriously the distinctively iconic way in which maps encode space. p. 496

The lesson for a semantics of maps is that the mere denotation of locations by points on the map is not su

cient to establish the spatial aspect of cartographic content. The spatial relations of points on the map

to one another must be iconically rendered into attributions of spatial relations in the geography. We therefore seek a geometric principle that might relate the 2D space of the map to the 3D space of the represented geography. As Rescorla (2009, 197–198) concludes “A map’s geometric structure is not just another element to be listed alongside its markers and coordinates … Rather, the markers and coordinates 18

stand in geometric relations, and those relations bear representational import.”

Scholars facing this juncture have consistently proposed that the relevant relation is one of isomorphism with respect to metric relations (supplemented, perhaps, by a scalar transformation). “A map represents how various objects and properties are distributed in physical space. It does so by purporting to replicate relevant geometric features of the spatial region it represents” (Rescorla 2009, 178). In other words, many 19

have found it natural to assume a semantic rule along the following lines. refer to the denotation or content of X.

Here I use the notation ⟦X⟧ to

(1)  Isomorphism Semantics A map M is accurate at a world w only if the distance between any two points p and p′ in M is equal to the distance between ⟦p⟧ and ⟦p′⟧ in w, times a scalar constant k.

But this apparently sensible proposal seems to be incompatible with Fact 2 and its corollary. For as we saw above, due to the vicissitudes of depth, distance on the 2D plane of the map is not generally proportional to distance in the 3D geography represented; isomorphism only holds when the geography itself is perfectly 20

at.

Thus the isomorphism semantics would entail that the map in Figure 22.5 is inaccurate, because the

distance relations between points on the map are not isomorphic to distance relations between the denoted locations in the world. And yet, Figure 22.5 is perfectly accurate (assuming the geography depicted in Figure 22.6): it doesn’t comment about distance relations in general, only distance relations as projected to the XY plane. That it would be ine ective as a map for hiking—but e ective as a map of voting districts—is independent of its accuracy conditions. Another way to put the point is that the isomorphism semantics neglects the role of viewpoint. It presupposes that the relevant spatial relation between map and world can be de ned exclusively in terms of their intrinsic geometries. Yet every map displays its geography relative to a viewpoint. At the very least, maps depict geography from above, and usually in a speci c orientation. The isomorphism semantics has no role for viewpoint. As a result, it misses the fact that, relative to such a viewpoint, relations of distance in the Z dimension can be lost in projection, as in Figure 22.5, whereas they would be explicit if the viewpoint were oriented to the side, as in an elevation drawing like Figure 22.6. p. 497

What is correct about the principle of metric isomorphism, of course, is that distance on the 2D plane of the map is isomorphic and proportional to distances in the 3D geography as projected onto a 2D plane. But then the principle of isomorphism is no longer doing any real work; we might as well just say that the 2D map is a projection of the 3D geography. Thus the intuition behind the isomorphism approach quickly leads us to adopt a projection semantics for maps. Perhaps this should be no surprise, as projection is one of the central themes of modern cartography, and because projection semantics was originally developed for ordinary pictures, where issues of depth are paramount. In any case, the general moral is that it is projection, not isomorphism, which should provide the semantic backbone of cartographic representation. I pursue this idea in section 3.

2.2. Markers The space of a map is lled with lines, colors, and textures, collectively known as markers. Every marker has a location on the map and belongs to a marker type. Marker types might include, for example, brown color ll or green color ll, or, in the case of lines, dotted lines or solid lines. The meanings of the various marker types are typically supplied in a legend or map key. Though there is considerable leeway in the kinds of meanings that can be assigned, they are usually properties or categories. Brown might mean land, for example, while green might mean forest.

Figure 22.7. A variety of marker types. Source: Excerpt from the Middlesex County Atlas (2002, 32–33). As a demonstration of the diversity of possible marker types, one can nd in the close-up of the Middlesex County map in Figure 22.7, at least eight di erent kinds of markers: p. 498

We can think of the syntax of a map marker as factored into two components: (a) marker type and (b) relative spatial location on the map. Nearly everyone writing on the semantics of maps has recognized that these two aspects of marker syntax express two kinds of content: (a) the marker type expresses a property; and (b) the location of a marker on the map helps pick out a spatially situated object in the world to which that property is attributed. (Here the notion of “picking out” is intentionally vague; this could be reference or it could be mere quanti cational description.) The kinds of object that a marker might pick out varies widely, potentially including landmarks, buildings, cities, rivers, patches of land, or swaths of geography. Not only do the two aspects of marker syntax express di erent kinds of content, but they do so in di erent ways. Marker types symbolically express their associated properties. That is to say, the semantic connection between marker types and properties ultimately depends on arbitrary juxtapositions of sign types with meanings. Yes, these associations are often motivated—blue is chosen for bodies of water, brown for land—

but the true semantic value is always xed by the list-like conjunction of marker types with properties rather than any quasi-pictorial interpretation. By contrast, the relative locations of markers on the map iconically express relative locations in the world, and these in turn are used to pick out the designated object. In other words, marker locations bear an intrinsic and natural relation to locations in the represented world. The connection between marker 21

locations and worldly location is not one of arbitrary juxtaposition; it is not symbolic.

We can put these observations together in the form of two further facts about map representation that a semantic theory must account for. Fact 3. Markers express properties by symbolic association of marker types with properties. Fact 4. Markers pick out objects for attribution by iconic association of the relative location of the marker on the map with relative locations of objects in the world. p. 499

Facts 3 and 4 work together: the properties of Fact 3 are attributed to the located objects of Fact 4. Much has 22

been made of the contrasts between maps and predicative language.

In these two facts, we see the depth of

both the similarities and the di erences. On one hand, unlike language, the very same syntactic element— the marker—both expresses a property and picks out an object, in part by exploiting iconic properties of the map. On the other hand, two di erent semantic mechanisms are responsible for contributing two bits of content—the property and the object. Thus the deeper structure of subject-predicate attribution is still preserved. In this way maps contrast with holistic theories of picture meaning, like that of Hopkins (1998, chap. 4), for example, where property and object are represented without di erentiation as part of a single situation, by a single semantic mechanism.

2.3. Tags Tags are symbols or symbolic expressions, including words and linguistic phrases, which are associated 23

with speci c subregions of a map.

Paradigm examples are names indicating the location and identity of

landmarks, but a variety of other subclausal phrases can be linguistic tags, including numerals, nouns, adjectives, as well as de nite and inde nite descriptions. (For now I’ll set aside indexical sentences, like “you are here,” which can also play a tagging-like role.) Nonlinguistic tags include a variety of specialized

graphical symbols, like ⌂, ✝ or ★, which may be listed in a legend, or conventional for a type of cartographic discourse.

In the excerpt from the Middlesex County map (Figure 22.8), for example, one nds names for towns and roads, a name for an entire railroad line, as well as numerals for local zip codes and a numeral for a local route.

Figure 22.8. Tagging as the annotation of a map with linguistic symbols. Source: Excerpt from Middlesex County Atlas (2002, 32–33). Much like markers, the role symbolic tags play is to contribute identifying and descriptive information p. 500

about particular objects for which the map provides spatial

information. An adequate semantics of

tagging must address three central facts about the use of tags in maps. Fact 5. The relative location of a tag on the map (partly) determines the relative location of the object in the world that it describes. Fact 6. There are a variety of ways, including placement at a variety of map locations, that a tag can determine the relative location of the object in the world that it describes. Fact 7. Tags vary widely in their semantic function, including predication, naming, and more. To illustrate these points, I’ll refer to maps of the landscape shown in Figure 22.9.

Figure 22.9. The landscape. Fact 5 has to do with the semantic signi cance of tag placement. The placement of tags within a map contributes to the content of the map by indicating which depicted objects are associated with the contents of the tags, with consequences for accuracy conditions. For example, relative to the landscape above, (3) is accurate. But swapping the position of “Eagle Hill” and “Turtle Pond” results in (4), which is not accurate

p. 501

(Figure 22.10). In

e ect, tags locate linguistically expressed properties within map space, just as markers

locate the properties they denote within map space.

Figure 22.10. Relative to the landscape, map (3) is accurate, but (4) is not. Yet the ultimate semantic contribution of tags can be achieved through a variety of expressive means; this is the point of Fact 6. In Figure 22.11, for example, the tag “Turtle Pond” is associated with the image by placement proximal to the part of the map it tags. “West River” is both proximal and aligned with the depiction of the river. “Black Mountain” is associated with a part of the image by spatial inclusion within it. 24

And “Little Mountain” and “Eagle Hill” are associated with the map through line linking.

(One can also see

numeral linking at the bottom of Figure 22.8.) Not only that, but the exact location of each tag is irrelevant to the content expressed. The content of Figure 22.11 is the same as that of Figure 22.12 even though most of the tags have been moved. Both the variety and exibility with which tags can be associated with map regions distinguishes them from markers, which 25

always comment on precisely the region of the map where they are located.

Figure 22.11. Tagging via line linking, inclusion, alignment, and proximity.

Figure 22.12. Local changes of tag position do not a ect content. The tension between Facts 5 and 6 can be resolved by recognizing that the various ways of tagging a region on a map all correspond to a single underlying relation of linking at the syntactic level. Inclusion, proximity,

26

alignment, and line-linking are all signals of the same syntactic relation.

On this model, the location of a

tag on the printed page is not itself part of syntax. Tags themselves have no location; they are associated with regions of the map plane by abstract syntactic links. Finally, Fact 7 highlights the variety of ways tags can work semantically. As Figure 22.13 demonstrates, nouns and names can both be tags, but because nouns and names denote objects in di erent semantic categories, the semantic signi cance of tagging must itself be allowed to vary. A landmark tagged with the p. 502

name “Black Mountain” expresses the content that the landmark is identical with ⟦Black Mountain⟧. An area

tagged with the noun phrase “marsh area” expresses the content that the area has the

property ⟦marsh

area⟧. More extreme variations are common. The numeral “46°” in Figure 22.13 expresses the content that

the tagged area has atmospheric temperature ⟦46°⟧. In Figure 22.8, the tag “08859” relies on the relation has the zip code.

Figure 22.13. Tags as names, as noun phrases, and as numerals for local temperatures. As these examples show, there is no single way to interpret a tag. Each tag is connected to the represented object by an implicit tagging relation. While the most common tagging relations are relations of naming and predication, a map-maker can introduce any number of others, like zip code or temperature, as convenience and recoverability allow. Since the operative tagging relation for a given tag is generally not explicitly 27

signaled, tagging relations are determined instead by context and local conventions.

Such relations appear

to be a species of the more general phenomena of multimodal coherence relations. These are structural links in a discourse that function to bind together independent elements from di erent modalities, such as 28

speech with iconic gestures, pictures with captions, or arrows with objects in a diagram.

Recognizing that

maps rely on the same kind of system of exible multimodal relations allows for a well-motivated and empirical adequate account of the semantics of tagging.

3. A Semantics for Maps A semantic theory for maps aims to explain how maps systematically express their cartographic contents. Braddon-Mitchell and Jackson (2006, 180–184), Camp (2007, 2018) and Blumson (2012), among others, have argued that map representation is in fact systematic in the way that invites semantic analysis. And a number of carefully worked out semantic theories for maps have already been developed by, for example, p. 503

Pratt (1993), Leong (1994), Casati and Varzi (1999, chap. 11), and Rescorla (2009). In

what follows, I hope

to build on these advances, while heeding the lessons about space, markers, and tags from the preceding discussion. Maps themselves are complex representational objects that function as the nexus point for a variety of interlocking interpretive rules and conventions. In this section I outline a semantic theory for maps that combines the key elements of pictorial projection, marker interpretation, and tagging and sets them within the common framework of possible world semantics. I again restrict my attention to the simplest of metric maps in planar projection, setting aside quasi-topological (subway) maps, globe projections, and the many, many sophisticated forms of marking and tagging that populate real-world cartography. Still, there are innumerably many planar, metrical map systems which vary in their speci c treatment of space, markers, tags, and more. There is no sense in attempting a semantics for all such map systems, or trying to formalize the key features that might unite them. Instead, I will develop a semantics for a particular, arti cially simpli ed or “formal” map system, what I will call map system S. My hope is that system S has enough in common with the natural map systems that a semantics for S may provide a template for other map semantics going forward.

3.1. Accuracy, Content, and Target 29

Maps are accurate or inaccurate relative to a map system, just as sentences are true or false in a language.

Whether a map is accurate relative to a system depends on the content the map expresses in that system and on how the world is. It also depends on what time and place within the world is supposed to be depicted by the map, and what position and orientation it is supposed to be depicted from. Collectively, these latter elements make up what we may call a viewpoint. A map is not simply accurate or inaccurate at a (possible) world, but at a world and a viewpoint. We can think of the core, cartographic content of a map, relative to a system, as the set of conditions under which it is accurate. In traditional linguistic theory, the content of a sentence is understood as the set of possible worlds relative to which that sentence is true. Since maps are accurate relative to a world and a viewpoint, we’ll treat the cartographic content of a map as the set of world-viewpoint pairs at which the 30

map is accurate. This is the innermost sphere of meaning for maps.

The content of a map as a whole can be factored into two aspects: singular content, which involves reference to particular objects; and attributive content, which consists of properties and relations that are attributed to objects, which are themselves either quanti ed or included in the singular content. Most of the cartographic content we care about is attributive: spatial relationships, types of geography, types of road, locations of borders. Explicitly represented singular content comes primarily from linguistic tags: names of cities, roads, and landmarks. There is the additional question, discussed in section 2.1, of whether maps also express indexical singular content—that piece of land, that road—even when they are not named or otherwise explicitly singled out. I am sympathetic with the idea that maps, and pictures generally, are vehicles of indexical reference in this p. 504

31

way.

Yet, as I argued above,

following Camp (2018), even if there is such singular content, this fact has

little bearing on the central question of how maps convey their spatial content. For the purposes of this analysis, I set aside the question of indexical singular content. Still, there is a clear sense in which a given map is, say, a map of Vermont, or a map of Vermont in 1800, or a map of your neighborhood following the proposed construction of the highway. These are descriptions of what we may call the target of a map. That target of a map is a particular world-viewpoint pair which the map is supposed to accurately represent. Often the map’s target is partially speci ed in a title or caption. The target of a map is not strictly part of its cartographic content but belongs to the sphere of meaning having to do with speech acts and discourse e ects. The content of a map is what the map says; its target is the situation in space and time that a given use of the map is supposed to say something about. When a map is intended to be accurate in the actual world, the target of the map consists of the actual world and a viewpoint located within the space of the actual world. Maps of past, present, and future geographies correspondingly have targets whose viewpoints correspond to di erent times in the actual world; 32

counterfactual and generic maps may have targets sampled from other possible worlds.

When maps do

target the actual world, they acquire a kind of particularity. They are directed at a speci c time and place. This goes a long way towards addressing the intuition that maps, by nature, are representations of 33

particular locations.

3.2. Viewpoints and Viewpoint Constraints A map is accurate at a given world only relative to a viewpoint: an oriented position in space and time. A viewpoint, in this sense, is not literally a point, but a complex structure which gives maps spatial and temporal perspective, and can be de ned, relative to a world, by a series of interlocking elements, 34

illustrated in Figure 22.14. p. 505

(5)  Elements of viewpoint (i) the position and shape of a projection source, relative to a world w; (ii) the position and shape of a cartographic plane, relative to the projection source; (iii) a bottom-to-top vector on the cartographic plane; (iv) a back-to-front vector on the cartographic plane: (v) a single temporal location for all these elements. Item (i), the projection source, establishes the position and orientation of the viewpoint. For the planar projection maps considered here, the projection source itself is a plane, while for linear perspective 35

projections, it would be a point.

Item (ii) gives the relative position of the cartographic plane, the surface

onto which features in the world are ultimately projected to constitute a map. When considering a particular map, the cartographic plane will always be proportional to the map itself. Elements (iii) and (iv) establish the orientation of the cartographic plane, and (v) identi es a temporal index for all.

Figure 22.14. Elements of viewpoint (shown from below). Generally, maps are to be understood as depicting the world from above, as de ned by the direction of gravity. There are also norms of orientation: customarily the top of the page is aligned with North, but other orientations can be achieved with a compass rose. We can understand these norms as providing conventionalized restrictions on the relationship between the world depicted and the position and orientation of the viewpoint from which it is depicted. Such viewpoint constraints amount to interpretive assumptions that help determine the content of a map, but do not depend on the map’s speci c 36

composition.

(6)  Cartographic Gravity Constraint G-Con(w, v) i

the back-to-front vector of v is anti-directional with the direction of Earth’s gravity

37

at v in w.

(7)  North Constraint N-Con(w, v) i

the bottom-to-top vector of v is co-directional with the south-to-north direction of

Earth in w, when projected to v. (8)  Compass Rose Constraint C-Con(w, v) i

the S-to-N vector of the compass rose in M, embedded at v, is co-directional with the 38

south-to-north direction of Earth in w, when projected to v.

I call (6) the Cartographic Gravity Constraint to distinguish it from the Gravity Constraint of ordinary pictures, where the top-to-bottom vector of the picture is normally aligned with the direction of gravity. The cartographic constraint holds for nearly all terrestrial maps, but alternative conventions that prioritize p. 506

other vectors will hold, for example, for maps of space stations. The North Constraint and the Compass Rose Constraint can come into con ict. In modern map-making, as I discussed in section 1, the North 39

Constraint tends to hold by default, but it is overridden by the presence of a compass rose.

 Camp (2018,

28–29) introduces the idea of an “unanchored” map, such as a plan for a possible estate with no settled building site, or a template for the layout of a chain of drug stores; in such cases, it seems that no orientational constraints apply. What viewpoint constraints apply to a given map can be signaled by an explicit sign (like a compass rose), by general or local convention, or by ad hoc indicators of the map-maker’s intention. For convenience, I’ll treat the operative viewpoint conventions as part of the current map system, allowing that they will vary

from system to system. The whole set of available viewpoint constraints may be de ned by the broader network of interpretive conventions from which particular map systems emerge. Although viewpoint constraints apply independently of the composition of the map itself, they are an important part of the apparatus of interpretation. They play an essential role in narrowing and xing cartographic content.

3.3. Projection Semantics In this section, I outline an approach to cartographic semantics that enlists geometrical projection, rather than metric isomorphism, as its organizing principle. Geometrical projection describes a family of methods for mapping a 3D scene onto a 2D surface, via a system of straight lines that are structured by a viewpoint. Familiar forms of projection from the study of depiction include perspective projection— where the lines of projection converge on a point—and parallel projection—where the lines of projection run orthogonal to a plane. More exotic projections are involved in large-scale cartography because of the challenges of mapping the sphere of the globe onto the 2D plane. In any case, we may de ne a projection function proj which takes possible worlds and viewpoints within those worlds as arguments and returns 2D projections as values. The idea behind a projection semantics is that, for a picture to be accurate at a world and a viewpoint, it 40

must be a projection of that world, from that viewpoint.

To work out the content of a picture is to

determine the kinds of worlds and viewpoints from which it could have been projected. Thus the content of a picture can be thought of as all of the world-viewpoint pairs from which the picture can be projected. Schematically, where P is a picture and ⟦P⟧ is its content, such a semantics takes the following form: (9)  Projection Semantics

a.  P is accurate at w, v i

p. 507

b. ⟦P⟧=

proj(w, v) = P

{⟨w, v⟩| proj(w, v) = P }

To adapt a projection semantics for maps, we will have to get into the de nition of projection a bit. There are two basic steps. First, a viewpoint is positioned in space relative to the world. This induces a set of projection lines, which extend from the viewpoint, through the cartographic plane, into the world. This e ectively sets up a correspondence between points in the world to points on the cartographic plane. In the method of projection illustrated in Figure 22.15, the projection lines are all orthogonal to a planar projection source and orthogonal to the cartographic plane. I assume a parallel projection like this for maps throughout this essay. In perspective projection, by contrast, the lines of projection converge on a point.

Figure 22.15. Construction of a map projection. Second, features in the world are mapped to features on the cartographic plane, following the lines of a projective correspondence. If, for example, a point in the world belongs to a body of water, then the corresponding point on the cartographic plane might be assigned the feature blue, and forests in the world might be paired with green. This projection of world features to map features is carried out for every line of projection. Conceived of this way, projection is a method for taking a region of the world, ranging over all of the basic features there, and transmitting them to the map. Of course, any number of di erent schemes for associating world-features with map-features are possible. To use projection as an interpretive principle, we reverse engineer this process. Following the two basic steps involved in projection, a projection semantics can be divided into two components. First the projection semantics proper tells us how we should interpret the spatial distribution of markers within the space of the map. Second, a marker semantics assigns meanings to each of the markers. Dividing the semantics into these two elements provides for a more compositional semantics, and more modular analysis. This, in e ect, is the plan for what follows. Before moving to state the projection semantics, let us rst de ne a minimal syntax for maps without tags, p. 508

what I’ll call a pure map. For present purposes, a pure

map is a 2D plane segment where distinguished

41

regions are associated with marking types. (10)  Syntax of pure maps in S

A structure M =⟨P , d, R, mark, V ⟩ is a pure map in S i : (i)  P is a set of points; (ii)  d is a Euclidean metric over P which de nes a 2D rectilinear space; (iii)  R is a set of signi cant regions of P; (iv)  mark is a function from R to sets of marking types; (v)  V is a pair of orientation vectors on P: bottom-to-top and left-to-right. Here P is the set of points from which the map is composed, and d supplies its basic spatial structure. R isolates those regions of the map which are semantically signi cant; I’ll assume that the regions in question are normally contiguous, but possibly overlapping, and correspond to psychologically natural 42

segmentations of the graphical space.

Next, mark associates each semantically signi cant region with a set

43

of marker types.

Finally, V gives the intrinsic orientation of the map (regardless of how the physical 44

vehicle is pinned to the wall).

To think of the pure map as a projection of a worldly scene, one has to put objects in the world into geometrical relations with the map itself. This requires notionally embedding the map in the space of the world being projected, where the embedded map is scaled up to match the size of the geography being 45

projected.

This is sensible if one keeps in mind that a “map” here is an abstract object constructed around

a 2-dimensional space. I’ll say that the embedding of a map M at world w and viewpoint v, written Mw,v, is identical to M except that the points that make up map (P) belong to the space of the world w. The internal 46

structure of the viewpoint v determines the scale, position, and orientation of the embedded map. We can now state the projection semantics, where markersS is de ned as the domain of ⟦•⟧M. (11)  Projection semantics for pure maps in S M is accurate at w, v in S i : for every region r in Mw,v: there is an object o in w such that: (i) for all projection lines ℓ from v: ℓ intersects r ↔ ℓ intersects o; (ii) for all τ ∈ markersS: τ ∈ mark(r) → ⟦τ⟧M (o) at w, v.

The semantics says that a map M is accurate at world w and viewpoint v if and only if, when M is embedded at w, v, every region in M corresponds to an object o in the world in the following way. (i) First, for any ray from v, it intersects with r when and only when it intersects with o. This sets up a spatial correspondence between regions in M and objects in w. (ii) Second, for any marker type available in S, if r has that marker type, then o has the property which is the denotation of that marker type. p. 509

The projection semantics has the same basic structure as the map semantics described by Leong (1994), Casati & Varzi (1999), or Rescorla (2009), since it divides labor between a geometrical constraint between map space and world space, on one hand, and interpretation of marker types, on the other. The virtue of the projection semantics is that it preserves correspondence of metrical relations in the XY dimension, without committing the content of map with respect to relations of depth in the Z dimension. Ultimately, projection semantics is a natural t for maps, not only because maps seem to extend the pictorial way of representing the landscape, but also because it makes sense of the facts about cartographic space reviewed in the last section. Of course, what I have described here is a very simple projection semantics for a very basic map system. There are in nitely many other potential methods of projection, and much of modern cartography has been devoted to the formal study of existing methods and the invention of new ones.

3.4. Marker Semantics The projections semantics has to be supplemented with a semantics for marker types. For the sake of illustration let us suppose that system S has ve marker types: (12)  Marker types in S

The marker semantics itself has the same structure as a linguistic lexicon. It takes the form of a list-like pairing of marker types with properties. We can de ne the interpretation of marker types in S as follows, where ⟦•⟧M is the specialized interpretation function, in S, for marker types: (13)  Marker semantics for S

In fact, this statement of the marker semantics is only approximate, because it sidesteps the crucial role of p. 510

viewpoint. For example, ▭ doesn’t just mean that there is unforested land in a given direction; after all, wherever there is unforested land, there

may be forested land, in the same direction, on the opposite side

of the earth. Instead, ▭ means, more nearly, x is part of unoccluded unforested land in w relative to v. And occlusion, in turn, isn’t de ned optically, but in terms of projection lines uninterrupted by land masses (since atmospheric phenomena might be optical occluders, but are not relevant). So, a more complete version of (i) above might be: (14) ⟦ ▭ ⟧ M = λx.λw.λv. x is unforested land in w and there is a projection line from v that intersects x without intersecting any other land masses.

In what follows I gloss over the important complications introduced by the viewpoint relativity of marker denotations.

3.5. Tagging Semantics 47

The next issue we must address is the interpretation of the symbolic tags which annotate a map.

As I

discussed in section 2.3, part of the challenge here is nding the right division of labor between syntax and semantics. I concluded there that tags are associated with regions of the cartographic plane by abstract 48

structural links.

The links which connect tags to regions need not be explicitly marked on the printed

page, but they are nevertheless signaled through a variety of means, as we saw in the initial discussion. The nal ingredient in the syntax of tagged images addresses the variety of tagging relations. I posit a set of relation-symbols which are, again, explicit in the syntax, but implicit on the printed page. Each link between a symbol and a region is associated with one such symbol. Formally, I treat the tag function as a mapping from picture regions to sets of pairs of tag symbols and relation-symbols. We can capture the core syntactic idea formally by adding the following structure to our de nition of a pure map, where R is the set of signi cant regions: (15)  tag is a (partial) function from R to the set of pairs 〈ρ, ϕ〉 where ρ is a relation-symbol, and ϕ is a tag. To incorporate tagging into our semantics for pure maps, we’ll add the following clause, where ⟦•⟧R is the

interpretation function for relation-symbols; ⟦•⟧L is the interpretation function for linguistic tags; and tagsS is the domain of ⟦•⟧R × the domain of ⟦•⟧L. Here r is a variable ranging over picture regions, o is a variable ranging over worldly objects.

(16) For all 〈ρ, ϕ〉 ∈ tagsS: 〈ρ, ϕ〉 ∈ tag(r) → ⟦ρ⟧R (o, ⟦ϕ⟧L) at w, v.

Clause (16) states that, for every possible relation-tag pair, if a region r is linked with that pair, then the p. 511

corresponding property is attributed to object o. The property in question is not simply ⟦ϕ⟧L, the content of the tag, but rather ⟦ρ⟧R (x, ⟦ϕ⟧L), the content

of the tag as it is modulated by the content of the relevant

tagging relation. I’ll represent identity and predication, the two most common tagging relations, by the relation-symbols as “id” and “pred” respectively, and the more specialized relation zipcode as “zip”: (17)  Relation-symbol semantics in S a. ⟦id⟧R = λx.λy.x = y

b. ⟦pred⟧R = λF.F

49

c. ⟦zip⟧R = λx.λz. the zipcode for region x is z

For example, if the tag is “Wantastiquet Mountain” and the relation is “id”, then, by (16) the semantic contribution of the tag can be derived as follows: (18) a. ⟦id⟧R (o, ⟦Wantastiquet Mountain⟧L) at w, v

b. (λx.λy.x = y) (o, m) at w, v c.  o = m at w, v

If the tag is “08816” and the relation is “zip,” the result is: (19) a. ⟦zip⟧R (o, ⟦08816⟧L) at w, v

b. (λx.λz. the zipcode for region x is z) (o, 08816) at w, v c. the zipcode for o is 08816 at w, v The semantics for tags can now be integrated into the projection semantics for pure maps, to yield a more nearly adequate semantics for a complete map system. This will have the desired e ect of allowing the content of symbolic tags to enter into the content of the tagged image at speci c, object-dependent locations in pictorial space.

3.6. Map Semantics We are nally in a position to compile the forgoing elements into a uni ed semantics for maps in the simpli ed map system S. Let us begin with syntax: (20)  Syntax for map system S A structure M =⟨P , d, R, mark, tag, V ⟩ is a map in S i : (i)  P is a set of points; (ii)  d is a Euclidean metric over P which de nes a 2D rectilinear space; (iii)  R is a set of signi cant regions of P; (iv)  mark is a function from R to sets of marker types; (v)  tag is a (partial) function from R to pairs 〈ϕ, ρ〉 of relation-symbols and tags;

(vi)  V is a pair of orientation vectors on P: bottom-to-top and left-to-right. p. 512

We are after a semantics for maps that is uni ed, but not uniform: as we’ve seen, maps bring together a diverse array of representational resources. The semantics for a system like S are de ned in terms of a series of interpretive sub-systems, including a method of projection for the pure map, a natural language semantics for linguistic tags, and denotational assignments for marking types, for non-linguistic tags, and for tagging relations. I’ll notate the various interpretation functions used in the nal semantics as follows: (21)  Interpretation functions • ⟦•⟧S … maps

• ⟦•⟧M … marker types

• ⟦•⟧L … linguistic tags

• ⟦•⟧R … tagging relations

Then we may think of the map system, S, as a structure containing all of these interpretation functions as components, in addition to a set of viewpoint constraints. (More elements would likely have to be added in a more fully formal treatment.) (22)  Map system S S =⟨  M, ⟦•⟧S, ⟦•⟧M, ⟦•⟧L, ⟦•⟧R, vp-constraintsS, …⟩

We can now state the semantics of S as follows, where markersS is de ned as the domain of ⟦•⟧M, and tagsS is 50

the domain of ⟦•⟧R × the domain of ⟦•⟧L. (23)  Semantics for map system S

For any map M in S: ⟦M⟧S = the set of all ⟨w, v⟩ such that:

for every map region r in Mw,  v: there is an object o in w:

(i) For all projection lines ℓ from: ℓ intersects r ↔ ℓ intersects o; (ii) For all τ ∈ markersS: τ ∈ mark(r) → ⟦τ⟧M(o) at w, v;

(iii) For all 〈ρ, ϕ〉 ∈ tagsS: 〈ρ, ϕ〉 ∈ tag(r) → ⟦ρ ⟧R (o, ⟦ϕ⟧L) at w, v; (iv) For all δ ∈ vp-constraintsS: δ (w, v).

(24)  Accuracy for map system S M is accurate at w, v, in S i

⟨w, v⟩

∈ ⟦M⟧S.

Clause (i) of (23) is the key projection clause, establishing a spatial correspondence between regions of the map’s cartographic plane and spatially located objects in the world. Clause (ii) requires that, for every region decorated with a given marker, its projective counterpart in the world instantiates the property which is the denotation of that marker type. Clause (iii) likewise requires that a region’s projective counterpart instantiates the property which results from interpreting the associated tag and tagging p. 513

relation. Finally,

clause (iv) ensures that the map’s content conforms with the system’s operative

viewpoint constraints. The resulting statement of cartographic semantics is compact, but it has none of the minimalist elegance of a Tarskian theory of truth. So be it: interpretive elegance went out the window with the invention of maps. Maps are not austere registers of homogenous information. They are maximalist, doing as much with as

they can within the con nes of the basic spatial substrate. Future semantic theories for naturalistic map systems, I am sure, will require the addition of further clauses and constraints, and of even greater complexity and variety than those envisioned here.

3.7. Exhaustivity There remains one issue in the semantics of maps which demands comment in part because it is the subject 51

of such extensive discussion in the literature.

This the question of exhaustivity: whether maps should be

interpreted as providing complete and exhaustive representations of their subject matters, or, whether they may be read as supplying merely partial or incomplete representations. Practically speaking, the question is: If a given marker type τ is absent at a given region on a map, does this imply that there is no object which is ⟦τ⟧ in the corresponding region of the world? According to the so-called absence intuition, which endorses 52

exhaustivity, the answer is yes.

The alternative is that the absence of τ merely indicates the map’s

quiescence on the question of the whether ⟦τ⟧ is realized in the corresponding region.

While some have thought that all maps, by de nition, are exhaustive, the truth is almost certainly more 53

complex than that.

Expectations of exhaustivity seem to vary by type of map, type of marker, subject

matter, and even among regions on a single map. At one extreme, in a mass-produced topographical map intended for hikers or surveyors, the absence of a depiction of a body of water almost certainly entails the absence of a corresponding body of water. At the other extreme, an informal map drawn on a napkin to given directions to the nearest gas station is free to include only landmarks the map-maker feels are relevant, without risking the representation of the absence of others. To a rst approximation, the assumption of exhaustivity seems to vary with the register—the degree of formality and implied rigor—of the map itself. Exhaustive maps appear to be the special provenance of mass-production in the age of information, where it is commonly assumed that any feature which can be surveyed, will be. Modern map consumers expect print and digital maps to possess a kind of encyclopedic authority. There remains the further question of whether exhaustivity should be viewed as a semantic assumption, built into cartographic content, or whether it belongs to the realm of discourse e ect or to classical 54

pragmatic implicature.

I won’t take a stand on this issue here, or try to improve on the empirical

generalization about when and where exhaustivity applies. Instead, I want to show how the availability of p. 514

exhaustive and

non-exhaustive interpretations of markers can be explained within the framework of

projection semantics. First, it should be noted that the projection semantics articulated here is non-exhaustive. The key formulation comes in the clause for the interpretation of markers: (25) For all τ ∈ markersS: τ ∈ mark(r) → ⟦τ⟧M(o) at w, v.

This clause requires that, for any marker type τ on any region r of the map, the object in the world which is the projective counterpart of r must have the property ⟦τ⟧. This is non-exhaustive, because it allows that there may be objects which have the property ⟦τ⟧ for which there is no corresponding region with τ.

Rescorla (2009, 181–196), along with Casati and Varzi (1999, 192–194), achieve exhaustivity by adding an additional clause to their semantics. The primary clause says that regions with a given marker τ must denote locations that have the property ⟦τ⟧; the supplementary clause says that all regions without τ must denote

locations that do not have the property ⟦τ⟧. In the semantics presented here, the same e ect is achieved by 55

strengthening the marker clause above to a biconditional. (26) For all τ ∈ markersS: τ ∈ mark(r) ↔ ⟦τ⟧M(o) at w, v.

What is interesting, in the context of a projection semantics, is that exhaustivity, as realized by the strengthened marker clause, really signals a return to the original understanding of projection. According to this conception, a method of projection impartially surveys features of the world and transforms them into image features. If we assume that a map is accurate only if it is a projection of the world, then an accurate map is an exhaustive record of the projectable features of the world. Exhaustivity just amounts to the assumption that the map really is the result of applying projection to the world. Here we might distinguish between forward-projection and reverse-projection. Forward-projection is an algorithm that impartially surveys features of a world, relative to a viewpoint, and returns a constraint on images—what an image must include to be a projection of that world. Reverse-projection is algorithm that impartially surveys features of an image and returns a constraint on a world and viewpoint—what the world must include to have projected that image. Both rely on the same projective geometry, but di er with respect to how they treat features of the world and the map. Forward-projection implies an exhaustive coding of world features; reverse projection implies an exhaustive de-coding of map features. In its loosest formulation, projections semantics says that an image is accurate at a world and viewpoint if and only if it is a projection of that world and viewpoint. We now see that the projection semantics I’ve developed here relies on reverse-projection, not forward-projection. It uses space projectively to establish the alignment of map regions and objects in the world. But markers on the map are surveyed impartially, and imposed back on the content, not vice versa. This was precisely to avoid the implication of exhaustivity. p. 515

Exhaustivity arises when it is assumed that maps are also forward projections of their subject matters: they are the result of impartially recording all relevant world features onto the cartographic plane. Such an assumption might be modulated for certain kinds of features, or for certain regions of a map; and a variety of semantic or discourse-level mechanisms might e ect such a modulation. What this suggestion illuminates is the nature of exhaustivity itself. It is not the result of an arbitrary supplemental constraint on map interpretation, nor a generalized pragmatic implicature, nor a principle of informational conservativity. Instead, it follows from the nature of map interpretation; it is a consequence of a strengthened understanding of the projective correspondence of map and world.

4. The Geography of Meaning I began this chapter by isolating the narrowest sphere of literal cartographic meaning as my subject matter. It is here, arguably, that maps are most clearly distinguished from other forms of representation. I went on to examine the roles that space, viewpoint, markers, and tags all play in the determination of a map’s total cartographic content. Finally, I attempted to provide, at least in outline, a projection-based semantics of maps that integrates and explains these factors. The value of such an investigation is not just to understand maps in themselves, but to understand how they relate to the broader geography of meaning. Typologies of representation which focus exclusively on the format of representational vehicles, or on the kinds of contents they express, miss out on what is arguably the central feature of any system of representation: the semantics by which it associates signs with contents. The deeper contours of the semiotic terrain are not constituted by what is represented, but by how it is represented. In comparing maps to language, scholars have noted that both are systematic and productive, but there has been debate about the extent to which maps employ predication, and further debate about the extent to 56

which they are propositional.

These are substantive philosophical questions. But I think we can cut to the

chase, to some extent, by directly comparing semantic theories for each type of representation.

With respect to their most basic constituents, semantics for map systems and languages have clear a

nities. The semantics of map markers have essentially the same structure as linguistic lexicons. And

tags, often enough, just are bits of language. Where the two kinds of representation truly pull apart is in the way that complex structures composed from basic elements are interpreted. Map semantics are based on natural spatial transformations of the cartographic plane. Projection from a viewpoint, applied uniformly to every point on the map, gives structure to the attribution of markers and tags. There are no such natural or uniform transformations in the semantics of (non-iconic) language. The p. 516

meanings of complex linguistic expressions are derived from the recursive application of composition rules to tree-like structures with lexical items at the nodes. The composition rules themselves assign structure types to logical operations, like function application, without regard to those structures’ spatial or dimensional properties. This, I believe, is the kind of contrast Camp has in mind when describing her requirements for an adequate semantics of maps: “rather than employing a combinatorial operation that is digital, universal, and asymmetrical, we need one that is holistic, spatial, and symmetrical” (Camp 2018, 41). With a semantics of maps in hand we can also begin to get a sense of where maps stand in relation to the surrounding terrain: not just of the distance between maps and language, but also of the more proximal representational forms. For example, previous isomorphism analyses suggested that maps belong to the same family as other isomorphism-based diagrams, like Venn and Euler diagrams. The present account, by contrast, suggests that maps are much more nearly like pictures in the way they structure space. What they have in common with diagrams is the more general feature of basing their semantics on a natural relation or transformation. The other thing that maps have in common with diagrams is the way they e ortlessly blend symbolic and iconic interpretation. In this respect, they join the wider world of multimodal discourse, where gestures, facial expressions, pictures, and diagrams are brought together with speech, annotations, labels, titles, and symbols. This, of course, is the natural setting of human communication. Maps exemplify this combinatorial instinct by bringing spatial transformations, symbolic markers, and multimodal discourse relations together into a uni ed cartographic arena. In the geography of meaning, maps are a crossroads.

Acknowledgements Thanks to Tendayi Achiume, Malihe Alikhani, Joshua Armstrong, Sheridan Bartlett, Elisabeth Camp, Sam Cumming, Catherine Hochman, John Kulvicki, Kevin Lande, Atsushi Shimojima, and Matthew Stone for discussion leading to the ideas presented here.

Notes 1.

By “First World” nations I mean to refer to former colonial powers, and by “Third World” nations I mean formerly colonized nations; see Achiume (2019, 1513–1514).

2.

This spherical model of meaning isnʼt intended to be an exact theory, much less an original one. Compare Panofskyʼs (1939, 5–8) more carefully worked out levels of iconology for pictorial art.

3.

See the frameworks described by Murray and Starr (2018, 2021) and Lepore and Stone (2015, chap. 14) for the scope of the discourse e ect and the kinds of mechanisms that govern it.

4.

The present taxonomy does not follow the classical division of communicative labor into semantics and pragmatics. Here I am guided by the arguments of Lepore and Stone (2015) to the e ect that the traditional domain of pragmatics can be factored without remainder into discourse e ects, on one hand, and ad hoc psychological and social engagement, on the other.

p. 517

p. 518

5.

See e.g. Wood (1992, chap. 1); Aguilera (2021).

6.

See extensive analyses of map connotation in e.g. Wood (1992, chap. 5) and MacEachren (2004, 330–51).

7.

Compare Taylor (forthcoming) on representations of social identity in the art world.

8.

See Wood (1992, chap. 2) and Harley (1988) for vivid studies of the political meaning in maps.

9.

Danforth (2014) and Williams (2016). Meier et al. (2011) provides psychological evidence for value associations with higher and lower spatial positions on a map.

10.

Crampton (1994).

11.

Hall (1992, 380); Wood (1992, 56–61).

12.

See, for example, work on diagrams (Shin 1994; Shimojima 2015), on speech with iconic gesture (Lascarides and Stone 2009a, 2009b), on images with linguistic captions (Alikhani and Stone 2019, 2018b), on iconic use of classifier constructions in sign language (Davidson 2015, 491–498), and iconic variables and predicates in sign language (Schlenker, Lamberton, and Santoro 2013, 103–120; Schlenker 2018). See Schlenker (2019) for an overview of recent work.

13.

This is a widespread but not universally shared view. See Lopes (1996, 94) or Kulvicki (2015, 149) for sympathetic statements. See Dilworth (2002, 191–192) for discussion.

14.

See e.g., Casati and Varzi (1999, 187): “Maps display no shortenings, no perspective e ects: a map can be conceived as close as it gets to a view from nowhere about the area it depicts.”

15.

Casati and Varzi (1999, 187).

16.

See Camp (2018, 36–41) for an extended discussion of this theme.

17.

Kulvicki (2020, 120) discusses and endorses this approach; Casati and Varzi (1999, 191) is a prominent example. See Camp (2018, 28–31) for a critical discussion.

18.

Likewise, 39: “Spatial relations in the map represent spatial relations in the world.” See Camp (2007, 161) for more on spatial relations in maps.

19.

Rescorla (2009, 178) o ers an explicit formulation much like the one in the text, and points to Russell (1923, 89) as an antecedent. Casati and Varzi (1999, 195) make topological isomorphism of the denotation function a prerequisite to semantic evaluation, rather than an explicit constraint on truth. (They acknowledge the need for metric constraints of the same kind on p. 196.) See Camp (2018, 39–40) for objections to Casati and Varziʼs way of building isomorphism into the truth conditions. Camp herself endorses something like an isomorphism semantics on p. 41.

20.

Leong (1994, 138) resolves the problem by explicitly idealizing all geographies as flat projections.

21.

The concepts of iconic and symbolic representation invoked here are developed in more detail in Greenberg (2021b).

22.

See e.g., Camp (2007); Rescorla (2009); and Camp (2018).

23.

Compare Leong (1994, 64–71), and Kulvicki (2015, 169–171) on labels and annotations in maps. The discussion here follows Greenberg (2019), which provides a general analysis of tags in maps and other visual media.

24.

Alikhani and Stone (2018a, 3555) call this indication.

25.

Leong (1994, 65) puts this point in terms of identity conditions of the map: local changes of tag location result in the same map.

26.

Compare Alikhani and Stone (2018b, 273). See Greenberg (2019) for discussion of how the di erent signals of linking should be resolved.

27.

See Armstrong (2016) for an account of local convention. Despite the flexibility, there are constraints on how tagging relations may be expressed. Identity and predication appear to be defaults. Other relations are inferred when these defaults are incoherent or otherwise ruled out in context. A further constraint is typographic consistency: tags with the

same typographic features are expected to encode the same tagging relation. Note the use of plain text versus italics in Figure 22.13.

p. 519

28.

On coherence relations for speech with gesture, see Lascarides and Stone (2009a, 2009b); for pictures with captions, see Alikhani and Stone (2018b, 2019); for arrows in diagrams, see Alikhani and Stone (2018a).

29.

See e.g., Leong (1994, 153–154) and Rescorla (2009, 178).

30.

See Blumson (2009) and Greenberg (2021a, sec. 1) for parallel accounts of pictorial content.

31.

See Lopes (1996, chap. 5) and Greenberg (2018) for defenses of this kind of view for ordinary pictures, and Kulvicki (2020, 120–123) for maps.

32.

See Cummins (1996); Greenberg (2018); and Pavese (2020) for development of the theory of target.

33.

Kulvicki (2020, 119–122) defends this intuition.

34.

See Ross (1997, chap. 5); Blumson (2009); and Greenberg (2021a, sec. 1) for parallel discussion of viewpoint in pictures.

35.

Globe projections require more elaborate notions of viewpoint. For example, the projection source might be located at the center of the globe, in front of, rather than behind the cartographic plane; and projection itself might take place in several discrete steps.

36.

Cumming, Greenberg, and Kelly (2017) and Cumming et al. (2021) identify a range of viewpoint constraints at work in film; they appear to be endemic to visual media. Dilworth (2002) discusses similar constraints for pictures (185–186) and maps (198–199), though Dilworth believes that, in general, map content depends on the physical orientation of the map, not its intrinsic orientation. See Leong (1994, 154) for a version of the North Constraint.

37.

When taking a projection of the ground, the map as read is facing up, away from the ground. Thus, the back-to-front vector of the cartographic plane faces away from (is antidirectional with) the direction gravity.

38.

The Compass Rose constraint requires a notional embedding of the map, including the compass rose itself, at the viewpoint; more on this idea in section 3.3.

39.

There are local exceptions to this, such as New York City street and subway maps, in which top-to-bottom is conventionally aligned with the lower-Manhattan to upper-Manhattan vector. Other default orientations were common in the precolonial era (Danforth 2014).

40.

For developments of projection semantics for pictures, see Greenberg (2011, 2013); Abusch (2015); Greenberg (2021a). Earlier analyses of depiction in terms of projection include Hopkins (1998, chap. 3); Kulvicki (2006, chap. 3); and Hyman (2006, 5).

41.

See Leong (1994, 41) and Kulvicki (2015, 151) for the distinction between a mapʼs syntactic structure and its physical form.

42.

The semantic function of image subregions is anticipated in Abuschʼs analysis of visual co-reference (Abusch 2012, secs. 3– 5; 2015, sec. 4). Iʼll assume here that the shapes of the regions in question are perfectly definite, though this is certainly an idealization in many cases.

43.

This definition of map syntax allows for many parameters of variation, but imposes few limits. A more complete and explanatory account would put substantive constraints on this structure. For example, Kulvicki (2015, 153–159) has proposed that marking types are divided into incompatibility classes, such that only one type per class can be associated with any given region. Leong (1994, 62) distinguishes basic categories of marker types—line, area, spot, and boundary— and requires (138) that certain marker types can only be paired with certain categories (e.g., dotted lines with lines, and colors with areas).

44.

Dilworth (2002, 3; 2003, 41–42) defends the idea that intrinsic orientation is a constitutive element of a picture.

45.

The embedded map, then, is much like the 1:1 scale map of Borgesʼs (1998) fable, except the fact that it is abstract would protect it from the inclemencies of weather.

46.

The orientation vectors of M match the orientation vectors of v following the right-hand rule.

47.

The syntax and semantics described in this section are essentially like those introduced by Leong (1994, 64–71, 154), except that Leong does not make provision for tagging relations. The key elements of the present analysis were originally presented in Greenberg (2019).

48.

Such links are a subsegmental variant of the text-to-image links posited by Alikhani and Stone (2018b).

49.

I assume that zip codes, as written, are interpreted, but they are not treated as numerals, but rather as numeral sequences; their contents are number sequences.

50.

The present analysis follows the precedent of Leong (1994, 154), whose semantics brings together a similar set of interpretive clauses.

51.

See, e.g., Pratt (1993, 81–83); Leong (1994, 145–153); Casati and Varzi (1999, 192–194); Rescorla (2009, 181–196); Blumson (2012, 428–430); Kulvicki (2015; 2020, 123–126); Bronner (2015); and Camp (2018, 31–35).

52.

Rescorla (2009, 181–183).

53.

Rescorla (2009, 181–196), following Casati and Varzi (1999, 192–194), defends the claim that exhaustivity is a definitional feature of maps. But other commentators have consistently drawn attention to counterexamples, e.g., Leong (1994, 149); Bronner (2015, 45); Kulvicki (2015, 152); and Camp (2018, 33). Work by actual cartographers does not seem to support universal exhaustivity either. See, e.g., Robinson (2019); Wood (1992, 79–94).

54.

Casati and Varzi (1999, 192–194), Rescorla (2009, 181–196), and Kulvicki (2020, 123–126) modeling develop semantic accounts; Pratt (1993, 81–83) and Leong (1994, 153–156) o er broadly discourse-level proposals; Blumson (2012, 428– 430), Bronner (2015), Camp (2018, 31–35) advocate for pragmatic explanations.

55.

Kulvicki (2015, 153–159) o ers an alternative analysis, also compatible with the semantics presented here, where exhaustivity is achieved through a refined theory of markers. Kulvicki divides marker types into “incompatibility classes:” one and only marker from each class can be associated with a given region. Thus a given region might be represented both as forest (green) and high altitude (textured lines), but not as both forest (green) and land (white). Where there is no visible marker for a given class, Kulvicki posits implicit “zero values.” Zero-values are markers that are interpreted as denoting the property of instantiating none of the other properties denoted by the same incompatibility class. Where 0 is a zero-value from incompatibility class Π: ⟦0⟧M = λx.λw.∀χ ∈ Π: χ ≠ 0 → ¬(⟦ χ ⟧M (x) at w).

p. 520

This is an unusual denotation, to be sure, but an interesting analysis of the problem. 56.

See e.g., Camp (2007); Rescorla (2009); Blumson (2012); Grzankowski (2015); Camp (2018).

References Abusch, Dorit. 2012. “Applying Discourse Semantics and Pragmatics to Co-reference in Picture Sequences.” Proceedings of Sinn und Bedeutung 17, 9–27. Google Scholar WorldCat Abusch, Dorit. 2015. “Possible Worlds Semantics for Pictures.” In Blackwell Companion to Semantics, edited by Lisa Mathewson, Cécile Meier, Hotze Pullman, and Thomas Ede Zimmermann. Forthcoming. New York: Wiley. Google Scholar Google Preview WorldCat COPAC Achiume, E Tendayi. 2019. “Migration as Decolonization.” Stanford Law Review 71: 1509–1573. Google Scholar WorldCat Aguilera, Mariela. 2021. “Heterogeneous Inferences with Maps.” Synthese 199 (1): 3805–3824. Google Scholar WorldCat Alikhani, Malihe, and Matthew Stone. 2018a. “Arrows are the Verbs of Diagrams.” In Proceedings of the 27th International Conference on Computational Linguistics, 3552–3563, edited by Emily M. Bender, Leon Derczynski, Pierre Isabelle. Sante Fe, New Mexico: Association for Computational Linguistics. Google Scholar Google Preview WorldCat COPAC Alikhani, Malihe, and Matthew Stone. 2018b. “Exploring Coherence in Visual Explanations.” In 2018 IEEE Conference on Multimedia Information Processing and Retrieval (MIPR), 272–277. IEEE. Google Scholar Google Preview WorldCat COPAC Alikhani, Malihe, and Matthew Stone. 2019. “ ʻCaptionʼ as a Coherence Relation: Evidence and Implications.” In Proceedings of the 2nd Workshop on Shortcomings in Vision and Language, 58–67, edited by Ra aella Bernardi, Raquel Fernandez, Spandana Gella, Kushal Kafle, Christopher Kanan, Stefan Lee, and Moin Nabi. Association for Computational Linguistics. Google Scholar Google Preview WorldCat COPAC Armstrong, Josh. 2016. “The Problem of lexical Innovation.” Linguistics and Philosophy 39 (2): 87–118. Google Scholar WorldCat Blumson, Ben. 2009. “Pictures, Perspective and Possibility.” Philosophical Studies 149 (2): 135–151. Google Scholar WorldCat Blumson, Ben. 2012. “Mental Maps.” Philosophy and Phenomenological Research 85 (2): 413–434. Google Scholar WorldCat Borges, Jorge Luis. 1998. “On the Exactitude of Science.” In Collected Fictions, translated by Andrew Hurley, 325. New York: Penguin. Google Scholar Google Preview WorldCat COPAC Braddon-Mitchell, David, and Frank Jackson. 2007. Philosophy of Mind and Cognition: An Introduction. Google Scholar Google Preview WorldCat COPAC Bronner, Ben. 2015. “Maps and Absent Symbols.” Australasian Journal of Philosophy 93 (1): 43–59. Google Scholar WorldCat Camp, Elisabeth. 2007. “Thinking with Maps.” Philosophical Perspectives 21 (1): 145–182. Google Scholar WorldCat Camp, Elisabeth. 2018. “Why Maps Are Not Propositional.” In Non-propositional Intentionality, edited by Alex Grzankowski and Michelle Montague, 19–45. Oxford: Oxford University Press.

Google Scholar

Google Preview

WorldCat

COPAC

Casati, Roberto, and Achille C. Varzi. 1999. Parts and Places: The Structures of Spatial Representation. Cambridge, MA: MIT Press. Google Scholar Google Preview WorldCat COPAC Crampton, Jeremy. 1994. “Cartographyʼs Defining Moment: The Peters Projection Controversy, 1974–1990.” Cartographica: The International Journal for Geographic Information and Geovisualization 31 (4): 16–32. Google Scholar WorldCat Cumming, Sam, Gabriel Greenberg, Elsi Kaiser, and Rory Kelly. 2021. “Showing Seeing in Film.” Ergo, an Open Access Journal of Philosophy 7: 730-756. Google Scholar WorldCat p. 521

Cumming, Samuel, Gabriel Greenberg, and Rory Kelly. 2017. “Conventions of Viewpoint Coherence in Film.” Philosopherʼs Imprint 17 (1): 1–28. Google Scholar WorldCat Cummins, R. 1996. Representations, Targets, and Attitudes. Cambridge, MA: MIT Press. Google Scholar Google Preview WorldCat COPAC Danforth, Nick. 2014. “How the North Ended Up on Top of the Map.” Al Jazeera America. http://america.aljazeera.com/opinions/2014/2/maps-cartographycolonialismnortheurocentricglobe.html. WorldCat Davidson, Kathryn. 2015. “Quotation, Demonstration, and Iconicity.” Linguistics and Philosophy 38 (6): 477–520. Google Scholar WorldCat Dilworth, John. 2002. “Varieties of Visual Representation.” Canadian Journal of Philosophy 32 (2): 183–205. Google Scholar WorldCat Dilworth, John. 2003. “Pictorial Orientation Matters.” British Journal of Aesthetics 43 (1): 39–56. Google Scholar WorldCat Greenberg, Gabriel. 2011. “The Semiotic Spectrum.” PhD diss., Rutgers University. Google Scholar Google Preview WorldCat COPAC Greenberg, Gabriel. 2013. “Beyond Resemblance.” Philosophical Review 122 (2): 215–287. Google Scholar WorldCat Greenberg, Gabriel. 2018. “Content and Target in Pictorial Representation.” Ergo, an Open Access Journal of Philosophy 5 (23): 865–898. Google Scholar WorldCat Greenberg, Gabriel. 2019. “Tagging: Semantics at the Iconic/Symbolic Interface.” In Proceedings of the 22nd Amsterdam Colloquium, 11–20, edited by Julian J. Schlöder, Dean McHugh & Floris Roelofsen. Amsterdam: ILLC. Google Scholar Google Preview WorldCat COPAC Greenberg, Gabriel. 2021a. “Semantics of Pictorial Space.” Review of Philosophy and Psychology 12 (4): 847–887. Google Scholar WorldCat Greenberg, Gabriel. 2021b. “The Iconic-Symbolic Spectrum.” Unpublished manuscript. Google Scholar Google Preview WorldCat COPAC Grzankowski, Alex. 2015. “Pictures Have Propositional Content.” Review of Philosophy and Psychology 6 (1): 151–163. Google Scholar WorldCat

Hall, Stephen S. 1992. Mapping the Next Millennium: The Discovery of New Geographies. New York: Random House. Google Scholar Google Preview WorldCat COPAC Harley, John Brian. 1988. “Maps, Knowledge, and Power.” In The Iconography of Landscape: Essays on the Symbolic Representation, Design and Use of Past Environments, edited by Denis Cosgrove and Stephen Daniels, 277–312. Cambridge, UK: Cambridge University Press. Hopkins, Robert. 1998. Picture, Image and Experience: A Philosophical Inquiry. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Hyman, John. 2006. The Objective Eye: Color, Form, and Reality in the Theory of Art. Chicago: University of Chicago Press. Google Scholar Google Preview WorldCat COPAC Kulvicki, John. 2006. On Images: Their Structure and Content. Oxford: Clarendon Press. Google Scholar Google Preview WorldCat COPAC Kulvicki, John. 2015. “Maps, Pictures, and Predication.” Ergo, an Open Access Journal of Philosophy 2: 149–174. Google Scholar WorldCat Kulvicki, John. 2020. Modeling the Meanings of Pictures. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Lascarides, Alex, and Matthew Stone. 2009a. “A Formal Semantic Analysis of Gesture.” Journal of Semantics 26 (4): 393–449. Google Scholar WorldCat Lascarides, Alex, and Matthew Stone. 2009b. “Discourse Coherence and Gesture Interpretation.” Gesture 9 (2): 147–180.

Leong, Mun-Kew. 1994. “Towards a Semantics for a Visual Information System.” PhD diss., Stanford University. Google Scholar Google Preview WorldCat COPAC Lepore, Ernie, and Matthew Stone. 2015. Imagination and Convention: Distinguishing Grammar and Inference in Language. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Lopes, Dominic McIver. 1996. Understanding Pictures. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC p. 522

MacEachren, Alan M. 2004. How Maps Work: Representation, Visualization, and Design. New York: Guilford Press. Google Scholar Google Preview WorldCat COPAC Meier, Brian P., Arlen C. Moller, Julie J. Chen, and Miles Riemer-Peltz. 2011. “Spatial Metaphor and Real Estate: North–South Location Biases Housing Preference.” Social Psychological and Personality Science 2 (5): 547–553. Google Scholar WorldCat Murray, Sarah E., and William B. Starr. 2018. “Force and Conversational States.” In New Work on Speech Acts, edited by Daniel Fogal, Daniel W. Harris, and Matt Moss. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Murray, Sarah E., and William B. Starr. 2021. “The Structure of Communicative Acts.” Linguistics and Philosophy 44 (2): 425–474. Google Scholar WorldCat Panofsky, Erwin. 1939. Studies in Iconology: Humanistic Themes in the Art of the Renaissance. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC

Pavese, Carlotta. 2020. “Practical Representation.” In The Routledge Handbook of Philosophy of Skill And Expertise, edited by Ellen Fridland and Carlotta Pavese, 226–244. London and New York: Routledge. Google Scholar Google Preview WorldCat COPAC Pratt, Ian. 1993. “Map Semantics.” Lecture Notes in Computer Science 716: 77–91. Google Scholar WorldCat Rescorla, Michael. 2009. “Predication and Cartographic Representation.” Synthese 169 (1): 175–200. Google Scholar WorldCat Robinson, Anthony C. 2019. “Representing the Presence of Absence in Cartography.” Annals of the American Association of Geographers 109 (1): 286–300. Google Scholar WorldCat Ross, Je . 1997. The Semantics of Media. Dordrecht:Kluwer. Google Scholar Google Preview WorldCat COPAC Russell, Bertrand. 1923. “Vagueness.” Australasian Journal of Psychology and Philosophy 1 (2): 84–92. Google Scholar WorldCat Schlenker, Philippe. 2018. “Visible Meaning: Sign Language and the Foundations of Semantics.” Theoretical Linguistics 44 (3–4): 123–208. Google Scholar WorldCat Schlenker, Philippe. 2019. “What Is Super Semantics?” Philosophical Perspectives 32 (1): 365–453. Google Scholar WorldCat Schlenker, Philippe, Jonathan Lamberton, and Mirko Santoro. 2013. “Iconic Variables.” Linguistics and Philosophy 36 (2): 91–149. Google Scholar WorldCat Shimojima, Atsushi. 2015. Semantic Properties of Diagrams and Their Cognitive Potentials. Stanford, CA: CSLI Publications. Google Scholar Google Preview WorldCat COPAC Shin, Sun-Joo. 1994. The Logical Status of Diagrams. Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Taylor, Paul. 2023. “Representation, Identity, and Ethics in Art.” In The Oxford Handbook of Ethics and Art, 693–708, edited by James Harold. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Williams, Caroline. 2016. “Maps Have ʻNorthʼ at the Top, but It Couldʼve Been Di erent.” BBC online. https://www.bbc.com/future/article/20160614-maps-have-north-at-the-top-but-it-couldve-been-di erent. Wollheim, Richard. 1987. Painting as an Art. London: Thames / Hudson. Google Scholar Google Preview WorldCat COPAC Wood, Denis. 1992. The Power of Maps. New York: Guilford Press. Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

23 The Semantics of Iconography and Code Words  John Kulvicki https://doi.org/10.1093/oxfordhb/9780192844118.013.44 Published: 22 May 2024

Pages 523–545

Abstract Philosophers have recently turned their attention to code words and dogwhistles. This chapter claims that some of those practices work exactly as iconographic interpretation works with pictures. Exemplars of it are not the kinds of code words that the recent literature focuses on. This reveals that there are two very di erent phenomena out there: spinning codes, which are the focus of most recent work, and distracting codes, the focus of this chapter. It can sometimes be hard to decide which kind of code is being employed, but having both in mind can explain more features of coded speech than having either by itself.

Keywords: iconography, code word, dogwhistle, semantics, picture, depiction, singular content, attributive content, reference Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction The term ‘iconography’ is more at home in certain parts of art history than it is in philosophy. Little philosophical work is dedicated to the topic, and within art history the term is used to describe multiple kinds of interpretation. Exemplary of the phenomenon is the process whereby one takes attributive aspects of a picture’s content, such as those representing a man shot through with arrows, and interprets the picture as representing St. Sebastian. I claim that this process is a semantic operation de ned over the attributive content of a picture that results in the picture having a singular content instead. This phenomenon is not limited to European Renaissance painting, and something like it shows up in most picture-making traditions (Taylor n.d.). By itself, this is an interesting and philosophically unexplored phenomenon. What makes it even more signi cant for present purposes is that we nd similar interpretive moves in language. Philosophers have recently turned their attention to code words and dogwhistles, and I claim that some linguistic coding practices work exactly as iconographic interpretation works. Exemplars of it are not the kinds of code words that recent literature focuses on. This reveals that there are two very di erent phenomena out there: spinning codes and, my focus, distracting codes. It can sometimes be hard to decide which kind of code is being employed, but having both in mind can explain more features of coded speech than having either by itself. Section 2 distinguishes what I call iconographic interpretation from a number of related practices with pictures. Section 3 o ers a simple semantic account of this kind of interpretation, and sections 4 and 5 consider related phenomena. Section 6 turns our attention to code words in language that seem to work p. 524

iconographically, and

section 7 shows how this proposal engages with existing discussions of code words

and dogwhistles. The chapter as a whole presents and expands upon some topics I work through in my (2020b, chap. 5).

2. Isolating Iconographic Interpretation Toward the end of his life, in about 1513, the Venetian painter Cima da Conegliano nished an altarpiece, most likely for the Hospital Brothers of St. Anthony, who were devoted to caring for the sick (Figure 23.1). Their namesake, Anthony Abbot, stands on a low plinth anked by Sts Lucy and Roch. Lucy, a martyr, is the patron Saint of the blind (Thurston and Attwater 1956, v.4, 548–549), while St Roch is renowned for ministering to victims of the plague (Thurston and Attwater 1956, 3:338). Anthony is considered the father of Christian monasticism, and he was also reputed to have healed su erers of skin ailments like ergotism (Thurston and Attwater 1956 1:104–109). None of these three ever spent time with any of the others. Lucy is thought to have lived in the early fourth century, in Syracuse; Anthony, between the third and fourth, in lower Egypt. Roch, born in Montpelier, is their junior by about a thousand years.

Figure 23.1. Cima da Conigliano (Giovanni Battista Cima), Three Saints: Roch, Anthony Abbot, and Lucy ca. 1513. Oil on canvas, 50.5x48in., Rogers Fund 1907. Metropolitan Museum of Art. Open access. https://www.metmuseum.org/art/collection/search/435902

What’s interesting about this case is that we can know the subjects of the painting based primarily on features the painting represents, along with local conventions for depicting them. There are no records of what any of these individuals looked like. Anthony Abbot is represented as an older man, with a long beard, carrying a sta

and, often, a bell. He is dressed in humble monastic robes. Lucy was a noblewoman,

martyred young, so she is depicted as young and well dressed, often holding an oil lamp. The palm frond indicates that she is a martyr. Because of a story about her eyes, she is sometimes represented without them, or holding a pair of eyes on a tray. Roch was also high born and is depicted as well dressed, but exposing a disease sore on his inner left thigh. He is reputed to have come down with the plague and been nursed by a dog that licked his wounds and brought him food. Roch is often represented by a man standing near a dog with a loaf of bread in its mouth, though not in this case. This example highlights a speci c pattern of interpretation that is semantically interesting, and relevant well beyond Renaissance altarpieces. The picture represents two men and a woman, for example, and the old man is dressed more modestly than either of the younger gures. Beards, sores, oil lamps, and palm fronds sit beside a raft of other features that this picture represents. Colors and shapes, lighting, footwear, background and oor. Some clusters of these features—older man, humbly dressed, with sta

and bell—

lead an appropriately learned audience to particular individuals, like Anthony Abbot. This is a code, in at least three ways. First, it is an operation de ned over meaningful, attributive aspects of p. 525

the image. A cluster of represented features somehow

stands for a saint. Second, those meaningful

aspects, or features, are not the meaning being communicated. One needs at least passing familiarity with stories of the Saints’ lives, along with pictorial conventions, in order to unpack the meaning coded within these attributive clusters. Notably, reading language is not required. There is no text on the painting, and

there needn’t have been any around it either. Let’s call the clusters of features needed to represent individuals like saints their “attributes” (Taylor n.d., chap. 4). Third, the connection between features and individuals is, in principle, arbitrary. Saints need not have just one attribute, as examples for St Lucy and St p. 526

Roch attest. In theory, pretty much any combination of features could be used as an attribute for any individual you like. Absent some connection to the individuals in question—something found in stories about their lives, or elsewhere—these codes become unreadable, but that doesn’t speak against their inprinciple arbitrariness. Let’s contrast this interpretive process with two others that seem to leave us in a similar place, by alternative routes. The contrasts between these three turn out to have interesting analogs in linguistic communication. First, a Man Ray photograph from 1922 depicts a woman sitting next to a painting. Inspection reveals many details of the person’s physiognomy, hair, facial expression, clothing, setting, and so on. Ditto for the painting, which one can clearly see depicts a woman sitting in a chair. Those who knew her could recognize Gertrude Stein in the photo, and those familiar with painting would recognize Picasso’s 1905 portrait of Stein as well. Those unfamiliar with either of these individuals could nevertheless say a lot about what they are like based on the photograph. Antecedent familiarity with individuals allows one to use an attributive content—the raft of features one extracts from the photograph—to isolate those individuals. “Oh, that’s Gertrude! And that portrait of her by Picasso.” We recognize friends, celebrities, and acquaintances in pictures all the time. As we will see later on, the process by which one gets to Stein here is quite di erent, semantically, than the process by which one gets to St Lucy. It’s fair to say that this Man Ray photo is “of” Stein and her portrait because Man Ray pointed his camera at Stein and her portrait. What matters right now is not whether photographs isolate individuals in a manner often absent from handmade images (see Walton 1984 and Lopes 1996, e.g.). Even with a handmade image, rich with detail, and perhaps made in the absence of both Stein and the Picasso, one could extract a rich set of features, and those features plausibly su

ce for distinguishing both Stein and Picasso’s portrait of her

from other things. So, the kind of interpretation of interest here is one akin to what happens when a rich description is recognized as denoting a speci c individual. Second, consider these two enamel Byzantine icons made around 900 years ago (Figures 23.2 and 23.3). One represents St Matthew, the other St John the Theologian (who might be identical to John the Evangelist). Aside from their gazes, the disposition of their right hands, and the pattern on their books, little distinguishes the pictures from one another attributively. The face, the beard, the clothing, the book cradled in the left arm, are all the same.

Figure 23.2. Unknown. Medallion with St Matthew from an icon frame. ca. 1100. Gold, silver, and enamel worked in cloisonné. Diam. 3.25in. Gi of J. Pierpont Morgan 1917. Metropolitan Museum of Art. Open access, public domain. https://www.metmuseum.org/art/collection/search/464545

Figure 23.3. Unknown. Medallion with St John the Evangelist from an icon frame. ca. 1100. Gold, silver, and enamel worked in cloisonné. Diam. 3.25in. Gi of J. Pierpont Morgan 1917. Metropolitan Museum of Art. Open access, public domain. https://www.metmuseum.org/art/collection/search/464549

What allows us to distinguish Matthew from John are the labels written onto the icons. One says “Saint John the Theologian” while the other says “Saint Matthew”. This labelling is not limited in place or time to the Byzantine era. Albrecht Dürer’s “Great Triumphal Car” from 1522 is another glorious European example, and we will see yet another by Biagio d’Antonio below. This case di ers from the rst two in that it o ers no attributive contents that help determine which individuals, if any, are involved in the representation, and it doesn’t really involve a code. No representation in this example is being deployed to mean something unlike what it ordinarily means. Yes, to understand this icon one must be minimally literate, but one need not have p. 527

learned, for example that ‘St Matthew’ means something like St George.

p. 528

In all of these cases, once the interpreting is done, we have representations that are in some sense “of” particular individuals like Gertrude Stein or St Matthew, while also being attributively rich. What separates these examples are the semantic mechanisms involved. I have chosen to call only the rst of these “iconographic” interpretation, though the others are interesting in their own ways, and art historians are fairly content calling at least the rst and the third “iconographic” (Taylor n.d.). Cynthia Freeland (2010), by contrast, suggests interesting connections between Byzantine icons and photographs. The next two sections try to be more explicit about how iconographic interpretation works with pictures, while section 5 sets the Man Ray example in its proper place, before we turn to language in section 6.

3. The Semantics of Iconographic Interpretation As of this writing there is no consensus on exactly what pictorial content is like, but we will not need agreement on that score to say worthwhile things about iconographic interpretation. Gabriel Greenberg (2018), for example, believes that all pictures have attributive and singular contents. I suggest (Kulvicki 2020b) that pictures fundamentally have attributive contents, but can acquire singular contents through operations de ned over those attributive contents. Many others seem to think that pictures have a mix of singular and attributive elements. Everyone who thinks pictures have contents at all thinks that they have attributive contents, and that su

ces to make the relevant points here.

Attributive content, as the name suggests, attributes features to something or other. “The car is red” attributes redness to the car. It also singles out some individual, the car, to which it attributes that property. Pictures represent, perhaps among other things, patterns of features, and in that sense have contents that are attributive. This makes pictorial content, at least in part, a lot like the content of a description. Erwin Panofsky (1955, 28), who was an important early theorizer of iconography, called these attributive contents the “natural subject matter” of pictures. Stick gure drawings show that pictures needn’t have terribly rich contents, but they can, and it’s reasonable to think that many pictures’ attributive contents outrun even a very rich lexicon of concepts. Presented with the rich attributive content, one then nds an attribute in it. The attribute is connected to some individual, who then becomes part of the picture’s content. That is, the picture now has a singular content. Generally, attributes are readily recognized features for which a target audience has concepts. It is useless to make a saint’s attribute a highly speci c color or shape for which people have no concepts, since in that case, interpreting iconographically would not be easy for the casual viewer. So, attributes will generally be easily articulated collections of features, like being male, reasonably well-dressed, and revealing a sore on the leg, or being female, well-dressed, and carrying an oil lamp and a palm frond. p. 529

Once one has identi ed an attribute, the picture acquires a singular content—the saint, in our examples—in addition to having its attributive content. Singular contents provide individuals of whom this or that feature can be predicated. Iconographic interpretation, on my view, delivers individuals as contents, and it does so based on an attributive content. It is a semantic operation de ned over a picture’s attributive content. So, iconographic interpretation is a way of moving from a representation that is (perhaps exclusively) attributive to one that includes a singular element. Understanding how this works involves taking a closer look at how pictures manage to have attributive contents in the rst place. Though people disagree, and the account that follows is one to which I am partial (Kulvicki 2020b, chap. 2), I try, for present purposes, to stay as neutral on controversial topics as I can while still making the relevant points. Pictures place colors at locations on their surfaces. Colors at locations a ect the chromatic features that di erent parts of a scene are represented as having. So, the presence of a color patch at a certain place on a picture surface constrains the ways such a picture can be interpreted. Subtle changes to the color, or to its location, can result in changes in the picture’s attributive content. At an extreme, we can imagine pictures placing inde nitely nely di erentiated colors at inde nitely small regions of their surfaces. And that’s why pictures can seem inde nitely attributively rich. Notice that we needn’t say that a color at a location represents that color at some represented location. That’s probably false (Hyman 2006; Newall 2006; Kulvicki 2020a). It’s just that, at some fundamental level, pictures represent chromatic features of spatially extended scenes. That fundamental level manages to make pictures attributively powerful well beyond chroma and place. Palm fronds, for example, manifest complex patterns of light, shadow, and color across a region of space.

Ditto oil lamps, faces, legs, sta s, bells, and eyes. By constraining the chromatic aspects of a scene, pictures can, as it were, describe them as including palm fronds and anything else one might encounter based on how it occupies space chromatically. And all such things are candidates for being included in the attributive content of the picture. Attributes, understood as clusters of features connected to particular individuals, are parts of pictures’ attributive contents, in the following sense. You can nd all the features that are included in the attribute— well-dressed woman holding a lamp and a palm frond, for example—and they will be a subset of the totality of features the picture attributes to the scene. There are many aspects of the Cima altarpiece that are irrelevant to interpreting it iconographically, even though they are relevant to interpreting it pictorially. Once we have carried out the iconographic interpretation, we have a representation with singular contents, like St Anthony Abbot, St Lucy, and St Roch. But what of the attributive content? It’s natural to think that to the extent that the picture has both attributive content and singular content, it attributes features to the individuals that constitute the singular content. But remember that there are three parts of the content we need to consider: the singular content, the attribute, or the part of the attributive content responsible for delivering the singular content, and the remainder of the attributive content. p. 530

In some cases, it is plausible that the attributive content is ignored altogether. The point of some pictures is to deliver the individual saint as a content, for contemplation and prayer. The point is not to say anything in particular about that saint, and certainly not about how they might have looked or dressed. That’s not to say pious viewers are blind to these pictures’ attributive contents. Quite the contrary. One needs the attributive content to get the singular content, for example. In addition, the painting could be very compelling and well-done when understood attributively. But in its role as an aid to devotional practice its function was completely singular. Art historian Charles Hope (1986) points out: In fteenth-century Italy religious pictures were divided into two basic categories: images (imagini) and stories (storie). An image was a representation of a person, for example a saint, and it was often used as a focus for devotion…. People who wanted to pray to the Virgin would have done so before an image, not a story of the Virgin. (804) Both images and stories were images, or pictures, in contemporary parlance, but the images weren’t interested in saying anything about the individuals that they represented, while the stories were. The Cima altarpiece is an image in Hope’s sense. The point is not to imagine a meeting of three gures who never met, or to tell us which of them was taller than the others, but to present these individuals to members of a society, the Hospital Brothers of St Anthony, for whom they were particularly signi cant. Though Hope isn’t speaking of Byzantine work in his paper, it is plausible that the rondels of Sts Matthew and John were playing a similar role. The point is not to say anything about the saints, so much as to o er them up for devotional practices. That’s why it matters little that the pictorial aspects of those pieces were pretty much identical. In the Byzantine case, it is the text, not attributive aspects of the pictures, that provides a singular content. One might think this frees the artist’s hand to depict Matthew and John in any way he might like. The lack of interest in doing this—and this is not an isolated example of the phenomenon —is telling. It’s not the attributive aspects of the pictures that matter, given the function of this image. So, attention to those details is somewhat beside the point.

In other cases, it seems as though the attributive content of the picture attributes features to the individuals identi ed iconographically. On the one hand, we have what Hope would call an image. With St Lucy, Cima had to present the relevant attribute as part of the pictorial content, but there was much else he had freedom to do that a ected how he presented Lucy to us. He could have presented Lucy from the left side, or the right. Looking up, or down. With any number of hair styles, colors of clothing, facial expressions, and lighting. None of those details are relevant to isolating her iconographically. So, the artist can manipulate them to tell us something about how he understands Lucy. It would be odd to suggest, by contrast, that Cima depicts Lucy as holding a palm frond and a lamp, Roch revealing a sore, or Anthony holding a cane. We only have a representation that includes Lucy as an element p. 531

in virtue of its having an attributive content

including a well-dressed woman holding a lamp. For those

aspects of attributive content to play an attributive role would suggest that all iconographic representations of Lucy also represent her as holding a lamp, for example. That’s implausible, and it doesn’t t with how such representations are generally interpreted, as will become clearer in the next case. The overall thought is that if an attributive element is used as part of a process to derive a singular content, then that element loses its attributive force. But much of a picture’s attributive richness is, as it were, available to the artist and can be used in the service of characterizing the individuals who are identi ed iconographically. In those cases, the picture is part of a communicative act akin to one that says, “Consider Sts Anthony, Lucy, and Roch as being like this.” ‘This’ refers to the complex attributive remainder. I say “Consider … ” because it is implausible that Cima was trying to tell us what these people looked like, or that he was capturing a way any of them actually dressed. It would be overreach to claim that this altarpiece expresses a proposition, but not to say it is an imperative, asking us to understand the saints in this or that way. On the other hand, we have what Hope would call a story. Consider the following Botticelli from around 1500 (Figure 23.4).

Figure 23.4. Botticelli (Alessandro di Mariano Filipepi). Three Miracles of St Zenobius. ca. 1500. Tempera on wood. 26.5x59.25in. John Stewart Kennedy Fund 1911. Metropolitan Museum of Art. Open access, public domain. https://www.metmuseum.org/art/collection/search/435729

This painting shows a uni ed space, as the architectural elements attest. Within that space, however, we have three examples of miracles that were supposed to have been performed by Saint Zenobius (see Thurston and Attwater 1956, 2:390), who was Bishop of Florence in the fourth and fth centuries. Each event was supposed to have taken place at a di erent time and place. On the left, he brings a dead child back to life. In the center, he brings a man who had been carrying relics back to life. And on the righthand side he blesses water and salt—you see him doing this in the far background—which St Eugenius, his deacon, then uses to bring a third person back to life. Notice Eugenius not only receiving the blessed items in the

p. 532

background, and using them in the foreground,

but also transporting them, in the middle ground. As

Hope (1986, 805) points out, it was completely ordinary at the time to use spatially uni ed scenes to depict multiple episodes in the life of a saint. This practice is related to what we nd in the Cima altarpiece, which presents individuals from di erent times and places occupying a uni ed time and space. What makes it di erent is that each part, as it were, describes an event in some detail. We see the concerned onlookers, the injured people, and so on. What is striking for our purposes is the way in which Zenobius is represented in each case: wearing full bishop’s robes. It is unclear that all of the miracles were performed while he was bishop, and the robes are, of course, fteenth-century raiment, not something Zenobius would have worn a thousand years earlier. The regalia serves, in this context, as an attribute that lets us identify the bishop across each of these scenes. He wasn’t the only bishop to be represented by Florentine artists, though he was an important one for that city, and he wasn’t always represented wearing such robes. Domenico Veneziano’s panel from 1422–1428, currently in the Fitzwilliam Museum at the University of Cambridge, makes this clear (see Lillie 2014). The important parts of the Botticelli painting are the stories it tells about what St Zenobius did. It identi es Zenobius, and attributes features to him, like kneeling and praying in front of injured people, surrounded by distraught family members and onlookers. The robes serve to single him out, but it is inappropriate to think Botticelli suggests that the relevant miracles were performed while Zenobius was so dressed. Attributes thus can serve to identify individuals across contexts that are meant to tell stories about them, as in Botticelli’s painting, but they lose their attributive force once they are used to do so. It’s interesting in light of the contrast between labelling, as we saw in the Byzantine rondels, and iconographic methods, to consider the following Biagio d’Antonio painting from 1482 (Figure 23.5).

p. 533

Figure 23.5. Biagio dʼAntonio. The Story of Joseph. ca. 1482. Tempera on wood. 27x59in. The Friedsam Collection, Bequest of Michael Friedsam 1931. Metropolitan Museum of Art. Open access, public domain. https://www.metmuseum.org/art/collection/search/435666

Similar to the Botticelli, this painting presents a spatially uni ed scene that depicts ve episodes in Joseph’s life (Thurston and Attwater 1956, 1:630–633). In this case, however, Joseph is represented without any attributes, and as wearing contemporary ( fteenth-century) clothing. We identify him as Joseph because he is labelled so, along with other characters in the ve stories (see Figure 23.6).

Figure 23.6. Biagio dʼAntonio. The Story of Joseph. ca. 1482, detail. Tempera on wood. 27x59in. The Friedsam Collection, Bequest of Michael Friedsam 1931. Metropolitan Museum of Art. Open access, public domain. https://www.metmuseum.org/art/collection/search/435666

By employing the labels, the d’Antonio painting frees up the usual attributive resources of the picture. None p. 534

are devoted to securing a singular content, so all of them

can plausibly be attributing features to St

Joseph. Doing this comes at a cost, of course. Those who are illiterate might be unable to interpret the painting, and we are using text in a way that mixes it with another sort of representation. The painting does not depict words, somehow written in the scene, but rather uses names to annotate parts of the picture. These uses of iconographic attributes are not limited to well-known stories of the saints’ lives. In one striking example from circa 1420 to 1430, by the G. Z. Master, the person who commissioned the painting, Pietro de’ Lardi, is depicted as kneeling in front of the Virgin and Child, being presented to them by St p. 535

Nicholas (Figure 23.7). The Virgin Lardi is identi ed by his appearance.

and Child, and St Nicholas, are identi ed via their attributes, while de’

Figure 23.7. Master G.Z. (possibly Michele dai Carri). Madonna and Child with the Donor, Pietro deʼLardi, Presented by St Nicholas. ca. 1400-31. Tempera and gold on wood. 45 7/8 x 43 5/8in. Bequest of Adele L. Lehman, in memory of Arthur Lehman 1965. Metropolitan Museum of Art. Open access, public domain. https://www.metmuseum.org/art/collection/search/437014

Figure 23.8. unknown. St Catherine Delivered from the Wheel. ca. 1375-1400. Oil and tempera on panel. 14 1/8x13in. A.A. Munger collection. Art Institute of Chicago. CC0 Public domain. https://www.artic.edu/artworks/12683/saint-catherinedelivered-from-the-wheel

We see close analogs of these practices in contemporary image-making. Perhaps the most salient example is comics, in which the central characters are often depicted via attributes like clothing and hairstyle. In these cases, the characters are often invented, but it’s possible to invent a character with a clear physiognomy, and thus make reference to that character work more like the Man Ray photo than the story of St Zenobius. But comics are often fairly sketchy when it comes to physiognomic detail. It’s much easier to p. 536

single out characters via more readily discernable traits, like clothes of a particular

sort, a hat, or a

hairstyle. And in the comics the pattern seems much like what we nd in the older examples of iconography. Though we might think of our favorite characters decked out in their standard clothes, it’s not necessarily part of the comic’s story that such characters always dress that way. Once attributive contents are involved in providing singular contents, they fail to maintain their attributive roles. Superheroes are an interesting exception. They are supposed to be wearing the same thing, whenever they are active in their roles, and that is an interesting way to let a story obviate the need for iconographic interpretation. So far, we have seen that iconographic interpretation makes use of attributive content to deliver a singular content, as part of what a picture represents. Such pictures can also attribute features to the individuals that constitute this singular content. But attributive content responsible for establishing the singular content ceases to play an attributive role. Iconographic interpretation shows up in pictures whose only role is presenting individuals for contemplation, or other acts, as well as in representations that are trying to tell viewers stories about them. The closest contrast to iconographic interpretation is the process of labelling. That provides a picture with singular contents, without burdening their attributive contents to do so. Both paths leave us with

representations that have both singular and attributive contents. It’s just that the iconographic case sacri ces some attributive potential of the picture in order to establish singular contents. The Man Ray photo of Stein and her portrait works di erently than both of these, as section 5 shows. Before moving on to that, however, we need to consider a variation on this theme, in which singular pictorial contents are mapped onto other singular contents.

4. Mapping Singular Contents to Singular Contents Do pictures ever get interpreted in such a way that singular contents are mapped onto other singular contents, in a way reminiscent of what happens with the more exemplary cases of iconography mentioned in the previous section? Perhaps. It’s controversial just how pictures wind up with singular contents generally, so bracket that issue and consider the following example. Pepe the Frog is a ctional character created in 2005 for a comic, Boy’s Club, by Matt Furie (2016). The character started showing up in internet memes soon after Furie created it. Memes are typically pictures involving both singular and attributive elements. Speci cally, a character is portrayed as saying or doing something or other. Pepe was a popular subject of these memes, but in 2016 the image was appropriated by right-wing supporters of Donald Trump’s presidential campaign. The character was soon included in the Anti-Defamation League’s Hate Symbols database. In the right-wing context, Pepe came to stand for an ideology, which in many cases was clearly White supremacist. p. 537

The interpretive game involved with a character like this is fairly complex. First, pictures with complex attributive contents—a froglike humanoid form, etc.—come to have a singular content associated with them: Pepe the Frog. To do this, one needs to be competent interpreting pictures attributively. But then the new singular content—Pepe—comes to stand for another singular content, for example, a White supremacist ideology, adherents to that ideology, or something like that. This mapping demands competence with the attributive contents of pictures, as well as the step, whatever it was, that gets from that to a singular content like Pepe. Someone unable to do either of those things could not helpfully interpret these images. The move from Pepe to a White supremacist ideology looks very similar to the move from a well-dressed woman with a lamp to St Lucy. It’s just that the content one starts with in this case is singular rather than attributive. As with the more standard examples of iconographic interpretation, once a part of a representation’s content is mapped onto something else, the representation is no longer interpreted as having that content. These representations aren’t representing a relationship between Pepe, or froglike humanoids, and White supremacy, for example. Representing Pepe was just the means to get to this other representation. This is exactly analogous to how the attributive cluster responsible for delivering St Lucy ceases to play an attributive role in the Cima altarpiece. When we turn to linguistic representations, in section 6, we will see that both attributive clusters and singular terms are the basis for reinterpretations that look a lot like what happens with iconographic interpretation in pictures.

5. Denotation in Photos and Handmade Pictures Man Ray’s photo captures details su

cient for identifying Gertrude Stein, and Picasso’s portrait of her. The

model for identifying Stein, however, is not a code that pairs her with an attribute or a linguistic label like a name. Instead, the picture acts as a complex description that attributes properties in su

cient detail to

identify Stein, perhaps uniquely. In that sense, the photo refers to Stein by denoting her, much as a detailed description might. Doing this does not require the picture to acquire a singular content, any more than an identifying description acquires a singular content by succeeding in referring to something. As mentioned, it’s reasonable to think that photos typically have singular contents because of their distinctive causal sources. Man Ray had to point his camera at something, and so that scene, and perhaps the individuals that constitute it, are plausibly what the photo is “of”. Plausibly, the photo attributes features to that scene. This is generally true of photos and might also be true of some handmade images (Lopes 1996, 2009). For now, though, that’s beside the point. What’s interesting about the photo is that it p. 538

does

something you might think a painting of a saint could not do: refer via denotation, or a description

that captures su

cient detail to distinguish the individual from others. We have no idea just what many of

these gures from the early Christian history looked like, so it’s very implausible that pictures could be made in their likenesses. There, is, however, a practice that latches onto gures from history in a manner analogous to the photo. Sometimes, a story is su

cient to identify the players in it. So, some paintings of stories identify such a

distinctive set of features that is makes sense to identify them as denoting a speci c event involving speci c individuals. St Catherine of Alexandria is reputed to have been sentenced to die by the Roman Emperor Maxentius (Thurston and Attwater 1956, 4:420–421). The sentence was to be carried out on a spiked wheel, but when they tried to do so, the wheel broke apart. Maxentius then had her beheaded. Unsurprisingly, Catherine is often represented by a well-dressed woman (she was supposed to have been from a wealthy family, even a princess) standing near a spiked wheel, often holding a palm frond. That is very much like examples we have already seen. But the fourteenth-century Spanish panel in Figure 23.8 works a bit di erently than standard cases of iconographic interpretation. It shows a man with a crown and sword pointing at a woman. Between them one sees a bladed wheel set on a support, a guard taking a nap, and an angel taking an axe to the wheel. This is not simply representing Catherine through her attributes. Instead, the artist depicts a scene from the life of St Catherine, and the details of the depicted scene are su

cient for one to zero-in on Catherine and

Maxentius. Without knowing what these people looked like, one can still denote them pictorially by representing a scene perhaps uniquely satis ed by them. This panel denotes a scene from the life of St Catherine, and so it is not iconographic interpretation, as I have been demarcating it. Instead, it is something much closer to learning that the photo is of Stein and Picasso’s portrait by noticing the features that the picture depicts.

6. Iconographic Interpretation in Language? The previous four sections unpack a number of ways in which pictures are used to serve complex representational ends. Recent work in philosophy of language considers a variety of “coded” ways in which messages—White supremacist, misogynist, anti-Semitic, and so on—can be communicated (Stanley 2015; Khoo 2017; Camp 2018; Saul 2018; and Santana 2022, e.g.). It has become clear that, not only are there multiple ways of thinking through any given phenomenon, but there are many, possibly distinct, phenomena out there worthy of explication. Some examples of this, I argue, are strictly analogous to what happens with iconographic interpretation. Speci cally, code words often have ordinary, attributive p. 539

contents, and the rule for interpreting them is that those

attributive contents—perhaps only part of the

content of the overall expression—deliver singular contents in their places. The clearest examples of this are not the code words most discussed in the literature. But once the exemplary cases are clear it becomes an interesting question whether some other cases, including those discussed quite a bit in the literature, might receive a similar account. The Anti-Defamation League maintains the “Hate on Display: Hate Symbols Database” here: https://www.adl.org/hate-symbols. Some of them are pictorial, in that to understand what they are you need to interpret them pictorially. Others in this list are purely linguistic. ‘14 Words’ is shorthand for a White supremacist phrase containing that many words. But it’s not just that, because ‘14 words’ is a piece of language that can nd its way into any number of ordinary expressions. And the point of the phrase for which it is shorthand is to stand for a White supremacist ideology, regardless of its ordinary attributive use. Consider an exchange in which someone says, of people he disagrees with, that “I suppose we have 14 words or so that we can share with them.” This ominous sentence is easy to interpret, at a basic level, and doing so is essential to interpreting it the extended iconographic sense I have in mind. The expression ‘14 words’, I propose, delivers a White supremacist ideology as a singular content. The original sentence in which the phrase occurs does not make a lot of sense if a singular expression for an ideology is substituted in there. Yes, perhaps he is suggesting that he will communicate the ideology to the relevant people. But more likely, it seems, is that the other content of the sentence is simply ignored. The point of using the expression is simply to present that ideology to his interlocutor in a way that expresses his endorsement of it. That can su

ce as a declaration of a position with respect to some other group. In some cases, perhaps the new

singular content makes perfect sense with the original surrounding contentful material, but in many cases it does not. And it is important that any connection to the original material seems contingent. One can often accomplish the very same conversational goal by simply uttering or writing ‘14 words’ absent other surrounding linguistic material. Both Elisabeth Camp (2018, 34) and Jenny Saul (2018, 366) point out that coded expressions needn’t express propositions, or even deliver content that, by itself, constitutes an articulate expression like a command. The code delivers a contentful part of a communicative act that can include many other meaningful parts. It’s essential to understanding this code that one be able to interpret the linguistic expression. It is not simply a symbol that needs to be learned by rote, though plenty of those exist on the Anti-Defamation League’s database. Evidence for this is that one need not use numerals to express this, one need not capitalize, or use a particular typeface. ‘Fourteen Words’ works, as can, in the right context, ‘ten plus four words’, and so on. An expression with an attributive content is noticed as such, and it is that attributive content that yields the singular content in its stead. Other White supremacist expressions, like ‘Blut und Ehre’ survive translation into ‘Blood and Honor’, ‘Blood & Honour’, and so on. Interestingly, this one does not survive reordering, so ‘Honor and Blood’ is far enough away not to raise worries about whether one is using a hate symbol.

p. 540

The following example uses a singular term in a coded manner, rather than an attributive expression, and so it’s more like the Pepe the Frog case than ‘14 words’ is. Recently, the Lincoln Project produced an advertisement against the 2021 Republican candidate for Virginia governor, Glenn Youngkin, who went on to win the election. (See https://lincolnproject.us/critical-race-card/). The video begins by asking, “Why is Glenn Youngkin suddenly talking about Critical Race Theory? … After all, it’s not taught in Virginia schools, not one.” They follow this up by reproducing Lee Atwater’s awful but telling quote (the epigraph in Saul 2018) about what to do when simple racial slurs become frowned upon: “So you stay stu

like ‘forced busing”, “states’ rights’, and all that stu .” Such expressions were cast by

Atwater as replacements for slurs. And the segment concludes by suggesting: Like Atwater, Youngkin is using coded language, because in 2021, he can’t get away with saying what he really wants to say. So, if you see Glenn Youngkin talking about critical race theory, make sure to let him know he’s not fooling anybody. (Lincoln Project 2021) The Lincoln Project’s suggestion is that in talking about why critical race theory is a big problem for Virginia, something worth ghting against, he is really saying that Black people are a problem. It’s also possible that he is saying that calls for racial justice and White accountability are a problem. For present purposes, settling the issue speci cally isn’t as important as noticing that we seem to have a code here, which works similarly to the earlier examples. In this case, ‘critical race theory’ is a singular term with a singular content, not an attributive one, and the thought is that a di erent singular content is substituted for it. Because di erent singular contents can perform similar functions in linguistic representations, this substitution can readily transform complaints about critical race theory into complaints about African Americans or calls for racial justice or both. But sometimes, it might be that uses of ‘critical race theory’ are just meant to present Black Americans or calls for racial justices in a disapproving manner. These examples t with the structure of iconographic interpretation in the following ways. First, they demand competence with the basic, non-iconographic, interpretation of the language. Meaningful aspects of the language, need to be understood as such. So, it seems like this interpretation is an operation that takes one kind of meaning, and delivers another. Second, they take attributive complexes or singular contents in the language and map them onto singular contents. Third, they are somewhat arbitrary, in that any number of phrases could acquire a similar role. And fourth, sometimes those singular contents then contribute to a new expression involving the remainder of the expression with which one started. But sometimes they do not, and the sole point of the non-iconographic content is to get one to the singular, iconographic content. In the cases we considered, that content can be something like an ideology, or supporters of that ideology, presented in a negative or positive light.

p. 541

7. Contextualizing the Proposal There are a number of phenomena that t under the general terms ‘code word’ and ‘dogwhistle’. And there are a number of competing accounts for each of those phenomena. Some insist that the point of code words is that they communicate some meanings that are hidden from some substantial part of one’s potential audience (Mendelberg 2001; Stanley 2015; Saul 2018; Camp 2018). Others suggest that meanings are communicated, but they are rarely hidden (Santana 2022). And yet others suggest that this isn’t about expressions having di erent meanings, so much as expressions a ecting the inferential or associative behavior of one’s audience (Khoo 2017; Saul 2018; Torices 2021). All parties to the conversation believe that code words allow those who use them a kind of cover from criticism. This is central to Tali Mendelberg’s (2001) important work that was an inspiration for many of the other accounts just mentioned.

Section 6 o ered an account of code words that’s di erent from those currently on the market, but it also highlights phenomena that have been peripheral to current discussions. Most code words discussed so far in the literature are “spinning codes” in that they seek to alter but not replace the message being communicated at the most basic level. The ones I have in mind are “distracting codes” in that when properly understood they can simply leave the original, basic-level message behind. In some cases, I think the literature confuses one kind for the other, even though distracting codes haven’t been the focus. Seeing all of this requires brie y comparing the current proposal to those aimed at the codes that spin. My model for distracting codes is not quite ambiguity, where a term has more than one meaning and it is unclear, in a given context, which one is intended. As far as I can tell, no one has unpacked such a view in detail. Justin Khoo (2017, 40n14) tentatively attributes something like this view to Mendelberg (2001) and then raises a number of e ective arguments against it (40–42). Saul (2018) suggests that “overt, intentional dogwhistles” communicate two meanings, but she doesn’t suggest a model according to which this amounts to ambiguity. In the cases I discuss, the “coded” interpretations are optional, in that many might not know the code, and even those who do might not know whether that was intended. But the relevant interpretation is an operation de ned over the ordinary, unambiguous, meaning of the relevant expression, like ‘14 words’. The model o ered here also does not suggest that secondary or alternative meanings are put into the conversation as “not-at-issue” contents, as Jason Stanley suggests (2015, 163; see Khoo 2017, 46, for worries). The “at-issue / not-at-issue” distinction seems valuable for understanding many linguistic practices, including some code words. But someone inserting ‘14 words’ into what seems like an innocuous sentence opens the possibility for a completely di erent kind of communicative act. Rather than one act with at-issue elements and not-at-issue elements, we get a di erent communicative act altogether. It singles out an ideology, or group of people, and presents it in an approving or disapproving manner. At the p. 542

basic level, notice, a sentence involving the phrase ‘14

words’ need not go anywhere near dodgy

ideologies. The cases Stanley worries about are those in which the conversation concerns a speci c topic, and the not-at-issue additions to the conversation a ect how people are able to think about that topic. The model I have in mind involves more of a wholesale change. For example, people criticized Youngkin for saying false, sometimes senseless, and often misleading things about critical race theory. But if the point of uttering ‘critical race theory’ is to achieve reference either to African Americans or calls for White accountability in a negative light, then the speci c things Youngkin says about critical race theory don’t really matter. His comments aren’t meant to be refracted by the lens of some not-at-issue content. Nor are we reinterpreting his statements with a new meaning for ‘critical race theory’. Most of his comments are just fodder for getting the expression ‘critical race theory’ out there. Replying to him by pointing out that his claims are false and misleading is missing the point altogether, and I take that to be the message of the Lincoln Project video. In images, as described above by Hope, the point is presenting the saint to those interested in praying to or contemplating them. Yes, one might appreciate the way Cima renders Roch’s clothing, but even a slapdash job could have accomplished the same iconographic end. With distracting codes, the material surrounding them seems to play a similar, inessential role. One needs to speak in complete sentences, lest one sound silly, but the speci c contents of the sentences matter little to the communicative goals. This is what makes some codes distracting. They change the topic, instead of spinning it. Distracting codes do preserve a limited deniability on the speaker’s part, because they can always say that they were making only the basiclevel claim. Moreover, distracting codes can interrupt otherwise helpful discussions of critical race theory because the term is now taken to refer to something it does not, in the basic sense, refer to. This is one route to what Jennifer Hornsby dubs “ine ability” (1995, 134–136). The words might still be there, but they cannot, in

certain contexts, be used to express what one wants to express. Compare the plight of the Renaissance painter who wishes to depict his niece, who happens to be young and well-dressed, holding a palm frond nearby a bladed wheel. The materials are there for doing so, but iconographic conventions yield a St Catherine, which is not what the imagined artist was after. Interrupting otherwise helpful discussions is an e ect of deploying distracting codes, and as such it can be a perlocutionary goal (Austin 1962) of such speech. A number of authors have unpacked spinning codes at least partly in terms of their e ects. For Khoo (2017, 47) they a ect our inferential behavior without needing a special semantic component to do so. He also suggests that they might involve associations, rather than inferences. Saul (2018) suggests that “covert intentional dogwhistles” or “implicit political communication” (Mendelberg 2001) bring preexisting racist attitudes “to bear where they might previously not have been drawn upon” (Saul 2018, 366). In covert cases, the audience is unaware that this is what is happening, and Mendelberg (2001) has data suggesting that when the in uence is made explicit, the e ect vanishes. Distracting codes have e ects, like all successful attempts at communication, but it seems like the e ects p. 543

they have are due to a semantic operation de ned over the

ordinary meanings of some expressions.

Because of that, they are also not implicit, even though some who use such expressions might unintentionally use them with the coded meaning (see Saul 2018, 365–371). In many cases, it might be hard to decide exactly what kind of code is being used. The clearest tell, in my opinion, is whether the coded interpretation seems closely linked to the underlying basic meaning, or oats free from it. In the former case, we usually have spinning codes, while in the latter we have distracting codes. These two kinds of code are species of an interesting genus, it seems, and Carlos Santana (2022) o ers a helpful general account that captures both sorts: An act of political communication is a dogwhistle to the extent that: a. It has a secondary, implicit meaning in addition to its surface meaning. b. The secondary meaning, but not the surface meaning, calls attention to a politically meaningful social category in a way that violates widely shared egalitarian norms. c. An ingroup audience approves of violating these norms, while an outgroup audience prefers adherence to them. d. The norm violation can be successfully denied by the dogwhistle user. (Santana 2022, 394) This account focuses on communicative acts, rather than speci c terms, but it does not specify the mechanism whereby any particular act achieves its twofold signi cance. That is by design, as there are many paths to this end, by Santana’s lights. The secondary meaning could be brought into the conversation via priming inferences, associations, not-at-issue contents, and so on. I suggest that, additionally, it could be brought in via a semantic operation de ned over a code word, as it is with ‘fourteen words’ and ‘critical race theory’ in Youngkin’s hands. Santana’s clause (b) is helpfully neutral as to whether the secondary meaning spins the original message, or replaces it with something else entirely. And the de nition also captures that these distracting codes are often aimed at groups, or, I would add, ideologies, and presenting them in this or that fashion.

8. Conclusion This chapter hopefully convinces those focused on language that there is much work to do sorting out the semantically complex things we do with pictures. What’s more, it hopefully convinces philosophers of language that attending to pictures can inform our understanding of what happens in language. Iconographic interpretation has received relatively little attention by philosophers working on images, much less philosophers of language, but it proves to be interesting for the way in which pictures use attributive contents to arrive at singular contents. Artists have developed many mechanisms for p. 544

accomplishing these ends. Some of them,

those I have called iconographic, are very much like codes,

while others make use of labels to accomplish similar semantic ends. One important part of this is that if an attributive content is used to secure a singular content, it ceases to play a meaningful attributive role in the representation so interpreted. Both of these processes are quite di erent from what we nd when a picture is understood to denote someone or something. In those cases, the picture singles out the individual by describing them in su

cient detail to distinguish them from other things.

The coded process behind iconography seems to be active in language as well, especially with what have been called code words and dogwhistles. In some cases, which I have called distracting codes, as opposed to spinning codes, the code word delivers a singular content. That content is often an ideology or the group of people supporting it, and they are presented in either an approving or disapproving manner. Importantly, distracting codes are their own communicative acts, not concerned to spin or alter one’s perspective on whatever the basic meaning of a sentence is. In many cases, the rest of a statement containing a distracting code word can simply be disregarded as unimportant to the conversational point. The one function that the underlying meaningful expression serves is to provide for a kind of deniability on the part of the speaker, who can disavow the distracting code in favor of the ordinary meaning of the sentence.

References Austin, J. L. 1962. How to Do Things with Words. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Camp, Elisabeth. 2018. “Insinuation, Common Ground, and the Conversational Record.” In Daniel Fogal, Daniel W. Harris, and Matt Moss, eds., New Work on Speech Acts. Oxford: Oxford University Press, 40–66. Google Scholar Google Preview WorldCat COPAC Freeland, C. 2010. “Photographs and Icons.” In S. Walden, ed., Photography and Philosophy: Essays on the Pencil of Nature. London: Blackwell, 50–69. Google Scholar Google Preview WorldCat COPAC Furie, Michael. 2016. Boyʼs Club. Seattle, WA: Fantagraphics Books. Google Scholar Google Preview WorldCat COPAC Greenberg, G. 2018. “Content and Target in Pictorial Representation.” Ergo 5: 865–898. Google Scholar WorldCat Hope, Charles. 1986. “Religious Narrative in Renaissance Art.” Journal of the Royal Society of Arts 134:804–818. Google Scholar WorldCat Hornsby, J. 1995. “Disempowered Speech.” Philosophical Topics 23 (2): 127–147. Google Scholar WorldCat Hyman, John. 2006. The Objective Eye. Chicago: University of Chicago Press. Google Scholar Google Preview WorldCat COPAC Khoo, J. 2017. “Code Words in Political Discourse.” Philosophical Topics 45 (2): 33–64. Google Scholar WorldCat Kulvicki, John. 2020a. “Colour and the Arts: Chromatic Perspectives.” In D. Brown and F. MacPherson, eds., Routledge Companion to the Philosophy of Color. London: Routledge, 91–106. Google Scholar Google Preview WorldCat COPAC Kulvicki, John. 2020b. Modeling the Meanings of Pictures. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Lillie, Amanda. 2014. ʻDomenico Veneziano, Saint Zenobius Restores to Life a French Widowʼs Sonʼ, published online 2014ʼ. In Building the Picture: Architecture in Italian Renaissance Painting. Exhibition catalogue. The National Gallery, London. http://www.nationalgallery.org.uk/research/research-resources/exhibition-catalogues/building-the-picture/placemaking/veneziano-miracle-of-saint-zenobius Google Scholar Google Preview WorldCat COPAC p. 545

Lincoln Project. 2021. Critical Race Card. Video, online only: https://lincolnproject.us/critical-race-card/ (accessed November 2021). WorldCat Lopes, Dominic. 1996. Understanding Pictures. Oxford: Clarendon Press. Google Scholar Google Preview WorldCat COPAC Lopes, Dominic. 2009. “Drawing in a Social Science: Lithic Illustration.” Perspectives on Science 17(1): 5–25. Google Scholar WorldCat

Mendelberg, Tali. 2001. The Race Card: Campaign Strategy, Implicit Messages, and the Norm of Equality. Princeton, NJ: Princeton University Press. Google Scholar Google Preview WorldCat COPAC Newall, M. 2006. “Pictures, Colour, and Resemblance.” Philosophical Quarterly 56:587–595. Google Scholar WorldCat Panofsky, Erwin. 1955. Meaning in the Visual Arts. Garden City, NY: Doubleday. Google Scholar Google Preview WorldCat COPAC Santana, Carlos. 2022. “Whatʼs Wrong with Dogswhistles.” Journal of Social Philosophy 53: 387–403. Google Scholar WorldCat Saul, Jennifer. 2018. “Dogwhistles, Political Manipulation, and Philosophy of Language.” In Daniel Fogal, Daniel W. Harris, and Matt Moss, eds. New Work on Speech Acts. Oxford: Oxford University Press, 360–383. Google Scholar Google Preview WorldCat COPAC Stanley, J. 2015. How Propaganda Works. Princeton: Princeton University Press. Google Scholar Google Preview WorldCat COPAC Thurston, H. and D. Attwater. 1956. Lives of the Saints. 4 vols. New York: Kenedy. Google Scholar Google Preview WorldCat COPAC Torices, J. R. 2021. “Understanding Dogwhistle Politics.” Theoria 36 (3): 321–339. Google Scholar WorldCat Taylor, P. n.d. “How Images Mean.” Unpublished manuscript. Google Scholar Google Preview WorldCat COPAC Walton, K. 1984. “Transparent Pictures.” Critical Inquiry 11 (2): 246–277. Google Scholar WorldCat

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

24 Visual and Linguistic Dogwhistles  Ray Drainville, Jennifer Saul https://doi.org/10.1093/oxfordhb/9780192844118.013.4 Published: 22 May 2024

Pages 546–570

Abstract The focus of the fast-growing philosophical literature on dogwhistles has been on linguistic dogwhistles. But there is a very important lacuna: the role of visual materials in dogwhistles, which is only glancingly touched on in the burgeoning dogwhistles literature. This chapter begins to rectify this oversight, with a detailed examination and taxonomy of the workings of visual dogwhistles, both on their own and as they interact with linguistic utterances. It shows that, like linguistic dogwhistles, visual dogwhistles may be categorized as Overt Code or Covert E ect and as intentional or unintentional. The discussion ranges from the history of art to modern political imagery and internet culture.

Keywords: dogwhistles, code words, implicit political messaging, racism, political communication, history of art, images, deniability Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction With the recent political turn in philosophy of language, several philosophers of language have begun to examine dogwhistles, albeit not always under that name (Stanley 2015; Khoo 2017; Saul 2018, 2019b, 2019a; 1

Henderson and McCready 2018, 2019, 2020; Santana 2021; Torices 2021). The exact de nition of ‘dogwhistle’ is a matter of debate, but—very roughly—these are utterances which function by concealing their meaning or intended e ect from at least some of the audience. This meaning of the term ‘dogwhistle’ originates in discussions of political advertising, and the primary focus of the literature has been on political manipulation. But this method of communicating is not con ned to manipulation in politics, and others sorts of dogwhistles have been discussed. However, there are still some very important gaps in this literature, one of which we aim to begin lling here.

Undoubtedly because they have been conducted by philosophers of language, the focus of the philosophical 2

discussion of dogwhistles has been on the linguistic. This is unsurprising, but it nonetheless represents a 3

very important lacuna. The best way to point to this lacuna is to brie y consider one of the key examples in this literature (which we will discuss in more detail later on), the Willie Horton ad used by the 1988 George H. W. Bush presidential campaign against Michael Dukakis. The words of this ad are largely irrelevant to the dogwhistle. The reason the ad is a dogwhistle is the image of Willie Horton. And yet, the role of images in 4

dogwhistles is only glancingly touched upon in the burgeoning dogwhistles literature. We propose to rectify this oversight, with a detailed examination and taxonomy of the workings of visual dogwhistles, 5

both on their own and as they interact with linguistic utterances.

p. 547

2. Background: Linguistic Dogwhistles Dogwhistles, some of which are also called ‘code words’ and ‘implicit political messaging’, have become a topic of interest due to their role in politics, especially in e orts to appeal to voters’ racist sentiments. In its original meaning, of course, a dogwhistle is a whistle that can only be heard by dogs, and not by people. The meaning at issue here is related to that: broadly, a dogwhistle is an utterance whose full meaning is only understood by—at most—a subset of audience members. (As noted, this characterization is broad. We will see that there are two very di erent sorts of dogwhistles that work in very di erent ways.) In its primary form, it will be carefully and intentionally constructed to function in this way. However, it is very important that—due to the concealment—people not in the know will unwittingly use or repeat dogwhistles. This is often part of the plan of those who create them, and it poses serious di

culties for those who hope to

combat dogwhistles and their e ects. The term ‘dogwhistle’ originates in media and online discussions (originally of opinion polling), and although it is often treated as a unitary phenomenon, there are two very di erent sorts of things which are referred to as ‘dogwhistles’. One sort of dogwhistle functions very much like a code. It is a message meant to be understood by only a part of the audience, but that subgroup in the know is meant to be fully aware of what’s going on. This we call an Overt Code dogwhistle, because to its intended audience it transmits an overt yet coded message. Another sort of dogwhistle is altogether stranger. It, too, has an intended audience. But that intended audience is meant to be a ected without being aware of it. Because this sort of 6

dogwhistle operates covertly upon its intended audience, we call this a Covert E ect dogwhistle. These two phenomena work very di erently, but both are popularly referred to as ‘dogwhistles’. They have in common 7

that their real purpose is meant to be concealed from at least part of the audience . The dogwhistles that are generally discussed, in both academic and popular literature, are ones where the concealed e ect runs contrary to some widely accepted social norm. Most often, this is a norm against racist speech. (See, for example, Saul 2018, and forthcoming, for discussions of how norms against racism play a role in dogwhistling.) Whether or not to build norm violation into a de nition of dogwhistling is a theoretical choice, and the decision that one makes will be a ected by one’s theoretical goals. Our focus here is a broad one, with a focus on the methodology of dogwhistling rather than on dogwhistling’s relationship to norms. Accordingly, we will not build norm violation into our de nition. However, we will occasionally note some of the ways that this decision a ects what we are saying.

p. 548

2.1. Overt Code Linguistic Dogwhistles Overt dogwhistles are designed to transmit at least two di erent messages, to at least two di erent audiences, at the same time. One message is generally an innocuous one, that would be judged acceptable by the audience at large; the other is one that merits concealment for some reason. The most commonly discussed Overt Code dogwhistles are racist ones. We get a very clear recent example of these from code words that racist groups on the internet have developed in order to communicate in such a way that they can avoid Twitter suspending their accounts. Very explicit codes have been developed and propagated, e.g. using ‘Google’ for Black people, ‘Skittle’ for Muslim, and so on (Kantrowitz 2016). The idea is that racists can communicate with each other, while other people—and algorithms—will not understand the racial associations of these dogwhistle terms. (These also have the advantage of being words that Twitter will be loath to censor, since they are names for major corporations.) If accused of racism, racist Twitter users can deny it by citing the standard meanings of the terms that they are using. Of course, once there are mainstream journalistic articles on the topic, these Overt Code dogwhistles become less e ective at concealing racist messages. It is worth noting for completeness, though, that not all dogwhistles will strike us as nefarious. Sometimes members of oppressed groups will want to transmit a coded message to one another. Polari, known as “Britain’s secret gay language” (Baker 2019), provides some great historical examples of non-nefarious Overt Code linguistic dogwhistles. This dialect developed out of language used, inter alia, in the theatrical and fairground worlds, and by the 1950s and 1960s provided a secret way that gay men communicated, which was not understood by outsiders—an important consideration at a time when gay sex was criminalized. Although Polari did involve some words that were speci c to Polari, other words were normal English words used with a meaning accessible only to insiders. These words function as Overt Code dogwhistles. Dudley Cave described Polari words as ‘secret passwords’. You could identify with other gay people if you thought they might be—you could drop a word in like “camping about” … Such code words would be likely to result in a nonplussed reaction if the hearer wasn’t gay so were a reasonably safe way of revealing one’s identity without being overly explicit, and could be shrugged o

as a simple

misunderstanding if someone had the wrong quarry. (Baker 2019: 128–129)

2.2. Covert E ect Linguistic Dogwhistles Covert e ect dogwhistles are quite di erent. They work on their targets without the targets being consciously aware of what is happening. The classic example of such a dogwhistle is the Willie Horton ad used by the p. 549

George H. W. Bush presidential

campaign against Michael Dukakis in 1988 and extensively studied by Tali

Mendelberg (2001). We know a lot about how this was designed to function thanks to the deathbed confession of its maker, Republican mastermind Lee Atwater. This ad featured the face of the convicted killer William Horton (called “Willie” in the ad, a deliberate change made by Atwater). It told the story of Horton committing rape and murder while on a furlough from prison, a program that was in place during Dukakis’s time as governor of Massachusetts. It did not use any racial terms of any sort. In an e ort to conceal the campaign’s origins, the ad was presented as the product of a small local organization and featured in one local TV market. However—as planned—the ad was picked up and shown over and over on the nightly news. When it was shown on the news, no racial aspects of the ad were initially discussed. It was either presented as an example of crime as a campaign issue or of negative advertising. During the time this advertisement aired, Americans were polled on their voting intentions and on their levels of racial resentment. Racial resentment is “A blend of anti-black a ect and the kind of traditional

American moral values embedded in the Protestant Ethic … a moral feeling that blacks violate such traditional American values as individualism and self-reliance, the work ethic, obedience, and discipline” (Kinder and Sears 1981: 416). It is measured by level of agreement with claims like the following (with racially resentful White Americans tending to agree more): It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well o

as whites.

(Tesler and Sears 2010: 19) Mendelberg studied what happened to racial resentment levels and voting intentions during the time when White Americans were increasingly exposed to the ad. She found that racial resentment levels did not change. However, there came to be an increased correlation between racial-resentment level and voting intention: more racially resentful White Americans became more likely to vote for Bush. The e ect was dramatic—it was, she argues, a key factor in causing Dukakis’s very large lead to fade away and turn into a 8

loss. What’s most interesting about her study, however, is what came a little later. Eventually, Jesse Jackson started speaking up about the role of race in the ad. His interventions were disrespectfully dismissed by the mainstream media, but they got a lot of airplay. Mendelberg studied their e ect. As soon as Jackson 9

highlighted the ad’s racial characteristics, the correlation previously observed started falling away.

Mendelberg’s theory of what took place is that the White Americans targeted by the ad harboured racial 10

resentment but did not want to see themselves as racist or as supporting racism.

Once Jackson raised the

topic of race, these people engaged in self-monitoring that kept the ad from working. The ad only worked if it could play on racial resentment outside viewers’ consciousness. It functioned by taking whatever racial attitudes people had and bringing them to bear on the topic under discussion (in this case, which candidate to vote for)—without their awareness. This is what makes it (in our terms) a Covert E ect dogwhistle: it is p. 550

crucial that the intended e ect not be consciously understood by its audience. Since

Mendelberg’s initial

study, other work has identi ed Covert E ect dogwhistle terms which function in the same way: ‘welfare’, ‘inner city’, and ‘government spending’ are well-studied examples. Use of these terms raises racial attitudes to salience and brings them to bear on whatever is under consideration—but without the conscious awareness of the person doing the considering (Hurwitz and Pe

ey 2005; Valentino et al.

11

2002).

In the Trump era, as the reader will be aware, it became less taboo to make overt reference to race in political speech. We don’t yet know all the factors that are playing a role in this, but one seems to have been the election of Barack Obama, which made many White Americans feel that anti-Black racism was a thing of the past. Experiments have shown that reminding White Americans of their vote for Obama increases their willingness to act on racial resentment. They have also shown that during Obama’s presidency, political speech making overt reference to race became more acceptable (Valentino et al. 2018). Importantly, though, the speech tested in these experiments is not as overtly racist as the speech of politicians like Donald 12

Trump.

So the particular examples in this section probably do not work in the same way now. However, the

way that they did work provides a clear illustration of what we mean by ‘Covert E ect dogwhistle’. The change over time in the reception of particular dogwhistles is in fact an important phenomenon which will 13

play a signi cant role in the discussion to follow.

2.3. Intentions and Dogwhistles Both Overt Code and Covert E ect dogwhistles may be used intentionally or unintentionally. So far, we have focused on their intentional use. But it quickly becomes apparent that this is not all that takes place. Part of the plan behind some of the most damaging intentional dogwhistles is that they will be spread unwittingly, as what Saul (2018: 376) has elsewhere called ‘ampli er dogwhistles’. So as discussed in Saul (2018), for example, once ‘government spending’ has come to be associated with racialized sentiments, innocent users of the phrase will not realize that they are introducing racialized connotations. Once ‘Google’ has come to mean Black person among certain groups on the internet, innocent utterances have the potential to communicate completely unintended messages if they are seen by the wrong audience. And this means that not-so-innocent people can claim (with varying degrees of plausibility) to be unaware of the underground meaning of the term that they use. This is very useful to those who propagate and use such dogwhistles. Importantly, the intentional/unintentional distinction crosscuts the Overt Code e ect/Covert E ect distinction. Either form of dogwhistle can be either intentional or unintentional. It will perhaps be helpful to have some simple examples to illustrate this: • Intentional Overt Code dogwhistle: Use of ‘Google’ to mean Black person by a member of far-right groups, carefully using the code the person has been taught. This is overt, because its intended p. 551

audience is meant to consciously grasp the

message. It is intentional, because the person using it

intends it to function in this manner. • Unintentional Overt Code dogwhistle: Use of ‘Google’ to mean the search engine by someone unaware of the far-right meaning, misinterpreted in an online forum. This is an unintentional use of a term that is an Overt Code dogwhistle. As a dogwhistle, the term is meant to function in an overt manner, with the intended audience consciously aware of the underground meaning. This usage, however, is an unintentional one, by someone unaware of this other meaning. • Intentional Covert E ect dogwhistle: There was a deliberate campaign to associate the term ‘welfare’ with Black people (Haney López 2015). The deliberateness makes this an intentional dogwhistle. However, this dogwhistle functions without the conscious awareness of those whose racial attitudes are activated by the term. This is what makes it a Covert E ect dogwhistle. • Unintentional Covert E ect dogwhistle: Because some people are unaware of the racial associations activated by the term ‘welfare’, and because welfare policy is a legitimate topic for political discussions, people may use the term ‘welfare’ without intending to activate racial associations, and without any awareness that they are doing so. These uses are unintentional because the people uttering the term do not intend it to function this way. They are covert because the e ects of this usage— activating the racial attitudes of the recipients—takes place without the awareness of those whose attitudes were activated.

3. Visual Dogwhistles Those familiar with the visual arts will have encountered the range of dogwhistles we have just described, though they will not, most likely, have thought of them in such terms. Artists have employed dogwhistles for a very long time—at the very least, since they have created works on behalf of rulers, religious hierarchies, the wealthy, and corporations to further these groups’ interests and self-conceptions. To illustrate this, we will provide examples from both the history of art and what is more broadly known as “visual culture”—that is, a range of works, from “high art” (such as unique panel paintings), to “low art”, work so ephemeral that their presence is barely registered (such mass-produced photography and illustrations in magazines). The concept of dogwhistles can help illuminate the tactics visual artists employ to nudge viewers into responding to representative art in ways conducive to produce speci c messages. The concept also lets us understand old concepts—such as iconography (to be explained in section 3.1.1)—in new ways. As with the dogwhistles outlined above, there are multiple types in the visual world, and they can be categorized as Overt Code or Covert E ect; and in addition, they may be purely visual or hybrid mixtures of text and image.

p. 552

3.1. Overt Code Visual Dogwhistles 3.1.1. Historical Before we discuss visual dogwhistles, it will be useful to take a short detour through the art-historical eld of iconography. Throughout the history of representative art, a common feature has been the use of repeated material that provides some sort of message. Let us take, for example, speci c headgear. In the European Middle Ages, a depicted person could be identi able as a king when the gure wore a metallic band, frequently ornamented with oral motifs that rise above that band. In contrast, peasants were frequently depicted with a close- tting cap held in place with a chinstrap. Over time and repeated exposure to such representations, viewers can come to di erentiate between kings and peasants quickly. And identi ed content extends beyond basic materials such as headgear: depending upon the gure’s perceived sex, and the objects she or he holds, we can identify the gure as a symbolic representation of ideas, e.g., Geometry, or The Church, depicted in human form. In turn, these details and gures build up into more complex quasi-narrative scenes as they interact with other gures both symbolic and real. Such personi cations survive to the modern day: a grey-green gure of a woman holding a torch and wearing a spiked crown is identi able as the Statue of Liberty, and its frequent representation in quasi-narrative scenes such as political cartoons has meant that it has come to represent the United States: what the statue does, or what happens to the statue, is transferred to mean that the same thing has been experienced by the United States. The study of these motifs and their use by visual artists is called iconography. Its use as a method of enquiry was most clearly de ned by Erwin Panofsky (1955/1982), who studied the didactic use of repeated symbolic representational imagery in the Middle Ages and Renaissance, but as noted, the same processes often 14

continue in contemporary imagery.

These representations most frequently contain personi cations (e.g.,

Justice depicted as a woman with a crown holding a scales and a sword), allegories (often multiple gures engaging in an activity, e.g., Liberty’s genitalia grabbed by Donald Trump implying that Trump is assaulting 15

the United States) , symbols (objects that convey a meaning deeper than their mere existence in a picture, e.g., the scales held by Justice signifying the evaluation of evidence), and attributes (an object associated with a gure that aids in identifying the gure, e.g., the torch of freedom held by the Statue of Liberty). As one might expect, iconographic motifs can only be successful if the intended audience can recognize the symbols and attributes present in a given representation—that is, they convey fairly obvious messages

when their content is recognizable as symbolic, and therefore, they are unlikely to be dogwhistles. However, an important evolution of Panofsky’s notion of iconography concerned the rendering of more naturalistic environments and its implication for the viewer’s reception of symbolic materials. He saw this development most clearly in fteenth- century Flemish oil panels and noted how it overturned the previous order of obvious symbolic content: “A way had to be found to reconcile the new naturalism with a thousand years of p. 553

Christian tradition;

and this attempt resulted in what may be termed concealed or disguised symbolism

as opposed to open or obvious symbolism” (Panofsky 1953/1971: 141). Thus, according to Panofsky, the small dog in the forefront of Jan van Eyck’s Arnol ni Wedding Portrait (1434; Figure 24.1) is a symbolic representation of marital delity, but its symbolic import is easily overlooked because it is “disguised” in the scene as just another exquisite detail in an otherwise convincingly rendered domestic interior 16

(Panofsky, 1934).

Although Panofsky did not use the term, for him, the dog was a dogwhistle, as was much

symbolic content in any work he thought would be likely to be interpreted as details in a naturalistic setting. Importantly for our current discussion, under this view, only a select few—those well-versed in theology— would be aware of the symbolism supposedly present in the panel. The vast majority of viewers likely would not recognize symbolic material as symbolic, and it is supposed that they would appreciate the panel purely as a marvellous depiction of the world. The panel—along with other instances of Panofsky’s disguised 17

symbolism—is thus an example of an Overt Code visual dogwhistle.

p. 554

Figure 24.1. Jan van Eyck (1434), Arnolfini Wedding Portrait. National Gallery, London. Source: Wikimedia, public domain (https://tinyurl.com/qf9gpjn). This is a place where our decision not to require norm violation matters. It is not at all clear that artists intended to “disguise” their use of symbolism because it violated some norm—indeed, the symbolism in the panel above seems well within its period’s philosophical, moral, and religious norms. It is, however, clear that Panofsky’s view depends upon symbolism being disguised, hidden by an overwhelming experience of naturalism, even though the use of these symbols does not appear to violate any norms. These, then, may be somewhat contentious examples of dogwhistles. 18

There are, however, many examples of norm-violating Overt Code visual dogwhistles in the history of art.

A prominent class of examples consists of imagery related to alchemy, a kind of proto-chemistry practiced throughout the medieval period up through the eighteenth century. Alchemical practitioners attempted to “transmute” metals from one to another, most famously to convert “base” materials such as lead into 19

gold.

In part, because it was thought that alchemists might be able to create gold—and thereby enrich

people or even destabilize whole economies—they wrote their books with tortuous symbolism to rebu unwanted readers (Principe 2013). Visual representations of alchemical experiments were typically depicted with equally esoteric symbolism, so that only the initiated could understand the content. For example, a passage in the fteenth-century alchemical tract Buch der Heiligen Dreifaltigkeit describes the “torment”

( ring and puri cation) of metals, and mixes quasi-biblical references with imagery of injured or tortured gures and household items that represent metals (Putscher 1986). In one illustration (Figure 24.2), a gure uses a lance to pierce Adam; the lance symbolizes puri cation through re, and Adam symbolizes 20

silver, the metal going through the heating process.

Figure 24.2. Anonymous (15th c.), Buch der Heiligen Dreifaltigkeit. Munich, Staatsbibliothek, BSB Cgm 598 (page turned 90° clockwise). Source: Munich Digital Library, marked for non-commercial use (https://tinyurl.com/y6nthbtl). Thus far we have discussed Overt Code dogwhistles which carry a message a knowledgeable viewer is meant to consciously understand, but an uninitiated viewer will miss. A further type of Overt Code visual dogwhistles—which to our knowledge has not yet been discussed in the literature on linguistic dogwhistles —is meant to give rise to a certain emotional or physiological reaction, consciously experienced, in its target audience—one which is at least deniable to a broader audience. These responses can stem from the manner of depiction and the poses employed by the artist, but we often have to guess at the viewer’s response: let us take an example from the work of “Salon” or “Academic” painters and sculptors of the nineteenth 21

century.

Salon artists’ depiction of nudity in Classical subjects is particularly relevant here, portrayed in large paintings intended for wealthy clientèle and galleries. Although these artists often painted well-trodden subject matter—gods and passages from history—they portrayed the naked female body rather more sexually than previous artists had. Alexandre Cabanel’s Birth of Venus (1863; Figure 24.3) is an example of this: in his portrayal of the goddess of love, the artist accentuates the gure’s sinuous curves through her 22

pose.

The curves and the pose, along with her veiled gaze at the viewer, suggest that the image is more 23

bedroom scene than birth scene.

Here, the painting seemingly conveys two messages: one, a super cially

Classical subject that makes it acceptable for public viewing, and another which was likely intended to p. 555

appeal to male sexual

p. 556 24

delectation.

Both are present on the canvas, but at its time of creation, only the mythological message

could be fully acknowledged publicly.

Figure 24.3. Alexandre Cabanel (1963), Birth of Venus. Musée dʼOrsay, Paris. Source: Wikimedia, labelled for reuse (https://tinyurl.com/y2lz88nx). This interpretation is not as far-fetched as it may initially appear. It is intimately connected with increasingly sexualized renditions of Venus/Aphrodite as part of a wider nineteenth-century fascination with the eroticized female. Depictions of Venus had become increasingly eroticized, not simply due to her identi cation as the goddess of love, but through associations with the ancient Greek courtesan and legendary beauty Phryne—associations which go back to the Classical era, but which were only visually explored by the Salon painters. In the Deipnosophist, Athenaeus (ca. 200 CE/1854: XIII.59/1874) relates that the sight of Phryne’s ritual public bath during the Poseidonia festival was the source for Apelles’s painting (now lost) of the birth of Aphrodite. Moreover, multiple stories about Phyrne transferred her supposedly available sexuality to the goddess. She was said to be Praxiteles’s model when he sculpted the Aphrodite of 25

Knidos—the rst votive statue of the goddess depicted naked.

In the Amores, Pseudo-Lucian (c. 300

CE/1967:13–16) relates several stories of men approaching this statue with sexual intent, the most notorious relating to a “stain” left on the gure’s thigh, put there by a man who snuck into the temple at night and ejaculated on it. Such stories—cleansed of some of the more prurient details—were revived in the nineteenth century and became part of sexual fantasies played out in di erent media. Classically oriented artists often referenced Phryne directly. Charles Baudelaire (1857/1993: 235) fantasized about multiple Phrynes cavorting with one another in the 1857 poem “Lesbos” (“Lesbos, where Phrynes seek each other out”), and the composer Camille Saint-Saëns (1893) wrote an opera about her in 1893. Phryne featured directly in a number of paintings in the period (e.g., by Gustave Boulanger [1850], Artur Grottger [1867], Henryk Siemiradzki 26

[1889], José Frappa [c. 1903], inter alia).

One of the better-known renditions is that by Jean-Léon Gérôme (1861), who painted Phryne before the 27

Areopagus (1861; Figure 24.4).

Phryne was placed on trial for impiety much as Socrates was; just as it

appeared she would lose both her case and her life, her defending magistrate tore o

her robe to “excite the

pity of her judges by the sight of her beauty”, such that they were not “able to stand the idea of condemning to death ‘a prophetess and priestess of Venus’ ” (Athenaeus, c 200 CE/1854:XIII.59). In Gérôme’s painting, the jurors are depicted reacting to the sudden display of nudity in di erent ways: ranging from shock and dismay, to clinical interest, to obvious pleasure at the view. Paintings directly relating stories of Phryne, then, already carry double messages: one that relates ancient stories for a cultivated audience, but another that o ers the vision of a beautiful woman who made her fortune through sex. More broadly, however, the Salon artists who painted sexualized renditions of Venus were well aware of these stories and drew upon the objecti cation of Phryne to render Venus, not just as the goddess of love, but as sexually available and sexually tempting, much as a courtesan was considered to be. Like paintings of Phryne herself, sexualized

paintings of Venus are at once clothed in Classical subject respectability and disrobed for the delight of the (male) viewer.

p. 557

Figure 24.4. Jean-Léon Gérôme (1861), Phryne before the Areopagus. Kunsthalle, Hamburg. Source: Wikimedia, labelled for reuse (https://tinyurl.com/yxz2k2qc).

3.1.2. Contemporary Overt dogwhistles are quite common in contemporary politicized imagery. One example of this would be “Pepe the Frog”. Originally a gure in Matt Furie’s cartoon “Boy’s Club”, the simply-drawn gure was coopted by the far or “alt right” online, having been incorporated into a cluster of memes referencing concentration camps and Christian identity movements (Pelletier-Gagnon and Trujillo Diniz 2018; 28

Zannettou et al. 2018).

Pepe also gured as a pro-Trump gure during the 2016 presidential campaign,

where the “smug Pepe” variant—wearing a sly smile and stroking its chin—served as a stand-in for Trump 29

in a picture that the candidate himself retweeted.

Ostensibly a tangential addition to the pro-Trump online

repertoire, the connection of Pepe to alt-right racism points to the likely dogwhistle: that the alt right 30

supported Trump and considered him a candidate who held views similar to theirs.

This intentional Overt Code dogwhistle also led to unintentional Pepe dogwhistles. And these unintentional dogwhistles served a valuable purpose for the alt right. The popular, pre-existing practice of sharing the Pepe imagery—most of which has nothing to do with racism or Trump—would, potentially at least, suggest that there was broader support for the alt right’s White supremacist views and for Trump than there in fact was. Moreover, people might unintentionally share images with alt-right dogwhistles thinking they were simply Pepe images. This helped to spread the alt right’s message and to increase its public prominence. p. 558

A similar mechanism (involving intentional and unintentional Overt Code dogwhistles) appears to have been in play in the co-opting of the ‘OK’ hand gesture in a symbol for White supremacy (the idea being that the gesture approximates the shape of the letters ‘w’ and ‘p’ in the phrase ‘White power’). Originally organized as a way of “owning the libs”, the gesticulating person’s intention has been purposely muddied so that the viewer cannot tell whether the symbol was ashed innocently or as a signal to fellow White supremacists (Neiwert 2018). This is useful in three ways. First, it o ers deniability to one who wants to send a signal to their fellow White-power enthusiasts. Second, this denial allows for mockery of those who have taken o ense, by insisting they are overinterpreting a well-known innocent gesture. Finally, the popularity of the ‘OK’ gesture could give an impression of a far greater adherence to White supremacy than actually existed in the general population.

More recently, during the initial phase of the protests following the 2020 police murder of George Floyd in Minneapolis, demonstrations were frequently attended by an assortment of gun-toting gures wearing camou age and gaudily patterned Hawaiian shirts. These were the “Boogaloo Bois”, individuals who anticipate a second Civil War in the United States, some of whom wish to accelerate the occurrence of this 31

con ict (Green 2020; Evans and Wilson 2020).

Largely organized online, they have attempted to evade

algorithmic detection by making reference to terms with a similar ring to ‘Boogaloo’: ‘Big Igloo’ and, more recently, ‘Big Luau’. In keeping with the latter term, they wear Hawaiian shirts. To the outsider, they seem like oddly dressed gures; to the in-group, they can identify like minds in a crowd. This is an Overt Code intentional dogwhistle, meant to help them pick out their compatriots while maintaining deniability. What is more, unwitting lovers of Hawaiian shirts will, of course, nd themselves unintentionally dogwhistling.

3.2. Covert E ect Visual Dogwhistles Covert e ect dogwhistles work on the audience without the target audience being consciously aware of their impact. In apolitical contexts, it has long been suggested that some advertising works on subconscious levels using a series of tricks to stimulate the viewer’s needs; claims centred on subliminal advertising have been largely debunked, but there appears to be solid evidence for subliminal priming, wherein the viewer is more likely to, e.g., select an object after it is presented for a few milliseconds (Packard 1957/2007; Nelson 2008; Pratkanis and Greenwald 1988; Elgendi et al. 2018). It is di

cult to prove whether unconscious

reception of visual material may successfully impart an intended message surreptitiously, whether or not its makers intended to do so. The Willie Horton case discussed earlier is an unusually clear one: Lee Atwater’s deathbed confession informed us of his intention, and the psychological studies showed us important things about its e ects. For most cases, sadly, we lack such veri cation. The examples here are therefore rather more speculative, although some appear to clearly employ or alter imagery to in uence the viewer’s perception.

p. 559

3.2.1. An Historical Example Peter Paul Rubens depicted Marie de’ Medici arriving at Marseilles, received by a personi cation of France 32

as part of a painted cycle celebrating her marriage to Henry IV of France (c. 1625; Figure 24.5).

Intended to

be hung in the Luxembourg Palace, the paintings in the series tower over the viewer, and it is Rubens’s use of perspective and composition that is relevant here. Shown arriving in France, Marie is represented on the canvas as if she were literally above the viewer: it is suggestive of the respective social positions of the future queen and the viewer in the period’s social hierarchy. Such an interpretation was probably intended p. 560

for a work presented in a palace setting, and made

more likely by Rubens’s extensive diplomatic work

reinforcing the status quo (Lamster 2010). The act of viewing the panel on its own reinforces a social hierarchy which places the queen above the viewer, and can be read as a covert method of manipulating viewers to unconsciously take on a subsidiary position in the hierarchy. This is, then, a plausible example of a Covert E ect visual dogwhistle.

Figure 24.5. Peter-Paul Rubens (1625), Arrival of Marie deʼ Medici at Marseilles. Palais du Luxembourg, Paris. Source: Wikimedia, labelled for reuse (https://tinyurl.com/y6ncyhg5).

3.2.2. Contemporary Examples The use of unconscious triggers in advertisements to awaken a viewer’s desires or needs is, as we have noted, in some ways contentious, but unconsciously registered elements of images nevertheless likely in uence the viewer’s perception of imagery, something long intuited by graphic designers. Roland Barthes (1977/1984) presents a classic example of this in his analysis of a photographic advertisement for Panzani, a 33

French brand selling Italian food products.

Both the Panzani logo and the composition of the pasta,

cheese, and fresh vegetables are bathed in red, white, and green colours, which help convey “Italianicity” for the viewer, since the colour scheme matches the Italian tricolour ag and the ingredients connote fresh 34

Italian cooking.

Barthes presents these as signs intended to evoke widely shared cultural references

(perhaps even stereotypes) in the viewer, and though the viewer is unlikely to register the Italian tricolour scheme consciously, the advertiser’s idea is clearly that referencing the tricolour will make viewers associate Panzani’s products with Mediterranean food, freshly cooked with quality ingredients. (This is another example that would not count as a dogwhistle if we required norm violation.) The aforementioned Lee Atwater was involved in the production of another political commercial for the 35

1992 US presidential election, usually referred to as the “Revolving Door” ad.

The piece again slams

Dukakis as being soft on crime for supporting Massachusetts’s prison furlough policy, and it depicts

prisoners walking in and out through a revolving door—suggesting that they are leaving, and then returning to, prison. Viewers of the commercial tend to think that a high proportion of the prisoners are Black when in fact, only two are (Jamieson 1992: 31). The increased racialized perception may arise because one of the two Black prisoners looks up at the camera as the prisoners pass through the door—he is the only one to do so—and this was likely perceived as threatening to the viewer. It seems that viewers unconsciously picked up on the prisoner’s “threatening” glance, which was enough to trigger racial biases in their reception of the commercial. Altered skin colour in political advertisements also give us some clear instances of Covert E ect visual dogwhistles, some intentional and some not. There is an extensive literature in psychology showing 36

linkages between negative stereotypes of Black people and darker skin colour.

There have also been

accusations that the political campaigns for White candidates have darkened the skin tone of Black 37

opponents in political advertisements.

 Messing et al. (2016) conducted a careful study of the e ects of

digitally altering Barack Obama’s skin tone. Participants rst viewed either an image of Obama with lightened skin or with darkened skin and then completed a standard task of negative stereotype activation (completing words related to negative stereotypes about Black people). They found that that those who had p. 561

seen the darker Obama showed

greater activation of negative stereotypes about Black people. We take

this to be a very plausible example of a Covert E ect visual dogwhistle: it is unlikely that participants are consciously aware of the factor—darkness of skin—which is producing the e ect of activating their stereotypes. What is less clear about darkened skin dogwhistles is whether they are intentional or not. The Willie Horton case is unusual, as we noted, in that Lee Atwater explicitly stated his intentions. In most cases, we lack this information. And in digital darkening cases, it is especially tricky to discern intention, because it is very easy to accidentally darken images—for example, taking imagery processed in a particular colour calibration and 38

placing it in a di erent colour space can radically change the colour range.

In addition, it is common in

political attack ads to alter imagery showing an opponent: colour is frequently desaturated, and colour balance is clipped so that darker colours are made more uniformly stark. We know, then, that these images are Covert E ect dogwhistles because of what they do. But in many cases we cannot be certain whether they are intentional or not.

3.3. Hybrid Visual/Linguistic Dogwhistles So far, we have been (mostly) writing as though there is a clear distinction between visual and linguistic dogwhistles. And there is in some cases: the phrase ‘inner city’, uttered in a radio broadcast, is purely linguistic. And a painting will generally be purely visual. But as we noted earlier, the classic dogwhistle example, the Willie Horton advertisement, is both visual and linguistic: the words discuss the prison policies in force during Dukakis’s time as governor and Willie Horton’s crimes, but say absolutely nothing about race. As we have seen, the ad’s most important e ect is the priming of racial attitudes, which takes place entirely through the images. (All of the other advertisements we discuss are also blends of words and images.) The idea that words and images often function together is not a new one. The media theorist W. J. T. Mitchell (2005) observed that those items we term “visual media” are rarely exclusively visual: they frequently incorporate audio (e.g., a video), or include touch (e.g., a sculpture), and that many visual media at the very least reference textual media (e.g., a painting of Phryne referencing a written story about her). Moreover, many visual media directly partner with other media to tell a story. It is often in the interplay between these modalities, e.g., spoken word and projected image that a message is conveyed, and were one to consider either spoken word or projected image exclusively, the message itself would fall apart. But the interplay

between di erent media can also provide the space to deny any objectionable intent, and political commercials in particular often operate in this hazy area. A nal example might serve to illustrate how imagery works alongside text. The QAnon conspiracy theory originally took hold on the fringes of American politics, but increasingly since 2020 has moved to the mainstream of the Republican Party with the election to the US Congress of such true believers as Marjorie Taylor Green and Lauren Boebert. Insofar as it coherent, the theory can be summarized as: an anonymous p. 562

gure known as “Q” (a pseudonym re ecting his security clearance to tra

level) has revealed a vast conspiracy

c, kill, and consume children; prominent members of the Democratic Party, in particular Hillary

Clinton, gure as the villains, as do international “globalist” gures such as George Soros. Donald Trump is working to destroy this entire evil infrastructure, and the villains’ arrests are imminent, but always delayed; the term “QAnon” refers to the community who support Q and Trump in this battle (LaFrance 2020). A number of Instagram in uencers have embraced the conspiracy theory, presented in the stylized “design look” popular on the platform—light earth tones, single dominant colours, loopy fonts, and low-detail imagery, all employing lters and all aligned with the bland aesthetics of an Ikea living room or hipster 39

co ee shop.

The aesthetics and the subjects—family, food, and fun—are therefore familiar and anodyne,

but the accompanying text and hashtags use themes familiar to QAnon, such as “WWG1WGA”, “The Great Awakening”, “Pain Is Coming”, and an overwhelming obsession with child tra p. 563

Given the speci c a ordances of Instagram as a platform and their

40

cking (Figure 24.6).

users’ device usage, few viewers will

see the full text of a post as they “heart” a given photo: full textual content is only accessible by clicking the “More” link, so hashtags appended at the end of a post will be hidden from view by casual scrollers. The result is that these in uencers gain exposure to their cause by people who upvote a photo, but who do not themselves necessarily see the text. QAnon adherents, however, interact with the post’s text, clearly understanding its intent (Ti any 2020). These posts, seemingly standard Instagram fare, then act as Overt Code dogwhistles for QAnon adherents.

Figure 24.6. Instagram post (mobile and desktop interfaces) by influencer Maddie Thompson. Note the reference to children and the use of the tags #greatawakening and #painiscoming, common Qanon tropes viewable by default only from the desktop interface. Source: Maddie Thompson Instagram account (https://tinyurl.com/y45vu5ar). Our contention here is that very many dogwhistles are mixtures of visual and linguistic elements. An analysis that focuses purely on the linguistic will miss a great deal of what is happening. Fortunately, it turns out to be fairly straightforward to apply the categories developed for linguistic dogwhistles to both images and the visual elements of mixed dogwhistles. It is also, we hope to have shown, illuminating.

4. Conclusion Any analysis of dogwhistles that focuses only on the linguistic side will miss a good deal of what is taking place in real examples of manipulation, be they political or otherwise: there is often no dogwhistle in the linguistic part of a political advertisement, but a very signi cant dogwhistle in the visual part, or in the juxtaposition of the two. Moreover, such analyses will fail entirely to see those dogwhistles which are wholly visual. And it is particularly important to notice and discuss the dogwhistle present in a Boogaloo adherent’s Hawaiian shirt, or a neo-Nazi’s use of the OK sign. To fully understand, analyze, and arrive at ways of responding to dogwhistles, all their elements must be understood. Applying dogwhistle classi cations from the philosophy of language to visual dogwhistles can help to clarify what happens in these cases. Considering visual cases also allows us to integrate important insights from art history and the study of visual culture more generally. We hope to have shown what a rich interdisciplinary territory this is for investigation, and also to have demonstrated the importance and urgency of doing so. Many of the most important elements in our political culture have always concerned things which politicians hesitate to say openly, but in our current political moment it is especially important to notice and appreciate how these things are being communicated in less open ways. The analysis of both linguistic and visual dogwhistles is vital to this endeavour.

Acknowledgements We are very grateful for immensely helpful discussions of this paper with audiences at the University of Arkansas, the University of Saskatchewan, UQAM, the HALO workshop, the ROGAP reading group, and the Words Workshop, as well as valuable feedback from the editors and a referee.

p. 564

p. 565

Notes *

There is currently no clear consensus, even amongst anti-racist scholars, about the capitalization of ʻWhiteʼ and ʻBlackʼ. We have chosen to follow the arguments of Kwame Anthony Appiah (2020), who argues for capitalizing both as a way of drawing attention to their socially constructed nature. However, within quotations we have followed the practice of the original authors.

1.

Stanley (2015), for example, discusses them as instances of propaganda. Our view is that while dogwhistles may sometimes serve as instances of propaganda, they need not. Some dogwhistles may be codes which facilitate secret communication in a particular group without any propagandistic function.

2.

One of us is amongst these philosophers of language, so perhaps ʻtheyʼ should be ʻweʼ.

3.

José Medina (2018) points to a similar lacuna in discussions of racist propaganda, noting that despite the importance of visual and mixed visual and linguistic propaganda, philosophical discussion has focused overwhelmingly on the linguistic. Medinaʼs paper is an extremely important discussion of more blatantly racist visual propaganda (e.g. lynching images) than that discussed here.

4.

Saul (2018, 2019a) are no exception to this.

5.

Where possible, this paper uses imagery labelled for reuse; in other instances, it makes direct reference to resources on the web using the “TinyURL” service for the readerʼs convenience.

6.

In the everyday sense of the term, all dogwhistles are covert. The term ʻCovert E ect dogwhistleʼ is used in technical sense when contrasted with ʻOvert Code dogwhistleʼ.

7.

Saul (2018, 2019) simply called these ʻcovertʼ and ʻovertʼ dogwhistles. However, the intervening years have revealed that many people found this terminology confusing. In order to provide further clarity, these are now called ʻCovert E ectʼ and ʻOvert Codeʼ dogwhistles. The new terminology emphasizes the large di erence in the way these two kinds of dogwhistles function.

8.

There is, however, debate over just how big a role it played relative to other factors. See Slides (2016).

9.

Jacksonʼs intervention occurred late in the campaign cycle. Indeed, Mendelberg argues that if the election had been held a couple weeks later Dukakis might have won.

10.

Although itʼs not the focus of Mendelbergʼs discussion, it is important to note that some White Americans do in fact feel perfectly comfortable with being racist or supporting racism. Covert e ect dogwhistles are not needed to reach these people, though Overt Code dogwhistles may be useful to avoid alienating other voters. For more on this, see Saul (2017, 2019b).

11.

A referee has expressed the concern that our understanding of covert dogwhistles may be so broad as to include baking cookies when one is trying to sell oneʼs house, in order to have an unconscious positive a ect on prospective buyers. This is the kind of example that might make one want to build norm-violation into the definition. Since there is nothing normviolating about cookies—cookies may be safely classified as norm-consistent—this case would very clearly not count as a dogwhistle once that was built in. However, the mechanism in such broader examples bears important similarities to norm-violating examples, hence our interest in leaving open the option of norm-consistent dogwhistles.

12.

For a discussion of factors that enabled the rise of Trump-like speech, see Saul (2019b).

13.

Questions have been raised about the accuracy of Mendelbergʼs picture and the existence of this sort of (in our terms)

Covert E ect dogwhistle (Huber and Lapinski 2006). We find Mendelbergʼs (2008) response convincing.

p. 566

14.

See also Roelof van Straten (1999), who, inter alia, has pursued ICONCLASS, an encyclopaedic database of iconographic materials.

15.

See Lennart Gäbelʼs “They Let You Do It”, a 2016 illustration made for Westdeutscher Rundfunk about the Trump campaign: https://tinyurl.com/y2hu7apl (accessed 3 December 2023).

16.

National Gallery, London. Available at https://tinyurl.com/zh2f63x (accessed 3 December 2023).

17.

Panofskyʼs view of symbolism as “disguised” or “driven underground” has been challenged, however. See Pächt (1956, 1977/1999); Delaissé (1957); and Bedaux (1986) for some fundamental criticisms of the idea. See also Seidel (1989) and Harbison (1990) for significantly di erent interpretations of the Arnolfini Wedding Portrait. This is not the place to delve into these challenges, but it is nevertheless clear in other contexts that the use of symbolic motifs provides examples of surreptitious messaging. What matters most is Panofskyʼs conception of  “disguised” symbolism.

18.

For a prominent example from the Renaissance, see Donatelloʼs (c. 1440) naked adolescent David (National Museum of the Bargello, Florence), available at https://tinyurl.com/ycfyy7cr (accessed 3 December 2023). This bronze sculpture has been interpreted by multiple scholars as an eroticized image of youth, for philosophical, political, and/or homosocial purposes; these are interpretations which certainly would not have been common amongst most people at the time of its creation (Randolph 2002: 138 ).

19.

Since the seven known metals were conceived of as composites and not elements, it was thought that an extended process of heating, purification, and combination could extract their essences, freeing practitioners to combine these essences and create other metals, notably gold. Alchemyʼs theoretical underpinnings are furthermore based upon the characteristics of the four elements, numerical symbolism, and a meaningful correspondence between varied materials and concepts, features common in medieval thought; thus, an analogy between two otherwise unrelated items was considered illuminating to both of them, e.g., a religious text and a metallurgical process. Alchemy evolved through the eighteenth century and expanded into a syncretic mix combining religious belief with quasi-philosophical ideas, the four humours, the seven planets, and the Kabbala, and rather a lot more besides, none of which was expressed plainly (Roob, 2014). Alchemyʼs later development will remain unexplored in this article, although it, too was visually represented with particularly opaque esoteric symbolism.

20.

Anonymous (c. 1450), Buch der Heiligen Dreifaltigkeit, Munich, Staatsbibliothek, BSB Cgm 598. See the two facing folios available at https://tinyurl.com/yxw4nhry and https://tinyurl.com/y6nthbtl, respectively (accessed 3 December 2023).

21.

Successful in their time, these artists are not widely known today, as the painting style and occasionally saccharine depictions of their subject matter has had little influence on the development of painting and sculpture in the past century. Salon art has, however, been influential in other spheres, notably in shaping cinematic depictions of the Classical world among others. For example, the cinematography in Ridley Scottʼs Gladiator is known to have been influenced by Lawrence Alma-Tademaʼs paintings of Roman scenes (Prettejohn et al., 2016).

22.

Musée dʼOrsay, Paris. Available at https://tinyurl.com/y5trsk3w (accessed 3 December 2023).

23.

This was coyly remarked upon at the time by the critic Théophile Gautier, who wrote: “The swell makes her body arch and accentuates her youthful charms all the more strongly” (Blühm 2011: 31). More recent art historians have also—rather coyly—noted the sexualized portrayal of the painting. The figure of Venus “hovers somewhere between an ancient deity and a modern dream” (Kern 1996: 101), a phrase more comprehensible if “modern dream” is a euphemism for a Playboy centrefold or a sexualized fantasy. Connections between art, the male viewer as the intended recipient of sexual availability, and the tropes of pornography have been noted before: see Berger (1972: 54–57).

24.

Some non-males surely enjoyed this too, but it was almost certainly not intended for this audience.

25.

See Anonymous, a er Praxiteles (c. 50 CE). Munich Glyptothek (Roman copy). Available at https://tinyurl.com/yylvxvza. (Accessed 3 December 2023).

26.

See: Boulanger (1850), available at https://tinyurl.com/yauz84nc; Grottger (1867), available at https://tinyurl.com/rejf44wp; Siemiradzki (1889), available at https://tinyurl.com/43e49jt5. There is also an earlier painting by J. M. W. Turner (1838; available at https://tinyurl.com/2tc8265h) ostensibly of Phryne, but the figures are

miniscule, and the artist focused on the landscape. Angelica Kaufmann painted Phryne still earlier, but unlike every other example here, Phryne is fully clothed in her panels. (Kaufmann [1794a, 1794b], available at https://tinyurl.com/yp2p6sy6 and https://tinyurl.com/yeys6hj8 respectively). (All links accessed 3 December 2023).

p. 567

27.

Kunsthalle, Hamburg. Available at https://tinyurl.com/yxz2k2qc. (Accessed 3 December 2023).

28.

The 2008 comic wherein Pepe employs the stock phrase “Feels good, man” may be found at Vice: https://tinyurl.com/yyglvgk7 (accessed 3 December 2023).

29.

“Canʼt Stump the Trump”, retweeted by Trump and preserved on the website Know Your Meme: https://tinyurl.com/y682rmb4 (accessed 3 December 2023).

30.

Pepeʼs insertion into a vehicle of communication is complex. The right-wing Pepe figure was first employed as a prank on journalists prior to turning into an actual alt-right symbol; it was also even used in the Hong Kong protests, and yet it concurrently remained a fan favorite without any political trappings whatsoever (Echevarria 2020).

31.

They call themselves “Boogaloo” a er the 1984 movie Breakinʼ 2: Electric Boogaloo. As with the co-opting of the “OK” hand signal or the Pepe character, their use of the term “Electric Boogaloo” is also a co-opting of online usage referring to anything with a second part, or the second in the series (e.g., “World War II: Electric Boogaloo”). For more, see Wiggins (2020).

32.

Palais du Luxembourg, Paris. Available at https://tinyurl.com/y6ncyhg5 (accessed 3 December 2023).

33.

Available at https://tinyurl.com/34u4uwmu (accessed 3 December 2023).

34.

According to Barthes, other elements conveying “Italianicity” include the fresh vegetables and the small mesh bag, other elements common to Italy, or at least to oneʼs stereotypes of Italy.

35.

The commercial may be seen on YouTube: https://tinyurl.com/4rsb5me2 (accessed 3 December 2023).

36.

For a useful summary of this literature, see Messing et al. (2016: 46–47, 55).

37.

See, for example, Kolawole (2008), who discusses whether the Hillary Clinton campaign darkened Obamaʼs face in a political ad.

38.

See Kolawole (2008) for side-by-side comparisons illustrating such an e ect.

39.

For more on the “design” aesthetic pursued on Instagram, see Manovich (2017: 67 ), who based his categorization on an analysis of approximately 15 million pictures posted to the platform.

40.

For an example, see https://tinyurl.com/y45vu5ar (accessed 3 December 2023). “WWG1 WGA” stands for “Where we go one, we go all”, a line from the obscure film The Squall and which serves as a rallying cry for QAnon.

References Anonymous, a er Praxiteles (c. 50 CE). Aphrodite of Knidos. Roman copy of a lost Greek original. Munich: Glyptothek. Accessed 3 December 2023. https://tinyurl.com/yylvxvza. Google Scholar Google Preview WorldCat COPAC Anonymous (c. 1450). Tormenting of the Metals. from Buch der Heiligen Dreifaltigkeit. Staatsbibliothek, BSB Cgm 598, Munich. Accessed 3 December 2023. https://tinyurl.com/y6nthbtl. Google Scholar Google Preview WorldCat COPAC Appiah, Kwame Anthony (2020), ʻThe Case for Capitalizing the B in Blackʼ, The Atlantic, accessed 3 December 2023, https://tinyurl.com/v4n58u3h. Athenaeus (c 200 CE/1854), Deipnosophistae, The Deipnosophists. Or Banquet of the Learned of Athens. 1854 ed. (London: Henry Bohn). Baker, Paul (2019), Fabulosa! The Story of Polari, Britainʼs Secret Gay Language (London: Reaktion). Google Scholar Google Preview WorldCat COPAC Barthes, Roland, ed. (1977/1984), ʻRhetoric of the Imageʼ, in Image, Music, Text (London: Fontana), 32–51. Google Scholar Google Preview WorldCat COPAC Baudelaire, Charles (1857/1993). “Lesbos”. Flowers of Evil. Transl. James McGowan. Oxford: Oxford University Press. 232–238. Google Scholar Google Preview WorldCat COPAC Bedaux, Jan Baptist (1986), ʻThe Reality of Symbols: The Question of Disguised Symbolism in Jan van Eyckʼs “Arnolfini Portrait”,ʼ Simiolus, 16/1, 5–28. Google Scholar WorldCat Berger, John (1972/1986), Ways of Seeing (London and Harmondsworth: British Broadcasting Corporation and Penguin). Google Scholar Google Preview WorldCat COPAC Blühm, Andreas (2011), Alexandre Cabanel: Tradition of Beauty (Munich: Hirmer Verlag). Google Scholar Google Preview WorldCat COPAC Boulanger, Gustave (1850). Phryne. Van Gogh Museum, Amsterdam. Accessed 3 December 2023. https://tinyurl.com/yauz84nc. Google Scholar Google Preview WorldCat COPAC Cabanel, Alexandre (1863), Birth of Venus. Musée dʼOrsay, Paris. Accessed 3 December 2023. https://tinyurl.com/y2lz88nx. Google Scholar Google Preview WorldCat COPAC Delaissé, Léon Marie Joseph (1957), ʻEnluminure et peinture dans les Pays-Bas: À propos du livre de E. Panofsky Early Netherlandish Paintingʼ, Scriptorium, 11/1, 109–118.

Donatello (c. 1440). David. Florence: Bargello. Accessed 3 December 2023. https://tinyurl.com/ycfyy7cr. Google Scholar Google Preview WorldCat COPAC Echevarria, Gina (2020), ʻHow This Frog Meme Became a Symbol of Hope and Hateʼ, Business Insider, accessed 3 December 2023, https://tinyurl.com/3tsr949a. Google Scholar WorldCat Elgendi, Mohamed, Kumar, Parmod, Barbic, Skye, Howard, Newton, Abbott, Derek and Cichocki, Andrzej (2018), ʻSubliminal Priming: State of the Art and Future Perspectivesʼ, Behavioral Sciences, 8/6, 54, https://doi.org/10.3390/bs8060054.

Google Scholar

WorldCat

Evans, Robert and Wilson, Jason (2020), ʻThe Boogaloo Movement is Not What Youʼd Thinkʼ, Bellingcat, accessed 3 December 2023, https://tinyurl.com/ycx482se. Google Scholar WorldCat Frappa, José (c. 1903). Phryne. Paris, Musée dʼOrsay. Accessed 3 December 2023. https://tinyurl.com/ycx5u2c4. Google Scholar Google Preview WorldCat COPAC p. 568

Gérôme, Jean.-Léon (1861). Phryne Before the Areopagus. Kunsthalle, Hamburg. Accessed 3 December, 2023. https://tinyurl.com/yxz2k2qc. Google Scholar Google Preview WorldCat COPAC Green, Jordan. (2020, May 28), ʻThey Want their Civil Warʼ: Far-Right ʻBoogalooʼ Militants are Embedded in the George Floyd Protests in Minneapolisʼ, Raw Story, accessed 3 December 2023, https://tinyurl.com/39k243 . WorldCat Grottger, Artur (1867) Phryne. Czartoryski Museum, Krakow. Accessed 3 December 2023. https://tinyurl.com/rejf44wp. Google Scholar Google Preview WorldCat COPAC Haney Lopez, Ian Haney (2015), Dog Whistle Politics: How Coded Racial Appeals Have Reinvented Racism and Wrecked the Middle Class (Oxford: Oxford University Press). Google Scholar Google Preview WorldCat COPAC Harbison, Craig (1990), ʻSexuality and Social Standing in Jan van Eyckʼs Arnolfini Double Portraitʼ, Renaissance Quarterly, 43/2, 249–291. Google Scholar WorldCat Henderson, Robert, and McCready, Elin (2018), ʻHow Dogwhistles Workʼ, In Arai, Sachiyo, Kojima, Kzuhiro, Mineshima, Koji, Bekki, Daisuke, Satoh, Ken and Ohta, Yuiko, eds., New Frontiers in Artificial Intelligence: JSAI-isAI Workshops, JURISIN, SKL, AI-Biz, LENLS, AAA, SCIDOCA, kNeXI, Tsukuba, Tokyo, November 13–15, 2017, Revised Selected Papers. Lecture Notes in Artificial Intelligence, no. 10838 (Cham: Springer), 231–240. Henderson, Robert, and McCready, Elin (2019), ʻDogwhistles and the At-Issue/Non-at-Issue Distinctionʼ, in Daniel Gutzmann and Katharina Turgay, eds., Secondary Content. Current Research in the Semantics/Pragmatics Interface 37 (Leiden: Brill), 244–245. Google Scholar Google Preview WorldCat COPAC Henderson, Robert, and McCready, Elin (2020), “Towards Functional, Agent-Based Models of Dogwhistle Communication,” in Proceedings of the Probability and Meaning Conference (PaM 2020), 73–77, Gothenburg: Association for Computational Linguistics. Google Scholar Google Preview WorldCat COPAC Huber, Gregory A., and Lapinski, John S. (2006), ʻThe ʻRace Cardʼ Revisited: Assessing Racial Priming in Policy Contestsʼ, American Journal of Political Science, 50/2, 421–440. http://www.jstor.org/stable/3694281. Google Scholar WorldCat Hurwitz, Jon, and Pe ley, Mark (2005), ʻPlaying the Race Card in the Post–Willie Horton Era: The Impact of Racialized Code Words on Support for Punitive Crime Policyʼ, Public Opinion Quarterly, 69/1, 99–112. Google Scholar WorldCat Jamieson, Kathleen Hall (1992), Dirty Politics: Deception, Distraction, and Democracy (New York: Oxford University Press). Google Scholar Google Preview WorldCat COPAC Kantrowitz, Alex (2016, Oct. 1), ʻRacist Social Media Users Have a New Code to Avoid Censorshipʼ, Buzzfeed News, accessed 3 December 2023, https://tinyurl.com/2wfsbvn8.

WorldCat Kaufmann, Angelica (1794a) Phryne Seduces the Philosopher Xenocrates (1794), Private collection. Accessed 3 December 2023. https://tinyurl.com/yp2p6sy6. WorldCat Kaufmann, Angelica (1794b). Praxiteles Giving Phryne His Statue of Cupid. Rhode Island School of Design Museum, Providence. Accessed 3 December 2023. https://tinyurl.com/yeys6hj8. WorldCat Kern, Stephen (1996), Eyes of Love: The Gaze in English and French Culture, 1840–1900 (London: Reaktion). Google Scholar Google Preview WorldCat COPAC Khoo, Justin (2017), ʻCode Words in Political Discourseʼ, Philosophical Topics, 45/2, 33–64. Google Scholar WorldCat Kinder, Donald R., and Sears, David O. (1981), ʻPrejudice and Politics: Symbolic Racism versus Racial Threats to the Good Lifeʼ, Journal of Personality and Social Psychology, 40/3, 414–431. Google Scholar WorldCat Kolawole, Emi (2008, March 6), ʻDid Clinton Darken Obamaʼs Skin?ʼ FactCheck.org., accessed 3 December 2023, https://tinyurl.com/3huc7fuy. WorldCat LaFrance, Adrienne (2020, June), ʻThe Prophecies of Qʼ, ʻShadowlandʼ, The Atlantic, accessed 3 December 2023, https://tinyurl.com/2j787jry. WorldCat p. 569

Lamster, Mark. (2010), Master of Shadows: The Secret Diplomatic Career of the Painter Peter Paul Rubens (New York: Anchor). Google Scholar Google Preview WorldCat COPAC Manovich, Lev (2017, October 4), ʻInstagram and Contemporary Imageʼ, manovich.net. Medina, José (2018), ʻResisting Racist Propaganda: Distorted Visual Communication and Epistemic Activismʼ, Southern Journal of Philosophy, 56/51, 50–75. Google Scholar WorldCat Mendelberg, Tali (2001), The Race Card: Campaign Strategy, Implicit Messages, and the Norm of Equality (Princeton, NJ: Princeton University Press). Google Scholar Google Preview WorldCat COPAC Mendelberg, Tali (2008), ʻRacial Priming Revivedʼ, Perspectives on Politics, 6/1, 109–123. Google Scholar WorldCat Messing, Solomon, Jabon, Maria and Plaut, Ethan (2016), ʻBias in the Flesh: Skin Complexion and Stereotype Consistency in Political Campaignsʼ, Public Opinion Quarterly, 80/1, 44–65. Google Scholar WorldCat Mitchell, William John Thomas (2005), ʻThere Are No Visual Mediaʼ, Journal of Visual Culture, 4/2, 257–266. Google Scholar WorldCat Neiwert, David (2018, Sept. 19), ʻIs That an OK Sign? A White Power Symbol? Or Just a Right-Wing Troll?ʼ, ʻHatewatchʼ, Southern Poverty Law Center website, accessed 3 December 2023, https://tinyurl.com/5hbj3hjx. WorldCat

Nelson, Michelle R. (2008), ʻThe Hidden Persuaders: Then and Nowʼ, Journal of Advertising, 37/1, 113–126. Google Scholar WorldCat Pächt, Otto (1956), ʻPanofskyʼs Early Netherlandish Paintingʼ, Burlington Magazine, 98/637, 110–116, 266–277.

Pächt, Otto (1977/1999), The Practice of Art History: Reflections on Method, trans. David Brit (London: Harvey Miller). Original German pub: Methodisches zur kunsthistorischen Praxis: Ausgewählte Schri en, 1977. Google Scholar Google Preview WorldCat COPAC Packard, Vance (1957/2007), The Hidden Persuaders (Repr. New York: IG Publishers. Google Scholar Google Preview WorldCat COPAC Panofsky, Erwin (1934), ʻJan van Eyckʼs Arnolfini Portraitʼ, Burlington Magazine, 64/372, 117–119, 122–127. Google Scholar WorldCat Panofsky, Erwin (1953/1971), Early Netherlandish Painting: Its Origins and Character (Repr. New York: Icon. Google Scholar Google Preview WorldCat COPAC Panofsky, Erwin (1955/1982), ʻIconography and Iconology: An Introduction to the Study of Renaissance Artʼ, in Meaning in the Visual Arts (Chicago: University of Chicago Press), 26–54. Google Scholar Google Preview WorldCat COPAC Pelletier-Gagnon, Jérémie and Trujillo Diniz, Axel Pérez (2018), ʻColonizing Pepe: Internet Memes as Cyberplacesʼ, Space and Place, 1–15. Google Scholar WorldCat Pratkanis, Anthony R., and Greenwald, Anthony G. (1988), ʻRecent Perspectives on Unconscious Processing: Still No Marketing Applicationsʼ, Psychology and Marketing, 5/4, 339–355. Google Scholar WorldCat Prettejohn, Elizabeth, and Trippi, Peter, eds. (2016), Lawrence Alma-Tadema: At Home in Antiquity (Munich: Prestel). Google Scholar Google Preview WorldCat COPAC Principe, Lawrence (2013), The Secrets of Alchemy (Chicago: University of Chicago Press). Google Scholar Google Preview WorldCat COPAC Pseudo-Lucian. (c. 300 CE/1967), Amores. A airs of the Heart. 1967 ed., Vol. 8 (London and Cambridge: Loeb). Google Scholar Google Preview WorldCat COPAC Putscher, Marielene (1986), ʻDas “Buch der Heiligen Dreifaltigkeit” und seine Bilder in Handschri en des 15. Jahrhundertsʼ, in Christoph Meinel, ed., Die Alchemie in der europäischen Kultur- und Wissenscha sgeschichte (Wiesbaden: Harrassowit), 151–178. Google Scholar Google Preview WorldCat COPAC Randolph, Adrian W. B. (2002), Engaging Symbols: Gender, Politics, and Public Art in the Fi eenth-Century Florence (New Haven, CT: Yale University Press). Google Scholar Google Preview WorldCat COPAC Roob, Alexander (2014), The Hermetic Museum: Alchemy and Mysticism (Cologne: Taschen). Google Scholar Google Preview WorldCat COPAC Rubens, Peter-Paul (1625). Arrival of Marie deʼ Medici at Marseilles. Palais du Luxembourg, Paris. Accessed 3 December 2023. https://tinyurl.com/y6ncyhg5. Google Scholar Google Preview WorldCat COPAC

Saint-Saëns, Camille (1893). Phryné. Google Scholar Google Preview p. 570

WorldCat

COPAC

Santana, Carlos (2021), ʻWhatʼs Wrong with Dogwhistlesʼ, Journal of Social Philosophy 53/3, 287–403. https://doi.org/10.1111/josp.12409. Google Scholar WorldCat Saul, Jennifer M. (2017), ʻRacial Figleaves, the Shi ing Boundaries of the Permissible, and the Rise of Donald Trumpʼ, Philosophical Topics, 2, 45/2, 97–116. https://doi.org/10.5840/philtopics201745215. Google Scholar WorldCat Saul, Jennifer M. (2018), ʻDogwhistles, Political Manipulation, and Philosophy of Languageʼ, in Daniel Fogal, Daniel W. Harris, and Matt Moss, eds., New Work on Speech Acts. Oxford: Oxford University Press, 360–383. Google Scholar Google Preview WorldCat COPAC Saul, Jennifer M. (2019a), ʻImmigration in the Brexit Campaign: Protean Dogwhistles and Political Manipulationʼ, in Carl Fox and Joe Saunders, eds., Media Ethics, Free Speech, and the Requirements of Democracy (Abingdon, UK: Routledge), 21–37. Google Scholar Google Preview WorldCat COPAC Saul, Jennifer M. (2019b), ʻWhat Is Happening to Our Norms against Racist Speech?ʼ Aristotelian Society Supplementary Volume, 93/1, 1–23, https://doi.org/10.1093/arisup/akz001. Google Scholar WorldCat Seidel, Linda (1989), ʻ “Jan van Eyckʼs Arnolfini Portrait”: Business as Usual?ʼ Critical Inquiry, 16/1, 54–86. Google Scholar WorldCat Siemiradski, Henryk (1889). Phyrne at the Festival of Poseidon in Eleusis. St. Petersburg: Russian Museum. Accessed 3 December 2023. https://tinyurl.com/43e49jt5. Google Scholar Google Preview WorldCat COPAC Slides, John (2016, 6 Jan.), ʻItʼs Time to Stop the Endless Hype of the “Willie Horton” Adʼ, ʻMonkey Cageʼ, Washington Post, accessed 3 December 2023, https://tinyurl.com/2b5tm6t8. Stanley, Jason (2015), How Propaganda Works (Princeton, NJ: Princeton University Press). Google Scholar Google Preview WorldCat COPAC Tesler, Michael, and Sears, David O. (2010), Obamaʼs Race: The 2008 Election and the Dream of a Post-Racial America (Chicago: University of Chicago Press). Google Scholar Google Preview WorldCat COPAC Thompson, Maddie (2020). Instagram post. Accessed 3 December 2023. https://tinyurl.com/y45vu5ar. Google Scholar Google Preview WorldCat COPAC Ti any, Kaitlyn (2020, 18 Aug.), ʻThe Women Making Conspiracy Theories Beautiful: How the Domestic Aesthetics of Instagram Repackage QAnon for the Massesʼ, ʻShadowlandsʼ, The Atlantic, accessed 3 December 2023, https://tinyurl.com/hjmkp3m. Torices, José Ramón (2021), ʻUnderstanding Dogwhistles Politicsʼ, Theoria, 36/3, 331–339. Google Scholar WorldCat Turner, Joseph Mallord William (1838). Phryne Going to the Public Baths as Venus; Demosthenes Taunted by Aeschines. Accessed 3 December 2023. https://tinyurl.com/2tc8265h. WorldCat Valentino, N., Hutchings, Vincent L., and White, Ismail K. (2002), ʻCues That Matter: How Political Ads Prime Racial Attitudes during Campaignsʼ, American Political Science Review, 96/1, 75–90.

Google Scholar

WorldCat

Valentino, N., Neuner, Fabian G., and Vandenbroek, L. Matthew (2018), ʻThe Changing Norms of Political Rhetoric and the End of Racial Primingʼ, Journal of Politics, 80/3, 757–771. Google Scholar WorldCat van Eyck, Jan (1434). Arnolfini Wedding Portrait. National Gallery, London. Accessed 3 December 2023. https://tinyurl.com/qf9gpjn. Google Scholar Google Preview WorldCat COPAC van Straten, Roelof (1994), An Introduction to Iconography (London: Taylor & Francis). Original German Pub: Inleiding in de iconografie, 1985. Google Scholar Google Preview WorldCat COPAC Wiggins, Bradley E. (2020), ʻBoogaloo and Civil War 2: Memetic Antagonism in Expressions of Covert Activismʼ, New Media & Society, https://doi.org/10.1177/1461444820945317. Google Scholar WorldCat Zannettou, Savvas, Caulfield, Tristan, Blackburn, Jeremy, de Cristofaro, Emiliano, Sirivianos, Michael, Stringhini, Gianluca, et al. (2018), ʻOn the Origins of Memes by Means of Fringe Web Communitiesʼ, arXiv, accessed 3 December 2023, https://arxiv.org/abs/1805.12512v3. WorldCat

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published online: 22 May 2024 Published in print: 02 May 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

10 Speech-Act Theory in Feminist Thought  Louise Antony https://doi.org/10.1093/oxfordhb/9780192844118.013.52 Published: 22 May 2024

Pages 211–244

Abstract In this chapter, I criticize attempts by Rae Langton, Jennifer Hornsby, and others to apply J. L. Austin’s theory of speech acts to accounts of the wrongs of pornography and rape. There are two main ideas: (a) that pornography is a verdictive or exercitive illocutionary act, and (b) that an assailant can “silence” a victim in their attempt to refuse sex by not providing “uptake.” The chapter argues that the rst idea is inconsistent with Austin’s theory of illocutionary acts. The second idea requires a reading of Austin that commits him to a host of counterintuitive consequences. The chapter concludes by pointing out that the normative claims these authors want to make about the wrongs of pornography and rape can be developed and defended without any appeal to Austin.

Keywords: speech act, pornography, rape, J. L. Austin, silencing, social construction, feminism, illocution, perlocution, convention Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction Austinian speech-act theory has come to play an important role in feminist theorizing about the subordination of women. The philosophers most responsible for the emergence of this trend are Rae 1

Langton and Jennifer Hornsby. Langton, in her 1993 paper “Speech Acts and Unspeakable Acts,”

introduced the idea that pornography should be regarded as an exercitive or verdictive speech act and, as such, as regulable action rather than protected speech. In the same paper, she appealed to the Austinian notion of “uptake” to explain a particular way in which women, or members of any disempowered group, can be subordinated, an idea taken up and elaborated by Hornsby in her 1995 paper “Disempowered 2

Speech.” The thought was that when disempowered individuals attempt to say things but are ignored or culpably misunderstood, they are being silenced or “illocutionarily disabled.” Hornsby and Langton, in their 3

“Free Speech and Illocution,” applied this analysis speci cally to cases in which an assailant forces sex on another person despite that person’s explicitly saying or otherwise indicating that they mean to refuse sex. These two ideas have spurred enormous philosophical discussion. Both the claims that pornography is an illocutionary act of subordination, and that rape involves silencing have received criticism. But both have also been defended and—especially in the case of the illocutionary account of silencing—extended to cover a number of subordinating acts and practices. The overarching suggestion that Austinian speech-act theory is pertinent to feminist theorizing has become entrenched. In this essay, I will try to swim against the tide. I will argue that Langton’s and Hornsby’s appeals to Austin involve serious misunderstandings of Austin’s work and serve to obscure his central contribution to the philosophy of language. Furthermore, I’ll argue, the appeal to Austin is, in any case, gratuitous. Although p. 212

the Austin-inspired

work has yielded some very careful, thoughtful, and original insights into forms of

subordination that involve speech, none of the phenomena described require mention of or are enriched by appeal to Austin’s work. Had Austin never written How to Do Things with Words, had he never identi ed the “performative” use of language, we would still be able to express everything that Langton and Hornsby and 4

their followers want to say about pornography, rape, and the subordination of women.

In section 2, I’ll focus on the ways that Langton and others have attempted to apply speech-act theory to issues about pornography. In section 3 2, I’ll look closely at the ways in which Hornsby, Langton, and others have deployed the notion of “uptake” to illuminate the thwarting of victims’ refusals of sex. In section 4 3, I’ll consider the general question whether Austin’s work o ers any real advance in our understanding of 5

either pornography or rape.

2. Is Pornography an Illocutionary Act? Issues around the production, sale, and consumption of pornography have long divided feminists, as Amia 6

Srinivasan documents in The Right to Sex. Since the 1970s, the main division has been—to put it crudely— between “anti-porn” feminists, like Andrea Dworkin and Robin Morgan (both founding members of the New York section of the organization Women Against Pornography) and “pro-sex” feminists like Ellen Willis. Feminists in the anti-porn camp argued that the production of pornography involved violence and coercion—which all feminists condemned—but also that the content of pornography promulgated a view of women and women’s sexuality that legitimated and encouraged the sexual subordination and mistreatment of women and that, at the extreme, promoted rape. As Morgan put it, “Pornography is the theory, and rape 7

is the practice.” Feminists in the pro-sex camp decried what they saw as regressive moralism in the antiporn movement. They argued that pornography could be a vital tool in women’s e orts to discover and reclaim their own sexuality, and that the anti-porn movement’s narrow focus on violent heterosexual pornography neglected new and liberatory forms of erotica being produced by lesbian and heterosexual women, gay men, and gender rebels. Crosscutting the feminist camps was the debate about the censoring of pornography: religious and social conservatives made common cause with anti-porn feminists in calling for the banning of the production and sale of pornography, while right-wing libertarians joined civil libertarians and pro-sex feminists in opposing any form of censorship—making for strange bedfellows on both sides. From this maelstrom, Catharine MacKinnon emerged with a unique analysis of pornography and a unique proposal for dealing with it legally. Her view of pornography developed against her background theory of gender, according to which the categories “man” and “woman” were, essentially, slots in a social hierarchy p. 213

organized by

sexual relations: to be a man was to be sexually subordinating and to be a woman was to be 8,9

sexually subordinated.

Pornography, she then argued, was the social textbook that trained both men and

women to take up their respective gender roles: men were taught to eroticize dominance, and women to eroticize submission. The pornographic narrative, MacKinnon argued, “constructed” women as beings who exist primarily for men, training men (by associating this narrative with strong sexual pleasure) to view 10

women as desirous of sexual subordination, thus naturalizing and legitimating sexual subordination.

In the legal domain, most discussion of pornography had, before MacKinnon, centered on con icts between the alleged harms of pornography (which included, for social conservatives, the matters of o ense to community standards and the promotion of “immoral” sexual activity) and the free speech rights of the producers of pornography. MacKinnon, however, sought to transform the legal debate. Building on her own landmark legal treatment of sexual harassment as a form of sex discrimination, she argued that the con ict between pornography and the well-being of women should not be conceptualized as a con ict between a right and a harm, but as a con ict between rights: the pornographer’s right to free speech, and a woman’s right against sex discrimination. Pornography, on this conception, was not only speech; it was an act that involved speech. Acts of discrimination—like posting a job advertisement that says “no Blacks need 11

apply”—are not protected by the First Amendment;

similarly, MacKinnon argued, the law should focus on

the pornographic practice rather than its product, and the incompatibility of that practice with women’s rights as equal citizens. In 1983, MacKinnon collaborated with Andrea Dworkin to craft a model law that would enable women who had been harmed by pornography to bring civil suits against pornographers, just as women harmed by discriminatory practices could sue for damages in civil courts. The language of the model law was incorporated into an amendment to the Code of Indianapolis and Marion County, Indiana (Indianapolis Code § 16-3(q)). But the MacKinnon-Dworkin strategy did not work. The Indianapolis ordinance was declared unconstitutional in district court, and the ruling was upheld on appeal at the Seventh Circuit Court. In his now famous (or infamous) opinion, the presiding appeals judge, Frank Easterbrook, cited several

reasons for his decision. Some had to do with the “vagueness” and “subjectivity” of the de nition of “pornography” in the ordinance, but the most important reason, Easterbrook said, was that the ordinance violated the First Amendment. Easterbrook was completely unmoved by the argument that the issue was a matter of what pornography did, as opposed to what it said. He did not deny that pornography caused harm—serious harm—to women, quoting, with approval, the following language from the ordinance: [P]ornography is central in creating and maintaining sex as a basis of discrimination. Pornography is a systematic practice of exploitation and subordination based on sex which di erentially harms women. The bigotry and contempt it produces, with the acts of aggression it fosters, harm 12

women’s opportunities for equality and rights [of all kinds]. p. 214

Perhaps, as some feminist writers allege, Easterbrook was striking a cynical pose in these passages. But whether or not these words were sincere, his choice of rhetoric at least makes clear that his ruling was based on the manner in which pornography produced its e ects, namely by promulgating a distinctive point of view. Yet this simply demonstrates the power of pornography as speech. All of these unhappy e ects depend on mental intermediation [my emphasis]. Pornography a ects how people see the world, 13

their fellows, and social relations. If pornography is what pornography does, so is other speech. In light of the feminist arguments we are about to consider, I want to emphasize that nothing in

Easterbrook’s ruling denies that pornography can “do things”—the judge’s point is that the mechanism by which pornography does things—the way in which it produces or maintains the subordination of women is by saying things. It may not be only speech, but Easterbrook’s point is that it is at least speech. And as long as pornography works through the mechanisms by which assertoric speech generally produces action—that is, by changing the mental states of the persons who hear it—it is functioning in a way that the U. S. Constitution protects. Although Easterbrook did not challenge, but rather accepted, the claim that pornography subordinated women, that claim had been challenged by Judge Sarah Barker in an earlier district court decision. She derided MacKinnon and Dworkin’s characterization of pornography as regulable, discriminatory behavior 14

as “sleight of hand.”

Philosopher William Parent seemingly agreed, calling the MacKinnon-Dworkin 15

de nition “philosophically indefensible.”

It was at this point in the dialectic that Langton entered the

scene. She saw in J. L. Austin’s theory of speech acts a way to interpret and vindicate MacKinnon’s de nition of pornography as an act of subordination rather than merely a cause of subordination, and thus to defend her against charges of confusion or incoherence. As Langton read him, Austin’s main goal in his de nitive work How to Do Things with Words was to point out the performative function of language; to show that one 16

could use language to perform actions, to “undermine the dichotomy between word and action.”

This,

according to Langton, is also what MacKinnon is aiming to do with her characterization of pornography: to 17

provide a “description of the actions constituted by pornographic utterances.”

In Langton’s view, then,

the main mistake made by MacKinnon’s critics was their failure to realize that language had this 18

performative potential, that in certain cases, speech can be both word and act.

But Langton’s interpretation of Austin is, importantly, incorrect. At the general level, she is wrong about Austin’s theoretical aims. He was not primarily concerned with making the general point that words could be used to do things; that much he took to be obvious. Rather, he wanted to characterize and discuss a particular way that language could be used to do things, a way that did not involve the making of a statement, and one that he felt had hitherto been neglected by philosophers. As he complains on early on: “[I]t was for too long the assumption of philosophers that the business of a ‘statement’ can only be to p. 215

‘describe’ some state of a airs, to ‘state some fact,’ which it

must do either truly or falsely” (HDTWW, 1).

This familiar function of language he will later call the “constative” function; it will be his main business to 19

characterize, in contrast, a “performative” function of language.

In Lecture 2, he registers his approval of

“the recent movement towards questioning an age-old assumption in philosophy—the assumption that to say something, at least in all cases worth considering … is always and simply to state something” (HDTWW, 12). Austin’s initial characterization of the class of utterances he is concerned with occurs in Lecture 1, in the section headed “Preliminary Isolation of the Performative.” A primary characteristic of the utterance types he is concerned with that is that “they do not ‘describe’ or ‘report’ or constate anything at all, are not ‘true’ or ‘false.’ ” (HDTWW, 5) A little later, he says that although these utterance types—now identi ed as “performatives”—“have on the face of them the look of ‘statements’,” they nevertheless are seen “when more closely inspected, to be quite plainly, not utterances which could be ‘true’ or ‘false.’ ” (HDTWW, 12–13) The example he then o ers is the utterance “I do,” made in the context of a marriage ceremony. “Here we should say that in saying these words we are doing something—namely, marrying, rather than reporting something, namely that we are marrying” (HDTWW, 13). I am belaboring this point for two reasons: rst, because Langton’s failure to recognize that Austin is making a distinction between the di erent ways speech can be action, rather than just a

rming that speech

can be action, leads her to misunderstand the nature of the illocutionary speech act and—quite importantly —the nature of the kinds of states of a airs an illocutionary act can bring about; second, because her speci c application of speech-act theory to the case of pornography requires her to treat pornography as simultaneously performative and constative, ignoring the distinction central to Austin’s project. I’ll make each point in turn. How is the notion of an illocutionary act bound up with the distinction between performative and constative uses of language? Let’s consider an example of a constative use of language. Suppose my friend and I are watching a show that he doesn’t particularly like, but that I do. In that context, I utter the sentence “Schitt’s Creek has won the 2020 Emmy for Best Comedy.” Now, everyone can agree that I have done various things: I’ve uttered a sentence; I’ve asserted a proposition; and I’ve informed my friend. I may also have implied that the show is quality television and insinuated that my friend’s judgment is awed. What I have not done is bring it about that Schitt’s Creek won an Emmy. That is, my saying that Schitt’s Creek won an Emmy did not make it so. The state of a airs that corresponds to what I have asserted is not brought into being by my saying what I say, nor can it be. In contrast, if the duly appointed representative of the Television Academy announces that, per the results of the members’ votes, “Schitt’s Creek has won the 2020 Emmy for Best 20

Comedy,”

 that individual has made it so. Austin is fundamentally concerned with how this second feat is

accomplished—how the representative could, simply “in saying” something, make that thing come about. That is why Austin, even before introducing the notion of an illocutionary act, takes up the question (and this is a section heading) “Can Saying Make It So?” It may, he says, sound “odd or even ippant at rst” to p. 216

say something like “To marry is to say a few “su

words” or “Betting is simply saying something,” but with

cient safeguards it may become not odd at all.” (HDTWW, 7) These “safeguards,” the characterization

of which take up most of the rest of the book, all have to do with facts about social conventions—that is, 21

intentional practices that may be more or less explicit, but that, once in place and accepted,

 create

categories of fact that would not otherwise exist. It is because the facts involved are—as we now say— socially constructed that it can be up to us whether or not they obtain. Austin lays out the general conditions that must hold for a saying to make something the case. He calls these conditions “felicity conditions”—conditions that must be satis ed for a performative use of language to be “happy.” There are two types of felicity conditions: the A/B type and the Γ-type. The A/B felicity conditions, but not the Γ-type conditions, are necessary conditions for the performance to be successful: if any one of these conditions fails to be satis ed, then no performative act has occurred, and—I would say—no new fact

has been created. Violation of either of the A/B conditions yields a “mis re,” whereas violation of a Γ-type condition yields only an “abuse” (HDTWW, 16). All of the A/B conditions—the necessary conditions—advert to the existence of conventions. Here are the A conditions: (A.1) There must exist an accepted conventional procedure having a certain conventional e ect [my emphasis], that procedure to include the uttering of certain words by certain persons in certain circumstances, and further (A.2) The particular persons and circumstances in a given case must be appropriate for the invocation of the particular procedure. (HDTWW, 14–15) The B conditions stipulate that (B.1) The procedure must be executed by all participants both correctly and (B.2) completely. (HDTWW, 15) In sum: Austin’s main objective in HDTWW is to distinguish the performative use of language from the constative use of language, and the distinguishing feature is that the performative use enables us to make things so “merely” in saying that they are so. The “things” over which we have such power are those things that depend on some “conventional procedure” that has “a certain conventional e ect.” Now let’s turn to Austin’s o

cial taxonomy of speech acts—things that we can do with words. Austin wants

to distinguish three kinds of actions that one can perform with words: the locutionary act, the illocutionary act, and the perlocutionary act. Although he introduces the pertinent distinctions in Lecture 8, he gives a concise characterization at the beginning of Lecture 9: We rst distinguished a group of things we do in saying something, which together we summed up by saying we perform a locutionary act, which is roughly equivalent to uttering a certain sentence p. 217

with a certain sense and reference. Second, we

said that we also perform illocutionary acts, such

as informing, ordering, warning, undertaking, &c., utterances which have a certain (conventional) force. Thirdly, we may also perform perlocutionary acts: what we bring about or achieve by saying something, such as convincing, persuading, deterring, and even, say, surprising or misleading. (HDTWW, 109) Although he doesn’t use this formula here, his shorthand for all this is that the locutionary act simply is the saying; the illocutionary act is what is done in saying something; and the perlocutionary act is done by saying something (Cf. HDTWW, 122–32). Austin has said earlier that his “interest in these lectures is essentially to fasten on the … illocutionary act and contrast it with the other two” and remarks that “there is a constant tendency in philosophy to elide this in favor of one or the other two” (HDTWW, 103). In Lecture 9, he says: “It is the distinction between illocutions and perlocutions which seems likeliest to give trouble,” and sets out to draw it. What’s the di

culty? Austin is aware that the “in doing/by doing” language cannot do the needed work, but he thinks

that what is signi ed by the prepositional di erence is often a distinction between—as I would put it—what is integral to an action and what is external to the action: “We have then to draw the line between an action we do (here an illocution) and its consequences” (HDTWW, 111). How is this to be done? In the case of physical (i.e., non-speech) acts, Austin points out that there is a cascade of equally apt ways of

characterizing an act that moves smoothly from factors completely internal to the actor (e.g., exing one’s nger) to factors involving the intentional aim of the act (shooting the gun) to consequences completely external to the actor (killing the rabbit). Austin needs there to be something that blocks a similar cascade from what is done or brought about simply in the performance of an illocutionary act, and what might result from the performance of such an act. What blocks the cascade, I contend, is the set of boundaries provided by the conventions that de ne the illocutionary act. These are also, of course, what specify the state of a airs that the performance of the illocutionary act can bring about. In other words, the answer to Austin’s question, “Can saying make it so?” is also the answer to his question about what demarcates the states of a airs brought about in saying something from the consequences (possibly) achieved by saying something. The only reason that the Television Academy representative can make it the case that Schitt’s Creek won the Emmy is that there is an 22

accepted conventional procedure—in this case, tallying the votes

and announcing the result—having a

certain conventional e ect, namely, Schitt’s Creek becoming the winner. It may be a further consequence of the representative’s announcement that the fans of the show are delighted, so that it will be true that to say, “The representative delighted the fans,” but this statement is not true in virtue of the representative’s illocutionary act—it is not part of what the convention speci es as the state of a airs that will be brought about by the performance of the prescribed speech in the prescribed circumstances. Nor could it be. Delighting fans is not the kind of state of a airs that can be brought about simply in saying p. 218

something, because it is not the kind of thing that convention can

govern. To articulate what it is in virtue

of which a fan is delighted, one must say something about the emotional condition of the fan, and that is constitutively independent of anything that anyone says, no matter what conventions are in force. In short, the speci cally illocutionary e ects of a speech act—the ones that are integral to, or inside, the illocutionary act—are those e ects that are conventional e ects of that act of saying; they are e ects that obtain only because of the background convention that de nes the act as one of some particular illocutionary type. I’ll go further and say that the kinds of changes in the world that an illocutionary act can bring about are changes that can only be e ected through conventional means. Austin writes: Certainly we can achieve the same perlocutionary sequels by non- conventional means, means that are not conventional at all or not for that purpose; thus I can persuade some one by gently swinging a big stick … Strictly speaking, there cannot be an illocutionary act unless the means 23

employed are conventional.

This, then, is my rst main objection to Langton’s claim that pornography is an illocutionary act of subordination. Subordination is a material condition that is constitutively (though obviously not causally) independent of anything that anyone says. It is not a state of a airs that can be brought about through the adoption of some convention, and it is a condition that can certainly be brought about by non-conventional means—say, by threatening a person with physical violence. Clearly, certain illocutionary acts—like disenfranchisement—can be instrumental to a program of subordination, but actual subordination involves 24

material relations of power.

Langton disagrees. She argues that, for example, a “legislator enacting apartheid law,” in saying (presumably in the legally correct circumstances), “Blacks are not permitted to vote,” not only performs the illocutionary act of disenfranchising Blacks, but also the illocutionary acts of ranking Blacks as inferior and 25

legitimating discrimination against them.

I quite agree that disenfranchisement is an illocutionary act. Voting

is a socially constructed practice, and the determination of who is eligible to vote is an essential part of whatever speci c conventions construct the practice. But “disenfranchisement”—setting it up so that, or changing things so that, someone cannot vote—does not entail, and therefore is not constitutively

connected to—ranking that person as inferior or legitimating discrimination against them. Doing these things is undoubtedly the aim of many acts of disenfranchisement, but we can pull things apart by considering the fact that there can be legitimate reasons for denying some people the right to vote, even if there are also reprehensible reasons. Suppose that we have just had a successful revolution against a terrible dictator who decreed that any of his relatives would be eligible to vote—even if they lived outside the country and had no interest in the country’s a airs. Our new government would be well within its rights to revoke the franchise of these relatives. In doing so, the government (because it is enlightened) passes no p. 219

judgment as to their “rank,” and does not in any way authorize or condone discriminatory treatment of those relatives. In cases where (despite the laudable aims of the

revolutionary council), people do come to

rank the relatives as inferior and feel justi ed in discriminating against them, these would be perlocutionary e ects of the illocutionary act of disenfranchisement. However, it could be argued that subordination can be the conventional e ect of a speech act, a condition brought about in the saying of something, and not just in consequence of the saying. Most occupations that have been at all institutionalized—like the occupation of being a college professor—have, as part of their conventional structure, a speci cation of chains of authority, with clearly speci ed responsibilities and powers. In academia, for example, deans often have the power to appoint department chairs, who in turn have a speci ed set of powers with respect to department members. But deans also have the power—not much exercised in a well-administered and collegial institution—to remove a person from the position of department chair, stripping them of the powers they possessed in that role. Removing a person from a position of relative power constitutes a form of subordination; thus deans have the authority to subordinate individuals in certain ways and under certain circumstances. Now suppose the rules of some particular institution specify that, if a dean wants to demote a department chair, the dean must call the o ending person into their o

ce and say, “I hereby subordinate you.” Wouldn’t this speech act, given the governing

conventions of the institution in which it is performed, and the authority of the person performing it, mean that the subordination of the errant department chair is a condition e ected in the dean’s performing their illocutionary act? This example brings us to issues about what type of speech act pornography might be, if it is a speech act at all. In my example, it appears that the dean’s illocutionary act is of the type Austin calls an “exercitive.” Here is his characterization of exercitives: [E]xercitives are the exercising of powers, rights, or in uence. Examples are appointing, voting, ordering, urging, advising, warning, &c. (HDTWW, 151) An exercitive is the giving of a decision in favour of or against a certain course of action, or advocacy of it. (HDTWW, 155) Since the power the dean is exercising is one that is clearly tied to the conventions—the rules—of the institution in which the dean serves, it would appear that the condition into which the errant chair has been placed is also tied to that convention; it is a socially constructed condition, which is why the dean can put the chair into it simply by performing the illocutionary act of declaring the chair to be subordinated. I agree. But I contend that this socially constructed condition—call it “institutional subordination”—is a di erent condition from the kind of subordination that women su er under patriarchy—what we can call “material subordination.” Institutional subordination involves the removal of only those powers and privileges that were originally speci ed within the conventional, institutional structure that de nes and creates the pertinent social roles. (The convention giveth, and the convention taketh away.) It may be part of the academic institutional framework that a demotion carries with it a reduction in salary, but even so, such p. 220

a reduction does not in itself constitute

material subordination. The chair might be impoverished by such

a reduction, but they need not be—they might be independently wealthy and not in need of their academic salary. That possibility demonstrates that the state of a airs that the dean brings about by saying “I hereby subordinate you” does not include the impoverishment of the chair. Another aspect of material subordination is, typically, a loss or absence of power. Again, while the chair does lose power when the dean demotes them, the loss is limited to only that power originally granted to the chair by the institutional structure. Most of the kinds of powerlessness experienced by persons in materially subordinate positions are di erent—they can include physical vulnerability, social invisibility, psychological stress, and impaired agency. Such conditions may result from the demotion, but they need not. For all that the chair has lost, they may retain physical power, the friendship and support of a network of friends, good mental health, and an ability to e ect their purposes. To motivate the distinction between institutional and material conditions, let me turn to another category of speech act: verdictives. Verdictives, in Langton’s view, are also candidates for the kind of speech act pornography performs. Here’s how Austin characterizes them: [V]erdictives are typi ed by the giving of a verdict, as the name implies, by a jury, arbitrator or umpire. (HDTWW, 151) Verdictives consist in the delivering of a nding, o

cial or uno

cial, upon evidence or reasons as

to value or fact, so far as these are distinguishable. (HDTWW, 153) 26

Verdictives characteristically—though not necessarily —are meant to track certain extralinguistic facts. Thus, judicial systems that provide for trials by jury are set up with the expectation that the jurors’ verdicts will tend to re ect the actual guilt or innocence of the defendant. No one in the United States in the 21st century should need to be told how easily and often this expectation is thwarted due to the ignorance, misinformedness, malignity, or corruption of the jurors. But even when jurors are given correct evidence, and are doing their very best to render an accurate assessment of the case, their verdicts can be mistaken—a defendant found guilty may not have committed the crime. Similarly, even a well-trained and scrupulous umpire can mistake a ball for a strike, and a diligent teacher can give a poor grade to a student essay they simply didn’t understand. What are we to say about the successfulness of the verdictive speech acts in cases 27

like this? In such cases, has the attempted verdictive act mis red, or only involved abuse?

I can nd no explicit answer to this question in Austin, but Austin gives us some clues in his discussion of 28

promising.

In Lecture 1, considering whether the issuance of a promise should be regarded as a statement

describing an intention, he notes that “it is appropriate that the person uttering the promise should have a certain intention, viz., here to keep his word” and also that “when such an intention is absent, speak of a ‘false’ promise” (HDTWW, 11). However, he continues: p. 221

Yet so to speak is not to say that the utterance ‘I promise that … ’ is false, in the sense that though he states that he does, he doesn’t…. For he does promise: the promise here is not even void, though it is given in bad faith. (HDTWW, 11) It’s clear, then, insincerity in promising should be counted as a Γ-type infelicity; one who makes an insincere promise has performed the speech act of promising, but they have done so unhappily. One is supposed to be sincere in making a promise, but insincerity does not prevent the promiser from incurring an obligation. Their promise does put them into a state of obligation, as the convention of promising provides, whether or not they intend to carry through (HDTWW, 11).

I suggest that we treat the problematic verdictives in a similar way. The jury’s verdict creates a certain condition—call it “judicial guilt”—which is supposed to track (what I’ll call) natural guilt. Judicial guilt is a socially constructed condition, the exact nature of which is speci ed by the conventions codi ed in the law. Jurors, by law, have the power to determine whether a duly charged person is judicially guilty or not. The jurors, however, have no power to determine whether a person is naturally guilty or not. Natural guilt is a matter of what a person has done or not done, independently of what any juror may believe or say. There is good support for this interpretation in Lecture 4, when Austin is again speaking of the consequences of someone’s making an insincere promise. An insincere promise, recall, is an unhappy promise, but a promise, nonetheless. He compares such a case with a case of a verdictive (which he is here introducing) like a jury’s arriving at a verdict: When we say ‘guilty’, this is happy in a way if we sincerely think on the evidence that he did it. But, of course, the whole point of the procedure in a way is to be correct; it may even be scarcely a matter of opinion …. Thus when the umpire says ‘over’, this terminates the over. But again we may have a ‘bad’ verdict: it may either be unjusti ed (jury) or even incorrect (umpire). So here we have a very unhappy situation. But still it is not infelicitous in any of our [Γ-A/B type] senses: it is not void (if the umpire says ‘out’, the batsman is out; the umpire’s decision is nal). (HDTWW, 43) In sum, I still want to maintain that the only conditions that a speech act can bring into being—the only kinds of things such that saying makes them so—are conditions de ned by conventions, so that, however much and in whatever way pornography may contribute causally to the subordination of women, it does not constitute their subordination. I turn now to the second of my main objections to Langton’s position, which has to do with the speci c claims Langton wants to make about the kind of illocutionary act she takes pornography to be. As I have mentioned, Langton says that pornography is either an exercitive or a verdictive speech act. (I don’t take issue with this disjunction, since Austin himself admits that the line between exercitives and verdictives can p. 222

be di

cult to draw.) I’ve argued that the material subordination of women is not the kind of state of

a airs—because it is not constitutively a social construction—that any such speech act brings about. What I want to point out now, however, is that the claim that pornography is either a verdictive or exercitive speech act is in con ict with other things that Langton wants to say about pornography. Langton, as I said, accepts, and wants to explicate MacKinnon’s overall view that pornography subordinates women. Moreover, she accepts what MacKinnon has to say about the mechanism by which this 29

subordination occurs.

Here’s MacKinnon’s story: pornography tells lies about women, such as the lie that

they desire to be sexually dominated. The ubiquity of pornography, along with its tremendous power to arouse men (construed in MacKinnon’s sense of sexual dominators) means that these lies become generally accepted, perhaps even by women. As men’s expectations about women’s sexual behavior and women’s own (perhaps uncomfortable) identi cation with their pornographic image, women actually become what pornography says they are—creatures who exist for, and desire domination by, men. The shorthand way to characterize this process is to say that pornography constructs women to be the way it says they are. Langton takes it to be a selling point of her illocutionary analysis that it explains how this sort of construction takes place. According to Langton, pornography carries an invidious message about women that somehow makes itself true. Pornographers characterize women in a certain way and thereby create a social reality for women in which they conform to this characterization—women come to be what the pornographers say they are. To understand this process, she argues, we can see pornography as either a verdictive or an exercitive speech act (HDTWW, 151–7). In promulgating its message, pornography e ectively 30

makes it the case that women are the way it says there are.

Now, there are a number of issues that arise in connection with this claim (and with MacKinnon’s claims) that have nothing to do with speech-act theory. The big issue is, What exactly is it that pornography makes true? Does Langton think that pornography makes it the case that women do desire violent and degraded treatment? Or that they actually are creatures who exist for the pleasure of men? Let me set these questions 31

aside, however.

I think it’s not implausible that one of the—I’d call them—causal consequences of the

widespread availability of certain sorts of heterosexual-male-oriented porn, and the subsequent e ect of that availability on the shaping of men’s desires and expectations, is that women acquire views about what is normal or usual sexual practice, and attempt to adhere to, or at least acquiesce in, the norms of sexual 32

behavior that they experience as applying to them. There is some empirical research to back this up.

But let

me ignore the subtleties of the details of the social construction of men’s and women’s sexuality, and return 33

to the (boring, I’m sorry!) subject of speech acts.

We’ve already seen Austin’s characterizations of verdictive and exercitive speech acts, and I’ve given one important reason why pornography cannot be a speech act of either kind. What I want to point out now, however, is that there is a contradiction between Langton’s contention that p. 223

pornography is a speech act of either of these types, and

the view that pornography works by saying false

things about women. If pornography asserts something about women, then it falls into the category of constative speech, whereas verdictive and exercitive speech acts are paradigmatically performative uses of speech. I’ve been arguing that Austin’s main purpose in HTDTWW was to challenge a view that he found to be prevalent in much of the philosophy of his time—namely, that “the business of a ‘statement’ can only be to ‘describe’ some state of a airs, or to ‘state some fact,’ which it must do either truly or falsely” (HDTWW, 1). This view, he argues (as we know) is false. “Utterances can be found” he says, that, despite employing “humdrum verbs in the rst person singular present indicative active,” “do not ‘describe’ or ‘report’ or constate anything at all, are not ‘true or false’ ” (HDTWW, 5). The examples he adduces are all, as it happens, of the type he will later identify as exercitives: saying “I do” in the context of a marriage ceremony, saying “I name this ship the Queen Elizabeth” under the conventional accepted protocol, bequeathing a watch, and making a bet. He comments: In these examples it seems clear that to utter the sentence … is not to describe my doing of what I should be said in so uttering to be doing or to state that I am doing it. None of the utterances is either true or false: I assert this as obvious and do not argue it. (my emphasis; HDTWW, 6) And again: When I say, before the registrar or altar … “I do,” I am not reporting on a marriage: I am indulging in it. (HDTWW, 6) Now, one might argue, against my view of Austin’s central aim, that while the distinction we see him drawing in these passages seems pretty crisp, it gradually erodes over the course of the lectures. He opens Lecture 11 by reminding the reader of two of the features that he had earlier said distinguished performative utterances from constative utterances: rst, that the former were instances of “doing something,” while the latter were “just” instances of “saying something,” and second, that while constatives were either true or false, performatives were “happy or unhappy.” But then he asks, “Were these distinctions really sound?”

Our subsequent discussion of doing and saying certainly seems to point to the conclusion that whenever I ‘say’ anything … I shall be performing both locutionary and illocutionary acts, and these two kinds of acts seem to be the very things which we tried to use, under the names of ‘doing’ and ‘saying’, as a means of distinguishing performatives from constatives. (HDTWW, 133) Austin goes on to argue that neither of these distinctions holds up. Stating, he concedes, is as much a “doing” as commanding or promising, and truth and falsity can be regarded as simply particular types of p. 224

34

felicity conditions.

The mistake of treating constatives

as non-performative, stems, Austin says, from a

tendency to focus, in the case of constatives, on the locutionary act and its central component, the declarative sentence. However, he writes, [o]nce we realize that what we have to study is not the sentence but the issuing of a sentence in a speech situation, there can hardly be any longer a possibility of not seeing that stating is performing an act. (HDTWW, 139) The upshot of all this reconsideration, however, is not that Austin repudiates the distinction he spent so long articulating in the rst few lectures, but that he reconceives it. Instead of developing a list of “explicit performative verbs” that could be used to distinguish performative utterances from constative utterances, he says now that what we need is “a list of the illocutionary forces of an utterance.” We must abandon, he says, “the dichotomy of performatives and constatives … in favor of more general families of related and overlapping speech acts” (HDTWW, 150). Well, things are looking bad for my interpretation. And there’s more: in the course of this reconsideration of the constative/performative distinction, Austin explicitly allows that there is no necessary con ict between (a) our issuing the utterance as being the doing of something and (b) our utterance being true or false (HDTWW, 135) For good measure, he provides an example to show that a single illocutionary act can possess two di erent forces at the same time: [C]ompare, for example, ‘I warn you that [the bull] is going to charge,’ where likewise it is both a warning and true or false that it is going to charge; and that comes in in [sic] appraising the warning just as much as, though not in the same way as, in appraising the statement. (HDTWW, 135–136) Despite all this, however, I still maintain that the main point of Austin’s book is to call attention to the performative as opposed to the descriptive use of language. Moreover, I maintain that despite his allowance that some utterances can have two di erent illocutionary uses at the same time, Austin maintains his original view that verdictives and exercitives cannot describe the state of a airs they bring about. First of all, note that Austin’s allowance that constatives are performatives does not entail that any performative is or can be a constative. He never takes back his claim that some performatives—or what he now wants to refer to as “performative uses of language”—bring about, but do not describe a particular state of a airs. That is, nothing in the later lectures requires a revision of what he says in Lecture 1: “When I p. 225

say, before

the registrar or altar … ‘I do,’ I am not reporting on a marriage: I am indulging in it” (HDTWW,

6). Austin really only mentions two kinds of cases in which he is willing to say that an utterance involves, simultaneously, a constative and a performative use of language. One kind is the sort presented at HDTWW 135-6 and cited 3 paragraphs above where a warning (a performative use of language) can at the same time be a description of a state of a airs; the other kind is when, in the making of an assertion (a constative use of language) one pre xes a performative verb, as in “I tell you that Biden is the legitimate president of the U. S.” This duality is a common feature of behabitives, e.g., apologies, congratulations, thanks, welcomes (HDTWW, 47). It may even be that some verdictives and exercitives fall into the category of performative/constatives—I think it is a mistake to think so, but let’s allow that it’s possible. Would this help Langton’s case? Would it enable her to reconcile (what she sees as) the performative and the constative function of pornography? No, because of the second important point: there must be a systematic relation between the state of a airs brought about by a performative and the state of a airs described by the constative. In Austin’s one example of a “dual-purpose” performative, the state of a airs described is expressed by the propositional complement of the performative act. In the case of the warning, the state of a airs brought about performatively is the fact that I have warned you that the bull is going to charge. In so warning you, I also describe that state of a airs. There is systematicity here—because expositives are expressed by verbs of propositional attitudes, they carry a propositional commitment with them. This restricts the states of a airs that can be the subject of a constative act. This relation between the performative—what one does in a

rming, warning, revealing, etc.—and the content of what is a

rmed, warned about, or revealed, places

limits on the relation between what is done performatively and what is done constatively. This relation does not hold between the state of a airs that Langton wants to say pornography brings about, i.e., the subordination of women, and the state of a airs that she wants to say that pornography says—e.g., that “women want to be sexually dominated.” There is, I contend, no illocutionary act that can simultaneously make it the case that (a) women are subordinated and (b) constatively expresses the proposition that women want to be sexually dominated. No doubt, if MacKinnon is correct about the way that pornography works, the promulgation of lies about women is intimately related to the subordination of women, but the relation is still an external, causal relation, and not the kind of internal relation that allows the assertion that the bull is going to charge to serve also as a warning. Those are my two main criticisms of Langton’s proposal and the Langton/Hornsby treatment of 35

pornography.

But there are several other important criticisms that I want to review here.

2.1. Could Pornography Be a Speech Act? Pornography is typically, a representational product: lms, still pictures, or stories. But representations are p. 226

not themselves acts of any kind. As Jennifer Saul argues, a 36

anything.

representation on its own does not say

It is only when the representation is deployed that it can do anything. And such a deployment

requires an agent, and a speci c context. Saul gives the example of a sign that has the sentence “I do” printed on it. Brandished by one agent in one context, the sign can be part of the speech act of marrying. Brandished in a di erent context, by a di erent agent, it can e ect the speech act of confessing to a crime. And what’s true of discursive representations is equally true of gestural and pictorial representations. My raising my hand in one context signals to a friend; in another context, it registers a vote. How, then, can pornography constitute a speech act: who is the actor, and what is the context that determines what the pornographic item “says”? There certainly are acts involving pornography: there is the production of it and the consumption of it. But it doesn’t seem promising to say that the producer of pornography is the speaker. This might seem plausible if we focus on pornographic performances, but the creation of pornography doesn’t necessarily involve an audience. Nor must it be produced with any communicative intent. (Whether you think that the aim of turning someone on ought to involve an intention to communicate, the fact is that it needn’t.) Of course, a lot of pornography involves lots of people: writers, directors, actors, technicians— but none of these individuals need be or are typically intended to be audience for the pornographic “message.” Now there are acts involved in the consumption of pornography—there is the promulgation, display, and viewing of pornographic materials. But these acts are even less plausibly regarded as speech 37

acts than the act of producing pornography.

2.2. Conventions and Authority As I’ve emphasized, Austin makes it a requirement for the performance of a speech act to be successful that [t]here must exist an accepted conventional procedure having a certain conventional e ect, that procedure to include the uttering of certain words by certain persons in certain circumstances. (HDTWW, 14) Many writers, myself included, have argued that whether or not pornography subordinates women, it does not do so in accordance with this condition. There is, in the rst place, no speci c convention for making a 38

pornographic “statement,” in the way there are conventions for making a bet, or christening a ship.

But suppose that we were to grant that the conventions of, well, conventional pornography do constitute it as pornography, that they are genuinely de nitive of pornography. These conventions could still not constitute pornography as an act of subordination. For Austin, it is essential to the successful performance of a speech act that the conventions governing a speech act of the pertinent type be accepted. There are two important aspects of “acceptance” for Austin, neither of which is present in the case of pornography. p. 227

The rst is that the conventions in play must be accepted by everyone involved in the conventions governing the speech act. If somebody issues a performative utterance, and the utterance is classi ed as a mis re because the procedure is not accepted, it is presumably persons other than the speaker who do not accept it. (HDTWW, 27) Suppose, Austin says, you are trying to pick sides for a casual game. When you say, “I pick George,” George refuses, saying, “I’m not playing.” That’s all George needs to do to prevent its being the case that he has

been “picked”—he simply indicates that he does not accept the convention of picking sides (or at least not in this case; HDTWW, 28). If there were to be a kind of “subordination” that could be brought into being through an illocutionary act, then, whatever the conventions might be that speci ed the conditions that constituted such “subordination,” all anyone would need to do to avoid being “subordinated” is to simply not accept those conventions. A woman who wished to avoid this conventional form of subordination could simply say “I’m not playing.” Let me emphasize, lest I be misunderstood, that I am not saying that a woman—or anyone— can simply will themself outside of a causal network that involves their material subordination. It has been part of my case that this sort of subordination—the sort that is rightly deplored and that ought to be struggled against—is something that typically happens against a person’s will. It is a matter of the person’s material circumstances, and not a matter of their state of mind. I’m saying that if we are not talking about a contingent causal process whereby pornography materially subordinates women, but rather about a process whereby a conventional form of subordination is meant to be enacted through a pornographic speech act, this would be a form of subordination that a woman could easily avoid simply by not accepting the conventions—by not agreeing to play the pornographic game. The second point about acceptance is that it is, for Austin, necessarily voluntary. No one can be forced to accept a convention—not even if they had accepted it “hitherto.” Austin says: “[I]t must remain in principle open to anyone to reject any procedure” (HDTWW, 29). He is adamant that the fact of acceptance is not something that can be simply read o

manifest behavior:

[F]or a procedure to be accepted involves more than for it merely to be the case that it is in fact generally used, even actually by the persons now concerned; and that it must remain in principle open for anyone to reject any procedure—even one that he has hitherto accepted …. Above all all [sic] must not be put into at factual circumstances; for this is subject to the old objection to deriving an ‘ought’ from an ‘is.’ (Being accepted is not a circumstance in the right sense.). (HDTWW, 29) Once again, let us suppose that there is some set of conventions governing the conditions under which pornography can “subordinate” someone. And let us suppose that there are many women who behave or p. 228

come to behave in ways that conform to the way that

“subordinated” women behave, according to these

conventions. That conformity would not, ipso facto, show that such women accepted the conventions. If a woman did not genuinely accept the conventions, then she would not be “subordinated.” The notion of authority as it functions in Austin’s theory must be understood in terms of its connection to conventions; and in terms of what is required of the persons potentially subject to them. As Austin explains, it is often a part of the conventions that constitute various types of speech act that certain individuals are granted authority to do various things. Jurors are granted the authority to declare a person (judicially) “guilty” or “not guilty,” certain o

cers of the state are granted the authority to perform marriages, and

Catholic clergy are authorized to perform certain sacramental rites. But Austin is perfectly clear that the kind of authority that is involved in these cases is dependent on the acceptance of the conventions that construct that authority. A ship’s captain is, by convention, authorized to issue orders to his subordinates and even to the passengers while the ship is at sea. If the ship is wrecked, however, and the survivors no 39

longer feel inclined to accept the authority of the captain, the captain’s authority disappears.

The question has arisen, within the critical discussion of Langton’s view, whether pornography has the requisite authority to perform an illocutionary act of subordination. I think that the bulk of this discussion involves the con ation of the notion of authority pertinent to the theory of speech acts, with the notion of 40

power.

Austin, however, clearly marks this distinction.

We can see this in what Austin has to say about the possible consequences of a person’s not accepting a convention. Regarding such a person—say, one who rejects a code of honor—Austin writes: “One who does so is, of course, liable to sanctions” (HDTWW, 29). The imposition of sanctions, however, would, in such a case, be an exercise of power, not a legitimate exercise of authority. It’s important to note, in general, that while authority is limited, power is not. As too many Americans know from personal experience, there is a very large di erence between the actions police o

cers are authorized to perform, and the ones they have

the power to perform. Similarly, there is a di erence between what actions a certain institutional role authorizes, and what actions a person holding that role has the power to perform. Langton points to empirical evidence that boys and men who are regularly exposed to pornography are more likely to “view women as inferior” and more likely to “accept rape myths as true.” These facts about the changes in men’s view of women, are, she says, poorly explained by an assumption that pornography has no authority; they are better explained by 41

an assumption that pornography does have authority, in a certain domain.

But the “authority” that Langton assigns to pornography is really a power—the power to be convincing. (And convincing, it should be noted, is a paradigm of a perlocutionary act.) It’s true that we philosophers often speak of “epistemic authority,” but such talk is ambiguous between reference to a person’s having an institutionally-backed authorization of some sort—my being credentialed as a professor by an accredited p. 229

university

makes me an “epistemic authority” in that sense—and reference to a person’s being regarded

as a reliable source of information in some domain. Institutional practices and markings can of course in uence people in their decisions about who to trust, but as recent US politics unfortunately shows, it is too easily possible for liars and charlatans to gain the social status of reliable informants without any authorization whatsoever. The question of who has “epistemic authority” in this de facto sense is, unlike the question of who has authority constructed by an accepted conventional practice, plainly a matter of the “ at factual circumstances.” This is not at all to deny that women—and other members of marginalized and slandered social groups— can be profoundly a ected—and badly harmed—by what people believe about them. Again, recent events in the United States make this vivid: the FBI reported that in 2019 there were 7,314 hate crimes, a 3 percent increase from the previous year. These include crimes directed against people and property in virtue of their race (3,963), Hispanic ethnicity (527, a 9% increase from the previous year), religion (1,521), sexual 42

orientation (1,195) and gender identity (198).

It is perfectly obvious that this rise in violence is connected

to the massive disinformation campaigns mounted by “news” outlets like Fox News, which o er their own “authorization” to the rantings of Q-Anon supporters and other fringe elements and are egged on by corrupt Republican politicians. It may even be correct to regard the social networks that look to right-wing outlets for their “information” as constituting a kind of conventional structure that “authorizes” people like Tucker Carlson and Joe Rogan to construct, and determine the creed of, a community of practice in a proper Austinian way. But such “authorities” are only authorities to the extent that their communities so regard them. Langton may be correct in seeing some similar emergent institutional/conventional structure in the pornography industry. But even if there were grounds for saying that pornography conferred, in a convention-backed way, a certain kind of authority upon the producers of pornography, thus vindicating Langton’s claim that their “utterances” are, in Austin’s sense, “authoritative,” the e ects of those utterances on women would still be downstream from the utterances themselves—perlocutionary e ects 43

rather than a part of what is done in saying what they say.

Here I return to the earlier point about what

kinds of states of a airs can and cannot be brought about in saying something. I am a properly authorized professor of philosophy. As such, I can declare that writing a paper is a requirement for passing my course, and my declaration makes it so. But my institutionally granted power to make things the case by declaration

does not extend to matters that simply cannot be made true or false by the adoption of conventions. My declaration that dualism is false does not—cannot—make it the case that dualism is false. (More’s the pity!)

3. Silencing The second way in which feminist philosophers have appealed to Austinian speech-act theory is in accounts p. 230

of the discursive dynamics involved in sexual assault. The idea, 45

elaborated by Jennifer Hornsby,

44

which was introduced by Langton,

and

depends on a notion that Austin introduces toward the end of HTDTWW,

the notion of “uptake.” Having drawn and elaborated the distinction between illocution and perlocution in terms of what is internal to the act of speech—what is done “in saying” something—and what is external to the act of speech—an e ect of the speech act, what is done “by saying” something, he makes a quali cation in Lecture 9. I have so far argued, then, that we can have hopes of isolating the illocutionary act from the perlocutionary as producing consequences, and that it is not itself a ‘consequence’ of the illocutionary act. Now, however, I must point out that the illocutionary act as distinct from the perlocutionary is connected with the production of e ects in certain senses …. …. Unless a certain e ect is achieved, the illocutionary act will not have been happily, successfully performed. This is not to say that the illocutionary act is the achieving of a certain e ect. [my emphasis] An e ect must be achieved on the audience if the illocutionary act is to be carried out. How should we best put it here?. …Generally the e ect amounts to bringing about the understanding of the meaning and force of the locution. So the performance of an illocutionary act involves the securing of uptake [emphasis original]. (HDTWW, 116–117) Langton and Hornsby point out that consent and refusal are illocutionary acts. (Austin lists consent as an example of a commissive illocutionary act (HDTWW, 158)—he’d presumably agree that refusal falls into this category as well.) Given the requirement of uptake, it follows that one cannot consent or refuse unless one’s audience understands what one is saying, and understands what speech act one is trying to perform. In particular, it follows that it is impossible for a woman to either consent to, or—more pertinently—refuse 46

sex from a man if the man does not recognize that she has refused.

Pornography, they continue, creates

conditions in which many men do not recognize a woman’s saying “I refuse” as a refusal—they interpret it as simply an initial move in a sexual language-game, a coy way of expressing consent. In such conditions, a woman’s attempts to refuse mis re—they are thwarted, because there has been no uptake. It is not that she has refused, and that her refusal has been ignored—it is as if she has not spoken at all. She has been silenced. This form of silencing has come to be known in the literature as “illocutionary disablement.” The Langton-Hornsby analysis has received an enormous amount of attention, both from critics and from defenders. The analysis has been elaborated or extended by many philosophers to cover a variety of ways in which the sti ing or ignoring of women’s speech gures in their oppression, and to understand, more 47

generally, the role of language in creating and maintaining social injustice.

I want to focus here on the

criticisms of the view. The rst issue has to do with the way in which Langton and Hornsby use the notion of “uptake.” I quoted Austin’s introduction of the notion at length, because I think it betrays his uncertainty about how to t the notion into his framework. The relevant passage is certainly confusing. He says rst: “Unless a certain p. 231

e ect is achieved, the illocutionary

act will not have been happily, successfully performed,” then

immediately denies that “the illocutionary act is the achieving of a certain e ect,” and then, seemingly, a

rms

precisely what he just denied: “An e ect must be achieved on the audience if the illocutionary act is to be carried out.” My own interpretation is that he wants the necessary e ect on the audience to be part of the background requirements that constitute the A/B felicity of the speech act type in question, rather than an e ect that the speech act must achieve to be successful: the constitutive requirements for several kinds of speech acts involve other people, and, indeed, other people who must (as per passages quoted earlier) accept the pertinent conventions. Thus, my betting requires there to be a person with whom I mean to bet, and who is willing to accept the bet. My marrying requires there to be someone for me to marry, who also intends to marry me. It is not that, by betting, I induce anyone to participate in the bet, nor that by marrying, I persuade my beloved to become my spouse. It is, furthermore, certainly part of acceptance that the other involved parties understand what’s being said, and they understand the signi cance of such saying in the circumstances at hand. And for some speech acts, it does seem reasonable to require that there be an audience for one’s performance. But the general requirement that there be an audience for the performance of a speech act of any type seems wrong, by Austin’s own lights. He says, of commissives in general: “The whole point of a commisssive is to commit the speaker to a certain course of action” (HDTWW, 67). It is not at all obvious why the making of such a commitment requires an audience, and quite a few of Austin’s own examples seem to be things that one can do in the privacy of one’s own study (or bathtub, for that matter): • undertake • am determined to • intend • mean to • shall • vow • embrace • favour (HDTWW, 157–158) I would add “pray.” Even if I, an atheist, am correct that God does not exist, I maintain that it is still true 48

that people pray.

An additional argument against taking uptake to be a strictly necessary condition for the performance of an illocutionary act is that doing so has a number of highly counter-intuitive consequences. Dan Jacobson argues that if we agree with Langton and Hornsby that a man’s failure to “take up” a woman’s attempted refusal of sex means that she literally has not refused, then, strictly speaking, if the man proceeds to have sex with the woman, it cannot be said that he did so with mens rea. The “fact” that she did not refuse sex could presumably be appealed to in order to prevent the conviction, or even the indictment of a person who 49

otherwise could be convicted of, or even charged with rape. p. 232

examples, including this one: a burglar

Alexander Bird gives some more mundane

ignores a homeowner’s posted sign that says “Beware of dog,”

believing, incorrectly, that the sign is not a warning but a blind. When he is attacked by the very real dog, he 50

cannot say that he was not warned.

A nal example is calling for help. This seems as good an example of an

illocutionary act as many on Austin’s list, and it is often—if not typically—performed in circumstances where there is no help immediately in sight. The cases I’ve just referred to are all ones in which there is no audience, because there is no other person available to serve as an audience. But a strict interpretation of the uptake requirement—and the interpretation that the Langton-Hornsby position needs—also entails that there are circumstances in

which there is another person (or persons) present when the locutionary act is performed, but in which the other person present fails to hear, understand, or attend to the utterance. I have pointed out that people often call for help when they don’t know whether anyone will hear them, and claimed that the speech act isn’t a failure if no one does. Similarly, I contend that one can successfully call for help even if everyone who does hear you does not understand the language that you are speaking. Finally, I don’t think that someone’s failure to hear me—whether it’s because they are deaf or because they aren’t paying attention—entails that I haven’t said anything. But let us put all this aside, and concede, for the sake of argument, that Langton and Hornsby are interpreting Austin correctly, and thus that a lack of uptake results in the mis re of an attempted illocutionary act. It is still a separate step, however, to say, as Langton and Hornsby do, that an audience member’s failure to provide uptake constitutes silencing. “Silencing” sounds like an activity, something that one engages in, or does. But if I fail to hear what someone is saying, or if someone says something in a language that I don’t understand (and there’s no one around who does understand), it seems odd to say that I’ve done anything. Now, Langton and Hornsby don’t have to say that silencing occurs in all cases in which a speech act is thwarted because of lack of uptake; they can allow that in the kinds of cases I’ve just mentioned, the “thwarted” is a no-fault matter, akin to cases in which some element of the background conditions required for a particular kind of speech act is, unbeknownst to the participant(s), has not been satis ed. (For example: An unnoticed clerical error resulted in the non-renewal of the license of the wedding o

ciant.) They could lay down additional conditions on a thwarting’s being a silencing. But what

would these conditions be? One might be that the would-be illocutionary disabler would have to take steps to prevent uptake, say, by stopping their ears, loudly singing “la la la,” or running away till they’re out of earshot. But of course, this condition wouldn’t be satis ed in the cases Langton and Hornsby are most interested in classifying as cases of “silencing.” According to them, pornography silences women by inculcating false beliefs in men’s minds to the e ect that no woman would refuse sex, and that apparent refusals are, as I put it above, simply coy moves in a complicated sexual game. As they envision the situation a man whose beliefs have been corrupted in this way thwarts a woman’s attempted act of refusal by failing to understand it. But then it looks like the man in the situation has done nothing wrong. It might have been wrong for him to consume pornography (although here there are lots of issues about individual culpability), p. 233

but given that he did, pornography has shaped his

social cognition in a way that makes it impossible for 51

him to understand the woman’s attempted refusal as a refusal. But now we come to a further di

culty in Langton and Hornsby’s account of silencing. If we accede to the

criteria their account presumes, so that a man who fails to take up a woman’s attempted refusal as a refusal counts as silencing her, then we need to consider the question whether silencing is always wrong. Suppose that I am passing a demonstration by White supremacists, and I stop my ears to avoid hearing the rantings of a Proud Boy. Have I silenced them? It appears, by the Langton-Hornsby criteria that I have, but if so, it also appears that I have every right to do what I’ve done. We are not obliged to listen to, much less respond to, anything that anyone happens to say in our presence. This is not the legal matter of whether or not one has the constitutional right to say various things, or to say them in various circumstances, but rather the moral question of whether one has the right to be listened to. It seems clear to me that while one might have this right in some cases, one doesn’t have it in many others. The di erence does not turn on any technical 52

features of speech act theory, but rather on a host of extralinguistic contextual matters.

If I am a student in a class, I have a right to be listened to, carefully and respectfully, by the teacher and by the other members of the class. Outside the classroom, neither the teachers nor the other students have any obligation to pay attention to anything I say—or if they do, it is only a matter of etiquette, with “obligations” that are easily overridden. (“Sorry, but I’m trying to concentrate on my knitting.”) But even my rights inside the classroom are conditional on my agreeing to treat the sayings of the other members of the classroom with the same degree of attention and respect that I expect from them. If I show myself to be

an uncooperative or disrespectful class member, say, by going on an anti-Muslim rant (as a student once did in my classroom), I can expect to be—I ought to be—“silenced.” In cases of intimate, or potentially intimate congress, each individual involved has a moral duty to ensure that their partner(s) consents fully to everything that transpires. Such a duty entails a duty to listen carefully and respectfully to anything their partner says, but also, more broadly, to nd out what their partner is and is not comfortable with. That is, anyone involved in an intimate or potentially intimate activity has a positive duty to verify their partner’s consent. Even a man whose judgment has been corrupted by exposure to pornography still has an obligation to make sure that his partner consents before he initiates any intimate act. This obligation has nothing to do with the conventions governing the performance of speech acts. All in all, it seems that that there are two options with respect to Austin interpretation: on the one side, we have considerations that militate against treating uptake as a necessary condition on the performance of an illocutionary act: 1. The original characterization of the illocutionary/perlocutionary distinction, which says that the illocutionary act must be complete “in the saying.” 2. The fact that Austin does not list uptake as one of the A/B felicity conditions. 3. The fact that Austin’s own examples of illocutionary acts fail the uptake condition, strictly construed. p. 234

On the other side, we have Austin’s own text stating clearly that an illocutionary act can only be successfully performed if there is uptake. What to say? I am inclined to agree with Bird, who says plainly that Austin is wrong about illocutionary acts. Uptake, he says, is just obviously not required for the successful performance of any and all illocutionary 53

acts.

4. Why Appeal to Austin? I’ve argued that Langton’s and Hornsby’s appeals to Austin’s theory of speech acts have been unsuccessful, in one way or another. Their claim that pornography is an illocutionary act of subordination involves a serious misreading of Austin’s text. In particular, they ignore Austin’s central requirement that an illocutionary act must be, rst of all, an actual act, and second, of a type constituted by a backing convention that is accepted by all. Their claim that an assailant’s failure to acknowledge a victim’ refusal as a refusal depends on a very strict reading of Austin’s discussion of “uptake.” Now, I’d rst like to point to an irony here. In order for Langton and Hornsby to get the result that pornography silences women by making men unable to hear their attempted refusals as refusals, they must rely on a very strict reading of a few passages in How To Do Things With Words, a reading that, moreover, introduces internal tensions into Austin’s view. But a similarly strict reading of Austin’s characterizations of performative illocutionary acts—in particular, the necessary role of accepted conventions, and the fact that exercitives and verdictives are not constative—defeats the speech-act analysis of pornography. If we set aside the requirement that illocutionary acts be constituted by conventions, or ignore Austin’s insistence that verdictives and exercitives are not constative, and do not describe the state of a airs they bring into being, why can we not set aside the requirement of uptake? But my main purpose in this section is to make the case that Austin’s theory of speech acts is simply irrelevant to the very real issues that surround the ubiquity of misogynistic pornography in our culture, and to the very real problem that sexual assault is still rampant in our culture, that the perpetrators of sexual assault frequently get away with it, and that women are still continually revictimized if they try to get

redress. Langton, in her reply to my “Against Langton’s Illocutionary Treatment of Pornography,” asserted that I think “it doesn’t matter whether pornography subordinates, or women are silenced, or rather, whether pornography is an illocution that subordinates, and whether women su er illocutionary disablement.” I do care, passionately, about the rst set of issues. I just do not see that the turn toward Austinian speech-act theory reveals anything about those social ills that cannot be adequately described without any appeal to technical philosophy of language. Langton and I have a number of disagreements. We disagree about how “pornography” should be de ned. I think that Langton has a very blinkered view of what pornography is, and in consequence, nds the p. 235

MacKinnon-Dworkin de nition adequate. 54

conventional pornography,

Langton’s discussions, which focus on cis-male-oriented,

do not show much awareness of the range of topic and treatment present in

pornography today. In particular, she neglects feminist, gay, and gender-non-binary pornographic material, and she also fails to consider the wide range of (even) popular mainstream pornography that focuses on female-to-male domination, as well as pornography that involves clearly consensual bondageand-domination and sadist-masochistic interactions. She does not address the aims or productions of feminist pornography. She does not tackle questions about the role that pornography has played in articulating and validating the erotic preferences and practices of individuals outside the gender binary. She and I may also disagree about the proper role of explicit, erotically stimulating material in the development and support of adult sexual enjoyment. I think that a clear-eyed view of the role of pornography in contemporary Western, heteronormative, ableist, racist culture yields a much more complex view of what “pornography” is and does right now. But what are the evils and dangers of contemporary pornographic culture that we need to address? Here, Langton and I are in agreement. One enormous problem is the role of misogynistic pornography in shaping the beliefs and expectations of young cis-men. There is ample evidence to con rm what Langton asserts is the case: that many boys and men are getting their ideas about what women want, and about what they can expect women to want to do, from heterosexist, misogynistic pornography, increasingly available on the internet. (I commend to the reader the compelling and insightful discussion of this problem in Amia 55

Srinivasan’s excellent collection of essays The Right to Sex. ) My point, here, is that the problem can be easily described without any recourse to speech-act theory. The problem is that, increasingly, boys and young men have access, via the internet, to materials that are profoundly arousing but that also feature images and narratives that are shaping their beliefs and expectations, about what girls and women want, sexually, and also conditioning their patterns of arousal in a way that conditions their sexual needs. This is the problem—how is speech-act theory supposed to help us understand or confront it? Relatedly, I do think that there is insight in MacKinnon’s view of (what Ian Hacking calls) the “looping” e ect of the promulgation of stereotypes of feminine sexual behavior—indeed, of the e ects of gender normativity in general. But this is a matter that has been a subject of discussion in feminist philosophy since 56

at least the early 1980’s.

What about the alleged contribution of speech-act theory to our understanding of the appalling dynamics involved in sexual assault? I do not, as I’ve made clear, think that speech-act theory has the clear implications for characterizing the harms perpetrated in these cases that Langton and Hornsby take there to be. But even if they did, it is pertinent to assessing the political value of the analysis to consider the scale of the problem that the speech-act analysis of silencing could cover. How often is it the case that a sexual assailant really does not recognize that he is acting against the wishes of his victim? The focus of the work on “silencing” is, presumably, cases in which the assailant knew his victim and could, with some minimal p. 236

plausibility, claim that he actually believed that

she wanted the sexual activity to take place. However, a

study by David Lisak, a clinical psychologist, indicates that a large percentage of rapes, including rapes where the assailant was known to the victim, involved drug- or alcohol-induced incapacitations.

Lisak surveyed 1,882 male college students between 1991 and 1998 and found 120 whose descriptions of their sexual activities met the criteria for rape or attempted rape, but who had not been charged. Eighty-one percent of these undetected rapists reported that they had had sex with women who were incapacitated because of drugs or alcohol …. Assaulting incapacitated women was, by far, the most common method of sexual assault — far fewer rapists said they had used threats or overt force for sexual intercourse (9 percent). Serial rapists and one-time rapists were 57

equally likely to assault women who were incapacitated because of drugs or alcohol.

Clearly, these data show that, in a large percentage of rape cases, the assailant is simply indi erent to the preferences of his victim, to the extent that assailants are perfectly willing to take advantage of their victims’ inability to register any preferences at all. (It’s not the case that assailants in this study were known to have or were presumed to have caused the incapacitation that they took advantage of.) So how much more do we know about the dynamics of rape if we add the speech-act analysis of “silencing”? Not much, I say. On the other end of things, I want to say something about the value of getting the philosophy straight. Langton, in her response to my criticism of her speech-act analysis of pornography, writes: Antony is a methodological pessimist, who doubts philosophy of language has much to o er the 58

political theorist.

I don’t know why Langton would infer from the fact that I criticize a particular proposal for applying a particular view in the philosophy of language to a particular issue, that I oppose any application of results in the philosophy of language to issues in politics or political theory. For the record, I think that insights from the philosophy of language have a great deal of relevance to current political issues and, in particular, a 59

great deal to o er those who strive to understand the role of language in the shaping of political opinion.

But I am, also, a philosopher of language, who believes that philosophical advances and achievements can stand on their own; that their value for our understanding of language and the mind are—provisionally, at least—independent of whatever application they may have in the extra-philosophical world. It’s important, I think, to understand these achievements on their own terms. I’m committed to the idea that all the true 60

things will cohere in the end,

but that doesn’t mean that we should try to interpret philosophical work

under the constraint that it has to make some readily discernible di erence to things we care about practically. Indeed, my faith that in the end “the truth cannot be an enemy to the truth” gives me p. 237

con dence that a rigorous consideration of

the philosophical questions, on their own terms, will

eventually cohere with and facilitate the achievement of our practical, moral, and political goals.

Notes 1.

Langton (1993, 283–330).

2.

Hornsby (1995, 127–147).

3.

Hornsby and Langton (1998, 21–37).

4.

In what follows, I will be relying on the Harvard University Press edition of J. L. Austinʼs How to Do Things with Words, edited by J. O. Urmson and Marina Sbisà (Austin 1955/1975), herea er cited in the text as “HDOTWW, page number.” All emphases within quotes will be Austinʼs unless marked as mine.

5.

I am not trying to provide a comprehensive survey of all the work inspired by the Langton-Hornsby appeal to Austin. For a more complete listing of this work, together with a discussion of connections between it and other feminist work on issues in the philosophy of language, see Jennifer Saul, Esa Diaz-Leon, and Samia Hesni (2022), https://plato.stanford.edu/archives/fall2022/entries/feminism-language/. Accessed September 10, 2023.

p. 238

6.

Srinivasan (2021).

7.

Morgan (1978, 163–169).

8.

See the essays in MacKinnon (1987), especially “Di erence and Dominance: On Sex Discrimination” (Ch. 2) and “Not a Moral Issue” (Ch. 13)

9.

MacKinnon was perfectly willing to embrace the implication of this view that some individuals socially recognized as men would count as women if they were in a sexually subordinated role—she thought this was likely true of some gay men; the same held, mutatis mutandis, for some individuals socially recognized as women.

10.

MacKinnon has virtually nothing to say about how the pornography “textbook” functions to socialize girls and women into eroticizing submission. (She might have cited a work by Ann Snitow that examines the narrative tropes in romance novels. See Snitow [1979, 141–161], https://doi.org/10.1215/01636545-1979-20-141). My own (nonscientific) observations suggest that girls and women are trained in multiple, largely nonpornographic ways to view themselves as beings who exist for others. (Note that this was the central theme of Simone de Beauvoirʼs existentialist feminism in The Second Sex). But on MacKinnonʼs own account, it is not womenʼs consumption of pornography that “makes” them into the kind of creature pornography says they are. It is not that they view pornography and identify with the women depicted, but rather, that they must interact with men whose preferences and expectations have been shaped by their identification with the men depicted (or with the imagined viewer of the pornographic scene).

11.

Although the same employerʼs saying “Blacks are inferior to whites” is protected speech.

12.

American Booksellers Association, Inc. v. Hudnut, US Court of Appeals for the Seventh Circuit: 771 F.2d 323 (7th Cir. 1985), sec. III, para. 6. https://law.justia.com/cases/federal/appellate-courts/F2/771/323/379919/.

13.

American Booksellers v. Hudnut, sec. III, para. 7.

14.

American Booksellers Assʼn, Inc. v. Hudnut, 598 F. Supp. 1316 (S.D. Ind. 1984) US District Court for the Southern District of Indiana—598 F. Supp. 1316 (S.D. Ind. 1984) November 19, 1984. https://law.justia.com/cases/federal/districtcourts/FSupp/598/1316/1476351/ (location 1330).

15.

W. A. Parent (1990, 205–211).

16.

Langton (1993, 28).

17.

Langton (1993). Thereʼs an interesting di iculty with Langtonʼs strategy: if the crucial point in MacKinnonʼs strategy is to establish that pornography is “conduct,” then it would seem that the case against it would be that, as conduct, it is not covered by the First Amendment—discriminatory job advertisements, for example, are not so covered. But MacKinnon explicitly denies that this is her strategy. Langton notes this (p. 28) and responds by distinguishing between the claim that “pornography is a kind of act” and the claim that “pornography is conduct.” She says that she is making the first claim, not the second, and so is not misinterpreting MacKinnon. I find it di icult to make the distinction Langton is drawing (except with respect to the syntactic frames of the two terms, act and conduct) or at least to appreciate its significance in this context, apart from the fact that making it allows Langton to square her interpretation of MacKinnon with what looks like an explicit disavowal of it by MacKinnon.

18.

Hornsby agrees with Langton that this is Austinʼs main theoretical point, but adds that his requiring “uptake” in order for an illocutionary act to have been performed shows that he is also fundamentally concerned with communication. Iʼll discuss this move below.

19.

On p. 98, he speaks of “our problem of the constative as opposed to the performative utterance.”

20.

The TV announcerʼs saying this does not make it so, as the debacle at the 2017 Academy Awards show, when the presenters announced the wrong movie as the best motion picture, demonstrated. “The 2017 Academy Awards telecast was coming to an end. The best picture had just been announced, the producers of ʼLa La Landʼ were making their acceptance speeches and journalists were finishing up their stories, hoping to go home soon. Then, suddenly, producer Jordan Horowitz was at the mike, waving a white card: “Thereʼs a mistake. ʻMoonlight,ʼ you guys won best picture. This is not a joke.ʼ ” Maria Puente, Andrea Mandell, and Bryan Alexander [2018], “We Were There: How the Worst Flub in Oscar History Went Down,” USA Today, February 28. https://www.usatoday.com/story/life/2018/02/28/we-were-there-how-

worst-flub-oscar-history-went-down/377305002/). Announcers are not umpires or judges; they are not properly authorized to perform the requisite exercitive speech act. 21.

Iʼll discuss this very important condition below.

22.

Which are, of course, also creatures of convention.

23.

HDTWW, 119. See also Lecture 10: “Illocutionary acts are conventional acts: perlocutionary acts are not conventional” (HDTWW, 121; emphasis in original).

24.

Iʼll discuss below the di erence between authority, the “acceptance” of which Austin makes a necessary condition for the performance of an illocutionary act, and power, which is independent of the existence or acceptance of conventions.

25.

Rae Langton (2009, 89).

26.

Austinʼs list of examples of verdictives (HDTWW, 153) includes “estimate,” “take it,” and “characterize.”

27.

John Searle, in response to this and other problems he sees with Austinʼs taxonomy of speech acts, proposes a di erent taxonomy involving a new taxonomic dimension: direction of fit. Many speech acts classified by Austin as verdictives, such as a jury verdict, become, on Searleʼs taxonomy, members of the new category “declaratives.” One of the distinctive features of speech acts in this class is that they have a bidirectional direction of fit—they are completely felicitous only if both (a) the declaration made fits the facts (word-to-world), and (b) the state of a airs declared comes into being (world-to-word). Searle never really tells us what this means for cases where one or the other direction fails—he says that a bad call from an umpire is still valid “for the purpose of baseball.” See Searle (1979, 1–23). In any case, Langtonʼs treatment of verdictives and exercitives is drawn from Austin, so Iʼll continue to concentrate on his own statement of the theory rather than Searleʼs revision of it.

28.

In Lecture 9, where Austin is pulling back from some of the distinctions between performatives and constatives that he drew, he considers the question whether performatives can, like constatives, be assessed in terms of truth and falsity. His answer is equivocal; he writes: “[T]here is an obvious slide toward truth or falsity in the case of, for example, verdictives…. We shall not say ʻtrulyʼ in the case of verdictives, but we shall certainly address ourselves to the same question.” (HDTWW, 11, 141) He never hints, however, that in a case where one might “slide toward” calling a verdict false or incorrect that the verdict would not have been e ective, that the verdictive speech act would not have come o . I see an analogy with the issuance of a false statement—the statement has been made, the illocutionary act has been performed, but unhappily. So truth and falsity, or their analogues, should be taken as falling into the Γ-type category of infelicities.

29.

The very fact that MacKinnon articulates a mechanism by which pornography e ects the subordination of women seems to me su icient reason for thinking that the illocutionary analysis is wrong, but I set that aside.

30.

A slightly di erent approach would be to treat the pornographer as having “makerʼs knowledge” of the inferiority of women. I am quite skeptical of this notion, but I will not rehearse my objections to it here.

31.

Mary Kate McGowen notes an internal problem with this package of ideas in MacKinnon—if pornography does “construct” women to be the way it says they are, then it looks like pornography canʼt be telling lies at all—women actually are desirous of sexual domination. McGowen believes that she can solve this problem by identifying a new kind of speech act, the “conversational exercitive.” See McGowan (2003, 155–189).

32.

See the research applying “script theory” to pornography: Bennett et al. (2019, 647–667); and Bridges et al. (2016). https://doi.org/10.1177/2374623816668275.

33.

Itʼs necessary, however, to emphasize that feminist philosophers Marilyn Frye and Sandra Bartky have long ago o ered accounts of the mechanisms by which gender is socially constructed. Frye explains how the promulgation of gender norms can, by creating incentives to conform to stereotypical gender ideals, generate “confirming instances” of those stereotypes. See Marilyn Frye (1983). Bartky discusses the ways in which sexist stereotypes and norms can “colonize” the consciousness, and the desire, of women, again, in ways that produce “confirming instances” of the stereotypes. See Sandra Lee Bartky (1990). More recently, Sally Haslanger has produced a series of important papers investigating the multiple ways in which “natural” kinds can be involved in the construction of social kinds. See especially, “Feminism in Metaphysics: Negotiating the Natural” and “Social Construction: Myth and Reality,” both in Haslanger (2012a).

34.

Austin ties this point to his anti-correspondentist, deflationary, and possibly conventionalist, view of truth: “It is essential

p. 239

to realize that ʻtrueʼ and ʻfalse,ʼ like ʻfreeʼ and ʻunfree,ʼ do not stand for anything simple at all; but only for a general dimension of being a right or proper thing to say as opposed to a wrong thing, in these circumstances, to this audience, for these purposes and with these intentions” (HDTWW, 145). And on HDTWW, page 149, he writes that “truth and falsity are … not names for relations, qualities or whatnot, but for a dimension of assessment.” Anyone who espouses a more substantialist view of truth has a basis, I think, for maintaining the constative/performative distinction in its original form.

p. 240

p. 241

35.

For a more detailed exposition of each of these criticisms, see my “Pornography and the Philosophy of Language” (Antony 2014); and “Be What I Say” (Antony 2016), both also reprinted in Antony (2022).

36.

Itʼs worth noting in this connection that many representations that are not intended to be pornographic can still be used for erotic purposes. Quite a few men of my generation report their nostalgia for the mild titillation they derived as boys from National Geographic and the Sears Roebuck catalogue. More seriously, it is an expressed concern among responsible filmmakers dealing with the topic of rape that explicit depictions of sexual violence might be consumed as pornography. See Lena Wilson (2017), https://theplaylist.net/problematic-history-rape-scenes-film-20171026/2/.

37.

See Jennifer Saul (2006, 61–80), doi:10.1111/j.1467-9264.2006.00146.x.

38.

There are, of course, conventional elements in many pornographic productions—standard poses for models, for example, and stock narrative themes. But these conventions do not constitute pornography as a speech act; they simply help to characterize the genre, and make the product legible. Indeed, they are not even necessary to somethingʼs being pornography.

39.

I know some of you are thinking about the television show Gilliganʼs Island. The fact is that the castaways never clearly established who their leader was, despite the Skipperʼs retaining the name “Skipper.” There was an election, which was won, in a surprise, by Gilligan, but he ended up repudiating his “presidency” once he realized that he lacked the power, despite having the authority, to get anyone to do anything. “President Gilligan,” episode 1, season1, aired October 31, 1964, Gilliganʼs Island Wiki, https://gilligan.fandom.com/wiki/President_Gilligan.

40.

Leslie Green raises this issue in Green (1998, 285–311). Langton responds in “Pornographyʼs Authority? Response to Leslie Green,” in Langton (2009d, 89–99), doi:10.1093/acprof:oso/9780199247066.003.0005. Mary Kate McGowen argues that pornographers have the requisite authority in (McGowan 2003, 155–189), doi:10.1111/j.1088-4963.2003.00155.x.

41.

Rae Langton (2009b).

42.

Data from the Southern Poverty Law Center: SPLC Responds to New FBI Report Showing Increase in Hate Crimes in 2019, November 16, 2020 https://www.splcenter.org/presscenter/splc-responds-new-fbi-report-showing-increase-hate-crimes2019.

43.

Alexander Bird, in his “Illocutionary Silencing,” provides a detailed discussion of various kinds of authority, and provides a taxonomy that is consistent with the distinction Iʼm drawing here. He also makes the important point that if pornography does have “practical” or “e ective” authority over anyone, it is over boys and men, not women (cf. note 10 above). E ects on women of pornographyʼs “authority” would therefore be causally downstream from the pornographic speech act, and thus perlocutionary. See Alexander Bird (2002, 1–15).

44.

Langton (1993).

45.

Hornsby (2002, 1–15).

46.

I will use the terms “woman” and “man” to designate the victim and the assailant, respectively, in an act of rape, as was common at the time of publication of the literature Iʼm discussing. But I fully acknowledge that women and non-binary individuals can also be rapists, and that men and non-binary individuals can also be victims.

47.

See, for example, Maitra (2009, 309–338); McGowan (2009, 389–407); Gruenberg (2014, 173–199); and Johnson (2020).

48.

Ishani Maitra agrees that it is implausible to treat uptake as a general condition on illocutionary acts, but she does think we should take it seriously as a condition on communicative acts. Maitra proposes to repair the account so as to support the Langton-Hornsby account of silencing by marrying Austinʼs account of speech acts to Griceʼs account of meaning. I see nothing wrong with this marriage, but I disagree with its providing support for the view that pornography silences women. Maitra notes that Hornsby seems to assume that all illocutionary acts are communicative. I agree, and believe that this mistaken assumption stems from Hornsbyʼs misunderstanding of Austinʼs main goal in HTDTWW, which, like Langtonʼs,

focuses on the fact that speech can also be action—a point, which, as Iʼve said, Austin regards as a truism. Hornsby goes further, and imputes to Austin a central concern with communication, which is, of course, essentially social. This is, I contend, a misreading. See Maitra (2009, 309–338).

p. 242

49.

Jacobson (1995). See also Nellie Wieland (2007, 435–456), doi:10.1080/00048400701572196. Although most jurisdictions have replaced “rape-by-force” definitions of “rape” with “rape-by-non-consent” definitions, persons accused of rape are still being acquitted when jurors or judges believe that thereʼs credible evidence that the accused did not know that the accuser had not consented. “[E]ven when rape is defined broadly, the mens rea of knowledge requires proof that the defendant in fact knew he was having sex without his partnerʼs consent. When framed in this manner, it is possible for the jury to both believe a womanʼs testimony that she was raped but not have evidence that the defendant knew the victim was not consenting.” Kari Hong (2018, 262), https://www.law.georgetown.edu/american-criminal-law-review/wp-content/uploads/sites/15/2018/04/55-2-A-NewMens-Rea-for-Rape-More-Convictions-and-Less-Punishment.pdf.

50.

Bird (2002).

51.

This is part of Jacobsonʼs argument that the Langton-Hornbsy analysis lets rapists o the hook.

52.

Samia Hesni agrees that Langton and Hornsby go wrong in treating any case in which uptake fails as a case of silencing, as a result of what she calls “internal tensions.” Hesni argues that their analysis incorrectly classifies as di erent two cases that ought to be classified together: they classify a case in which an individual refuses sex but is taken to be consenting as a case of illocutionary disablement, but not a case in which an individual is understood to be refusing but where the refusal is ignored. Hesni proposes an alternative account of speech acts drawing on a Gricean notion of speaker meaning to bring these cases together. See Samia Hesni (2018, 947–976).

53.

See the section “Is Uptake Necessary for Refusal?” in Bird (2002, 10–12).

54.

And even so, misses many ubiquitous themes, which include feminine domination. The image of the dominatrix is so wellknown that it has become a familiar trope in mainstream movies (e.g., Batman Returns, Love Actually, The Out-of-Towners (1999), and Everything Everywhere All at Once) and television series (e.g., Family Guy, Desperate Housewives, Billions, and Life in Pieces). I donʼt say that any of these movies or shows presents a realistic treatment of fem/dom culture, but the good-natured handling of the B&D theme suggests both general familiarity with and acceptance of this sort of sex play.

55.

Srinivasan (2021). See also the episode of the podcast Science Vs. devoted to the empirical e ects of pornography: “Pornography,” Science Vs., ABC podcast, June 17, 2015, https://www.abc.net.au/radio/programs/sciencevs/pornography/8604926.

56.

See Frye (1983), Bartky (1990), and Haslanger (2000).

57.

Liah Libresco, « Rapes Assisted by Drugs Are All Too Common” FiveThirtyEight, July 8, 2015. https://fivethirtyeight.com/features/rapes-assisted-by-drugs-or-alcohol-are-all-too-common/.

58.

Langton (2011, 379–440).

59.

Iʼll mention just two writers who brilliantly apply their expertise in linguistics and philosophy of language, respectively, to matters of pressing political concern: Sally McConnell-Ginet and Jennifer Saul. Iʼm a huge fan, in both cases. Of course, I do not mean to imply that either of them would agree with the substantive conclusions Iʼve argued for here.

60.

See Antony (2016, 21–39).

References Antony, Louise. (2014). “Pornography and the Philosophy of Language.” In The Philosophy of Pornography: Contemporary Perspectives, edited by Lindsay Coleman and Jacob Held. Lanham, MD: Rowman and Littlefield. 147–176. Reprinted in Antony (2022). Antony, Louise. 2016. “Things Oughta Make Sense.” Presidential address to the Eastern Division of the American Philosophical Association, Jan. 8, 2016. In The Proceedings and Addresses of the American Philosophical Association, vol. 90, November 2016, pp. 21–39. Google Scholar WorldCat Antony, Louise. (2017). “Be What I Say.” In Beyond Speech: Pornography and Analytic Feminist Philosophy, edited by Mari Mikkola, 59–87. Oxford: Oxford University Press. Reprinted in Antony (2022). Google Scholar Google Preview WorldCat COPAC Antony, Louise. (2022). Only Natural: Gender, Knowledge, and Humankind. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Austin, J. L. (1975). How to Do Things with Words. Second Edition, edited by J. O. Urmson and Marina Sbisà. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Bartky, Sandra Lee. (1990). Femininity and Domination: Studies in the Phenomenology of Oppression. London: Routledge. Google Scholar Google Preview WorldCat COPAC Bennett, Margaret, Brittny J. LoPresti, Rory McGloin, and Amanda Denes. (2019). “The Desire for Porn and Partner? Investigating the Role of Scripts in A ectionate Communication, Sexual Desire, and Pornography Consumption and Guilt in Young Adultsʼ Romantic Relationships.” Western Journal of Communication 83 (5): 647–667. doi:10.1080/10570314.2018.1564934. Google Scholar WorldCat Bird, Alexander. (2002). “Illocutionary Silencing.” Pacific Philosophical Quarterly 83 (1): 1–15. Google Scholar WorldCat Bridges, Ana J., Chyng F. Sun, Matthew B. Ezzell, and Jennifer Johnson. (2016). “Sexual Scripts and the Sexual Behavior of Men and Women Who Use Pornography.” Sexualization, Media, & Society 2 (4). https://doi.org/10.1177/2374623816668275. Google Scholar WorldCat Frye, Marilyn. (1983). The Politics of Reality. Trumansburg, NY: Crossing Press. Google Scholar Google Preview WorldCat COPAC Green, Leslie. (1998). “Pornographizing, Subordinating, Silencing.” In Silencing: Practices of Cultural Regulation, edited by Robert Post, 285–311. Los Angeles, CA: Getty Research Institute. Google Scholar Google Preview WorldCat COPAC Gruenberg, Angela. (2014). “Saying and Doing: Speech Actions, Speech Acts and Related Events.” European Journal of Philosophy 22 (2): 173–199. Google Scholar WorldCat Haslanger, Sally. (2000). “Feminism in Metaphysics: Negotiating the Natural.” In The Cambridge Companion to Feminism in Philosophy, edited by Miranda Fricker and Jennifer Hornsby, 107–126. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC p. 243

Haslanger, Sally. (2012a). Resisting Reality: Social Construction and Social Critique, 183–218. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC

Haslanger, Sally. (2012b). “Social Construction: Myth and Reality.” In Resisting Reality: Social Construction and Social Critique, 183–218. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Hesni, Samia. (2018). “Illocutionary Frustration.” Mind 127 (508): 947–976. Google Scholar WorldCat Hong, Kari. (2018). “A New Mens Rea for Rape: More Convictions and Less Punishment.” American Criminal Law Review 55 (2): 259–332, at 262. https://www.law.georgetown.edu/american-criminal-law-review/wp-content/uploads/sites/15/2018/04/55-2-ANew-Mens-Rea-for-Rape-More-Convictions-and-Less-Punishment.pdf. Google Scholar WorldCat Hornsby, Jennifer. (1995). “Disempowered Speech.” Philosophical Topics 23 (2): 127–147. Google Scholar WorldCat Hornsby, Jennifer. (2002). “Illocutionary Silencing.” Pacific Philosophical Quarterly 83 (1): 1–15. Google Scholar WorldCat Hornsby, Jennifer, and Rae Langton. (1998). “Free Speech and Illocution.” Legal Theory 4 (1): 21–37. Google Scholar WorldCat Jacobson, D. (1995). “Freedom of Speech Acts? A Response to Langton.” Philosophy and Public A airs 24 (1): 64–78. Google Scholar WorldCat Johnson, Casey Rebecca. (2020). “Mansplaining and Illocutionary Force.” Feminist Philosophy Quarterly 6 (4). Google Scholar WorldCat Langton, Rae. 1993. “Speech Acts and Unspeakable Acts.” Philosophy and Public A airs 22 (4): 293–330. Reprinted in Langton (2009c). Google Scholar WorldCat Langton, Rae. (2009a). “Pornographyʼs Authority? Response to Leslie Green.” In Sexual Solipsism: Philosophical Essays on Pornography and Objectification. Oxford Academic (2011 online ed.). https://doi.org/10.1093/acprof:oso/9780199247066.003.0005. Reprinted in Langdon (2009c). Google Scholar Google Preview WorldCat COPAC Langton, Rae. (2009b). “Pornographyʼs Divine Command? Response to Judith Butler.” In Sexual Solipsism: Philosophical Essays on Pornography and Objectification, edited by Rae Langton, 103–116. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Langton, Rae. (2009c). Sexual Solipsism: Philosophical Essays on Pornography and Objectification. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Langton, Rae. (2011). “Replies.” Jurisprudence 2:379–440. Google Scholar WorldCat MacKinnon, Catharine. (1987). Feminism Unmodified: Discourse on Life and Law. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Maitra, Ishani. (2009). “Silencing Speech.” Canadian Journal of Philosophy 39 (2): 309–338. Google Scholar WorldCat McConnell-Ginet, Sally. (2011). Gender, Sexuality, and Meaning: Linguistic Practice and Politics. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC

McGowan, Mary Kate. (2003). “Conversational Exercitives and the Force of Pornography.” Philosophy and Public A airs 31 (2): 155–189. doi:10.1111/j.1088-4963.2003.00155.x. Google Scholar WorldCat McGowan, Mary Kate. (2009). “Oppressive Speech.” Australasian Journal of Philosophy 87 (3): 389–407. Google Scholar WorldCat Morgan, Robin. (1978). Going Too Far: The Personal Chronicle of a Feminist. New York: Random House. Google Scholar Google Preview WorldCat COPAC Parent, W. A. (1990). “A Second Look at Pornography and the Subordination of Women.” Journal of Philosophy 87 (4): 205–211. Google Scholar WorldCat Saul, Jennifer. (2006). “Pornography, Speech Acts and Context.” Proceedings of the Aristotelian Society 106 (2): 61–80. doi:10.1111/j.1467-9264.2006.00146.x. Google Scholar WorldCat Saul, Jennifer, Esa Diaz-Leon, and Samia Hesni. (2022). “Feminist Philosophy of Language.” The Stanford Encyclopedia of Philosophy (Fall 2002 edition). Edited by Edward N. Zalta and Uri Nodelman. https://plato.stanford.edu/archives/fall2022/entries/feminism-language/. WorldCat Searle, John. (1976). “A Classification of Illocutionary Acts.” Language in Society 5 (1): 1–23. Google Scholar WorldCat p. 244

Snitow, Ann Barr. (1979). “Mass Market Romance: Pornography for Women Is Di erent.” Radical History Review 1 (20): 141–161. https://doi.org/10.1215/01636545-1979-20-141. Google Scholar WorldCat Srinivasan, Amia. (2021). The Right to Sex: Feminism in the Twenty-First Century. New York: Farrar, Straus and Giroux. Google Scholar Google Preview WorldCat COPAC Wieland, Nellie. (2007). “Linguistic Authority and Convention in a Speech Act Analysis of Pornography.” Australasian Journal of Philosophy 85 (3): 435–456. https://www.tandfonline.com/doi/abs/10.1080/00048400701572196. Google Scholar WorldCat Wilson, Lena. (2017). “The Long, Problematic History of Rape Scenes in Film.” The Playlist, October 26. https://theplaylist.net/problematic-history-rape-scenes-film-20171026/2/.

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

11 Pornography as Oppressive Speech  Mari Mikkola https://doi.org/10.1093/oxfordhb/9780192844118.013.34 Published: 22 May 2024

Pages 245–264

Abstract One of Catharine MacKinnon’s central claims is that pornography is not “only words.” Rather, pornographic speech subordinates and silences women. Using Austin’s speech-act theory, Rae Langton argues that pornographic speech thus has the power to oppress women. The speech-act theoretic defense of MacKinnon’s position has dominated much of Anglo-American philosophizing about pornography over the past thirty years. This chapter examines and evaluates the philosophical legacy of the speech-act approach to pornography’s putative harms. It considers what this approach amounts to, its philosophical and practical tenability, and whether pornography can be said to have the kind of authority to make good the speech- act approach.

Keywords: pornography, feminism, Langton, subordination, silencing, speech act theory, authority Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction Although perhaps intuitively surprising, the view of pornography as oppressive speech is widely discussed in philosophy. Insofar as its manufacture and distribution are protected by free speech legislation (at least in the United States), the courts have decreed pornography to be speech. Or more accurately, pornography is a form of expression that includes spoken words and utterances, public recordings, written words, nonverbal images, and other means of expression. As speech or expression, anti-pornography feminism holds that pornographic speech depicts the subordination of women. However, in championing their antipornography stance, Catharine MacKinnon and Andrea Dworkin famously advanced a view of pornography doing more: it is a practice of sex discrimination (Dworkin 1981; MacKinnon 1987, 1989a, 1989b, 1993). The problem with pornography isn’t what it says or represents, but what it does. In short, pornography celebrates, promotes, and legitimizes sexualized violence against women. It eroticizes male dominance and female submissiveness, and puts this forward as the apparent truth about sex (MacKinnon 1987, 171). MacKinnon and Dworkin hence de ned pornography as “the graphic sexually explicit subordination of women through pictures and words” (MacKinnon 1987, 176). However, pornographic speech not only subordinates women; it also silences women. In making violence the apparent truth about sex, pornography prevents women from saying otherwise. It devastates women’s attempts to articulate experiences of sexualized violence, which due to pornography are just seen as part of ordinary sexual realities—women are thus “stripped of authority and reduced and devalidated and silenced” (MacKinnon 1987, 193). This view that pornography is subordinating and silencing speech came under sustained attacks from prominent liberal philosophers in the 1980s and 1990s. Ronald Dworkin (1991) famously claimed that the p. 246

view was based on a “dangerous confusion”

between negative and positive liberty: the distinction

between enjoying freedom from some interference and having the liberty to do something. Our negative liberty may be restricted in a manner that is consistent with free speech protections. However, the view that pornographic speech silences women seeks to argue against it by appealing to women’s positive liberty to be heard and to make speech acts. Such guarantees are not within the remit of the law, though, and the law need not protect the freedom to make certain claims, such as those of cranks and at-earthers. William Parent held that the Dworkin-MacKinnon de nition of pornography is philosophically indefensible because subordination is “an action or a practice engaged in by human beings and directed against other beings … the logic of ‘subordinates’ requires that it have some human action or actions as a subject” (Parent 1990, 208). As books, magazines, and images are not human beings, anti-pornography feminism commits a category mistake: pornographic speech simply is not the sort of thing that can subordinate. Joel Feinberg (1985) in turn argued that feminist arguments against pornography made in terms of a ‘harm principle’— that pornography harms women by inciting male violence against women—are unsuccessful. The case allegedly fails because empirical evidence does not demonstrate that pornographic speech incites violence 1

against women.

Rae Langton famously defends the cogency of the MacKinnon-Dworkin position in her classic article “Speech Acts and Unspeakable Acts” (1993). She articulates further defenses together with Jennifer Hornsby (see Hornsby and Langton 1998). Langton draws on J. L. Austin’s (1962) speech-act theory to make good the idea that pornographic speech does something harmful. Austin argued that our statements can (and do) do more than simply make true or false claims about the world—sometimes, we perform actions other than just speaking with our utterances. With this in mind, Austin divides speech acts into locutions, perlocutions, and illocutions: the speaker’s locution (the words uttered) can perform some illocutionary action (in uttering something the speaker’s locution can count as øing), and the locution can have some perlocutionary e ects (by uttering something the speaker’s locution can cause further extralinguistic e ects). Subsequently, Langton argues that pornographic speech illocutionarily subordinates and silences women. In saying something about women, pornographic speech does something other than make mere utterances. It

functions like the speech of a priest who just in declaring, “I pronounce you a married couple,” performs the action of marrying. Pornographic speech, however, performs harmful actions. It subordinates and silences women in ranking them as inferior, in legitimating discrimination against them, and in depriving women of important free speech rights (Langton 1993, 305–313). This allows us to see that the Dworkin-MacKinnon position is not philosophically indefensible, and that it does not rest on a category mistake. The speech-act theoretic defense of the Dworkin-MacKinnon position has dominated much of AngloAmerican philosophizing about pornography over the past thirty years, and the literature discussing it is by now extensive. In this chapter, I will examine and evaluate the philosophical legacy of the speech-act defense. There are two important caveats worth immediately noting. First, the chapter deals only with p. 247

debates that take

pornographic speech to be oppressive. It does not consider so-called pro-pornography

views or views that take pornography to have emancipatory force. This is for the simple reason that those views do not treat pornography as speech in the manner that anti-pornography feminism typically does (for a discussion of pornography as liberation, see my 2019b, chap. 7). Second, this chapter only discusses pornographic speech in the sense relevant for speech-act theory. Philosophers have also debated whether pornographic speech counts as speech in a legal sense, which is a separate issue. After all, not all speech in the speech-act sense counts as speech in a legal sense—that is, as speech that deserves protection (for more on the legal debates, see my 2019b, chap. 4).

2. Speech-Act Approach to Pornography The basic premise of Langton’s argument, supported by US courts, is that pornography is speech. As speech or expression, it depicts the subordination of women. However, pornography does more, according to MacKinnon and Dworkin: it causes women’s subordination. Even critics of MacKinnon and Dworkin accept this much. However, the constitutive claim that pornographic speech is a form of subordination has been said to be philosophically untenable, highly indefensible, and deeply incoherent. By contrast, Langton argues, with the help of Austin’s speech-act theory, that the claim is perfectly coherent and philosophically 2

cogent. We perform a whole range of actions with our words (e.g., warning, promising, refusing). Austin identi es performatives as a type of speech act, and he divides such acts into locutions, perlocutions, and illocutions. When we utter a sentence that has a particular traditionally conceived meaning, we have performed a locutionary act. But with our words, we can do more: by uttering something our words can cause extralinguistic perlocutionary e ects. And in uttering something our words (or locutions) can count as actions of some sort—we have performed some illocutionary action, where our action is constituted by the words intentionally uttered in their usual meaning. For instance, take the example of warning. Imagine that I notice a re break out in a bakery that I am passing by. I utter the right words intending to warn the customers (“Danger, re!”), who understand my words and my intention to warn them. Warning the customers was my intended illocutionary action, which my locution successfully pulled o . Now, I also intend to bring about some extralinguistic e ects: that the customers leave the shop for safety. Provided that they do, I will also have succeeded in bringing about my intended perlocutionary e ects. Furthermore, speech in the strict sense is not necessary for performing such actions. Imagine that due to an illness I have lost my voice. Instead of shouting “Danger, re!”, I write those words on a sign and run into the shop. This act of expression (if successful) will have performed the same action as my uttering the words. p. 248

According to Langton, critics of MacKinnon and Dworkin have failed to grasp pornography’s illocutionary force. This is what the constitutive subordination and silencing claims home in on. Illocutionary force hinges on satisfying certain felicity conditions. First, whether the speaker’s intentions are satis ed; after all, illocutionary speech acts are not just idle noises but intentionally performed actions. Second, whether the

speaker achieves uptake: that the hearer recognizes the particular intended illocution being performed. Third, that the speaker is authoritative relative to the intended illocution’s domain. (Not all successful illocutions require satisfaction of these conditions, but the ones relevant to pornography do.) Consider again the earlier example of warning. For my illocutionary act to succeed, I must intend to warn (rather than merely make use of my vocal cords for no apparent reason), the audience must recognize my intention to warn (rather than think I am joking), and I need to be somehow authoritative to do so (something about me must not prevent me from being taken seriously). Consider the well-known example of being an actor on stage and seeing a re break out behind the audience. I shout “Danger, re!” intending to warn, but the audience takes this to be part of the play. In this case, there is a failure of uptake since the audience does not recognize my intended illocution to warn and I lack the right kind of authority—the audience does not take my utterance seriously since they think my utterance is part of the play (i.e., they think that I am merely acting). I have not successfully performed the illocutionary act of warning; I have merely attempted to warn. How, then, can illocutionary speech acts subordinate? Langton considers a legislator in Pretoria during the apartheid who utters, “Blacks are not permitted to vote,” in a context of enacting a piece of legislation. This utterance does not merely report some prior state of a airs. Nor does it merely have the result that Black South Africans won’t show up at polling stations on election days. Rather, the utterance enacts a piece of legislation that is subordinating in making it the case that Black South Africans are deprived of voting rights (Langton 1993, 302). The speech act subordinates in virtue of (1) unfairly ranking Black people as inferior; (2) legitimating unjust racial discrimination; and (3) illegitimately depriving Black South Africans important powers and rights (303). Some of these subordinating illocutionary speech acts (like ranking and valuing) are, as Austin put it, verdictive in giving a verdict about something. Others (ordering, permitting, prohibiting, authorizing, and enacting laws) are exercitive speech acts: they confer or deprive powers and rights (304). In virtue of what, then, might pornographic speech subordinate in a parallel manner? The mere depiction of subordination does not su

ce; lms and reports can depict women’s subordination

without counting as such. Rather, Langton holds that pornography is (1) verdictive speech that ranks women as inferior sex objects whose purpose is to “service” men’s sexual needs; and (2) it is exercitive speech that legitimates sexualized violence against women in depicting women’s degradation in a manner that endorses, celebrates, and authorizes such degradation (307–308). Since verdictive and exercitive speech are authoritative acts, pornographic speech must be able to authoritatively rank women as inferior and legitimate sexualized violence against them—and Langton holds that it does. Contra liberal critics of antip. 249

pornography

feminism, pornographers are not a “powerless minority, a fringe group especially

vulnerable to moralistic persecution” (311). Instead, they are the ruling power when it comes to the domain of sex. What grants pornographic speech illocutionary force is its authoritativeness in the realm of sex. This might suggest that harmful pornographic speech should be countered by better speech that undercuts pornography’s false depictions of and lies about women’s sexuality, and undermines its authoritativeness. However, if pornographic speech has the power to silence women, such counterspeech looks set to fail. That is, pornography is said to be exercitive, not only in legitimating sexualized violence, but also in depriving women of important free speech rights—in silencing them. According to MacKinnon (1987, 1993), in creating a worldview where violence is eroticized, it becomes part of “ordinary” sex. This creates a climate where women are viewed as non-credible: when women try to speak out against sexualized violence or report assaults, they are often disbelieved, ridiculed, dismissed, and blamed for those assaults. This ts with the idea that pornography causes women’s silencing: thinking that one will not be believed, in fact, causes victims of sexual assaults to stay silent. But Langton has alone (1993) and together with Hornsby (1998) made a stronger claim. Pornography also constitutes women’s silencing; hence, there is a free-speech argument against pornography. The manufacture and consumption of pornographic speech may interfere with women’s ability to perform certain actions with their speech, where this amounts to a peculiar kind of silencing. Langton argues that there are three types of silencing. Locutionary silencing takes place when no

locutions are made (for instance, when I am gagged). Perlocutionary frustration takes place when the speaker voices the appropriate words, but fails to secure their intended extralinguistic e ects. Illocutionary disablement takes place when the speaker utters the appropriate locution, but fails thereby to perform the intended illocutionary act. For instance, imagine a judge passing a sentence, who utters the right words in the right context (e.g., “I nd the defendant guilty” uttered in a court of law). The judge intends to perform the illocutionary act of sentencing, but imagine that, unbeknownst to the judge, they have been disbarred during the proceedings and have lost the power to enact sentencing. In this case, the intended illocution fails or (as Austin put it) “mis res” because a requisite success condition of authority relative to a domain has not been met. This failure on the illocutionary level seemingly results in perlocutionary frustration too. If the judge has lost their authority to sentence people, their words will not secure the desired extralinguistic e ects: the defendant won’t be imprisoned or detained as a result of the judge’s utterance, because the judge no longer has the power to enact sentencing. The mis ring of the judge’s intended illocution hinges on the lack of authority. But illocutions may mis re for other reasons too. Take again the case of being an actor on stage and seeing a re break out. Here we have an uptake failure since the audience fails to recognize my intended illocution to warn. Note that this di ers from the ‘boy who cried wolf’–type cases, where the audience takes up the intended illocution to warn, but believes the warning to be insincere. Taking sincere illocutions as insincere ones represents a di erent kind of failure for Austin: they are abuses, not mis res. p. 250

With this in mind, how might pornographic speech silence women? Having speaker authority and securing uptake are central for illocutionary success; and (the thought goes) pornographic speech may block the satisfaction of these felicity conditions. Pornography does not silence women in the locutionary sense. But it may prevent women from performing some intended illocutionary actions with their speech. In particular, this may happen in sexual contexts, where pornographic speech is authoritative. And so, it may silence women’s refusals of unwanted sex. First, pornography may shape men’s conceptions of women’s sexuality in ways that result in perlocutionary frustration of refusal. When faced with unwanted sex, the pernicious e ects of pornographic speech may result in a woman’s locution ‘No!’ as being recognized as a refusal, but simply ignored. Women’s speech fails to have the desired extra-linguistic e ects because perpetrators see no reason to respect women’s own wishes in the sexual domain. Second, pornographic speech may shape men’s conceptions of women’s sexuality by creating communicative conditions that result in the illocutionary disablement of refusals. In this case, pornographic speech prevents women from successfully performing the illocutionary speech act of refusing: if pornographic speech prevents the locution ‘No!’ from being taken to be a refusal in a sexual context, due to which sex is forced on the speaker, she has not successfully performed the illocution of refusing the unwanted sex. This sort of uptake failure is exempli ed (though not exclusively) in cases of the kind: “I thought she was being coy and meant yes when she said ‘no’ ”. Women’s speech is illocutionarily disabled when certain locutions with which they intend to perform actions like refusing sex mis re: when women’s communicative e orts are interfered with by pornographic speech. This should not be merely incidental and idiosyncratic, but a symptom of some systematic failure. Although the systematicity condition has been critiqued as underdeveloped (McGowan et al. 2016), we can tentatively characterize it as follows: an instance of uptake-failure meets the condition “if it is nonidiosyncratic, it is brought about by systematic features of society, and others would make a similar mistake under similar circumstances” (McGowan et al. 2016, 76). To specify this further, consider Hornsby’s view that illocutionary speech acts are communicative. For her, a condition of linguistic communication is reciprocity (Hornsby 1995a, 1995b). This obtains when interlocutors “recognize one another’s speech as it is meant to be taken,” which ensures that the attempt to perform some speech act has been successful (1995b, 224). In this sense, communication is reciprocal and turns on the speaker being able to do with their words what they intend to and on the hearer recognizing the speaker’s intention. We can now better understand what happens when illocutionary silencing happens: the

promulgation of certain false views about women (and their sexuality) interferes with reciprocity, which is a precondition of communication. Such false views can render women relatively powerless in communicative exchanges. Their speech is not subsequently made ine able (there is no locutionary silencing). Nevertheless, women’s speech is in another sense inaudible: their ability to communicate has been hampered (Hornsby 1995a). Illocutionary disablement crucially involves such communicative interference.

p. 251

3. Tenability of the Speech Act Approach Although it has been extremely in uential, not everyone agrees with Langton’s analysis of what pornographic speech does. A signi cant question is whether pornography is speech, or speech of the kind that the speech-act approach assumes. Louise Antony argues that a good deal of pornography is not speech at all (although, ultimately, Antony thinks that the speech-act approach fails on di erent grounds). Speech-act theory is meant to illuminate verbal communication, but extending this model “metaphorically to cover actions that do not involve speech” is a stretch too far (Antony 2011, 389). Antony’s point is that much of pornography is pictorial and involves no words. But according to Austin, for some act of expression to be an illocutionary act, it must be a locutionary act: One cannot illocute without locuting, and pornographic pictorial representations do not locute. Whatever the courts have declared, by Austin’s own lights, much of pornography does not involve exercitive or verdictive utterances that could count as illocutionary silencing. At most, we can say that the ‘message’ of pornography might implicitly be exercitive 3

or verdictive, but this (the argument goes) ill ts Austin’s original account (Antony 2011, 390).

Elsewhere Antony (2017) argues that taking some pornographic speech to be exercitive/verdictive is in tension with another oft-made anti-pornography claim: that pornography tells lies about women. In telling lies about women, pornography is putatively reporting or describing some supposed way women are: pornographic speech misrepresents women’s ‘nature’ as being inferior, of lower value, and as suited to service men’s sexual needs. But anti-pornography feminism also holds that pornography makes this the case in ranking women as inferior and in legitimating violence against women. However, Antony argues, some use of language cannot simultaneously both falsely report something about women (describe some state of a airs) and make women t the description (bring about that state of a airs). For instance, as the chairperson of a jury, my utterance “We nd the defendant guilty” may either report or bring about a state of a airs. This depends on the context: if I utter it in the context of passing judgment in a court of law, it has illocutionary (performative) force. If I utter this to a friend post-trial, the utterance merely reports (describes) a state of a airs. But I cannot simultaneously do both in the same context. Or to put this di erently: for pornography to tell lies about women, there must be some prior way that women are (e.g., women do not enjoy violent sex). But the supposed power that pornography has to shape sexuality con icts with this. If it shapes both male and female sexuality, pornography no longer tells lies about women: it has constructed a reality where women t the pornographic image. Pornography’s reporting and constructing of women’s sexuality are in tension with one another, and one of them must go. (McGowan [2005] and Mikkola [2010] discuss this tension in MacKinnon; Jenkins [2017] o ers an alternative formulation of this claim.) p. 252

Perhaps surprisingly, Hornsby (2014) also argues that pornography is not speech, although we can use Austin’s speech-act theory to understand what pornography does. I say “surprisingly” because Hornsby together with Langton defends the silencing claim against critics. Hornsby’s objection hinges on her general view of language, brie y introduced earlier: the central function and point of language is communication. This is not equivalent to producing some (even meaningful) noises; rather, communication involves a “meeting of minds,” where the hearer recognizes what the speaker is intending to say and to do with their words. This is the main thrust of legal free-speech principles: to make sure that we can engage in meaningful communication with one another to advance democratic deliberation and knowledge-seeking

activities. However, according to Hornsby, pornography is not communicative in this sense. First, we cannot isolate some message that is being communicated on some occasion; second, even if we could isolate the message, we cannot identify whether it is the one the producer intended to convey; third, we cannot discern whether the consumers become aware of the intended and conveyed message (Hornsby 2014, 132). Pornography consumption alters its consumers’ attitudes and beliefs, and it conceivably has e ects like disabling women’s speech, where the mechanism for achieving this is some sort of mental conditioning (Hornsby 2014, 132). But pornography does not constitute communicative illocutionary actions: it does not feature in reciprocal “meetings of minds.” Since communication is the hallmark of speech, we should not think of (non-communicative) pornography as speech. Others have critiqued Hornsby and Langton (1998) in a di erent sense. For them, when pornography prevents women’s locution ‘No!’ from securing the required uptake, the locution fails to satisfy the required success conditions for counting as a refusal. This constitutes silencing of women’s refusals in the 4

illocutionary disablement sense. However, some reject that uptake is necessary for illocutionary success. Alexander Bird, for instance, argues that uptake is unnecessary for successful illocutions in general and for successful illocutionary refusals in particular, where ‘uptake’ means “the appreciation by an audience of the 5

intended illocution of the speaker” (Bird 2002, 2). Rather, for Bird, successful illocutions depend on “the words [locuted], their normal meaning and the context alone” (2002, 13). If this is right, securing uptake is neither here nor there. (Daniel Jacobson [1995] makes a similar claim; see Hornsby and Langton [1998] for a response.) Consider Bird’s argument in more detail. He distinguishes institutional illocutions from non-institutional ones, and argues that the former require no uptake for success. For institutional illocutionary success, we simply require a locution (verbal or written) in its usual meaning in the relevant context to be executed in conformity with some established conventions. The same is allegedly true of non-institutional illocutions and even of those that are “intimately bound up with communication” (Bird 2002, 8). Bird considers Jacques, the conceited chef, who falsely views himself as being world-class and thinks that his customers cannot get enough of his cooking. The in ated and warped view of Jacques’s cooking skills is the result of undue praise from overenthusiastic restaurant critics, friends, and family. This has had the pernicious e ect p. 253

that when

Jacques o ers his customers more food and they refuse saying ‘No, thank you’, Jacques thinks

that they want more. He simply thinks that his customers are being coy about accepting more. So, “when Sara says ‘No’ intending to decline an o er of food, there is not even a icker of uptake in Jacques. On the contrary, he takes this as a reason to give her yet more” (Bird 2002, 11). Sara intends to refuse, but Jacques fails to appreciate her intended illocution. On the Hornsby-Langton view, Jacques’s failure to take up Sara’s intended refusal means that Sara has not refused: she has merely tried to. By contrast, Bird claims that this “is not, I believe, how most people would read [the] case. She refused all right—only Jacques’ arrogance and boorishness … prevented him from seeing that she was refusing” (2002, 11). Most people’s intuitions (according to Bird) count against the view that uptake is necessary for successful refusals. I do not share this intuition though. As I see it, Bird is mistaken about the role of uptake since he does not su

ciently appreciate the communicative nature of refusals. Of course, successful illocution and

communication are not equivalent. However, illocutionary refusals are not just intimately connected to communication; they are essentially communicative. Hornsby is right (I contend) that a hallmark of communication is reciprocity: when it fails, interlocutors are talking at cross-purposes. This is precisely what happens when illocutionary refusals mis re, which (as Hornsby and Langton originally argued) demonstrates that such mis rings involve communicative failures. When one is faced with unwanted sex (or food) and utters ‘No!’, the usual way to read this situation is to take one’s locution as intending to communicate something to someone else. The speaker aims to express herself to another with the view to producing some e ects: that the hearer grasps the locution as a refusal and stops. The utterance does not simply express the speaker’s disapproving opinion about the sexual advances. Rather, the locution is the

verbal equivalent to kicking and punching one’s adversary. As I argue elsewhere in more length (see Mikkola 2011), the locution ‘No!’ has a dual aim: to refuse sex and to communicate this to the hearer. Insofar as this is the case, uptake is necessary for illocutionary success at least of refusals. McGowan (2009a) identi es a di erent aw in Bird’s case: he takes Hornsby and Langton to claim that refusals are purely communicative, when they are merely highlighting the importance of communicating sexual refusals. If all that matters for illocutionary success is communication, McGowan holds, we get the wrong results. She o ers the following example. Imagine that Cindy successfully communicates to Carl that Sally is unwilling to have sex with him. On the purely communicative view of illocutions, Cindy would have refused sex on Sally’s behalf. However, Cindy’s communication of the refusal would not count as Sally refusing sex with Carl because only Sally can issue such a refusal. If the silencing claim is committed to a purely communicative view, it looks prima facie implausible. But (McGowan holds and rightly so) this is not Hornsby and Langton’s view. In taking refusing to be an excercitive speech act, it is a sort of communication-plus speech act, where ‘plus’ highlights the relevance of speaker authority. That is, uptake is crucial for the communication-part and authority for the ‘plus’-part of the speech act. And the two can work in concert: sometimes, what blocks uptake is precisely that one is not considered to be authoritative as p. 254

a speaker. Just think

back to the case of the actor trying to warn an audience of a re. Precisely because

one’s role as an actor prevents the audience from recognizing the intended illocutionary intention to warn, the locution does not count as a warning. In this case, there is a failure of uptake that is intertwined with speaker authority. All of these aspects are lost in Bird’s view that successful illocutions depend on “the 6

words [locuted], their normal meaning and the context alone” (2002, 13).

The silencing claim also allegedly diminishes rapists’ responsibility, which supposedly undermines the idea of pornography being oppressive speech. Although the speci cs of sexual assault legislation di er for di erent jurisdictions, a rape conviction must typically satisfy two jurisprudential conditions: the actus reus (the guilty act) and the mens rea (the guilty mind) requirements. With these requirements in mind, some jurisdictions hold that if a man engages in sex with a woman honestly believing that she consented, however unreasonable this belief may be, he does not do so with ‘a guilty mind’. Bird maintains that if the silencing claim holds, alleged perpetrators are less culpable or even exonerated in that they won’t have committed rape. Daniel Jacobson also takes the phenomenon of illocutionary disablement to have the “strange and troubling consequence” that if women fail to illocute refusals and sex is forced on them, we “cannot call this rape” (1995, 77; italics original). This supposedly renders the Hornsby-Langton view practically indefensible. However, this line of argument is deeply awed. The silencing claim is in trouble only if the lack of refusal entails consent. After all, the determination of rape does not turn on whether the victim refused, but on whether they consented to sex. For example, in English law, if the victim did not consent, the actus reus of rape is satis ed—whether the victim refused is neither here nor there. So, we must not con ate consent with non-refusal: a failure to illocute refusal does not entail that one has illocuted consent to sex instead. Granted, neither Bird nor Jacobson explicitly state that consent is entailed by the absence of refusal. But the “if no refusal, then no rape” argument only gets o

the ground and seems plausible if we

accept this implicit and false extra premise. Nellie Wieland (2007) has put forward a variant of this argument. The silencing claim takes pornographers as authoritative convention-setters in sexual discourse: women face illocutionary silencing because their ‘No!’ is taken to mean yes. However, accepting this (according to Wieland) diminishes perpetrators’ culpability because they cannot satisfy the mens rea requirement: [I]f there is a convention of women meaning yes by uttering “no” in relevant [sexual] contexts, then it is no longer the case that men misinterpret women as meaning yes by uttering “no” in contexts of unwanted sexual encounters; but rather, men correctly interpret “no” as meaning yes

in these contexts given the establishment of the convention. The result is that these speakers have no words to express their refusal of an unwanted sexual encounter. (Wieland 2007, 445) If pornographic speech is responsible for such a meaning switch, so that women’s attempted refusals are honestly taken as granting consent, the mens rea requirement will not be ful lled. If the convention set by p. 255

pornography (that ‘no’ means yes) is stable and

not easily countered, Wieland continues, it makes

women illocutionarily silenced and perpetrators interpretatively disabled. This leaves us with no “reason for thinking that a rapist should be able to understand a victim’s meaning that p by uttering ‘x’ ” (Wieland 2007, 452), which allegedly excuses rape further. Wieland is right to claim that if Hornsby and Langton are correct about women’s illocutionary disablement, this ill ts legislation that requires mens rea for rape conviction. Wieland even accepts that women have been silenced, if the perpetrator can reasonably claim to having obtained consent. But, she holds, “this is crazy and has got to be wrong” (2007, 453). I agree; but Wieland’s argument does not speak against the silencing claim or show that it is wrong. Rather, I contend, a law that allows a defense along the lines of “I honestly thought she meant yes when she said ‘no’ ” is wrong. In fact, Wieland’s argument seemingly supports the silencing claim: if interpretative disablement is widespread and at least partly due to pornographic speech, the e ects of such speech are even more pernicious than anti-pornography feminists initially realized. It demonstrates the need for sexual communication education to eradicate illocutionary silencing and interpretative disablement. That said, although Wieland’s analysis of interpretative disablement has pre-theoretical appeal, it is probably due to various factors rather than just pornographic 7

speech. The link to pornographic speech remains to be empirically established.

4. Alternative Accounts of Silencing The main thrust of the Hornsby-Langton account is that silencing in the illocutionary disablement sense involves communicative interference: the hearer fails to recognize the speaker’s illocutionary intention (what sort of a speech act the speaker aims to perform), where this is brought about in a systematic manner by pornographic speech. This is a novel sense of silencing in addition to the more familiar case of perlocutionary frustration, where a woman’s ‘No!’ is recognized as a refusal but ignored. Mary Kate McGowan has identi ed yet further ways in which women’s refusals might be silenced. In her (2009a), McGowan develops (what I am calling) authority-failure silencing, which is constituted by the hearer’s failure to grant the speaker authority over some relevant domain. What goes wrong in failed sexual refusals is that the speaker fails to exercise her authority over the domain of sexual access to her body. McGowan (2014) argues for sincerity silencing. Pornographic speech may shape men’s conceptions of women’s sexuality in ways that, when faced with unwanted sex, women’s locution ‘No!’ is taken as an insincere refusal. This is a variant of the “I thought she was being coy and meant yes when she said ‘no’ ” case. To clarify: in the illocutionary disablement sense of silencing, the perpetrator does not take a woman’s ‘No!’ as a refusal at all. In the sincerity silencing case, they take ‘No!’ as a refusal but think that the woman is not sincerely refusing. In the former case, we have a failure of the speaker’s illocutionary intention condition; in the latter, of the speaker’s sincerity condition. McGowan (2017) explicates a still further type of silencing: true feelings-recognition failure. This is a case where p. 256

a woman says ‘No,’ sincerely intending to refuse sex, but although the addressee recognizes her sincere refusal, the addressee nevertheless falsely believes that refusing is not what the speaker’s “deep self” really wants. Suppose, for example, that the addressee believes that the woman will change her mind as soon as she realizes how amazing he is.

(McGowan 2017, 49) Such a case, McGowan holds, is distinct in kind from the other types of silencing in that the illocutionary intention, authority, and sincerity conditions have been ful lled. These types of silencing raise a question though: Is pornographic speech responsible for any of them? McGowan (2017) suggests that pornographic speech indeed causes the requisite types of silencing: it conceivably causes its consumers to have false beliefs, and it sexually conditions them in ways that hinder the recognition of the relevant conditions (intention, authority, sincerity, and true feelings conditions). McGowan further considers pornography’s constitutive role in oppressive silencing. She holds that pornography enacts norms that silence. For instance, relative to true-feelings- recognition silencing pornography enacts norms that comply with certain rape myths (e.g., women really want sex even when they think otherwise). Deep down women want to be ‘taken by force,’ but they have simply been socialized to hide this. Abiding by norms like this may then “systematically prevent the recognition of the speaker’s true feelings of refusal” (McGowan 2017, 56). Now, these ways of bringing about silencing may certainly be to some extent grounded in pornography consumption; but as of yet, the empirical evidence does not back up the claim that pornography in particular causes and constitutes these types of silencing. Plenty of gendered norms and role expectations conceivably interfere with the satisfaction of the above conditions, and these types of silencing were commonplace well before the onset of contemporary pornography production and consumption on mass scale. In fact, a central case of silencing that McGowan (and Maitra) make use of comes from Jane Austen’s Pride and Prejudice, which in no way connects to pornographic 8

speech. A more convincing case, then, must be made of pornography’s role in women’s silencing.

5. Pornographyʼs Authority The foregoing does not yet establish that pornographic speech in particular undergirds illocutionary disablement. For instance, gender norms, expectations, romance novels, and the media generally also enact communicative conditions that seemingly interfere with women’s ability to do things with their words and men’s ability to understand women’s utterances. In order to show pornography’s role in (re)creating such communicative conditions, we need to establish that pornography is somehow—as Langton puts it— authoritative over the domain of sex. However, the idea that pornographic speech and pornographers are in some requisite sense authoritative has generated much resistance. I will discuss next some important contributions to the debate. p. 257

Langton takes there to be empirical evidence for pornography’s authoritativeness insofar as young people (and young men in particular) are increasingly treating pornography as sex education. In her (2017) Langton discusses the 2013 report from the UK O

ce of the Children’s Commissioner (Maddy et al. 2013).

This large-scale report found that many respondents take pornography to educate them about sex and that such education is in many ways problematic. Consumption of pornography was linked to heterosexual boys’ unrealistic expectations about sex, with corresponding feelings in girls that they must submit to boys’ expectations irrespective of their own wishes. (For similar results, see Paul 2006.) The report o ers compelling evidence for the claim that many young people consider pornographic materials to o er 9

authoritative guidance on sex.

However, Antony holds that this sense of authority is not that of Austin’s. ‘Authority’ as Langton’s employs it can only be used to determine whether pornographic speech as a pernicious type of sex education has perlocutionary e ects (Antony 2011, 396). But this is not the sense of authority Austin had in mind when thinking about illocutionary speech acts; rather, his sense was about linguistic competence and about authority as a competent language user. Judith Butler (1997) also argues that the appeal to authority fails, irrespective of any empirical issues and whether pornographic speech has real-world e ects on people. Her

view pertains to the internal workings of language: there is a ‘gap’ between speech (or speakers’ intentions) and its e ects on recipients, because speech can always be “resigni ed.” Meanings are not stable in the sense presumed by advocates of the speech-act theory because the conventions governing speech are always open to subversion. One of the most celebrated examples is that of ‘queer’: it has undergone a resigni cation from a pejorative term to proud self-identi cation. The act of appropriating the term subverted the earlier problematic meaning. And so, bluntly put, whatever meaning-intentions pornographers might have, they lack the authority to enforce them because audiences can always engage in reappropriation or subversion of pornographic meanings. For instance, feminist audiences can ‘turn the tables’ and ‘read’ pornographic materials in subversive ways, thus disrupting pornography’s pernicious meanings. The sheer possibility of subverting pornographic meanings shows it to lack the requisite sort of authority presupposed by the speech-act approach: if it were to have it, its authority would be “divine”. Langton (2009, chap. 5) denies that her approach assumes a divine conception of authority, where speakers’ intentions always transpire in their intended e ects. (See also Schwartzman [2002] for a critique of Butler.) She accepts that feminists can o er subversive readings of pornography; but she holds that this does not undermine her analysis because silencing is a context-dependent phenomenon. So even though in some contexts subversively interpreted pornographic materials do not silence, in other contexts they may well do. Furthermore, Langton takes Butler’s hopes for deconstructive and subversive readings of pornography to be hyperbolic: It is not easy to see how such readings are supposed to reveal to pornography’s usual consumers, and their partners, the “unrealizability,” the self-defeatingness, of pornography’s norms. How p. 258

does a reading of pornography “against itself” help

those women who are abused and exploited

in its making? How does it a ect those men who want their sex lives, and their partners, to resemble ever more closely what pornography o ers them? (Langton 2009, 115) For Langton, pornography’s authority is greater than the force of deconstruction. An argument to the best explanation demonstrates this: perlocutionary e ects on our normative beliefs about sex and our subsequent actions are better explained by assuming that pornographic speech has authority contextdependently than by assuming that “pornography has no authority” (Langton 2009, 111). Even though I share Langton’s skepticism about the e

cacy of deconstruction, her critique of Butler does

not entirely succeed. Langton draws an opposition between pornography having no authority and pornography having authority (in a context), and proceeds to reject the former view. However, this way of demarcating the disagreement isn’t entirely right. As I see it, Butler does not hold that pornography has no force; she simply denies that this force is somehow unique to pornography. Or to put the point di erently: Butler does not deny that pornography has power to (re)produce meanings. Antony helpfully elucidates the nature of pornography’s power to be “a matter of being able to arrange contingencies, to make things be the way you want” (2017, 79). Pornographic meanings cannot be (re)produced without discursive force and power. But Butler’s point is that pornography is not alone in having such power: even those in oppressed social positions have power—otherwise resistance would be impossible. Still, pornography does not have authority in the sense of having some special purview over the (re)production of meanings, because feminist readings of pornography can subvert pornographic meanings. In her response to Butler, however, Langton fails to note this subtle distinction between power and authority, and the two end up talking past each other. One might critique Langton for confusing authority with power in other ways too (Antony 2014, 2017; Drabek 2016). Being authoritative as a speaker to enact women’s subordination and silencing can only take place within some established and institutionalized conventions; but since in pornography such conventions are missing, Langton’s approach is undermined. Antony (2014) elucidates this with a

distinction between conventional and de facto subordination. Conventional subordination requires authority relative to some domain; for instance, being a legislator in Pretoria during the apartheid came with authority in the requisite sense to enact laws that deprived Black South Africans of voting rights. This sort of authority is necessary for performing exercitive speech acts and it is institutionally governed. An ordinary White South African could have made the same locutions, but they would not thereby have enacted discriminatory laws. Still, ordinary White South Africans had power over Black South Africans by virtue of the social and political structures that were in place. They were powerful in being able e ectively to treat Black South Africans as inferior, despite lacking the necessary conventionally governed authority over some relevant domain(s). This suggests that we should shift our focus: Langton takes pornographers to have p. 259

authority over the domain of sex in o ering pernicious sex education. But since there are

no clear

conventions framing this sort of authority in the required sense, we must look to pornography’s audiences to see what is really doing the work. This is where power comes in, in shaping the beliefs and attitudes of men is not the full extent of the power of the pornographers. Insofar as the members of pornography’s audience are powerful—insofar as pornography exists within a patriarchal society—the consumers of pornography have the ability to shape reality itself. Expecting certain behavior from women, they can coerce women into displaying it. (Antony 2014, 163) In short, pornography’s authority is rather limited, and it comes apart from the kind of authority typically required to make verdictive and exercitive speech acts. Instead, the background structures that give consumers power to shape social reality in certain ways is of central importance. For Antony, focusing on this part of the story nevertheless in no sense hinges on speech-act theory. A similar thought seemingly undergirds Nancy Bauer’s (2015) critique of Langton. Bauer asks: Why is it that pornography apparently makes people see the world in a certain way and thus has perlocutionary e ects in the world? Why is it that consumers “acquiesce to the pornographer’s point of view” (Bauer 2015, 80)? These e ects for Bauer cannot be explained in terms of pornographers’ authority. Instead, a more fruitful feminist analysis requires that we refocus our attentions away from speakers and their illocutionary intentions toward hearers and why certain perlocutionary e ects are at all possible. Bluntly put: problematic perlocutionary e ects are not due to producers’ authority, but due to consumers granting pornographic speech particular powers to shape actual social relations. Leslie Green’s (1998) critique echoes this thought as well. Pornography has power, not by virtue of what it says, but rather as a function of “the 10

whole social context in which it occurs” (Green 1998, 294).

Langton (2009, chap. 4) has responded to Green in a manner that (I hold) also responds to Antony and Bauer. At times, Langton does run power and authority together in a manner that suggests she is con ating the two. However, looking at her response to Green dispels this concern. Langton draws the following analogy: imagine a powerful religious organization has authority over a small, but socially powerful minority. Now imagine that this organization condemns homosexuality and has the authority to convince its followers that homosexuality is a sin. Even though the church in this case does not have the authority to hire and re those with di ering views, the socially powerful members of the society convinced by the church’s message can enact such subordinating practices. That is, they have the power to make the church’s teachings ‘real’ in the world. In this sense, even those who reject the church’s authority are still indirectly under its jurisdiction. So we can see a sort of two-tier way in which subordination is enacted: rst from the church to the ruling group via the former’s authority; then from the group to the wider society via the exercise of social power. Power and authority work in concert. This tells us something important about Langton’s speech-act approach: even if pornography’s authority is limited and consumers’ power is more p. 260

in uential, it does not

su

ce to focus on the latter alone. A deeper social explanation of pornography’s

e ects in the world should also bear in mind the sort of authority that pornographic speech has over sex.

Irrespective of whether pornographers wish to be authoritative about sex, they are so for many—especially for young men. And even if pornographers’ authority derives from less conventionally regulated and more di use social arrangements, this does not undermine Langton’s claim that pornographic speech has authority over the domain of sex. We must simply remember that producers’ authority is intertwined with consumers’ (social) power. One might argue in Langton’s defense that there is still a further sense of authority that does not fall prey to some of the above critiques. McGowan (2003; see also her 2019) argues that the sort of authority required to make exercitive speech acts need not be governed by institutionalized conventions, such as in the case of enacting laws. She draws a distinction between Austinian and conversational exercitives. Although the exercise of the former requires some conventionally and institutionally governed background conditions to be in place, the latter does not. Ordinary conversations are governed by loose conventions and ordinary speaker authority xes what is permissible within such conversations. Moreover, this involves exercitive speech acts, albeit not obviously. For instance, if I say to someone, “My neighbor’s cat Lucky keeps sneaking into my house unnoticed,” I have made Lucky the most salient cat in this conversation. This (according to McGowan) is a conversational exercitive and I have xed what is permissible in the conversation to follow. If my interlocutor then replies, “That cat is crazy,” referring with ‘that cat’ to some cat other than Lucky, they have in a very ordinary sense done something conversationally impermissible. In a sense, my interlocutor has made an impermissible ‘move’ in our conversation. Understanding pornography with such conversational exercitives in mind avoids the problem of institutionalized authority because we need not hold that pornographic speech is governed by institutionalized conventions. There are perfectly ordinary ways in which we ‘authorize’ certain moves in everyday conversations; and so, the apparent lack of pornography’s institutionalized authority does not yet undermine the speech-act approach. Drawing on this idea, Langton herself has suggested that we should understand what pornography does on a pragmatic model, where its illocutionary force hinges on pornography altering normative facts about what is permissible and possible in a conversational context. Subsequently, we can see how speech can change beliefs and norms without requiring that the speaker satisfy strong Austinian felicity conditions for exercitives (Langton 2012, 84).

6. Methodological Remarks In this chapter, I have considered in uential arguments to make good the idea that pornographic speech is oppressive in silencing and subordinating women. Despite being extensively debated, it remains to be established that pornographic speech is responsible for women’s silencing. In fact, philosophical debates p. 261

about silencing often have

merely a tenuous connection to pornography. Moreover, discussion often

advances by appealing to intuitive ‘gut-feelings’ about speci c cases. This is especially so with respect to debates about the success conditions of refusals. Now, although I, too, have appealed to intuitions to justify particular views, ultimately, I do not nd this methodologically conducive to settling the matter. Much of the debate ends up trading intuitions, and it seems highly unlikely that we can achieve some consensus on which intuitions will be the ‘right’ ones given how idiosyncratic they tend to be and how greatly our background beliefs and commitments in uence the assessment of examples used. Methodologically, I would thus recommend against relying on intuitions as justi cations for ‘our’ views about subordinating and silencing speech: immediate ‘gut-feelings’ about particular cases may a ord valuable starting points for further inquiry, but do not a ord particularly reliable and helpful ways to settle disputes.

Notes

p. 262

1.

For more on these liberal critiques and an introduction to the relevant free speech debates, see Watson (2010) and West (2013).

2.

Note that she does not aim to argue that it is true, only that it is philosophically defensible. Still, Langton (1993) also takes the claim to be true if we grant “some not entirely implausible empirical assumptions” (300).

3.

Although for a view that there are covertly exercitive speech acts, see McGowan (2009b, 2019).

4.

This also helps to explain why Hornsby thinks that speech-act theory is still relevant when thinking about pornography: although pornography itself cannot be analyzed with such a theory, its e ects (like failures of refusals) can be.

5.

In what follows, I take the expression “oneʼs locution ʻpʼ performs the illocution of øing” as shorthand for “oneʼs locution ʻpʼ successfully performs the illocution of øing.” As I see it, unsuccessful performance of an illocution is just a failure to perform that illocution (just as if one unsuccessfully performs the act of voting, one simply has not voted).

6.

Whether pornographic speech has illocutionary force has also been debated relative to the matter of context: what context is relevant for fixing illocutionary force? In thinking about the relevant context for determining such force, a parallel is typically drawn between pornographic recordings and multipurpose signs involved in delayed communication (like a sign that reads ʻI doʼ that can be used to perform a variety of speech acts). If we conceive of pornographic speech as being akin to multipurpose signs, it has illocutionary force only context-specifically. Once we take this claim seriously, Jennifer Saul (2006) argues, Langton must settle for a restricted claim: only some viewings count as the illocutionary subordination of women because only some viewing contexts are subordinating ones. Hence, pornography as such cannot be said to have the illocutionary force to subordinate women. Not everyone agrees though. Claudia Bianchi (2008, 2014) argues that the relevant context is that of intended or expected viewings, which would enable Langton to make the stronger claim. I have argued against Bianchi (Mikkola 2008, 2019a).

7.

Others have raised doubts about whether the linguistic conventions Wieland identifies do obtain (although some discuss this point without directly referring to Wieland). For more, see Antony (2011); Grünberg (2014); Maitra and McGowan (2010).

8.

Furthermore, I am not entirely convinced that McGowan has o ered genuinely alternative types of silencing. I argue for this in my 2019b, chapter 3.

9.

The educational value of pornography is likely to di er significantly for heterosexual and non-heterosexual youth. Gender non-conforming youth and those who fall outside heteronormative sexualities more o en find pornographic depictions of non-heteronormative sexualities valuable educationally and in legitimating their identities (Albury 2014). Disability activists have also welcomed pornography with atypically functioning bodies in order to disrupt the stereotype of ʻdisabledʼ people as asexual. In these ways, pornography is also authoritative but in beneficial ways—something Langton does not discuss. (I consider these cases in more detail in my 2019b, chapter 7.)

10.

Linda LeMoncheck (1997) also holds that the reason pornographic lies about women (that women are ʻnaturalʼ sexual subordinates of men) have any purchase is that such lies are already part of gendered stereotypes and role expectations (130). This is where pornographyʼs power derives from.

References Albury, Kath. 2014. “Porn and Sex Education, Porn as Sex Education.” Porn Studies 1 (1–2): 172–181. doi: 10.1080/23268743.2013.863654. Google Scholar WorldCat Antony, Louise. 2011. “Against Langtonʼs Illocutionary Treatment of Pornography.” Jurisprudence 2: 387–401. Google Scholar WorldCat Antony, Louise. 2014. “Pornography and the Philosophy of Language.” In The Philosophy of Pornography: Contemporary Perspectives, edited by Lindsay Coleman and Jacob M. Held, 147–176. Lanham, MD: Rowman and Littlefield. Google Scholar Google Preview WorldCat COPAC Antony, Louise. 2017. “Be What I Say: Authority versus Power in Pornography.” In Beyond Speech: Pornography and Analytic Feminist Philosophy, edited by Mari Mikkola, 59–90. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Austin, J. L. 1962. How to Do Things with Words. Oxford: Clarendon Press. Google Scholar Google Preview WorldCat COPAC Bauer, Nancy. 2015. How to Do Things with Pornography. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Bianchi, Claudia. 2008. “Indexicals, Speech Acts and Pornography.” Analysis 68:310–316. Google Scholar WorldCat Bianchi, Claudia. 2014. “How to Do Things with (Recorded) Words.” Philosophical Studies 167:485–495. Google Scholar WorldCat Bird, Alexander. 2002. “Illocutionary Silencing.” Pacific Philosophical Quarterly 83:1–15. Google Scholar WorldCat Butler, Judith. 1997. Excitable Speech: A Politics of the Performative. New York and London: Routledge. Google Scholar Google Preview WorldCat COPAC Coy, Maddy, Liz Kelly, Fiona Elvines, Maria Garner, and Ava Kanyeredzi. 2013. “Sex without Consent, I Suppose That Is Rape”: How Young People in England Understand Sexual Consent. Report. O ice of the Childrenʼs Commissioner, London. Accessed September 5, 2023. https://assets.childrenscommissioner.gov.uk/wpuploads/2014/02/Sex_without_consent_I_suppose_that_is_rape.pdf. Google Scholar Google Preview WorldCat COPAC Drabek, Matt L. 2016. “Pornographic Subordination, Power, and Feminist Alternatives.” Feminist Philosophy Quarterly 2 (1): 2. Google Scholar WorldCat Dworkin, Andrea. 1981. Pornography: Men Possessing Women. London: Womenʼs Press. Google Scholar Google Preview WorldCat COPAC Dworkin, Ronald. 1991. “Liberty and Pornography.” New York Review of Books, April 14, 1991, 12–15. p. 263

Feinberg, Joel. 1985. O ense to Others: The Moral Limits of the Criminal Law. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Green, Leslie. 1998. “Pornographizing, Subordinating, and Silencing.” In Censorship and Silencing: Practices of Cultural Regulation, edited by Robert C. Post, 285–311. Los Angeles: Getty Research Institute for the History of Art and the Humanities.

Google Scholar

Google Preview

WorldCat

COPAC

Grünberg, Angela. 2014. “Saying and Doing: Speech Actions, Speech Acts and Related Events.” European Journal of Philosophy 22 (2): 173–199. Google Scholar WorldCat Hornsby, Jennifer. 1995a. “Disempowered Speech.” Philosophical Topics 23 (2): 127–147. Google Scholar WorldCat Hornsby, Jennifer. 1995b. “Speech Acts and Pornography.” In The Problem of Pornography, edited by Susan Dwyer, 220–232. Belmont, CA: Wadsworth. Google Scholar Google Preview WorldCat COPAC Hornsby, Jennifer. 2014. “Pornography and ʻSpeech.ʼ” In The Philosophy of Pornography: Contemporary Perspectives, edited by Lindsay Coleman and Jacob M. Held, 129–146. Lanham, MD: Rowman and Littlefield. Google Scholar Google Preview WorldCat COPAC Hornsby, Jennifer, and Rae Langton. 1998. “Free Speech and Illocution.” Legal Theory 4: 21–37. Google Scholar WorldCat Jacobson, Daniel. 1995. “Freedom of Speech Acts? A Response to Langton.” Philosophy and Public A airs 24:64–79. Google Scholar WorldCat Jenkins, Katharine. 2017. “What Are Women For: Pornography and Social Ontology.” In Beyond Speech: Pornography and Analytic Feminist Philosophy, edited by Mari Mikkola, 91–112. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Langton, Rae. 1993. “Speech Acts and Unspeakable Acts.” Philosophy and Public A airs 22:293–330. Google Scholar WorldCat Langton, Rae. 2009. Sexual Solipsism: Philosophical Essays on Pornography and Objectification. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Langton, Rae. 2012. “Beyond Belief: Pragmatics in Hate Speech and Pornography.” In Speech and Harm: Controversies over Free Speech, edited by Ishani Maitra and Mary Kate McGowan, 72–93. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Langton, Rae. 2017. “Is Pornography Like the Law?” In Beyond Speech: Pornography and Analytic Feminist Philosophy, edited by Mari Mikkola, 23–38. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC LeMoncheck, Linda. 1997. Loose Women, Lecherous Men: A Feminist Philosophy of Sex. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC MacKinnon, Catherine A. 1987. Feminism Unmodified. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC MacKinnon, Catherine A. 1989a. “Sexuality, Pornography, and Method: Pleasure under Patriarchy.” Ethics 99 (2): 314–346. Google Scholar WorldCat MacKinnon, Catherine A. 1989b. Toward a Feminist Theory of the State. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC MacKinnon, Catherine A. 1993. Only Words. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC

Maitra, Ishani, and Mary Kate McGowan. 2010. “On Silencing, Rape and Responsibility.” Australasian Journal of Philosophy 88 (1): 167–172. Google Scholar WorldCat McGowan, Mary Kate. 2003. “Conversational Exercitives and the Force of Pornography.” Philosophy and Public A airs 31 (2): 155– 189. Google Scholar WorldCat McGowan, Mary Kate. 2005. “On Pornography: MacKinnon, Speech Acts, and ʻFalseʼ Construction.” Hypatia 20:23–49. Google Scholar WorldCat McGowan, Mary Kate. 2009a. “On Silencing and Sexual Refusal.” Journal of Political Philosophy 17 (4): 487–494. Google Scholar WorldCat p. 264

McGowan, Mary Kate. 2009b. “Oppressive Speech.” Australasian Journal of Philosophy 87:389–407. Google Scholar WorldCat McGowan, Mary Kate. 2014. “Sincerity Silencing.” Hypatia 29: 458–473. Google Scholar WorldCat McGowan, Mary Kate. 2017. “On Multiple Types of Silencing.” In Beyond Speech: Pornography and Analytic Feminist Philosophy, edited by Mari Mikkola, 39–58. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC McGowan, Mary Kate. 2019. Just Words: On Speech and Hidden Harm. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC McGowan, Mary Kate, Ilana Walder-Biesanz, Morvareed Rezaian, and Chloe Emerson. 2016. “On Silencing and Systematicity: The Challenge of the Drowning Case.” Hypatia 31 (1): 74–90. Google Scholar WorldCat Mikkola, Mari. 2008. “Contexts and Pornography.” Analysis 68:316–320. Google Scholar WorldCat Mikkola, Mari. 2010. “Is Everything Relative? Anti-realism, Truth and Feminism.” In New Waves in Metaphysics, edited by Allan Hazlett, 179–198. Basingstoke, UK: Palgrave-Macmillan. WorldCat COPAC Google Scholar Google Preview Mikkola, Mari. 2011. “Illocution, Silencing and the Act of Refusal.” Pacific Philosophical Quarterly 92:415–435. Google Scholar WorldCat Mikkola, Mari. 2019a. “Fixing Pornographyʼs Illocutionary Force: Which Context Matters?” Philosophical Studies 177: 3013–3032. https://doi.org/10.1007/s11098-019-01357-2. Google Scholar WorldCat Mikkola, Mari. 2019b. Pornography: A Philosophical Introduction. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC C cc Parent, William. 1990. “A Second Look at Pornography and the Subordination of Women.” Journal of Philosophy 87:205– 211. Google Scholar WorldCat Paul, Pamela. 2006. Pornified: How Pornography Is Damaging Our Lives, Our Relationships, and Our Families. New York: Times Books. Google Scholar Google Preview WorldCat COPAC

Saul, Jennifer. 2006. “Pornography, Speech Acts and Context.” Proceedings of the Aristotelian Society 106:229–248. Google Scholar WorldCat Schwartzman, Lisa. 2002. “Hate Speech, Illocution, and Social Context: A Critique of Judith Butler.” Journal of Social Philosophy 33: 421–441. Google Scholar WorldCat Watson, Lori. 2010. “Pornography.” Philosophy Compass 5: 535–550. doi:10.1111/j.1747-9991. 2010.00292.x. Google Scholar WorldCat West, Caroline. 2013. “Pornography and Censorship.” The Stanford Encyclopedia of Philosophy (Fall 2013 Edition), edited by Edward N. Zalta. http://plato.stanford.edu/archives/fall2013/entries/pornography-censorship/. Google Scholar Google Preview WorldCat COPAC Wieland, Nellie. 2007. “Linguistic Authority and Convention in a Speech Act Analysis of Pornography.” Australasian Journal of Philosophy 85 (3): 435–456. Google Scholar WorldCat

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

12 Pronouns and Gender  Cameron Domenico Kirk-Giannini, Michael Glanzberg https://doi.org/10.1093/oxfordhb/9780192844118.013.35 Published: 22 May 2024

Pages 265–292

Abstract This chapter introduces readers to the empirical questions at issue in debates over gendered pronouns and assesses the plausibility of various possible answers to these questions. It has two parts. The rst is a general introduction to the linguistics and psychology of grammatical gender. The second focuses on the meanings of gendered pronouns in English. It begins with a discussion of some of the methodological limitations of empirical approaches to the topic and the normative implications of those limitations. It then argues against three simple theories of the semantics of gendered pronouns in English and proposes an alternative that fares better: the Gender-First View. Finally, it discusses the singular use of ‘they’ and its connection to nonbinary gender identities.

Keywords: pronouns, gender, sex, agreement, identity, semantics, psychology Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction In February 2022, social media app TikTok updated its community guidelines to include a provision prohibiting users from posting, uploading, streaming, or sharing “content that targets transgender or non1

binary individuals through misgendering or deadnaming.” In contrast—and just two months later— Tennessee’s House of Representatives passed a bill specifying that teachers and other employees of public schools are “not required to use a student’s preferred pronoun when referring to the student if the preferred 2

pronoun is not consistent with the student’s biological sex”; and Nicholas Meriwether, a professor at Shawnee State University in Ohio, won $400,000 in a settlement with his employer after refusing to use the 3

preferred pronouns of one of his students.

The present handbook appears during a period of historic controversy in the English-speaking world regarding the nature of gender and the use of gendered pronouns. And though the emergence of this controversy has been shaped by broader political and cultural trends (best analyzed by historians and political scientists), the arguments to which representatives of both sides appeal often turn on empirical questions about the meanings of gendered pronouns in English (best analyzed by philosophers and 4

linguists).

In keeping with the goals of a handbook of applied philosophy of language, our aim is to introduce readers to the empirical questions at issue in debates over gendered pronouns and to assess the plausibility of various possible answers to these questions. This project is primarily descriptive rather than normative: we are interested in describing the actual conventions governing the use of pronouns in languages, with a focus p. 266

on English. We are not arguing for any particular conception of what the ideal

conventions might be. That

said, we will stop at various points to discuss the normative implications of our descriptive claims for debates about the use of gendered pronouns. We have written above of ‘gender’ and ‘gendered pronouns’, as well as of ‘biological sex’. These terms invite confusion, and it is worth clarifying at the outset what we mean when we use them. There are two important distinctions to be made here. First, we distinguish between gender as a property of persons (personal gender) and gender as a property of linguistic expressions (grammatical gender). Grammatical gender is a theoretical posit in linguistics that is primarily intended to explain certain morphosyntactic processes of agreement. Personal genders, on the other hand, are generally taken to be socially constructed 5

categories, akin to professional categories like surgeon and legal categories like parent.

The second important distinction is between personal gender and sex. In contrast to personal gender, which 6

is generally taken to be a social phenomenon, sex is generally understood to be a biological phenomenon. Exactly how best to de ne sex is a complicated matter. Following Gri

ths (2021), we note brie y that

biologists, particularly evolutionary biologists, characterize sex in terms of gametes. Many species have phenotypes that are involved in producing larger gametes and phenotypes that are involved in producing smaller gametes. Organisms producing smaller gametes are classi ed as male; those producing larger gametes are classi ed as female. Although it is not in keeping with evolutionary biology, (human) medicine often characterizes sex in terms of chromosomes. In humans, individuals with one X chromosome and one Y chromosome are, in general, male; individuals with two X chromosomes are, in general, female. In the few cases where we need to represent what an informed non-expert might think about sex, we will use chromosomal sex as an example, even if this is not quite what evolutionary biology might tell us. We adopt a convention of using ‘male’ and ‘female’ to pick out sex categories and ‘man’, ‘woman’, ‘boy’, and ‘girl’ to pick out personal gender categories, without taking any further stand on what those categories are. Note that these terms belong to our semantic metalanguage; we are not interested here in the semantics of e.g. ‘female’ or ‘woman’ in English. Finally, we note that existing literature in linguistics often contrasts grammatical gender with natural gender. As we understand this term, it is meant to pick out whatever non-grammatical properties are semantically implicated by a given language’s grammatical gender system. Talk of natural gender is useful because it allows linguists to state certain generalizations about grammatical gender without having to answer the di

cult question of whether the languages they are studying are semantically sensitive to

personal gender, sex, or other properties. (As we will see below, languages make use of a wide range of natural gender properties.) That said, in what follows we will try whenever possible to use more speci c terms (‘personal gender’ or ‘sex’) rather than ‘natural gender’. Given that grammatical gender is a theoretical posit internal to the science of language, it might reasonably be wondered why it is implicated in contemporary controversies about personal gender. The answer to this question is that, while grammatical gender is a property of linguistic expressions, it is in many cases

p. 267

associated with

those expressions in a non-arbitrary way which seems to be sensitive to something like

sex and/or personal gender. When it comes to English pronouns like ‘she’ (grammatically feminine) and ‘he’ (grammatically masculine), it seems clear that linguistic conventions place constraints on the sorts of individuals to whom they may felicitously be used to refer: one cannot, for example, felicitously use ‘she’ to refer to Peter Geach, nor ‘he’ to refer to Elizabeth Anscombe. Yet it is not clear whether the felicitous use of pronouns like ‘she’ and ‘he’ is governed by the personal gender of the referent (as TikTok’s revised community guidelines seem to suggest), the sex of the referent (as the Tennessee bill seems to suggest), or some more complicated property or properties. Therefore, though gendered pronouns are clearly gendered in the grammatical sense, when it comes to semantics, it is not clear whether they are gendered, sexed, or neither—this is the main question we will consider here. To emphasize the openness of this question about the semantics of pronouns, we will avoid using terms like 7

‘masculine’ and ‘feminine’ when referring to English pronouns. Instead, we will refer to the English pronoun ‘she’ and its in ected forms as S-series pronouns and to the English pronoun ‘he’ and its in ected forms as H-series pronouns. Let us also introduce the term S-properties as a label for whatever properties of an individual license the use of S-series pronouns to refer to her; and let the term H-properties be de ned in the same way, mutatis mutandis. Stated in our terminology, the crucial question about gendered pronouns in English is what the corresponding S-properties and H-properties are—whether personal gender, sex, or something else. Our discussion divides naturally into two parts. The rst part, consisting of sections 2 and 3, is a general introduction to the linguistics and psychology of grammatical gender. Readers who are familiar with these topics may wish to focus their attention on the second part of the chapter, which speci cally concerns the semantics of gendered pronouns in English. We begin this second part by discussing some methodological limitations of empirical approaches to our topic and the normative implications of those limitations (sections 4 and 5). Relying on our own semantic judgements as native speakers of English, we then argue against three simple theories of the semantics of S- and H-series pronouns in English and propose an alternative, the Gender-First View (sections 6 and 7). Finally, we discuss the singular ‘they’ and its connection to nonbinary gender identities (section 8). Section 9 concludes.

2. Gender: An Overview In keeping with our e ort to ground our observations in linguistic facts, we start with a brief overview of the linguistics of grammatical gender. For us, the main observation is that grammatical gender is a feature of nouns that sorts them into noun classes in accordance with their involvement in agreement patterns. Some 8

common examples include noun/pronoun agreement in English, as in: p. 268

(1)

a. The man reached his destination. b. *The man reached her destination.

We also see agreement between nouns and predicate adjectives in languages like French:

(2)

a. LʼhommeM est grandM. (“The man is big.”) b. La chaiseF est grandeF. (“The chair is big.”)

Cross-linguistically, we also nd a variety of other forms of agreement involving gender. We invite the 9

interested reader to pursue this variety in the linguistics literature.

In addition to driving agreement patterns and sorting nouns into classes, gender features are among what linguists often call phi-features (or ᵠ-features). Phi-features typically include person and number along with gender. They are distinguished from other types of features by having semantic content. In contrast, other grammatical features need not have semantic content. Whatever features make a noun a noun and not a verb, for example, are not part of its meaning. Phi-features both play a role in syntax and have semantic 10

content.

As we de ned S- and H-properties, it follows immediately that the semantic contents of

pronouns in English (their phi-features) re ect them. We will look at the wider range of contents of gender features cross-linguistically in a moment. What is the status of the content carried by grammatical gender features? One natural idea is that the semantic contents of phi-features, including gender, are presupposed. This is a technical notion, but the main idea is that presupposed content is what is taken for granted in saying something, rather than the ‘pro ered content’, which is what the speaker is adding by saying what they say. A useful example is the presupposition of a change-of-state verb like ‘stop’:

(3)

Sam stopped smoking a. Presupposition: Sam smoked in the past. b. Pro ered content: Sam no longer smokes.

In some way, the speaker takes for granted that Sam used to smoke, and adds that he or she no longer 11

does.

Presuppositions are often identi ed through projection patterns. What is taken for granted stays in place whether or not you negate a sentence, for instance, or whether you put the presupposition-carrying material in the antecedent of a conditional or a polar question. We thus see:

(4)

a. Sam stopped smoking. b. Sam did not stop smoking. c. If Sam stopped smoking, it is great. d. Did Sam stop smoking?

p. 269

12

All of these indicate that Sam used to smoke.

We see broadly the same behavior with the gender features of pronouns in English:

(5)

a. Sam respects her. b. Sam does not respect her. c. If Sam respects her, I should too. d. Does Sam respect her?

All likewise indicate that the referent of ‘her’ has S-properties. 13

So, it is an appealing idea that the contents of phi-features, including gender features, are presupposed.

That view has been advanced, e.g. by Cooper (1983) and Heim and Kratzer (1998). But we should note that Cooper’s main interest is quanti cation, and that Heim and Kratzer’s book is a textbook, trying to cover a wide range of material. Neither is focused on gender or phi-features. It is well known that when we look in more detail, gender features do not obviously project exactly like standard presuppositions. For example, we see:

(6)

a. Bill thought that Sam stopped smoking. b. Bill thought that she was a linguist.

In these environments, the presupposed content is typically that Bill believes or thinks the presupposition of the embedded constituent. Thus, it is indicated by (6a) that Bill believes that Sam used to smoke. But (6b) 14

seems di erent. It still seems to indicate that the referent of ‘she’ has S-properties.

Another problem for the view that phi-features are presupposed is that gendered pronouns show di erent behavior when they are bound. Presuppositions project under quanti cation. For instance:

(7)

a. Every student stopped smoking. b. Of all the students, only John stopped smoking.

The presupposition of ‘stopped’ projects here. Both of these seem to presuppose that every student used to smoke. Matters here are complicated by the fact that the properties of presupposition projection under quanti cation are disputed. Some have argued that what is presupposed here is only that some students 15

used to smoke.

Regardless of exactly what is projected, however, something clearly is.

In contrast, gender features seem to disappear under binding:

(8)

Of all the students, only Mary aced her homework.

Here we need agreement between ‘Mary’ and ‘her’ locally. But we see no indication that all the students bear S-properties linked to ‘her’. The gender feature of ‘her’ does not seem to project up to the whole quanti ed p. 270

sentence. This is in marked contrast to the

behavior we saw a moment ago. Likewise, in cases where the

class is partly men and partly women, either of the following is acceptable (Heim 2008):

(9)

a. Every student in the class voted for himself. b. Every student in the class voted for herself.

Again, we see that gender features are not looking like presuppositions, at least, not at rst glance. It should be noted that the issue here is not restricted to gender. Bound pronouns can seem to lose their phifeatures. We see this in:

(10)

Only I did my homework.

The rst-person feature seems to be lost here: ‘I’ seems to be functioning like a bound variable, and its 16

person feature seems to have disappeared.

Somehow, when bound, pronouns can shed their phi-features,

and speci cally for us, they can shed their gender features, at least if the context allows. That is not typical presuppositional behavior. There has been a great deal of work on what is happening in these cases with bound pronouns. We refer interested readers to Heim (2008) and Sudo (2012) for overviews and proposals. Sudo, in particular, argues at length that despite the problems we have noted, gender features are presupposed. Many authors have suggested that binding somehow eliminates phi-features. Heim expresses dissatisfaction with all the available options, and calls for more research to understand the phenomenon better. For rough and ready purposes, we suggest that one can think of the semantic content of gender features as presupposed. Outside of binding cases, it behaves more or less like a presupposition. The binding cases present an unresolved set of problems, as do the delicate properties of projection under attitude verbs. So, we can say that gender features are presupposed, but do note that this is very rough, and ready only in some situations. So much for the projective properties of gender features. We turn now to examining grammatical gender from a cross-linguistic perspective. In English, nouns are not grammatically gendered. Some nouns, e.g. ‘man’ and ‘woman’, are seen as carrying natural gender, re ecting the personal genders and/or sexes of the things they pick out. But we really only nd phi-features on pronouns. ‘He’, ‘she’, ‘it’, and ‘they’ carry phifeatures typically labeled masculine, feminine, neuter, and common, respectively.

Many languages show richer gender systems, and these systems also show important variety. Gender features, in contrast to number features and person features, typically carry contents related to personal gender, sex, animacy, humanness, or animalhood. We mention a few examples, drawn from work of Kramer (2020). Her broad cross-linguistic survey notes, among many other data points, that Sochiapan Chinantec (Otomanguean: Chinantecan) assigns gender using animacy: animate nouns are assigned one gender, and p. 271

inanimate nouns another. As she also notes, many

Niger-Congo languages have one gender for human-

denoting nouns and one for non-human-denoting ones. Some languages assign gender to nouns seemingly arbitrarily, at least outside of nouns with clearly gender-speci c semantic content. Spanish is an example. Some languages assign gender based in part on morphology. Russian does this for ‘lower animals’ and inanimate objects. Some languages, like Hungarian, seem to have no gender at all, and do not have gendered pronouns. Some languages assign grammatical gender in a way that con icts with natural gender. Polish, for instance, marks the terms for ‘girl’ and ‘puppy’ as grammatically neuter. Some languages assign grammatical gender to inanimate objects that may be associated with natural gender in some social group. So we nd in French ‘la jupe’ (the skirt, feminine) and ‘le pantalon’ (the pants, masculine). Grammatical 17

gender is, clearly, a rich and varied linguistic phenomenon.

For all their variety, gender systems do seem to have a semantic core. This is what Kramer calls the semantic core generalization: that grammatical gender systems always have some nouns whose gender is semantically predictable, such that: Grammatical gender is always assigned to at least a subset of nouns on the basis of animacy, humanness, and/or social gender for humans/sex for animals. 18

(Kramer 2020, 47)

When we turn to English pronouns in later sections, it is the details of this semantic core that will concern us. Before leaving our overview of the linguistic properties of gender, we pause to comment brie y on the use of ‘they’ as a singular pronoun. This has become increasingly common in the past several decades among some groups. Sometimes it is o ered as a gender-neutral pronoun (or least one marked as common, rather 19

than H-series or S-series). It is also o ered as an appropriate pronoun for nonbinary individuals.

This use is complicated by the fact that ‘they’ appears to carry a plural number feature for many English speakers. Even so, ‘they’ can appear to function as if singular in bound and anaphoric environments. Consider, for example:

(11)

a. i. Every parent believes their child is a genius. ii. Every parent believes his child is a genius. b. i. Somebody made a large donation, but they donʼt want to reveal their identity. ii. Somebody made a large donation, but she doesnʼt want to reveal her identity.

Note that the ‘they/their’ versions here are perfectly natural. They are perhaps most natural in cases where the quanti er ranges over a group which could include people of di erent genders or the target for anaphora is not speci ed for gender. But a least some English speakers also nd bound uses of ‘they’ 20

acceptable when gender is speci ed, as in:

(12)

p. 272

Every man said they were happy.

We have already discussed the fact that in quanti ed environments, pronouns can sometimes seem to lose their phi-features. It may be that we are seeing the same phenomenon here. Cases like these are a long-standing part of English. The more recent phenomenon is the use of ‘they’ as a singular pronoun in deictic environments, or, more generally, where it is not bound or anaphoric, and the personal gender or sex of the intended referent is clear to speakers. For instance, we often currently see ‘they’ used in certain speech communities as a preferred (singular) pronoun for nonbinary individuals. Some speakers no doubt nd deictic singular uses of ‘they’ awkward, even if they otherwise see themselves as part of a community like the ones we just described. Presumably, such speakers hear ‘they’ as marked plural and in deictic environments see it as awkward to use it for singular reference. However much they might want to conform to community ideas about gender-neutral pronouns, they simply hear ‘they’ as plural. Changing these sorts of facts about one’s idiolect can be hard to do. This is likely because pronouns belong to what linguists call the ‘closed class’ part of language, which also includes (for example) tenses, modals, and aspect markers. In contrast, the ‘open class’ part of language includes the major lexical categories like nouns and verbs. It has long been observed that the open class categories are open in that it is easy and quick to add to them. To add a new noun or verb, we merely need to nd an interesting new idea, attach a word to it, and see if it catches on. In contrast, to add a tense to a language is not something we can do so easily. Tense systems do change, but only at the glacial pace of language change. Just how quickly 21

closed class expressions can change is not fully understood.

Even though pronouns are closed class items, pronoun systems do change over time. Certainly, earlier forms of English had pronouns like ‘thee’ and ‘thou’, which in Middle English acquired marking for informality. We put aside any linguistic prescriptivism, which might insist on how singular pronouns ‘should’ be used. Our own suspicion is that our language is in ux, with mounting pressure for conventionalizing singular deictic ‘they’. With that, we will end our brief overview of the linguistics of gender marking. Our concern in what follows is not primarily with the syntactic realization of gender features or their role in agreement, but rather with their semantic and social properties. Nonetheless, it is important to remember that one of the main ways gender is identi ed cross-linguistically is via these purely syntactic properties. Thus our investigation here concerns the semantic properties of a feature in language that is as much syntactic as semantic. Gender is part of grammar, and it can show great variety and complexity. When we turn to the semantics of English pronouns below, we should remember that they show us one speci c case of how gender appears in language. They exhibit interpretable features that can relate to personal gender and/or sex, but they also show agreement properties and other aspects of the inner workings of a human language.

p. 273

3. Grammatical Gender and Psychology In this section, we consider psychological work on the cognitive relationship between grammatical gender, personal gender, and sex. Does a word’s grammatical gender a ect how speakers think about its referent? This might seem obvious for English. We might assume that if we refer to someone with an H-series pronoun, then we automatically think of them as having a particular personal gender and/or sex. Perhaps we do. But consideration of other languages and work in cognitive psychology shows that the connection between grammatical gender and how we think about something is not always simple. This question arises most vividly for languages that mark masculine or feminine gender on nouns whose referents are not the right types of thing to possess personal gender or sex. Do such languages implicitly guide speakers to think of certain things as masculine/male or feminine/female because of grammatical gender marking, even if those things are non-human animals, plants, artifacts, or inanimate objects? 22

Many have assumed not.

But more recently, a number of psychologists and psycholinguists have asked 23

whether we can see subtle e ects of grammatical gender marking in cognition.

Proponents of speci c

views about what S- and H-properties are might hope to nd some support for their view in these kinds of experiments. Might we in some subtle way think of things as speci cally male or female, or as men or women, when the words we use for them are marked for gender? A note of caution before proceeding: the results we will survey here are not all consistent, and the experimental designs di er substantially. As with any experiment, one can and should ask about their designs, the quality of the data, and the strength of the e ects. These empirical results should be taken with great care and caution, and, when possible, should be used with input from experimentalists who can help us interpret them and use them well. Our main question in this section is whether languages with rich gender marking and some arbitrary gender assignment trigger e ects of grammatical gender when subjects think about objects without natural gender or sex (inanimate objects, artifacts, etc.). But one has to probe for this without relying on speakers’ naming or categorizing the objects, which would simply re ect the genders their languages assign. One needs to nd other ways to track the role of gender in thought. Keeping our note of caution in mind, we start with the example of a much-cited set of studies from Boroditsky et al. (2003). In one of their experiments, Boroditsky et al. used a list of object nouns that have the opposite grammatical gender in German and in Spanish. Native speakers of German and Spanish were asked to provide the rst three adjectives that came to mind to describe a named object. These were rated for whether they were masculine or feminine descriptors. The nding was that speakers’ descriptors followed the grammatical gender of the noun. p. 274

This appears to be a nding that grammatical gender a ects how we think about things. We can, as always, ask how strong such a result is. We can ask about the coding of descriptors, the sample size and population, and any number of other standard questions. And, we will see, other experimental designs have shown di erent results. But this result holds out the tantalizing prospect that when you hear a gendered descriptor, you think about the thing as if it has some form of natural gender. Other ndings are not so clear. Here is an example, involving a very di erent experimental design, due to Vigliocco et al. (2005). In this experiment, subjects were o ered three words and asked to judge which two were most similar in meaning. Vigliocco et al. compared the judgements of speakers of Italian and of German, which mark nouns for grammatical gender, with the judgements of speakers of English. They found that for Italian speakers, meaning similarity is a ected by grammatical gender for animal terms, but not artifact terms, where the results were similar to English speakers. For German speakers, they found no di erence with English speakers for either animals or artifacts. Of importance here is that Italian has two

grammatical genders, while German has three grammatical genders and marks all diminutives as grammatically neuter even if their referents have natural gender. Hence, it makes a less consistent mapping between grammatical gender features and natural gender. A nal experiment compared Italian and English speakers, but replaced words with pictures. Vigliocco et al. found no e ect of gender in this case. In sum, Vigliocco et al. found at best highly limited e ects of grammatical gender. They found them only for two-gender languages and then only for animal terms, not artifacts. In other experiments, they also found the e ects to be highly task-speci c. Generally, though the experiments are di erent in design, they do not seem to nd the strong e ects of gender that Boroditsky et al. (2003) did. This reminds us of our note of caution. These are complicated experiments, and to our knowledge, there is not a large body of related work with which to compare them. Vigliocco et al. were careful to formulate two di erent hypotheses about how gender might a ect thinking. One is that the e ects of gender are not really distinctive. There is a well-known e ect in language learning 24

of learners wanting to associate similar morphosyntactic forms with similar meanings.

This could help

simplify the language-learner’s task. So one possibility is that when it comes to languages with rich grammatical gender marking, learners simply look for any similarity they can relate to gender marking. A second hypothesis is that speakers interpret grammatically gendered nouns directly as indicating sex or personal gender. This would work most easily for nouns for humans or animals. One version of this second hypothesis predicts e ects only for nouns denoting humans, animals, and anything else that might easily be conceived of as having sex or personal gender. A more general version predicts that it would apply to any noun. (This is close to the hypothesis explored by Boroditsky et al.) The second hypothesis in either form predicts that languages that mark grammatical gender di erently will produce di erent associations with gender. The rst hypothesis does not. Overall, Vigliocco et al. did not nd strong support for the rst p. 275

hypothesis for gender (though there is evidence for it in other domains). Nor did they nd support for

the

stronger version of the second hypothesis. They report only nding evidence for the modest and constrained version. One more example of work in this area is from Maciuszek et al. (2019), focusing on Polish. As Maciuszek et al. describe it, Polish has three main grammatical genders (masculine, feminine, and neuter), and also a rich system of gender-marking morphology and two further genders for plural forms (masculine-personal and 25

non-masculine-personal). It shows signi cant arbitrary gender assignment.

Maciuszek et al. take as a

starting point Vigliocco et al.’s hypotheses about the ways gender can a ect cognition. In one experiment, they used the three-word paradigm of Vigliocco et al., but tried to be more careful about other measures of similarity between words in a triple. Also, instead of using English as a comparison, they computed statistics directly from Polish speaker data. They found similar e ects to those found in Vigliocco et al. for German, though with some more details about semantic classes. But they also ran two other experiments. One was based on a variant of the Implicit Association Test (IAT). This work found more support for the rst of Vigliocco et al.’s hypotheses: that similar forms are associated with similar meanings. In another experiment, they used a paradigm of asking speakers to assign masculine or feminine voices to objects. They compared objects presented as pictures with presentation of corresponding nouns. Both inanimate objects and animals were used. Here they found a strong e ect of grammatical gender on how objects are conceptualized. Overall, they nd a complex situation, where aspects of grammar and cognition interact in a number of ways, and di erent hypotheses seem to be supported by results of di erent sorts of tasks. We think this illustrates the need for caution. Di erent studies, even with similar designs, show somewhat di erent results. We suggest that philosophers wait for more clarity and stability before relying on emerging empirical results in this area. We end this section with one more intriguing nding. A result from Segal and Boroditsky (2011) suggests that certain grammatical gender assignments seem to pick up at least some metaphorical signi cance in cases of personi cation. For example, the word for ‘sun’ in Spanish (‘el sol’) is grammatically masculine

while the corresponding word in German (‘die Sonne’) is grammatically feminine, and the word for ‘death’ is grammatically masculine in German (‘der Tod’) while the corresponding word is grammatically feminine in Spanish (‘la muerte’). Might Spanish speakers and German speakers depict the sun or death di erently for this reason? Here we know a little. Segal and Boroditsky (2011) found a strong correlation between personi cation in art and assigned grammatical gender. So, at least when it comes to art or metaphor, speakers can recognize and exploit arbitrary grammatical gender assignments. It is not easy to decide how this relates to our understanding of language, as the relation of metaphor interpretation to other aspects of 26

language is not a well-understood matter.

But at least this nding suggests that hearing a grammatical

gender might trigger some metaphorical thinking about natural gender. In light of the results just described, our best assessment, with due caution, is that given the complicated nature of the experiments and the con icting nature of the results, not much should be concluded with p. 276

certainty. Our brief overview of some

results from psychology, together with our brief overview of some

results from linguistics, reveals a complicated empirical situation. We have seen that languages mark gender in many di erent ways. These can relate to personal gender and/or sex, but can also re ect animacy, humanness, or other properties. Even in languages where gender marking does relate to personal gender and/or sex, it can also show signs of arbitrary marking for many nouns. When we look at languages with grammatical gender marking that at least sometimes re ects personal gender and/or sex, we can ask if the presence of such marking triggers thinking related to personal gender or sex in speakers. We have seen that the psychological results on this are, so far, incomplete. There may be some such e ects, in some cases. At the same time, it may just be that language learners see similar morphosyntactic marking as an indicator of similar meaning, and are only accessing their ideas about personal gender and/or sex to nd some similarities. As we turn to the English pronoun system and issues surrounding it, we would do well to remind ourselves that it is one among many di erent gender systems in language. We should be careful about drawing too many conclusions about how the marking of gender relates to thought from any one example.

4. Methodological Limitations We have seen that English lacks any grammatical gender on nouns and has a system of gendered pronouns corresponding roughly to personal gender and/or sex. In addition to S-series and H-series pronouns, there are the neuter it and the plural they, which is sometimes glossed as ‘common’ gender, as it can refer to groups with members of di erent genders and/or sexes. Thus English appears to be limited in the richness of its gender system, but to be no exception to the cross-linguistic generalization that most gender systems have a semantic core that is tied to something like personal gender and/or sex (or animacy, etc.). In the remainder of the chapter, we look more carefully at the situation with English in its current social environment. It is tting to begin this discussion by mentioning our limitations. Semantic theories are built around a number of data points, including judgements of assertability, truth, synonymy, and entailment. These semantic judgements can be supplemented with data about syntax, morphology, and so on, to ll in a picture of how a language works. The toolkit has expanded over the years, but that has been and remains its 27

core.

When we face aspects of language that relate to current social issues, we need to remind ourselves

that the methods of linguistics work best when they target hypotheses between which ordinary speaker– hearers can easily distinguish, when diachronically stable linguistic conventions prevail at the community level, and when questions of speech are divorced from questions of politics. In the case of gendered pronouns, in the setting of the English language and the politics of the United p. 277

States, none of these conditions are met. Typical English speakers

likely do not draw a sharp conceptual

distinction between sex and gender. The minimal amount of grammar we see in the English gender system

o ers us few stable points from which to start describing subtle aspects of semantics. Cognitive psychology o ers us few solid results to build on—if there are subtle e ects on thinking triggered by grammatical gender, we do not yet understand them. At the same time, the increasing visibility of trans and nonbinary individuals has resulted in the need to coordinate at the community level on linguistic conventions governing the use of gendered pronouns in cases involving individuals exemplifying combinations of sex and personal gender that, fty or a hundred years ago, would never have become salient to the typical English speaker–hearer. And the broader conversation about how to coordinate on these conventions is bound up with heated political discussions of feminism and the rights of trans people. These observations suggest to us that any descriptive inquiry into the semantics of gendered pronouns in English should proceed with humility and caution. It is likely that the idiolects of individual speakers di er with respect to what property or properties license the application of gendered pronouns. It is also likely that the phi-features of gendered pronouns in the idiolects of many speakers are underspeci ed with respect to whether they pick out sex, personal gender, or some other, more complicated property. Some speakers may report patterns of judgements about the acceptability of uses of gendered pronouns that consistently suggest one theory of their phi-features as opposed to another; the judgements of other speakers may not t into any coherent pattern. Di erences in pronoun usage may also re ect di erent views about what constitutes sex or personal gender rather than or in addition to di erent internalized semantic theories for S- and H-series pronouns. At the community level, it may be that considerable agreement between speakers exists about how gendered pronouns may be acceptably used, but it also possible that no convention yet exists. These are empirical questions, answerable only by eliciting the 28

judgements of a large and diverse body of English speakers.

To summarize: when considering English pronouns, we face an empirically di

cult situation. We have

limited grammar to work with, as the gender system of English is minimal. Results from psychology are uncertain. And we face a political and social situation that can render linguistic judgements indeterminate or hard to probe for. But it does not follow that nothing at all can be said about the semantics of English gendered pronouns. Below, we discuss the pattern of our own intuitions about cases and argue that they favor one possible theory of the phi-features of gendered pronouns over certain others. To the extent that our intuitions are shared by others, our discussion can be interpreted as evidence that the theory we favor correctly describes the meaning of gendered pronouns in English. But it should be kept in mind that the intuitions are, rst and foremost, our own, and that drawing any rm conclusions about the grammar of English would require empirical work beyond the scope of the present discussion. Before turning to our intuitions, however, we pause to consider the normative implications of the possibility that there might be no single convention governing the use of gendered pronouns in English.

p. 278

5. Anything Goes? If di erent speakers might be guided by di erent internalized grammars for gendered pronouns, and if there might not at present exist any community-level conventions or clear psychological tendencies to which one could have recourse in deciding which uses of gendered pronouns are correct and which are not, does it follow that all ways of using gendered pronouns are somehow on a par? In particular, does it follow that there is nothing wrong with misgendering trans individuals by referring to them using pronouns other than those they prefer (e.g., using ‘she’ to refer to a trans man or ‘he’ to refer to a trans woman)? In this 29

section, we argue that the answer to this question is negative.

If the conventions of English as it is currently used do not decide whether the phi-features of gendered pronouns pick out sex, personal gender, or something less speci c, then using gendered pronouns to misgender someone cannot be said to be incorrect according to the conventions of English. Nevertheless, we argue that misgendering uses of gendered pronouns can be assessed in other ways, and can be found objectionable, or, at least, dispreferred. We can nd a normative dimension to misgendering, even if it does not derive from the current conventions of English. We will focus on two aspects of misgendering uses of gendered pronouns, drawing on the literature on pragmatic approaches to the pejorativity of slurs. First, Bolinger (2017) o ers an explanation of the pejorative e ects of slurs in terms of constrastive choice. Bolinger’s idea is that the use of di erent expressions—even ones with the same semantic content—can be probabilistically associated by listeners with di erent information about the speaker. To take a simple example, if a speaker chooses to use the lexical item ‘aubergine’ rather than the lexical item ‘eggplant’ to refer to an eggplant, hearers will likely infer that she is British (note that this inference depends on the assumption that the speaker could just as easily have tokened the other expression; when this assumption is not justi ed, the inference does not go through). But the information probabilistically associated with uses of expressions can also pertain to the beliefs, a ective attitudes, and political orientations of a speaker. This, on Bolinger’s view, explains why freely choosing to use a slur rather than its neutral counterpart is o ensive: it signals to hearers that the speaker harbors pernicious beliefs, attitudes, or political views. Whether or not they are linguistically correct, it is plausible that misgendering uses of gendered pronouns are probabilistically associated in the minds of many English speaker–hearers with certain beliefs, a ective attitudes, and/or political commitments regarding trans individuals. These could range quite widely, depending on facts about the individual doing the misgendering. In some cases, misgendering could simply indicate a general insensitivity to or misunderstanding of current social and political issues. We might think this when talking to an elderly person, who would have grown up in an environment where current issues about gender were never spoken of, if they were recognized at all. In other cases, the tissue of beliefs, p. 279

attitudes, and commitments

could be much stronger. In our current social and political environment,

misgendering someone might signal dislike of people who prefer pronouns di ering from those corresponding to their sex. In more extreme cases, it could, for instance, signal a belief that trans people are mentally ill or a political commitment to passing legislation to deny them appropriate medical treatment and exclude them from gendered spaces corresponding to their gender identities. It is not our job to comment on politics here. But to the extent that signaling such beliefs is hurtful, misgendering might also be hurtful. And to the extent that any of the beliefs in question could be held to be normatively 30

objectionable, so, too, could misgendering as an expression of them.

Second, Herbert (2017) emphasizes the fact that hearing an utterance of a slur (even a slur being mentioned 31

rather than used) can produce harmful e ects; “pernicious associations,” as she describes them.

Thus

hearing a slur might raise to salience the troubled history of the use of that slur, the associated complex of prejudicial beliefs about the group it targets, and/or certain well-known instances of violence against that group. Hearing a slur might also cause members of the audience involuntarily to recall unpleasant personal experiences involving the slur. These associational e ects of slurs play an important explanatory role, since they provide an account of how uttering a slur can be o ensive even when the slur is quoted. Though in such cases the speaker cannot correctly be said to have used the slur or applied it to any individual or group, if she could have made the same point without uttering a word which would conjure pernicious associations in members of her audience, she can legitimately be criticized for failing to show concern for her 32

interlocutors.

The same e ects can be seen with misgendering. Of course, the e ects of any particular case of misgendering will depend on the psychological facts about the individuals involved, and perhaps also the political and social situation. But here is a generalization drawn from current research: trans individuals consistently report that misgendering uses of gendered pronouns that target them conjure negative

33

emotions and associations which are plausibly just as unpleasant as those triggered by slurs.

So, just as

associational o ense explains how slurs can be o ensive even when they are quoted (and thus semantically inert), the possibility of associational o ense related to misgendering uses of gendered pronouns explains how they might be o ensive even if they are linguistically correct in the idiolect of the speaker. We leave it up to the reader to assess how the considerations we have just described bear on the moral properties of misgendering. We do hope to have shown, however, that a lack of clear conventions about how English pronouns should be used would not by itself settle the question of whether misgendering can be objectionable.

6. Three Simple Theories Bracketing concerns about whether there is a stable community-level convention governing the use of p. 280

gendered pronouns, we turn now to a discussion of our own

intuitions. The pattern of our judgements

about cases leads us to reject a number of simple proposals about the semantic content of the phi-features of gendered pronouns in English, and to endorse the more nuanced theory described in section 7 below. To the extent that our intuitions are shared by readers, they have reason to believe that their idiolects are best described by our preferred theory as well. The two simplest defensible accounts of the contents of the phi-features of S- and H-series pronouns are: (Gender-Only View): S-series pronouns are appropriate if and only if the referent is a woman or girl; Hseries pronouns are appropriate if and only if the referent is a man or boy. (Sex-Only View): S-series pronouns are appropriate if and only if the referent is female; H-series 34

pronouns are appropriate if and only if the referent is male.

According to (Gender-Only View), the phi-features of gendered pronouns are exclusively sensitive to personal gender; according to (Sex-Only View), they are exclusively sensitive to sex. A third possible account, based on the rst two, is: (Ambiguity View): Each S-series pronoun is ambiguous between a lexical item which is correctly described by (Gender-Only View) and a lexical item which is correctly described by (Sex-Only View). The same is true of each H-series pronoun. According to (Ambiguity View), whereas it would at rst seem that there is a single English personal pronoun ‘she,’ in fact there are two: one which can appropriately be used to refer to an individual just in case that individual is a woman (or girl), and one which can appropriately be used to refer to an individual just in case that individual is female. We will brie y discuss each of these three views in turn, sketching its advantages and then describing the considerations we take to show that it is not viable as a nal theory of English gendered pronouns.

6.1. The Gender-Only View There is much to recommend (Gender-Only View). Most simply, it seems natural to refer to men by ‘he’ and women by ‘she’. Moreover, (Gender-Only View) explains the practice of referring to (binary) trans individuals using their preferred pronouns. Since trans men are men and trans women are women, it is appropriate to refer to trans men using H-series pronouns and to trans women using S-series pronouns. Importantly, the explanation it o ers here is that these uses are linguistically correct given the relevant claims about personal gender. To use anything but an H-series pronoun for a trans man or an S-series p. 281

pronoun for a trans

woman is to presuppose something false about that individual’s gender. For this

reason, (Gender-Only View) has a neat explanation of what is objectionable about misgendering a trans person. Unfortunately, (Gender-Only View) is also subject to a number of di

culties. It fails to predict the felicity of

using gendered pronouns to refer to beings which possess sex but not gender, such as the higher nonhuman vertebrates. Along similar lines, on the assumption that infants are too young to assume gendered social roles or to have developed internal gender identities, it struggles to explain the common practice of referring to infants using S-series pronouns if they have female genitalia and H-series pronouns if they have male genitalia. On (Gender-Only View), it would seem to be a conceptual or linguistic mistake to refer to one’s newborn child or one’s pet using gendered pronouns, whereas this practice is common and relatively uncontroversial.

6.2. The Sex-Only View The advantages and disadvantages of (Sex-Only View) are almost precisely the inverse of those of (GenderOnly View). (Sex-Only View) straightforwardly predicts the felicity of using gendered pronouns to refer to beings which possess sex but not gender. On the other hand, it fails to predict that it is linguistically correct to refer to binary trans individuals using their preferred pronouns, and thus that there is anything mistaken about misgendering uses of gendered pronouns. Indeed, it predicts that referring to binary trans individuals using their preferred pronouns is linguistically incorrect. The best the proponent of (Sex-Only View) can say about this issue is that we might have pragmatic reasons, such as those discussed in section 5, to refrain from using linguistically correct gendered pronouns to refer to trans individuals. We take this to be a signi cant consideration against (Sex-Only View), since it seems to us that using binary trans individuals’ preferred pronouns is linguistically correct in addition to being courteous. Consider, for example:

(13)

a. If Jonah is a transgender man, he is the first transgender man at his company. b. *If Jonah is a transgender man, she is the first transgender man at her company.

Our judgement is that (13b) is a linguistically incorrect use of S-series pronouns (not simply, for example, a discourteous one). (Sex-Only View) cannot accommodate this intuition. Similarly, it often seems that we can be ignorant about an individual’s sex without being ignorant about which gendered pronouns are the linguistically correct ones to use to refer to that individual. The possession of a Y chromosome is, at least ceteris paribus, associated with male sex. But we judge the choices of gendered pronouns in the following examples felicitous because correct:

p. 282

(14)

a. I donʼt know how many Y chromosomes Joan has because she hasnʼt told me. b. I donʼt know whether Joan is trans; it would be rude to ask her out of the blue.

Again, (Sex-Only View) cannot predict these judgements. It is also possible to use gendered pronouns to refer to individuals with no sex. It is easiest to nd examples of this in ction, but we note that the ctional cases show no linguistic resistance. We naturally refer to ctional humanoid androids like C-3PO of the Star Wars franchise and Marvin the Paranoid Android from Douglas Adams’s The Hitchhiker’s Guide to the Galaxy using H-series pronouns. Though these things do not exist, we nd no linguistic problems with the ctions. A similar point can be made about inanimate objects: it is acceptable to refer to watercraft using S-series pronouns, for example. These observations are di

cult

35

to reconcile with (Sex-Only View).

6.3. The Ambiguity View Because it holds that English gendered pronouns are ambiguous between a gender-only-type meaning and a sex-only-type meaning, (Ambiguity View) predicts felicity in any case where either of those two views would predict felicity. Unlike (Gender-Only View), then, (Ambiguity View) predicts that it is acceptable to use gendered pronouns for beings which possess sex but not gender. Unlike (Sex-Only View), moreover, (Ambiguity View) predicts that it can be correct to refer to a trans individual using their preferred pronouns. In these respects, (Ambiguity View) represents an improvement over both (Gender-Only View) and (SexOnly View). But the advantages of (Ambiguity View) should not be overstated. Indeed, we include it here more for completeness than because we think it a likely competitor view. Like (Sex-Only View), it fails to predict that there is anything mistaken about misgendering uses of gendered pronouns—for this reason, it fails to predict the intuitive contrast between (13a) and (13b). And it does no better than (Gender-Only View) and (Sex-Only View) at explaining how gendered pronouns can felicitously be used to refer to inanimate objects which possess neither gender nor sex. Perhaps more importantly, as the ambiguity view is a claim about the semantics of a natural language, we would like to nd independent evidence that English pronouns are ambiguous—and we do not see any such evidence. For these reasons, we think that (Ambiguity View) is implausible as a semantic theory. We must look elsewhere for a satisfactory theory of English gendered pronouns.

7. Gender First None of the three simple theories we have just canvassed stands up to scrutiny, at least from the perspective p. 283

of our personal linguistic intuitions. In this section, we argue that

this is not at all surprising. There is no

reason to expect that the correct theory of the phi-features of English gendered pronouns will be simple. Adding some complexity, we propose an alternative view which we nd more plausible. We have seen that grammatical gender is important for noun/pronoun agreement in English and for noun/adjective agreement in other languages like French. The rst of these processes appears to require a match between the grammatical gender of the pronoun and the personal gender and/or sex of its referent;

the second requires agreement between the grammatical gender of the adjective and the grammatical gender of the noun it modi es. But things get more complicated as soon as we consider gendered pronouns in languages with a grammatical gender distinction for nouns. In many such languages, the lexical items which are used as personal pronouns are also obligatorily used to refer anaphorically to inanimate objects with the corresponding grammatical gender. In German, for example, the word ‘Kamera’ (camera) is grammatically feminine, and so the appropriate pronoun for referring to a single camera is ‘sie’ (she):

(15)

Wo ist meine KameraF? SieF ist im Schrank. (“Where is my camera? It is in the cabinet.”)

At the same time, ‘sie’ is the appropriate deictic pronoun in German to apply to women. So the phi-features of gendered pronouns in languages which, like German, require agreement between the grammatical gender of a pronoun and the grammatical gender of its nominal antecedent must obey a complex disjunctive rule: if the referent is a person, their acceptability depends on some set of social and/or biological facts about that person; if the referent is not a person, their acceptability depends on the grammatical gender of 36

the nominal antecedent.

Phi-feature complexity of this type is common in natural language. Similar

examples could have been constructed using the languages of the Romance family, for example. For languages with more complex noun classes or gender features not related to personal gender or sex, as we reviewed in section 2, matters can get even more complicated. What examples like (15) show is that it is possible and indeed cognitively natural for humans to internalize complicated and disjunctive rules governing the acceptability of pronouns. So, while nothing excludes a priori the possibility that the correct theory of English personal pronouns is as simple as (Gender-Only View) or (Sex-Only View), we assign that possibility a low probability, even setting aside the considerations discussed in section 6. Some of the data we reviewed in section 3 also support this perspective. Once we discard the thought that the correct theory of English gendered pronouns must be simple, a vast terrain of theoretical possibilities reveals itself to us. We will not presume to explore this terrain fully; instead, we will discuss one view of the semantics of English gendered pronouns we nd especially appealing because of its t with our linguistic intuitions about particular cases. According to the view we p. 284

favor, the phi-features of English gendered pronouns are sensitive to gender when they are used to

refer

to gendered beings, sensitive to sex when they are used to refer to beings with sex but not gender, and potentially sensitive to still other facts when used to refer to things possessing neither sex nor gender: (Gender-First View): If the referent of an English personal pronoun belongs to a gendered category, Sseries pronouns are appropriate if and only if the referent is a woman or girl, and H-series pronouns are appropriate if and only if the referent is a man or boy. Otherwise, if the referent belongs to a sexed category, S-series pronouns are appropriate if and only if the referent is female, and H-series pronouns are appropriate if and only if the referent is male. In other words, according to (Gender-First View), the S-properties and H-properties di er depending on what sort of thing the referent is. Note that, as we understand the notion of belonging to a gendered category, it does not require an individual to actually possess a gender. Normal adult humans belong to a gendered category, and so a given gendered pronoun is appropriate for them only if they possess the corresponding gender. There may be some normal adult humans who do not have a gender; (Gender-First View) then predicts that it is not linguistically correct to refer to such individuals using either ‘he’ or ‘she’. We consider the semantic issues raised by such nonbinary identities in section 8.

Note that (Gender-First View) is compatible with many di erent accounts of when it is appropriate to use gendered pronouns to refer to objects possessing neither sex nor gender. We think this is as it should be, for English. For instance, it is probably a one-o

convention, not amenable to interesting systematization, that

it is acceptable in English to refer to watercraft using S-series pronouns—and there are probably a variety of these sorts of one-o

37

conventions across natural language.

So a complete account of the conventions

governing the use of gendered pronouns will be very complex and disjunctive; (Gender-First View) captures just the part of that complexity which governs the application of English gendered pronouns to things which possess personal gender and/or sex. Observe that (Gender-First View) reproduces the advantages of both (Gender-Only View) and (Sex-Only View) without being subject to their disadvantages. Like (Gender-Only View), it explains our intuitions about examples like (13a, 13b) and (14a, 14b) and provides an account of the incorrectness of misgendering uses of gendered pronouns. Like (Sex-Only View), it explains why it is often correct to refer to non-human animals using gendered pronouns. If ctional humanoid androids can be said to have a gender, it explains why it is appropriate to refer to them using gendered pronouns; if they cannot, it assimilates them, like watercraft, to the hodge-podge of non-sexed, non-gendered entities which might feature in one-o conventions governing the acceptability of gendered pronouns. Finally, though it accounts for variation and complexity, it does not posit a brute lexical ambiguity. For these reasons, we nd (Gender-First View) to be 38

a good candidate for the correct view of English gendered pronouns.

p. 285

8. Nonbinary Identities and ʻTheyʼ We noted in section 2 that some communities use ‘they’ as a singular deictic pronoun (in addition to its bound and anaphoric uses). We now turn to a brief discussion of the semantics of such uses. Running parallel to the recent increase in discussions of issues concerning binary trans individuals and the use of gendered pronouns has been the emergence of dialogue about nonbinary trans identities and the 39

gender-neutral singular deictic pronoun ‘they’.

In this section, we will focus on the relationship between ‘they’, nonbinary identity, and the semantics of gendered pronouns. In particular, we will be interested in assessing an argument against (Gender-First View) and (Gender-Only View) from two premises related to nonbinary identity. First, assume that what it is to be nonbinary is to lack a gender. Second, note that it is sometimes appropriate to refer to nonbinary individuals using gendered pronouns. Thus, it is not always unacceptable to token:

(16)

Dallas told me that she identifies as nonbinary.

If (16) is acceptable, and if Dallas lacks a gender in virtue of being nonbinary, then (this argument proceeds) S-series pronouns must be licensed by features of individuals other than gender. Thus the acceptability of (16) might be thought to favor (Sex-Only View) over alternatives. We think there are a number of ways of resisting this argument. First, the claim that all nonbinary people lack a gender is very strong, and perhaps dubious. Dembro

(2020), for example, provides testimony from a

number of individuals who identify as both nonbinary and women. But if nonbinary individuals can be gendered, (Gender-First View) and (Gender-Only View) have straightforward explanations for the acceptability of sentences like (16).

Even when nonbinary individuals do not identify with any gender category, the acceptability of sentences like (16) might be due not to their semantic well-formedness but instead to the fact that the relevant nonbinary individuals have explicitly permitted the use of semantically inappropriate pronouns to refer to them as a way of acknowledging the di 40

‘they’.

As Dembro

culty many members of society experience with using the singular

(2020, 9) remarks:

In our current society, saturated in exclusive, binary divisions, there is no possibility of never taking gender norms to be relevant to oneself. Public spaces, such as toilets and locker rooms, legal institutions, social clubs, language, and marketing, to name but a few places, are heavily gendered, and gendered not only according to the binary, but in a way that leaves someone attempting to navigate these structures no choice but to pick a side. Moreover, because all (or nearly all) p. 286

genderqueer persons

were socialized as either men or women, and often are perceived as men

or women, only self-applying the norms of ‘a person wearing people clothes’ is not possible. In cases where conventions for usage are in ux and communities face complex issues about how to establish stable uses, acceptability judgements can re ect more than just semantic or syntactic wellformedness. When nonbinary individuals who do not identify as gendered permit others to use gendered pronouns to refer to them as a way of acknowledging the impossibility of avoiding being perceived as gendered, sentences like (16) are predicted to be acceptable though not semantically well-formed. Again, then, the acceptability of sentences like (16) does not demonstrate the falsity of (Gender-First View) or 41

(Gender-Only View).

Indeed, far from counting against (Gender-First View) or (Gender-Only View), we think that sentences like (16) actually form the basis of an argument against (Sex-Only View). For in cases where a nonbinary person neither identi es with a gender category nor explicitly permits others to use gendered pronouns to refer to them, we judge that sentences like (16) are not acceptable: the only acceptable singular pronoun to use in such cases is the gender-neutral ‘they’. This fact is easily explained by (Gender-First View) and (GenderOnly View), but (Sex-Only View) struggles to accommodate it, since on (Sex-Only View) gendered pronouns can be correctly employed to refer to nonbinary individuals regardless of facts about their gender identities.

9. Conclusion English gendered pronouns are controversial. Yet if our arguments above are sound, perhaps they should be less so. There is little to recommend the simplistic idea that the linguistic correctness of S-series and Hseries pronouns must be sensitive in all cases either exclusively to sex or exclusively to personal gender, but our intuitions about sentences like (13a, 13b) and (14a, 14b) suggest that any plausible semantics will predict that the correctness of S- and H-series pronouns turns on personal gender whenever they are used to refer to normal adult humans. Even if there is at present no stable convention among English speakers governing the use of S- and H-series pronouns, the considerations discussed in section 5 suggest that all speakers, regardless of idiolect, have reasons to use their addressees’ preferred pronouns which resemble their reasons not to needlessly mention racial slurs. Together, these considerations suggest that there is no strong linguistic or normative argument to be made in favor of practices of misgendering trans people. But we also note that the issues of English pronouns do not by any means exhaust the range of issues surrounding gender in language. The range of grammatical gender systems, the complexity of our cognition of gender marking, and the richness of social and political environments in which we nd languages and thinkers are vast. We hope to have shed some light on pressing contemporary questions about pronouns in p. 287

English

and their relation to gender, but also to have indicated a little of where these questions t into a

broader perspective.

Acknowledgements Thanks to Ernie Lepore, Elin McCready, and Matthew Stone for very helpful comments on previous drafts of this chapter.

Notes

p. 288

1.

https://www.tiktok.com/community-guidelines (accessed 5/20/2022).

2.

An Act to Amend Tennessee Code Annotated, Title 49, Chapter 6, relative to student pronouns, HB2633, 112th General Assembly (2022).

3.

As reported in Lavietes (2022).

4.

Of course, the grammatical and semantic aspects of gender are only some of the many ways gender and language interact. Sociolinguistics has explored a great number of these. For an introduction to this work, we refer readers to Eckert and McConnell-Ginet (2003); and Hellinger, Bußmann, and Motschenbacher (2001–2015).

5.

See, for example, Haslanger (2000); Dembro (2016); and Barnes (2020).

6.

For a discussion of some of the complexities involved in this distinction, see Dembro (2021).

7.

Though such pronouns are clearly grammatically masculine or feminine, using terms like ʻmasculine pronounʼ may suggest a view on which the felicity of English pronouns like ʻheʼ is sensitive exclusively to the gender of the referent. Better to have terms which do not invite misinterpretation in this way.

8.

The * marking the second sentence indicates that it is judged somehow bad by speakers. It is not specific about what makes the sentence bad.

9.

See, for instance, den Dikken (2011); and Corbett and Fedden (2016).

10.

For extensive discussion, see Sudo (2012).

11.

A good reference on presupposition is Beaver (1997). For some interesting new developments, see Tonhauser et al. (2013).

12.

It remains a controversial issue in the foundations of presupposition just what the ʻindicateʼ relation is. Perhaps the best we can say, for our purposes, is that it is some form of implication, understood generally. Some theories see it as a feature of contexts in which an utterance is felicitous. Others see it as a combination of entailment and implicature. We will not pursue this matter here. See again Beaver (1997), as well as Simons (2006).

13.

Note that agreement does not, in any way, project. Rather, the idea is that the semantic content of a phi-feature might project like presupposition.

14.

For presupposition projection from attitudes, see Heim (1992), and the many references therein. For an extensive discussion of how these issues relate to gender, see Sudo (2012).

15.

Again, see Beaver (1997) and Sudo (2012) for overviews. This is a lively topic of current research.

16.

This example is probably due first to Partee (1989). See Heim (2008) for more references.

17.

Some readers may wonder what it means to say, as we have above, that some languages assign grammatical gender using (e.g.) animacy. A er all, if a grammatical phenomenon has to do with animacy rather than personal gender and/or sex, in what sense is that phenomenon grammatical gender? This raises an important methodological point, which is that, from the perspective of linguistics as a science, there is no interesting di erence between grammatical gender systems which are sensitive to personal gender and/or sex and morphosyntactic systems of noun classifiers which are sensitive to other phenomena like animacy. For this reason, linguists o en refer to all such morphosyntactic systems as gender systems. For more on gender across languages, see again Corbett and Fedden (2016).

p. 289

18.

Kramer (2020, 46) clarifies that she “use[s] the term social gender for the property of human beings indexed by grammatical gender”—that is, for whatever the S-properties and H-properties turn out to be.

19.

Some readers may find the idea of nonbinary gender identities unfamiliar. A good resource to consult on this topic is Dembro (2020). We will have only a little to say about the complex social and political issues involved. Our main focus is on the semantics of ʻtheyʼ.

20.

In fact, the two authors of this paper disagree about this example, which illustrates how much variation in judgements we can find.

21.

Any good linguistics textbook will expand more on the di erence between closed and open classes. Some more recent theoretical work includes Abney (1987) and Baker (2003). For some reflections on a case where language change appears to have happened unusually quickly, see Doron (2015).

22.

See e.g. István (1959).

23.

Note that this is an instance of the more general question of how language a ects thought. The idea that language broadly influences thinking is the Whorf-Sapir hypothesis (Whorf 1956). It has been widely criticized (e.g. Pinker 1994). Recent years have seen a number of developments of more carefully drawn ideas of how language can a ect thought and a great deal of experimental work on the issue. See Gumperz and Levinson (1996) and Gentner and Goldin-Meadow (2003). For a somewhat critical overview see Gleitman and Papafragou (2012).

24.

This is a version of the syntactic bootstrapping hypothesis. See, e.g., Landau and Gleitman (1985); and Fisher et al. (1991).

25.

A brief glance at the literature suggests that the gender system of Polish has been a controversial issue. See, for instance, Swan (2015).

26.

For an overview and many insights, see Camp (2006).

27.

See the first chapter of Chierchia and McConnell-Ginet (2000) for a good overview of the data and methods of semantics. The papers in Maienborn et al. (2011) give an overview of more recent developments.

28.

Moreover, as we noted in section 2, it might be that some speakers initially have trouble producing certain individualsʼ preferred pronouns, even when they regard those pronouns as correct.

29.

The prefix ʻmisʼ in ʻmisgenderingʼ might be taken to suggest that applying the term to a given use of a pronoun carries the normative implication that the use was linguistically incorrect. As we understand the term ʻmisgenderingʼ, however, it is purely descriptive: misgendering an individual is simply referring to them using pronouns other than the ones they prefer. Adopting a purely descriptive definition of ʻmisgenderingʼ is important for our purposes, since, as we have seen, there may be no stable community-wide conventions governing the linguistically correct use of pronouns in certain cases.

30.

For a recent treatment of misgendering which appeals to a similar mechanism, see Davis and McCready (2020).

31.

On the subject of how the pejorative force of slurs interacts with quotation, other work has come to similar conclusions— see, for example, Anderson and Lepore (2013).

32.

To be clear, we are not taking a stand on the correct theory of slurs. We merely find these ideas about slurs helpful for addressing the issue of misgendering.

33.

For a recent study of the e ects of misgendering on trans individuals, see Gunn (2020). The participants in Gunnʼs study described the feeling of being misgendered variously as “a black cloud hanging over my head,” “a sinking feeling in my chest and a rock in my stomach,” and “being punched in the gut” (Gunn 2020, 38–39).

34.

We remind readers that we have adopted a convention of using man and woman to pick out personal genders and male and female to pick out sexes.

35.

Note that the convention of using S-series pronouns to refer to watercra is also evidence against (Gender-Only View). Whether (Gender-Only View) can be reconciled with our observation about humanoid androids depends on whether such beings can correctly be said to possess a gender. Oneʼs answer to this question will depend on the particular account of gender one endorses: on Haslangerʼs (2000) social account, for example, they will fail to possess a gender; whereas on Jenkinsʼs (2016) account of gender as identity, some fictional humanoid androids might be men or women.

36.

Or, in cases lacking a salient nominal antecedent, on the grammatical gender of the most common noun used to refer to items of that type.

37.

We speculate that something similar is happening in languages that show a substantial number of arbitrary grammatical gender assignments.

38.

Of course, (Gender-First View) is a simplification suitable for a handbook discussion. It is in close sympathy with the important work of McConnell-Ginet (e.g., 2014, 6). We quote her at length: My own research, especially McConnell-Ginet ([1979] 2011), shows that gender in English, while not ʻgrammaticalʼ in the fullest sense because pronouns are the only agreement targets, is not really ʻnaturalʼ either. English-like languages have what I now call notional gender systems: pronominal usage cannot be understood without considering sociocultural gender and the ideas about sex and sexuality current at a given time. And it is such gender ʻnotionsʼ that can be embedded in and a ect agreement phenomena, especially but not only pronouns, even in languages where grammatical gender predominates. We recommend McConnell-Ginetʼs work to interested readers.

39.

As we mentioned in note 19 above, readers who find these issues unfamiliar might wish to consult Dembro (2020).

40.

We pause to emphasize that we take semantic well-formedness not to be a prescriptive matter. In the case in question, we can suppose that the pronoun semantically requires a referent of a particular gender (or sex), and yet is being used to refer to an individual that lacks it. This is a problem within the semantics, not a matter of prescription.

41.

Methodologically, note that we believe our linguistic judgements are sensitive to whether a given example sentence is linguistically well-formed in addition to whether it is acceptable in a social setting. Thus (16), which is acceptable but not linguistically well-formed, strikes us as di erent than (13a), which is acceptable because linguistically correct.

p. 290

References Abney, Steven P. (1987). ʻThe English Noun Phrase in Its Sentential Aspectʼ. PhD dissertation, Massachusetts Institute of Technology. Google Scholar Google Preview WorldCat COPAC Anderson, Luvell, and Lepore, Ernie. (2013). ʻWhat Did You Call Me? Slurs as Prohibited Wordsʼ. Analytic Philosophy 54:530–563. Google Scholar WorldCat Baker, Mark. (2003). Lexical Categories. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Barnes, Elizabeth. (2020). ʻGender and Gender Termsʼ. Noûs 54:704–730. Google Scholar WorldCat Beaver, David. (1997). ʻPresuppositionʼ. In Johan van Benthem, and Alice ter Meulen, eds., Handbook of Logic and Language, 939– 1008. Amsterdam: Elsevier. Google Scholar Google Preview WorldCat COPAC Bolinger, Renée Jorgensen. (2017). ʻThe Pragmatics of Slursʼ. Noûs 51:439–462. Google Scholar WorldCat Boroditsky, Lera., Schmidt, Lauren A., and Phillips, Webb. (2003). ʻSex, Syntax, and Semanticsʼ. In Dedre Gentner and Susan Goldin-Meadow, eds., Language in Mind, 61–79. Cambridge, MA: MIT Press. Google Scholar Google Preview WorldCat COPAC Camp, Elisabeth. (2006). ʻMetaphor in the Mind: Cognition of Metaphorʼ. Philosophy Compass 1:154–170. Google Scholar WorldCat Chierchia, Gennaro, and McConnell-Ginet, Sally. (2000). Meaning and Grammar: An Introduction to Semantics. 2nd ed. Cambridge, MA: MIT Press. Google Scholar Google Preview WorldCat COPAC Cooper, Robin. (1983). Quantification and Syntactic Theory. Dordrecht: Reidel. Google Scholar Google Preview WorldCat COPAC Corbett, Grenville. G., and Fedden, Sebastian. (2016). ʻCanonical Genderʼ. Journal of Linguistics 52:495–531. Google Scholar WorldCat Davis, Christopher, and McCready, Elin. (2020). ʻThe Instability of Slursʼ. Grazer Philosophische Studien 97:63–85. Google Scholar WorldCat Dembro , Robin A. (2016). ʻWhat Is Sexual Orientation?ʼ Philosophersʼ Imprint 16 (3): 1–27. Google Scholar WorldCat Dembro , Robin. (2020). ʻBeyond Binary: Genderqueer as a Critical Gender Kindʼ. Philosophersʼ Imprint 20 (9): 1–23. Google Scholar WorldCat Dembro , Robin. (2021). ʻEscaping the Natural Attitude about Genderʼ. Philosophical Studies 178:983–1003. Google Scholar WorldCat den Dikken, Marcel. (2011). ʻPhi-Feature Inflection and Agreement: An Introductionʼ. Natural Language and Linguistic Theory 29:857–874. Google Scholar WorldCat

Doron, Edit., ed. (2015). Language Contact and the Development of Modern Hebrew. Leiden: Brill. Google Scholar Google Preview WorldCat COPAC Eckert, Penelope, and McConnell-Ginet, Sally. (2003). Language and Gender. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Fisher, Cynthia, Gleitman, Henry, and Gleitman, Lila R. (1991). ʻOn the Semantic Content of Subcategorization Framesʼ. Cognitive Psychology 23:331–392. Google Scholar WorldCat Gentner, Dedre, and Goldin-Meadow, Susan, eds. (2003). Language in Mind. Cambridge, MA: MIT Press. Google Scholar Google Preview WorldCat COPAC Gleitman, Lila R., and Papafragou, Anna. (2012). ʻNew Perspectives on Language and Thoughtʼ. In Keith J. Holyoak and Robert G. Morrison, eds., The Oxford Handbook of Thinking and Reasoning, 543–568. New York: Oxford University Press. Gri iths, Paul E. (2021). ʻWhat Are Biological Sexes?ʼ Preprint. PhilSci Archive. http://philsci-archive.pitt.edu/19906/. WorldCat Gumperz, John J., and Levinson, Stephen C., eds. (1996). Rethinking Linguistic Relativity. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC p. 291

Gunn, Hamish A. (2020). ʻThe Experience of Misgendering among Trans and Gender Diverse Peopleʼ. PhD dissertation, University of Massachusetts, Boston. Google Scholar Google Preview WorldCat COPAC Haslanger, Sally. (2000). ʻGender and Race: (What) Are They? (What) Do We Want Them to Be?ʼ Noûs 34:31–55. Google Scholar WorldCat Heim, Irene. (1992). ʻPresupposition Projection and the Semantics of Attitude Verbsʼ. Journal of Semantics 9:183–221. Google Scholar WorldCat Heim, Irene. (2008). ʻFeatures on Bound Pronounsʼ. In Daniel Harbour, David Adger and Susan Bejar, eds., Phi Theory: PhiFeatures across Modules and Interfaces, 35–56. Oxford: Oxford University Press. COPAC Google Scholar Google Preview WorldCat Heim, Irene, and Kratzer, Angelika. (1998). Semantics in Generative Grammar. Malden, MA: Blackwell. Google Scholar Google Preview WorldCat COPAC Hellinger, Marlis, Bußmann, Hadumod, and Motschenbacher, Heiko. (2001–2015). Gender across Languages. 4 vols. Amsterdam: John Benjamins. Google Scholar Google Preview WorldCat COPAC Herbert, Cassie. (2017). ʻExclusionary Speech and Constructions of Communityʼ. PhD dissertation, Georgetown University. Google Scholar Google Preview WorldCat COPAC István, Fodor. (1959). ʻThe Origin of Grammatical Genderʼ. Lingua 8:1–41. Google Scholar WorldCat Jenkins, Katharine. (2016). ʻAmelioration and Inclusion: Gender Identity and the Concept of Womanʼ. Ethics 126:394–421.

Kramer, Ruth. (2020). ʻGrammatical Gender: A Close Look at Gender Assignment across Languagesʼ. Annual Review of Linguistics 6:45–66.

Google Scholar

WorldCat

Landau, Barbara, and Gleitman, Lila R. (1985). Language and Experience: Evidence from the Blind Child. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Lavietes, Matt. (2022). ʻProfessor Who Wouldnʼt Use Trans Studentʼs Pronouns Wins $400K Settlementʼ. NBC News, April 19. https://www.nbcnews.com/nbc-out/out-news/professor-wouldnt-use-trans-students-pronouns-wins-400k-settlementrcna24989. Maciuszek, Jozef, Polak, Mateusz, and Świa̧tkowska, Natalia. (2019). ʻGrammatical Gender Influences Semantic Categorization and Implicit Cognition in Polishʼ. Frontiers in Psychology 10:2208. Google Scholar WorldCat Maienborn, Claudia, von Heusinger, Klaus, and Portner, Paul, eds. (2011). Semantics: An International Handbook of Natural Language Meaning. Vol. 1. Berlin: De Gruyter. Google Scholar Google Preview WorldCat COPAC McConnell-Ginet, Sally. ([1979] 2011). ʻPronouns, Prototypes, and Personsʼ. In Gender, Sexuality, and Meaning: Linguistic Practice and Politics, 185–206. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC McConnell-Ginet, Sally. (2014). ʻGender and Its Relation to Sex: The Myth of “Natural” Genderʼ. In Greville G. Corbett, ed., The Expression of Gender, 3–38. Berlin: De Gruyter. Google Scholar Google Preview WorldCat COPAC Partee, Barbara H. (1989). ʻBinding Implicit Variables in Quantified Contextsʼ. Proceedings of the Chicago Linguistics Society 25:342–365. Google Scholar WorldCat Pinker, Steven. (1994). The Language Instinct. New York: William Morrow. Google Scholar Google Preview WorldCat COPAC Segal, Edward, and Boroditsky, Lera. (2011). ʻGrammar in Artʼ. Frontiers in Psychology 1:244. Google Scholar WorldCat Simons, Mandy. (2006). ʻFoundational Issues in Presuppositionʼ. Philosophy Compass 1:357–372. Google Scholar WorldCat Sudo, Yasutada. (2012). ʻOn the Semantics of Phi Features on Pronounsʼ. PhD dissertation, Massachusetts Institute of Technology. Google Scholar Google Preview WorldCat COPAC Swan, Oscar. (2015). ʻPolish Gender, Subgender, and Quasi-Genderʼ. Journal of Slavic Linguistics 23:83–122. Google Scholar WorldCat p. 292

Tonhauser, Judith, Beaver, David, Roberts, Craige, and Simons, Mandy. (2013). ʻToward a Taxonomy of Projective Contentʼ. Language 89:66–109. Google Scholar WorldCat Vigliocco, Gabriella, Vinson, David P., Paganelli, Federica, and Dworzynski, Katharina. (2005). ʻGrammatical Gender E ects on Cognition: Implications for Language Learning and Language Useʼ. Journal of Experimental Psychology: General 134:501–520. Google Scholar WorldCat Whorf, Benjamin L. (1956). Language, Thought, and Reality. Cambridge, MA: MIT Press. Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

13 How to Do Things with Gendered Words  E. M. Hernandez, Archie Crowley https://doi.org/10.1093/oxfordhb/9780192844118.013.36 Published: 22 May 2024

Pages 293–319

Abstract With increased visibility of trans people comes increased philosophical interest in gendered language. This chapter aims to look at the research on gendered language in analytic philosophy of language so far, which has focused on two concerns: (1) determining how to de ne gender terms like ‘man’ and ‘woman’ such that they are trans inclusive, and (2) if, or to what extent, we should use gendered language at all. We argue that the literature has focused too heavily on how gendered language can harm trans people and has not considered how trans people use gendered language to create meaning and joy for ourselves. Pulling from the literature in sociolinguistics, we look at examples of how trans people use language to make their lives better by gaining recognition, playing with gendered language, nding joy in gendered language, and taking control of de nitional power, concluding that debates about gendered language need to consider not only how such language harms trans people but also how trans people use it for our own liberation.

Keywords: transgender, feminism, gender and language, pronouns, philosophy of language, sociolinguistics Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

Why should a sequence of words be anything but a pleasure? —Gertrude Stein

1. Introduction Languages are like gardens. They are simultaneously cultivated and wild. We make decisions about how they are shaped, but they go in their own direction just as naturally. Gardens may cross-pollinate, invasive species may enter and take over, and we can transplant cuttings or full plants from one garden to another. It is through these cycles of contingencies and change that they seemingly develop both in our control while still wildly out of our control. We have the power to intervene in their natural growth, but it can be di

cult

and come at a cost. One such intervention into language has been the feminist intervention. In her survey of feminist philosophy of language, Mary Kate McGowan (2021) argues that one of the primary concerns has been feminist criticisms of language, such as the gender-neutral use of terms like ‘man’ and ‘men.’ Some feminists, according to McGowan, argue that terms like ‘man’ can’t be gender neutral; others contend that there is ambiguity between both a gender-speci c and gender-neutral form of ‘man,’ but that this ambiguity is one shaped by sexism. Such criticism leads to suggestions of how to alter our language, either through creating new terminology or through using our preexisting linguistic tools in di erent ways, such as substituting ‘human’ for ‘man.’ Altering our language to make it less sexist is often part of the feminist project. p. 294

Philosophers’ investigation into gendered language has been rmly rooted in this feminist tradition. As McGowan (2021) argues, feminist philosophers have put philosophy of language to use in understanding gendered slurs, generics, forms of silencing, subordinating speech, and hate speech. However, the literatures on slurs, generics, silencing, subordinating speech, and hate speech are so broad and signi cant that they warrant more attention than we can give them here. The focus of this chapter is on other forms of gendered language, speci cally as it relates to transfeminist issues. We look at two growing areas of research in analytic philosophy that speci cally concern trans people: (1) determining how to de ne gender terms like ‘man’ and ‘woman’ such that they are trans inclusive, and (2) if, or to what extent, we should use gendered language at all, given the harm of misgendering. But, one may wonder, what about other ways gendered language is used? How do trans people, ourselves, use gendered language to convey meaning, and is that all we do with it? Trans people are often incredibly creative with the terminology we use, coining new terms like ‘mapa,’ ‘dic-clit,’ and ‘titty skittles.’ Gendered language is often a tool of self-determination for trans people, and we nd inventive ways to use, not only our terminology, but our voices, grammar, and discursive styles to communicate and play with gender. Insofar as new literatures on gendered language are concerned with trans people, we nd there is something missing from philosophers’ scope of analysis. The chapter begins by looking at the two aforementioned areas of research on gendered language. We argue that the literature has focused too heavily on how gendered language can harm trans people and has not considered how trans people use gendered language to create meaning and joy for ourselves. Pulling from the literature in sociolinguistics, we look at examples of how trans people use language to make our lives better by gaining recognition, playing with gendered language, nding joy in gendered language, and taking control of de nitional power. Philosophical debates about gendered language need to consider not only how such language harms trans people but also how trans people use it for our own liberation. We conclude by re-examining these two debates in light of the way trans people use language and suggest new areas of research.

2. The Story So Far In her landmark article “Sexism,” Marilyn Frye (1983) draws our attention to how deeply gender is encoded in English (17–40). The gendered aspects of many languages force us to constantly communicate what gender/sex we and others are, making gender/sex appear far more relevant than necessary. On Frye’s view, sexism involves marking everyone in one (and only one) of “the two sexes,” so it is of vital importance that we can tell who is who, and clearly communicate that with our language. Both of the literatures on gendered language that we are looking at are rooted in these observations. The rst primarily concerns how we t people into the de nitions of ‘woman’ and ‘man,’ and the second concerns how to respond to the overwhelmingly gendered aspects of language. p. 295

The rst literature nds its contemporary form in Jennifer Saul’s (2012) observation that terms like ‘woman’ serve as both sex terms and gender terms. That is, ‘woman’ functions in some contexts as a biological term picking out things like certain hormone levels, XX chromosomes, and vaginas, and in other contexts, picking out social roles in a hierarchical society. In response, Saul suggests a contextualist 1

analysis of the term ‘woman’ (201). Unfortunately, Saul argues that her analysis excludes trans women. Take, for example, a trans woman, call her Carla, who has not had vaginoplasty. On Saul’s contextualist analysis, if Carla is in a social context where she needs to use a changing room, then the relevant standard is her self-identi cation as a woman, and thus the sentence “Carla is a woman” is true. However, if Carla is in a medical context where she may need a pap smear or to be screened for testicular cancer, the sentence “Carla is a woman” is false. In this case, Carla is not “relevantly similar” to cisgender women since the presence of a cervix is the relevant standard. This is a problem because it trivializes trans women’s claim to being women (213). Importantly, Saul believes this de nition fails on moral and political grounds, not on semantic grounds (204). The de nition does not fail because it fails to track the di erence in ‘woman’ as a sex term and gender term, but because of moral and political reasons to not marginalize trans women. Therefore, Saul concludes terms like ‘woman’ and ‘man’ are politically signi cant terms because understanding their proper use involves not just a semantic analysis but a political one as well. This problem raises a methodological issue for analytic philosophy of language since it treats semantic analysis as distinct from these other concerns. For this reason, Saul does not believe contextualism can solve this issue, because what we need is a way to account for how political concerns a ect the semantics. Talia Mae Bettcher (2013b) o ers one solution to this problem: the multiple-meaning view. According to Bettcher, we need a de nition of ‘woman’ that both includes trans women and does not trivialize their standing as women. Her solution is to say that there are two de nitions of ‘woman’ at play, and importantly, they’re rooted in di erent communities (240). First, there’s the dominant de nition of ‘woman’ that tracks how the term is used in dominant communities, which are often oppressive to trans women. Second, there’s what she calls the resistant de nition of ‘woman,’ which follows from how trans women have rede ned the term in trans subcultures. Both de nitions are rooted in di ering metaphysics, with the dominant de nition rooted in a (false) view of gender (i.e., gender is dimorphic and tracks genitalia) that is harmful and oppressive to trans people (243). Therefore, we can say that the resistant de nition is the better, more accurate de nition. This helps explain how the term ‘woman’ may function as a sex term in speci c contexts, but also why trans women can reject that de nition of ‘woman’ because it is founded on misconceptions about gender and sex in the dominant society. Bettcher’s solution, then, takes on Saul’s suggestion to look beyond the contextualist project to get us out of the methodological problem. Esa Díaz-León (2016), however, is concerned that the multiple-meaning view accepts the existence of exclusionary concepts, namely, the dominant de nition of ‘woman.’ Díaz-León believes that to resolve the p. 296

concern about politically signi cant terms,

we need to rid ourselves of exclusionary concepts. Toward

that end, she argues for a revised version of Saul’s contextualism that gets rid of these concepts. According to Saul’s contextualism, what will determine if ‘X is a woman’ is true are “the standards that are relevant in that context” (2016, 249). Examples of such standards are what genitalia someone has (for medical contexts) or what someone’s gender identity is (for other social contexts). Díaz-León claims that this is where political and moral concerns become relevant for the meaning of terms like ‘woman,’ since there may 2

be di ering views on what standards are relevant in any particular context. Questions about which standards are relevant for the context will become crucial, and the moral and political considerations will determine, in part, how to adjudicate which standards to take up. On this view, political considerations, such as trans women’s legitimate claim to womanhood, can a ect the truth conditions of a sentence because those political considerations count as the “objective” features of X’s context. In all cases where a trans woman makes the claim “I am a woman,” then the fact that she identi es as a woman will be relevant for determining the standards relevant to her context, making the sentence true. A number of philosophers, including Bettcher (2017), Laskowski (2020), and Chen (2021), argue that DíazLeón’s (2016) contextualism is actually a form of invariantism, where the de nition does not actually shift from context to context. As Hsiang-Yun Chen (2021) puts it, “one might claim that misgendering can never be accepted in any context and that genuine self-identi cation is the most valuable consideration in all contexts,” making it such that the de nition of ‘woman’ never shifts (584). If there is no context where “Carla is a woman” is false, then the view collapses into invariantism. The literature on politically signi cant terms is a debate over how to render gender terms trans-inclusive. The contemporary literature begins with Saul, Bettcher, and Díaz-León, however, many other philosophers 3

have also weighed in on how to resolve this problem. The common theme that runs through this literature is that philosophers nd it important to include trans people and avoid harming us when de ning gender terms. The political project of resolving the problem of politically signi cant terms is the acceptance and inclusion of trans people in mainstream, or dominant, de nitions, and trying to prevent further marginalization. This brings us to the second literature on gendered language, which focuses on how gendered our language should be. One key to this debate is understanding how gendered language can harm trans people. Stephanie Kapusta (2016) argues that trans people are vulnerable to a linguistic form of moral harm via gender-term deployments, or what is known more broadly as misgendering (502). According to Kapusta, misgendering involves using “gender terms that exclude transgender women from the category woman, or that hierarchize that category in a way that marginalizes transgender women” (503). A trans woman, for example, can be misgendered when someone uses ‘he/him/his’ pronouns for her and by theories (like Saul’s contextualist analysis) that marginalize trans women. In recognizing how gender-term deployments can cause moral harm, the ethics of using gendered language has become a major focus. Robin Dembro p. 297

and Daniel Wodak (2018) argue that the harms of misgendering a ect not only trans

women and men, but genderqueer and nonbinary people as

well. Their analysis of misgendering builds

on Kapusta’s (2016), locating the harm of misgendering in denying someone’s gender. On their view, it is not a moral duty to a rm trans people’s genders; it is only a moral duty to avoid denying their gender (Dembro

and Wodak 2018, 383–384). This distinction is important because it means that using gender-

neutral pronouns (e.g., ‘they/them/theirs’) for someone cannot constitute misgendering. Dembro

and

Wodak go on to argue that not only is it morally permissible to use gender-neutral pronouns for everyone, we have a moral duty to avoid using gender-speci c pronouns for anyone. They argue that using genderneutral pronouns for everyone is the most pragmatic solution, because it protects the privacy of queer and 4

trans people and ghts gender essentialism. While they do not hold this as an exceptionless duty, they do believe the few exceptions do not defeat the moral duty (387).

Dembro

and Wodak (2021) further develop this view, arguing that it is not only gendered pronouns we

should refrain from using, but also gendered honori cs, generics, and su

xes. They argue that English

should not be any more gendered than it currently is raced, and since English currently does not have racespeci c pronouns, honori cs, generics, or su

xes, we should not have gender-speci c ones either. Again,

this conclusion follows from how these gendered (and raced) terms harm marginalized people. They argue that these gendered and raced terms stigmatize and stereotype social groups, exclude people who don’t t the gender binary of man–woman (and race “binary” of white–black), and present privacy issues by forcing people to disclose gender and race details about themselves or their partners. E. M. Hernandez (2021) argues that we should be skeptical of the conclusion that we ought to make English more gender neutral. While agreeing with Dembro

and Wodak (2018) that the harm of misgendering is due

to denying someone’s gender rather than failing to a

rm that gender, e argues that gender a

has positive moral value. This positive moral value of gender a moral reasons to gender-a

rmation still

rmation makes it so that we often have

rm trans people. Furthermore, subversive uses of gendered language (e.g.,

using gender-neutral pronouns for cis people and preferred pronouns for trans people) can challenge gender essentialism as well as using gender-neutral pronouns for everyone can. Hernandez, therefore, rejects the conclusion that there is a moral duty to use gender-neutral pronouns for everyone. E remains agnostic on whether we should have the long-term goal of gender-neutralizing English, since what linguistic practices would be present in a society with gender equality is an open question. Finally, Quill Kukla and Mark Norris Lance (2023) argue for a pragmatics of gender ascriptions that cuts across both of these literatures. On their view, gender ascriptions such as “Carla is a woman” or gendered pronouns help alter and (re)organize social space. When we call someone a woman, we place that person in social space to be treated as a woman. Importantly, such ascriptions don’t metaphysically constitute someone’s status as a particular gender. Disagreements about sentences like “Carla is a woman,” then, are negotiations or disputes about how someone should t into a normative social space, and not part of a debate about some antecedent truth about Carla’s womanhood. This view rejects the underlying semantic p. 298

project that Saul (2012), Bettcher (2013b),

and Díaz-León (2016) are engaged in by arguing that

sentences like ‘X is a woman’ are not declaratives but primarily performative utterances, while also rejecting Dembro

and Wodak’s arguments (2018, 2021) by illustrating how using someone’s self-

identi ed gender terminology is both ethically and pragmatically correct. What is noticeable about these two debates is how they focus on the harm gendered language does to trans people. The rst is primarily motivated by including trans women and men on the same semantic footing as cis women and men, and the second is focused on how we should use gendered language, and both projects are aimed at avoiding misgendering trans people. Even Hernandez’s (2021) project, while focusing on morally good ways to use gendered language, is still primarily focused on how people talk to and about trans people as opposed to how trans people talk to others and about themselves. This trend, we believe, mirrors one noted by Kukla (2018) about the language of sexual negotiation. Kukla argues that philosophers’ focus on consent has made them miss other aspects of the ethics and pragmatics of sexual negotiation. By becoming so focused on how to avoid wronging people in sexual contexts, they have failed to think about how to promote good sex. In e ect, philosophers’ focus on the harms of failed sexual negotiation distorts what problems are at issue. Although we do not think these two debates have distorted the issues faced by trans people, we do believe this focus has been narrow. These debates are primarily about how those in power should use linguistic resources to include trans people in their worlds. The whole study of gendered language, as it relates to trans people, becomes a study of how to include. We do not think this is unimportant, but it ignores how trans people use gendered language ourselves. If we ip the focus of analysis onto trans people, we nd that gendered language can also be a tool for disruption, a way to undercut the harms trans people face. Trans people do all sorts of things with gendered language,

and turning our attention to how we use gendered language to make our lives better not only complicates these current debates but opens entirely new avenues of study.

3. Transing Language To bring trans uses of gendered language into conversation with these debates, we will pull from the sociolinguistics literature that has been documenting and analyzing trans and gender-nonconforming people’s linguistic practices. Emerging work within a trans linguistic framework centers the work and analytic lenses of trans researchers and fundamentally works to “uplift collective movements of transgender joy and liberation” (Konnelly 2021, 79; see also Zimman 2020). We draw examples from this body of research to demonstrate the ways in which transgender and gender-nonconforming people use elements of gendered language to create moments of joy and gender a

rmation. Because we focus on

providing examples from global gender-diverse communities, our philosophical analysis of these various uses of language will be fairly cursory. Our aim is not to develop a full philosophical or pragmatic theory of p. 299

how trans

people use language for their betterment, but to shine a light on areas of research that deserve

more attention. Language and gender are deeply interconnected. Every level of language is potentially lled with gendered meanings and implications, from the sounds of spoken languages to grammatical structures to larger stylistic features of language that communicate the speaker’s gender or index gendered meanings. Philosophical debates on gendered language have primarily focused on gendered lexical items (e.g., ‘man,’ ‘woman,’ ‘he,’ ‘she’) because philosophers of language, in the analytic tradition, are primarily interested in referential meaning, or how language refers to objects in the world. However, we believe that philosophers should be aware of the other kinds and locations of gendered meaning in language in order to address the 5

questions about gendered language philosophers have taken up. First, many of the ways trans people use gendered language to create understanding, recognition, joy, and a

rmation occur on these other levels of

language that often get overlooked by philosophers. Second, if we are concerned about language being too gendered, we shouldn’t look simply at one level of analysis, but should see other ways gender is communicated—especially when it is concealed and often goes unnoticed. Opening the scope of analysis beyond referential meaning means that we’re not solely concerned with terms that pick out speci cally gendered referents, but also with the things trans people do with language. Therefore, we’re following a well-known intervention into philosophy of language made by J. L. Austin (1975) in his collection of lectures How to Do Things with Words, wherein he demonstrates that language does not just represent the world; rather, utterances can be actions that have various e ects on the world. Language can be used to put people in obligation to one another, marry one another, and place one another in social positions (cf. Austin 1975; Kukla and Lance, 2023). Notably, Judith Butler’s (1990) work on gender performativity draws on Austin to explore how “gender is … instituted through a stylized repetition of [habitual] acts” (179). Gendered identities, then, are expressed and reinforced through performative acts. Like Austin, Butler stresses the ways in which performatives have e ects in the social landscape. In attempting to look beyond how gendered language can harm trans people, we explore how trans people do things with gendered words. Our exploration of gendered language therefore makes two departures from the current literature on the topic. The rst departure we make is to look beyond the impact of referential meaning and toward how trans people do things with language. The second departure is to go beyond the gendered features of the lexical level of language to explore the other levels on which language can be used to communicate gendered identities.

Sociolinguists have explored the ways in which various levels of language have gained gendered 6

associations through the semiotic process of indexicality. Social indexical meaning can potentially be tied to all levels of language: the phonetic, morpho-syntactic, lexical, and discursive levels. The phonetic level focuses on the sounds of spoken languages (and equivalent features of signed languages). Phonetic features, such as pitch, often index the gender of the speaker. For example, in many English-speaking contexts, p. 300

higher-pitch voices are associated with women/femininity and lower-pitch voices are

associated with

men/masculinity. Further, these gendered linguistic associations are why voices in the middle range are often perceived as “androgynous” by listeners (Azul 2013). Gender and language are also intertwined through grammatical forms and sentence structures, on what is called the morpho-syntactic level. For example, the use of tag questions (‘That’s the room, right?’ or ‘You’re coming, aren’t you?’) is often associated with women (Lako

1973). In addition to these smaller units of language, there are larger units

of language that are also part of how we do and communicate gender; these occur on the discursive level. One example is how di erent speech styles convey gendered identities, such as how people perceive “gossip” versus “locker room talk” as gendered styles of interaction. Although both generally function to share ingroup information and build camaraderie, “gossip” is typically perceived as “feminine,” and “locker room talk” is perceived as “masculine.” In our examples, we draw from these various levels of language to explore how gendered linguistic features function as a resource for trans, nonbinary, and gendernonconforming language users. We take the things trans people do with gendered words to fall into three overlapping (and likely not comprehensive) categories: recognition use, playful use, and joyous use. When we say that these are the things trans people do with language, we mean that these are the ways trans people use language to accomplish things for themselves. There is another sense in which trans people do things with language, such as rejecting certain terms or forms of speech (e.g., ‘biologically male’), replacing old terms with new ones (e.g., 7

‘niece/nephew’ with ‘nibling’), and appropriating/reclaiming forms of language (e.g., creaky voice, and even the term ‘transgender’ itself). These uses of language are often for the sake of accomplishing one or more of the other things we take trans people to do with language.

3.1. Recognition Use One of the most basic things trans people do with gendered language is use it to gain recognition. It is worth noting that these are the kinds of concerns that Dembro

and Wodak (2018, 2021) count in favor of

gendered language. Gendered language can be used to gain recognition for gender identi cation, which is necessary for naming systems of oppression, but also can aid in accessing resources and making ourselves intelligible to others. As Butler (2004) notes, To nd that you are fundamentally unintelligible (indeed, that the laws of culture and of language nd you to be an impossibility) is to nd that you have not yet achieved access to the human, to nd yourself speaking only and always as if you were human, but with the sense that you are not, to nd that your language is hollow, that no recognition is forthcoming because the norms by which recognition takes place are not in your favor (30). So, although this category is somewhat obvious, it is worth attending to because the importance of recognition cannot be overstated. p. 301

The most straightforward example of using gendered language for recognition use is by adding terminology to ll hermeneutical gaps. One form of harm trans people face is what Miranda Fricker (2007) calls “hermeneutic injustice”, “the injustice of having some signi cant area of one’s social experiences obscured from collective understanding owing to a structural identity prejudice in the collective hermeneutical resource” (155). Most of us are raised knowing only the binary genders of man and woman, as both

exclusive and exhaustive, and this makes up the collective hermeneutical scheme. Creating terms like ‘genderqueer’ and ‘nonbinary’ helps give people the terminology to de ne themselves and to understand others who are outside of the gender binary. Furthermore, adjectives like ‘trans’ and ‘cis’ help delineate what we may call di erent modalities of gender (i.e., whether one’s gender identity aligns with their gender assigned at birth) (Ashley 2022). But this is not the only way in which trans people use language to gain recognition. The uptake of gendered forms of language for recognition use is highly apparent in communities whose languages have grammatical gender-marked morphologically. Orit Bershtling’s (2014) work demonstrated the linguistic practices of a community of genderqueer speakers of Hebrew, whose robust morpho-syntactic gender-agreement system requires grammatical gender agreement between nouns, verbs, and adjectives via in ectional morphology. Although this stricter linguistic system poses challenges for genderqueer users of Hebrew, Bershtling’s ethnographic interviews with a community of genderqueer speakers of Hebrew demonstrated how participants would selectively alternate between masculine- and feminine-gendered morphology to articulate their genderqueer identities as uid, changeable, and emergent. These genderqueer individuals “[use] the available linguistic resources to lead change and instruct others in their surroundings about alternative gender possibilities” (58). Similar trends have been found in work on nonbinary French (Knisely 2020; Tudisco 2021). This work has traced the wide array of linguistic strategies that nonbinary French users employ to navigate gendered morphology. While the “o

cial” French

language is closely regulated by the Académie Française, nonbinary individuals in both online and inperson communities nd ways to subvert grammatical rules by alternating gendered morphology and “the nonnormative use of binary grammatical gender to index non-binary identities” (Tudisco 2021, 1). Through the strategy of using gendered morphology in nonnormative ways and outing expected grammatical agreement rules, nonbinary users of these languages make their identities salient to those around them, challenging assumptions about the relationship between grammatical and social gender and rejecting the imposition of a linguistic gender binary. Another example of how communities take up and recontextualize terminology for recognition use is provided in Jenny Davis’s (2014) research with various Two-Spirit community groups. Indigenous peoples take up localized gender terminology from their respective tribal a

liations in conjunction with

terminology such as ‘gay,’ ‘queer,’ and ‘transgender.’ For example, some people use both ‘transgender’ or ‘queer’ and ‘Two-Spirit’ to describe themselves; others use portmanteaus such as ‘indigiqueer.’ This p. 302

functions to locate contemporary Two-Spirit communities within both local

and traditional

understandings of gender and sexuality. Their combination of local, tribal-speci c terminology with more generalized concepts were tools that these community members used to articulate the “both/and approach to gender and sexuality, and to femaleness and maleness [of Two-Spirit people]” (63). It is noteworthy that by appropriating English gender and sexuality terms to expand their hermeneutical resources, Two-Spirit communities are creating legibility both outside and inside their indigenous communities (Davis 2014). Furthermore, gender-diverse communities are using gendered linguistic resources to construct and recognize their identities even in linguistic communities where there are no gendered pronouns. In ethnographic interviews with Indonesian tombois and their girlfriends in the lesbi community, Evelyn Blackwood (2014) explores how community members eschew the gender-neutral pronouns of Indonesian and would instead have others refer to the tombois using either their names or the masculine pronoun from the Mingangkabu language (‘wa’ang’), or refer to them as ‘cowok’ (‘guy’). These strategies functioned as a way to highlight the tombois’ masculine gender. Even while using language to reinforce the tombois’ masculinity, the tombois and their girlfriends subvert normative expectations that the girlfriends would partner with men with whom they could have children, challenging community cultural beliefs about both gender and family structure. What’s particularly noteworthy about this usage is that Indonesian tombois are

lling a hermeneutical gap in their context, not by creating new terminology for unrecognized concepts, but by swapping or appropriating terms to mark unrecognized concepts. The previous examples demonstrate how gender-diverse communities use gendered linguistic features such as labels, pronouns, and morphology to construct and express their identities to their communities and to challenge normative gender expectations. But using gendered linguistic features for recognition use is not limited to just taking up various gendered terms; within trans linguistic research, sociophoneticians have looked at what trans and gender-diverse people are doing creatively with their voices. There are various features of the voice that have strong ties to gendered perceptions. Although pitch is typically the most salient, another vocal feature, ‘creaky voice,’ is strongly associated with young women, and a fronted pronunciation of the phoneme /s/ is often perceived as feminine (Calder 2019; CampbellKibler 2011; Zimman 2013). Therefore, normative gender expectations might assume that there would be some sort of alignment between these gendered features. However, Lal Zimman’s (2017a) research shows how some trans masculine people engage in what he calls “stylistic bricolage,” whereby queer trans men use features that simultaneously index masculinity and femininity. In this example, we can see how the gendered linguistic features function together as tools through which trans individuals can creatively express their gender. Similarly, trans feminine speakers often go through voice feminization training, not only to raise the pitch of their voices, but also to make them sound “brighter” and more stereotypically feminine, thus using gendered linguistic features as tools for social identity recognition (Cli ord 2019; English 2021). p. 303

These various features of language serve as ways for trans people to gain recognition. But how does language accomplish this? We nd one clue by looking at what Kukla and Lance (2023) say about gender ascriptions. They argue that gender ascriptions are illocutionary acts that place one in social space. Uttering the sentence “we are nonbinary” places us (the authors) in social space as nonbinary people; or, if we say “the actress Hunter Schafer is a woman,” we are placing her in social space as a woman. Notably, the felicity conditions for these two utterances are di erent. The rst-person gender ascription that we are nonbinary functions as a request, or a demand, that we be taken as nonbinary, which then requires uptake from a second party to succeed. A third-person gender ascription like “Hunter Schafer is a woman” functions di erently, because it cannot be seen as a petition or request since Hunter herself is not making the utterance. According to Kukla and Lance, we need Hunter’s uptake for the ascription to succeed—and we will happily wait to hear from her—because we do not have the relevant ethical authority to place anyone other than ourselves into gendered social space. We suggest that not only do straightforward utterances like “we are nonbinary” or “Hunter Schafer is a woman” count as illocutionary acts, but that the other features of language noted in this section may count as well. For example, selectively alternating between masculine- and feminine-gendered morphology in Hebrew is a way to place oneself as genderqueer in social space. Similarly, trans women raising the pitch of their voices and making themselves sound “brighter” is a way to place themselves as women in social space. What is notable about using linguistic grammar and phonetics as illocutionary acts is that they are covert. Kent Bach and Robert Harnish (1979) argue that some illocutionary acts are covert since the intention of the act need not be recognized by the parties perceiving the utterance. Whereas uttering “we are nonbinary” clearly communicates intention, using morpho-syntactic or phonetic aspects of language to place oneself in social space is covert because the intention not only does not need to be recognized to succeed, in cases like 8

trans women feminizing their voice, the utterer may not want anyone to notice the intention.

Recognition use is fairly broad and serves as a basis for much of what trans people do with language. Not only does it play a role in altering referential meaning (by lling hermeneutical gaps), such uses can function as illocutionary acts to place oneself in social space. As we will see in the next sections, recognition use underlies a lot of the other things trans people do with language.

3.2. Playful Use Trans people often use gendered language in playful ways. Playfulness, as an attitude, is an important part of decolonial feminist practices. María Lugones (2003) characterizes playfulness as an “attitude that carries us through [an] activity … [and] turns that activity into play,” and that involves an openness to surprise and doesn’t presuppose rules for how to play (95–96). For Lugones, this attitude of playfulness is important in p. 304

part because of how it helps marginalized people explore the worlds of others who have power

over them

without being “taken in” by those perspectives. When you face marginalization, it is too easy to be in a context where someone else’s perception of you becomes how you perceive yourself. Being playful, for Lugones, involves not placing yourself rmly in any one world or context, leaving you open to selfconstruction, a lightheartedness to create opportunities for growth without becoming trapped (96). This attitude is protective because keeping yourself from being rooted in a dominant world or context helps you maintain a part of yourself that is more authentic while also exploring how you’re perceived in these more hostile worlds. This aids in self-construction by allowing you to learn things about yourself—that is, how your attributes even in resistant worlds are born out of experiences in the dominant world. Engaging in this re ective work with a playful attitude can help create the space necessary to understand the dominant world without being taken in by it. Playful language can involve (a) taking up this attitude when de ning and creating terminology or (b) using creative terminology to cultivate this attitude. What makes (a) valuable is how the playful space allows for altering language in a way that can subvert gendered norms and expectations. The ways in which trans people often use gendered language exist in this playful space, where things are not taken so seriously as to inform o

cial policy proposals, while still having the language a ect the linguistic landscape in trans

communities. Whereas what makes (b) valuable are the ways in which it helps foster playfulness, as Lugones explores. Ultimately, navigating dominant institutions and contexts can be stressful and even harmful for trans people, so having ways to cultivate playfulness can ease the di

culty of engaging in these

worlds and create the necessary space to protect oneself from various forms of psychological harm. One context in which the playful use of language can be seen is the terminology that trans people use to talk about our bodies. Some trans people choose to use gender-neutral terms for body parts, such as ‘genitals’ or ‘chest’; others take up typically gendered terms for body parts in ways that challenge the essentialist linking between physiological traits and certain gendered identities, a practice which serves to highlight the discursive construction of sex (Zimman 2014). This primarily happens through the recontextualization of typically gendered/sexed body parts (e.g., ‘dick,’ ‘clit,’ ‘cock’). For example, despite not having plastic surgery to alter their genitals, some trans masculine people refer to their genitals as ‘dicks’ and ‘cocks,’ and some trans feminine people refer to their genitals as ‘clits.’ Beyond recontextualizing existing terminology for body parts, a common playful move within trans communities is to create portmanteaus of gendered terms to develop new words to refer to genitals. One newer innovation is the creation of the su

x ‘-pussy’:

with trans masculine people saying they have a ‘boypussy’ (Zimman 2014) or trans feminine people referring to surgically enhanced genitals as their ‘robopussy’. In trans masculine communities in the 1990s, the term ‘dic-clit’ was in wide circulation, a portmanteau of dick and clitoris, though it has now generally declined in circulation (Zimman 2014). Furthermore, as Elijah Edelman and Lal Zimman (2014) explored, trans men looking to have sex with men used a wide array of body part terminology—including the term ‘bonus hole’ to refer to their genitals. This naming frames the trans body as extra desirable because of this p. 305

“bonus” which

“o er[ed] sexual value not otherwise provided in a typical homoerotic exchange between

men” (686). This insight has also been explored by Talia Mae Bettcher (2013a) in her work on interpretive intimacy about how trans women can alter the social meaning of their bodies during sex. It is noteworthy that Bettcher considers such activities to be playful, stating that “within a sexual context, there is an element of playfulness that opens the doors of possibility, paying less attention to the constraints of social reality” (55). Ultimately, all of these examples of language play give trans people discursive agency over

their bodies, with which many trans people often have a fraught relationship. This kind of language play allows new and creative ways for trans people to connect to their bodies. This form of playful use is also accomplished by appropriating/reclaiming various gender terms. For example, the terms ‘transsexual,’ ‘biological woman,’ and ‘autogynephile’ have all been used in various ways to marginalize and oppress trans women (Serano 2020). All of these terms have been appropriated by trans people, using them as playful in-community jokes to reveal how shallow trans-exclusionary positions are. One famous example of this is Natalie Wynn, a trans woman YouTuber who goes by “Contrapoints,” who regularly calls herself a “biological woman” to poke fun at trans-exclusionary arguments she encounters. Part of the appropriation rests on the fact that many trans people, through medical interventions like hormone replacement, actually do t some biological de nitions of ‘woman’ and ‘man,’ undermining trans-exclusionary arguments that trans women are di erent from “real” biological women. Medicalized terminology is another space where trans people play with language. Many trans people spend a signi cant amount of time navigating medical institutions to alleviate gender dysphoria and promote 9

gender euphoria. Such institutions are sites of stress and even harm for trans people, and by playing with language, they cultivate a playful attitude that makes navigating them manageable. One example of how they do this can be seen in the way they describe and de ne various medical interventions. Florence Ashley (2020) notes the variety of names trans feminine people use to refer to hormone replacement therapy, including “titty pills, titty skittles, smartitties, chicklets, anticistamines, mammary mints, life savers, tit tacs, breast mints, femme&m’s, antiboyotics, trans-mission uid, and the Notorious H.R.T.” (75). Similarly, trans masculine people often refer to top surgery as ‘titty chop.’ And both trans feminine and trans masculine people refer postsurgery to their ‘designer’ bodies, for example, as having ‘designer nipples’ or a ‘designer vagina,’ and may refer to their surgeon as the “designer.” This playful language shifts the doctor-patient role to one of designer and client, providing trans people with a way of conceptualizing gender-a

rmative care as nonpathologized.

When someone uses language with a playful attitude, the attitude aids in executing speech acts that do not require the speaker to have any particular authority, what Mary Kate McGowan (2019) calls covert exercitives. Standard exercitives require the utterer to have the relevant authority to enact the speech act (e.g., a baseball umpire having the authority to call a runner “out”). McGowan argues that there are also exercitives p. 306

that don’t require this authority; instead, such exercitives are moves in a game that

enact permissibility

facts, a kind of norm, that dictate what one can then go on to do in some norm-governed activity. For example, a speaker may say something insulting about a mutual acquaintance in a conversation, enacting a permissibility fact that it is okay to speak negatively about that mutual acquaintance in that conversation (of course, a second speaker may disagree with the initial insult, thereby canceling the covert exercitive and making it impermissible to continue speaking negatively about the mutual acquaintance). Trans people often do not have the relevant kind of authority to initiate changes to norms in the dominant society, and furthermore, in many cases of playful use of gendered language, they do not want to alter certain permissibility facts (e.g., implement policy or make it permissible for cis people to call someone a ‘transsexual’ or to talk about our boypussies). Covert exercitives therefore function to change permissibility facts within our own subcommunities, altering how we talk about and understand gender. Rowan Bell (forthcoming), for example, analyzes how members of US Ballroom culture (a community of trans people of color) repurpose and create gender norms to be legible to one another and to foster a safe space to be openly queer. This is a phenomenon in many trans subcommunities, and it is often accomplished through the playful use of gendered language. Something notable about these covert exercitives is that they do require some level of uptake from community members. That’s because the norm-governed activity they are altering is a playful one that involves acceptance from community members. For example, if someone does not understand that play is occurring, that person may misunderstand the use and cancel it, like the second speaker in McGowan’s example, who cancels the covert exercitive of insulting the mutual acquaintance.

What is interesting about these playful language games is that they are structured di erently from the norm-governed activities that McGowan (2019) covers. McGowan argues that norm-governed activities have two sets of norms: the norms that govern the activity, which do not change, or g-norms, and the norms that are enacted by covert exercitives, which alter what moves one makes in the activity, or s-norms. Playful language games, though, may also alter g-norms. Lugones (2003) argues that playful activities do not presuppose speci c rules (95). Following Lugones, we want to suggest that the governing norms for the activity can be altered based on moves within the game itself. G-norms declaring that certain moves are not acceptable can similarly be altered, reconstituting the activity itself, which, importantly, leaves us open to surprise and allows us, as Lugones has argued, to remain lighthearted and open to self-construction. There are interesting overlaps between recognition use and playful use. All the examples we have explored so far have a playful quality either because they are rooted in a playful attitude or are a way to cultivate a playful attitude. But many of these playful uses are also recognition uses. In using language to gain better recognition and legibility, invented terms like ‘bonus hole’ create legibility for trans men in homoerotic 10

exchanges with other men. p. 307

Appropriations of terms like ‘biological woman’ by trans women also help

draw attention to how feminists after the second wave reject the sex-gender

distinction presupposed by

trans-exclusionary writers (Fausto-Sterling 2008; Stryker 2017). Furthermore, mixing gendered morphology, as in the examples of genderqueer Hebrew and nonbinary French, is often also a playful activity for trans people who want to disrupt gendered language. There is an important di

culty to note with playful use. There is not one resistant world, or trans

community, and there are many trans people who exist outside of such trans communities. These playful forms of language use often succeed given how they are understood or taken up in trans communities. For those who exist outside a trans community, it will be di

cult to play with language because it often fails to

be recognized as playful in the dominant world. (We will say more about the interplay between the dominant world and resistant worlds when we discuss de nitional power in section 3.3.) As Lugones (2003) mentions, in the dominant world, speci cally in White-Anglo feminist spaces, people take her to be extremely serious, not because she acts di erently but because that world does not allow for a Latina to be anything but serious (93). Lugones refers to this position as lacking health in those worlds; this is the same lack of health that interrupts a trans person’s attempts to playfully engage with language outside trans communities. Playful use is also not a valuable tool for trans people who are primarily concerned with making themselves legible to mainstream society, which is sometimes true for trans women and men. For some trans people, there is so much distress centered around their gender ascription being taken up that their playful use of gendered language runs the risk that they will not be taken seriously in mainstream society as a woman or a man. This is, notably, not a problem with gendered language but with the world itself. Trans people have to navigate preexisting tensions in the world to survive, and this di ers for di erent trans people based on our connection to community, economic and social resources, and preferences. It is a tension worth recognizing because it does limit playful use, but the limitation doesn’t come from gendered language, but gendered oppression more directly.

3.3. Joyous Use While playful use of language provides an opportunity for trans people to playfully recon gure binary categories and medical structures, gendered language can also be used to create trans joy and other positive a ects that help trans people ght against our marginalization. What makes this joy signi cant, in part, are the conditions in which trans people nd joy. As Talia Mae Bettcher (2019) puts it, “We trans people live an ‘everyday’ shot through with perplexity, shot through with WTF questions. We live in the WTF” (651; emphasis Bettcher’s). She elaborates,

What does it mean to say that I’m a woman, I’ve wondered? And why does so much [emphasis Bettcher’s] appear to hinge on it? How do I make sense, for example, of being assaulted in the p. 308

middle of Santa Monica Boulevard by someone who wanted

to prove I was really a man? Why do

people want to kills us? WTF? … the WTF is so all-embracing, so personal, indeed, existential in nature. (651–652) The WTF is practically the trans condition. Trans journeys are guided by the pursuit of making sense of how to t into a world that not only doesn’t understand us, but actively, violently, tries to keep us from living in it. The WTF can be taken as an example of what Sandra Bartky (1990) calls psychological oppression, which refers not to the social barriers that inhibit people’s movement in the world but to the psychological barriers that “weigh” one down and promote feelings of inferiority (11). Staggering rates of depression and suicide among trans people speaks to how signi cantly the WTF questions weigh us down (Toomey, Syvertsen, and Shramko 2018). Finding joy under such conditions is signi cant because the joy keeps us alive. One of the ways trans people nd joy is in nding gendered language that resonates with us. Part of the trans journey is nding the language that accurately describes how a trans person experiences gender. Often, for trans and gender-diverse people, one part of the process of gender discovery is coming to nd what language “feels good.” For many people, nding the “right” words that a

rm an individual’s

gender is an exciting process that can be coupled with experiences of gender euphoria. While gender euphoria is often related to physical experiences, it can also come from being referred to by one’s proper name, pronouns, or terminology (Beischel et al. 2022; Hernandez 2021). Furthermore, experiencing gender euphoria from stereotypical gendered linguistic features is not limited to trans men and trans women. For example, some nonbinary people feel gender euphoria when using or playing o example, some nonbinary people might use ‘boyfriend’ and ‘girlfriend’ or feel a

“gendered” terms. For rmed when they are called

‘sir’ or ‘ma’am.’ For many, this experience occurs when this language challenges the typical gendered 11

language they might be ascribed.

Unlike recognition use, which often involves illocutionary speech acts, and playful use, which often involves covert exercitives, joyous use is not accomplished through a particular set of speech acts. Whether a language user experiences joy in gendered language largely depends on their own history with, and exploration of, gendered language. To extend an analogy David Lewis (1983) makes between conversations and baseball games, there is nothing about the various moves in baseball that necessarily create joy in the players’ experience of it. It’s easy to imagine a child who hates baseball, but whose parent forces them to play because it is the parent’s favorite sport, so every time the child is in the out eld they’re excruciatingly 12

bored and annoyed.

The joy some of us feel in engaging in baseball (or any other sport) comes from the joy

that we take in the activity; it is in one sense a side e ect of the various actions that involve playing the sport. But notice also that the fun derived from playing sports is also a su

cient reason to play the sport in

the rst place. The same is true for the joyous use of gendered language. Joy is not an action accomplished by speech, but it is the result of certain engagements with language that make the engagement worth doing in the rst place. The joy that a trans person derives from using gendered language may be a stronger p. 309

motivator for using

gendered language than to gain recognition or to play with non-normative sense of

gendered speech. Joyous use overlaps signi cantly with recognition use and playful use. Given the WTF, nding the terminology that accurately captures your gender experiences can pull double-duty. It not only opens up the possibility for more resources and legibility, but it can also bring a kind of clari cation that promotes feelings of euphoria and joy (a feeling of joy from clari cation that probably every philosopher, and perhaps

some linguists, are intimately familiar with). Finding the right words might not create a sense of joy in every trans person, but it is a way that many of us nd joy in gendered language. Given that play is fun, the playful use of gendered language is also fun. While the playful attitude can help a marginalized person to feel at ease in uncomfortable settings and to nd new ways to understand oneself, it also is an attitude that can create joy. The playful terminology that trans people use to describe our bodies are de nite sites of joy. We often use humor to cope with our situation, and naming hormone medications things like ‘titty skittles’ or ‘antiboyotics’ are clear examples of the kind of humor we use to not only put ourselves at ease but also to nd joy in the mundane and toilsome aspects of our lives—like being on rigid medication schedules. Before we reconsider the two debates we began with, we think it is important to re ect on how trans language use involves taking back de nitional power. The uses of gendered language that we have explored so far all have to do with how various features of language, from the phonetic to the discursive, create social meaning within communities. Our focus has been on trans communities, but as Bettcher points out, trans communities are subcultures that are, to some extent, cut o

from more dominant communities. Although

meaning is created socially, not all linguistic communities have the same power to in uence broader trends in language—or, as with African American English, they do not get the credit for their in uence. The problem posed by politically signi cant terms is an example of this—it is the di erential power that makes such terms politically signi cant. If trans people had equal de nitional power, the problem could not arise, since de nitions of gender terms like ‘man’ and ‘woman’ would be co-de ned by both cis and trans people. Part of what makes such terms politically signi cant is that one group with disproportionately more power is trying to solve or account for the needs of a group with less. So, where does this leave trans users of language? Marilyn Frye (1983) drew philosophers’ attention to how heavily gendered English was, but she also drew our attention to how meaning—not just of words, but social meaning—primarily came from patriarchal power (80). For feminists to move forward, Frye argues, we must “rely on ourselves to make meaning … [to be] capable of weaving the web of meaning which will hold us in some kind of intelligibility” (80). When Bettcher (2013b) talks about trans subcultures and resistant de nitions of ‘woman,’ she is pointing us toward this need. Except that instead of narrowly conceiving of the problem of patriarchal power as one of men versus women, Bettcher expands the scope to include how trans people are kept from creating meaning. There is a dominant context, with dominant de nitions, and what trans people are doing in their p. 310

communities is resisting this web of meaning, weaving a web of our own. Whether Bettcher is right that the multiple-meaning view resolves the issue of politically signi cant terms, she is right that there are dominant and resistant communities that use language to develop ways of making the world intelligible. The lesson we learn from Frye and Bettcher is that to survive, we have to create meaning for ourselves, which is precisely what trans people do with language. Linguistic self-determination, a common value within many US trans communities, is the prioritization of linguistic self-identi cation (Zimman 2017b). Trans folks in these communities hold that individuals are the ones who will be the ultimate authority over their gender (and therefore over the language used to describe it). Part of this process of self-determination is the ability for trans people to nd the language that they feel accurately describes their experience of gender. The underlying linguistic norms of trans communities distributes de nitional power for the sake of self-determination. Notably, linguistic self-determination is exactly what Bettcher (2009) is pointing toward when talking about rst-person authority of trans people, and the ethical (as opposed to epistemological) grounding of trans people de ning themselves. Furthermore, Kukla and Lance (2023) give us an underlying pragmatics and ethics of this self-determination when arguing that gender ascriptions are about placing someone in social space.

Linguistic self-determination underlies what we have called self-recognition and joyous use of language by trans users. What is particularly valuable is when the terminology that was brought up in trans communities makes its way into the dominant structures of meaning. In 2019, the American Dialect Society named the locution ‘(my) pronouns’ as word of the year, and singular “they” (speci cally for its use with nonbinary 13

language users) as word of the decade (American Dialect Society 2020).

Merriam-Webster also chose the

nonbinary use of “they” as word of the year (Merriam-Webster 2019). Furthermore, Ben Zimmer, chair of the New Words Committee of the American Dialect Society, says in the announcement: The selection of “(my) pronouns” as Word of the Year speaks to how the personal expression of gender identity has become an increasing part of our shared discourse. That trend is also re ected in singular “they” being chosen as Word of the Decade, with a growing recognition of the use of “they” for those whose identities don’t conform to the binary of he and she. (American Dialect Society 2020) The use of ‘my pronouns are…’ has moved outside the trans and queer communities to become a far more widespread linguistic practice. Listing one’s pronouns has become so common that they are regularly listed in social media bios, leading some apps (e.g., Instagram) and websites (e.g., LinkedIn) to create a separate bio eld where users can list their pronouns. This has simultaneously popularized the use of the singular “they” to recognize nonbinary people (Conrod 2019; Konnelly and Cowper 2020). While “nonbinary” hasn’t been named word of the year (yet), it has also entered dominant usage beyond queer and trans communities, and has obviously in uenced the way pronouns have become such a signi cant topic. p. 311

There is surely something politically signi cant when language use originating in trans communities enters into the dominant varieties of English. Determining what and how these linguistic items are doing, politically speaking, when moving from trans subcultures to the dominant culture is a project too big for our space here. But the way trans people use language to claim de nitional power is something signi cant that we do with gendered language, even if it is not a phenomenon we fully understand.

4. Reconsidering the Debates Both of the philosophical debates we began the chapter by discussing are motivated by not wanting to misgender trans people, whether it’s through de ning our terms in a way that is not exclusionary or through removing gendered language that often leads to misgendering. Our aim is not only to open up these discussions about how such gendered language may harm trans people, but also to consider how gendered language can help trans people. None of what we have explored yields decisive conclusions to these debates, or even necessarily nudges us in a particular direction. The uses of language we pull from sociolinguistics are fairly broad, and trans linguistics itself is still fairly new, with work centering on trans communities increasing over the past decade and solidifying with Zimman’s formulation of a “trans linguistic framework” (Zimman 2020). While what we have done here has been very broad, we believe there are a number of interesting avenues for philosophers of language to explore.

4.1. Politicaly Significant Terms, Reconsidered The problem of politically signi cant terms has spawned answers that both recognize the underlying issue and attempt to resolve it (Díaz-León 2016; Ichikawa 2020); other answers try to draw our attention to how questions of political signi cance are separable from questions of semantics (Laskowski 2020; Chen 2021). There are also views that use the articulation of the problem to provide an account of gendered language that sidesteps or dissolves the problem (Bettcher 2013b; Kukla and Lance, 2023). Which route you take seems to turn on methodological commitments (i.e., the political encroaches on the semantic vs. these items can and should be treated separately) as well as political issues (e.g., what counts as a liberatory theory for trans people?). The way trans language users de ne terms aligns most nicely with Bettcher’s (2013b) approach to the problem, and we can see Kukla and Lance’s (2023) project as articulating the underlying pragmatics of such an approach. While we are sympathetic to this way of viewing things, it remains open whether it provides us with a solution to the problem Saul (2012) raises. It makes sense that trans people would alter language as a way to understand our experiences, and then build a new linguistic community around those uses. However, one may think that there is still a signi cant problem here: even if p. 312

the on-the-ground

attempts at semantic justice are approaching the problem in one way, that approach

is orthogonal to the importance of politically signi cant terms. Such a stance is a perfectly reasonable position to hold, but we are skeptical of its value. The issue of politically signi cant terms arises out of a particular view of language and semantic analysis that Ludwig Wittgenstein ([1953] 2009) famously called into question. Wittgenstein drew our attention to how the picture of language that captures the attention of analytic philosophers involves assuming that “reality must, and our concepts should, conform to the demands of logic” (Scheman 2022, 1). Furthermore, Wittgenstein warns that “the preconception of crystalline purity can only be removed by turning our whole inquiry around. (One might say: the inquiry must be turned around, but on the pivot of our real need.)” ([1953] 2009, sec. 108). Naomi Scheman (2022) argues that the logical must that Wittgenstein draws our attention to is particularly important when considering issues raised by oppression, of which trans exclusion and violence are certainly instances. In our own attempt to shine light on other, more positive uses of gendered language we have taken exactly this kind of approach. That is, we have pivoted away from our preconceptions of what our gendered terms should mean and focused instead on how trans people use gendered language in general. Importantly, the way trans people use gendered language is based in “our real need”; it is based in gaining recognition, subverting norms through play, nding ways to deal with harmful aspects of the dominant culture, nding much-needed joy in the face of the WTF, and rede ning the landscape of the dominant culture that is trying to exclude us. While we nd it is important for our language to be inclusive, there are ways trans people are doing this in our own communities, and we believe that this should constitute a more signi cant focus of philosophers’ research. The examples we pull from sociolinguistics serve as evidence for the value of reorienting our analysis to the way gendered language is used “on the ground.” In addition to concerns about the orientation of our analysis, there are other interesting questions raised by the practices of trans language users. The debate of politically signi cant terms is, obviously, about terms, but there is more to language than just terms. Kukla and Lance (2023) arguably already draw our attention to discursive use and how the pragmatics of gender ascriptions are caught up with an ethics of gendering. But what about the other units of language? Are there politically signi cant phonemes? Is there politically signi cant syntax? Well, let’s consider what trans language users do with our voices. The ways in which we use language communicates gendered meaning. The clearest example is probably how trans feminine people, and 14

particularly trans women, train their voices to be higher in pitch and brighter.

This alteration

communicates gendered meaning by conveying how they want to be read by others. Of course, many trans feminine people and trans women do not change their voices; and in e ect, they are less likely to “pass as

cis” and may face more misgendering. Similarly, trans masculine people taking testosterone who experience a change in their acoustic voice often have to navigate how their metaphorical voice (e.g., p. 313

opinions, perspectives) is given new “weight” in social interactions

(Crowley 2021). Is the meaning

attributed to higher or lower voices politically signi cant? How do we broaden or alter the meaning of how people speak given the political reasons? Do we simply undo the assumption that a high, “bright” voice is feminine and that lower voices are masculine? But where would that leave trans people who do alter their voice for gendered recognition? Arguably, we already know there is politically signi cant syntax. Gendered features are built into the grammar of languages like Hebrew and French. We have discussed how trans people in these linguistic communities have altered their language use to mismatch masculine and feminine parts of language to make the languages genderqueer. Furthermore, there have been attempts to do this in Spanish by providing gender-neutral options such as replacing the -o and -a su

xes with -x or -e (López 2021). But which

direction do we go in to solve the problem of politically signi cant syntax? Do we create gender-neutral versions of words so that the language does not have gender on the morpho-syntactic level? Or do we play with it to subvert and hopefully remove the gendered meaning? Can we even remove the gendered meaning while keeping the same words? That is, can we make a language like Hebrew, with masculine and feminine nouns, more like Danish where there are gendered nouns, but they don’t track masculinity/femininity, only distinguishing between which article they take?

4.2. How Much Gender? Reconsidered In looking at how trans people use gendered language to improve our lives, we are not denying that there are clear ways that gendered language also hurts us. There are a number of forms of gendered language use that we think should be limited or removed, especially when the use of such language involves assuming someone’s gender. The multitude of ways that trans people use gendered language to improve their lives, however, certainly complicates this question. In the philosophical literature, this debate began with a focus on pronouns, and then broadened to other lexical items like honori cs, su

xes, and generics (Dembro

of why these lexical items to the side (since Dembro

and Wodak 2018, 2021). Putting the question

and Wodak [2021] think some lexical items like ‘son’

and ‘daughter’ are perfectly ne), there can still be a question of why stop at the lexical level? The preceding discussion of the political signi cance of phonetic and morpho-syntactic levels can push us to question whether these aspects of language should be made gender neutral. If we’re going to gender-neutralize some parts of English, why don’t we just gender-neutralize everything? Responses to Dembro

and Wodak (2018,

2021) argue that they go too far (Hernandez 2021; Kukla and Lance, 2023). But what if they haven’t gone far enough? Why should pitch communicate gender? Why should a fronted /s/? Why label certain nouns as feminine and others as masculine? Why have any gender at all? p. 314

However, what can we make of this project given all the good things trans people do with gendered language? The playful use of gendered language shows that removing gender from language isn’t the only way to ght essentialism, as Hernandez (2021) has argued. The joy that trans people get from playing with gendered language and nding the language that works for them is an additional reason that we might want to resist fully gender-neutralizing language. These reasons become all the more convincing when we consider how some gender-non-normative language users, when presented with a gender-neutral language like Indonesian, add gendered pronouns into their language. They add exactly what some argue we have a moral duty to subtract. Why should we remove resources from trans people to make their lives more enjoyable and understandable? Why remove the options to play with and subvert the language that is so often used to harm us? Can’t we have a little gender, as a treat?

The ways in which trans people use gendered language does not give us answers to the question of how gendered language should be. (This is not surprising since the way people actually use language could not tell us the way people ought to use language.) However, it should draw our attention to many more factors that should be considered when arguing for one position over another. How are we demarcating what to neutralize and what to leave as is? What are the moral or linguistic reasons for using/keeping those particular linguistic items? In posing these questions, we mean to communicate how much more detailed and robust arguments for or against gender-neutralizing language need to be.

5. Conclusion What we have aimed to accomplish is to broaden the philosophical discussion about gendered language. We believe there are a lot of exciting and interesting projects for philosophers to engage in when it comes to gendered language, and almost all of them have both a practical and a political impact. This is not to dismiss the current debates; in fact, we think the answers to the questions we have raised may prove enlightening and insightful for the problems and puzzles that have already been recognized. However, it also appears that there are more uses of gendered language than have been dreamt of in our current philosophies. We began by saying language is like a garden—it can be pruned and watered, but we don’t have full control of where it grows. Trans people’s language use is a quickly growing vine and should not be pruned too early. Overpruning plants or overworking the soil can kill the garden, and great beauty, function, and balance can arise through happy accidents and by carefully watching as things grow. The work done by philosophers of language and linguists is valuable for understanding how things grow and questioning how we may want to shape things in the future. But it is through our attention to its growth that we gain insight into all language can be.

p. 315

Notes 1.

Saulʼs (2012) contextualist analysis is: “X is a woman” is true in a context C i X is a human and relevantly similar (according to the standards at work in C) to most of those possessing all of the biological markers of female sex” (201).

2.

Díaz-Leónʼs (2016) revised contextualist analysis is: “X is a woman” is true i X is human and relevantly similar to most females, where what counts as relevantly similar to most females depends on “objective” features of Xʼs context, including instrumental, moral, and political considerations having to do with how X should be treated (regardless of who utters the sentence or what their beliefs are)” (251).

3.

N. G. Laskowski (2020) argues that gender terms are polysemous, with the aim of severing the political significance from the semantics, and opting instead to say that whatever is wrong with our semantics should then be criticized as a project of ethics. Jonathan Jenkins Ichikawa (2020) argues that since contexts can o en be manipulated by power and create what he terms contextual injustice, if we do have a context where “Carla is not a woman” is true, that context is an unjust one and therefore worse than the trans-inclusive context. So, while the truth conditions of “X is a woman” may change from context to context, we have the standing to argue that some contexts are unjust contexts that should be altered. Hsiang-Yun Chen (2021) argues that the contextualist project is ill-fitted to resolve these debates. Instead of trying to collapse all the moral, political, and practical matters at hand into the semantics of ʻwoman,ʼ we should recognize the limitations of philosophical semantics and treat these questions head-on as the moral and political concerns that they are.

4.

Dembro and Wodakʼs (2018) arguments in more detail are as follows: First, they argue that recognizing genderqueer and nonbinary individuals puts us in a position to either have a specific “nonbinary pronoun,” which is inegalitarian given the wide variety of nonbinary genders, or have an abundance of pronouns for each gender identity, which they claim would be unmanageable by a linguistic community. Therefore, using gender-neutral pronouns for everyone is the best option (389). Second, gender-neutral pronouns help protect the privacy of trans and queer people by not forcing them to out

themselves by using gendered pronouns to refer to themselves or their partners (392). Third, given that gendered language reinforces sexist oppression and gender essentialism, removing gender-specific pronouns would aid in fighting these social harms (395).

p. 316

5.

Other forms of meaning beyond referential meaning are mostly important for the debate on “how much gender is too much gender,” as Dembro and Wodak (2021) would put it, but, as we discuss in section 4, other forms of meaning raise interesting questions along the same lines as politically significant terms.

6.

Within sociolinguistics and linguistic anthropology, indexicality is the process through which one sign (e.g., word, sound, syntactic construction etc.) points to (or indexes) another object in the context in which it occurs (e.g., some sort of social meaning like class, gender etc.). See Silverstein (1976).

7.

Creaky voice, popularly called ʻvocal fry,ʼ is a mode of phonation that is raspier than a typical speaking voice.

8.

That other aspects of language beyond words can have illocutionary force is a suggestion that requires more defense, but which goes beyond the scope of this current chapter. In mentioning it here, we just mean to suggest that the pragmatics that Kukla and Lance (2023) are noting could likely be applied to the other examples of trans language use we discuss in this section.

9.

Gender euphoria is “a distinct enjoyment or satisfaction caused by the correspondence between the personʼs gender identity and gendered features associated with a gender other than the one assigned at birth” (Ashley and Ells 2018).

10.

There is a tangentially interesting way that using ʻbonus holeʼ serves as a kind of speech act, because it can serve as what Kukla (2018) calls an invitation for what forms of sex may be explored in a particular encounter. Although this topic is outside the scope of the current chapter, there are interesting ways in which the terms trans people use for our bodies orient others to engage with our bodies.

11.

How trans joy is created with language is one of the phenomena that the linguist of our authorial pair is currently researching using ethnographic group interviews with trans communities in the US South.

12.

One half of our authorial pair, however, loved baseball and had a .550 batting average in little league.

13.

The runners up for word of the year were “ok boomer,” “cancel,” and “Karen.”

14.

The linguist among us wants to be clear that altering the pitch or brightness of oneʼs voice is not strictly speaking a ʻphonemeʼ even though it is part of phonetics.

References American Dialect Society. (2020). “2019 Word of the Year Is ʻ(My) Pronouns,ʼ Word of the Decade Is Singular ʻTheyʼ as voted by American Dialect Society.” Press release, January 3. https://www.americandialect.org/wp-content/uploads/2019-Word-of-theYear-PRESS-RELEASE.pdf. Ashley, Florence. (2020). “Surgical Informed Consent and Recognizing a Perioperative Duty to Disclose in Transgender Health Care.” McGill Journal of Law and Health 13:73. Google Scholar WorldCat Ashley, Florence. (2022). “ʻTransʼ Is My Gender Modality: A Modest Terminological Proposal.” In Trans Bodies, Trans Selves, edited by Laura Erickson-Schroth, 2nd ed. 22. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Ashley, Florence, and Ells, C. (2018). “In Favor of Covering Ethically Important Cosmetic Surgeries: Facial Feminization Surgery for Transgender People.” American Journal of Bioethics 18 (12): 23–25. Google Scholar WorldCat Austin, J. L. (1975). How to Do Things with Words. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Azul, D. (2013). “How Do Voices Become Gendered? A Critical Examination of Everyday and Medical Constructions of the Relationship between Voice, Sex, and Gender Identity.” In Challenging Popular Myths of Sex, Gender and Biology, edited by Malin Ah-King, 77–88. Cham: Springer. Google Scholar Google Preview WorldCat COPAC Bach, K., and Harnish, R. M. (1979). Linguistic Communication and Speech Acts. Cambridge, MA: MIT Press. Google Scholar Google Preview WorldCat COPAC Bartky, Sandra L. (1990). Femininity and Domination: Studies in the Phenomenology of Oppression. New York: Routledge. Google Scholar Google Preview WorldCat COPAC Beischel, W. J., Gauvin, S. E., and van Anders, S. M. (2022). “ʻA Little Shiny Gender Breakthroughʼ: Community Understandings of Gender Euphoria.” International Journal of Transgender Health 23 (3): 274–294. Google Scholar WorldCat Bell, Rowan. (forthcoming). “Being Your Best Self: Authenticity, Morality, and Gender Norms.” Hypatia. Google Scholar WorldCat Bettcher, Talia Mae. (2009). “Trans Identities and First-Person Authority.” In Youʼve Changed: Sex Reassignment and Personal Identity, edited by Laurie J. Shrage, 98–120. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC p. 317

Bettcher, Talia Mae. (2013a). “Trans Women and Interpretive Intimacy: Some Initial Reflections?” In The Essential Handbook of Womenʼs Sexuality, edited by Donna Castañeda, 51–68. Santa Barbara, CA: Praeger. Google Scholar Google Preview WorldCat COPAC Bettcher, Talia Mae. (2013b). “Trans Women and the Meaning of ʻWoman.ʼ” In The Philosophy of Sex: Contemporary Readings, edited by Raja Halwani, Alan Soble, Sarah Ho man, and Jacob M. Held, 6th ed., 223–250. Lanham, MD: Rowman and Littlefield. Google Scholar Google Preview WorldCat COPAC Bettcher, Talia Mae. (2017). “Trans Feminism: Recent Philosophical Developments.” Philosophy Compass 12 (11): Article e12438. Google Scholar WorldCat

Bettcher, Talia Mae. (2019). “What Is Trans Philosophy?” Hypatia 34 (4): 644–667. Google Scholar WorldCat Bershtling, O. (2014). “Speech Creates a Kind of Commitment.” In Queer Excursions: Retheorizing Binaries in Language, Gender, and Sexuality, edited by Lal Zimman, Jenny Davis, and Joshua Raclaw, 35–61. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Blackwood, Evelyn. (2014). “Language and Non-normative Gender and Sexuality in Indonesia.” In Zimman, Davis, and Raclaw, Queer Excursions, 81–101. Google Scholar Google Preview WorldCat COPAC Butler, Judith. (1990). Gender Trouble: Feminism and the Subversion of Identity. New York: Routledge. Google Scholar Google Preview WorldCat COPAC Butler, Judith. (2004). Undoing Gender. New York: Routledge. Google Scholar Google Preview WorldCat COPAC Calder, J. (2019). “The Fierceness of Fronted/s: Linguistic Rhematization through Visual Transformation. Language in Society 48 (1): 31–64. Google Scholar WorldCat Campbell-Kibler, K. (2011). “Intersecting Variables and Perceived Sexual Orientation in Men.” American Speech 86 (1): 52–68. Google Scholar WorldCat Chen, Hsiang-Yun. (2021). “Contextualism and the Semantics of ʻWoman.ʼ” Ergo: An Open Access Journal of Philosophy 20 (8) 577– 597. Google Scholar WorldCat Cli ord, Lily. 2019. “Late Acquisition of Gendered Phonetics: Voice Feminization in Transgender Women.” Poster presentation. New Ways of Analyzing Variation 48 Conference (NWAV 48), October 10–12, Eugene Oregon. Conrod, Kirby. (2019). “Pronouns Raising and Emerging.” PhD diss., University of Washington. Google Scholar Google Preview WorldCat COPAC Crowley, Archie. (2021). “The Weight of the Voice: Gender, Privilege, and Qualic Apperception.” Toronto Working Papers in Linguistics 43 (1): 1–14. Google Scholar WorldCat Davis, J. L. (2014). “More Than Just ʻGay Indians.ʼ” In Zimman, Davis, and Raclaw, Queer Excursions, 62–80. https://doi.org/10.1093/acprof:oso/9780199937295.003.0004. Google Scholar Google Preview WorldCat COPAC Dembro , Robin, and Wodak, Daniel. (2018). “He/She/They/Ze.” Ergo: An Open Access Journal of Philosophy 5: 371–406. Google Scholar WorldCat Dembro , Robin, and Wodak, Daniel. (2021). “How Much Gender Is Too Much Gender?” In The Routledge Handbook of Social and Political Philosophy of Language, edited by Justin Khoo and Rachel Katharine Sterken, 362–376. New York: Routledge. Google Scholar Google Preview WorldCat COPAC Díaz‐León, Esa. (2016). “Woman as a Politically Significant Term: A Solution to the Puzzle.” Hypatia 31 (2): 245–258. Google Scholar WorldCat Edelman, E. A., and Zimman, Lal. (2014). “Boycunts and Bonus Holes: Trans Menʼs Bodies, Neoliberalism, and the Sexual Productivity of Genitals.” Journal of Homosexuality 61 (5): 673–690. https://doi.org/10.1080/00918369.2014.870438. Google Scholar WorldCat

English, Brooke. 2021. “Gender on the Line: Agency, Power, and Fundamental Frequency of Transfeminine Individuals.” Paper presentation. Lavender Languages and Linguistics Conference, May 21–23, San Francisco California. Fausto-Sterling, A. (2008). Myths of Gender: Biological Theories about Women and Men. New York: Basic Books. Google Scholar Google Preview WorldCat COPAC p. 318

Fricker, Miranda. (2007). Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Frye, M. (1983). The Politics of Reality: Essays in Feminist Theory. Crossing Press. Google Scholar Google Preview WorldCat COPAC Hernandez, E. M. (2021). “Gender-A irmation and Loving Attention.” Hypatia 36 (4): 619–635. Google Scholar WorldCat Ichikawa, J. J. (2020). “Contextual Injustice.” Kennedy Institute of Ethics Journal 30 (1): 1–30. Google Scholar WorldCat Kapusta, S. J. (2016). “Misgendering and Its Moral Contestability.” Hypatia 31 (3): 502–519. Google Scholar WorldCat Knisely, K. A. (2020). “Le français non-binaire: Linguistic Forms Used by Non-binary Speakers of French.” Foreign Language Annals 53 (4): 850–876. Google Scholar WorldCat Konnelly, L. (2021). “Nuance and Normativity in Trans Linguistic Research.” Journal of Language and Sexuality 10 (1): 83–92. Google Scholar WorldCat Konnelly, L., and Cowper, E. (2020). “Gender Diversity and Morphosyntax: An Account of Singular They.” Glossa: A Journal of General Linguistics 5 (1): 1–19. Google Scholar WorldCat Kukla, Quill Rebecca, and Lance, Mark Norris. (2023). “Telling Gender: The Pragmatics and Ethics of Gender Ascriptions.” Ergo: An Open Access Journal of Philosophy 9: 42. https://doi.org/10.3998/ergo.2911 Google Scholar WorldCat Kukla, Rebecca (2018). “Thatʼs What She Said: The Language of Sexual Negotiation.” Ethics 129 (1): 70–97. Google Scholar WorldCat Lako , R. (1973). “Language and Womanʼs Place.” Language in Society 2 (1): 45–79. Google Scholar WorldCat Laskowski, N. G. (2020). “Moral Constraints on Gender Concepts.” Ethical Theory and Moral Practice 23 (1): 39–51. Google Scholar WorldCat Lewis, David. (1983). “Scorekeeping in a Language Game.” In Semantics from Di erent Points of View, edited by Rainer Bäuerle, Urs Egli, Arnim Stechow, 172–187. Berlin: Springer. Google Scholar Google Preview WorldCat COPAC López, Á. (2021). Direct and Indirect Non-binary Language in English to Spanish Translation. Google Scholar Google Preview WorldCat COPAC Lugones, María. (2003). Pilgrimages/Peregrinajes: Theorizing Coalition against Multiple Oppressions. Lanham, MD: Rowman and Littlefield. Google Scholar Google Preview WorldCat COPAC

McGowan, Mary Kate. (2019). Just Words: On Speech and Hidden Harm. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC McGowan, Mary Kate. (2021). “Feminist Philosophy of Language in the Analytic Tradition.” In The Oxford Handbook of Feminist Philosophy, edited by Kim Q. Hall and Ásta, 184–197. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Merriam-Webster. (2019). “ʻTheyʼ Is Merriam-Websterʼs Word of the Year 2019.” Merriam-Webster Dictionary (online). https://www.merriam-webster.com/words-at-play/woty2019-top-looked-up-words-they. Saul, Jennifer. (2012). “Politically Significant Terms and Philosophy of Language.” In Out from the Shadows: Analytical Feminist Contributions to Traditional Philosophy, edited by Sharon L. Crasnow and Anita M. Superson, 195–216. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Scheman, Naomi. (2022). “Disrupting Demands: Messy Challenges to Analytic Methodology.” Journal of Social Philosophy 53 (4): 473–493. Google Scholar WorldCat Serano, J. (2020). “Autogynephilia: A Scientific Review, Feminist Analysis, and Alternative ʻEmbodiment Fantasiesʼ Model.” Sociological Review 68 (4): 763–778. Google Scholar WorldCat Silverstein, M. (1976). “Shi ers, Linguistic Categories, and Cultural Description.” In Meaning in Anthropology, edited by K. H. Basso and H. A. Selby, 11–55. Albuquerque: University of New Mexico Press. Google Scholar Google Preview WorldCat COPAC Stryker, Susan. (2017). Transgender History: The Roots of Todayʼs Revolution. Berkeley: Seal Press. Google Scholar Google Preview WorldCat COPAC Toomey, R. B., Syvertsen, A. K., and Shramko, M. (2018). “Transgender Adolescent Suicide Behavior.” Pediatrics 142 (4): 1–8. Google Scholar WorldCat p. 319

Tudisco, J. (2021). “Queering the French Académie: Reclaiming Linguistic Authority for Trans and Non-binary People.” Toronto Working Papers in Linguistics 43 (1): 1–15. Google Scholar WorldCat Wittgenstein, Ludwig. ([1953] 2009). Philosophical Investigations. Oxford: Wiley-Blackwell. Google Scholar Google Preview WorldCat COPAC Zimman, Lal. (2013). “Hegemonic Masculinity and the Variability of Gay-Sounding Speech: The Perceived Sexuality of Transgender Men.” Journal of Language and Sexuality 2 (1): 1–39. Google Scholar WorldCat Zimman, Lal. (2014). “The Discursive Construction of Sex: Remaking and Reclaiming the Gendered Body in Talk about Genitals among Trans Men.” In Zimman, Davis, and Raclaw, Queer Excursions, 13–34. Google Scholar Google Preview WorldCat COPAC Zimman, Lal. (2017a). “Gender as Stylistic Bricolage: Transmasculine Voices and the Relationship between Fundamental Frequency and /s/.” Language in Society 46 (3): 339–370. https://doi.org/10.1017/S0047404517000070 Google Scholar WorldCat Zimman, Lal. (2017b). “Trans Peopleʼs Linguistic Self- Determination and the Dialogic Nature of Identity.” In Representing Trans: Linguistic, Legal and Everyday Perspectives, edited by E. Hazenberg and M. Meyerho , 226–248. Wellington, NZ: Victoria University Press.

Google Scholar

Google Preview

WorldCat

COPAC

Zimman, Lal. (2020). “Transgender Language, Transgender Moment: Toward a Trans Linguistics.” In The Oxford Handbook of Language and Sexuality, edited by Kira Hall and Rusty Barrett, 1–23. Online edition. Oxford Academic. https://doi.org/10.1093/oxfordhb/9780190212926.013.45. Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

14 Beyond Pronouns: Gender Visibility and Neutrality across Languages  Iz González Vázquez, Anna Klieber, Martina Rosola

Author Notes

https://doi.org/10.1093/oxfordhb/9780192844118.013.37 Published: 22 May 2024

Pages 320–346

Abstract The aim of this chapter is to explore some trans and feminist concerns about the gendered aspects of languages beyond English, focusing on Spanish, Italian, and German. Historically, discussions about gendered language have often challenged the ways in which language can make women (in)visible, by addressing the implicit and explicit androcentrism and sexism in our language, in what can be called the visibility project. Recently, questions surrounding trans-inclusiveness and the possibility of avoiding gender markers altogether have become more prominent, often highlighting the issue of pronouns (e.g. the singular ‘they’). This can be called the gender-neutrality project. However, anglophone philosophy o ers few discussions on how gender-neutrality or visibility strategies can be applied to languages that are more gendered than English. The chapter starts by showing how gender expresses itself in the grammatical structure of the three sample languages, and identi es some of their speci c visibility and neutrality strategies that could be (or are already) applied to them. This leads to the ‘innovative project’ which constructs nuanced and novel ways to transform gendered grammatical structures. Depending on their application, these strategies can ful l gender-neutrality goals in some cases and visibility goals in others. This highlights, among other things, the ways in which genderqueer people have attempted to shape their linguistic resources in di erent linguistic contexts, while also opening up avenues for new research when it comes to discussions of gender and language.

Keywords: gendered languages, gender neutrality, gender visibility, trans-inclusive language, Italian, German, Spanish Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction The relationship between gender and language has been a topic in feminist discussions for some time now. Historically, these discussions have belonged to what we call the visibility project: they confronted the ways in which language can render women invisible (or, conversely, unduly visible) by addressing the implicit and explicit androcentrism and sexism in our language. Recently, this discussion has expanded to questions surrounding how to accurately refer to genderqueer people (that is, those who do not, or do not exclusively, identify as either men or women). Although the relationship between gender and language has received some attention within philosophy, the literature on this topic has mainly focused on the gendered aspects of English—with the question of 1

whether we ought to use gender-neutral pronouns for everybody guring prominently. This chapter will not weigh in on this speci c discussion, but it takes its starting point earlier: the almost exclusive focus on English means that we have very little idea about how an application of these various strategies looks like for languages that are more gender-loaded. In this chapter, we aim to open up a discussion about this by 2

focusing on Spanish, Italian, and German.

The general argument that underlies this chapter is that languages ought to be inclusive of women and genderqueer individuals. However, we think that there is no single strategy for how to achieve this. On the contrary, we believe that many di erent strategies can serve this purpose and do it in di erent ways. We divide these strategies into three projects. The rst two we’ve already mentioned: visibility and genderneutrality. The third is the innovative project, which groups strategies that employ novel linguistic resources p. 321

for the inclusion of women and genderqueer people. However, the

linguistic representation of any one

group will depend on how these strategies are used. Hence, as we will show, innovative strategies can be implemented in line with either the visibility or the gender-neutrality project. While the aim of this chapter is not to tease out all of the ways in which these projects are compatible or con ict, we do point to some of these issues throughout. So, while we don’t necessarily argue for the superiority of one strategy over the others—especially provided the linguistic context-sensitivity of our discussion—we do highlight the fact that some strategies of gender inclusion are more successful for certain purposes than others (e.g., some, but not all, include genderqueer people while others challenge gender stereotypes e

ciently). Overall, the

case we make in this chapter with the analysis of our sample languages is not only to show how things work in Spanish, Italian, and German speci cally but to highlight the broader point that there doesn’t seem to be a xed answer to the question of how to represent genderqueer individuals in language. This representation will depend on the language and context. Of course, a larger sample of languages could give us an even more interesting and nuanced picture. However, we still hope to broaden the philosophical discussion on gender and language by highlighting the ways in which the strategies used to make gender visible or neutral don’t necessarily work in the same way in every language. This means that trans people in diverse language contexts need to nd di erent positive ways to change and manipulate their language. If our philosophical discussion wants to make claims about the ways in which genderqueer people can nd a language that has room for them, we should broaden our discussion: a focus on English misses the reality for many genderqueer people, particularly the reality and struggles of those who nd themselves in language contexts that provide fewer resources to move beyond a grammatical gender binary. And this might impact our philosophical arguments: for example, if our claim is that we ought to use gender-neutral pronouns for everybody, but we are confronted with a language where this is much harder to do than in English, where does that leave speakers of those languages, particularly those who have found new, innovative, and nuanced strategies to adapt what their language o ers? It is our hope to open up a discussion about these issues.

We start by providing a short overview of the central issues we’re concerned with. Next, we show how gender is marked in the grammar of our three sample languages and identify some visibility and genderneutrality strategies speci c to these languages. Several caveats arise: while gender neutrality isn’t always easy to achieve in our sample languages, there are also issues with some visibility strategies—e.g., as we’ll see in more detail later, they might enforce binarism, or only ensure the visibility of women in language, while not o ering satisfying options for genderqueer people. This not only shows that the discussion has to be much broader than a mere focus on pronouns would suggest, but also leads us to identify a third project: the innovative project. As mentioned earlier, this project consists in transforming gendered grammatical structures and introducing novel linguistic resources into the language. We show how, depending on how these strategies are applied, they can serve either gender-neutrality or visibility goals.

p. 322

2. What are the central issues? In her paper ‘Sexism’, Marilyn Frye (1983) calls attention to the fact that ‘[s]ex-identi cation intrudes into every moment of our lives and discourse, no matter what the supposedly primary focus or topic of the moment is’ (19–20). According to her, this suggests that gender is always relevant, even when it shouldn’t be, which is particularly evident in the gendered elements of language that force speakers to focus on what Frye calls ‘sex-marking’: we can’t use singular pronouns, according to her, without being forced to take into account the ‘sex’ of a person. Now, as Saul and Díaz-León (2016) point out, like many feminists of the early 1980s, Frye doesn’t consider the possibility that personal pronouns like ‘he’ or ‘she’ could be a matter of gender rather than sex. She also doesn’t consider trans issues or gender-neutral pronouns. However, Frye highlights an important point here—a point we think still holds: how we mark gender (or sex) is deeply entrenched in how we use the grammatical structures of our languages. And this goes for some languages more than for others. So, even though our discussions have shifted, and we now tend to focus on gender rather than sex, the general questions Frye points to haven’t ceased to be an issue. And, of the various 3

elements of language that feminists have scrutinized, in this chapter, we will discuss the issues that arise 4

from grammatical gender.

So, why is grammatical gender such a problem? Why do we need strategies to tamper with language in the rst place, and what are these strategies supposed to address? To make this clear, we want to highlight three structural issues and a widespread misuse that require attention in our sample languages—arguably in di erent ways than in English: generic masculines, overextended masculines, misgendering, and unmarked masculines. These issues are by far not new, but they continue to manifest central reasons as to why our languages can be exclusive. The rst problem is of a similar nature to the one that Frye points out: marking gender entrenches all aspects of our speech. This becomes especially apparent given that, in our sample languages, almost every linguistic element is gendered and, with very few exceptions, there are no gender-neutral su

xes or third-

person singular pronouns. This is not just a matter of bad use, but it constitutes a structural problem: speakers of German, Italian, and Spanish have to use gendered terms even when they talk about ‘generic’ or unknown individuals. Conventionally, this can be done using masculine terms as in ‘man is a mammal’ or ‘I don’t know who broke in: he was gone before I arrived’. Notice that while in English, one could opt for gender-neutral alternatives to ‘man’ and ‘he’ in the previous sentences, this is not always possible for our sample languages precisely because they lack gender-neutral su

xes or pronouns. The use of masculine

terms or forms for ‘generic’ or unknown individuals is referred to as generic masculine.

Another consequence concerns the gender of terms for mixed-gender groups: just as in the previous case, a p. 323

speaker of our sample languages has to choose a gender for such

terms and, again, the grammatical

convention is to use the masculine, regardless of the gender proportions in the group. So, while one man is su

cient to require the masculine, 99% of women are not enough for the feminine, which is only used

when all group members are women. These instances are called overextended masculines. Moreover, that our sample languages lack suitable pronouns and forms to refer to genderqueer people gives 5

rise to a structural form of misgendering: whereas in English, there is the option of the singular ‘they’ (despite the resistance it often encounters), similar options are not available in other languages. Finally, the use of masculine professional titles for women (which Formato 2019 calls unmarked masculines) is much discussed in Italian, German, and Spanish speaking contexts. An example of this type is ‘Maria 6

Montessori fu medico’ (‘Maria Montessori was a doctor [masc]’). Unlike the three aforementioned issues (generic and overextended masculine and misgendering), that depend on the feature of our languages (e.g., heavy gender-loading and binary grammar), unmarked masculines are not structural: the feminine professional titles can be formed using the standard morphological rules when they are not attested already. In this case, then, speakers do have the chance to use forms that correspond to the referent’s gender. For this reason, we’ll focus on generic and overextended masculines and on misgendering (for a discussion on the unmarked masculines, see Formato 2019). One concern about generic and overextended masculines has to do with the possibility of confusion—that it may not always be clear whether a masculine form refers only to men or to people of other genders too. This hypothesis, though, is often rejected as unlikely, even by critics of the generic and overextended uses of the masculine (see, e.g., Cameron 1998). One of the most paradigmatic contexts in which the masculine is taken to unambiguously refer to anyone is in law, where it’s also especially important to avoid misinterpretation. Yet there have been several cases where masculine words in laws—while phrased as a ‘generic’ reference, allegedly referring to everybody—have been interpreted as referring to men only. As a result, women were not granted the rights provided by laws due to the masculine wording. E.g., Barbara Pezzini (2015) reports that, in 1906, Italian women’s request to be included in electoral rolls was rejected, even though the electoral law did not require voters to be men, and the Statuto Albertino (1948), the Constitution e ective in Italy before 1948, stated that all [masc] subjects [masc] were equals before the law and equally bene tted from civil and political rights. According to the highest court of appeal, though, explicit legislation was needed to extend this right to women, indicating that the masculine terms in the Statuto Albertino and in the electoral law were, in fact, gender speci c. Marguerite Ritchie reports similar cases in Ireland, England, and Canada, where the courts interpreted the occurrence of masculine terms and pronouns in legal acts as indicating that women were not covered by those laws and were not, therefore, entitled to voting rights. Strikingly, this happened notwithstanding the Interpretation Act, which explicitly stated that masculine words ‘shall be taken to include Females … unless the contrary … is expressly provided’ (quoted in Ritchie 1975, 694). One court even ruled that a single woman did not fall into the description ‘every man’ ‘for purposes of election, but treated p. 324

her as included within

the masculine terminology for the purpose of imposing a penalty against her’

(696). Recently, a similar argument was used in a lawsuit against Hillary Clinton. The claimant argued that a woman cannot become president of the United States because the president is always referred to with ‘he’ 7

in the Constitution (see Lieberman 2008; Wallace v. Cegavske 2016). As these examples show, generic and overextended masculines can and did lead to (genuine or alleged) confusion. Hence, this raises concerns about the ambiguity of these usages.

The issues of generic masculines, overextended masculines, and misgendering have been addressed by what we call the visibility and gender-neutrality projects. As mentioned already, the visibility project has historically challenged the ways in which language has made women invisible by, amongst other things, the use of generic masculines and overextended masculines. This has been done in various ways and sometimes with diverging approaches. E.g., on the one hand, we have what Cameron (1992) calls ‘theoretical reformism’, i.e., ‘the belief that fairly minor changes in the surface forms of words are su

cient to solve the

problem of sexism in language’ (ibid., 103). According to her, reformists such as Casey Miller and Kate Swift (1980) suggest that a central problem of androcentric and sexist speech is that it distorts reality and is outdated, and, thus, we should simply move away from androcentric words, e.g., by replacing ‘mankind’ with ‘humanity’, and adopting gender-fair or neutral language. However, this, to Cameron, oversimpli es important ideological and political features of people’s linguistic behaviour. For instance, where a theoretical reformist might argue that saying ‘astronaut’ instead of ‘spaceman’ makes it easier to imagine women in space, too, Cameron stresses that this doesn’t help us if the word ‘astronaut’ in our society is also used as if it were masculine (120). Hence, Cameron proposes a di erent way to make women visible: according to her, we should be honest that suggesting linguistic change has ideological reasons and a speci c political utility. To this end, Cameron uses feminine generics: ‘I do not want my use of pronouns to slip by unnoticed: I want readers to think about it, and to act on their conclusions’ (126). The gender-neutrality project, in turn, challenges ways in which language has made genderqueer people invisible, also by the use of generic masculines and overextended masculines, with the addition of the issue of misgendering. For instance, Dembro

and Wodak (2018, 2021) have argued that the use of ‘they’ can be a means to avoid

gender entirely in English. In ‘He/She/They/Ze’, they defend two main claims: ‘We have a duty not to use binary gender-speci c pronouns (he or she) to refer to genderqueer individuals’ and ‘We have a duty not to use any gender-speci c pronouns to refer to anyone, regardless of their gender identity’ (2018, 372). They propose gradually eliminating gender-speci c pronouns. This proposal has an interesting consequence— namely, that we’d use ‘they’, not just as a pronoun for genderqueer individuals, but as a gender-neutral singular pronoun for everyone. However, Dembro

and Wodak ‘leave open whether our case for this

generalizes to other natural languages with gender-speci c pronouns or more extensive systems of grammatical gender; the elimination of he and she in English is much less drastic than the elimination of grammatical gender in, say, Portuguese’ (2018, 374–375). p. 325

8

Similarly, in Spanish, Italian, and German, we lack suitable pronouns and linguistic forms to refer to genderqueer people. So, how might we address these issues?

As mentioned previously, strategies involving the introduction of novel linguistic resources have been proposed for these languages. We call this third project innovative. This project is an original proposal that has the potential to address issues in line with both the visibility and the gender-neutrality projects, depending on how the strategies are used. For instance, López (2020) provides an interesting perspective on these issues by focusing on the language used for trans and genderqueer individuals in both Spanish and English, and includes two types of proposals about how to translate from English to Spanish: Indirect Nonbinary Language (INL) and Direct Non-binary Language (DNL). INL is an indirect approach to avoid marking gender. This includes modifying sentences, either by choosing gender-neutral words that don’t mark gender or by restructuring sentences to avoid marking gender. For example, using ‘estas personas leen’ (these people read), which, albeit is grammatically feminine, can be used for people of all genders, instead of ‘ellos leen’ (they [masc] read). DNL, on the other hand, is a more direct approach that involves using new grammatical resources to make genderqueer people visible, e.g., by substituting gendered su

xes with morphemes such as ‘-e’ and ‘-x’. This results in ‘elles leen’ (they read) instead of ‘ellos leen’

(they [masc] read). Such a solution would be helpful when translating from English to Spanish, especially when the translation involves genderqueer individuals. Under López’s proposal, then, the ‘-x’ and ‘-e’ would be used as a third grammatical gender in a language that only has two binary grammatical genders available. We call these strategies innovative because they introduce novel linguistic resources that can contribute both to making gender visible and obscure it, depending on how they are used. This highlights a central but rarely acknowledged point: the issue of gender neutrality goes far beyond the topic of pronouns—it extends to articles, adjectives, and nouns (and, as Hernandez and Crowley, this volume, suggest, also requires that we take into account more than grammar). While there are convincing arguments to rid languages— speci cally English—of grammatical gender, this is arguably harder, provided the more gender-loaded grammatical systems of our sample languages. However, we want to show that, despite these limitations, speakers have found ways to use binary grammatical gender structures in creative and innovative ways. Some are concerned with the visibility of women or marginalized genders more generally, others focus on achieving gender neutrality, and others ‘tamper with language’, to use Cameron’s words by introducing new morphological or phonological resources in the language. By introducing the innovative grouping, we hope to draw out another possible area for future research, which might explore whether these strategies can serve the aims of making women and genderqueer individuals visible, or of gender-neutrality strategies, or a combination of these, as well as the implications of such proposals. So, what do these di erent visibility, neutrality, and innovative strategies look like in practice, and how do they fare with our heavily gender-loaded sample languages in more detail? p. 326

The following sections address this in two ways: rst, we provide a short grammatical overview of what grammatical gender looks like in Spanish, Italian, and German. Next, we introduce speci c strategies that deal with gender exclusion in these languages.

3. Gender in the Grammar of German, Italian, and Spanish German, Italian, and Spanish are languages with heavy gender loading—that is, the extent to which gender gets marked in the grammatical structure of a language. And despite the di erences between these languages, there are several similarities: for one, not just pronouns but all nouns, even those with inanimate referents, are gendered, alongside certain articles, adjectives, and, in some cases, past participles, all of which must agree in gender and number with the noun they are controlled by. The general idea in German, Italian, and Spanish is that an adjective and article referring to a masculine noun must be masculine too. However, though the grammatical genders of nouns with inanimate referents are arbitrary 9

(see, e.g., Corbett 1991), those of personal nouns (usually) re ect the referent’s social gender. That is, generally, a noun will be grammatically masculine when referring to a man and feminine when referring to a woman. What about the aforementioned generic and overextended masculines, and misgendering? In what follows, we present the main strategies used in German, Italian, and Spanish to address the issues of generic masculines, overextended masculines, and misgendering.

3.1. German German and English belong to the same linguistic family and have important similarities. However, they also di er signi cantly in some respects. As shown in Table 14.1, in German, articles, nouns, and adjectives su

10

xes have di erent masculine and feminine forms:

Table 14.1. German gendered morphology Definite article

Sing.

Plur.

p. 327

Indefinite article

Fem

Masc

Fem

Masc

German

Die gute Lehrerin

Der gute Lehrer

Eine gute Lehrerin

Ein guter Lehrer

English

The good teacher

German

Die guten Lehrerinnen

English

The good teachers

A good teacher Die guten Lehrer

–

–

–

As you can see in this table, articles’ and nouns’ endings have varying masculine and feminine forms, as do adjectives (here, when using an inde nite article). The adjective endings are a bit more complex; however, making nouns ‘feminine’ is usually quite straightforward: you simply add the su

x ‘-in’ to the masculine

word stem. In addition, in several parts of the German-speaking world (in Austria, for instance), it is common in speaking to place a gendered article before a proper name (see Table 14.2).

Table 14.2. Articles preceding names in some German speaking dialects Fem.

Masc.

German

Die Olivia spielt Fußball

Der Marco spielt Fußball

English

(The [fem]) Olivia is playing football

(The [masc]) Marco is playing football

This is not something you will usually nd in textbooks: it’s a spoken and colloquial use, and not common in every German-speaking region.

Next, let’s move on to pronouns. Most relevant here are the singular 3rd-person pronouns: ‘er’ (he), ‘sie’ 11

(she), ‘es’ (it).

All plural pronouns are gender neutral; e.g., the 3rd-person plural pronoun ‘sie’ (they

[plur]) can be used for all genders. See Table 14.3 for some examples.

Table 14.3. German gendered pronouns 3rd-person pronouns

Sing.

Plur.

Fem.

Masc.

German

Sie liest

Er liest

English

She reads

He reads

German

Sie lesen

English

They read

Although the 3rd-person plural ‘sie’ is gender neutral, there isn’t a straightforward way to transform it into a gender-neutral singular pronoun. This is partly because ‘sie’ is, as we have seen, also the singular feminine pronoun; but it is also because gender is marked in multiple ways in the German language (e.g., word su

xes) and, for the most part, conforms to a strict binary. The more complex our sentences become,

the harder it would be to stick with ‘sie’ as gender neutral.

p. 328

3.2. Italian Italian belongs to the Romance language family. But as in German, in Italian, adjectives, articles, and nouns are either masculine or feminine (see Table 14.4). Again, personal nouns re ect the referent’s gender, with some exceptions (loan words, dead metaphors, meaning extension, etc.):

Table 14.4. Italian gendered morphology Definite articles

Sing.

Plur.

Fem

Masc

Italian

La brava maestra

Il bravo maestro

English

The good teacher

Italian

Le brave maestre

English

The good teachers

I bravi maestri

Unlike in German, the word stem in Italian is not simply treated as identical to the masculine noun; instead a gendered su

x, typically ‘-a’ for the feminine and ‘-o’ for the masculine, is added to the stem.

Recall the region-speci c German habit of adding the gendered article in front of proper names. A similar use appears in Northern Italy (see Table 14.5):

Table 14.5. Article preceding names in some Italian speaking dialects Fem.

Masc.

Italian

La Olivia gioca a calcio

Il Marco gioca a calcio

English

(The [fem]) Olivia plays football

(The [masc]) Marco plays football

Moreover, in Italian, even in writing, women’s surnames, but not men’s, are almost always preceded by an article: one would say ‘Machiavelli’ but ‘la Montessori’ (the [fem] Montessori). This usage, then, is asymmetrical. Finally, let us turn to personal pronouns. Again, the singular 3rd-person pronouns, ‘lui’ (he) and ‘lei’ (she), 12

are the most relevant.

See Table 14.6 for some examples:

Table 14.6. Italian gendered pronouns 3rd person pronouns

Sing.

Plur.

Fem.

Masc.

Italian

Lei legge

Lui legge

English

She reads

He reads

Italian

Loro leggono

English

They read

The 3rd-person plural pronoun ‘loro’ (they) is gender neutral but, as in German, not easy to use for singular individuals. Notice, however, that subject pronouns can be omitted in Italian in most cases, so one can simply write ‘mangia’ ([s/he] eats) without any pronoun. Moreover, possessive pronouns do not agree with the owner but with the owned. Hence, it is possible to talk about someone while avoiding personal pronouns altogether.

p. 329

3.3. Spanish Spanish, like Italian, is a Romance language. Spanish is one of the most widely spoken languages worldwide, and thus has many variants. We will focus on the Spanish spoken in Mexico, which is not homogenous either, due in part because of its interaction with more than sixty other Mexican national languages (Gobierno de México, n.d.). Just as in German and Italian, all Spanish nouns are gendered, and articles and adjectives must agree with them (see Table 14.7):

Table 14.7. Spanish gendered morphology Definite articles

Sing

Plur

Fem.

Masc.

Spanish

La buena maestra

El buen maestro

English

The good teacher

Spanish

Las buenas maestras

English

The good teachers

Los buenos maestros

As we’ve seen with both Italian and German, di erent articles and adjectives are used for men and women in both plural and singular forms. And, as in Italian, feminine nouns end in ‘-a’, and most masculine nouns end in ‘-o’. So it will come as no surprise that pronouns for the most part are also gendered (see Table 14.8), though to a lesser extent compared to Spanish spoken in Spain, which additionally includes 2nd-person plural gendered pronouns: Notice that plural pronouns are gendered too, with the interesting exception of the 2nd-person plural, where a gender-neutral formal pronoun is available. As in Italian, though, pronouns can often be omitted in Spanish. We’ll come back to this in the strategy section. For now, let’s look at some examples in Table 14.9 to make clear how they are applied:

Table 14.8. Spanish pronouns 1st person Fem. Sing.

Plur.

2nd person Masc.

Fem.

3rd person Masc.

Gender-neutral

Fem.

Masc.

Spanish

Yo

Tú

Ella

Él

English

I

You

She

He

Spanish

Nosotras

Ellas

Ellos

English

We

Nosotros

Ustedes You

They

Table 14.9. Use of Spanish gendered pronouns

3rd person singular

1st person plural

3rd person plural

Fem

Masc

Spanish

Ella lee

Él lee

English

She reads

He reads

Spanish

Nosotras leemos

Nosotros leemos

English

We read –

–

Spanish

Ellas leen

Ellos leen

English

They read

To recap: German, Italian, and Spanish are heavily gender loaded in a binary way. A word’s grammatical gender usually re ects its referent’s gender, and for the most part, only masculine and feminine are available. Thus these languages lack a grammatical gender that can be used to refer to genderqueer individuals. On top of that, there is still an ongoing struggle regarding female visibility in the language itself, owed to the lasting dominance of the generic and overextended masculine. We’ll now move on to consider some of the ways in which German, Italian, and Spanish speakers have nevertheless attempted to use and/or extend the resources of their languages to create visibility or genderneutral strategies.

4. Beyond Sexism: Strategies in German, Italian, and Spanish In what follows, we present the main strategies used to refer to genderqueer people, and to avoid generic and overextended masculines, in German, Italian, and Spanish.

p. 331

4.1. Visibility The rst family of strategies aims speci cally at making women visible in language. There are various ways to achieve this in our three languages:

1.  Splitting. This strategy amounts to repeating the same term in both masculine and feminine forms. For instance, in German, to talk about a mixed-gender group one would say ‘die Lehrer und Lehrerinnen’ (the [plur] teachers [masc] and teachers [fem]) or ‘der Lehrer und die Lehrerin’ (the [masc, sing] teacher [masc] and the [fem, sing] teacher [fem]). Whereas in the plural one can use a gender-neutral ‘die’, the singular requires both gender-speci c articles. This strategy, akin to the English use of ‘he or she’, poses, however, a new problem: what gender should come rst? Furthermore, far from correctly referring to genderqueer people, it reinforces binarism because it repeats the masculine and feminine forms as if they exhaust all of the possibilities. In addition, since in German, Italian, and Spanish (almost) every noun, article, and adjective is gendered, repeating both forms means doubling most elements of a sentence. For example, in Italian, ‘i maestri sono bravi’ (the [masc] teachers [masc] are good [masc]) becomes ‘i maestri e le maestre sono bravi e brave’ (the [masc] teachers [masc] and the [fem] teachers [fem] are good [masc] and good [fem]). The result is clearly redundant, much more so than its equivalent in English, where only personal pronouns are repeated. It is partly for this reason that splitting has been criticized. Redundancy is particularly problematic in contexts where brevity is crucial, such as titles, slogans, advertising, and any text subject to a word limit. But as we will see, the following variants of this strategy (article splitting, ‘gendern’, and slash) are a bit less redundant. 2.  Article splitting. In Italian and Spanish, sometimes two di erent articles are used at the same time, even with a noun that is gendered in one speci c way: e.g., ‘las y los maestros’ and ‘le e i maestri’ (the [fem] and the [masc] teachers [masc]), respectively. Notice, however, that the noun is still gendered. There are some terms in Italian and Spanish which have a single form that can refer to individuals of 13

any gender, namely epicenes and semi-epicenes.

Using these nouns, we can construct phrases such as

‘il e la farmacista’ (the [masc] and the [fem] pharmacist). For other nouns, however, one has to choose a gender category for the articles. Traditionally, two factors determine this: the generic masculine and proximity agreement. Under proximity agreement, closest words must have matching gender. Hence, if the closest article is masculine, so is the noun, according to both criteria. If, however, the closest article is feminine, the two criteria prescribe di erent outcomes: ‘i e le maestri’ (the [masc] and the [fem] teachers [masc]), according to the generic masculine, and ‘i e le maestre’ (the [masc] and the [fem] teachers [fem]), according to proximity agreement. p. 332

3.

14

‘Gendern’ (gendering). In German, a more concise variant of splitting is to place an underscore (known as a ‘gender-gap’) or an asterisk (known as a ‘gender-star’) before a feminine su

x: e.g., 15

‘die Lehrer_innen’ (the teachers [masc_fem]) or ‘die Lehrer*innen’ (the teachers [masc*fem]). similar strategy, called ‘Binnen I’, involves capitalising the initial ‘i’ of the feminine su

A

x: e.g., ‘die

LehrerInnen’ (the teachers [mascFem]). It’s important to note that, if we understand this as simply joining up feminine and masculine forms, this strategy is not much of an improvement for genderqueer people. This becomes even more striking in the singular, where one has to list both the masculine and feminine articles in German: e.g., ‘der_die Lehrer_in’, ‘der*die Lehrer*in’, ‘der/die LehrerIn’ (the [masc, fem] teacher [masc, fem]). There are other interpretations of this strategy, however, which we will come back to shortly. It is also of note that, although not impossible, it is di

cult to transpose this strategy into speech. One option is to make a very short and subtle pause

where the symbol sits in writing: one could say ‘Lehrer [short pause] innen’ rather than just ‘Lehrerinnen’. 4.  Slash. To talk about referents of unknown gender in all three languages, it is possible to use a forward slash. In German, this results in ‘der Lehrer/die Lehrerin’ (the [masc] teacher [masc]/the [fem] teacher [fem]), for example. In Italian and Spanish, the slash allows one to avoid repeating the whole noun as it can divide articles and word endings only: e.g., ‘il/la maestro/a’ and ‘él/la maestro/a’ (the [masc]/the [fem] teacher [masc]/[fem]), respectively. Notice that something similar is commonly

done when both the singular and the plural are correct, as in ‘segna la/e risposta/e corretta/e’ (mark the [sing]/[plur] correct [sing]/[plur] answer(s) [sing]/[plur]). However, while this solution might work on paper, it is di

cult to use orally. Even in writing, it becomes challenging when moving past

the use of an article and a noun. To illustrate this point, consider the following example in Spanish: ‘Los maestros están encargados de informar a los alumnos’ (The [masc] teachers [masc] are in charge [masc] of informing the [masc] students [masc]), following this strategy, would become ‘Los/Las maestros/as están encargados/as de informar a los/las alumnos/as’ (The [masc]/[fem] teachers [masc]/[fem] are in charge [masc]/[fem] of informing the [masc]/[fem] students [masc]/[fem]). In Spanish, this has fallen out of favour due to its complexity. In Italy, however, it is quite common in administrative forms, where the gendered elements are usually con ned to a few words. 5.  Generic femininum. A nal strategy that foregrounds women speci cally is the generic femininum. That is, using the feminine instead of the generic and overextended masculines to refer to unknown individuals and mixed-gender groups. In Italian, for instance, one would say ‘qualcuna ha una penna?’ (does anyone [fem] have a pen?) instead of ‘qualcuno ha una penna?’ (does anyone [masc] have a pen?), and ‘le maestre’ (the [fem] teachers [fem]) instead of ‘i maestri’ (the [masc] teachers [masc]) when talking about a mixed-gender group of teachers. This solution is clearly as concise as the generic and overextended masculines, but it is not widespread, possibly for fear of p. 333

misunderstanding, given

that the feminine usually refers to women only, or because, due to

sexism, people resist being referred to with feminine terms. Interestingly, however, The University of 16

Leipzig adopted the feminine ‘Professorin’ for all professors (The Local, 2013). On top of the di

culties posed by each of these strategies, none are particularly well suited to refer to 17

genderqueer individuals.

This, however, doesn’t mean that these strategies can’t be used alongside 18

strategies that are more inclusive of genderqueer individuals.

This group of strategies, which we will look

at next, are the gender-neutrality ones, which also have the potential to be concise (depending on its concrete application).

4.2. Gender Neutrality The second family of strategies, as previously mentioned, proposes ways to ‘neutralize’ our language in terms of gender. There are various ways to achieve this in our three languages:

1.  Neutral plural. In German, it is possible to transform a gendered noun into a gender-neutral one by forming a neutral plural. This is achieved by adding the su

x ‘-enden’ to the noun root. For instance,

‘die Lehrenden’ (the ones who are teaching) is not gendered. However, this can only really be used (more or less) straightforwardly in the plural form. In the singular, you’d either have to go with ‘die Lehrende’ (the [fem] teacher) or ‘der Lehrende’ (the [masc] teacher), with each article indicating the teacher’s gender. Again, using the neuter article ‘das’ is not really an option, due to its pejorative connotations—except for the few cases of neuter personal nouns, e.g., ‘das Mädchen’ (the [neuter] girl [neuter]) would be such a case. 2.  Nouns with xed gender. Certain German, Italian, and Spanish nouns have a xed grammatical gender that does not re ect the referent’s gender. For instance, the German ‘Person’ (person) is feminine, and ‘Individuum’ (individual) is neuter, but both can refer to people of any gender. E.g., one might say ‘die Person’ (the person) to refer to anyone regardless of gender and use it to build genderneutral periphrasis. Speci cally in German, however, this construction is not available for all gendered nouns, and sometimes speakers are required to do some verbal gymnastics to make this work. An example in German would be ‘activist’ (Aktivist [masc]/Aktivistin [fem]), which could be substituted with ‘die Person die sich aktivistisch engagiert’ (the person who is engaged in activism). This is, admittedly, not a very snappy way to refer to somebody; it would be somewhat easier to come up with a new gender-neutral word altogether. In other cases, we can make use of other recently developed word constructions: e.g., a gender-neutral way to say ‘doctor’ (Arzt [masc]/Ärztin [fem]) would be to say, e.g., ‘ärztliche Fachperson’ (which roughly translates to ‘medical expert’). You can see that here we had to add ‘Fach’, to the word ‘Person’, as speaking about an ‘ärztliche Person’ (medical p. 334

person) wouldn’t make much sense. ‘Fachperson’, however, can be a neutral way to

indicate that somebody is an expert in something. 3.  (Semi-)Epicenes without gendered satellite elements. Another way to avoid gender marking in Italian and Spanish is to use epicenes and semi-epicenes when possible. These are nouns that have a single form for referents of any gender. Here, gender is only ever visible in satellite elements (namely, articles and adjectives that agree with the noun), some of which are in fact gender-neutral (e.g., numerals greater than one and adjectives ending with ‘-e’/‘-i’), while others, in certain contexts, are optional (e.g., articles in conjunctions) or indistinguishable from one another (e.g., in Italian, de nite articles before vowels). So, a sentence with epicene nouns and no gendered satellite elements lacks explicit gender marking. 4.  Collective nouns. Another gender-neutral strategy is to use collective nouns, which refer to people through their function or participation in a group, instead of as individuals. In particular, these nouns don’t re ect the gender of the group members: unlike the Italian ‘i cittadini’ (the [masc] citizens [masc]) or the Spanish ‘los niños’ (the [masc] children [masc]), which are masculine for all-men and mixed-gender groups but feminine for all-women groups, ‘cittadinanza’ (citizenry) and ‘infancia’ or ‘niñez’ (childhood) can refer to any single-gender and mixed-gender groups alike. 5.  Gender-neutral adjectives and pronouns. In Italian and Spanish, a few adjectives and pronouns lack gender marking, e.g., in Italian, ‘qualunque’ (any), ‘chi’ (who), or ‘coloro’ (those who). One can use them to create gender-neutral expressions, for instance ‘chi legge’ (the one who reads) instead of the generic masculine ‘il lettore’ (the [masc] reader [masc]). Similarly, we’ve seen that in German, the plural article is always ‘die’ (the [plur]), regardless of the referent’s gender. 6.  Avoid personal reference. Finally, in all of our sample languages, one can modify a sentence to avoid personal nouns altogether. For instance, turning a personal sentence into an impersonal one, as in ‘è necessario aspettare fuori’ (it is required to wait outside) instead of ‘i parenti sono tenuti ad aspettare fuori’ (the [masc] relatives [masc] are required [masc] to wait outside).

While these strategies allow one to avoid reference to the referent’s gender, it is not always possible to employ these devices. For one, there aren’t collective nouns for all groups: as we saw, there is no straightforward neutral plural for the German ‘doctor’, and Italian lacks a collective form for, e.g., ‘i condomini’ (the [masc] residents [masc]). Moreover, these strategies heavily impact word choice or sentence structure, and are often di

cult to employ because they require the speaker to do a lot of mental

gymnastics on the spot and rearrange certain words. Drastically rephrasing sentences, as these strategies sometimes require, carries the risk of changing the expressed meaning or unintentionally conveying 19

(audience-)implicatures.

For example, if one says ‘persone che curano’ (people who heal) instead of

‘medici’ (doctors [masc]), their audience might take the lengthy expression as outing the maxim of manner (‘be brief’, as discussed in Grice 1989) in an attempt to implicate that the people at issue are not real p. 335

doctors. But

even when it’s possible to employ these devices so that no unwanted meaning arises, de-

gendering frequently involves redundancy, which constitutes a problem in some contexts, as discussed above. Redundancy is avoided by another group of strategies which consist in substituting gendered su

xes with

di erent symbols or sounds. Since they introduce new linguistic resources, we call these innovative strategies.

4.3. Innovative What is noteworthy about these strategies is that most of them aim at providing a way to refer to genderqueer individuals, but they can also be used to avoid generic and overextended masculines. As we will see, they can make gender visible or obscure it, depending on how they are used. There have been numerous proposals of this kind in our three sample languages.

1.  ‘-@’. In Italian and Spanish, one possibility is to use ‘-@’ instead of a gendered su

x. In Spanish,

for instance, this results in ‘ell@ lee’ ((s)he reads). This symbol is reminiscent of both ‘-o’ and ‘-a’, which are the most common masculine and feminine singular su

xes in these languages. This

symbol was not originally used to refer to genderqueer people, although in Italy some do use it in this way. 2.  ‘-*’. On top of dividing masculine and feminine endings in German, as mentioned above, the asterisk ‘-*’ can be used in Italian, Spanish, and with some German words, to replace gendered su

20

xes entirely.

In Italian, for example, one can write ‘maestr*’ (teacher) instead of the singular

masculine ‘maestro’ or feminine ‘maestra’, as well as instead of the masculine and feminine plurals ‘maestri’ and ‘maestre’. It should be noted that the applicability of this strategy is a bit more limited in German. If one were to write ‘Lehrer*’ (teacher), they would simply be writing the masculine word for teacher, with an added star, which would not ful ll the same neutralizing function as it does in the Italian example. It does work with certain words, however: for example, instead of writing ‘Lieber’ (dear [masc]) or ‘Liebe’ (dear [fem]), one could write ‘Lieb*’ (dear), with the asterisk taking the place where a masculine or feminine ending (here, ‘-er’ or ‘-e’, respectively) would otherwise be. While this works better in Italian and Spanish, the asterisk has various limits. For one, it doesn’t di erentiate between the singular and the plural. And it has a heavy impact on the language in general, much more than the singular ‘they’ pronoun in English, since asterisks must be used for every personal noun and all of its satellite elements. Moreover, no sound corresponds to the asterisk, 21

which is thus restricted to writing.

3.  ‘-x’. ‘Ellx lee’ (they read) is a gender-neutral Spanish sentence built with the ‘-x’ su

x. This

solution was proposed for Spanish and has been adopted, though not widely, for Italian too, where p. 336

some also use ‘-y’ in the same way. This

strategy continues to be used in Spanish, even though

some argue that the ‘-x’ cannot be pronounced. But as López (2020) argues, this is not the case: ‘It is, indeed, pronounced in several di erent ways: traductorx (translator) could be/traduk’torks/, /traduk’toreks/, /tradukto’rekis/ or /traduk’tore/.’ Interestingly, some people advocating for gender-neutral strategies denounce this solution as incorrect: ‘these symbols … make communication impossible, since they are not linguistic signs, there is no way to give pronunciation or meaning to the sentences when reading them’ (Instituto Nacional Electoral, our translation). 4.  ‘-e’. A proposal that has gained considerable uptake in Spanish is the gender-neutral su

x ‘-e’. As

authors such as López (2020) have pointed out, this proposal dates back to at least 1976. As you may have noticed in the grammar section, the gender-neutral 2nd-person plural pronoun, ‘ustedes’, ends with ‘-es’. However, some masculine terms end with ‘-e’ too, such as ‘jefe’ (boss [masc]). Since words like these are gendered, we need to change other parts of our sentences to re ect that we are using them in a gender-fair way. For example, in ‘el jefe está afuera’ (the [masc] boss [masc] is outside), one has to change the article ‘el’ to ‘le’, i.e., ‘le jefe está afuera’, in order to signal that they are using ‘jefe’ as gender-neutral. Despite this di

culty, it remains a

popular option, especially in genderqueer communities. 5.  ‘-u’. A similar solution for Italian involves using ‘-u’ instead of gendered singular and plural su

xes, e.g., in ‘maestru’ (teacher). Like ‘-*’, ‘-u’ does not di erentiate between singular and

plural forms, and so carries with it all of the problems that this implies, as seen above. But, unlike the asterisk, ‘-u’ is a natural addition to spoken language. However, this usage con icts with other functions of ‘-u’ in southern Italian dialects, where it is a masculine su Italian sound could be adapted as a gender-neutral su

x. What’s more, no other

x, as an alternative su

x has to be

constituted by vowels to maintain Italian word structure and prosody. Since Italian word endings are typically not stressed, only ve vowels in the Italian repertoire could be used as su

xes: ‘a’, close-

mid ‘e’, ‘i’, close-mid ‘o’, and ‘u’. Among these, the ‘-u’ is the only vowel that is not already used as a gendered su

x.

6.  ‘-ə’/‘-з’. Recently, a di erent solution has become quite popular in Italy: the ‘schwa’ as a genderneutral su

x. The schwa is a mid-central vowel, and Luca Boschetto, who initially put forward this

solution, proposed to use the short schwa (/ə/) for the singular, and the long schwa, which they 22

indicate with the symbol /з/, for the plural (see Boschetto 2019a). neutral su

Used in this way, this gender-

x, unlike ‘-*’ and ‘-u’, distinguishes between the singular and the plural. However,

vowel length is not a meaningful distinction in spoken Italian (see Gra

and Scalise 2003, 104), so

the di erence is evident in writing but not in speech. Moreover, other supporters of this strategy, such as the linguist Vera Gheno (2022) and the publisher E equ (2020), use the short schwa only. In this variant, the su is that ‘-ə’ is di p. 337

x does not distinguish singular and plural forms, even in writing. Another issue

cult to write as it is not present on many keyboards. This problem can be overcome

by creating a macro for it, as described

on Boschetto’s (2019b) website. Moreover, many mobile

keyboards have recently started to include it. One of the biggest di

culties for this proposal is that

the schwa, while being pronounceable, does not belong to the phonological repertoire of standard 23

Italian.

This has two consequences: it is di

cult to pronounce and recognize for many Italian

native speakers, and it involves not only a morphological change (as with ‘-*’ and ‘-u’), but also an expansion of the phonological repertoire, which is arguably much harder to achieve. This clearly constitutes an obstacle for the schwa proposal. Notwithstanding these di

culties, this strategy has

spread quickly and is being used outside of the activist circles that have adopted other strategies, such as ‘-*’ and the ‘-u’, over the years. 7.  Plural ‘-ai’. Activist Rubynia Reubens (2021) proposes the use of ‘-ai’ as an Italian plural genderneutral su

x. Unlike ‘-*’, ‘-ai’ does have a corresponding sound, namely the diphthong /ai/, which,

unlike the schwa, already belongs to the Italian phonological repertoire (it occurs, for instance in ‘baita’, lodge). Hence, this proposal, just like other innovative strategies, in uences the morphology, but does not require a phonological change, which, as noted, is particularly di

cult to achieve.

Moreover, unlike ‘-u’, it does not con ict with other usages. However, this proposal only concerns plural forms and, therefore, should be combined with another solution for the singular, and must therefore confront that solution’s own limitations. 8.  Alternation. In quite a di erent strategy to the ones described above, many German, Italian, and Spanish genderqueer people use both masculine and feminine forms in alternation. For instance, in German, one can say ‘Ich warte auf jemanden. Er ist zu spät. Oh warte, da ist sie!’ (I am waiting for someone. He is late. Oh wait, there she is!), referring to one single person by using the masculine 24

pronoun ‘er’ rst and the feminine ‘sie’ afterward.

This strategy is popular in some genderqueer

communities in our sample languages because it does not require the introduction of new neo25

morphemes.

However, unlike other innovative strategies, this one cannot be extended to address

the sexist practice of, for example, addressing women and mixed-gender groups with the masculine. 9.  Neo-pronouns. Since there is no equivalent for the singular ‘they’ in German, Italian, and Spanish, one option is to introduce neo-pronouns. Some German speakers have adopted pronouns such as 26

‘hen’,

‘hän’, ‘ze’, ‘sier’, etc., which function exactly like ‘er’ (he) and ‘sie’ (she). For example, you

could say ‘Ze geht einkaufen’ (Ze is going shopping), referring to a genderqueer person. In Italian, neo-pronouns are often based on the various gender-neutral su

xes, so ‘l@i’, ‘l*i’,

‘ləi’, and ‘lai’ are alternatives to the gendered ‘lui’ (he) and ‘lei’ (she). Notice, though, that the ‘u’ strategy cannot be used to create a neopronoun, since the existing gendered ones di er by their central vowels, and ‘lui’ is already a masculine form. While pronouns can often be omitted in Italian, declaring one’s pronouns is becoming increasingly widespread in activist circles in Italy.

However, people tend to mention English pronouns in this context, which suggests that this habit is in uenced by practices in English-speaking countries. p. 338

10.

‘Gendern’ (again). Recall the German strategy of using an underscore or asterisk before a feminine su

x. The ‘gender-gap’ and ‘gender-star’ are often interpreted in a broader way: some

argue that the underscore symbolizes room for those ‘in-between’ men and women and that the asterisk multiplies the possibilities beyond ‘man’ and ‘woman’ (e.g., see AG Feministisch Sprachhandeln HU 2015, 20). Unfortunately, this meaning often gets lost, partly because these strategies are usually associated with the visibility goal. In many ways, these strategies don’t problematize the gender binary, and may entail misleading messages about gender-queerness: e.g., not all genderqueer people experience their gender as ‘somewhere in-between man and woman’. 11.  Dynamic Underline. Here, the idea is to introduce an underline into words, somewhat similar to the ‘gender-gap’, but in a non-static way. The idea is that this can challenge the conventional assumption that people can only fall into two gender categories (and are cis by default), and refer to genderqueer people in new ways. The dynamic underline ‘moves’ through words, so rather than writing ‘Lehrer_innen’ (teachers) one might encounter ‘Leh_rerinnen’ or ‘Lehrerin_nen’. It can also be introduced in pronouns, articles or adjectives (e.g., d_ie Lehreri_nnen). Again, however, it is di

cult to apply this in spoken language, and it encounters resistance in professional contexts, even

more so than a static underline (e.g., see Hornscheidt 2012; AG Feministisch Sprachhandeln HU 2015). All the innovative strategies presented share some advantages and some limitations. They can all be used to refer to genderqueer people, don’t require changes in word choice or sentence structure, and are not redundant, a feature that makes them suitable for length-constrained contexts such as titles or slogans. However, they all require a change at least at the morphological level. Such changes are not always straightforward, especially when it comes to elements where masculine and feminine are distinguished by more letters, as in the Italian su

xes ‘-tore’ (for masculine) and ‘-trice’ (for feminine), or the articles ‘il’

(for masculine) and ‘la’(for feminine): which should constitute the basis for the innovative form? On top of these technical issues, innovative strategies, by modifying su

xes, make it di

cult to recognize

words for people with dyslexia, as well as for screen readers used by people with visual impairments (Iacopini 2021). While screen readers can be programmed to read innovative su

xes correctly, a solution

for people with dyslexia is much harder to nd. This is especially problematic as it puts anti-sexist and antiableist needs in opposition.

5. Where to go from here? So far, we have provided a (limited) overview of the discussion surrounding gender and language, p. 339

introduced the heavily gender-loaded grammars of German, Italian,

and Spanish, and discussed various

proposals that respond to these binary grammatical systems: visibility, gender-neutrality, and innovative strategies. We’ve seen that feminist and trans concerns in our languages have to go far beyond pronouns, and involve broader grammatical structures in multiple ways. So, where do we go from here? This concluding section will give an outlook on some further discussions and future perspectives. First, we’ll brie y revisit the central issues introduced in the beginning, then think about the question of whether there are certain strategies we should prefer over others, and nally make a few remarks about how it is essential to understand that the kind of speech practices we discussed in this chapter are social practices—practices that evolve as time goes on.

At the beginning of this chapter, we focused on three structural loci of sexism and trans-exclusion in German, Italian, and Spanish: generic masculines and overextended masculines, both of which render women and gender minorities invisible through the use of words that encode androcentrism and masculine terms; and misgendering genderqueer people, due to the lack of linguistic resources to correctly refer to them. Of these, the issues caused by the overextended masculine are speci c to heavily gender-loaded languages, while the others are shared by lighter gender-loaded languages such as English. We focused especially on the issues relating to grammatical gender, for which diverse strategies have been proposed. As we’ve seen, some of these were put forward to avoid the generic and overextended masculine, but are not e ective in referring to genderqueer individuals; some even reinforce binarism. The strategies that can be used to refer to genderqueer people consist in avoiding gender markers or introducing new linguistic resources into the language. Since gender loading is pervasive in German, Italian, and Spanish, these solutions do not simply concern pronouns, but involve new word endings and articles. Indeed, it is quite pointless to use gender-neutral pronouns while having to mark every noun, article, and adjective with a binary gendered su

x. That is to say, in gender-loaded languages, linguistic reforms extend far beyond

pronouns. One important question is whether it’s best to use one and the same strategy to address all of the issues, or if we should adopt separate solutions to refer to generic individuals, genderqueer people, and mixed-gender groups. And are there certain strategies we should prefer over others? Many activists and scholars endorse repeating the same terms in various gendered forms, with the argument that this best ensures the visibility of all genders. For instance, Giuliana Giusti argues, similar to Cameron, that neutralization ‘can work when [women’s] visibility is not brought into question by stereotypes …. It is not functional, however, to ght gender stereotypes …. We know, indeed, that naturally epicene languages such as English are not exempt from a masculine interpretation of all plurals, especially of those of prestigious or stereotypically masculine role names. De-gendering, thus, would favor a biased interpretation of gender stereotypes’ (Giusti 2022, 11; our translation). That is, according to Giusti, making language neutral is not the best choice if we need to challenge gender stereotypes. Indeed, she cites p. 340

psycholinguistic evidence (such as Gabriel et al.’s 2008 and Misersky et al.’s 2014), which shows that biased interpretations were more common in languages where gender is not marked: gendered forms allow us to make the presence of women explicit. Other experiments (Moulton et al. 1978; Khosroshahi 1989; and Gastil 1990) found that the gender-neutral pronoun ‘they’ evokes more masculine images than compounding both masculine and feminine pronouns (e.g., ‘he or she’). Notice, however, that these experiments are now dated, and participants might have responded according to a binary paradigm, without considering the possibility of genderqueer individuals. Giusti’s point is also intuitively shown by the famous surgeon riddle: a father dies and his son is brought to the hospital where, however, the surgeon says ‘I cannot operate on him: he’s my son.’ Recently, the BCC got various creative responses to this, which show that the interviewees still simply assumed that the surgeon had to be a man (see The Fifth Floor, BBC Sound, 2018). This assumption, not presupposed by the word itself, as Eckert and McConnell-Ginet (2013) point out, is part of our conceptual baggage, namely ‘ “meaningful” but non-semantic associations with a word’ (2013, 164). Notice that this association would not have been possible if the word ‘surgeon’ had a distinct feminine form in English, as ‘doctor’ does in German, Italian, and Spanish. This is Giusti’s point: pervasive grammatical gender can be a resource in challenging gender stereotypes such as the one assumption that a surgeon must be a man.

In light of this, we could opt for separate solutions and use gender-neutral or innovative strategies to refer to genderqueer individuals, but visibility strategies for generic individuals and mixed-gender groups. As observed above, however, splitting and its variants usually consist in repeating a term’s masculine and feminine forms, which reinforces binarism. To avoid this, one could also add the term’s innovative form, e.g., in saying ‘tutti, tutte, e tuttu’ (everyone [masc], everyone [fem], and everyone [innovative]), a salutation used by some in Italy. This combines the need to make both women and genderqueer people visible. However, repeating all three forms is complex in heavily gender-loaded languages, and so the feasibility of such a strategy beyond salutations is unclear. But even if triple splitting were viable, it would not be an egalitarian solution. Indeed, as argued by Dembro

and Wodak (2018), men and women would then have a dedicated linguistic form, while all

genderqueer identities would be lumped together. To avoid this issue, the genderqueer form could occupy all the roles: to refer to genderqueer people, generic individuals, and mixed-gender groups. One would simply say ‘tuttu’ (everyone [innovative]) rather than ‘tutti, tutte, e tuttu’ (everyone [masc], everyone [fem], and everyone [innovative]). In response, one could embrace Giusti’s point and worry that this solution shadows women and fails to challenge gender stereotypes. However, while gender-neutral strategies that rely on existing linguistic resources may sometimes go unnoticed or, as argued earlier, be misunderstood by the audience, innovative strategies are extremely visible and clearly signal the speaker’s intention to refer to people of any gender, beyond the binary. Some people, indeed, argue that we ought to use the options that cause the most disruption in order to trigger a thought process about the reasons behind it (e.g., see Cameron 1992: 125–126 on feminine generics, or Hornscheidt 2012: 293–321 on speech p. 341

interventions). If this is correct, innovative strategies might elicit di erent

responses from those

recorded in the psycholinguistic experiments conducted so far with gender-neutral elements like ‘they’. This hypothesis seems to us worth exploring in further studies. In turn, one might worry that using the innovative form for everybody would make men and women the exception. Indeed, as Alma Graham (1974) has observed, ‘If you have a group half of whose members are A’s and half of those members are B’s and if you call the group C, then A’s and B’s may be equal members of group C. But if you call the group A, there is no way that B’s can be equal to A’s within it. The A’s will always be the rule and B’s will always be the exception—the subgroup, the subspecies, the outsiders.’ However, while the masculine refers to those who are already considered the rule, using the innovative form for all could contribute to making genderqueer people visible and empower them. Finally, it is still an open matter which strategy works best to refer to genderqueer individuals in each language. This is especially relevant given that genderqueer people are starting to get their gender recognized on birth certi cates, as recently happened in Mexico for the rst time (Hilkinger 2022). This is very important for various reasons, including the fact that there is no agreed-upon way, institutionally or otherwise, to talk about genderqueer people. While presenting the various options, we highlighted their advantages and limitations. However, no strategy in particular stands out as the best. Hence, one might also wonder whether it is even desirable to nd a ‘once and for all’ solution, especially given that language is constantly changing. As we’ve highlighted in the introduction, the point of this chapter is not to argue for the superiority of one strategy over the other—even though we did point out some problems that will be easier or harder to resolve, depending on the goals the strategies ful ll. This point, however, seems particularly important to us. If we look at how conversations about gender and exclusion have developed in our sample languages over the past years alone, it is predictable that more is to come. Linguistic interventions—and the innovative strategies we introduced for sure can count as such—are not detached from the social reality in which they arise, and this social reality will form new demands as time goes on.

We think, therefore, that it is essential to stress that the speech practices and interventions we’ve discussed in this chapter are social practices that evolve and change over time. It is partly due to recognizing this that some proponents of innovative strategies argue that they are not proposing a permanent solution, but an experiment (see, e.g., Sciuto 2021 and Gheno 2022). And, as Cameron highlights, halting the process of linguistic change at one particular point is not straightforward: Feminists have often reminded those critics who complain about women changing the language that the language changes by itself, and this process cannot be halted. Of course, this is not quite accurate—language does not exist independently of speakers, and it is they who introduce innovations—but the point is well taken. Our vocabulary and semantics cannot remain xed for all time, as some commentators, even today, might prefer. New words are coined or borrowed or made out of combined parts from existing words; the meanings of old words gradually shift. (Cameron 1992, 112–113) p. 342

At the same time, we understand that this solution might be somewhat unsatisfying. If we want to do right by women and genderqueer individuals, it is very useful, to say the least, to have concrete strategies to point to, recommend, or even insist upon. It is important for many genderqueer people that institutions and broader communities of speakers adopt a way to refer to them, and constantly putting out di erent proposals can make this more di

cult. In this respect, studying these strategies and converging on the

most suitable might increase the chances that any of them ‘stick’. However, understanding the fact that our approaches might have to change as time goes on shouldn’t deter us from doing the best we can with the available solutions at a given moment. Advocating for awareness that language is changing, and that it will keep changing, seems to be just as important as using the most inclusive terminology for everyone. Of course, there are still many points to consider. But we hope that, by pointing to some of the issues surrounding gender in our sample languages, our discussions can spark new research that look into making heavily gender-loaded languages less androcentric, binary, and fairer to women and genderqueer individuals alike.

Notes *

First and foremost, we are very grateful to Jennifer Saul for her input and feedback throughout the di erent stages of working on this chapter. We also want to thank Emma Bolton for both her input as well as for proofreading our chapter. Last but not least, many thanks to the editors, Luvell Anderson and Ernie Lepore, as well as to an anonymous referee, whose valuable comments greatly improved our chapter.

1.

The relationship between gender and language has also received significant attention in linguistics and psychology, as well as in activist spaces outside the academy.

2.

We have chosen these as they are the authorsʼ native languages.

3.

For an overview see e.g., Saul and Díaz-León (2016).

4.

When we talk about grammatical gender here, we refer to the linguistic notion concerning agreement patterns based on classes of nouns. Of course, gender as it appears in the social world is central to this discussion as well, given that the complexity of gender is one of the reasons why we need to have this linguistic discussion in the first place. However, how to precisely define and understand gender is a research area on its own, one we cannot go into detail with here (see Mikkola 2008 for an overview).

5.

See Kapusta (2016) for a discussion of the wrong of misgendering in general.

6.

When relevant, we indicate grammatical gender and number in square brackets.

p. 343

p. 344

7.

Wallace v. Cegavske. Supreme Court of the State of Nevada. October 19, 2016. https://www.casemine.com/judgement/us/5914d702add7b0493486a84b#.

8.

See more discussions on the history of the singular ʻtheyʼ in Bodine (1975).

9.

An interesting exception in Italian concerns ʻautomobileʼ (car), which is feminine because Gabriele DʼAnnunzio argued that it should be, since cars were supposed to be driven by men. See Dizionari Corriere (n.d.).

10.

Although there is a neuter option for articles and pronouns, its application is more or less fixed for certain words and contexts, and using it to refer to people in other cases can have pejorative connotations.

11.

The pronoun ʻesʼ is equivalent to the English ʻitʼ. You can use it in impersonal sentences such as ʻes regnetʼ (itʼs raining). In some exceptional cases, ʻesʼ can refer to humans.

12.

Traditionally, these were Italian direct object pronouns, but they are increasingly substituting the traditional subject ones ʻegliʼ (he) and ʻellaʼ (she).

13.

Nouns that are indeclinable for gender both in the singular and in the plural are called epicenes, while those which have one form in the singular but two in the plural (one feminine and one masculine) are called semi-epicenes by Formato (2019).

14.

Some, though, argue that article splitting is ungrammatical as, in Italian, articles cannot be joined with a conjunction (see, for instance, Giusti 2022).

15.

These options are not exhaustive, and people tend to come up with new ways to make clear the ʻgenderingʼ e ect. E.g., Lehrer:innen or Lehrer/innen are also popular. For an overview see, e.g., AG Feministisch Sprachhandeln HU (2015, 13).

16.

Itʼs important to highlight that this strategy can be problematic for genderqueer people: using the feminine form to refer to generic or unknown individuals and to groups with genderqueer members gives rise to a discrepancy analogous to the one produced by the generic and overextended masculines. Moreover, the generic femininum, just like the other visibility strategies, does not solve the problem of structural misgendering. While this solution makes women (especially) visible, it can contribute to the invisibilization of genderqueer people in language.

17.

An exception might be the use of underscore (or asterisk or slash) in singular German. E.g., a genderqueer German speaker might be referred to by using the underscore, as in ʻMeine Freund_in ist ein_e Lehrer_inʼ (My friend [masc_fem] is a [masc_fem] teacher [masc_fem]). While this mentions both the masculine and feminine, as when referring to mixedgender groups, for instance, here the idea is to create something that ʻgoes beyondʼ that, a gender reference that is neither masculine or feminine. Weʼll come back to this shortly.

18.

The claim here is not that the visibility project is less important than the gender-neutrality project. It is that visibility strategies alone are unable to contribute to the visibility of genderqueer people. We do not make a claim about the importance of one project over the other. This would exceed the aims of this chapter, which is to expand the philosophical conversation about the relationship between gender and language to languages other than English.

19.

Saul (2002) defines audience-implicatures as claims that the audience believes the speaker is trying to conversationally implicate. Hence, it is possible to convey an audience-implicature without wanting to, or even realizing that one has.

20.

This is true in most cases, namely when the su ix consists of one vowel. However, as we will see in the discussion, some Italian nouns have longer su ixes that cannot be replaced by one symbol.

21.

Some people who use the asterisk in writing elide the last vowel when talking. However, this is technically a di erent gender-neutral strategy, namely the apocope, and presents di erent issues, especially a change in prosody.

22.

Notice, however, that the IPA symbol /з/ corresponds to the open-mid central unrounded vowel.

23.

It is, however, a phone of some Italian regional varieties and dialects, such as Sicilian and Neapolitan.

24.

This example is oriented around one mentioned by Dembro and Wodak (2018, 385). They use this to show how, unlike in the case of the singular ʻtheyʼ, uttering a sentence like this with alternating pronouns in English causes problems. However, in languages such as German, using multiple pronouns can help genderqueer individuals to de-specify their gender, so to speak. Weʼll say a bit more about this later on.

25.

Rather, this strategy uses existing morphemes in novel ways. So, even though it does not introduce new morphemes, this strategy introduces new usage rules and can, therefore, be considered an innovative strategy.

26.

ʻHenʼ is also the Swedish gender-neutral pronoun. This has been adopted in some German contexts as well, but is much less used and accepted than in the Swedish context (see Frankfurter Allgemeine Zeitung 2015).

References AG Feministisch Sprachhandeln der Humboldt Universität zu Berlin. (2015). Was tun? Sprachhandeln—aber wie? W_Ortungen statt Tatenlosigkeit. https://feministisch-sprachhandeln.org/. WorldCat Bodine, Ann. (1975). ʻAndrocentrism in Prescriptive Grammar: Singular “They”, Sex-Indefinite “He”, and “He or She”.ʼ Language in Society 4 (2): 129–146. https://doi.org/10.1017/S0047404500004607. Google Scholar WorldCat Boschetto, Luca. (2019a). ʻCome si scriveʼ. Italiano Inclusivo (website). https://italianoinclusivo.it/scrittura/. WorldCat Boschetto, Luca. (2019b). ʻStrumentiʼ. Italiano Inclusivo (website). https://italianoinclusivo.it/strumenti/. Google Scholar WorldCat Cameron, Deborah. (1992). Feminism and Linguistic Theory. 2nd edition. London: Palgrave Macmillan. Google Scholar Google Preview WorldCat COPAC Cameron, Deborah. (1998). The Feminist Critique of Language: A Reader. London: Psychology Press. Google Scholar Google Preview WorldCat COPAC Corbett, Greville G. (1991). Gender. Cambridge, UK: Cambridge University Press. https://doi.org/10.1017/CBO9781139166119. Google Scholar Google Preview WorldCat COPAC Dembro , Robin, and Daniel Wodak. (2018). ʻHe/She/They/Zeʼ. Ergo, an Open Access Journal of Philosophy 5 (14): 371–406. https://doi.org/10.3998/ergo.12405314.0005.014. Google Scholar WorldCat Dembro , Robin, and Daniel Wodak. (2021). ʻHow Much Gender Is Too Much Gender?ʼ In The Routledge Handbook of Social and Political Philosophy of Language, edited by Justin Khoo and Rachel Sterken, 362–376. New York: Routledge. Google Scholar Google Preview WorldCat COPAC Dizionari Corriere. (n.d.). ʻAutomobile: Un nome dalla storia bisexʼ. In Dizionari—Corriere Della Sera. Accessed March 31, 2022. Milano: RCS Mediagroup. https://dizionari.corriere.it/dizionario-si-dice/A/automobile.shtml. Google Scholar Google Preview WorldCat COPAC Eckert, Penelope, and Sally McConnell-Ginet. (2013). Language and Gender. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC E equ. (2020). Lo schwa secondo noi. https://www.e equ.it/lo-schwa-secondo-noi/. WorldCat Formato, Federica. (2019). Gender, Discourse and Ideology in Italian. Cham, Switzerland: Palgrave Macmillan. Google Scholar Google Preview WorldCat COPAC Frankfurter Allgemeine Zeitung. (2015). ʻEr, sie, “hen”.ʼ Frankfurter Allgemeine Zeitung. https://www.faz.net/aktuell/gesellscha /hen-schweden-fuehrt-geschlechtsneutrales-pronomen-ein-13539757.html. WorldCat Frye, Marilyn. (1983). The Politics of Reality: Essays in Feminist Theory. Trumansburg, NY: Crossing Press. Google Scholar Google Preview WorldCat COPAC p. 345

Gabriel, Ute, Pascal Gygax, Oriane Sarrasin, Alan Garnham, and Jane Oakhill. (2008). ʻAu Pairs Are Rarely Male: Norms on the

Gender Perception of Role Names across English, French, and Germanʼ. Behavior Research Methods 40 (1): 206–212. https://doi.org/10.3758/brm.40.1.206. Google Scholar WorldCat Gastil, John. (1990). ʻGeneric Pronouns and Sexist Language: The Oxymoronic Character of Masculine Genericsʼ. Sex Roles 23 (11– 12): 629–643. https://doi.org/10.1007/bf00289252. Google Scholar WorldCat Gheno, Vera. (2022). ʻSchwa: Storia, motivi e obiettivi di una propostaʼ. Lingua Italiana, March 21. Treccani. https://www.treccani.it/magazine/lingua_italiana/speciali/Schwa/4_Gheno.html. WorldCat Giusti, Giuliana. (2022). ʻInclusività della lingua italiana, nella lingua italiana: Come e perché. Fondamenti teorici e proposte operativeʼ. DEP—Deportate, Esuli, Profughe. Rivista Telematica Di Studi Sulla Memoria Femminile 48 (1): 1–19. Google Scholar WorldCat Gobierno de México. (n.d.). Catálogo de las Lenguas Indígenas Nacionales. Instituto Nacional de Lenguas Indígenas. https://www.inali.gob.mx/clin-inali/. Accessed March 28, 2022. Google Scholar Google Preview WorldCat COPAC Gra i, Giorgio, and Sergio Scalise. (2003). Le lingue e il linguaggio: Introduzione alla linguistica. 2nd edition. Bologna: Il Mulino. Google Scholar Google Preview WorldCat COPAC Graham, Alma. (1974). Letter. The Columbia Forum 3 (4): 618. https://hdl.handle.net/2027/mdp.39015008922596. Google Scholar WorldCat Grice, Paul. (1989). Studies in the Way of Words. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Hilkinger, Aiko. (2022). ʻMexico Issuing First Non-binary Birth Certificate Is the Futureʼ. Remezcla, February 23 2022. https://remezcla.com/culture/mexico-non-binary-birth-certificate/. WorldCat hornscheidt, lann. (2012). feministische w_orte: Ein lern-, denk- und handlungsbuch zu sprache und diskriminierung, gender studies und feministischer linguistik. Frankfurt: Brandes & Apsel Verlag. Google Scholar Google Preview WorldCat COPAC Iacopini, L. (2021). Lo schwa (ǝ) che rende lʼinclusione inaccessibile. Webaccessibile.org. https://www.webaccessibile.org/approfondimenti/lo-schwa-%C7%9D-che-rende-linclusione-inaccessibile/ WorldCat Instituto Nacional Electoral. (n.d.). Diagonales y Arrobas. Accessed March 28, 2022. https://igualdad.ine.mx/lenguajeincluyente/tema-diagonales.html. WorldCat Kapusta, Stephanie Julia. (2016). ʻMisgendering and Its Moral Contestabilityʼ. Hypatia 31 (3): 502–519. https://doi.org/10.1111/hypa.12259. Google Scholar WorldCat Khosroshahi, Fatemeh. (1989). ʻPenguins Donʼt Care, but Women Do: A Social Identity Analysis of a Whorfian Problemʼ. Language in Society 18 (4): 505–525. https://doi.org/10.1017/s0047404500013889. Google Scholar WorldCat Liberman, Mark. (2008). ʻSex-Neutral “he”: The Constitutional Questionʼ. Language Log https://languagelog.ldc.upenn.edu/nll/? p=18.

WorldCat López, Ártemis. (2020). ʻYou, Me, Hir, and Non-binary Languageʼ. ATA Spanish Language Division, October 7. https://ataspd.org/2020/10/07/you-me-hir-and-non-binary-language/. WorldCat Miller, Casey, and Kate Swi . (1980). The Handbook of Nonsexist Writing. London: Womenʼs Press. Google Scholar Google Preview WorldCat COPAC Misersky, Julia, Pascal M. Gygax, Paolo Canal, Ute Gabriel, Alan Garnham, Friederike Braun, et al. (2014). ʻNorms on the Gender Perception of Role Nouns in Czech, English, French, German, Italian, Norwegian, and Slovakʼ. Behavior Research Methods 46 (3): 841–871. https://doi.org/10.3758/s13428-013-0409-z. Google Scholar WorldCat p. 346

Moulton, Janice, George M. Robinson, and Cherin Elias. (1978). ʻSex Bias in Language Use: “Neutral” Pronouns That Arenʼt ʼ. American Psychologist 33 (11): 1032–1036. https://doi.org/10.1037/0003-066x.33.11.1032 Google Scholar WorldCat Pezzini, Barbara. (2015). ʻImplicito ed esplicito nel rapporto circolare tra genere e dirittoʼ. In Lucia Morra and Barbara Pasa (eds.) Questioni di genere nel diritto: Impliciti e crittotipi. 201–234. G. Giappichelli Editore. Google Scholar Google Preview WorldCat COPAC Reubens, Rubynia. (2021). ʻPlurale in -aiʼ. Instagram video. https://www.instagram.com/p/CQWDtPRqPQt/. Google Scholar WorldCat Saul, Jennifer. (2002). ʻSpeaker Meaning, What is Said, and What Is Implicatedʼ. Noûs 36 (2): 228–248. https://doi.org/10.1111/1468-0068.00369. Google Scholar WorldCat Saul, Jennifer, and Esa Díaz-León. (2016). ʻFeminist Philosophy of Languageʼ, The Stanford Encyclopedia of Philosophy, Edward N. Zalta & Uri Nodelman (eds.). https://plato.stanford.edu/entries/feminism-language/. Google Scholar Google Preview WorldCat COPAC Statuto Albertino. (1848). https://it.wikisource.org/wiki/Italia,_Regno_-_Statuto_albertino. WorldCat The Fi h Floor (2018). Riddle Me This. BBC Sounds. https://www.bbc.co.uk/sounds/play/p062qc0c. WorldCat The Local. (2013). Uni switches to feminine professor titles for men. The Local. https://www.thelocal.de/20130605/50095/ WorldCat

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

p. 349

15 Privacy, Critical De nition, and Racial Justice  Anita L. Allen https://doi.org/10.1093/oxfordhb/9780192844118.013.38 Published: 22 May 2024

Pages 349–363

Abstract This chapter points the way to a philosophical approach to conveying the meaning of privacy called critical de nitional facilitation. This approach contrasts to another “analytic de nitional prescription,” which has been commonplace in the eld of philosophy for decades and which critical de nitional facilitation ought to supplement or supplant. The motivation for shifting emphasis from prescriptive analysis to critical facilitation—a shift from an emphasis on de nitional meaning to de nitional signi cance is the political and urgent nature of privacy discourse in contemporary life. Privacy discourse is a way of thinking, speaking, and acting with serious implications for moral, human, civil, and legal rights. This chapter addresses the relationship between de nitional facilitation, politics and justice for African Americans.

Keywords: privacy, data protection, critical definitional facilitation, politics, African American, race, essentially contested concepts, definition, analytic, ambiguity, vagueness Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction I want to point the way to a philosophical approach to conveying the meaning of privacy that I call critical de nitional facilitation. This approach contrasts to another approach, analytic de nitional prescription, which has been commonplace in the eld of philosophy for decades and which critical de nitional facilitation ought to supplement or supplant. The motivation for shifting emphasis from prescriptive analysis to critical facilitation—a shift from an emphasis on de nitional meaning to de nitional signi cance, a distinction rst popularized by E. D. Hirsh (Hirsh 1984; see Carter and Marshall 1955)— is the political and urgent nature of privacy discourse in contemporary life. Privacy discourse is a way of thinking, speaking, and acting with serious implications for moral, human, civil, and legal rights (Zubo

2019). Here I address the

relationship between de nitional facilitation, politics, and justice for African Americans.

2. The Myth of Muddle Varied popular, constitutional, tort, and statutory legal uses of “privacy” were in circulation by the nal third of the twentieth century. Americans used the word “privacy” broadly to denote di erent kinds of conditions and powers—informational, physical, decisional, propriety, associational, and intellectual— relating to a host of values and things valued—including freedom, dignity, independent thought, p. 350

tranquility, limited government, secrecy, intimacy, equality, identity, and reputation (Allen 2013). This conceptually complex reality of ordinary and o

cial discourse would grow to trouble some analytical

philosophers who were committed to linguistic clarity and precision. Moreover, socially conservative philosophers, jurists, and policymakers alike were disturbed that the new privacy jurisprudence might be both “confused” but also deployed for unwelcome progressive causes tied to race, criminal defendants’ rights, euthanasia, and birth control and abortion rights. Much of the earliest philosophical work respecting privacy focused on how to precisely de ne “privacy” and to escape confusion, ambiguity in usage, and vagueness. What does the term denote and connote? What, in broader senses, does it mean? Is it a value or a state of a airs? How, if at all, should “privacy” be used in ethics, policy, and law? Is the term being used appropriately by the law courts? Of course, philosophers also engaged in theorizing about the ethical and practical value of privacy. Why is it important? What good does it do and what are its downsides and risks? Taking a stand on the value of privacy seemed to presuppose a de nitional perspective about what is or should be talked about under the rubric of “privacy.” Philosophers presumed some sort of de nitional analysis was necessary to reign in privacy usage, wary that both de nitions and the methodology and goals of de nition were themselves philosophically divergent and contestable. Like most nouns in the English language, “privacy” is both vague and ambiguous. (Sorensen 2018, sec. 2) For this reason, it is understandable that the concept of privacy seemed murky and confusing to some of the academics who rst sought to make sense of it through the e orts of developing a philosophical de nition of the term “privacy.” With a substantial body of privacy scholarship from many disciplinary perspectives currently at the philosopher’s disposal, any sound basis for continuing to describe privacy as a special muddle has completely disappeared. Today, as numerous scholars have explained, like it or not, the language of privacy clearly, broadly, variously, and controversially encompasses concerns about seclusion, solitude, anonymity, secrecy, con dentiality, data protection, data security, repose, reserve, bodily integrity, intimacy, intimate relationships, control over attributes of identity and likenesses, reputation, and autonomy. There is little point to continuing to declare some of this usage illegitimate and prescribing circumscribed uses; it makes more sense to seek to understand what those speaking and writing about privacy are attempting to convey, and to facilitate their just causes. This is because what are framed as

privacy-related values demonstrably encompass vital freedoms and virtues, from moral independence and political immunities to trust. Academics began puzzling over privacy in the 1950s. The language of privacy rst became a focal point of sustained philosophical analysis in the United States in the 1970s, prompted by a dramatically increased volume of popular and legal applications of the term “privacy” in relation to government surveillance, computer technology, Bill of Rights protections, sexual freedom, and healthcare (Schoeman 1984). Academics greeted the arrival of privacy in everyday discourse with mixed enthusiasm. Not only did the rise p. 351

of “privacy” coincide with the bid for heightened individual rights and

greater equality opposed by some

conservatives and traditionalists, but academics who felt the need to precisely de ne the term found the task confounding. De ning “privacy”—under any de nition of “de nition” (Gupta 2021)—was not to be a quick and easy victory, because doing so has generally meant describing necessary and su

cient conditions

for its proper use, where coherence with paradigm cases of proper usage is a chief guide rail. Rather than see their di

culty as identical to the di

culty scholars have had de ning other terms

commonly used in normative discourse, such as “freedom” and “equality,” some pronounced that “privacy” referred to uniquely vague, ambiguous, or indeterminate conditions, values or rights—for example, as “the Cheshire cat of values: not much substance, but a very winning smile” (Epstein 2014: 149).

3. Elite Scholars Weighed In and Leaked Out University of Chicago’s Edward Shils published an elegant and in uential article in 1966 arguing that privacy is essential for respecting our humanity and civility, but led o

by characterizing privacy as a

“vague idea” that is hard to get into “proper perspective” (Shils 1966: 281–306). He intended no disparagement of privacy or the task of de ning it. At the dawn of the information age, in one of the very rst books (if not the rst book) about privacy published by an American, Harvard Law School professor Arthur Miller, decried an “assault on privacy,” but at the same time asserted that the “concept of privacy is … exasperatingly vague and evanescent” (Miller 1971: 25). Miller likewise intended no disparagement. Notable academics took a di erent direction to conclude that privacy talk could or should be avoided to escape its vagueness and ambiguity. In 1975, in one of the early philosophical papers about privacy, prominent Massachusetts Institute of Technology analytic moral philosopher Judith Jarvis Thomson quipped that “the most striking thing about the right to privacy is that nobody seems to have any very clear idea what it is” (Thomson 1975: 295). A gesture of disparagement, Thomson suggested that the law could do without privacy talk. Several years later, in the wake of progressive and liberal uses of privacy discourse, Robert Bork, a Yale Law School professor and later a President Ronald Reagan–era federal judge, dismissed privacy talk as ‘‘intellectually empty” and “utterly specious,” as well as “unprincipled” and “unconstitutional” (Boyle 1991). These observations from elite, in uential academics, some disparaging of privacy, some not, helped encourage a generation or two of optimistic scholarly e orts to carefully de ne “privacy” for ethics and law (see, e.g., Allen 1988: 5–34). William Parent, for example, made an attempt to prescribe a tidy de nition of p. 352

privacy su

cient for legal

uses, built on a belief that “American privacy jurisprudence is in conceptual

shambles” and that the “courts continue to work with spurious and sometimes even irreconcilable de nitions” (Parent 1983: 305–338). Ruth Gavison, an Israeli student of philosopher H. L. A. Hart, published an important article in 1980 in the Yale Law Journal (Gavison 1980). Gavison’s article critically and analytically engaged the philosophical literature to suggest what she considered a structured, clearheaded way to think about privacy in common-law doctrines and constitutional law doctrines by noting that many instances of “not letting people alone” cannot be described as invasions of privacy (e.g., requiring people to pay taxes (438). Gavison argued that important conceptual clarity can be gained by distinguishing “privacy”

in a sense of freedom from government prohibitions on choice (e.g., to use birth control, have an abortion, marry outside your race (436) from privacy proper, de ned as she did as limited access to persons and personal information (e.g., seclusion, secrecy, solitude, and anonymity (428). Decades later, echoing Parent and Gavison, sounding the same note, Daniel Solove prescribed a taxonomy of privacy to address his belief that legal applications of privacy were “in disarray” but subject to tidying (Solove 2006: 477). At the same time that they spurred optimistic e orts to carefully de ne privacy, the group of early elite scholars created the stubborn and unwarranted precedent of describing privacy discourse as subject to potentially incurable and pernicious vagueness, ambiguity, and confusion. Gavison’s in uential article caste shade on referring to reproductive rights or the right to die as “privacy” rights, as had the earlier article by Judith Jarvis Thomson. Thomson claimed that what are termed “privacy” rights could be reframed as liberty or property rights, rendering vague and ambiguous privacy discourse super uous (Thomson 1975: 295–314). These philosophers implicitly construed the normative belief that privacy is invaded by legal restrictions on what one can do with one’s own body as imprecise thinking. In e ect, by virtue of their particular prescriptive de nitional analytics, Gavison and Thomson were “taking sides” against a progressive movement in the law built on more capacious understandings of privacy than they condoned— understandings that facilitated the just cause of shifting power, inter alia, from men to women, from doctors and hospitals to families, and from straight people to LGBTQ peoples. Writing in the 1970s before the full ascendency of applied, public, and critical philosophy, trends that underscore the real-world implications and impact of philosophical work, Thomson may have had no real expectation that her law-related suggestions would have signi cant practical in uence. The same could be said of Gavison. However, as lawyers, Bork and Miller could have expected their characterizations of privacy to have clout beyond the academy and to be put to both practical and political uses. Over time, a stream of prescriptive de nitional analytic precedent, some disparaging of privacy talk and despairing, seeped across disciplines and leaked through crevices in the walls of the university. It united with a stream of calls by technology innovators, business, and the government for nuanced understandings of what it means to respect and protect privacy. Among the seepage was a thesis, still in circulation in legal and policy circles today, that privacy is intractably vague.

p. 353

4. Intractable Vagueness Thesis: Not “So Amorphous as to Defy Description” Since the 1970s, many philosophers and legal theorists have believed they could function as dispassionate housekeepers, focused on the task of cleaning up untidy legal and popular privacy discourse (DeCew 1997, 26–80). Under the banner of ideologies of privacy, private property, and private choice, they appeared to assume they could clearly de ne privacy and its value without taking sides in a political debate tied to the harms of patriarchy, racism, homophobia, and colonialism. Philosophers rarely considered or owned up to the political impetus or implications of their analyses of privacy’s meaning in their publications; they did not articulate their philosophical ideas in relation to the political dynamics of power. The battle for bodies was thus reduced to an abstract debate about conventional and ideal uses of words. In the spirit of reigning in privacy discourse, privacy prescriptivist Adam Moore has raised concerns about an “over expansive conception of privacy” in an assessment of some of my work (Moore 2013). He took exception to my characterization of professional obligations of con dentiality as unwanted “privacy” (Allen 2011). Moore thought it “vacuous” to describe laws that require professionals to keep quiet about information they learn in the context of their professional role as imposing obligations of privacy on the professional. Yet describing the obligation to keep a patient’s or client’s information con dential as coerced

privacy is a way to call attention to the political balance of power that lies underneath norms of professional silence: they give patients and clients ethical and legal power over the silenced professionals. In response to digital-era problems, new generations of experts and commentators have searched the literature for existing de nitions of privacy or tried to “de ne” privacy themselves. An assumption has been that one cannot protect consumer or citizen privacy if one can’t say (exactly) what “privacy” means. It takes work to understand how the term “privacy” is used and to defend perspectives on conventional and ideal uses. Although no one expects to reduce the concepts “liberty,” “solidarity,” “free expression,” or “justice” to a simple de nitional phrase subject to universal agreement, it seems that many people have expected something di erent of “privacy” and are frustrated by the de nitional challenges the term poses. As a result, one nds would-be de ners of privacy throwing up their hands and resorting to what I call the “intractable vagueness of privacy thesis.” This is the assertion that the term “privacy” is too vague and ambiguous for practical uses. The intractable vagueness of privacy thesis has its roots in the early philosophical accounts of privacy such as those of Judith Jarvis Thomson, Ruth Gavison, and William Parent. However neutrally intended, the thesis suits the political needs of privacy law opponents, be they traditionalists who want to weaken fundamental rights jurisprudence that might be used in support of progressive social causes, or prop. 354

technologists who nd privacy constraints a drag on lucrative innovation. Upon re ection, the thesis

is

peculiar. First of all, it opportunistically applies a standard to privacy that is not applied to other political and moral concepts. In addition, it ies in the face of the fact that “privacy” (and its foreign-language equivalents) is now used with practical success throughout the substantive law in a number of jurisdictions, especially the United States and the European Union, and in international law. Practical goals of privacy and data protection have been achieved through the utilization of privacy discourse in policy, government, “

business, and other sectors (Allen 2019a). Privacy” may be ambiguous or vague, but it is not uniquely or too much so for pursuing and achieving just causes through law or policy, both replete with words that are equally vague or ambiguous. Think of the terms “trust,” “equity,” or “reasonable.” The Supreme Court of India dealt a blow to the intractable vagueness of privacy thesis in the landmark 2017 case Puttaswamy v. Union of India, establishing a constitutional right to privacy for the world’s largest democracy. The Court had been asked to decide whether an o

cial biometric identi cation card (called the

Aadhaar card), which was required for citizens who were seeking government services and incorporated iris scans and ngerprints, violated a constitutional right to privacy (see Duraiswami 2017). Lawyers defending the Aadhaar card on behalf of the Indian government argued that there can be no constitutional right to privacy, because privacy is not expressly mentioned in the constitution and because privacy is “so amorphous as to defy description” (Allen 2019b). Although the Court extensively cited philosophical scholarship, it did not conclude that broad usage of privacy or sometimes niggling de nitional debates about the precise meaning of “privacy” supported the claim that privacy de es description. Thus the intractability thesis that privacy is singularly and fatally vague did not impede the recent establishment in India of a privacy jurisprudence based on the belief that “privacy lies across the spectrum of protected freedoms” and “re ects the basic need of every individual to live with dignity” (Puttaswamy. v. Union of India). Privacy may be vague and ambiguous, but not perniciously so. Justice can be sought and built on privacy’s foundation—1.3 billion Indians now have a right to privacy that is materially a ecting lives for the better.

5. Contested, Essentially Contested, and Political In the digital age, the eld of scholarship is, in the words of Ari Waldman, “teeming with di erent conceptions of privacy.” The multiplicity of active conceptions of privacy inside and outside academe does not warrant a conclusion that privacy, which is certainly what philosophers of language would term vague and ambiguous, is perniciously so. A recent e ort by a group of scholars to illuminate the basis of disputes about the meaning of privacy in the context of digital-age concerns, one that seeks to turn decades of p. 355

de nitional disputes and variation into a “generative” virtue, argued that privacy is

a concept of the sort

W. B. Gallie called “essentially contested” (Mulligan et al. 2016). According to Gallie, many concepts are essentially contested, meaning that their proper use inevitably involves endless disputes about their proper uses on the part of their users (Gallie 1955). Privacy indeed seems to t the bill fairly well, but by no means uniquely. A great many other concepts have been plausibly described as “essentially contested,” including the concepts of art (Gallie 1956: 97–114), freedom (Day 1986), philanthropy (Daly 2012), power (MacDonald 1976), stakeholder (Astley and Sachdeva 1984), racism (Harris 1998), rule of law (Waldron 2002), and medicine (McKnight 2003). But there is another feature of privacy discourse that bears emphasis. Not only is the concept of privacy contested, and perhaps essentially contested, but it is also emmeshed in politics. Attempting to de ne privacy more than twenty- ve years ago in the face of reductionist claims that “the more expansive conceptions of privacy are vague, ambiguous, or indeterminate,” Madison Powers pointed out that “conceptual positions can serve strategic political interests,” and that “conclusions on de nitional matters may be motivated partly by partisan positions on the underlying substantive issues, such as abortion, or by doctrinal disagreements about methods of constitutional adjudication, rather than by concerns about philosophical clarity” (Powers 1996: 371). Powers was delicate; one could say bluntly that conceptual positions prescribing proper and ideal uses of the term “privacy” serve strategic political interests and are often motivated by partisan politics. Academic contributions to the characterization of privacy discourse as hopelessly vague, ambiguous, specious, and dispensable have had political implications. They have hampered liberal, progressive, and critical agendas of racial, sexual, and reproductive equality and impaired e orts to criticize and hold accountable pro t-seeking innovators, personal-data-dependent businesses and Big Tech platforms. Stressing the politics of language and the contested nature of privacy’s meanings, a putatively politically neutral and neutral-seeming prescriptive conceptual analysis of privacy dictating ideal uses can only take us so far. What is to be commended instead is a politically self-aware methodology of critical de nitional facilitation. Rather than engage in armchair prescriptive linguistic analysis to correct privacy talk—because it is deemed confused, vague, and even misguided—philosophers should accept that the concept of privacy is and will be de nitionally contested, and own up to the political struggles embedded in privacy talk, by excavating what people are seeking through privacy discourse and doing something practical about it. For example, our work as philosophers of language could facilitate an understanding of how the recent antiracist, anti-surveillance movements use the discourse of privacy to express opposition to monitoring, surveillance, awed racial recognition technologies and inequitable uses of our data toward algorithms that make decisions based on stereotypes of criminality, irresponsibility, and threat. Taking this on is the essence of critical de nitional facilitation. Like feminist philosophy of language, as described by Andrea Nye, my approach seeks to discover something “about the speaking subject, about the societies in which speakers do and could live, and about the possibility of understanding in language” (Nye 1986: 45–51).

p. 356

6. Political Dimensions and Dynamics We use words to do something as well as to say something. Philosophers were reminded of this important truth by J. L. Austin (1962: 25). Since Austin there has been a growing appreciation of the politics of language and how, importantly, philosophy gets situated within that politics. Words and statements are often political—chosen to mobilize and immobilize, spin, shift, mystify, manipulate, and distort (Edelman 1985: 10–19). An abortion rights attorney once told me that in the 1960s, she and her colleagues speci cally chose to advocate for women’s reproductive rights using the language of privacy rather than of equality because at the time (the early 1960s), women’s privacy was felt to be a more conservative value than women’s equality, and thus less likely to fail with conservative male judges and lawmakers. The strategy worked for a while, long enough to see the decriminalization of birth control and abortion. But the right wing quickly came see that privacy discourse had become a tool of gender equality and a powerful weapon for toppling the social regime of patriarchy, moralism, and heteronormativity—so they turned against it. Philosophers claiming privacy talk to be vague, ambiguous, and confused, knowingly or unknowingly, won right-leaning political allies and advanced right-wing politics. One of the things we do with privacy discourse is to express that something deeply felt and politically urgent is at stake. Philosophers and others claiming privacy to be coherently grounded in fundamental individual constitutional, human, and moral rights advance progressive and leftist politics. Professional philosophy has been accused of ignoring the language of African Americans (McClendon 2004: 307). The role of privacy discourse in relation to African American advocacy and scholarship merits greater attention. For example, African Americans were perhaps the rst minority group on whose behalf the Supreme Court recognized that privacy rights were implicit in constitutional guarantees of free association, since privacy in group association may be necessary for groups to freely associate (NAACP v. Alabama, 357 U.S. 449 (1958)). The Court did so to help express that it was of vital, desperate importance that a civil rights organization not be required to hand over its membership list to a racist, segregationist Alabama regime dead-set on its immobilization. The lives and livelihoods of Blacks and their allies, potentially threatened by violence and loss of employment, were at stake. Introducing privacy into the First Amendment freedom of association jurisprudence on behalf of Blacks was a political strategy, and a highly successful and precedential one. It remains to be seen whether privacy discourse from the mouths and pens of African Americans themselves or their allies can counteract digital-age racism and bias as e ectively as it counteracted 1950s racial injustice and inequity on behalf of the NAACP. The political dimensions of the language of privacy in both law and philosophy merit further elaboration. There is a dynamic politics to privacy discourse in American law, suggesting that the precise de nition of privacy is far less important than the just and unjust uses to which the term and concept are variously put.

p. 357

Consider the legal uses

of “privacy” in the nineteenth century in the United States. In the rst half of that

century, the language of privacy was deployed to shield from prosecution violence against both enslaved Black people and married White women. A man who shot and wounded an enslaved Black woman in his service was held not prosecutable in North Carolina due to the “privacy” of the (rei ed) master/slave relationship (State v. Mann, 13 N.C. 263 (N.C. 1829)). Later in North Carolina, a husband who beat his wife was held not prosecutable due to the “privacy” of family and marriage (State v. Rhodes, 61 N.C. 453 (1868)). In the second half of the century, privacy was deployed in an attempt to shelter the feelings of sensitive White men and pedestalized White women from the e rontery of unwanted publicity and commercial uses of their photographs, and later to promote a utopian vision of gender equality. Disturbed by tabloid news coverage of family matters and sex scandals, Louis Brandeis and Samuel Warren argued for judicial recognition of a common law right to privacy, a right to be let alone (Warren and Brandeis 1890; see Allen and Mack 1990). A few years later, Charlotte Perkins Gilman urged the economic reorganization of homes to a ord opportunities for personal privacy to women con ned to domestic roles in their own homes (Gilman

1898/1970). A legal ideal previously employed to immunize violence against enslaved Blacks and married women was appropriated only decades later (without comment or apology for the past) to protect the feelings and sensibilities of White people assaulted by business and economic structures. In another dynamic reversal, in the second half of the twentieth century, the language of privacy was deployed by progressive reformers to empower Blacks struggling for equal rights, women seeking control over their reproductive health and LGBTQ communities ghting for sexual freedom and marriage equality (Allen 2012b). It was in the wake of progressive and liberal uses of privacy discourse that Judge Robert Bork began dismissing privacy talk as nonsense (Boyle 1991: 263). No one was disparaging privacy talk as “nonsense” when it was deployed on behalf of angry weapon-wielding white slavers and husbands. Less well appreciated, there is also a dynamic politics to privacy discourse in philosophy. The two are related, to the extent that the discipline of philosophy, which largely dwells inside the academy, is responsive to trends and potentially impacts developments outside the academy, in the policy and legal spheres. When lawyers and policy makers were not yet much discussing privacy, philosophers were not yet much discussing privacy either. Philosophic perspectives on ideal uses of the term “privacy”—whether knowingly and intentionally politically partisan or the fruits of seemingly apolitical linguistic conceptual analysis—may help persuade the lawyers, judges, and other non-philosophers who encounter such perspectives to endorse or debunk privacy-related claims and arguments. Philosophic perspectives on ideal uses of “privacy” thus have implications for the range of authoritative rhetorical options available for defending laws under which people around the globe will live. If heeded, philosophic analyses of the language of privacy prescribing or eliminating particular uses of “privacy” potentially set back the causes of real people with real concerns addressed by legislators and courts. Conversely, analyses of privacy may facilitate courts and lawmakers extending the protection of the law sought in the name of privacy. p. 358

The politics of privacy discourse in law and philosophy matter greatly in the digital era. The fate of privacy is one of the most widely voiced concerns spawned by digital life. Users of smart phones, social media, and other internet platforms continually share personal information including locations, plans, wants, needs, and opinions. Data gathering, storage, aggregation, and algorithmic analysis are taking place in the business and government sectors. Individuals are surveilled, pro led and manipulated. Highly sensitive information and tiny bits of data about quotidian behaviors are market commodities and political assets. In this context, alarms are sounded about a “weaponization” of technology (Foucek 2020), “surveillance capitalism” (Zubo

2019), “racial capitalism” (Leong 2013: 2151–2152), and the “ban-opticon” (Bigo 2008;

see Bentham 1791/1995: 10–11) that permit governments to easily control, identify, and oppress marginalized groups within a society. Privacy protection has been called for as urgent protection from fundamental assaults on freedom, dignity, and justice. New signi cance can thus be attached to the fact that the judge who created the positive legal right to privacy tort in the United States in 1905, arrestingly characterized the deprivation of privacy as a kind of slavery. In his ruling in Pavesich v. New England Life Ins. Co. (122 Ga. 190 (Ga. 1905)), Judge Andrew Jackson Cobb pointed out that the victim of privacy invasion is “no longer free, and that he is in reality a slave, without hope of freedom, held to service by a merciless master” (see  Allen 2012c). “Privacy recognizes the autonomy of the individual and the right of every person to make essential choices which a ect the course of life” (Puttaswamy v. Union of India). Theorists claim that premodern and merely modern liberal de nitions, concepts, and theories of privacy such as Dworkin’s, which seem to imply the need for choice and consent-based data policies, cannot support the needs of the emergent global digital society (Allen 2012c). Notwithstanding this, as exempli ed by the biometric-identi cation-card case in Puttaswamy and “privacy” activism in support of a data-protection statutes and agencies, Enlightenment liberal ideas are guiding the courts and lawmakers down the path to substantive digital age justice. Consistent with this, there may be respects in which certain views about the de nition of privacy stand in

the way of needed change in the digital economy. Big data and AI technology strain traditional ways of thinking about privacy rights. Those liberals and libertarians who de ne privacy as “choice” in the service of autonomy and dignity may balk at the idea that we should abandon the pervasive North American “notice and consent” model of privacy protection (Solove 2013). Those who would like the corporate sector to maintain the upper hand they have enjoyed since the rise of the internet may politically advocate for retaining choice-based de nitions of privacy precisely to allow the status quo to continue in which individuals—who do not always understand that they are sharing data, read privacy notices, or otherwise exercise their opt-out rights very frequently— are deemed to voluntarily give up privacy in exchange for goods and services. If privacy means something, such as restricted access or contextual integrity of data ows or trust, then it would be less easy to escape criticism. One way revenge pornographers, social media companies, and government benefactors alike seek to silence p. 359

complaints about the loss of data privacy and gain the

upper hand is to claim that data they have taken

control of was shared consensually by the data subject. Another way, of concern here, is to declare that nothing properly termed “privacy” is importantly at stake in what data subjects or their advocates frame as a privacy matter. For example, the controverted claim is made that, by de nition, there can be no “privacy” rights in public places to constrain police use of facial recognition technologies to identify suspects or persons of interest (Nissenbaum 1997; Rosen 2000). “Privacy in public” is supposed to be a simple oxymoron, foreclosing privacy objections to the deployment of facial recognition technology based on images collected in public places. Similarly, it is claimed that there can be, by de nition, no “privacy” rights respecting the “deidenti ed” information sought after by researchers, pharmaceutical companies, or social tracking data aggregators. De-identi cation supposedly takes the personal and the private out of the equation by de nition. In these and other ways, assertions about the meaning of privacy are designed to shut the opponents of controverted government and business practices up. As noted earlier, something analogous happened before the digital age. Privacy rights to control birth control and abortion decisions were granted to women, quite often increasing their power over their bodies and lives and in their relationships with domineering men and families. One strategy to combat the conferral of power to women was to claim that the right to birth control and abortion were not properly explicated as fundamental “privacy” matters, which placed accountability for personal medical decisions outside the scope of governmental regulation, but are instead ordinary liberties and subject to rational state regulation. Many argued that the US Supreme Court had made a mistake in creating a fundamental rightsbased jurisprudence of so-called decisional privacy; that jurisprudence, having been grounded on the conceptional mistake of de ning the absence of government interference with decisions as a matter of “privacy” rather than “liberty.” For social conservatives, whom 1960s liberals expected to be swayed by notions of privacy because of its connotations of families and property, this idea of individual “privacy” rights threatened to upset the society dominated by cisgendered White men. New “privacy” rights to minority group association, interracial marriage, birth control, abortion, and medical choice heralded sweeping progressive change. An unguarded White male undergraduate I encountered in the 1990s laid bare the gender politics at play in the legal-right-to-abortion rights debates, declaring, un-self-consciously, that legally granting women abortion choice put the decision about pregnancy “in the hands of the least quali ed person.

7. Critical Definitional Facilitation The discourses of privacy are political through and through. There can be no politically neutral de nitional analyses of “privacy” packaged for application to practical problems. The aim of philosophers should be to articulate what is at stake in privacy discourses, rst, and then to seek the de nitional clarity needed do the p. 360

work of making deep

normative values, such as freedom and anti-subordination real. In short,

philosophers have thought they should de ne the meaning of privacy and then explain its value. But actually we should be explaining the value of the things that people call privacy, and then o er analytical clari cation of the terms as needed to get practical work done. The shift from prescriptive meaning to practical signi cance is a shift from mere linguistic analysis to political analysis that includes linguistic analysis. One source of great concern is the way in which people of color in the United States are subjected to extraordinary degrees of pro ling and surveillance. African Americans believe that they lack the privacy they deserve merely because of their race, low social status, need for public services, and residence in urban communities. The use of error-prone facial recognition and the unfair algorithms are described as privacy problems. The Uighurs, a Muslim minority in China, are subjected to extreme surveillance as a mechanism of social control and exclusion. What if the philosophy of privacy began here—with the plight of Blacks and Uighurs and then moved to de nition in a service of justice? We would begin on this approach by characterizing why it matters if minorities are surveilled by the government for purposes of repression. I have argued that privacy goods are extremely important, and worthy of characterization as foundational goods of the sort one might view as the basis of fundamental rights, natural rights, and human rights (Allen 2011; 2012c: 1187). A strong way to characterize the value of many forms of privacy is to say that privacy is a tool of anti-subordination, standing in opposition to the horrors of chattel slavery. This characterization must be reconciled with the feminist critique of privacy that has emphasized that women have been harmed when privacy notions support the unfettered domination of women’s lives by husbands and fathers; and with the sense that privacy is an elitist notion, conferring privileges on those who can a ord the houses, encryption, and social standing that confer practical privacy. For privacies to be tools of anti-subordination, they must be possessed by all, including those who subsist in the lowest economic classes, castes, and minority groups. After a rich characterization of why and how the material conditions people are seeking through the discourses of privacy matter, the new approach would move on to stipulate de nitions that capture the concerns. As a group, Blacks’ privacy concerns are nuanced and subject to change, of course (Gandy 1993). The privacy Blacks are complaining about today is freedom from pervasive, selective, accusatory, anxietyproducing and chilling observation, harassment, and violence by o

cials. Theirs are classic physical and

informational privacy concerns about the panopticon society, where the “privacy” at stake is amenable to de nition as restricted access to people and information about them (Allen 2022). Also of concern to Blacks today is freedom from interference private life, data privacy on social media platforms, and privacy in healthcare and at work. Some might be tempted to argue, in the mode of Thomson or Gavison, that African Americans who think of privacy so expansively are confused about the proper de nition of privacy. But since it’s the political problem that is the practical problem, philosophers should p. 361

understand their role as helping to characterize Blacks privacy claims on their own

terms, as privacy

claims. We do this by linking what they are saying to what is otherwise said about privacy, to the language of privacy overall. When de nitional prescription is replaced by de nitional facilitation, philosophers foreground practical politics, methodologically legitimating urgent concerns about bodily integrity, autonomy and government interference that are commonly framed as privacy concerns because of their

deep relationship to personhood, personality, and the intimacies that make us more t for social roles and the responsibilities of citizenship. None of this is to say that the adoption of privacy discourse should always go unchallenged. There will be occasions when, to get the political work done, the best move would be to replace the language of privacy with one that is likely to be more persuasive. Although I disagree, some feminists have argued that it would have been better to use an equality discourse rather than a privacy discourse to obtain abortion rights in the 1970s (Allen 2005). In any case, de nitional facilitation would have us be guided in the choice of alternative discourses by the demands of the underlying values and goals as articulated by a ected groups (Browne 2015), not principally by the dictates of abstract, putatively apolitical prescriptive analysis of words.

References Allen, Anita (1988), Uneasy Access: Privacy for Women in a Free Society (Totowa, NJ: Roman and Littlefield). Google Scholar Google Preview WorldCat COPAC Allen, Anita (2005), ʻRe-writing Roe v. Wadeʼ, in Jack M. Balkin, ed., What Roe v Wade Should Have Said (New York: New York University Press), 92–108. Google Scholar Google Preview WorldCat COPAC Allen, Anita (2011), Unpopular Privacy, What Must We Hide (New York: Oxford University Press). Google Scholar Google Preview WorldCat COPAC Allen, Anita (2012a), Configuring the Networked Self: Shared Conceptions and Critiques, Concurring Opinions. https://concurringopinions.com/archives/2012/03/configuring-the-networked-self-shared-conceptions-and-critiques.html. WorldCat Allen, Anita (2012b), ʻFirst Amendment Privacy and the Battle for Progressively Liberal Social Changeʼ, University of Pennsylvania Journal of Constitutional Law, 14/4, 885–927. Google Scholar WorldCat Allen, Anita (2012c), ʻNatural Law, Slavery, and the Right to Privacy Tortʼ, Fordham Law Review, 81/3, 1187–1216. Google Scholar WorldCat Allen, Anita (2013), ʻOur Privacy Rights and Responsibilities: Replies to Criticsʼ, ʻTribute to Anita L. Allenʼ, special edition of American Philosophical Association Law and Philosophy Newsletter, 13/1, 19–27. WorldCat Allen, Anita (2019a),ʻPhilosophy of Privacy and Digital Lifeʼ, Presidential Address delivered at the 115th Annual Eastern Division Meeting of the American Philosophical Association, 93 Proceedings of the American Philosophical Association, pp. 21–38. Google Scholar WorldCat Allen, Anita (2019b), ʻSynthesis and Satisfaction: How Privacy Scholarship Mattersʼ, Theoretical Inquiries in Law, 20/1, 343–366. Google Scholar WorldCat Allen, Anita (2022), “Dismantling the Black Opticon: Privacy, Race Equity, and Online Data Protection Reform,” Yale Law Journal Forum, 131, 907. Google Scholar WorldCat Allen, Anita, and Mack, Erin (1990), ʻHow Privacy Got Its Genderʼ, Northern Illinois Law Review, 10/3, 441–478. Google Scholar WorldCat p. 362

Astley, W. G., and Sachdeva, P. (1984), ʻStructural Sources of Intraorganizational Power: A Theoretical Synthesisʼ, Academy of Management Review, 9/1, 104–113. Google Scholar WorldCat Austin, J. L. (1962), How to Do Things with Words (Oxford: Clarendon Press). Google Scholar Google Preview WorldCat COPAC Bentham, Jeremy (1791/1995), Panopticon: The Inspection House (repr. London: Verso). Google Scholar Google Preview WorldCat COPAC Bigo, D. (2008), ʻGlobalized (In)Security: The Field and the Ban-Opticonʼ, in D. Bigo and A. Tsoukala, eds., Terror, Insecurity and Liberty: Illiberal Practices of Liberal Regimes a er 9/11 (London: Routledge), 10–48, 10. Google Scholar Google Preview WorldCat COPAC

Boyle, James (1991), ʻA Process of Denial: Bork and Post-Modern Conservatismʼ, Yale Journal of Law and the Humanities, 3/2, 263. Google Scholar WorldCat Browne, Simone (2015), Dark Matters: On Surveillance of Blackness (Durham: Duke University Press). Google Scholar Google Preview WorldCat COPAC Carter, R., and Marshall, T. (1955), ʻThe Meaning and Significance of the Supreme Court Decreeʼ, Journal of Negro Education, 24/3, 397–404. Google Scholar WorldCat Daly, S. (2012), ʻPhilanthropy as an Essentially Contested Conceptʼ, Voluntas: International Journal of Voluntary and Nonprofit Organizations, 23/3, 535–557. Google Scholar WorldCat Day, Patrick (1986), ʻIs the Concept of Freedom Essentially Contestable?ʼ, Philosophy, 61/235, 116–123. Google Scholar WorldCat DeCew, Judith Wagner (1997), In Pursuit of Privacy, Law, Ethics and the Rise of Technology (Ithaca, NY: Cornell University Press). Google Scholar Google Preview WorldCat COPAC Duraiswami, D. (2017), ʻPrivacy and Data Protection in Indiaʼ, Journal of Law & Cyber Warfare, 6/1, 166–186. Google Scholar WorldCat Edelman, Murray (1985), ʻPolitical Language and Political Realityʼ, in PS: Political Science and Politics, 18/1, 10–19. Google Scholar WorldCat Epstein, Dimitry, Roth, Merrill C., and Baumer, Eric P. S. (2014), ʻItʼs the Definition, Stupid! Framing of Online Privacy in the Internet Governance Forum Debatesʼ, Journal of Information Policy, 4, 144–172. Google Scholar WorldCat Foucek, Chynna (2020), ʻIs Data Privacy a Privilege? The Racial Implications of Technology-Based Tools in Governmentʼ, Techdirt. https://www.techdirt.com/articles/20200622/07255444756/is-data-privacy-privilege-racial-implications-technology-basedtools-government.shtml. Gallie, W. B. (1955), ʻEssentially Contested Conceptsʼ, Proceedings of the Aristotelian Society, 56, 167–198. Google Scholar WorldCat Gallie, W. B. (1956), ʻArt as an Essentially Contested Conceptʼ, Philosophical Quarterly (1950-), 6/23, 97–114. Google Scholar WorldCat Gandy, O. H. 1993, ʻAfrican Americans and Privacy: Understanding the Black Perspective in the Emerging Policy Debateʼ, Journal of Black Studies, 24/2, 178–195. Google Scholar WorldCat Gavison, Ruth (1980), ʻPrivacy and the Limits of Lawʼ, Yale Law Journal, 89/3, 421. Google Scholar WorldCat Gilman, Charlotte Perkins (1898/1970), Women and Economics (repr. New York: Source Book Press). Google Scholar Google Preview WorldCat COPAC Gupta, Anil (2021), ʻDefinitionsʼ, in The Stanford Encyclopedia of Philosophy (Summer 2021 Edition), Edward N. Zalta (ed.), https://plato.stanford.edu/archives/sum2021/entries/definitions/. Google Scholar Google Preview WorldCat COPAC Harris, Leonard (1998), ʻThe Concept of Racism: An Essentially Contested Concept?ʼ Centennial Review, 42/2, 217–231.

Google Scholar

WorldCat

Hirsh Eric Donald (1984), ʻMeaning and Significance Reinterpretedʼ, Critical Inquiry, 11/2, 202–225. Google Scholar WorldCat Leong, Nancy (2013), “Racial Capitalism”, Harvard Law Review, 126/8, 2151–2226. Google Scholar WorldCat p. 363

MacDonald, K. I. (1976), “Is ʻPowerʼ Essentially Contested?” British Journal of Political Science 6/3, 380–382. Google Scholar WorldCat McClendon, John, H. (2004), ʻPhilosophy of Language and the African American Experience: Are There Metaphilosophical Implications?ʼ, Journal of Speculative Philosophy, 18/4, 305–310, 307. Google Scholar WorldCat McKnight, C. James (2003), ʻMedicine as an Essentially Contested Conceptʼ, Journal of Medical Ethics, 29/4, 261–262. Google Scholar WorldCat Miller, Arthur R. (1971), The Assault on Privacy: Computers, Data Banks and Dossiers (Ann Arbor: University of Michigan Press). Google Scholar Google Preview WorldCat COPAC Moore, Adam (2013), ʻCoercing Privacy and Moderate Paternalism: Allen on Unpopular Privacyʼ, American Philosophical Association Law and Philosophy Newsletter: Edition in Tribute to Anita L. Allen, 1–27, 9–12. Mulligan, Deirdre K., Koopman, Colin, and Doty, Nick (2016), ʻPrivacy Is an Essentially Contested Concept: A Multi-dimensional Analytic for Mapping Privacyʼ, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 374/2083, 1–17. Google Scholar WorldCat Nissenbaum, Helen, 1997. ʻToward an Approach to Privacy in Public: Challenges of Information Technology,ʼ Ethics & Behavior, 7/3, 207–219. Google Scholar WorldCat Nye, Andrea. (1986), ʻFrench Feminism and Philosophy of Languageʼ, Noûs, 20/1, 45–51. Google Scholar WorldCat Parent, William A. (1983), ʻA New Definition of Privacy for the Lawʼ, Law and Philosophy, 2/3, 305–338. Google Scholar WorldCat Powers, Madison (1996), ʻA Cognitive Access Definition of Privacyʼ, Law and Philosophy, 15/4, 369–386, 371. Google Scholar WorldCat Puttaswamy, K. S. (Justice Ret.) & Anr. v. Union of India, (2017), 10 SCC 1, AIR 2017 SC 4161, 26. ALE at 244. Google Scholar WorldCat Rosen, Je rey (2000), ʻPrivacy in Public Placesʼ, Cardozo Studies in Law and Literature, 12/1, 167–191. Google Scholar WorldCat Schoeman, Ferdinand D. (1984), Philosophical Dimensions of Privacy: An Anthology (Cambridge, UK: Cambridge University Press). Google Scholar Google Preview WorldCat COPAC Shils, Edward (1966), ʻPrivacy: Its Constitution and Vicissitudes,ʼ Law and Contemporary Problems, 31/2, 281–306, 281. Google Scholar WorldCat

Solove, Daniel (2006), ʻA Taxonomy of Privacyʼ, University of Pennsylvania Law Review, 154/3, 477. Google Scholar WorldCat Solove, Daniel. (2013), ʻIntroduction: Privacy Self-Management and the Consent Dilemmaʼ, Harvard Law Review, 126/7, 1880– 1903. Google Scholar WorldCat Sorensen, Roy (2018), ʻVaguenessʼ, The Stanford Encyclopedia of Philosophy (Summer 2018 Edition), Edward N. Zalta (ed.) (Stanford University: The Metaphysics Research Lab). Google Scholar Google Preview WorldCat COPAC Thomson, Judith Jarmon (1975), ʻThe Right to Privacyʼ, Philosophy and Public A airs, 4/4, 295–314. Google Scholar WorldCat Waldron, Jeremy (2002), ʻIs the Rule of Law an Essentially Contested Concept (In Florida)?ʼ, Law and Philosophy, 21/2, 137–164. Google Scholar WorldCat Warren, Samuel D., and Brandeis, Louis D. (1890), ʻThe Right to Privacyʼ, Harvard Law Review, 4/5, 193–230. Google Scholar WorldCat Zubo , Shoshana (2019), The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power (New York: Public A airs). Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

CHAPTER

p. 364

16 Gender-Neutrality and Family Leave Policies  Jules Holroyd, Matthew J. Cull https://doi.org/10.1093/oxfordhb/9780192844118.013.39 Published: 22 May 2024

Pages 364–387

Abstract Robin Dembro

and Daniel Wodak (2018, 2021) argue that we have a duty to use gender-neutral

pronouns, but they do not extend this argument to all other aspects of our language. This chapter evaluates the extent to which gender-neutral language is desirable in the context of parental leave schemes, taking as a case study the parental leave schemes found at a higher education institution in the United Kingdom. It argues that the considerations Dembro

and Wodak take as speaking against

gender-speci c pronouns and other gender-speci c aspects of language also strongly speak against gender-speci c language in the context of parental leave policies. As a project in non-ideal theory, the chapter argues against the framing of existing policies which refer to ‘maternity’ and ‘paternity’ leave and for moving to the language of ‘parental’ or ‘family’ leave. The fact that the majority of those giving birth are women does not provide decisive reasons for framing policies in gender-speci c terms. Moreover, given the welcome move to facilitate shared parental leave, any concerns for gender equity, and equity for carers, are better served by requesting demographic information for monitoring purposes rather than by policies that refer to gender-speci c parenting roles.

Keywords: parental leave, maternity leave, pregnancy, philosophy of language, gender-neutral language, institutional reform, misgendering, trans-inclusion, family, higher education Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction Robin Dembro

and Daniel Wodak (2018) provide a compelling argument for the claim that we have a duty

to use gender-neutral pronouns (they, them) rather than gender-speci c pronouns (she, her, he, him), and that we should also reject certain other aspects of gendered language (su Dembro

xes, generics, and honori cs;

and Wodak 2021). Using gender-speci c pronouns, they argue, violates privacy and supports

essentialist assumptions. It perpetrates the harms of exclusion and misgendering. Using gender-neutral pronouns, they argue, avoids these issues and is a better way of proceeding than other options we have, such as using ‘they’ alongside gender-speci c pronouns as a third ‘catch all’ gender term, or proliferating many di erent gender pronouns. (More on each of these considerations to follow.) They are clear that their argument targets only gender-speci c pronouns (Dembro gendered aspects of language (pronouns, su

and Wodak 2018) and other quite speci c

xes, generics, and honori cs). Whether other gender-speci c

aspects of language should be eliminated or retained depends on substantive argument as to whether those aspects of language are useful in articulating and resisting gender oppression or, rather, contribute to it (Dembro

and Wodak 2021).

In this chapter, we consider the extent to which these considerations, and others, apply to gender-speci c language as used in the context of family leave policies, taking as our starting point—and as an instructive 1

case study—family leave policies found in a United Kingdom higher education institution. In the United Kingdom, family leave policies are underpinned by government funded schemes to enable parents to take time o

work, with some nancial support, to care for their infants. Central government frames these

leaves in gender-speci c terms: maternity and paternity leave are available. This gender-speci c framing thus informs how institutional policies are presented, as we will detail here. This provides a particularly p. 365

challenging test case for Dembro

and Wodak’s arguments, as, in addition to the gender-speci c

language embedded in central government policy, the processes of childbirth and infant childcare are socially strongly gendered, and the majority of those birthing babies are women. Likewise, in the United Kingdom in 2023, the majority of those with primary responsibility for the feeding and care of young infants are women, and both biological pressures (related to breastfeeding) and social pressures (related to social conceptions of motherhood) shape these patterns. Moreover, giving birth to and caring for children—and taking leave to do so—is correlated with various patterns of inequity, so that it is important to track the gendered dimensions of family leave policies. However, we will argue that notwithstanding these considerations—and in some instances because of them —there are strong reasons in favour of institutions adopting gender-neutral language in the context of family leave policies. We set out these considerations, and outline a working model for how such policies could be articulated. In section 2, we outline the essential features of the institutional case study with which we are working: family leave policies at a UK institution of higher education. In section 3, we take up the reasons for moving to gender-neutral language presented by Dembro

and Wodak, and consider the extent

to which they apply in the context of family leave policies. We argue that the considerations apply, and that this supports a rst proposal for gender-neutral institutional policies. In section 4, we consider objections to moving to gender-neutral language related to the need to identify gender discrimination and patterns of gender inequity related to family leave. This enables us to nesse our proposal. In section 5, we defend this proposal against further objections. Our aim is limited in scope: we are not proposing revisions that would address all inequities to do with family leave. Our focus on language reform leaves out a number of other issues, including the amount of leave that is taken, how it is distributed, the rate at which the leave is paid, how many people to whom it is available, as well as broader issues around workplace culture that support or isolate people taking leave. Substantive changes are needed to address issues in these related areas, but they 2

will not be the focus of this paper.

2. Context We here present the leave entitlements in the United Kingdom, framing them initially in gender-speci c terms to re ect the policies as they are currently stated. 3

As of September 2023, people who give birth and meet certain eligibility conditions are entitled to fty-two weeks maternity leave, a signi cant portion of which is nancially supported by statutory maternity pay (up 4

to thirty-nine weeks). Two weeks must be taken after the birth of the child, and the mother can then decide how much of the remaining fty weeks to take. Paternity or partner leave, on the other hand, is considerably more limited—two weeks of nancially supported leave can be taken within eight weeks of the birth of the p. 366

child. Since 2015, parents have also had the option of taking

Shared Parental Leave, whereby the fty

weeks to which the mother is entitled (after the initial two weeks of maternity leave) can be split between the parents, who can take leave simultaneously or consecutively, with a considerable degree of exibility. Thirty-seven of these weeks are nancially supported. These existing provisions, then, are strongly unequal: the person who gives birth has a far greater entitlement to leave, and even with the option of 5

sharing parental leave, it is ‘theirs’ to share. In contrast, the father or partner is entitled to very little leave. Employers are constrained by the entitlement framework we have just outlined, but they can supplement

the nancial support attached to this leave if they choose. Many higher education institutions in the United 6

Kingdom o er enhanced leave packages. For example, providing that certain service conditions are met,

the institution in our case study o ers its employees who take maternity leave eighteen weeks of full pay. Two weeks of paternity or partner leave are also paid at full pay. The Shared Parental Leave policies are also reasonably generous. Provided certain eligibility criteria are met, and that the mother is entitled to statutory maternity pay from the government, a partner who is employed by the institution may be o ered up to eighteen weeks of leave at full pay. Analogous leave packages are also available to adoptive parents, and those who become parents through surrogacy arrangements. Note, then, that the leave packages are framed in gender-speci c terms at two stages: rst, in central government policy, which sets out entitlements to maternity, paternity or partner leave and pay structures, as well as Shared Parental Leave (couched in terms of the mother and her partner). At the second stage, gender-speci c language is used in institutional-level policy, again, framed in terms of maternity and paternity leave, and Shared Parental Leave split between the mother and her partner. The gendered nature of these family leave policies have the following implications: rst, only the person who gives birth can take ‘maternity leave’, so called. Second, all people who give birth must take ‘maternity leave’, so called. Parents who identify with the role ‘mother’ but did not give birth (e.g. partners in lesbian relationships) can only take ‘paternity/partner leave’. People who give birth but do not identify with the role ‘mother’ are nonetheless institutionally required to identify themselves as the mother taking ‘maternity leave’. The existing framework, then—both in central government policy, and in its institutional manifestation—uses language that is strongly cisnormative and heteronormative. Various legal frameworks governing trans parenting in the United Kingdom have already been identi ed as problematically exclusionary. A recent review (White 2018) of these legal frameworks focuses on the Gender Recognition Act (2004), the Human Fertilisation and Embryology Act (2008), and the Births and Deaths Registration Act 1953. The review argues that in the existing legislation ‘there is an absence of consideration for trans people as parents in any capacity’ (White 2018: 4). It notes that trans men or non-binary people who give birth are consistently referred to as the ‘mother’ of the child, with no legal options for alternative 7

self-de nition (3–6). Similar problems, we argue, beset the frameworks for family leave schemes, aimed p. 367

at supporting new parents. Thus, at present, some parents’ access to this support is participation in schemes which misgender them, as we argue in section 3.

conditional on their

The arguments we present apply to both stages: they provide reasons both for central government to change their policies to gender-neutral framings, and for institutions that provide leave packages to do so, as well. Since central government policy currently constrains institutional policy, the ideal would be for it to change; but we note that this is not necessary for institutions to frame their own policies in gender-neutral terms, and we will make some proposals for how they can do so.

3. The Case against Gender-Specific Language in the Context of Family Leave We do not wish to deny that gender-speci c terms capture the statistical norm. However, we want to evaluate the desirability of doing so. What is problematic about gender-speci c language? In this section we bring to bear the arguments that Dembro

and Wodak (2018, 2021) advance, and show how they can be 8

applied, with some revision, to gender-speci c family leave schemes.

3.1. Entrenchment of Binary Gender and Exclusion of Non-Binary Identities First, Dembro

and Wodak (2021) raise the concern that use of gender-speci c terms assumes two discrete

and exhaustive genders, and excludes non-binary people. By making it the only option that people designate themselves ‘man’ or ‘woman’, this assumes that these are the only two gender categories; nonbinary individuals, such as people who consider themselves agender, bigender, or a di erent gender yet, for example, are obscured by using only these two gender-speci c categories. The same applies in the context of family leave: the use of just two gender-speci c terms (‘maternity’ and ‘paternity’) assumes that these are the only two parent roles, and that these two roles correspond to two discrete gender roles (man and woman). Indeed, gender-speci c content is built into the terms ‘maternity’ and ‘paternity’: part of the semantic content of ‘paternity leave’ is ‘leave for fathers’, and the semantic content of ‘father’ takes us directly to the gender-speci c man in relation to his child. Note here, however, that we are not taking issue with the gender-speci c language of ‘man’ and ‘woman’—as we note below, there are some important reasons for retaining this gendered aspect of language. Our focus is on the corollary gendered parenting roles ‘mother’, ‘father’—and in particular, the institutionally enshrined leave schemes associated with them: ‘maternity’ and ‘paternity’ leave. With respect to these leave schemes, non-binary individuals are obscured, insofar as they do not identify p. 368

with either of the two gender-speci c parenting roles. In the context

of institutional leave schemes,

those who do not identify with either of those binary parenting roles have no adequate parenting role options available to them. Bennett and Fu (2015) describe how Cara Jeiven, who identi es as genderqueer, ‘felt pressured to solidify her gender identity. “I knew I didn’t feel like a ‘mother’ and didn’t necessarily want to be a ‘dad,’ ” she says. “I suddenly felt like I had to choose.” ’ Thus Jeiven’s identity is obscured, both socially and institutionally. In forcing people to ‘pick’ a side of the binary, as Jeiven felt pressured to do, these leave schemes misgender 9

people. This is clear in the case of many non-binary individuals, who are forced to take on the label of ‘mother’ or ‘father’, with their attendant gendered connotations, despite not identifying as either a man or a woman. Similarly, such policies require that pregnant trans men label themselves as on maternity leave. And trans women may be forced to label themselves as on paternity leave, once again misgendering. An additional concern arises in the context of gender-speci c family leave terms: there may be individuals who do not consider themselves non-binary (and are thus willing to identify as man or woman) but who do not embrace gendered roles in their daily lives, and are therefore resistant to the terms mother or father,

maternity or paternity. These individuals’ gender identities are not obscured, but their preferred parenting roles are, if they are required instead to designate themselves as either mother or father. Shared Parental Leave fares little better. Since the leave is framed in terms of the mother splitting her leave entitlement with her partner, at least one side of the gender binary persists here. Moreover, since the other partner remains entitled to the two weeks paternity leave, alongside the allocation of Shared Parental Leave agreed with the mother, this model hardly transcends the gender binary. It assumes that one parent is the 10

mother, and the other, who is entitled to paternity leave, might seek to share some of her maternity leave.

The consequences of these exclusions are to entrench a system of binary gender and to commit the harms of misgendering. Where non-binary gender identities are obscured, the marginalization and oppression of non-binary gender identities are entrenched. Where preferred parenting roles that do not conform to binary gender-speci c parenting roles are obscured, this entrenches the pervasion of the binary gender system across domains of life. In contrast, a policy that frames leave in gender-neutral terms—parental or family leave (of maximum of 11

fty-two weeks or a maximum of two weeks )—would avoid these exclusions, and avoid entrenching binary gender.

3.2. Privacy A second set of concerns about the use of gender-speci c terms has to do with privacy. Dembro

and Wodak

(2018, 2021) argue that the use of gender-speci c pronouns—rather than gender-neutral ones—can often put individuals in a position of revealing facts about themselves and their lives that they may prefer to keep p. 369

private. For example,

privacy about one’s personal relationships may be threatened. If assumptions are

made about the gender of an individual’s partner (‘will she be joining you on holiday?’), then the interlocutor has to correct them (‘yes, he will’), and thereby reveal facts about their sexuality that they might prefer to keep private, especially in a context of homophobia. Alternatively, they might let the assumption stand, and risk being seen as deceptive should the correct information later come to light. This places individuals in a ‘disclose or deceive’ dilemma (Dembro

and Wodak 2018: 393).

Privacy with respect to one’s gender identity is similarly threatened by the use of gender-speci c pronouns. If assumptions are made about one’s gender identity (‘Sam said she’d present the paper today; go ahead and start, Sam’), then the interlocutor has to correct them (‘I use he/him pronouns’), revealing facts about their gender identity they might prefer to keep private, especially against a background of transphobia. If they choose to let the assumption stand, they risk being seen as deceptive should the correct information later come to light (a particularly dangerous risk for transgender people, to whom the transphobic trope of ‘deceiver’ is often applied (see Bettcher 2007)). Analogous privacy concerns arise in the context of gender-speci c terminology in family leave policies. Privacy about one’s personal relationships may be threatened due to the assumptions made in having the standard leave packages framed as ‘maternity’ and ‘paternity’. For example, the heteronormative framing puts individuals in lesbian relationships either in a position of having to deceive—go along with the assumption that their partners are fathers, hence men—or disclose that their partner, the person who is taking the nominal ‘paternity’ leave, is not a father and not a man; a potentially perilous disclosure in a homophobic work environment. Note that as long as ‘paternity’ comes with gendered assumptions—that the person taking paternity leave is a father, and hence a man—this dilemma will not be avoided by attempts to use gender-neutral language alongside the gender-speci c parenting roles (‘will your partner be taking paternity leave?’). Such utterances still import the assumption that the partner is a father and a man.

Privacy about one’s gender identity may also be threatened by policies that frame the leave that individuals take as ‘maternity’ or ‘paternity’ leave. For example, a pregnant person who is transgender faces assumptions about their gender identity when taking up maternity leave—that they are a mother, and a 12

woman.

These assumptions could either be corrected, disclosing the person’s status as (for example) a

trans man. Or the pregnant person could go along with these assumptions, at the risk of being accused of deception should it later be disclosed that they are not a woman. As noted, this latter option also involves forced misgendering and attendant harms. Does Shared Parental Leave avoid any of these privacy concerns? Recall that in actuality, the framing of Shared Parental Leave hardly transcends the gender binary, relying on gender-speci c parenting roles that have as their semantic content speci c gender identities (mother/woman, father/man). The same privacy concerns will arise, then. However, it should be acknowledged that, in the case of the mother’s partner’s leave, the problems are parasitic on the attachment of paternity leave to Shared Parental Leave, and could be p. 370

avoided if paternity leave were not framed in gender-speci c terms.

This side of Shared Parental Leave,

then, provides a glimpse of how gender-neutral language in parental leave policies might work: framed in gender-neutral terms (as ‘parental’ or ‘family’ leave) with no gendered assumptions about the role or about the gender of the person taking up the role. A gender-neutral leave scheme, which refers to those taking the leave simply as ‘parents’ or even ‘family’ would not force people into ‘disclose or deceive’ dilemmas, and so would not pose a threat to privacy the way gender-speci c leave schemes do.

3.3. Anti-essentialism Dembro

and Wodak advance the concern that gender-speci c language contributes to gender

essentialism. They point to evidence that linguistic markers of gender (gender-speci c pronouns and other gendered components of language, such as su

xes like the ‘-ess’ in actress, waitress, etc.) entrench beliefs

about the relevance of gender. Such gendered language communicates that gender is signi cant, or has some important explanatory role. And whilst gender is sometimes an important explanatory feature, it is not always so in the way that is implied by gender-speci c language. In the context of family leave, the worry here would be that the gendered labelling of parenting roles (‘mother’ and ‘father’, and ‘maternity’ and ‘paternity’) communicates that one’s gender has important implications for how one performs that parenting role. In particular, it contributes to the message that the roles are di erent. Moreover, against a backdrop of such starkly di erential leave entitlements, it risks 13

transmitting the message that mothers should be doing more parenting.

Thus, in addition to contributing

to essentialist beliefs about the relevance of gender to one’s role in parenting, it also contributes to harmful stereotypes about men and women in relation to parenting: that women are the natural nurturers, that men’s role is to merely ‘help’ occasionally (cf. Dembro

and Wodak [2021] on the claim that gendered

language can contribute to stereotypes and stigma). One might object here that gender is relevant to one’s parenting role, either because women give birth or because women take maternity leave, which is more generous than men’s. So if the language used re ects that and transmits that information, it is not a problem. Each claim warrants a response. First, we do well to remember that whilst many people who give birth are women, not all are—trans men, agender, non-binary people also give birth. The use of gender-speci c language obscures this fact (as we have argued). Second, the fact that women take maternity (rather than parental) leave and that it is a considerably longer leave entitlement than is available to fathers is indeed relevant to the social role of parenting that many people take up; but it is precisely these institutional features (the entitlements to each parent, and the genderspeci c language in which the policies are framed) whose relevance one can question, asking whether

gender should be signi cant to one’s parenting role in the way that current societal arrangements have it. p. 371

We argue that it should not be. Finally, the essentialist concern, moreover, is that the implication is not only that gender is explanatorily relevant, but

that it is so because of some essential property of women. This

problematic belief may be transmitted alongside other beliefs about the current, contingent social relevance 14

of gendered parental roles.

A gender-neutral framing of leave schemes as ‘parental’ or ‘family leave’ does not transmit any beliefs about di erential parenting roles, or about who—which people of which, if any, gender—should take up any particular role. As such, gender-neutral terminology would avoid this essentialist worry, along with any attendant stereotypes about parenting roles.

3.4. State (or Institution) Gender Marking Elsewhere, Dembro

(2020) has raised a concern about state gender-marking, and suggested that the onus

should be on those who endorse this to explain why it is so important. One concern in this domain is related to privacy, but, in particular, the potential harms that can come from the state, or other institutions that have considerable power over us, having information of a personal nature about us. A similar concern may arise here: that there are risks in one’s employer having certain kinds of personal information. To the extent that this is information we think it is important to have control over when and to whom we disclose, the issues arising here are those addressed in the section on privacy, above. Gender-speci c family leave schemes may force people to disclose information about themselves and their relationships they would otherwise prefer to keep private. Another concern in this region is to do with the extent to which institutional or state use of categories entrenches them as part of our social world. This worry has a

nities with the worry about essentialism mentioned above. We want to tease out distinct

and speci c problems with institutions—such as employers and central government—deploying genderspeci c language in the context of family leave schemes. First, doing so serves to reify the parenting roles— ‘mother’ and ‘father’. As social constructs, this institutional rei cation serves to entrench and perpetuate such roles. Binary parenting roles are normalized and given institutional backing, whilst alternative parenting roles gain no traction. Second, and relatedly, institutional deployment of gender-speci c familial roles also tacitly signals support for those binary parenting roles. This may have a psychological impact on those who are excluded—they are not recognized, and only gain access to institutional support on the condition of framing their parenting roles in terms that do not ‘ t’. Moreover, for the employees of that institution, these binary roles are inescapable—access to the support to which they are entitled is 15

conditioned on subscribing to these (perhaps ill- tting) roles.

But there is another side to this concern, which focuses on the extent to which it might be important for institutions to have certain information about their employees. For example, it might be important for an institution to have information about who takes family leave, to examine the relationship between those who take these leaves and other important equity outcomes (pay, professional advancement, and so on). In p. 372

16

particular, given the relationship between parental leave and gender pay gaps,

one might

think it

particularly important that institutions are able to scrutinize the relationship between the leave that women and gender non-conforming people take and these other equity-related outcomes. This is an important concern. For it to have bite, though, we would need to see that the existing schemes (framed in genderspeci c terms) do in fact do a good job of enabling us to track the relationship between women taking leave and other gender equity related outcomes; at least as good as the alternative proposal for gender-neutral leave schemes. We note this concern, and return to it shortly, after having presented our proposal. There, we elaborate on the concern, and explain how our proposal is better placed to meet this concern than existing gender-speci c leave policies.

3.5. Gender Neutrality as the Best Option So far, we have advanced considerations against framing family leave in gender-speci c terms (‘maternity’ and ‘paternity’) and suggested that gender-neutral terms (‘parental’ or, better, ‘family leave’ simpliciter) would avoid these problems (though we are yet to present a fuller proposal on this; see section 4). Why insist on a gender-neutral framing instead of rather than alongside gender-speci c parental leave schemes? Indeed, one might think that Shared Parental Leave accomplishes something like this by giving people the option to take parental leave, rather than leave with a gender-speci c parenting label (as we noted earlier, the current framing of Shared Parental Leave is unsatisfactory, since it is framed in terms of the mother and her partner and insofar as it is attached to the two weeks of maternity or paternity leave). One might suggest that, if it were made genuinely gender neutral, (shared) parental leave could be o ered alongside maternity or paternity leave. This would be the analogue of an option that Dembro

and Wodak

(2018) consider for pronouns, whereby ‘they’ is available as a third option alongside the gender-speci c pronouns ‘he’ and ‘she’. In the context of pronouns, they worry that this strategy is problematic, in that this third category subsumes everyone who does not identify as ‘she’ or ‘he’. It is therefore inegalitarian—this strategy a ords men and women speci c pronouns, but all other genders are given the ‘catch-all’ pronoun ‘they’. This also reinforces the idea that there is something natural or inevitable about binary gender categories, whilst everything falling outside these two categories is treated as ‘other.’ Do similar worries apply in the context of using a gender-neutral ‘parental’ or ‘family’ leave scheme for all those who do not want to take up gender-speci c parenting roles? As we have argued, gender-speci c schemes do tend to convey the idea that there is something natural about mothering (and fathering), and the risk is that a third ‘parental’ scheme to catch ‘other’ parents would entrench this. Moreover, this strategy also catches all parenting roles other than mother/father in this third ‘other’ category, and so risks similar inegalitarian implications. Moreover, insofar as ‘mother’ remains a role that is only available to the person who gave birth, partners who do not give birth but who nonetheless identify with this role are also p. 373

marginalized—whether they are a lesbian or

bisexual woman in a relationship with another woman who

is giving birth, a woman in a relationship with a trans man who is giving birth, or otherwise. A scheme that is, by default, gender neutral for all does not privilege heterosexual relationships or a ord the preferred gender-speci c parenting role only to those who meet a heteronormative ideal. Another alternative that Dembro

and Wodak consider is whether to proliferate pronouns, using as many as

needed to capture multiple gender identities. We might consider similarly proliferating the family leave terms, so they re ect multiple familial roles corresponding to multiple gender roles. Dembro

and Wodak

argue that proliferating pronouns is infeasible. The same is true, we think, of proliferating roles in the context of family leave, but for di erent reasons. First, the parenting role may not track gender or gender identity (there need not be any one-to-one correspondence between parenting role and gender). Second, with pronouns, there are so many options, Dembro

and Wodak argue, that we might easily get it wrong;

our ability to learn new pronouns is limited (given facts about our cognition and how this aspect of language works), so this option raises the real risk of misgendering people. In the context of parenting roles, the infeasibility of proliferating labels for parental roles is somewhat di erent. There does not seem to be the extensive linguistic repertoire for naming alternative parenting roles. As Andrea Bennett writes, ‘I’ve remained label-less … there’s nothing equivalent to “dad” for me … non-binary folks have adopted pronouns, like “they” and “ze,” to carve out space for ourselves in language. Parenting labels could use a 17

similar revision.’

In the absence of such labels, however, there is no proliferation of labels for policies to

18

adopt.

Given this, the best option seems to be to frame leave policies in gender-neutral terms, simply as family leave. In the next section, we substantiate this claim by outlining a proposal for gender-neutral leave policies and then, in section 5, defending it.

4. A Proposal for Gender-Neutral Leave Schemes We have suggested that there are compelling considerations against gender-speci c leave schemes, and that these considerations provide some support for a gender-neutral leave scheme. A gender-neutral scheme would avoid cisnormative and heteronormative assumptions about parenting. But what would such a scheme look like, and how would it be implemented? Recall that, for now, we are working within the constraints of the UK parental leave entitlements, currently framed as fty-two-weeks entitlement to the mother, and two weeks to the father, with the possibility that the mother can share some of the fty weeks 19

(post two-week maternity leave) with her partner (Shared Parental Leave).

We propose that the various leave schemes are instead framed, in institutional policy, in gender-neutral terms. This means framing the leaves as parental or family leaves for both parents (the parent who gives p. 374

birth, who is entitled to two-plus- fty weeks, and the

parent entitled to two weeks, with possibility of

shared weeks). Since the leave schemes are presently asymmetrical, there would be little chance of confusion—those taking leave would simply need to ll out the form relevant to them (family leave with 2 + 50 entitlement, or family leave with 2 + entitlement). Moreover, since there are other ways of referring to people than ‘mother’ or ‘father’, e.g. by their proper names, there is little chance that the scheme would be 20

beset by confusion about who is taking what portion of leave.

A family leave scheme framed in gender-neutral terms and focusing on the present entitlement would avoid the concerns that attach to gender-speci c leave schemes—exclusion, violation of privacy, essentialism, and the infeasibility of alternatives. Nor do the leave schemes themselves engage in institutional gender marking, and so the attendant worries about privacy and essentialism here are also avoided. Dembro

and Wodak introduce the following norm in relation to gender-speci c pronouns:

Denial: We have a duty to not deny others’ gender identities by using third person pronouns that misrepresent the referent’s gender identity (2018, 383). There is a corollary norm for other gendered aspects of language in relation to parental leave: Denial*: We have a duty not to deny others’ gender identities by using gender speci c language to 21

describe roles that misrepresent the referent’s gender identity.

Gender-speci c leave schemes that insist on framing leave-takers as ‘mothers’ or ‘fathers’ violate Denial*. As we have already noted, gendered parenting roles have the e ect of forcing people who may not identify with a binary gender as having to pick an ill- tting parenting role which is in con ict with their gender identity. Gender-neutral family leave schemes, such as those we have proposed, do not. One might object that by framing leave schemes in gender-neutral terms, we are violating denial by failing to embrace the gender-speci c roles that some parents do keenly embrace. There are two responses to this, one semantic, the other practical. First, as a matter of semantics, ‘family’ or ‘parent’ doesn’t deny the possibility of gender-speci c instantiations of this role in the way that gender-speci c language does with respect to gender-neutral roles. This is evidenced by the use of ‘family’ and ‘parent’ in many contexts in which the gender of the person ful lling the role is not relevant or known—such as, for example, in the locution ‘parents’ evening’ at schools, and ‘family fun run’ at local parks. People who identify as ‘mothers’ or ‘fathers’ take themselves to be subsumed under this ‘parent’ locution, and do not nd themselves misgendered by doing so. Second, as a practical matter, there are ways of implementing gender-neutral family leave schemes that in fact meet a stronger norm than Denial* and show that an on-paper gender-neutral policy is capable in

p. 375

practice of embracing a range of gender-speci c much. Consider also the norm of a

roles—but only when the people taking leave choose as

rmation introduced by Dembro

A rmation: We have a duty to a

and Wodak (2018):

rm others’ gender identities by using third person pronouns that

represent the referent’s gender identity (382). Dembro

and Wodak deny that we have a duty to a

rm others’ gender identity by using the correct

pronouns, since we have various ways of referring to people (by their names, titles, and so on) that do not 22

misgender them.

The analogue to this for our concern with family leave schemes would be:

A rmation*: We have a duty to a

rm others’ gender identities by using gender-speci c language

to describe roles that adequately represent the referent’s gender identity. If we had such a duty, then gender-neutral family leave policies would violate it by failing to a

rm the

gender-speci c parenting roles that those who identify with the roles mother and father take on (it would also fail to represent other non-binary gender-speci c roles). It seems clear that we don’t have a duty to a

rm people’s parenting roles: nothing is going wrong when schools advertise ‘parents evenings’ rather

than ‘mothers and fathers evenings’, for example, or when local parks advertise ‘family fun runs’ rather than ‘mothers and fathers fun runs’. However, there may be something problematic with insisting on using the gender-neutral ‘family leave’ or the term ‘parent’ to describe someone who strongly identi es as a mother (or father) and asks to be referred to as such. But note that the proposed gender-neutral policies need not do this. Whilst a gender-neutral on paper policy is the most inclusive option (for reasons outlined in section 3), implementing the policy—which requires downloading the relevant leave form and providing the required personal information (name, expected week of birth or adoption, etc.)—could very easily include a box for people to write in the parenting label they choose to adopt, which could then be used in all communications. Thus, a partner in a lesbian relationship who is entitled to two weeks of leave and identi ed as a mother could—instead of being misgendered as a father taking paternity leave—indicate that they wish all correspondence to identify them as the mother of the child. Or someone who is entitled to twoplus- fty weeks of leave and identi es as a parent, could indicate that they wish all correspondence to identify them as the parent of the child. Or a non-binary person who is entitled to two-plus- fty weeks of leave and indicates a preferred parenting role of ‘baba’ could indicate that they wish all correspondence to 23

identify them as the baba of the child. could very easily ensure that a

Thus, in practice, implementing a gender-neutral leave scheme

rmation is also met—which is a nice, respectful, thing to do, even if there is

no duty to do so. Indeed, this practice actively welcomes the expression of people’s chosen identities, and thus amply satis es A rmation*. So the institutional use of alternative parental roles comes with certain p. 376

good upshots, as we alluded to in section 3.4 above (see note 15): in the process of social construction of non-binary parenting roles, and in the positive psychological impact on parents who may have their preferred parenting role institutionally a

rmed. Whether this way of implementing the policy is taken up

will ultimately depend on a weighing of these gains with any costs—in particular, the bureaucratic costs associated with storing and circulating this information as appropriate. Note that this option also comes with additional risk of misgendering in a harmful way: it is one thing to have gender-neutral parental leave policies; it is another thing to ask people to specify what language they would like to have used and to ensure adherence to this. To ask for preferences and then ignore them (because the information fails to be appropriately ltered from Human Resources departments to line managers) would be a particularly problematic failure. If no indication is given (or if an institution decides not to adopt a practice of giving people the option of specifying their preferred parenting role), the gender-neutral term ‘parent’ could be used as a default— which would not misgender, since it is compatible with non-neutral instantiations, as we argued above— but this may not be frequently needed, since as noted, there are multiple ways of referring to someone, for

example by using their proper name. Since what is chie y relevant, from the institutional point of view, is the amount of leave taken, it is straightforward to communicate about this in terms which do not make 24

reference to the speci c family role at all.

Consider:

Dear Jules, we have received noti cation of your intention to take up your entitlement of 2 + 50 weeks of leave following the birth of your child. Speci cally, we understand that you intend to take the 2 weeks that are mandatory following birth, and then 25 of the following weeks. Your partner will then take the remaining 25 weeks of leave, according to the Shared Parental Leave (SPL) scheme. Please ensure that they also communicate with their employer about the 2-week entitlement, as well as any other enhanced entitlements they may have in relation to the 25 weeks of SPL. When considering both the policy and the possibilities for implementation in practice, then, we see that the duty not to misgender is met and that there are in fact ways of a

rming the chosen familial roles that

individuals choose to take on.

5. Defending the Proposed Policy The proposal we have made does not face the concerns that we set out for gender-speci c family leave policies. It does not exclude, obscure, or misgender non-binary people and trans men. It respects privacy by not requiring of employees that they provide information about the personal relationships or gender identity. It further avoids entrenching essentialist assumptions about women’s and men’s roles in parenting. Finally, it is more feasible than alternatives which might append a gender-neutral option onto the existing gender-speci c structures, or which might proliferate parenting roles. Our arguments have p. 377

been enormously aided by drawing on Dembro

and Wodak’s

work on gender-neutral pronouns (2018)

and other aspects of language (2021). We have suggested that such a family leave scheme would avoid denying anyone’s gender identity, and that there are ways of implementing the policy that would enable the institution to actively a

rm it, by o ering the option of providing details of preferred parenting role labels.

There are other things speaking in favour of our proposal also: in virtue of enabling people to enter their preferred parenting role labels, it can acknowledge family structures other than heterosexual nuclear families (though to the extent that leaves are restricted to two parents, there is a limit to the ability to do this). Before our case for gender-neutral family leave schemes is complete, though, we should also consider the potential costs of advancing a gender-neutral scheme of family leave. In particular, one might argue that there is important work done by having the vocabulary of ‘maternity’ and ‘paternity’ available, and that a complete evaluation of the merits of gender-neutral leave must take this into account. Accordingly, before dispensing with gender-speci c language in this domain, we must ‘look to the ways that gender-speci c terms in particular socially function, and decide from there which terms are important for anti-sexist work’ (Dembro

and Wodak, 2021, 371). What might be the gains of having the language of ‘maternity’ and

‘paternity’ available, and in particular, entrenched in institutional policy?

5.1. Tracking Gender and Other Equity-Related Outcomes One important task to which such language might be put, it could be argued, would be to track the extent to which those who take maternity leave su er adverse outcomes in terms of other equity-related outcomes— gender pay gaps, and so on. We have already mentioned that this might stand as a reason in favour of institutional gender marking, whereby one’s employer keeps track of maternity leaves and monitors the relationship between these leaves and other outcomes. This could be important data to have in the context of anti-sexist work, which enables us to detect inequities and formulate policies to address them. However, we note that existing practices are not well positioned to gather such data or evaluate its relationship to equity-related outcomes. First, insofar as current policies exclude non-binary people, it isn’t possible to gain any understanding of the relationship between leave taken by non-binary people and other equity-related outcomes. This problem attaches both to the family leave part of the data and to the data gathering about relevant equity-related outcomes. For example, data about gender pay gaps should attend to whether non-binary people experience pay gaps, also. Second, feminists might be particularly concerned to identify the relationship between the parental leaves that women take, and the relationship to other equity-related outcomes, such as gender pay gaps and professional advancement. Indeed, this is precisely what studies such as the one by Costa Dias et al. (2018) aim to scrutinize. However, it is worth noting that, at least as current data gathering stands at higher education institutions such as our own respective p. 378

institutions, it is very di

cult to tease out the relationship between the leave

that women take and these

other equity-related outcomes. This is because the current framing of leave policies fails to track all, and only, the women who take family leaves of any kind. First, it fails to track all the women who take family leave. There will be some women who take family leave as either ‘paternity/partner leave’ or ‘Shared Parental Leave’ qua the ‘mother’s partner’, who may not be accurately identi ed as women who take leave. Second, it will not track only women who take leave if the relationship studied is between those who take maternity leave and other equity-related outcomes. There will be some non-binary people and some men who give birth, yet are designated as taking ‘maternity leave’. In fact, these non-binary people and men may be particularly vulnerable to discrimination, both qua person who gave birth and qua non-binary persons or trans men. This will not be tracked in a system that misgenders non-binary people and trans men as ‘mothers’. Such a policy both fails to be trans inclusive and is unhelpful for understanding how trans people and women are disadvantaged by family leaves. So, whilst gender-speci c language may be important for gathering relevant data about those who take parenting leave and other equity-related outcomes, it should be clear that the existing framing in terms of ‘maternity’ and ‘paternity’ does not serve this goal. How might this purpose better be served, then? We propose that institutions seeking to gather and scrutinize data of this sort for the sake of equity monitoring should instead commit to gathering data alongside information about leave. This is a simple matter of including certain questions about gender on the forms people submit to notify their institutions of their intention to take family leave. For example, information could be asked about gender/gender identity, as well as sexuality, age, disability status, and race, since the intersectional nature of oppression means that it is important to trace how the intersection of these vectors of oppression may all be implicated alongside the impact of taking family leave. In addition to enabling institutions to better monitor the impact of leaves of gender-related outcomes, this would enable institutions to analyze carefully the contribution of such time out of the workplace on, for example, race pay gaps, the career advancement of disabled scholars, and so on. Of course, all this requires a commitment, at institutional level, to appropriately analyzing and scrutinizing this sort of data, and taking 25

steps to change policy or its implementation where inequities are visible.

Crucially, it must be clear that providing this personal information, alongside information about leave, is optional. This is important to ensure that no individual is in a position of being forced to reveal this personal

information, and privacy is threatened. Moreover, the collection of this data for equality monitoring is completely compatible with the information not being available to managers in a way that is connected to any individual. Contemporary best practices for hiring collect demographic data, but hide it from hiring panels so that no demographic pro le can be connected to any given individual candidate. We suggest that 26

institutions adopt a similar approach when collecting demographic data about family leave.

p. 379

5.2. Further Objections to Gender-Neutral Leave Policies We next defend and re ne our proposal in light of ve further objections, concerning the relationship of our proposal to wider feminist aims.

5.2.1. Anti-Feminist? One might suggest that it is anti-feminist to remove talk of ‘women’ and ‘maternity’ leave from institutional policies around pregnancy. But as we have argued, these proposals actually serve feminist goals well. First, they enable better tracking of the relationship between women who take family leaves and other equity-related outcomes, by accurately tracking the genders of those who take leave, making the women who take leave when their partners give birth more visible. Second, our proposal avoids entrenching essentialist assumptions about women’s parenting roles. Third, our proposal protects women from having to disclose personal information about their gender or relationships if they choose not to. Fourth, our proposed policies are trans inclusive and so serve the goals of all women, not just cisgender women. Moreover, our proposals do not prohibit individuals and institutions from talking about women, motherhood, and maternity leave, since we propose that individuals should have the option of identifying ‘mother’ as their preferred parental role in lling out the noti cation of leave form. The crucial point is that people get to decide what parenting role label to take up, rather than having one of the two binary roles imposed upon them.

5.2.2. Devaluing Motherhood? Another claim that our proposal might be anti-feminist in is based on the suggestion that it devalues motherhood. Feminists have called attention to the way that childcare, and particularly women’s role in childcare, is invisibilized, unrecognized, and unrewarded (see Ferguson 2020). In recommending a move towards gender-neutral language, aren’t we contributing to obscuring women’s important work in childcare? This proposal does not devalue motherhood, but it does decentre it. This is because (a) not all women embrace the role of ‘mother’, and (b) not all parents embrace either role—‘mother’ or ‘father’. Moreover, whilst we acknowledge the importance of recognizing the labour that predominantly women do in parenting, it is feasible to expect that our proposals, by focusing on the labour of carers and pregnant people rather than their genders, actually render the work done by women—and all parents who are primary carers —more visible. Unpacking the nature, causes, and implications of the gendered division of labour and devaluation of women’s work is beyond the scope of this chapter; however, few would deny that part of the problem with women’s unpaid and unrecognized childcare labour is precisely the fact that this work is 27

feminized—it is seen as women’s work.

A scheme that sees the parenting of infants in gender-neutral

terms could perhaps counter this tendency. Another very important step in challenging this gendered aspect of childcare would be the taking up of parental leave by men, which would both expose them to the p. 380

experiences of the full-time primary parenting of an

infant and eventually block the discriminatory

tendency to assume that women (but not men) will take leave from the workplace after the birth of a child. Framing family leave in gender-neutral terms, rather than as a scheme which takes as its starting point mothers’ greater entitlement to leave, may facilitate this step. As such, our proposal does not devalue

motherhood, but rather works to undermine the feminization of parenting work and renders this important work more visible and appropriately valued.

5.2.3. Pregnant People? We have not focused much on a very important embodied aspect of parenting—namely that, in many cases, it follows pregnancy and the birth of a child. Does our proposal obscure the fact that parenting follows pregnancy and thereby do a disservice to women who experience pregnancy and childbirth? Firstly, we note, once again, that not only women experience pregnancy; trans men and non-binary people also do. Their participation in pregnancy is rendered more visible by using language that does not exclude them. Secondly, we note, again, that some people who identify as women prefer not to use the terminology of ‘mother’ and ‘maternity’ to describe their role in gestating, birthing, and raising infants. So there is no necessary connection between moving to gender-neutral language and obscuring pregnancy. And at least some people’s pregnancies are rendered more visible by removing the gendered focus on discourse about pregnancy and parenting (see Surkan 2015; Summer 2014). Gender inclusive language around pregnancy, alongside the use of second personal (‘you’) or gender-neutral pronouns ‘they/them’ to refer to the person 28

giving birth are therefore preferable in institutional information.

Accordingly, our proposal does not 29

render pregnancy and birth, or women, invisible. It just acknowledges that not only women give birth.

It

also acknowledges that women are involved in parenting in other ways (as partners of the birthing parent). So it does better justice to the full range of women’s experiences and birthing experiences. Moreover, recall once again that we are not ‘banning’ from institutional discourse the language of ‘mothers’ or ‘maternity’. A person who feels that, throughout pregnancy and parenting she is a mother could indicate as much on her leave noti cation form, and systems should be instantiated to respect this (letters about the leave and discussions with line managers can be guided by these expressed preferences). So, there is no reason to suppose that our proposals ignore or obscure the embodied aspects of pregnancy that often precede family leave.

5.2.4. Care Networks Existing gender-speci c leave policies, at both the UK state level and in higher education institutions such as our own frame the leave that may be taken based on the assumption that only one or two people will take that leave. That is, models for the care of young children assume either a traditional nuclear family, featuring a mother and father, or a single parent. We suggest that this assumption is (a) unwarranted, and (b) unwelcome. As Jessica Clarke notes, writing about the US context, but nonetheless equally applicable to the United Kingdom, the assumption that childcare is undertaken solely by one or two parental gures is unwarranted: p. 381

Many pregnant people rely on networks of extended family and friends for support, rather than a parent partner. This may be particularly true for families of color. Professor Melissa Murray explains that “[w]ithin the African-American community, for example, parents frequently share caregiving responsibilities and material resources with community members in an arrangement known colloquially as ‘other-mothering.’ In Latino communities, compadres—literally ‘coparents’—play a central role in the child’s spiritual upbringing and often are expected to share the parents’ caregiving responsibilities. Yet workplace accommodation laws, like the FMLA [Family 30

and Medical Leave Act], are “unrealistically focused on parenthood as the locus of caregiving.”

Given this actual variation in families beyond one- and two-parent models, we suggest that the genderspeci c formulation of leave packages in the United Kingdom are additionally unwarranted in assuming that there are only one or two parents to be considered when developing family leave policies.

Moreover, we suggest that the assumption that there will be only one or two parents is undesirable, for two reasons. First, existing policies erase the labour of co-parents and wider networks of care pointed to by Clarke. We ought to be able to implement policies that recognise this labour. Second, existing leave policies are not neutral between conceptions of the good, but promote a particular gendered form of the nuclear family. Instead, we ought to leave space for social forms that reform or outright reject the family. Here we are thinking of families formed by those who are polyamorous, or political radicals who wish to reject the 31

family form entirely, raising children communally.

We do not want to suggest that forms of social

organization beyond the one- or two-parent family unit are superior here; nonetheless, we think that concerns of equality should lead us to leave space for such other social forms in policy. These varied actual and potential care networks are better accommodated by policies framed in genderneutral language. After all, if a polyamorous commune of ve pregnant women wants to jointly raise their children, it seems absurd and oppressive to insist that one of the ve be nominated as the ‘father’ of any one child. There is, however, a further worry lying in the background here: supposing that the adage is correct and that ‘it takes a village to raise a child’, as demonstrated by the large care networks pointed to by Clarke, what justi cation do we have for keeping language that references parents? That is, if the care work undertaken in birthing and raising a child is sometimes performed by the child’s parents, grand-parents, siblings, aunts and uncles, family friends, and a wider social circle, we might worry that a policy that only allows parents to take the leave is unjusti ed. Instead, we might think that whoever performs the labour of birthing or raising the child ought to be entitled to leave. To be maximally inclusive, the framing of ‘family leave’ would avoid assumptions about those taking the leave being the biological or social parents of the child. That framing opens up the possibility of moves away from nuclear familial structures, which existing 32

policies assume and perpetuate.

If that’s so, then perhaps the language of ‘parent’ should be excluded from our ‘parental’ leave policies, p. 382

leaving us with a set of care leave policies that should be neutral

with respect to the gender of the carer

and particular relation of the carer to the child. Our ideal model of ‘parental’ leave, then, is one in which the various sorts of caregiving networks receive ample state and institutional support, with ample resources to allow people to gestate and care for children in the ways that best suit them. That said, this ideal is not one that lends itself well to implementation by individual institutions in the legal framework of the United Kingdom. Indeed, it seems hard to see how one would even begin to develop policies that would support such a model of care at the institutional level without signi cant reform at the legislative level. Our recommendation in section 4 did not fully incorporate this concern. We take ourselves to have been doing non-ideal political philosophy of language: we recognize that the ideal is some way o , and we hope that the implementation of our proposed recommendations can at least serve as a stepping stone towards this more just way of organizing care labour. Gender-neutral language, of course, is not the end of reproductive justice. Rather, it is one part of a larger program to build a more just society for those seeking to give birth to and raise children.

5.2.5. Feasibility Is what we are proposing really feasible? Does introducing gender-neutral language introduce infeasible complexity or confusion into the process of notifying one’s institution of family leave? We have argued that 33

it does not. The policies and their implementation can be relatively straightforward.

Moreover, the

existing arrangements are highly infeasible for those who are erased by them. And they re ect and entrench undesirable normative assumptions about gendered parenting roles, heteronormative relationships, and nuclear families. Our proposal does better at avoiding these.

One issue that might raise concerns over feasibility is that, in the absence of a legal framework that supports these changes, institutions may face issues in implementing a more just family leave system. After all, an institution that exists within a state that legislates that only cisgender women may take certain kinds of leave will face signi cant obstacles to adopting our recommendations. Whilst it is clear that some legal frameworks will make these changes more di

cult to achieve than others, as we have demonstrated with

the example of a higher educational institution in the United Kingdom, institutions have the ability to develop improved policies even in the face of somewhat restrictive legal frameworks. Ultimately, institutions should navigate the legal frameworks they nd themselves in so as to implement the best practices feasible.

6. Conclusion Dembro

and Wodak’s arguments lay a promising foundation for examining the ways in which gender-

neutral language may be desirable. Consideration of the extent to which gender-neutral language is desirable in family leave policies is particularly challenging. We have imported important considerations, p. 383

and raised new ones, for considering incorporated in institutions’ o

whether gender-speci c language for parenting roles ought to be

cial family leave policies, and have argued that it should not be. Gender-

neutral policies, accompanied by practice that facilitates the expression of preferred parenting or familial roles, better serve the need to monitor gender discrimination, and avoid entrenching gender oppression and language that imports cis- and hetero-normative assumptions.

Acknowledgements Thanks to the Feminism Reading Group at the University of She

eld, and to the participants in the Words

Workshop run out of the University of Pittsburgh, for great discussion and feedback on early drafts of this paper.

Notes 1.

These are fairly indicative of parental leave schemes in the UK, though the financial support o ered is better than some leave packages in the private sector.

2.

There is a considerable history of feminist argument for language change, much of it has focused on false gender neutrality and arguing for the need, in some contexts, for gender specificity. See Mercier (1995) and Vetterling-Braggin (1981). We take this project—arguing for gender neutrality in a specific domain—to be perfectly compatible with other contexts in which gendered language might be important or even required as a corrective to false gender neutrality. Indeed, Mercierʼs (1995: 252–253) paper advances support for gender-neutral pronouns.

3.

These conditions are having worked a requisite number of hours in the previous year and made the relevant tax contributions. This makes the support inaccessible to women who are not able to work for various reasons, e.g. uncertain citizenship status. Mothers who donʼt meet the eligibility criteria because they havenʼt made the relevant tax contributions may be entitled to a lesser amount of maternity allowance. Some women, however, are entitled to very little support: asylum seekers, for example, can only receive a one-time payment of £300 to cover the costs associated with the birth of their child (essential items such as nappies, bottles, clothes, and so on). We think that this obvious injustice must be urgently rectified, and that asylum seekers should be given support that enables them to flourish.

4.

Currently in the United Kingdom (as of September 2023), this amounts to 90% of oneʼs average weekly earnings per week for six weeks and, therea er, £172.48 per week or 90% of oneʼs average weekly earnings for the remaining 33 weeks, whichever is lower.

5.

Under conditions of gender oppression, and given the various physical burdens associated with birth, it is not obvious that this entitlement for the person who gives birth is wrong. However, the lack of adequate leave for the partner of the person giving birth is deeply problematic.

6.

These conditions are in fact deeply problematic—in particular, the requirement for a year of continuous service before enhanced leave can be received. In other research, Holroyd and Clark have found that this service condition is particularly problematic for those with fixed-term contracts and has sometimes led to considerable stress and uncertainty. We have recommended, in line with best practice in the sector, that this service condition be scrapped.

7.

See McConnell and YY v Registrar General (2020) EWCA Civ 559, for a prominent recent case upholding this restriction on UK birth certificates. A number of other EU countries have similar rules, Sweden being a notable exception (van den Brink and Dunne 2018: 61). The question of anti-discrimination protections for pregnant trans people across the EU is vexed, and may depend on which version of the so-called pregnancy Directive (92/85/EEC) is employed. As van den Brink and Dunne point out, the English and Finnish versions of 92/85/EEC use neutral language (though, relevantly, the Finnish version does not use neutral terms for maternity leave), whilst the Dutch, French, and German versions refer to female workers (van den Brink and Dunne 2018: 100).

8.

Dembro (2020) also considers gender specific voting lists in the US. Some considerations are raised there that do not appear in the papers, and which are highly pertinent to some of our concerns below—we note that we are drawing on Dembro ʼs ideas at that point also.

9.

On the nature and harms of misgendering, see Kapusta (2016) and Cull (2020).

10.

Not only is the woman assumed to be the primary caregiver, but the woman is also framed as in charge of the organization of care, and hence the domestic sphere. This assumption thereby reinforces a particular traditional sexist stereotype about womenʼs role in society.

11.

Of course, family leave that a orded a more equitable split between the two parents would be more desirable still, but we set aside those pressing concerns here, since our focus is on the linguistic dimensions of leave policies rather than the entitlements themselves.

12.

See Surkan (2015) for an articulation of the complex norms navigated (and violated) during pregnancy as a trans man.

13.

For an articulation and critical evaluation of such gender norms, see Williams (2001).

14.

Note that to genuinely undermine such gendered assumptions, more thoroughgoing changes are needed to address unequal entitlements. This does not undermine the reasons for moving away from gender-specific language but provides reasons to do much more.

15.

Conversely, institutional recognition and deployment of non-binary parenting roles would have constructive power with respect to alternative roles, and stand to confer psychological benefits upon those whose alternative parenting roles are given tacit institutional support. We return to this point in section 4.

16.

See e.g. Costa Dias et al. (2018, 24).

17.

Bennet (2019); see also Bennett and Fu (2015).

18.

Were there, however, one might wonder about the original infeasibility concern arising—that the risks of misgendering someone by ascribing the wrong parental role would loom large.

19.

As noted, we of course also support more equitable and generous parental leave for both parents, and have argued for this elsewhere. However, our focus here is on the linguistic framing of existing schemes. Institutions can implement this step now, and doing so is an important inclusive step—even if there are many other steps to take. We stress that the adoption of a more inclusive linguistic framework should not be used as a cover for institutions to fail to take these further steps (see Ahmed 2012).

20.

And were the leave schemes more uniform—each parent being entitled to the same amount of leave, say—there would still be easy ways of avoiding confusion about whose leave is at issue—the person who gave birth, or the person who did not, in the case of parenting following gestation. In the case of adoption or surrogacy, there would be other mechanisms

p. 384

for distinguishing whose leave is at issue, such as proper names or parenting see details of our proposal implementation in section 4.

p. 385

roles that are chosen and implemented—

21.

Thereʼs another issue here, which concerns using gendered roles that the person doesnʼt embrace whether or not they misrepresent the referentʼs gender identity, e.g. calling cisgender women mothers when they prefer to be called parents. Is there a duty to avoid this? This is obviously not as problematic as the harms of misgendering, so we donʼt formulate the duty to accommodate this. But certainly, if there are ways to avoid using misnomers for parentsʼ roles, it is a good thing to do. We argue below it can easily be achieved.

22.

For a fuller discussion, and endorsement, of this norm, see Hernandez (2021).

23.

See Bennett (2019).

24.

The best outcome would be for central government to o icially adopt gender-neutral terms in framing its policies. But in the absence of that, institutions have the power to shield their employees from the harmful e ects of gender-specific language by using gender-neutral terms in framing their own policies. Human resources can be the interface between the gender-neutral policies and any gender-specific information that needs feeding back to central governmental institutions that currently use gender-specific language, such as His Majestyʼs Revenue and Customs (HMRC) in the UK.

25.

We do not want to overstate the extent to which these changes might alter the trends that are visible on the basis of existing data. The point is, rather, about the integrity of the data—whether it captures all those it purports to—and whether the categories involved are deployed in ways which misgender or respect gender identity (and parenting role).

26.

The one exception to this is the preferred parenting label, which would be used in communications with the employee taking family leave.

27.

See Fontana and Schoenbaum (2019) on various problematic aspects of this, including the exclusionary impact on trans and gender non-conforming people.

28.

Exactly how to make language around pregnancy and perinatal care gender inclusive, however, is very complicated. For discussion, and for the argument that pluralistic linguistic strategies are needed to meet the moral and communicative goals of these contexts, see Cull, Holroyd and Woollard (n.d.).

29.

For excellent resources related to trans pregnancy, see the Trans Pregnancy Project: https://web.archive.org/web/20230406031543; https://transpregnancy.leeds.ac.uk/; https://ruthpearce.net/tag/transpregnancy/

30.

Clarke (2019, 196).

31.

For recent work on family abolition and radical visions of care, see OʼBrien (2020); (2019); Lewis (2019); and Gri iths and Gleeson (2015).

32.

Note that this issue is not straightforwardly solved by emphasizing parent as a social role rather than a biological relation. A er all, the thought here is not that the language ʻparentʼ is troubled by people other than the biological progenitors of a child taking on the parental role. Rather, the thought is that a variety of relationships are possible that provide labour in raising a child (sometimes in combination with a central parental figure or figures). Now, it might be possible to do some conceptual engineering here to broaden the meaning of ʻparentʼ so that it encompasses a wide variety of relationships, but as things stand, ʻparentʼ has a narrower definition, and it is unclear whether such an engineering project is desirable.

33.

We have already noted that feasibility will be a consideration for institutions to confront in determining how to gather and deploy information about parentsʼ parenting language preferences.

p. 386

References Ahmed, S. (2012), On Being Included: Racism and Diversity in Institutional Life (Durham, NC: Duke University Press). Google Scholar Google Preview WorldCat COPAC Bennett, Andrea (2019, May 10), ʻIʼm a Non-binary Parent. There Still Isnʼt Space for Meʼ, Xtra Magazine, https://www.dailyxtra.com/im-a-non-binary-parent-there-still-isnt-space-for-me-154990, accessed 9 Jan. 2021. WorldCat Bennett, Andrea, and Fu, Kim (2015), ʻBeyond Mom ʻnʼ Popʼ, Hazlitt Magazine, https://hazlitt.net/feature/beyond-mom-n-pop, accessed 9 Jan. 2021. WorldCat Bettcher, Talia Mae (2007), ʻEvil Deceivers and Make-Believers: On Transphobic Violence and the Politics of Illusionʼ, Hypatia, 22/3, 43–65, http://www.jstor.org/stable/4640081. Google Scholar WorldCat Clark, J., and Holroyd, J. (n.d.) ʻParental leave in UoS: Policy and Practice, Equity and Inclusionʼ (Manuscript. Available on request). Google Scholar Google Preview WorldCat COPAC Clarke, Jessica (2019), ʻPregnant People?ʼ Columbia Law Review Forum 119/173, 173–199. Google Scholar WorldCat Costa Dias, Monica, Joyce, Robert, and Parodi, Francesca (2018), ʻThe Gender Pay Gap in the UK: Children and Experience in Workʼ, IFS Working Papers, No. W18/02, Institute for Fiscal Studies (IFS), London. https://ifs.org.uk/sites/default/files/output_url_files/MCD_RJ_FP_GenderPayGap.pdf. Google Scholar WorldCat Cull, M. (2020), ʻEngineering Genders: Pluralism, Trans Identities, and Feminist Philosophyʼ, PhD thesis, University of She ield. Google Scholar Google Preview WorldCat COPAC Cull, M., Holroyd, J., and Woollard, F. (n.d.) Caring for Everyone: Inclusive Language in Perinatal Care (Manuscript (unpublished). Available on request). Google Scholar Google Preview WorldCat COPAC Dembro , Robin (2020) ʻCategorising Candidatesʼ slide presentation, provided in personal correspondence. Google Scholar Google Preview WorldCat COPAC Dembro , Robin, and Wodak, Daniel (2018), ʻHe/She/They/Zeʼ, Ergo 5/14, 371–406 https://doi.org/10.3998/ergo.12405314.0005.014. Google Scholar WorldCat Dembro , Robin, and Wodak, Daniel (2021), ʻHow Much Gender is Too Much Gender?ʼ In Justin Khoo and Rachel Sterken, eds., Routledge Handbook of Social and Political Philosophy of Language (London: Routledge), 362–376. Google Scholar Google Preview WorldCat COPAC Ferguson, Susan (2020), Women and Work: Feminism, Labour, and Social Reproduction (London: Pluto Press). Google Scholar Google Preview WorldCat COPAC Fontana, David, and Schoenbaum, Naomi (2019), ʻUnsexing Pregnancyʼ, Columbia Law Review, 119 (2), Available at https://columbialawreview.org/content/unsexing-pregnancy/. Google Scholar WorldCat

Gri iths, K. D., and Gleeson, J. J. (2015), ʻKinderkommunismus: A Feminist Analysis of the 21st-Century Family and a Communist Proposal for Its Abolitionʼ, Ritual. 1.0, Available at https://isr.press/Gri iths_Gleeson_Kinderkommunismus/index.html. Google Scholar WorldCat Hernandez, E. M. (2021). ʻGender-A irmation and Loving Attention.ʼ Hypatia, 36(4), 619–635. Google Scholar WorldCat Kapusta, S. (2016), ʻMisgendering and Its Moral Contestabilityʼ, Hypatia 31/3, 502–519. Google Scholar WorldCat Lewis, Sophie (2019), Full Surrogacy Now: Feminism against the Family (London: Verso). Google Scholar Google Preview WorldCat COPAC Mercier, A. (1995), ʻA Perverse Case of the Contingent a Priori: On the Logic of Emasculating Language (a Reply to Dawkins and Dummett), Philosophical Topics, 23/2, 221–259. Google Scholar WorldCat OʼBrien, M. (2020), ʻTo Abolish the Familyʼ, Endnotes 5, 361–417. Google Scholar WorldCat OʼBrien M. (2019), ʻCommunizing Careʼ, Pinko 1, https://pinko.online/pinko-1/communizing-care. Google Scholar WorldCat Summers, A. K. (2014), Pregnant Butch: Nine Long Months Spent in Drag (Berkeley CA: So Skull Press). Google Scholar Google Preview WorldCat COPAC p. 387

Surkan, K. (2015), ʻThat Fat Man Is Giving Birth: Gender Identity, Reproduction and the Pregnant Bodyʼ. In N. Burton, ed., Natal Signs: Cultural Representations of Pregnancy, Birth and Parenting (Bradford, CAN: Demeter Press), 58–72, doi:10.2307/j.ctt1rrd8tc.7. Google Scholar Google Preview WorldCat COPAC van den Brink, Marjolein, and Dunne, Peter (2018), Trans and Intersex Equality Rights in Europe—a Comparative Analysis. Report prepared for the European Commission, Directorate General for Justice and Consumers, Luxembourg: Publications O ice of the European Union. Vetterling-Braggin, Mary (ed.) (1981). Sexist Language: A Modern Philosophical Analysis (New Jersey: Littlefield, Adams). COPAC Google Scholar Google Preview WorldCat White, F. R. (2018), Trans Pregnancy: An International Exploration of Transmasculine Practices of Reproduction Law and Policy Review United Kingdom, https://web.archive.org/web/20221009154427/https://transpregnancy.leeds.ac.uk/wpcontent/uploads/sites/70/2018/05/Trans-Pregnancy-policy-review-UK.pdf, accessed 19 Jan. 2021. WorldCat Williams, Joan (2001), Unbending Gender: Why Family and Work Conflict and What to Do about It (New York: Oxford University Press). Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

17 In Praise of Spoken Philosophy  Esa Saarinen https://doi.org/10.1093/oxfordhb/9780192844118.013.40 Published: 22 May 2024

Pages 388–412

Abstract This chapter locates the growing lack of relevance for philosophy as an academic discipline and as a contributing force to society at large, as linked to the dismissed status the profession’s dismissal of spoken philosophy and orality as a form of expression. Taking inspiration from Socrates, from Plato’s reservations about written philosophy, and from Merleau-Ponty’s idea of “the total function of the philosopher” and paying close attention to the signi cance of the often-disregarded questions of style in the conduct of philosophy, the chapter lays out a case for new type of university philosophy and a philosophy at large that aids people in the conduct of their lives. From the point of view of the applied philosophy of language and metaphilosophy defended here, this new approach places a high priority on addressing people as people, in their everyday naivety, contradictions, “animal spirits,” tendencies toward non-openness, as well as their hopes and strivings for something “transcendent” The chapter argues that philosophy, and the use of language in conducting philosophy, should meet humanity head-on as it seeks to renew itself. It explores and argues for philosophy as an enterprise that wants to be relevant to the lives of people as they are actually lived, as opposed to o ering mere abstractions. The result is a call for the renewal of spoken philosophy, orality, and special kind of life-philosophical lecturing as a contributing force in the actual lives of people.

Keywords: spoken philosophy, orality, Socrates, Merleau-Ponty, life philosophical lecturing Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction Was it an accident that Socrates only operated orally?

Did the rst philosopher engage in spoken philosophy as a secondary choice and for incidental reasons, feeling, say, inferior as a writer? In this chapter I explore the possibility that one reason for the growing perceived lack of relevance of philosophy as an academic discipline and as a contributing force to society at large, is linked to the dismissed status of the profession’s dismissal of spoken philosophy.

2. The Total Function of the Philosopher Maurice Merleau-Ponty, one of the celebrated greats of the twentieth-century philosophy, opens his inaugural lecture as Chair of Philosophy at Collège de France by noting that the institution he has the honor of entering “has been charged with the duty, not of giving its hearers already-acquired truths, but the idea of free investigation”. That is a credo many philosophers would likely endorse. What about the implementation? Merleau-Ponty is harsh: What he [the modern philosopher] says enters rst of all into an academic world where the choices of life are deadened and the occasions for thought are cut o . p. 389

He o ers these words as a broad overall assessment, with the utmost seriousness. He nails the case with two striking further statements: The philosophy placed in books has ceased to challenge men (34). and The modern philosopher is frequently a functionary, always a writer (33). Since he said these words, in 1952, things have hardly improved for professional philosophy. Is the decline real? One option is to take the view that philosophy is not about “the choices of life” and “occasions for thought” that “challenge men” but is primarily a scholarly enterprise. It serves civilization and the rational culture as one expert eld among others, and as such legitimately proceeds by the criteria the university institution has developed to fruition as the kind of dominantly article-based written philosophy we enjoy today. If that makes us philosophers “functionaries”, so be it. Without o ering scholarly review on the various forms in which such a metaphilosophy of philosophical functionarism is endorsed, Merleau-Ponty makes clear he is talking about a practice of philosophy he strongly rejects. Here Merleau-Ponty highlights something he calls “the total function of the philosopher” as part of his “praise of philosophy”, turning to a philosopher whose name is familiar to everybody: even the philosophical writers whom we read and who we are have never ceased to recognize as their patron a man who never wrote, whole never taught, at least in any o

cial chair, who talked

with anyone he met on the street. … We must remember Socrates. What are the chief features of philosophy thus conceived? Crafting the answer from what Merleau-Ponty is saying and what is generally accepted knowledge of Socrates, the following description seems appropriate: When operating from the total function of the philosopher, the philosopher challenges people and aims at an in uence on their choices in life.

In this chapter, I shall be using this understanding of a possible metaphilosophy, some challenges and opportunities which I seek to investigate. The focus is on the conduct of us philosophers, as opposed to our scholarly debates, say, on Socrates or Merleau-Ponty scholarship, theory of virtues, or the various schools of meta-philosophy. Does the practice of a Socratian philosopher call for applied philosophy of language that challenges some of the foundational beliefs our profession holds dear?

p. 390

3. In Praise of Philosophy If you want to challenge people in their way of life, you need to reach them. As an empirical fact, it can be noted that article-based written philosophy that addresses the highly specialized expert cultural audience of academic philosophy has become the dominant mode of expression in academic philosophy in our times. Since the days of Merleau-Ponty’s inauguration lecture, the role of the journal format has gained ground from books, making even the general reading public more distant. The article format makes academic philosophy close to completely inaccessible to most people. What was the reason for Merleau-Ponty’s scruples regarding written philosophy? Certainly not writing per se. Himself a master writer, Merleau-Ponty turns to Henri Bergson as someone to point the way. The predecessor of Merleau-Ponty who received the Nobel Prize in literature in 1927 “in recognition of his rich and vitalizing ideas and the brilliant skills with which they have been presented”, certainly did reach beyond his fellow philosophy professionals. In front of the Provost and other dignitaries of Collège de France the new philosophy professor Merleau-Ponty does not hesitate to talk “In Praise of Philosophy”, as he entitles his talk, even if that enterprise of the mind and human creativity in some important ways deviates from those of the sciences. With his subsequent writing, which is often hard to classify, Merleau-Ponty indeed demonstrated how philosophy can create a context for the thinking of one’s thinking, and for challenging that thinking, beyond any particular expert sector of the academic librarian’s. In front of the provost of Collège the France and other dignitaries of the highly esteemed academia, Merleau-Ponty demonstrates and asserts his project to reach out beyond the philosophical academia.

4. Life Worth Living For academic philosophy, the gure of Socrates is di

cult, if not an outrage. For a contemporary junior

philosopher struggling to meet his “key performance indicators” for a promotion, the actions of the most famous philosopher of world history sharply fail to serve as a model. As someone who never wrote anything, Socrates’s practices are a call for self-in ected disaster. Writing on what makes “life worth living” and on “a state of mind in which such a man lives”, as opposed to “those who are really not philosophers but have only a coating of opinions, like men whose bodies are tanned by the sun” (Seventh Letter 340c–d), Plato asserts: “There is no writing of mine about these matters, nor will there ever be one. For this knowledge is not something that can be put into words” (341c). These are strong words coming from someone considered one of the greatest writers of all time. p. 391

Plato thus seems to be making an explicit case against written philosophy, and giving us a strong reminder of its limitations. But the allure and glow of Plato’s written word is so majestic and the fame of his corpus so colossal that the professional philosopher nds a quick, welcome refuge from the nasty feeling caused by Plato’s statements against writing.

Nor will the academic philosopher feel worried about the dominant use of dialogue as a form of expression in Plato—a form never used in current academic journals. The modern philosopher might think that the use of dialogue as the form of written philosophy is a matter of stylistic concern and does not touch the substance. The beef is elsewhere, the modern journal-publishing writer-philosopher could think, even after it is pointed out that a host of informed readers have found the use of dialogue to have an important, even fundamental role in Plato’s philosophy. That there is certain unmistakable liveness and animation in the dialogues of Plato, as well as in the gure of Socrates and in the activities of Socrates also described by others, is nothing more than a bag of ornamental gimmicks, not of concern to serious present-day philosophers on the appointment committees. Matters of style do not touch anything signi cant in philosophy, it could be argued. If Bergson won a Nobel Prize for his style of writing, good for him personally. The case is philosophically incidental—just as with Bertrand Russell, who won the prize in 1950, “in recognition of his varied and signi cant writings in which he champions humanitarian ideals and freedom of thought”. It could be said that Russell’s “major philosophical works” never became bestsellers the way Unpopular Essays or Why I Am Not a Christian did, and indeed, “On Denoting” from 1905 was nothing less than an article in Mind.

5. Life as Lived It is uncontroversial that Picasso’s Guernica can touch people, as can Beethoven’s Moonlight Sonata and Tolstoy’s War and Peace. The Little Prince has touched many. As an empirical fact, it can be noted that people can be touched, with results that sometimes transform their lives. How should professional philosophy approach this possibility? By shying away, is one answer. Indeed, an excellent way to tackle the unease created by the actions of Socrates, the reservations about writing expressed by Plato, and the apparent success of philosophers such as Bergson or Bertrand Russell, with their style that reaches out to the general audience is to insist that connecting to people is irrelevant for philosophy. Philosophy is a content a air, it could be suggested, and not everyone is interested in its contents. How about “life as lived”, unmistakably a theme in Plato’s Apology, “Seventh Letter,” and a host of other writings? While “life as lived” is a theme most people might nd of interest, the same is not true of “life as lived” as a theme in Plato. The latter is what excites professional philosophers. But aspirations that relate to themes p. 392

are di erent from those aspirations that drive life as

lived. In particular, sensibilities that are at the core

of the latter might be irrelevant to the former. As an empirical fact we can note that few people hail as desirable a life as lived without love as a rst-person experience, along with appreciation, awe, compassion, amazement, and countless other aspects of a worthy life experience. Yet in many contexts of conducting philosophy, emotions are treated as an obstacle, a nuisance and distraction from the essential. Themes of signi cance should be approached with due detachment from the rst-person perspective and one’s subjective experience. Themes are discussed and analyzed best from without and impersonally, the thought runs, without drawing explicitly from one’s subjective sensibilities, feelings, and a ects. They are too elusive, ephemeral, and vague to serve as the foundation for anything academically solid. Yet given the clearly personal character of the speeches of Socrates in Apology, and given that his life is at stake, a student might be surprised to learn that her philosophy professor does not seem much intrigued by

the a ectual side of the goings-on, at least not in the classroom. The professor’s communication is completely devoid of emotion, of the kind one would expect is humanly appropriate given that a pending execution is implicated. That initial bewilderment will pass as the student realizes that the philosophy professor in fact follows a university paradigm, and the same can be witnessed in literature and music theory classes. A ects like empathy, compassion, and awe may be part of the human condition, but the cognitive imperatives of university life encourage a decidedly detached approach. Did the student not start to feel increasingly drowsy when her professor of French literature, who has published an often-cited article on The Little Prince, went through his insights?

6. Re-Reading Platoʼs Apology When Merleau-Ponty provocatively claimed, “The philosophy placed in books has ceased to challenge men”, his reference was to people’s lives as they are lived, as opposed to propositions about or theories of such lives. The contrast between what Merleau-Ponty calls “the true philosopher” and a “functionary” marks a dramatic change of perspective, focus, ontology, frame of reference, methods, and style as well as the interface with oneself, the human condition, and other people. Our interest lies in the implications of the shift for the practice of philosophy and for the possibilities that emerge for speech in the process. Merleau-Ponty was not an idiot. When he stated that in the hands of the functionary philosophers, “the choices of life are deadened and the occasions for thought are cut o ,” (1988: 33) he did not mean thoughts wouldn’t continue in lively intellectual debates in learned circles of his time. He did express his regret that those expert-cultural debates fail to challenge people and do not bear on people’s lives as they unfold in actual lived reality. p. 393

Few would question that Socrates at least tried. Indeed, among the gures in world history people can identify who might have meant something to them from the point of view of challenging one to live a life more closely examined, Socrates must stand out as prominent among people of the past 2,500 years. Touching and challenging people, with what he called a gad y kind of insistence, is one of the cornerstones of the Socratic approach to philosophy—the way of life, he made clear in his trial, that “as long as I draw breath and am able, I shall not cease to practice”. One way to read Apology is with a view to such an orientation in life. That readily opens for the philosopher a domain not often appreciated by academic philosophical practices: the present moment. For a professional pianist, the chance to re-enter Chopin’s “Scherzo No 2” is a welcome option. In fact, it is part of the reason she enjoys her profession. What a delight it is to nd something new in the familiar piece during a live concert performance—what a thrill to be able to give a touch of that something magical in real time to real people! Knowledge of the score of “Scherzo No. 2” is just one step for the pianist. It is through her sensitivities that she becomes enchanted by something in the piece that had always been there, but is now brought out freshly. She creates a re-entry to the score in the present moment. Is it not conceivable that a philosopher could do something similar with Apology? The present moment, with its unique concerns, tensions, and aspirations, does set the stage for a re-entry, a possibility that something in the text will come alive for that reader as she searches for deeper meaning in life.

We are here alluding to some familiar, even obvious, but hard-to-describe features of human experiencing. Are these not perhaps of interest to philosophy, as a point of reference for one’s practices? Whatever the process that took Paul Ekman, the famed emotion and facial expression psychologist, by storm when he met Dalai Lama in 2000, we have it on his own word and according to testimonies of others, that the event left him visibly moved on the spot and a ected him the rest of his life. A person can achieve insight into life and undergo change in one’s life, even dramatically. Erik H. Erikson in Young Man Luther (1993) sets himself the considerable challenge of trying to describe the internal struggles of a human being, something that is hard to carry out reliably, especially centuries after the fact. It might be that some of phrases Erikson uses will seem clumsy by the year 2200, but surely what Erikson aims to describe is a human possibility that is with us today, and will continue to be with us as long as there is humanity.

7. The Pivotal Question Is it possible that by overtly emphasizing the “intellect” and the conceptual, themes-and-propositions and theories, we as philosophers inadvertently end up diminishing life, instead of enriching it? I’m reminded of a remark by a famed non-philosopher friend of mine from an Ivy League university, as he politely explained p. 394

why he had

stopped going to philosophy seminars. He felt there was too much “negativity” in the

meetings, too strong a methodological faith in “criticism” as the chief instrument to move forward. If you go to Plato’s Apology, you will encounter themes that have been found to be conceptually complex during subsequent centuries of analysis. Concerning anything Socrates brings up in extant documents of his speeches at the trial, discussions of themes such as “justice”, “virtue”, “citizenship”, and “soul” continue. The self-generating nature of intellectual, conceptual, and theoretical discussion is reinforced by the institutions that have built incentives for spiraling up such discussions. Those discussions are remote, almost completely cut o

from the everyday lives of most people. As

academic discussions of “virtues” continue in some of the best funded, most esteemed universities in the United States, the vast majority of the people of the United States remain untouched. The intellectually brilliant analyses and debates conducted as standard philosophical practice at universities do not stir interest, even among colleagues from other humanities departments. Is the human, all-too-human, reality of people in Athens, Georgia, or Athens, Greece, or the Athens of any other location, too mundane a reality for the liking of a professional philosopher of the day, too full of banality and intellectual naivety, to warrant attention? What kind of philosophical writing, acknowledged by tenure processes, could have reached and a ected anybody that stormed the US Capitol on January 6, 2021? Acute attention to what in fact drives people, documents of Socrates tell us, occupied him day in and day out. From the point of view of our Socratian, or “Merleau-Pontyan” meta-philosophy, and the applied philosophy of language that goes together with them, a key question concerns the status we want to assign to people as people, in their everyday naivety, contradictions, “animal spirits”, super cialities, desires, my-biases, tendencies for non-openness, hopes, hatreds, strives towards something “transcendent”, etc. To what extent do we think serious philosophy, and the use of language for conducting philosophy, should address head-on our humanity in that sense? This is the pivotal question.

8. Pointing Beyond Academic Hair Splitting “I think a lot of people would say,” the Nobel laureate replied, “that I have tried to put people back to the center of economic theorizing.” I had asked Professor Edmund Phelps, the recipient of the 2006 award in economics about his contribution, as he saw it. I had found exciting his work on what he called “mass ourishing” and “good economy,” in Phelps’s vision on something that pointed far beyond standard economics. His opus magnum (Phelps 2013) has the telling subtitle “How Grassroots Innovation Created Jobs, Challenge and Change”. p. 395

Phelps’s words startled me. I recalled his Nobel acceptance speech: “The neo-classical growth theory was conspicuous in having no people in it.” Still, somehow hearing the point directly from the tall, smiling, generous academician, not three feet way from me, hit me as a kind of home run in thinking. How is it possible, a aneur could ask, that people were absent from economic theorizing in the rst place? Isn’t it people who create the whole of economic phenomena? But a map can become more exciting than the territory it is a map of. I submit that most people would nd the following distinction comprehensible: 1. Economic theory that takes people into account as people 2. Economic theory that takes people into account as abstractions It might take some e ort to make someone admit that the contrast between the two statements in fact is not so clear as one might think on the face of it. It might take some ve minutes for you to explain this to a student. He or she need not read Quine’s “Two Dogmas of Empiricism” to see the point. In fact, if she were to try to read the classic article of academic analytic philosophy, she might be more confused as to what the point is, as compared to your explanation. Part of the reason is that during those few minutes of your explanation, she would be drawn into a shared space of communication in a present moment. She would experience you and the tone of your voice, your silences and punctuations—the stimulating, compassionate teacher in you. When reading Quine’s article, she would focus on the text. The experience would be dramatically less multidimensional compared to the oral situation in which she, for instance, connects with your eyes in the experience of the shared now-moment. We do know, based on social neuroscience, that things happen in her as she listens to you that would not happen in her when reading Quine’s “Two Dogmas”. Five minutes of reading an intellectually highly condensed article is a di erent experience from ve minutes of discussion. Yet the discussion is also condensed, only di erently. Only a portion of that happens is in the explicit, verbal, or conceptual dimension. In comprehending propositions such as (1) and (2), we talk about the understanding of natural language. The understanding of natural language, as LePore (1982) emphasized in an early paper, is the focus of Ernest LePore and Donald Davidson’s groundbreaking theorizing. But are people as people, and actions as actions, part of “Davidson’s program”? People as people and actions as actions are certainly part of Socrates’s program. “He is there in life,” as Merleau-Ponty (1988, 36) put it. People as people and actions as actions are what Apology takes pain to convey to us as what Socrates is struggling with in passages such as this:

“I shall not cease to practice philosophy, to exhort you and in my usual way to point out to any one of you whom I happen to meet: Good Sir, you are an Athenian, a citizen of the greatest city with the p. 396

greatest reputation for both wisdom and power;

are you now ashamed of your eagerness to

possess as much wealth, reputation and honors as possible, while you do not care for nor give thought to wisdom or truth, or the best possible state of your soul”. Then, if one of you disputes this and says he does care, I shall not let him go at once or leave him, but I shall question him, examine him and test him, and if I do not think he has attained the goodness that he says he has, I shall reproach him because he attaches little importance to the most important things and greater importance to inferior things. I shall treat in this way anyone I happen to meet, young and old, citizen and stranger. (29e–30a) The realm that Socrates struggles to reach in Apology is living one’s life in some such sense as a human being might understand as parallel and similar to her own. This is how many people a ected by Apology have understood it throughout the centuries. For that to happen, people need to have connected not only with the text but with what it implies. There is a necessity to the fact that as long as you live your life, it is you who live it. You live your life in some intuitive “being trapped and without breaks” sense. You continue to live as long as you are alive, without a chance for a timeout in the haven of pure concepts. What Merleau-Ponty found lacking in “the philosophy placed in books [that] has ceased to challenge men” is philosophy that would follow the lead of Socrates and address life in that sense. “For in the last analysis each one of us knows for his own part that the world as it is, is unacceptable”, Merleau-Ponty (1988: 34) observes. These words, if uttered by one’s son or daughter with due emotional conviction, would not lead you to intellectual snobbery. You wouldn’t start a conceptual hair splitting on this ground or that. You would feel that your child’s life is moving in a direction that you would feel grateful to reinforce. Right from the start, Merleau-Ponty sets the tone of his inaugural speech to point beyond academic hairsplitting. He is not making the case for “universal unacceptalism” or any other ism. He points to a universe, imperative for a “true philosopher” to enter, which does not reduce to concepts, propositions, or intellectual stances. The point of such philosophy is life itself, and people as people.

9. The Socratic Call Most people piss in the morning. Even the stoutest Clarity First conceptualist proceeds to his morning actions promptly and without waiting for solid conceptual clarity regarding what he is up to. When it comes to actually living one’s life, the philosopher is “thrown to the world” just like everyone else. He will remain at a loss just like everyone else, will stay human, all-too-human, and be forced to go through the struggles of life that are often conceptually opaque, paradoxical, and unsolvable. The philosopher might be at home in the realm of the particular concepts of his expert domain, but that’s di erent from being at home with a toothache. When it comes to living one’s life, the philosopher does not have the luxury of a legitimated footing in an Absolute or First Cause, any more than Thomas of Acquinas or Bishop Berkeley did. p. 397

The philosopher needs to live his life, even in the absence of a conceptual self-legitimating description of that life. The rst philosophy of anybody’s life necessarily contains non-philosophy, and its form is a verb. The Socratic call is to acknowledge that non-philosophy.

10. Ordinary Language Philosophy in Athens “How to do things with words”, J. L. Austin asked, and created a sensation. In Austin’s time, during the heyday of debates that assumed philosophy is concerned with descriptive statements to this or that e ect, as well as with their logic, it was striking for many professional philosophers to be reminded that, in actual reality, language can be used to a host of other purposes beyond the descriptive—that “meaning” can be a function of “use”, and thus involve aspects philosophers had overlooked when presenting a comprehensive “philosophy of language”. Austin did not ask: How to do philosophy, of the kind Socrates exempli es, and Merleau-Ponty advocates, with words. If Austin had asked this question, he probably would have turned to the use of language by Socrates, noting its key characteristics in action-terms. The unit of analysis of such a hypothetical Austinian analysis of the ordinary language of Socrates would have recognized situations that are, without exception, oral. Furthermore, the situations would have been recognized as ordinary, even necessarily so, because Socrates makes it clear that his philosophizing can and will be launched anywhere and everywhere, in everyday situations in Athens, where he just happens to meet a fellow citizen, or a stranger, young or old. Socrates is not going to wait for a right moment to arise or require his partner to ful ll some preset intellectual standards regarding the clarity of thinking. Any encounter will do, and there is no kind of habitual or institutionalized organization for the context in which Socrates is going to engage in “how to do philosophy with words”. The set-up is oral and ordinary, and any citizen is a potential participant. This is the action space of Socratian philosophy, one that Merleau-Ponty, in his inaugural speech, describes with such admiration. If philosophy is thought of as a sport, the opportunities to score are vastly bigger than the concepts-andpropositions-loving-theme-devoted philosopher had imagined. Still, the conceptual purist will remain lukewarm. He won’t feel comfortable discussing life as lived even if he might be willing to debate what particular “philosophies of life” as doctrines propose. As a professional philosopher, he might signal knowledge of Stoicism, but with that his scope for practicing philosophy has grown only marginally. Even after covering some explicit “philosophies of life”, the lives of people who p. 398

attend the class are likely to remain as untouched as before. Nothing personal: modern

philosophy

classes do not aim to be “spiritual exercises” in the sense in which Pierre Hadot (2002) has demonstrated was the case in antiquity. And people are not that interested in “philosophies of life” to start with. The connection is not there. Even when the occasional book, say, on existentialism, makes its unlikely way up to the New York Times Bestseller list, most people in the United States remain ignorant that the book exists. Philosophy books are not that widely read, even among people who read books. Almost all people, in contrast, are interested in their lives. On some important and real level people care about their lives, even if few care to engage with philosophies of life as doctrines. So much worse for most people, the philosophical purist could conclude. He returns to his communicative arsenal developed for the miniature audiences that value the critical, analytical, and propositional methods of academic philosophizing—people he has access to and with whom he connects. “That’s democracy”, he might say, as did, in fact, one of the greatest of American contemporary philosophers when I asked why he thought hardly anybody knew of him even in his home state of Massachusetts.

11. Socratian Oral Philosophy If one assumes the understanding of natural language should take the place in the context of a Socratian engagement and dialogue of the kind alluded to in Apology, as opposed to a context where someone utters, for instance, “There’s a bird on the tree,” more would be at stake than just an intellectual understanding of the language that is used. One striking aspect of Socratian oral philosophy, an aspect of it that is not implicit or hinted but brought into focus absolutely in the open, is the sheer magnitude of the potential impact level that is laid out for the “total function of the philosopher”. The horizon of impact for philosophy as Socrates envisions it is to potentially transform the lives of people, the lives of any and all, and including the philosopher himself. How Socrates acts in the face of a death sentence is evidence for the truth of what he says, some would say. This way of understanding “truth” might point beyond Tarski’s theory of truth, but I believe many would hope Apology will be also read in the future, with an understanding that acknowledges “truth” in Socrates’s words, and based on what he is reported to have done at his trial. A transformative intent that reaches out to the lives of actual people as people, however ill-informed they might be about philosophical theories of virtues or the methods of conceptual analysis, is certainly a possible meta-philosophical position. As a premise for philosophical action, such thinking will lead to a philosophy of communication that sees no reason to limit itself to writing articles for academic-expert audiences. We would enter the domain of understanding natural language from an angle that is di erent from what we are likely to discuss as “semantics” amongst our philosopher of language colleagues. The stakes are now higher than anything you p. 399

encounter

in a promotion process at a university, or in a philosophy colloquium. The leverage of

understanding the use of language, when what is discussed is your potential execution as judged by some 501 fellow citizens, is certainly a phenomenon of distinction. If the theme of a life worth living is brought to bear on one’s life’s actual dramas and tragedies, in the context of one’s everyday, with a Socratian seriousness and esh and blood, the stakes explode to higher reaches compared to what we encounter in any conceptual debate on a theme. A debate that takes place on the conceptual level does not lead, per se, to the death of a human being, whereas in actual life, in the present case, that did happen. Once Socrates is dead, the debate regarding something he said or implied can continue, but Socrates is dead. What could the highly able professional philosophers of Cornell, Harvard, Yale, Stanford, and UC Berkeley have done di erently for one of the persons who stormed the Capitol to stay home? I submit the answer is not, “Write more sharply argued articles for philosophy journals.” Nor, “Nothing.”

12. The Road Less Travelled Understanding what Socrates and Plato are after in Apology will certainly have to build on our intellectual, rational, analytic, and cognitive faculties. Yet more is involved. Let us recall the key contrast that professional philosophy so easily dismisses. It is one thing to read Apology and become impressed by the integrity and truth of Socrates, and another to engage with the text when preparing for philosophy symposium on Plato. In the former case, we talk about the communication that takes place between Apology and the reader, in the midst of her life; and in the latter, about communication within a speci c academic-expert community, conducting the a airs of their academic life.

As far as we know, the possibility of understanding something expressed in a natural language, say, the text of Apology, and then turning that experience into something transformative on the level of one’s life as lived, is a uniquely human possibility. However di

cult it would be to describe what such transformation

amounts to conceptually, psychologically or neuroscienti cally, or explicate the exact historical instances that demonstrate the phenomenon, it seems that Socrates in Apology points in this direction. It seems uncontroversial that a philosopher could set herself the challenge to serve such a call. It is easier to measure students on their knowledge of Plato’s early dialogues than with regard to their human growth. The case is generic: it is likewise easier to help someone to progress as a basketball player than as a human being. It is unclear what “human growth” even amounts to, and the results will be seen perhaps only after years or decades. John Wooden attempted it. His stated aim as coach of the UCLA Bruins, in the twelve-year period that lead to ten championships, was to contribute to the lives of his players beyond basketball. p. 400

“The Wizard of Westwood” considered himself to be “a teacher”. Few departments of a university, outside of philosophy, could cite an example as towering as Socrates as pointing to the John Wooden way of being a teacher, to that road less travelled.

13. The Socratian Transformative Intent If we take the Socratian transformative intent as a starting point, two conclusions follow: 1. The success criteria of one’s philosophizing cannot be de ned in isolation from other people, most of whom are non-philosophers; 2. The success criteria of one’s philosophizing will involve parameters beyond objectively identi able “content”. In Apology, Socrates does not o er the trial a lecture on the human nature. He does not present or defend himself as an investigator of meaning or of the human condition. Quite explicitly and proudly, he presents himself as a man of change. At this de ning moment for philosophy, “the total function of the philosopher” links with living a life with re ection, challenge, and connection with other, rather than with the creating of propositions or the making of theories thereof. The total function of the philosopher, I submit, is a matter of orientation. It is concerned with how one perceives oneself vis-à-vis other people and one’s life. It calls for a re ective stance to life, in the context of a life unfolding in the present moment and lived in the midst of uncertainties, paradoxes, and injustices, in and among fellow citizens from whom one does not isolate. After being sentenced to death but o ered a chance to leave the city and stay alive, Socrates does not yield. He does not give up his philosophical way of life. He does not escape. Socrates refuses the invitation to enter his own bubble and continue to live excluded from his fellow citizens. Right to the end, the rst philosopher de nes his actions on connectivity to others and to himself from within.

14. Two Meta-Philosophies for Applied Philosophy of Language Let us distinguish two meta-philosophies for applied philosophy of language: A. In a philosophical practice language should be used primarily to express propositions for the p. 401

philosophical expert community to scrutinize and evaluate.

Development takes place through the

peer evaluative expert discussion, much the same way as does the paradigm in other sciences. B. In a philosophical practice, language should be used primarily to enhance the possibilities of re ection by the people who participate in the practice. Development takes place within people as internal dialogue. This distinction is intentionally heuristic. Position B, as exempli ed by Socrates, is illustrated in Apology in lines such as this: “For I go around doing nothing but persuading both young and old among you not to care for your body or your wealth in preference to or as strongly as for the best possible state of your soul” (Apology 30a–30b). It is trivial to recognize that from the paradigm of position A, one can raise smart objections to B. Whatever Socrates expresses, indicates, or exempli es in Apology can be challenged by position A criteria, just as one can read, study, and analyze any human artefact intellectually, in a bubble, and without any e ect on the conduct of one’s life at home or vis-à-vis one’s fellow citizens. But B refers to an option on how to live a life.

15. The Sauna Question “Why are you guys just talking with one another, instead of us non-philosophers”, my Special Lady said around 1984 in our sauna. Our love had ashed to ames. This question came to haunt me. When it comes to enriching one’s life through deliberate thinking and the re-thinking of one’s beliefs and convictions, prejudices, emotional responses, patters of reaction, features of personality, the various habits, routines, and ways of thinking that make us what we are, it seems uncontroversial that anyone could bene t. It is part of “the human condition” to have access to an endowment of thinking and of the thinking of one’s thinking, to faculties that promise better life as a result of better thinking. Why would philosophy bypass the possibility? Socrates did not. One way to read his legacy is to say that the philosopher’s task is to create conditions under which people can improve their lives, “souls”, as Socrates put it, using the faculty of reasoning and deliberation. Should a philosopher pick up the Socratian challenge only in Greece? Socrates certainly did not say so, but there have been times when Latin was assumed to be the preferred language of the truly signi cant communication. In our time, the assumption is that articles written for expert audiences is the ideal in philosophy. Why restrict? One reason is provided by the organizational and institutional context in which one’s life as a philosopher takes place. If one operates in a church at the time of St. Acquinas, the requirements are di erent as compared to a modern university context. Stepping beyond the Church or the University to the marketplace of one’s city will certainly call for communicative practices that are more generally shared. If p. 402

one’s philosophical practices are to resonate communicatively with people

of the marketplace, creativity

is called for—of a kind your institution did not prepare you for.

16. Bringing Back the Context If “the total function of the philosopher” turns us to other people, as opposed to an institutionally de ned, select group of clerics or academics, we end up studying ourselves as human beings among other human beings.

As a philosopher examines one’s life as a human being, one of the rst surprises is provided by the something nobody can overlook in one’s actual living—context. Considering all its trivialities, it might be tempting for a philosopher, in his desire to embrace the abstract, to overlook the context in which one’s life takes place. But why to de-contextualize philosophical communication a priori? Why narrow the broadband before the fact, why cut down the forms of potential connectivity? Again, that Plato often sets his dialogues in a context might be thought to represent his cute, but ultimately irrelevant, preferences. That they create accessibility for people can be considered insigni cant in contrast to the real philosophical “content”. It can be argued that it is incidental that Socrates raises the question of courage speci cally with Laches, the war hero, rather than with, say, a playwright. The case does not show sensitivity to the discursive context on the part of Socrates or Plato. The alternative view that I nd intriguing says that context matters. Context is part of human communication. Diogenes Laertius’s Lives of the Eminent Philosophers is not much appreciated in academic philosophy curriculums, because he “frequently focuses on trivial or insigni cant details of his subjects’ lives while ignoring important details of their philosophical teachings” (as Wikipedia puts it on November 13, 2021). Diogenes Laertius is not known to be reliable as a historian, but maybe along with the context, imagination also counts—the possibilities that emerge in the mind of the reader. Diogenes Laertius communicates. Montaigne has certainly served the re ection of life’s possibilities for many, but his Essays are routinely pushed to margin in standard expositions of Western philosophy. Modern academic philosophy is fond of focusing on examples that are extremely remote from anything anybody is likely to encounter—like a trolley on the way to hit people, with you at the wheel. How about real-life examples? If you take your job as a philosopher to be to communicate with other human beings qua human beings, as opposed to experts, it would seem natural to turn to empirical work on thinking for increased understanding. You might be struck by what you nd in the eloquent, richly argued writings of the likes of Jerome Bruner, the famed psychologist-thinker, who has impressively brought into focus the fundamental signi cance of narrative for human thinking. As a result, in line with position B above, you might be encouraged to inject stories to your philosophical p. 403

communication. You might even be struck by the question of the

metalevel narrative that you as a

philosopher generate for the people with whom you engage in a given context in your philosophical practice. The metanarrative of intellectual brilliance and paternalizing elementary-school-teacher-ism will probably serve only particular contexts, as does that of a razor-sharp propagator of some set of xed views. If the idea is to enhance people’s thinking as it connects with their actual life, we would likely encounter with enthusiasm many philosophers such as Sapho, whom many people have found and do nd inspiring despite the doubts regarding her doctrinal impact. If imagination is allowed as a resource from which to draw indigenous innovation to one’s life, one can envision those strolls with an Aristotle and his bunch. How stimulating it can be to walk around and discuss. How wonderful it is to allow your thinking to dwell and to wander, to move disrupted, with pauses and silences, while hearing others dialoguing excitedly, creating a shared moment of ow and mental jamming. Like generations of people before us, we could be delighted by the autobiographical way Augustinus in Confessions manages to talk to us in some such direct way as to question the dimension of faith in our life, irrespective whether we are “Christians”. We could look with awe to Kierkegaard’s passion with his various pseudonyms, and to return to Tractatus with the kind of solemn admiration that lled us when we rst read its concluding lines in high school.

Were we not thrilled when we learned about Wittgenstein’s “lectures”, given for twenty people or so, always with the title “Philosophy”, at which he spoke with an intensity nobody would forget? We could recall what George Steiner said of another “anti-academic academic”—Heidegger: “Witnesses, such as Löwith, as Gadamer, as Hannah Arendt, are of one voice in saying that those who did not hear Martin Heidegger lecture or conduct his seminars can have only an imperfect, even distorted notion of his purpose.” (Steiner 1991, xiii). When deliberating on such philosophical practices that leak out from the box of academic philosophy, it seems strange to suggest that we should limit our philosophical considerations and communication to decontextualized and impersonal, linear and article-formatted expert-directed written text only.

17. Life as Not an Accessory to Philosophy With or without the help of professional philosophy, many if not most people at least tacitly do in fact use deliberation to live their lives better. For many this happens when getting children. For some a sudden tragic disruption in life calls for renewed re ection on one’s ways and manners. Most people, at the deathbed of a loved one, adopt something like a Socratian stance when viewing life towards the uncertain future. But do you continue to examine your life? p. 404

Do you challenge your reasoning

and intellectual powers to live up to the current realities, instead of

continuing business as usual and the life of “the everyman”? Each of the 501 Athenians in the deciding body at the trial of Socrates probably felt fairly pleased with themselves as thinkers. On some obvious intuitive and signi cant level, it is close to impossible for a human being not to think. But at the same time, it is easy enough to leave that thought unthought. Whether a person deliberates his thinking from the point of view of broader perspectives, asking questions such as, “Is this right”, “Am I being driven by hatred motives”, “What happens if this becomes a rule”, as embedded to the issues and tensions of one’s own life, including the immediate situation, will vary from person to person and situation to situation, but the option remains for all as part of the condition of life—human life. Socratian meta-philosophy, the total function of the philosopher, builds on the assumption that anyone can deliberate more carefully, more deeply, more truthfully, and more consistently, and that the philosopher’s task is to build practices on that premise. This the back and the bone of meta-philosophy B. The cornerstone of the present chapter is the claim that in Apology a particular kind of conduct of philosophy is advocated, articulated, and exempli ed, one that reaches out to us today as a radical challenge. Thus understood, life is not an accessory to philosophy, and philosophy is not ornament to life; they are intertwined throughout, and should together be perceived as forming the fundamentals of a life worth living. Life worth living concerns everybody, encompasses anybody, and as a challenge is part of the human condition itself. “The total function of the philosopher” points beyond expertism, privilegism, and functionarism and situates the philosopher among those cultural forces that do not ask for race, color, status, degree, IQ, institution, or a particular site in order to launch, engage, contribute, and grow. Thus doing philosophy approaches art, perhaps especially the performative arts, and takes some gentle but respectful distance from what’s optimal in natural science. Inspiration from therapeutic practices might also be forthcoming, but that is a question not explored here.

18. Bad News to Abstraction Enthusiasts The deeper we push the practice of philosophy towards real people living their lives in the midst of their every day, the less secure are sophisticated abstractions as instruments of communication. This is bad news for abstraction enthusiasts. If the chief source of excitement and the assumed professionalism of the philosopher comes from his abilities to manage, play around, and discourse on expert concepts with other experts, the everyday as a forum of encounter is not inviting. p. 405

In human interaction and communication, especially in non-verbal art, such as music, one can make use of an abstract concept such as “love” without becoming trapped by the concept or its explicit representations. One can demonstrate “justice” powerfully from a theatre stage without using the word “justice”. One can stimulate others to think for themselves without making them read Kant’s “What Is Enlightenment?” or saying to them, “Think for yourself!” One can have a conversation with a nouveau riche friend, and without any explicit reference to phenomena such “modesty”, “greed”, or “altruism”, might save the life of his suicidal daughter, who feels rejected by her super-performing parent. To have a Socratian impact on the other, to have an e ect on somebody in the realm of “know thyself” and “life worth living”, using one’s intellect, reasoning skills, and the thinking of one’s thinking, there is no clear-cut criteria for how concepts should be used. For instance, despite their unquestionable status in the Western cannon, even the following words do not carry far as they are read out in most situations: Good Sir, you are an Athenian, a citizen of the greatest city with the greatest reputation for both wisdom and power; are you not ashamed of your eagerness to possess as much wealth, reputation and honors as possible, while you do not care for nor give thought to wisdom or truth, or the best possible state of your soul. (Apology 29d–29e) At the same time, in the right context, even a simple word can help to open someone’s eyes to something life changing—for instance, by starting to attach “little importance to the most important things and greater importance to inferior things”. For such a Socratian e ect to emerge in the path of a human being, philosophy as an explicit doctrine is like some speci c theory in a therapy context: not a necessity. Academic philosophy is more a back-o

ce a air.

A certain intention is at the front line. A certain interaction to emerge within the person, with life itself, is of the essence. This is where “the total function of the philosopher” emerges. That benevolent intention of contribution with the use of language and rational reasoning, in a context of humanly tuned interaction, is not often highlighted as a cornerstone in the philosophy and the meta-philosophy of Socrates, but I submit that is what is involved in “the total function of the philosopher” of the kind Merleau-Ponty describes.

19. What Infant Research Teaches Us Academic philosophers might nd the concepts “intention”, “agency”, “initiation”, “self-regulation, “expectation”, “pre-verbal experience”, and “interaction” opaque and confused, but the scientists of early human development consider them useful, indeed, fundamental. p. 406

Here’s how Beatrice Beebe and Frank Lachmann, two groundbreaking researchers who have brought the ndings of infant research to bear on therapeutic treatment for adults, summarize some of the fundamental

ndings of themselves and those of their colleagues: There is extensive experimental evidence that, from birth and even before, babies form expectations of predictable event[s] … infants are biologically prepared to detect regularity, generate expectancies, and act on these expectations. (Beebe and Lachman, 2002, 150) Is self-regulation in interaction with one’s mother, and the reading of intentions, in some such sense as depicted by infant research, relevant for understanding what might take place when spoken language is used as part of “the total function of the philosopher”? I believe we here enter the territory of human growth—and of contextuality, communication—that has been bypassed by philosophy but should be taken seriously as part of the challenge of Socrates. The total function of the philosopher calls for intentions, as perceived by us already pre-verbally, to be acknowledged as an ally in the connectivity, communication, and attunement one can have with one’s fellow human beings in the context of an oral philosophical practice. Let’s not pretend we could postpone. There could be intellectual and conceptual reasons why we should halt some promising action until clarity is achieved on some named prerequisites. But why should we follow the moral of intellectualism as opposed to the moral of our grandparents, the moral of ordinary life since time immortal, that says we ought to live our lives as well as we can, together with other fellow humans, challenging ourselves as part of our humanity while aiming at a better life, and using our rational faculties to scrutinize and deliberate on what’s right in the particular contexts where our life takes place, trying to make sense of it all. A possible moral of life says that we cannot wait, but must engage in striving towards betterment right now with whatever sense-making and reasoning skills we have. When as infants we tune to the intentions of our mothers and other caretakers, we do that by forming expectations and by joining into “moments of meeting” and “mutual recognition”, to use some key words from work of Louis Sander (2008), one of the groundbreaking pioneers of systemic infant research. It is now well-established that the key to early development is the “bi-directional dyad” in which “inner and relational processes are co-created in tandem” (Beebe et al. 2005, 11). The mother and the infant form an interactional system in which “ ‘individuals’ do not exist apart from the totality of their interpersonal relationships” (10). Before they master any words or concepts, the infant’s systems and attunement skills are the key to “relationships in development” (Seligman 2017). The “communicative musicality” that unfolds (Malloch and Trevarthen 2010) is part and parcel of the non-verbal process of the cradle of thought, the moment-to-moment dyadic system that infant researchers insist is “the most basic unit of interest” (Beebe et al. 2005, 10). p. 407

The di erence with a philosophy student could hardly be stronger. The student is not invited to “mutual recognition” and into a “co-created bidirectional dyad” in which her “self-regulation” develops, along with the interactional, ne-tuned, moment-to-moment process in which both are constantly in uenced by the other. In a philosophy class, the student experiences a potentially hostile environment, where her “initiations” reduce to externally accented expressions of questions or calls for clari cations, but her capabilities for intersubjective connectivity, demonstrably rich since birth in its initiating powers, attunement to connection, and as a vehicle of development, are treated as nil. The baby and the mother, viewed developmentally, are “beyond the doer and the done to”, as should also be the case in therapy, as Jessica Benjamin (2017) aptly puts it. A student in interaction with a philosopher could also be viewed in such terms, but that would require that the developmental perspective be assigned the priority, and the philosopher would surrender his I-am-the-doer stance in an assumed dualistic doer-

done-to metalevel narrative. It would require the philosopher to step down from his one-sided, selfproclaimed position of authority towards “moments of meeting” and “mutual recognition”. The philosopher would have to take the other seriously as someone who can grow and develop indigenously, moment-to-moment with the dyad unfolding in the present moment, and as someone who in uences the philosopher in real time. In oral philosophical practices informed by the infant research, the philosopher can’t remain detached, disconnected, and untouched. The “total function of the philosopher”, thus envisioned, calls for humbleness from the philosopher, unimaginable for him qua one-directional “doer-to” authority and informer-of-content. It calls for acknowledging that human growth and development are fundamental to philosophy as a relational and lived phenomenon. In the hands of a Socratian philosopher, the dyadic model of human growth depicted by infant research and cherished by relational therapists, will take advantage of the higher cognitive human faculties of the intellect and the verbal, yet holds onto the co-created nature of where the development comes from. The foundation for growth remains the same as in the beginning of life—relational.

20. The Oral Philosopher as a Mid-Wife What kind of challenges for philosophy are in o

ng, if growth and development are envisioned in relational

terms? Speaking somewhat bluntly, and pointing beyond “content” and “themes”, the parameters that emerge as particularly relevant for the purposes of the present chapter concern communication, context, reception, and the living presence. p. 408

In Plato’s dialogue Theatetus, the theme is “knowledge”, and in Symposium, it is “love”. In both dialogues Plato describes the context in which the themes are developed, eloquently indicting the various ways Socrates communicates and relates to the people at hand. In these aspects of the dialogue that relate to context, communication, and reception in the orally expressed cues and perspectives of Socrates, and following the propositions-centered meta-philosophy, the philosophical core is assigned to the “content”. When it comes to a live musical performance, few would deny the relevance of the context, communication, reception, or living presence for the proper account of the event, the experience, or of what the music is all about. It is not enough to point to the score and the notes. If the possible practices of an oral philosopher are approached from the angle of performing arts, it would be natural to consider such phenomena as: i. The philosopher’s communication with the audience ii. The philosopher’s use of the context as a platform for the philosophical practice iii. The reception of the philosopher’s communication iv. The living presence the philosopher and the audience co-create While perhaps somewhat clumsily formulated here, critical for our concerns in this chapter is the claim that (i)–(iv) point towards something that is philosophically relevant in spoken philosophical practices that seeks to facilitate human growth through philosophy.

Such relevance is particularly pronounced, if what is addressed is life itself with a Socratian seriousness and hope for impact. Now “reception” is of the essence. Surely, contextual factors are critical. Clearly, how the speaker manages to connect and communicate with the audience is relevant. Such parameters point beyond “content” and “the literal”. If the total function of the philosopher involves reaching out to people in order to touch them, while discussing life’s fundamentals, trivially oral practices o er possibilities of intersubjective nature that the philosopher should try to master. Nothing wrong with attempts to use Symposium for launching more love to the world, but an oral philosopher realizes how weak any text per se is in front of a live audience. You need to tune your delivery to touch people, if your passion is to create in an auditorium a living moment in which people could engage in thought processes of their own that bear on love, with potential e ect in the way love actually becomes embodied and experienced by them and the people in their lives. People can engage with their thinking in such a mode. People can engage in the thinking of their thinking in a way that makes a di erence to their thinking and actual practice. A philosopher can aim to facilitate that. Enter the “midwife” metaphor, favored by Socrates but awkward from the point of view of the written philosophy-of-position-taking. p. 409

The idea in viewing the role of a philosopher as that of a midwife is to emphasize the active and primary role of the philosopher’s partner in the growth-enhancing dyad with the non-philosopher. In a university class, the philosopher makes the students gure out what the teacher already knows: the student is supposed to give a second birth to what was originally the philosopher’s baby. Not so in the Socratic midwife format. It’s not your baby, and you don’t even beget the child. The student gives birth to the insight herself, and the baby might not have anything to do with you—the birth could take place years later, with few traces of where it originated. If the idea is the other giving birth to the child, it makes no sense for you as the midwife a priori to restrict your supportive actions to only what’s in a formbook or xed as a template. Cases vary, and each birth has its speci c features in the unfolding living presence. In your philosophical practice it could sometimes work if you used your own life as an opening of the re ections. Or used examples from popular culture. Music might help to create rhythm. In all instances, you are likely to want to make the other feel respected, recognized as a human being, ready for the challenge. You can use silences, accent, and intonation for increased attunement and shared intersubjectivity. Your seriousness and felt authenticity, honesty, and warm humanity might be singularly important. Yet it is the other that is in the focus. You serve.

21. Empirical Findings When I started as a professor at the Helsinki University of Technology in 2001, I did not ll a pre-existing philosophy chair. There wasn’t one at the university. I was invited to join the faculty at the Systems Analysis Laboratory with a broad task-assignment from the head of the lab, Professor Raimo P. Hämäläinen, to teach whatever I thought tting for engineering students. We came up with the title “Filoso a ja systeemiajattelu” (Philosophy and Systems Thinking) for a course that would be o ered to anybody at the university.

Filoso a ja systeemiajattelu was conducted in Finnish. Some six thousand students took the course before my retirement in 2021, more than 95 per cent of them voluntarily. In 2013 we began putting videos of the lectures, of 2 hours and 15 minutes of e ective time, on YouTube. As I am writing this (beginning of September 14, 2023), the number of views of the lectures is 2.15 million. The rst lecture of the 2021 season, in which I try to explain the basic idea, has 717 209 views with an average watch time of 29 minutes 22 seconds. When I presented the concluding lecture of Filoso a ja systeemiajattelu for the last time, Helsingin Sanomat, the main newspaper in Finland, streamed it on their webpage. 101,000 people joined the live stream. The lecture was attended live on Zoom by the 536 of listed and 896 open webinar participants. As of this writing the lecture has 68 797 views on YouTube. Finland is a country of 5.5 million, and the language is not spoken outside Finland to any signi cant degree. p. 410

At least, in the country of Finland, there is a demand for something one could call “life-philosophical lecturing”. I note that the thousands of essays on “the landscape of my life” that the students submitted at the end of the course make for reading that few experienced pedagogues would read without being touched. Through this example, I try to make the following claim: it is possible to carry out philosophical lecturing, an oral philosophical practice, that people can relate to in a way that makes it possible for them to engage in the thinking of their thinking and to re ect on their lives in way that makes a di erence (by their own standards). I believe we professional philosophers should acknowledge that possibility. I stress that “content” is only one part of what counts in a successful Socratian lecture of the kind I believe I have conducted. Content is important, but the main challenge is elsewhere. The prime challenge is provided by what I believe is the pivotal purpose of “the total function of the philosopher”: human growth. The philosopher is acting as a midwife. The primary point is not to instruct the audience. The chief aim is to help people to re ect themselves and to think of their thinking on themes that matter, in real time. This is not a dream. I believe my experience shows that it is possible to do philosophical lecturing that people from various walks of life can appreciate and relate to without giving the hearers already-acquired truths. It is possible to present to people the idea of free investigation in the context of their own lives. I submit the example demonstrates how to do philosophy with words, philosophy of the Socratian kind, and with the kind intention Merleau-Ponty alluded to with “total function of the philosopher”.

References Beebe, Beatrice, Steven Knoblauch, Judith Rustin, and Dorienne Sorter (2005), Forms of Intersubjectivity in Infant Research and Adult Treatment (New York: Other Press). Google Scholar Google Preview WorldCat COPAC Beebe, Beatrice, and Frank M. Lachmann (2002), Infant Research and Adult Treatment: Co-constructing Interactions (Hillsdale, NJ: Analytic Press). Google Scholar Google Preview WorldCat COPAC Benjamin, Jessica (2018), Beyond Doer and Done To. Recognition Theory, Intersubjectivity and the Third (London and New York: Routledge). Google Scholar Google Preview WorldCat COPAC Erikson, Erik H. (1993), Young Man Luther (New York: W. W. Norton). Google Scholar Google Preview WorldCat COPAC Hadot, Pierre (2002), What Is Ancient Philosophy? (Cambridge, MA: Harvard University Press). Google Scholar Google Preview WorldCat COPAC Laertius, Diogenes (2018), Lives of the Eminent Philosophers, trans. Pamela Mensch, ed. James Miller (New York: Oxford University Press). Google Scholar Google Preview WorldCat COPAC LePore, Ernest (1982), ʻIn Defence of Davidsonʼ, Linguistics and Philosophy, 5, 277–294. Google Scholar WorldCat Malloch, Stephen, and Colwyn Trevarthen (2010), Communicative Musicality: Exploring the Bais of Human Companionship (New York: Oxford University Press). Google Scholar Google Preview WorldCat COPAC Merleau-Ponty, Maurice (1988), ʻIn Praise of Philosophyʼ, in In Praise of Philosophy and Other Essays (Evanston, IL: Northwestern University Press), 3–70. Google Scholar Google Preview WorldCat COPAC Phelps, Edmund S. (2006) “Macroeconomics for a Modern Economy”, Prize Lecture, December 8, 2006, www.nobelprize.org, accessed 14.9.2023. Google Scholar WorldCat p. 411

Phelps, Edmund (2013), Mass Flourishing: How Grassroots Innovation Created Jobs, Challenge, and Change (Princeton, NJ: Princeton University Press). Google Scholar Google Preview WorldCat COPAC Plato (1977), Apology, translated by G.M.A. Grube, in Complete Works, edited by John M. Cooper (Indianapolis, IN: Hackett). Google Scholar Google Preview WorldCat COPAC Plato (1977), Letters, translated by Glenn R. Morrow, in Complete Works, edited by John M. Cooper (Indianapolis, IN: Hackett). Google Scholar Google Preview WorldCat COPAC Sander, Louis (2008), Living Systems, Evolving Consciousness, and the Emerging Person. A Selection of Papers from the Life Work of Louis Sander, edited by Gherardo Amadei and Ilaria Bianchi (New York: The Analytic Press). Google Scholar Google Preview WorldCat COPAC Seligman, Stephen (2017), Relationship in Development: Infancy, Intersubjectivity, and Attachment (London: Routledge).

Google Scholar p. 412

Google Preview

WorldCat

COPAC

Steiner, George (1991), Martin Heidegger (Chicago: University of Chicago Press). Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

18 American Sign Language, Audism, and Power: Boxed Ears and Swiveling Fists  Teresa Blankmeyer Burke https://doi.org/10.1093/oxfordhb/9780192844118.013.53 Published: 22 May 2024

Pages 415–433

Abstract “Language is power” is a well-used catchphrase, yet it is rarely deployed to denote di erences in the power of language modality. Linguists speak of language modalities, meaning how the language is expressed—the modalities include spoken language, written language, and signed language. The default language modality is spoken language, followed by written language. Although signed languages are for the most part recognized by the public as languages today, thanks to the work of signed language linguists and others, such as the late William Stokoe, false beliefs about signing language modality persist. This chapter takes up the topic of language modality expression as an impetus for discrimination, focusing in on the ways in which the interactions and expectations of two language modality di erences, signed and spoken, impact the communication experiences of signing Deaf people in terms of power. In identifying the inequities in signed and spoken language modality interactions through the discussion of several examples, it will identify and analyze three issues (which are related): language modality and direct access to language, language modality and symmetry of communication, and, nally, language modality and partial access to language.

Keywords: audism, signed language, power, language modality, deaf Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction This chapter is a translation of a paper that was originally going to be presented in American Sign Language (ASL). Despite the e orts of Connie Di ee at the University of Memphis and DeafConnect of the Midsouth, who worked hard to provide interpreter support, there was no team of ASL interpreters available that day. There are lots of reasons for this, but one of them is tied to deaf geography. There are cities where the pool of interpreters is deep. These are cities with sizable deaf populations—large urban centers, but also towns where state residential schools for deaf children are located. Sometimes even the size of the city doesn’t matter—I have a professor colleague in New York City who struggles to nd interpreters for her classes and faculty meetings and goes without communication access far more often than you might imagine. Even in Washington, DC, a city where there are more deaf professionals than anywhere else in the world, we frequently have shortages—I get emails from my university (the only bilingual liberal arts university in the world for deaf and hard of hearing people) several times a month announcing periods (days or weeks) when interpreter shortages will occur. Due to the absence of ASL/English interpreters and to my voice privilege as a Deaf person who speaks, the paper was given in spoken English, with some ASL content, which is glossed here in this paper using upper-case English, per the usual convention in ASL linguistics. Contributing to the irony of having one’s power of language choice undercut at a conference on language and power was that the paper was delivered during the 2017 International Week of the Deaf, September 18–24, proclaimed by the World Federation of the Deaf. The theme for 2017 was “Full Inclusion with Sign Language.” When p. 416

writing a paper in English that I will simultaneously translate into

American Sign Language, I make

speci c and careful choices about sentence structure and construction that make it easier to render the 1

translation. I have kept these in place.

2. A Requiem for Magdiel Sanchez On Tuesday, September 19, 2017, Magdiel Sanchez, a deaf man living in Oklahoma City was shot in his own 2

front yard. He was sitting on his front porch when his father drove up to the house. Shortly afterward, police arrived to follow up on a hit-and-run accident his father was allegedly involved in. At least three neighbors shouted at the police o

cers to inform them that Magdiel Sanchez was deaf and could not hear

them—apparently he had a metal pipe in his hand when he got up, though by all accounts he was holding it at his side and not wielding it in a threatening or aggressive way. Both police o

cers took action—one used

a Taser; the other shot to kill. Magdiel’s father sat in his truck and watched police o

cers kill his deaf son 3

despite the multiple shouts from neighbors informing police that Magdiel was deaf.

Deaf people know that an encounter with police can go very badly if they do not have voice privileges or even if they do, since the deaf accent is often viewed by law enforcement as evidence that a person is under the in uence. A deaf person who is making audible noise but not words frequently gets handcu ed. This is the equivalent of putting a gag on a hearing person during an arrest and expecting them to communicate. When I rst read about Magdiel’s Sanchez’s killing (on Deaf twitter), I felt numb. Another deaf man shot by police because the wrong language modality was presumed—auditory, not visual. Again. Magdiel Sanchez Terrence Crutcher Darnell Wicker Randall Waddel

Edward Miller John T. William 4

And so many more.

This is a familiar story—a deaf person, most frequently a deaf man of color, does not respond to shouted commands from police. Police read the deaf person’s lack of response as de ance or resistance, interpret this response as a threat, and take down the deaf person. The (so-called) lucky ones who only get tackled or tasered are handcu ed and taken to jail, where they may be deprived of communication access for hours, if not days. I rst became personally acquainted with deaf people who had experienced these injustices when I was a graduate student in philosophy also working part-time for a deaf social services agency. That experience shapes my work in deaf philosophy—I am an ivory tower academic today, but my roots are in community work and activist advocacy. As we know from the social justice movement Black Lives Matter, outrageous killing of innocent people by p. 417

cops is not a phenomenon that is restricted to deaf people. The

abuse of law enforcement power in killing

Black citizens is a longstanding systemic problem that has urgently needed to be addressed for centuries. The abuse of the power of law enforcement that results in the deaths of innocent citizens has many strands —I con ne my remarks to the case of deaf people because that is the focus of my paper, but I am equally mindful of the lurking threat of potential violence by law enforcement that permeates Black and Brown peoples’ consciousness every day. And of course, some Black and Brown are also deaf people, who are doubly jeopardized in other law enforcement contexts because of their appearance and their (deaf) 5

behaviors.

Returning to my focus on deaf people, when police or other law enforcement o

cials kill a deaf person, this

is typically viewed as a tragedy that could have been avoided. It is framed as a problem with a solution that gets pushed back on us deaf people, not on law enforcement. We are told to learn to speak (use their language, even though it is much more di

cult for a deaf person to

6

learn to speak than for a (sighted) hearing person to learn to sign). We are told to learn to read lips (even 7

though studies show that only 30 to 40 percent of English speech is visible on the mouth). We are told to register with emergency services (yes, let’s develop a registry so that deaf and disabled people are easier to 8

locate—that went well in Nazi Germany). We are told to wear a beta device that will warn police that they are dealing with a deaf person (as if a cop with a drawn gun is going to stop to check his phone or tablet if it 9

beeps). We are told to put our arms up any time we see law enforcement (what about the times when they 10

are shouting at us from behind?).

Yet the refrain of “hands up, don’t shoot” for deaf people comes with the

proviso that for deaf people, “hands up” can be both a survival action and a silencing directive; and for deaf people with no voice privileges, shouting “don’t shoot” is not an option, for to shout without speech can lead to being handcu ed and, literally, being silenced. We are told to put placards in our cars and how to move to show police o

11

cers we are deaf.

We are told to use our time to advocate for better training of law enforcement (in addition to the other accessibility advocacy we already do every day to get our society to follow the Americans with Disabilities 12

Act—a civil rights law that is over thirty years old).

We are told to get involved with law enforcement

training (I’ve done this myself—there’s an irony in being appointed to do court-ordered deaf-awareness training for law enforcement sta

and having to argue for access to signed language interpreters in order to

do this). And we are told to alter our bodies through surgery (cochlear implants have variable e ectiveness, 13

and do not guarantee access to language comprehension).

We are expected to speechread spoken language

while under stress, when it is known that adding this variable reduces our ability to comprehend the 14

message.

The upshot of the message to deaf people who are furious about yet another deaf person killed by

law enforcement is this: This is YOUR problem. Fix it by xing yourselves.

Magdiel Sanchez was killed in his own front yard because of language modality bias and assumptions based on how people should respond to spoken language—even when we are deaf. p. 418

Quoting from an Associated Press story: OKLAHOMA CITY. Oklahoma City police o

cers who opened re on a man in front of his home as

he approached them holding a metal pipe didn’t hear witnesses yelling that he was deaf, a department o

cial said Wednesday.

Magdiel Sanchez, 35, wasn’t obeying the o

cers’ commands before one shot him with a gun and

the other with a Taser on Tuesday night, police Capt. Bo Mathews said at a news conference. He said witnesses were yelling “he can’t hear you” before the o

cers red, but they didn’t hear

15

them.

Let me repeat this. Captain Bo Mathews said that Magdiel Sanchez, a deaf man, wasn’t obeying o shouted commands. But the o

cers’

cers, both hearing men who use spoken language, were given a pass for not

responding to shouts that Magdiel Sanchez was deaf because THEY DIDN’T HEAR THEM . Rest in peace, Magdiel Sanchez.

3. Deaf Terminology and Concepts “Language is power” is a well-used catchphrase, yet it is rarely deployed to denote di erences in the power of language modality. Linguists speak of language modalities, meaning how the language is expressed—the modalities include spoken language, written language, and signed language. (There are also receptive modes of language: reading and listening—whether listening with one’s ears or eyes.) The default language modality is spoken language, followed by written language. While signed languages are for the most part recognized by the public as languages today, thanks to the work of signed language linguists and others, 16

such as the late Dorothy Casterline, false beliefs about signing language modality persist.

This paper takes

up the topic of language modality expression as an impetus for discrimination, focusing in on the ways in which the interactions and expectations of two language modality di erences, signed and spoken, impact communication experiences of signing Deaf people in terms of power. In identifying the inequities in signed and spoken language modality interactions through discussion of several examples, I will identify and analyze three issues (which are related): language modality and direct access to language, language modality and symmetry of communication, and, nally, language modality and partial access to language. Before I launch into my talk, I’d like to introduce some of the terms I will be using. The word deaf has been used in deaf studies in two di erent ways—Deaf with a capital D indicates Deaf people who know and use a signed language as a primary language. But not all deaf people use sign language, and so deaf with a lowercase d is used to indicate the group of deaf people who have bodies that have di erent responses to auditory p. 419

noise than is considered to be species-typical. There is currently a challenge within the

eld of deaf

studies to jettison this distinction given a common misunderstanding of these categories as “signing deaf” 17

and “speaking deaf” and rigidly assigning deaf people to one, and only one, of these categories.

But there

are many deaf people who sign or speak depending on the context—sign only, speak only, sign and speak simultaneously, and so forth. Deaf people are a diverse group of people with a multitude of communication styles using signing, speaking, and written modalities. These modality choices may be made for deaf people when they are children and their parents determine what language modality their child will use (signed or spoken), or they may be political choices made by thoughtful deaf adults who refuse to speak even though they have voice privileges.

Let me note that in terms of variation on audiograms, which is very much a medicalized perception of deafness, 48 million US Americans have a signi cant “hearing loss,” which is about 15 percent of the population. The term “hearing loss” is used in the medical literature, and it is one I prefer to avoid when writing about deaf people, in part because not all of us have experienced our state of being in the world as a 18

loss; additionally, many of us do not view this as a loss but as a di erence or, in some cases, a gain.

In signing deaf communities and spoken language communities another categorical distinction is also frequently used, which sets apart deaf people from “hard of hearing” people. There is a social movement in part of the signing deaf community to resist these categorizations as an imposition by mainstream hearing culture norms based on medicalized auditory status that give preferential status to those who are more like culturally hearing people—for example, those who choose to use voice privileges and amplify their residual hearing. Rejecting these categories and considering all people who have varied responses to auditory signals as ‘deaf’ reframes, uni es, and strengthens the advocacy e orts of deaf people, as opposed to setting up arti cial dichotomies and categories that pit us against each other in our quest for access, often framing the di erences in a way that results in internecine battles over limited resources, instead of ghting for su

cient resources to provide access for all deaf people in multiple language modalities: spoken, signed,

and written. In this paper, I will use the term deaf to refer to all people who have variable access to auditory content. (In other words, I will not be using the phrase ‘deaf and hard of hearing’ but will instead use deaf to refer to this group.) I will also use modi ers when appropriate, such as ‘signing deaf’, ‘sighted deaf’, ‘deafblind’, ‘nonsigning deaf’ and so forth. I will use similar modi ers when discussing people who have access to auditory content: ‘culturally hearing people’, ‘nonsigning hearing people’, ‘signing hearing people’, and so forth. I will also use the acronym ASL to refer to American Sign Language. There are numerous sign languages—I expect that much of what I say will apply to other signed-language-using communities, but my focus is on the community I know best—ASL users. Finally, another term I will use is audism, that A-U-D-I-S-M , which refers to discrimination against deaf people. This includes individual acts of discrimination as well as institutionalized audism of systems, practices, and institutions established to bene t those with culturally hearing privileges, including the use of spoken voice, audition, and spoken p. 420

language(s). Given the similar

sounds of the words audism and autism, I want to be sure that this

distinction is noted. Do note that this is a language modality issue as well as a language di erence, plus also gure out the technical meaning. In ASL these signs are quite di erent and are not easily confused. (Demonstrate in ASL for the audience: AUTISM

19

20

and AUDISM . ) The sign for audism looks like a box around

the ear. I’ve engaged in a little ASL-English wordplay with my title, using an English phrase ‘boxed ears,’ that is also a description of an the ASL word audism. Boxing a person’s ears is an act of dominance, as is audism. Swiveling sts is another description of ASL and can refer to several di erent words with this shape (Demonstrate in ASL:  RADICAL, RESIST, PROTEST, REBEL

21

 VERY STRONG NEGATION ).

4. Hurricane Irma and Signed Language Interpreters In September 2017, two incidents involving signed language interpreters at televised press conferences dealing with the emergency response to Hurricane Irma became news. In the press conference held by Florida governor Rick Scott, nonsigning (mostly culturally hearing) people were trans xed by the facial 22

expressions of the Certi ed Deaf Interpreter (CDI),

Sam Harris, whose interpretation was pitch perfect in

capturing the urgency of the critical lifesaving message. 23

The public response initially disparaged Harris’s interpreting product as “dramatic,” “overdone,” “fake”. Signing deaf people, on the other hand, rolled our eyes and sighed, knowing that we’d once again have to articulate basic grammatical features of ASL only to be corrected by nonsigners that their account of 24

“expressiveness” trumped (ahem) our understanding of the language we use every day. nonsigning philosophers are not immune from this behavior.]

[Aside—

The other interpreted incident involved a smaller press conference in Manatee County, where the person on the screen interpreting, Marshall Greene, was not a signed language interpreter, but a relative of a deaf person who was recruited to “interpret” because he happened to be on site. Now, deaf people are used to a wide range of interpreter quality, including some pretty abysmal interpreting, but this was especially egregious. This was not a local social event, but a press conference providing critical information during a live broadcast dealing with a natural disaster. Deaf people were watching the broadcast to get information about whether to stay in place or leave, where to go if mandated to evacuate, and so forth. The word salad that was part of the “interpretation” included such snippets as “bear monster,” “pizza,” and “pray wait water.” The decision that led to the lack of lifesaving information reaching signing Deaf people was not one based on a lack of resources—highly quali ed certi ed signed language interpreters were on standby. Instead, a variety of factors, including folk beliefs about signed language interpreting (that it is easy, that anyone who communicates with a deaf person using signs is quali ed to interpret, that something is better p. 421

than nothing when

it comes to deaf people) and that it would be easier to use someone who is already in

the building) contributed to this debacle. In each of these examples, nonsigning hearing people made judgments about the quality of the signed language interpreting provided to deaf people. This is the norm for deaf folk, who frequently have very little say about our communications accommodations once a request is made or a need identi ed. And this is true even in the case of very savvy, highly educated deaf professionals. (Aside: I had a micro-stando

just last

night at Reagan National Airport—an airport that has dozens of deaf people traveling through it every day. I “won,” but it wasn’t until I got on the plane that I could breathe easy and no longer worry about retaliation.) In most cases, it is hearing nonsigners who make these decisions about who will provide signed language interpretation, and they often make these judgments based on information that is conveyed to them in spoken or written English by other hearing people. These can include cold calls to develop exclusive contracts to interpret for a particular business or institution or website public relations materials. Not surprisingly, nonsigning hearing people are unable to assess whether the signed language interpreter product is exceptional, unacceptable, or functional. It is not uncommon to learn about interpreting contracts based on congeniality in a culturally hearing context or by attractive promotional materials, rather than relying on the local deaf community assessment of professional skills. In most cases, the people making determinations about what communication access deaf deserve know nothing about the di erence between a CT, a NIC-Master, or an NAD III certi cation, and which of these certi cations assess the skills that best match the communication styles of the person who has requested interpreting services. This is a direct e ect of institutionalized audism, an injustice which rests on a foundation of paternalistic beliefs about the capacities and abilities of deaf people conjoined with assumption that culturally hearing people are superior. Ignorant nonsigning hearing people prioritize their desire to perform acts that allow them to see themselves as “charitable” and “a good person” by helping deaf people, thereby trumping the deaf person’s knowledge about signed language interpreting and their own communication preferences. The two cases of signed language interpreting at Hurricane Irma press conferences illustrate a range of epistemic injustice responses involving signed language modality and its users. On the one hand, a solid ASL interpretation by the native deaf signer and CDI Sam Harris is ridiculed and assessed as “not really interpreting” or objecti ed as a “performance”; this is even as signed language users exulted at the exceptionally clear product, which was not only easily comprehended by native ASL users, but possessed enough iconicity and classi ers (all grammatically correct) that a deaf immigrant with a di erent native sign language, such as Cuban Sign Language, would be able to get the gist of the message, though not the details. Yet the testimony by deaf native users of ASL commending one of their own was repeatedly disregarded in the early social media and internet news discourse, with (presumably) nonsigning hearing people dismissing such authoritative claims from self-identi ed native deaf signers and replacing them p. 422

with

assertions of their own (completely uninformed) judgments about the quality and legitimacy of the

interpreted product. In the case of the lifeguard “pseudo-interpreter” a similar phenomenon occurs to that of Thamsanqa 25

Jantjie, the interpreter at Nelson Mandela’s memorial service.

Both times, deaf people immediately spoke

out on social media about the terrible quality of pseudo-interpreting, but these initial observations were disregarded. In the case of Nelson Mandela’s funeral service, which was not a matter of life and death, a deaf member of the South African parliament (one of the only signing deaf politicians of that stature in the world —we have never had a signing deaf member of Congress in the United States) spoke out on Twitter, but her comments were not attended to until a critical mass of deaf people made so much ruckus on international 26

deaf twitter that hearing nonsigners had to pay attention.

I provide these examples of Magdiel Sanchez, Sam Harris, and the lifeguard pseudo interpreter Marshall Greene as a sampling of the injustices foisted on signing deaf people. It is unusual that three separate instances would be featured in the mainstream media over such a short time span—I have a glimmer of hope that this may be a moment for us when public awareness of the lives of deaf people turns a corner. I now turn to an assessment of three inequities of language and communication that emerge from these examples and illustrate them in a context that is more familiar to this audience—that of the academy and the deaf academic.

5. Having Direct Access to What Is Being Said Deaf academics with voice privileges frequently encounter an ethical dilemma when another deaf person shows up at a conference. Typically, deaf academics with voice privileges (and I include myself in this group) rely on their adeptness at speaking and speechreading (these are separate, but I’ll treat them together for the purposes of this section) to succeed in the academy. (More about that later too!) When I am invited to a philosophy or bioethics event, and it is advertised to the public, and advertised as interpreted, I have mixed feelings about this. On the one hand, I am glad that the event is posted as accessible; on the other hand, I know that this means additional work for me. What do I mean by this? Deaf academics, especially those working in places that are centered around spoken language, rarely have the opportunity to have direct access to academic content. Almost always, our entry into academic discourse is mediated through an interpreter, be it in a graduate school seminar, a conference, or in the classes we teach. In almost all cases, that interpreter is not a member of our academic discipline (deaf academics working in elds such as linguistics, education, deaf studies, or interpretation and translation studies are p. 423

the exception), and so our access to content is mediated, not only by the fact of another

person

interpreting what is spoken to us, but also by the fact of a person who is not a scholar in our eld and is making decisions about what to convey, what to omit due to time constraints, and what to paraphrase. For this reason, my policy (especially now that I am tenured) is to switch my presentation from spoken English to ASL if there is a deaf signer in the audience. This comes with costs. The costs are borne by everyone. I bear the cost of having to prepare my talk in two languages (and the time that takes) and the stress of not knowing until I walk into the room which of the two languages I will use. Our interpreters (for indeed, they are ours, not mine alone) do not know which language I will present in and will have to divide their preparation between two languages. There are also costs borne by the audience members. If I opt to give my talk in ASL to an audience of nonsigning philosophers and a signing deaf academic, the interpreted talk’s focus shifts away from the technical philosophy I would have used in spoken English to one aimed at a more general academic audience. (The deaf academic is most likely not a deaf philosopher—there are very few of us in the world (that is, those with academic credentials in philosophy), and of those who know ASL well enough to follow a philosophy

talk in the language, there are literally just a handful. I know them all, and given the rare opportunities for deaf academics to interact in conference spaces, we would have made prior arrangements in advance to meet up, so their presence at my talk would not be a surprise, and we would have negotiated in advance the language used for the public talk. If a deaf person who is not an academic shows up (I do a lot of work in deaf bioethics advocacy, and for this reason people are sometimes curious about what I do and want to meet me), I have another decision to make. Do I make my talk more accessible in two ways—by providing direct access and by providing the doxa “the well-informed educated layperson knows” that I would provide to an audience of grassroots deaf activists, that would be absolutely tting, but that for an audience of highly educated philosophers and other academics may come across as a bit condescending? But this works the 27

other way as well. If I am doing Deaf philosophy

(and the reasons to do this in ASL are cogent and

compelling), then I must be a kind of interpreter myself, not of the kind that conveys content from one natural language to another, but of cultural doxa—what is widely known and assumed by nonsigning hearing academics is not the same body of knowledge that is widely known and assumed by signing deaf people—academics or no. I mentioned my rank and tenure status as a consideration of the work involved in calculating which language to use. When I was a junior academic, I was keenly aware of stereotypes about deaf people and our 28

purported (in)abilities.

I knew that were I to give an academic presentation in ASL, people would make

assessments about my own abilities based on the lters of language choice added by the ASL-English interpreters. For this reason, I was extremely reluctant to use ASL—philosophers and bioethicists were evaluating me. Having the ability to not only use the right jargon, but also the phrasing that indicates one has been properly socialized and trained in a discipline that attends to hierarchy (far more than I believe is good for it, but that’s another story) allows people to feel more comfortable about approaching a deaf p. 424

philosopher who not

only uses her voice privileges, but who also uses phrases that ag insider status

(e.g., “I have a view…,” “I have a worry,” “ceteris paribus,” “veil of ignorance,” “akrasia,” “hand or nger?”). And those direct, unmediated philosopher-to-philosopher conversations go some distance in providing future opportunities down that road. Given that I had voice privileges, and this was already known about me, I did not believe that I had the luxury of choosing to go “voice o ” in academic venues and not risk harm to my career in the process. Deaf academics working in spoken-English academia must determine how they will address languagepower imbalances in the academic infrastructure. This includes choice of language (ASL or English for North American anglophones—a word that itself privileges the modality of spoken language through speech and hearing), language symmetry (more about this in the section 5), and how to adjudicate the challenges of constant partial access to language. Partial access to language shows up in interpreted interactions due to the interpreter not having the deep knowledge of philosophical concepts and the absence of the depth and breadth that comes from reading hundreds of philosophical texts. It shows up in CART (Communication Access Realtime Translation) captioning human keystroke error due to the phonetic nature of the stenographic keying system. And it shows up in speechreading errors based on mistaken phoneme and 29

morpheme guesswork.

30

Other constraints include the ability to choose one’s interpreter

and interpreter

availability, including the challenge for Deaf academics of color working with White female interpreters, who may not match or re ect their own identities, which can result in a lack of experiential knowledge of 31

the power imbalances in the academy faced by Deaf academics of color.

6. Symmetry It is possible to communicate across language modalities, but there is an inherent danger in making this choice for deaf people with voice privilege. As such a person, I can choose to speak to a nonsigning hearing person, telling them I am deaf, and o er a pad of paper for them to write their answer. Most people reject this request, deploying their audist intuition, and speak to me instead, telling me that I should lipread them. (This is a common knowledge that signing and nonsigning deaf people all know—it isn’t just my experience!) The asymmetry of modality—that I speak to you, but ask you to write to me, includes an unspoken request that you do more work by writing. It is slower and revealing. Some people are ashamed of their illegible handwriting, or their weak literacy skills, or their poor spelling or bad grammar, or any number of things and may opt to continue to speak in order to avoid a situation where these characteristics are revealed. And yet, I suspect there is more to it. Here’s why: when I approach a nonsigning hearing person with pad of paper in hand and write my request to them, they immediately accede. One explanation for this might be rooted in a symmetry of e ort—we are now both perceived as equally contributing our p. 425

labor to the task of written dialogue

communication. Using the frame of audism, a person of lower social

standing (that is, a deaf person) is not asking the nonsigning hearing person to do more work than she is, but is asking that they contribute equally to the labor of communication. I’ve addressed the bene ts to me regarding the choice of speaking over signing in professional settings with a majority composition of nonsigning hearing people (and alluded to the discomfort that this choice raises), but I haven’t directly addressed the bene ts to you. All things being equal, for the deaf person, speaking and speechreading involve more work than for hearing people. This applies even if the deaf person is operating in her L1 ( rst acquired) language. When one does not hear all of the speech sounds, but knows that they must be articulated, there is additional cognitive load in remembering that (to use my own example) the fricatives must always be attended to, and especially where inattention would possibly contribute to ambiguity, such as with the addition of a ‘s’ sound to make a claim plural (possibly universal) or singular (particular)—important distinctions for the philosopher. There is the additional cognitive load of remembering how to pronounce words that one has never really ‘heard’ except in one’s head—in my case, these tend to be words I’ve encountered later in life, and those tend to be proper names or terms of art in my elds—for example, “Berkeley” ( ngerspell in ASL as BARK-LEE ), “akrasia” ( ngerspell in ASL as AK-RUH-SEE-AH ), and “ceteris paribus” ( ngerspell in ASL as KET-TER-IS PAIR-EH-BUS ).

The work I put into speaking is miniscule compared to the work I put into speechreading. The literature on the language acquisition of deaf individuals tends to focus on two modalities—spoken language and signed language. Yet many deaf academics often o er a narrative that runs like this—learning spoken language, then building on spoken language through the experience of reading and writing, and ultimately grounding one’s sense of language competence through the one modality that doesn’t involve any guesswork—written language. Consider one folk explanation for this: people who excel in the academy have a number of skills, but one skill is the ability to take in considerable amounts of information in written form and to engage with it in written form. When a signing deaf person uses a signed language as L1, the lack of a corresponding written language means that the deaf person is probably working in his L2 when he writes. When an oral deaf person is using a spoken language as L1, that language has a written form, so the deaf person is working in the dominant mainstream language of academic discourse. For many of us, the language modality that is our L1 (so to speak) is written language. The inequity of language modality symmetry also includes such communication accommodations as CART captioning—this written language modality is asymmetrical and unidirectional. It gives me your words in my L1 of written language, but it is a communication accommodation that is only available to deaf

academics with voice privilege. Unless of course, you are the kind of deaf academic who doesn’t mind being silenced. Or, you are the exceptional person who can disrupt the conventions of auditory academia su

ciently that people will pay attention to you when you write your question of comment down during the

Q&A, and do not incur subtle social penalties for the additional time this takes in a venue where numerous p. 426

activities are tightly crammed

into a schedule. By using signed language interpreters, a deaf academic

without voice privileges acquires a voice, even if it is not the voice she believes best represents her knowledge and disciplinary training.

7. Partiality The conditions that make it more likely one will become a successful deaf academic are conditions that also make it hazardous. I have written previously about full and partial access to language. What do I mean by partial access to language? Consider the auditory metaphor of the dropped or distorted cell phone signal. The idea is that a continuous stream of information is broadcast, but the receiver is only able to acquire part of that signal, and the content of what signal is dropped or distorted is random. For deaf people, the lived experience of partial access to communication isn’t occasional, but a dominant part of our communication landscape in all live face to face interaction. Signed languages, unlike spoken languages, di er in that the person receiving the signal must deliberately focus her gaze on the speaker. A icked glance at a person entering the room may result in losing a crucial piece of information. Fortunately, ASL uses repetition and redundancy to compensate for this, and also cultural norms around asking for repetition do not carry negative valence. Visual language modalities require the use of one’s eyes, which have muscles that tire, the muscles of the ears do not tire in this way. In conversational settings, full access to language is the norm for native or uent ASL users, provided that social conventions of requests for repetition and clari cation occur. Sometimes these do not. Non-native signers, especially deaf non-native signers, have been socialized through shaming audist hearing cultural practices that asking for repetition or clari cation beyond an occasional request is annoying or rude and carry over this social habit into the signing deaf community. Consequently, ‘blu

ng’ the practice of

pretending that you have understood when you have not, is a learned behavior for many deaf people with voice privileges. Concomitantly, deaf people also adopt the communication strategy of withdrawing from conversation in spoken language modality settings with partial access, since the (audist) social conventions of spoken English users also involves ridiculing, shaming, avoiding, or ignoring people who make remarks that are o point. Shooting, whether with a gun or shooting o

one’s mouth to make a pithy remark at the expense of

another, is a way that audism reinforces the deeply entrenched assumption and expectation that full access to a language modality is a given. Partial access to language modality is not named or included as a form of diverse experience, despite its pervasiveness within the deaf community. Until the phenomenology of partial access to language and language modalities is acknowledged and incorporated into the array of possible explanations for seemingly ‘aberrant’ behavior, deaf people will continue to be harmed, and sometimes killed for this inequity. Turning back to the academy. Partial access to interpreted academic events is the norm and deaf people p. 427

accept it—in part, I suspect, because we’ve had to ght so hard

to just get access. This is not intended as

a slam against signed language interpreters, but should be taken as a report on the state of the eld, including the lack of resources available to serve (among others) the growing class of deaf professionals. Historically, signed language interpreters provided access in settings where the professionals were hearing and the consumers were deaf. Hearing professionals, cognizant of stereotypes about deaf people being less informed and having limited cognitive abilities, would tailor their conversation to a layperson’s vernacular,

which the interpreter would then interpret. This is not the case for interpreted academic events. Here, interpreters are working at the height of their abilities to convey content in a discipline that they typically are unfamiliar with, and doing it at a speed that is much faster than conversational interpreting. (People read papers quickly—it took me a long time to gure out that this wasn’t necessarily a preferred norm of the academy but was often due to nerves or trying to t a long paper into a shorter time space.) The interpreters’ lack of knowledge about the state of play in a current academic debate, including the names of key article or book titles or the names of those engaging in the scholarly discussion, impacts their ability to interpret this content. The cognitive demand on signed language interpreters is enormous (studies of the brain activity of signed language interpreting show the whole brain lighted up), and unlike listening to a conversation, where the utterance of an unfamiliar name is often picked up, it is much harder to do that when you are simultaneously interpreting what is heard into not only another language, but another language with a di erent modality. My interpreters are instructed to mouth the phonemes as best they can —because I am an excellent speechreader and because I do know the literature and the players, I can usually augment the partial information I receive. For deaf academics whose L1 is English, signed language and spoken language communication is always partial, but in di erent ways. The non-native user of ASL contends with both a non-native language and non-native language modality, each of which impacts the native language user’s intuition. This intuitive capacity for one’s native language is instrumental to full access—even after years of language use, the nonnative user will still miss things on occasion, whether it is a new word or an unfamiliar idiom. The deaf native English speaker who is also a speechreader misses spoken words—remember that only 30 to 40% of spoken English is visible to speechreaders—and must compile meaning from this partial information. Sometimes this is very di

cult when a critical word is missed—for example, the speechreader may not

realize until she is well into the conversation that she has missed the word “not” and must then go back and reconstruct her understanding as she is also trying to speechread and follow the current stream of words being spoken.

8. Conclusion I have identi ed and brie y analyzed just three inequities of language modality choice: the inequity of p. 428

indirect access to what is said (via an interpreter or other

intermediary), the inequity of asymmetrical

language modality interactions, and the inequity of partial access to language across language modalities. There is much more work to be done in this area of language modality inequity—I have barely scratched the surface. Returning to the killing of Magdiel Sanchez: These language modality inequities were all present. 1.  Magdiel Sanchez had no direct access to the language or communication modality being used. The police used spoken language in their attempts to engage with him. It isn’t clear to me whether the reports that Magdiel Snachez was QUOTE “nonverbal” END QUOTE refer to his lack of voice privileges or whether he did not possess a language, including a signed language. This is yet another instantiation of the pervasiveness of audism in reporting about deaf people, here lumping together all manual movement as communication without distinguishing whether it was language or something else. 2.  Police communication with Magdiel Sanchez also involved an asymmetrical inequity. Only spoken language was used, and no other language modalities were o ered—not even a one word sign that could have saved Magdiel’s life, the sign for DEAF . This one word should be a standard part of all law enforcement training. And right now, I am going to ask all of you to raise your index nger, point it to your ear, and your mouth, and shape your facial expression into a questioning countenance, tilting

your chin up and slightly raising your eyebrows as you mouth the word deaf (DEMONSTRATE. REPEAT. REPEAT AGAIN) . Knowing this one sign may someday save a deaf person’s life.

3. I guarantee you that the inequity of partial access to language was a factor in the killing of Magdiel Sanchez—whether it was trying to puzzle out the words shouted by law enforcement or by his neighbors. Sometimes people cannot decipher the partial content—important pieces are missing or the stress of the situation takes away from cognitive processing, or time has run out. Magdiel Sanchez’s time ran out in part because of language modality inequities. I o er this paper as a small attempt to salvage something helpful from a horri c incident.

Acknowledgments I extend much gratitude to Luvell Anderson for the invitation to think and write on this topic at the 36th Spindel Conference on Language, Power and Society, and to the incredible slate of speakers at the 2017 Spindel Conference for their helpful feedback and discussion. This paper is only slightly changed from the original version presented in September 2017; social events occurring after this time (including the global racial reckoning following the murder of George Floyd in 2020) have not been addressed.

p. 429

Notes 1.

For more information about the process of the production of scholarly work by bilingual Deaf Academics working in American Sign Language and English, see Teresa Blankmeyer Burke, ʻA Philosophical Analysis of ASL/English Publishing in Crip Authorshipʼ, edited by Mara Mills and Rebecca Sanchez 2023. New York, New York: NYU Press, 2023).

2.

As I was preparing the first version of this paper for the 2017 Spindel Conference on Language and Power at the University of Memphis, another deaf man, Magdiel Sanchez, was killed by a police o icer during the week of the conference.

3.

There is some confusion in the news reports as to whether Sanchez was also ʻdevelopmentally disabled,ʼ ʻnonverbalʼ (an ambiguous term that can refer to non-speaking by choice or lacking full capacity for expressive language), or ʻlanguage deprivedʼ—an egregious harm that can happen because of bad luck of being born in a place where deaf education is not available. In such cases, even a deaf person who is fully capable of acquiring language may be unable to do so if the opportunities for learning language arenʼt presented early enough.

4.

For a current list of police violence against deaf individuals, please see the “HEARD Log of Police Violence & Discrimination Against Deaf People,” last updated June 2020, https://docs.google.com/spreadsheets/d/1HZ6YLtXzRNiEsu2RCfEUb1WmsCwM4Pn89ikpAwE4b-Q/edit#gid=1519942027.

5.

Teresa Blankmeyer Burke. “Flying While BAD (Brown and Deaf),” Duke City Fix, November 28, 2015. http://web.archive.org/web/20151128030021/http://www.dukecityfix.com/profiles/blogs/flying-while-bad-brown-anddeaf

.

6.

Diane Lillo-Martin and Jonathan Henner, “Acquisition of Sign Languages,” Annual Review of Linguistics 7 (2021): 395–419. https://doi.org/10.1146/annurev-linguistics-043020-092357; and Kathryn Davidson, Diane Lillo-Martin, and Deborah Chen Pichler, “Spoken English Language Development among Native Signing Children with Cochlear Implants,” Journal of Deaf Studies and Deaf Education 19, no. 2 (April 2014): 238–250, https://doi.org/10.1093/deafed/ent045.

7.

Nicholas A. Altieri, David B. Pisoni, and James Townsend, “Some Normative Data on Lip-Reading Skills,” Journal of the Acoustical Society of America 130, no. 1 (2011): 1–4, https://doi.org/10.1121/1.3593376.

8.

Horst Biesold, Crying Hands: Eugenics and Deaf People in Nazi Germany. 1999 (Washington, DC: Gallaudet University Press, 1999); and Eva F. Kittay, “Deadly Medicine: Project T4, Mental Disability, and Racism,” Res Philosophica 93, no. 4 (October 2016): 715–741, https://doi.org/10.11612/resphil.1568.

p. 430

9.

See Maria Polletta, “Avoiding ʻA Deadly Situationʼ: Phoenix Police Learn How to Interact with Deaf Residents,” AZCentral, April 17. 2018, https://www.azcentral.com/story/news/local/phoenix/2018/04/17/deadly-situation-phoenix-police-gettraining/519437002/ and Alejandra Molina, “Measure Aims to Alert Police That a Driver Is Deaf or Hard of Hearing,” The SUN, June 30, 2018, https://www.sbsun.com/2017/06/03/measure-aims-to-alert-police-that-a-driver-is-deaf-or-hard-ofhearing/.

10.

Marion Hersh, James Ohene-Djan, and Saduf Naqvi, “Investigating Road Safety Issues and Deaf People in the United Kingdom: An Empirical Study and Recommendations for Good Practice,” Journal of Prevention & Intervention in the Community 38, no. 4 (2010): 290–305, https://www.tandfonline.com/doi/abs/10.1080/10852352.2010.509021.

11.

Minnesota Commission of the Deaf, DeafBlind, and Hard of Hearing, “Communication Card for People Who Are Deaf/Hard of Hearing & Police O icers,” May 3, 2018, https://mn.gov/deaf-commission/news/?id=1063-338186

.

12.

Amiel Fields-Meyer, “When Police O icers Donʼt Know about the ADA,” The Atlantic, September 26, 2017, https://www.theatlantic.com/politics/archive/2017/09/the-steadily-problematic-interactions-between-deaf-americansand-police/541083/.

13.

Isabelle Boisvert, Mariana Reis, Agnes Au, Robert Cowan, and Richard Dowell, “Cochlear Implantation Outcomes in Adults: A Scoping Review,” PloS One 15, no. 5 (2020): Article e0232421, https://doi.org/10.1371/journal.pone.0232421.

14.

Julia E. Feld and Mitchell Sommers, “Lipreading, Processing Speed, and Working Memory in Younger and Older Adults,” Journal of Speech, Language, and Hearing Research 52, no. 6 (2009): 1555–1565. https://doi.org/10.1044/10924388(2009/08-0137).

15.

Ken Miller, “Witnesses Yell ʻHe Canʼt Hear Youʼ as Cop Shoots Deaf Man,” AP News, September 20, 2017, https://apnews.com/article/ok-state-wire-north-america-us-news-ap-top-news-oklahoma-city1ebeb1f11d694090bcc0d2e93df7f086.

16.

Dorothy Casterline, who passed away this year, was a deaf woman of color who is only just now getting recognition for the work she did. See her NYT obituary for an interesting reference.

17.

I have not fully worked out my position on this issue, which is reflected by my decision to split the di erence. From this point in the chapter I use the lower case deaf for all deaf people, whether they use a signed language or spoken language. I will specifically note the language modality when it is needed. Prior to this point, I follow the conventional practice of Deaf used to represent a signing individual with auditory deafness and deaf representing a deaf or hard of hearing person who does not sign. In cases of identity claims I follow the stated preference of the person. For proper nouns or expressions, I follow the Deaf/deaf distinction as well, i.e. ʻDeaf Twitterʼ.

18.

Teresa Blankmeyer Burke, “Armchairs and Stares: On the Privation of Deafness,” in Deaf Gain: Raising the Stakes for Human Diversity, ed. H-Dirksen L. Bauman and Joseph J. Murray, 3–22. Minneapolis: University of Minnesota Press, 2014. http://www.jstor.org/stable/10.5749/j.ctt9qh3m7.6.

19.

“AUTISM,” Signing Savvy, https://www.signingsavvy.com/sign/AUTISM/9337/2, accessed October 28, 2023.

20.

“Signs for AUDISM,” Handspeak, https://www.handspeak.com/word/2489/, accessed October 28, 2023.

21.

“REBEL/strike/protest/defiant,” YouTube, https://www.youtube.com/watch?v=NzhYpb6dgfs, accessed October 28, 2023.

22.

For more information about deaf ASL/English interpreters, please see National Deaf Center on Postsecondary Outcomes, ʻDeaf Interpretersʼ, September 16, 2023, https://nationaldeafcenter.org/resources/accessaccommodations/accomodations101/interpreting/deaf-interpreters/.

23.

Lee Jackson, “A Deaf Perspective on Hurricane Irma Interpreters,” MT & Associates Sign Language Practice (website), October 16, 2017, https://www.mtapractice.com/2017/10/16/deaf-perspective-hurricane-irma-interpreters/

.

24.

Akira Okrent, “Why Great Sign Language Interpreters Are So Animated,” The Atlantic, November 2, 2012, https://www.theatlantic.com/health/archive/2012/11/why-great-sign-language-interpreters-are-so-animated/264459/.

25.

Teresa Blankmeyer Burke, “Charlatan Interpreter Sparks Outrage at Mandelaʼs Service,” https://feministphilosophers.wordpress.com/2013/12/10/charlatan-interpreter-sparks-outrage-at-mandelas-service/;

and Burke, “Defer to the Deaf Person: Interpreters and Quality Control,” Feminist Philosophers (blog), December 11, 2013, https://feministphilosophers.wordpress.com/2013/12/11/defer-to-the-deaf-person-interpreters-and-qualitycontrol/.

p. 431

26.

Alva Noe, “On Failing to Spot Gibberish,” 13/7 Cosmos and Culture, National Public Radio, December 21, 2013, https://www.npr.org/sections/13.7/2013/12/20/255872895/on-failing-to-spot-gibberish

.

27.

For more on Deaf philosophy, please see Teresa Blankmeyer Burke, “Doing Philosophy in American Sign Language: Creating a Philosophical Lexicon,” Philosop-Her (blog), April 15, 2015, https://politicalphilosopher.net/2015/04/10/featured-philosop-her-teresa-blankmeyer-burke/; Burke, “Uncovering Deaf Women Philosophers,” The Blog of the American Philosophical Association, December 2, 2020, https://blog.apaonline.org/2020/12/02/uncovering-deaf-women-philosophers/; and the ASLCore Philosophical lexicon, https://aslcore.org/philosophy/.

28.

Teresa Blankmeyer Burke, ”Thereʼs a Deaf Student in Your Philosophy Class—Now What?,” Teaching Philosophy 30, no. 4 (2007): 421–442.

29.

For example, see de re/sunbeam example in the post by Teresa Blankmeyer Burke at Disabled Philosophers, September 13, 2011, https://disabledphilosophers.wordpress.com/2011/09/13/teresa-blankmeyer-burke/.

30.

Teresa Blankmeyer Burke. “Choosing Accommodations: Signed Language Interpreting and the Absence of Choice,” Kennedy Institute of Ethics Journal 27, no. 2 (2017), doi:10.1353/ken.2017.0018.

31.

Lynn Hou and Rezenet Moges, “ʻSORRY HARD UNDERSTAND STRONG ACCENT !ʼ: Racial Dynamics of Deaf Scholars of Color Working with White Female Interpreters,” in Thinking with an Accent, edited by Pooja Rangan, Akshya Saxena, Ragini Tharoor Srinivasan, and Pavitra Sundar (Berkeley: University of California Press, 2023), https://doi.org/10.1525/luminos.148.i.

Bibliography Altieri, Nicholas A., David B. Pisoni, and James Townsend. “Some Normative Data on Lip-reading Skills.” Journal of the Acoustical Society of America 130, no 1 (2011): 1–4. https://doi.org/10.1121/1.3593376. Google Scholar WorldCat ASLCore Philosophy Lexicon. https://aslcore.org/philosophy/. WorldCat Biesold, Horst. Crying Hands: Eugenics and Deaf People in Nazi Germany. Washington, DC: Gallaudet University Press, 1999. Google Scholar Google Preview WorldCat COPAC Boisvert, Isabelle, Mariana Reis, Agnes Au, Robert Cowan, and Richard Dowell. “Cochlear Implantation Outcomes in Adults: A Scoping Review.” PloS One 15, no. 5 (2020): Article e0232421. https://doi.org/10.1371/journal.pone.0232421. Google Scholar WorldCat Burke, Teresa Blankmeyer. “A Philosophical Analysis of ASL/English Bilingual Publishing.” In Crip Authorship: Disability as Method, edited by Mara Mills and Rebecca Sanchez, 259–273. New York: NYU Press, 2023. Google Scholar Google Preview WorldCat COPAC Burke, Teresa Blankmeyer. “Armchairs and Stares: On the Privation of Deafness.” In Deaf Gain: Raising the Stakes for Human Diversity, edited by H-Dirksen L. Bauman and Joseph J. Murray, 3–22. Minneapolis: University of Minnesota Press, 2014. http://www.jstor.org/stable/10.5749/j.ctt9qh3m7.6. Google Scholar Google Preview WorldCat COPAC Burke, Teresa Blankmeyer. “Charlatan Interpreter Sparks Outrage at Mandelaʼs Service.” Feminist Philosophers (blog), December 10, 2013. https://feministphilosophers.wordpress.com/2013/12/10/charlatan-interpreter-sparks-outrage-at-mandelas-service/. WorldCat p. 432

Burke, Teresa Blankmeyer. “Choosing Accommodations: Signed Language Interpreting and the Absence of Choice.” Kennedy Institute of Ethics Journal 27, no. 2 (2017): 267–299. doi: 10.1353/ken.2017.0018. Google Scholar WorldCat Burke, Teresa Blankmeyer. “Defer to the Deaf Person: Interpreters and Quality Control.” Feminist Philosophers (blog), December 1, 2013. https://feministphilosophers.wordpress.com/2013/12/11/defer-to-the-deaf-person-interpreters-and-quality-control/. WorldCat Burke, Teresa Blankmeyer. “Doing Philosophy in American Sign Language: Creating a Philosophical Lexicon.” Philosop-Her (blog), April 10, 2015. https://politicalphilosopher.net/2015/04/10/featured-philosop-her-teresa-blankmeyer-burke/. WorldCat Burke, Teresa Blankmeyer. “Thereʼs a Deaf Student in Your Philosophy Class—Now What?” Teaching Philosophy 30, no. 4 (2007): 421–442. Google Scholar WorldCat Burke, Teresa Blankmeyer. “Uncovering Deaf Women Philosophers.” The Blog of the American Philosophical Association, December 2, 2020. https://blog.apaonline.org/2020/12/02/uncovering-deaf-women-philosophers/. WorldCat Davidson, Katherine, Diane Lillo-Martin, and Deborah Chen Pichler. “Spoken English Language Development among Native Signing Children with Cochlear Implants.” Journal of Deaf Studies and Deaf Education 19, no. 2 (April 2014): 238–250. https://doi.org/10.1093/deafed/ent045. WorldCat

Feld, Julia E., and Mitchell Sommers. “Lipreading, Processing Speed, and Working Memory in Younger and Older Adults.” Journal of Speech, Language, and Hearing Research 52, no. 6 (2009): 1555–1565. https://doi.org/10.1044/1092-4388(2009/08-0137). Google Scholar WorldCat Fields-Meyer, Amiel. “When Police O icers Donʼt Know about the ADA.” The Atlantic, September 26, 2017. https://www.theatlantic.com/politics/archive/2017/09/the-steadily-problematic-interactions-between-deaf-americans-andpolice/541083/. HEARD. “HEARD Log of Police Violence & Discrimination against Deaf People.” https://docs.google.com/spreadsheets/d/1HZ6YLtXzRNiEsu2RCfEUb1WmsCwM4Pn89ikpAwE4b-Q/edit#gid=1519942027. WorldCat Marion Hersh, James Ohene-Djan, and Saduf Naqvi. “Investigating Road Safety Issues and Deaf People in the United Kingdom: An Empirical Study and Recommendations for Good Practice.” Journal of Prevention & Intervention in the Community 38, no. 4 (2010): 290–305. https://www.tandfonline.com/doi/abs/10.1080/10852352.2010.509021. Google Scholar WorldCat Hou, Lynn, and Rezenet Moges. “ʻSORRY HARD UNDERSTAND STRONG ACCENT! ʼ: Racial Dynamics of Deaf Scholars of Color Working with White Female Interpreters.” In Thinking with an Accent, edited by Pooja Rangan, Akshya Saxena, Ragini Tharoor Srinivasan, and Pavitra Sundar. Berkeley: University of California Press, 2023. https://doi.org/10.1525/luminos.148.i. Google Scholar Google Preview WorldCat COPAC Jackson, Lee. “A Deaf Perspective on Hurricane Irma Interpreters.” MT & Associates Sign Language Practice (website), October 16, 2017. https://www.mtapractice.com/2017/10/16/deaf-perspective-hurricane-irma-interpreters/. WorldCat Kittay, Eva F. “Deadly Medicine: Project T4, Mental Disability, and Racism.” Res Philosophica 93, no. 4 (October 2016): 715–741. https://doi.org/10.11612/resphil.1568. WorldCat Lillo-Martin, Diane, and Jonathan Henner. “Acquisition of Sign Languages.” Annual Review of Linguistics 7 (2021): 395–419. https://doi.org/10.1146/annurev-linguistics-043020-092357. Google Scholar WorldCat Miller, Ken. “Witnesses Yell ʻHe Canʼt Hear Youʼ as Cop Shoots Deaf Man.” AP News, September 20, 2017. https://apnews.com/article/ok-state-wire-north-america-us-news-ap-top-news-oklahoma-city1ebeb1f11d694090bcc0d2e93df7f086. WorldCat p. 433

Minnesota Commission of the Deaf, DeafBlind, and Hard of Hearing. “Communication Card for People Who Are Deaf/Hard of Hearing & Police O icers,” May 3, 2018. https://mn.gov/deaf-commission/news/?id=1063-338186. WorldCat Molina, Alejandra. “Measure Aims to Alert Police That a Driver Is Deaf or Hard of Hearing.” The SUN, June 30, 2018. https://www.sbsun.com/2017/06/03/measure-aims-to-alert-police-that-a-driver-is-deaf-or-hard-of-hearing/. WorldCat Noe, Alva. “On Failing to Spot Gibberish.” 13/7 Cosmos and Culture, National Public Radio, December 21, 2013. https://www.npr.org/sections/13.7/2013/12/20/255872895/on-failing-to-spot-gibberish. WorldCat Okrent, Akira. “Why Great Sign Language Interpreters Are So Animated.” The Atlantic, November 2, 2012. https://www.theatlantic.com/health/archive/2012/11/why-great-sign-language-interpreters-are-so-animated/264459/. Polletta, Maria. “Avoiding ʻA Deadly Situationʼ: Phoenix Police Learn How to Interact with Deaf Residents.” AZCentral, April 17,

2018. https://www.azcentral.com/story/news/local/phoenix/2018/04/17/deadly-situation-phoenix-police-gettraining/519437002/. Risen, Clay. “Dorothy Casterline, Who Codified American Sign Language, Dies at 95.” New York Times, August 16, 2023. https://www.nytimes.com/2023/08/16/us/dorothy-casterline-dead.html

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

CHAPTER

p. 434

19 Resistance and Reclamation: Notorious Thugs  Luvell Anderson https://doi.org/10.1093/oxfordhb/9780192844118.013.19 Published: 22 May 2024

Pages 434–450

Abstract For some, the idea that words like ‘thug,’ ‘terrorist,’ and ‘urban’ carry racial connotations might come as a shock. They contend these words simply refer to vicious criminals, violent dissidents, and geographic regions. However, this chapter argues that these kinds of terms do carry racial overtones and are often used to justify coercive treatment of those who are so labeled. It then presents two broad strategies for addressing the unjust negative valuations based on being classi ed by such expressions: refusal and resistance. The chapter will demonstrate these strategies by focusing on the term ‘thug’.

Keywords: thug, appropriation, resistance, refusal, methodology, racialized term, slur, code word, dogwhistle Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction “It goes without saying … that language is … a political instrument, means, and proof of power. It is the most vivid and crucial key to identity: It reveals the private identity, and connects one with, or divorces one 1

from, the larger, public or communal identity.” James Baldwin makes this powerful insight in his essay “If Black English Isn’t a Language, Then Tell Me, What Is.” It captures something important about how language is intertwined with social and political life. For some, the idea that ‘thug’ carries racial connotations nowadays will come as a shock. They contend the word is just a way of referring to vicious criminals, illustrated by its application against individuals of various racial backgrounds. Judy Muller 2

(2009) expresses this point of view in her Hu ngton Post article “A Thug by Any Other Name.” Muller describes a discussion during a dinner party in which one member of the party, a young African American woman, objected to Whites using ‘thug’ to refer to the late Zimbabwe president Robert Mugabe. She insisted that it was racist. Muller, who is White, disagrees:

This is one term that should be free of PC shackles. To rob the language of this excellent noun is tantamount to linguistic thuggery. This is a time-honored term that should remain available for identifying all those who qualify as “cruel or vicious ru

ans.”

Muller concedes that she may have cultural blocks that keep her from understanding her dinner companion’s perspective. She identi es those same blocks when it came to her ignorance of the di ering p. 435

perspectives of the criminal-justice system between Blacks

and Whites. Muller also expressed

bewilderment at the use of ‘thug’ in the mouths of rappers. She writes, “But I was still confused; if rappers 3

chose to identify as “thugs,” then why was it racist to use the term in a broader context?”

Unlike brazenly explicit slurs, terms like ‘thug,’ ‘inner city,’ ‘urban,’ and ‘terrorist’ carry racial overtones, even if in a subtle way. Following Patrick O’Donnell, I will regard these kinds of expressions as racialized terms, terms that make race-speci c interpretive options salient. These terms subtly complement their more explicit sibling, the slur. Both terms function to identify and exclude their targets as persons for whom we should express full moral concern. I will focus on ‘thug’ with the hope that broader lessons can be drawn from thinking about the role and function of such terms and responses to them. I argue that one response to this term has been an attempt to re-appropriate it to undermine its oppressive e ects. Re-appropriations in this mode constitute a form of resistance. I then show that Muller’s confusion over rappers’ use of the expression is due to a failure to recognize that a di erent speech act is being performed.

2. A Brief Note on Method In his 1955 William James Lectures–turned-book How to Do Things with Words, John Langshaw Austin attempted to redirect philosophers of language to attend to the panoply of language uses. Austin considers speech beyond mere assertion as well as the role social institutions play in felicitous speech acts. I think Austin was right to push in this direction; however, I don’t think he pushed it far enough. Ultimately, Austin fails to interrogate the signi cance the social plays in linguistic interaction robustly. Reckoning with the power of terms like ‘thug,’ understanding what makes them so e ective requires more than attention to their linguistic properties. Several theories of derogatory language attempt to explain its nature linguistically, focusing on content, whether directly or indirectly expressed, implicated, or presupposed. Some approach such language with the speaker’s attitudes primarily in mind. However, this kind of attened approach runs into problems when the complexity that was initially bracketed for simplifying reasons gets reintroduced. A theory can only be powerful if it has application for the things it purports to be about. If the messiness of lived experience seriously weakens its application, this goes against the view’s explanatory power. What I nd troubling about many approaches to derogatory speech by philosophers of language is a misplaced emphasis on linguistic properties coupled with the under-theorization of the social conditions in which these speech acts occur. This leads to an error of mislocation. I concur with Pierre Bourdieu: By trying to understand the power of linguistic manifestations linguistically, by looking in language for the principle underlying the logic and e ectiveness of the language of institution, one 4

forgets that authority comes to language from outside. p. 436

In my view, a signi cant part of the story is the “outside” Bourdieu references—namely, social conditions. It is not my aim to defend this here, but it is my contention that much contemporary philosophy of language —even that bit that styles itself as the social and political philosophy of language—pays insu

cient

attention to the social. In short, the social philosophy part of the enterprise needs to be more fully integrated into our theorizations. Plenty of people acknowledge the importance of the social, but it rarely shows up in any robust way in the work. One cost of this neglect is the construction of views that misdescribe the way derogatory language functions and, in turn, lead to misinformed strategies for addressing it. For instance, theorizing slurs on a model that posits them as primarily means for expressing ideas or attitudes might mean ignoring other e ects that are not addressed with counter ideas and attitudes. While I do not o er a worked-out methodology for integrating social philosophy and social theory with philosophy of language here, I want to note the importance of considering this more fully—that is, if you are interested in providing fruitful analyses of the relevant phenomena. Philosophers of language interested in social philosophy of language must attend to these issues with greater care, learning about and drawing on already existing work. I attempt to make a start on that in what follows.

3. Racialized Terms Racially demeaning someone can take direct, indirect, subtle, or vivid forms. Readers will be familiar with slurs as paradigm instances of derogatory speech. In case you are tempted to analyze terms like ‘thug,’ ‘terrorist,’ or ‘gang member’ as slurs, let me spend a few moments dispelling that notion. Slurs are o ensive expressions that target race, nationality, religion, gender, sexual orientation, politics, immigrant status, occupation, geographic region, and other demographics. Many people de ne slurs as a conventional way of derogating targets and typically contrast them with a neutral counterpart, purportedly co-referring 5

expressions that lack the derogatory aspect of slurs (e.g., Hornsby, 2001). ‘Thug,’ ‘terrorist,’ and similar terms do share a

nities with slurs. For one, they admit of race-speci c uses.

Verena Erlenbusch-Anderson, for instance, explores how the term ‘terrorist’ gets applied di erently in the 6

media and politics along racial and ethnic lines. Secondly, these terms are clearly pejorative, meant to express something negative about their targets. Despite these similarities, I think that ‘thug’ is not a slur. First, it does not appear to have a neutral counterpart. ‘Thug’ is what we might call essentially pejorative; its function is to pick out a reference class whose members possess some negatively valenced normative trait. In this case, the class is of those individuals who behave criminally. There does not appear to be a way of referring to this same group of people in a neutral way. p. 437

7

Suppose we deny, along with Lauren Ashwell, that neutral counterparts are essential for identifying slurs. ‘Thug’ still behaves di erently in various contexts than other expressions we recognize as slurs. First,

considering that ‘thug’ is understood—at least in the US context—to refer mainly to Black people, we might think people would be more reticent to use it, as opposed to some euphemism. I concur with John 8

McWhorter that ‘thug’ is itself a euphemism, “a nominally polite way of using the N-word.” Given this, if ‘thug’ were on a par with the N-word we might expect the former to be treated similarly to the latter. But that is not the case. Another di erence is in a kind of exibility that applies to ‘thug’ that usually is not applicable to slurs. Terms like ‘thug,’ ‘criminal alien,’ and ‘poor kids’ can refer to di erent subsets of groups depending on context. For example, although ‘thug’ has come to indicate race-speci c senses primarily, it can occasionally be used in racially non-speci c ways. If Tucker Carlson decries the lowly state of the National Basketball Association due to an increasing thug presence, this gives o

racist vibes. Because the NBA

predominantly comprises Black players, I believe this would be a reasonable reaction to Carlson. In contrast, when Cornel West declares that the US government is full of thugs and gangsters, racial connotations are usually absent. Slurs do not usually exhibit this kind of referential exibility. People can indeed attempt to

use a slur in a nonstandard way, but the e ect is often metaphorical. Case in point: various videos have emerged from time to time of non-Black adolescents referring to each other as “niggas” in much the way some Black speakers do. In these situations, the non-Black speakers are not using the expression in its standard derogatory form, but their use still provokes thoughts of impropriety in a way that West’s use of ‘thug’ typically does not. I suppose one could claim West’s use of ‘thug’ is like Black use of the N-word—membership in the target group is relevant for how we interpret the expression’s use. However, the racial connotations of ‘thug’ can, in certain contexts, be mitigated even when used by non-members of its typically race-speci c use. A headline from a Leicestershire news website reads, “Thug stalked and throttled the mother of his 9

children.” The article is accompanied by a picture of a White woman with her hand raised in front of her face. The website and story are produced for a general audience and authored by a White man. The news story plausibly manages to access a racially non-speci c use of ‘thug’ in the way West does. If it were not so, we would expect broadly recognizable norms against the use of explicitly anti-Black slurs, especially by non-Black people and organizations, to generate public discontent and outcry. For these reasons I believe ‘thug’ is not a slur. Because ‘thug’ often expresses racial overtones yet is exible enough in meaning to allow room for plausible deniability, you could be forgiven for thinking it might be best classi ed as a code word or dogwhistle. Theorists such as Tali Mendelberg (2001), Ian Haney Lopez (2015), Jennifer Saul (2018), and Justin Khoo (2017) have provided accounts of racially coded language—dogwhistles—language that accesses p. 438

existing racial resentment to make surreptitious racial appeals. Jennifer Saul, for instance, usefully distinguishes overt, covert, intentional, and unintentional dogwhistles. An overt intentional dogwhistle is a speech act designed, with intent, to allow two plausible interpretations, with one interpretation being a private, coded message targeted for a subset of the general audience, and concealed in such a way that this general audience is unaware of the existence of the second, coded interpretation. (2018, 362) Saul illustrates this kind of dogwhistle with an example from George W. Bush’s 2003 State of the Union speech: (1) Yes there’s power, wonder-working power, in the goodness and idealism and faith of the American people. The phrase ‘wonder-working power’ is meant as an overt intentional dogwhistle for Evangelical Christians. According to Saul, there are two possible messages Evangelicals can take away from Bush’s utterance. The rst message is simply a translation: Yes there’s power, the power of Christ, in the goodness and idealism and faith of the American people. (362) The second message is that Bush identi es with them and speaks their language. Saul thinks both are instances of overt intentional dogwhistles. Saul (2018) describes covert dogwhistles as “a dogwhistle that people fail to consciously recognize” (365). The infamous Willie Horton ad, run in George H. W. Bush’s 1998 presidential campaign, which targeted a prison furlough program Michael Dukakis oversaw as governor of Massachusetts, usefully illustrates the covert intentional dogwhistle. The ad showed a picture of Horton, an African American man who, while out on furlough, raped a White woman and stabbed her husband. Though race was not explicitly mentioned, it

was clear to many that the ad drew on racial tropes about Blackness and criminality to stoke fear in White voters. An unintentional dogwhistle, according to Saul (2018), is an “unwitting use of words and/or images that, used intentionally, constitute an intentional dogwhistle, where this use has the same e ect as an intentional dogwhistle” (368). Dogwhistles of this sort are passed on by unwitting others and achieve similar e ects to the original intentional one. Given that ‘thug’ seems to allow for racial and non-racial interpretations, Saul may call it an overt intentional dogwhistle. While this is an initially plausible explanation, there are reasons to reject it. In his article “When Code Words Aren’t Coded,” Patrick O’Donnell distinguishes code words from racialized terms. Racial code words, says O’Donnell, are expressions that “trigger post-semantic inferences or 10

associations which get across racialized content.”

Recall Jennifer Saul’s non-racial example from the

George W. Bush campaign speech: (1) Yes there’s power, wonder-working power, in the goodness and idealism and faith of the American people. p. 439

Bush employs a phrase, wonder-working power, that resonates with Evangelical Christians in a way not necessarily transparent for those not in that group. The attitudes or ideas associated with the phrase for Evangelicals is triggered when Bush utters it, perhaps signaling that he identi es with them. In the case of racially coded language, the speech acts as a kind of wink to the relevant crowd that the speaker shares their beliefs or attitudes. This is essentially what the Republican strategist Lee Atwater acknowledged in his 11

explanation of the switch from explicitly racist language to more subtle language.

According to Atwater,

terms like ‘states’ rights’ and ‘busing’ are proxy terms meant to appeal to White voters who experience racial resentment. What’s crucial about code words is that they are an indirect way of conveying information. O’Donnell argues that terms like ‘thug,’ contrary to coded language, are more direct in their e ects, though, like code words, play on racial resentment. He describes the contrast as follows: i) Racialized terms involve direct or predicative relations between a term and a racialized group whereas code words involve indirect inferential or associational relations, and ii) Racialized terms elicit racial resentment by making salient race-speci c interpretive options, 12

whereas code words function by making salient race-neutral interpretive options.

O’Donnell characterizes the di erence between the two sorts in terms of the directness of race-speci c interpretations made available. What makes distinguishing code words from racialized terms di

cult is that we are trying to track di erent

types of e ects they can have on an audience. Both appear to operate on a person’s antecedently held attitudes—racial resentment. The di erence is in the directness of the relation between the term and the racialized group. If racialized terms are more direct about their connection to speci c groups, why are people like Muller confused about them? One explanation is that they are acting in bad faith. Many are fully aware of the expression’s racial overtones and are simply feigning ignorance. Newt Gingrich appears to do this when he responds to Juan Williams’s pushback against statements Gingrich made that Black Americans should demand “jobs, not food stamps” and that “poor kids” lack a strong work ethic. Williams asked if Gingrich could see that such statements would insult Black Americans, and Gingrich responded that he could not see 13

that.

In the context of the statements, it was clear ‘poor kids’ referred to Black kids particularly.

However, suppose that Muller was not acting in bad faith and was genuinely confused about thug’s racial connotations. Should we think the connection between the term and a group must always be transparent to count as a racialized term? This depends on a couple of factors. One factor concerns the scope of the ignorance of this connection. If the connection is broadly recognized, then Muller’s confusion would simply re ect a lack of competence on her part. Another factor concerns the genesis of a connection’s purported obscurity. It sometimes happens that a term’s racial connotations are systematically hidden from select p. 440

14

groups.

Recall again Atwater’s strategic use of “states’ rights”

in lieu of explicit slurs. His goal was to

trigger racial resentment in White voters without employing language that made race transparent. Claiming “states’ rights” lacked salient racial connections would be mistaken. Describing ‘thug’ and similar expressions as racialized terms provides a better way of explaining how they work than do models that treat them like slurs or code words, which, in turn, opens space for more e ective ways of responding to particularly problematic uses.

4. Thug Narratives Before moving on to possible strategies of response, I want to highlight the historical association between blackness and criminality. The ubiquity of that association and how it works helps to explain why Muller and others do not perceive it. I begin with an autobiographical tale. When I was eleven years old I had a run-in with the law. I was a school crossing guard and had walked to my post with my friend, who also was a crossing guard. We stood at the crosswalk for a few minutes before a police o

cer pulled up. Neither of us was wearing the identifying mark of a school crossing guard—i.e., an

orange plastic belt that draped across the torso. The o

cer aggressively probed us with questions. “What

are you doing out here at this time of the morning?” “We are crossing guards,” we replied. Peering in my direction, the o

cer continued, “Where are your patrol belts?” “Mine is in my book bag,” quipped my

friend. “Put it on,” the o

cer demanded. “I forgot mine at home,” I sheepishly responded. The o

cer

barked at me, “Go get it.” When I returned “O The o

cer Friendly” was gone, and my friend lled me in on their conversation once I left.

cer told him not to hang around with “my kind” and that I was trouble. Given the brief interaction

between the o

cer and me, the most plausible basis for his judgment was my being a Black male. Somehow,

he was able to deduce that I was a troublemaker during our brief encounter. In truth, it is doubtful the o

cer and I even needed to speak for him to come to that conclusion. His

judgment is one made quite often when it comes to Black people. It is one that has a long tradition in the United States. From the earliest days of colonial America, inherent Black inferiority and criminality have been a persistent story. Khalil Gibran Muhammad details the cloaking of this prejudice in scienti c jargon in 15

The Condemnation of Blackness.

Muhammad notes how many people writing on the so-called Negro

problem used statistics to justify claims about inherent Black criminality. Even today, the notions of crime and blackness are so tightly wound that many people can hardly escape thinking ‘black’ when hearing ‘crime.’ The ease with which this association is made suggests that it emerges from a set of background beliefs and habits that govern our social perception. Social psychologists have done important work demonstrating the robustness of the troublesome p. 441

association. For example, in their article “Seeing Black: Race, Crime,

and Visual Processing,” Eberhardt

et al. demonstrate the frequency and robustness of the association by conducting a series of experiments 16

involving object recognition and face priming.

Other social psychologists have also demonstrated this

association through various experiments, e.g., Devine (1989); Payne (2006); Correll et al. (2002) and

Greenwald, Nozek, and Banaji (2003). Even more work suggests that the association may be automatic; Payne (2006); Payne, Lambert, and Jacoby (2002). Eberhardt et al. also claim that “the Black racial category functions as the prototypical associate for a 17

number of ostensibly race-neutral concepts, such as crime, jazz, basketball, and ghetto.”

This claim,

coupled with the results of work like Payne (2006), helps answer why Muller and others like her fail to see any racial connotations with respect to ‘thug.’ As studies have shown, the association is made automatically so the speaker need not consciously think Black people are criminals for the use of ‘thug’ to reinforce or re ect that association. The racialized nature of ‘thug’ exacerbates a danger already present with words of its type more generally. As stated earlier, ‘thug’ is a pejorative expression that licenses negative evaluation judgments. When they are perceived as being aptly applied, such terms can further be used to justify certain forms of punitive treatment. If widely perceived as apt, calling someone a racist can mean things like social ostracization or being red, for instance. Classi catory pejoratives like ‘thug,’ ‘racist,’ ‘misogynist,’ ‘transphobe,’ as well as slur words, are value-laden; standard uses are meant, in part, to signal that the object of the pejorative is someone who is outside of some acceptable social order. Societies ordered by White domination take this up a notch since they treat non-pejorative, non-White subject classi cations like ‘Black,’ ‘gay,’ ‘trans*,’ etc., as socially deviant—in certain contexts and to varying extents. In our current case, the racialization of ‘thug’ works to destabilize the social acceptability of ‘Black’ by casting a shadow of suspicion over that group’s members. I think it is clear that Black people have good reason to be concerned about mainstream representation that reinforces this association between blackness and criminality. And if, as I referenced earlier, people like John McWhorter are right about ‘thug’ being a slightly nicer way of saying the N-word, there is good reason to respond to this kind of language. Thus, given the robustness and automaticity of the association between blackness and criminality and the function of ‘thug’ as justi cation for punitive treatment, what can be done about it? How should Black people respond to such an insidious and pervasive language? In the following sections I explore two broad responses: refusal and resistance.

5. Linguistic Refusal What makes terms like ‘thug’ so dangerous is that people tend to believe that subjecting those labeled as thugs to liberty-restricting treatment is justi ed. Thugs are criminals, and criminals forfeit their right to uninterrupted freedom due to their socially disruptive actions. Language is not only used descriptively; it p. 442

serves a number of functions,

including subjugation. The term’s pejorative force makes it ripe for

weaponization, a tool for bypassing calls to engage in the back-and-forth of reason giving. ‘Thug’ and similar terms carry evocative weight that sways audiences emotionally. If a term like this becomes associated with a particular group of people in the broader social imagination, it can have profound e ects on how members of this group are viewed and treated in that society. Recognizing this, some people attempt to counter weaponized uses by resisting them. I wish to highlight two broad types of response: refusal and resistance. Before delving into what resistance might entail, I rst want to turn attention to the other strategy—refusal. There are various ways to understand refusal. For example, Bonnie Honig (2021) provides a feminist theory of refusal based on three concepts—inoperativity, inclination, and fabulation. She grounds her view in a reading of Euripides’s tragedy the Bacchae. In the play, the women (1) refuse to work, (2) leave the city to live otherwise, and (3) return with a list of demands (Honig 2021, 1). Honig conceives of these three things as belonging to an arc of refusal. Accounts of refusal can also be found in works by Fred Moten (2003), Herbert Marcuse (1969), and Lindsey Stewart (2021).

In An Essay on Liberation, Herbert Marcuse says “the Refusal” makes people “reject the rules of the game 18

that is rigged against them.”

We can borrow this sentiment to characterize the general response I aim to

identify. I should rst distinguish my intended focus from what are often described as speech acts of refusal. Houck Noël and Susan Gass describe these speech acts as ones by which a speaker “denies to engage in an 19

action proposed by the interlocutor.”

As an example, consider this refusal of a request:

John: Would you be able to help with my homework tonight? Mary: Sorry, I can’t. This small exchange features an initiating act (the request) followed by a refusal response. In these situations, the initiating act typically asks or demands something of the other party. John, in this case, calls on Mary to do something. Mary’s response is a refusal in that she declines to comply. We can regard examples like this as utterance-level refusals. Additionally, I think there are also discourse-level refusals. ‘Discourse’ has several de nitions, but what I intend to refer to here is something that goes beyond language. James Gee (1991) nicely captures what I have in mind. He describes (big D) Discourse as “ways of being in the world, or forms of life which integrate words, acts, values, beliefs, attitudes, social identities, as well as gestures, glances, body positions and 20

clothes.”

He further describes Discourse as “a sort of ‘identity kit’ which comes complete with the

appropriate costume and instructions on how to act, talk, and often write, so as to take on a particular social 21

role that others will recognize.”

Big D Discourse is contrasted with small d discourse—“connected 22

stretches of language that make sense, like conversations, stories, reports, arguments, essays.”

Big D

Discourses include discourses while also being more than language. p. 443

There are, then, two levels at which refusal in the relevant sense can take place—utterance-level and Discourse-level. An example of the former is represented by former National Football League cornerback Richard Sherman’s response to being called a thug: “I was on a football eld showing passion … I wasn’t 23

committing any crimes or doing anything illegal.”

Sherman was called a thug—and worse—on Twitter

after delivering an enthusiastic interview with Erin Andrews after making a game-saving defensive play in 24

an NFC Championship game.

Sherman essentially performed an utterance-level refusal, denying the

term’s application in his case. Though this may seem like a sensible strategy to employ, I think it poses some signi cant challenges as a general strategy. First, denying that the label applies actually leaves in place what was initially objectionable, the purportedly natural association of blackness and criminality. If Sherman’s response had worked, all it would have accomplished was his (momentary) exoneration while leaving the problematic association in place. Second, Sherman’s response ultimately implies that he accepts the legitimacy of the 25

association. Philosophers of language refer to this implication as either an implicature or a presupposition. Now, with either one of these phenomena, it is possible to deny one’s implicit endorsement. But if the

speaker fails to explicitly object to the implication, it is left in place and assumed to be the case. Thus, there is a real danger to simply denying the term’s application. Avoiding the dangers associated with Sherman’s response requires a Discourse-level refusal. What this means is a refusal to accept the presumed values and recommended ways of being that undergird ‘thug’s’ meaning in its speaker’s world view. This is perhaps made all the more important when that world view converges with the values and social order sanctioned by the governing state. Opting out of “the game” is a refusal to accept or entertain a particular Discourse’s authority. That is, as far as possible, one refuses to legitimate the broader discourse in which racialized terms like ‘thug’ get their meaning. Perhaps its inclusion in a broader politics of refusal is needed to make such an approach e ective. This may take the form of opting out of one game coupled with opting into an entirely di erent one. I won’t detail

what such politics would look like here, but I envision something akin to what Lindsey Stewart describes in The Politics of Black Joy: Zora Neale Hurston and Neo-Abolitionism (2021). She describes a politics based on nding an alternative basis for value and meaning “within the group” rather than on gaining political recognition from White worlds. In short, refusal—on this view—is sustained by creating a community of language users who share an alternative Discourse with di erent values and subject positions.

6. Linguistic Resistance A possibly di erent response to racialized terms like ‘thug’ is resistance. What kind of actions should count as acts of resistance? There are di ering opinions on the matter. For some, the agent’s action must be aimed p. 444

at reducing oppression. Others believe

the agent’s actions must reduce some constraint or, at least, set

actions to reduce those constraints in motion. It is beyond the scope of this chapter to argue for a theory of resistance, so in what follows, I will largely borrow from an extant view. Much of what we consider an act of resistance depends on how we answer certain questions about the agent’s mental state, the object at which their actions are aimed, the conditions or circumstances surrounding the agent, etc. For instance, on some views, the agent must perform an action with the intention of reducing oppression. The account of resistance that I adopt here is found in Lawson and McGary (1993). In Between Slavery and Freedom, Howard McGary o ers a view of resistance that allows for the counting of subtle acts by slaves during American chattel slavery—for example, the breaking of farm tools or suicide— as resistant acts. According to McGary’s account, an action undertaken by some agent A would be judged as an act of resistance if (a) The condition of constraint (positive or negative) exists against agent S and the group in which S is a member, and S has a general conception of these constraints, and (b) S’s action A is intentional under S’s description of A which may or may not be the same description under which S’s action A reduces the constraint, and (c) S’s action A under an appropriate description is one that could reduce the constraint directed at the group of which S is a member, and (d) The causal process that S sets in motion with his action A is one that reasonable persons, who are similarly constrained and aware of their constraints, would also be likely to set in motion if they 26

wanted to reduce these constrictions.

As McGary notes, the constraint need not be one of oppression. Intuitively, one can resist situations that are not themselves oppressive. For example, a child may resist her mother’s attempts to show her a ection in front of her school friends. That is surely not an oppressive situation. I also believe it is a genuine act of resistance and not just a metaphorical extension of the concept. Condition (b) also allows for an agent to engage in resistance without intentionally aiming to reduce constraints, even though under certain descriptions that action might be characterized as reducing constraints. I do think there is an ambiguity present in the use of constraint that it will be helpful to make clear. The view does not make it explicitly clear whether we are to understand the constraints as external or internal. External constraints are largely legal and social, such as redlining practices in housing or Jim Crow segregation. It is these kinds of constraints that probably leap to mind most readily. Internal constraints, on the other hand, are typically psychological. We might view one aspect of attempts by slaveholders to deny their slaves education, for instance, as an attempt to erect an internal constraint, one aimed at stunting the

development of slaves’ mental lives. Internal constraints aim to diminish one’s self-valuation, often by imposing ideas and understandings that reduce one’s imaginative capacity. I do not see why we should discount internal constraints as proper targets of resistance. In its most basic sense, resistance intuitively p. 445

targets attempts

by someone to control others. The means of control might be physical or mental.

Recognizing both has implications for understanding the conditions of McGary’s view. One nal thing to mention is that McGary’s view does not require the constraint actually be reduced for an act to count as resistance. Clause (c) of the account contains an epistemic modal, ‘could.’ This of course leaves open the possibility that the action could not reduce constraints. I think a viable theory of resistance should make room for that possibility; though an action might ultimately be ine ective, it still makes sense to classify the attempt as resistance. The considerations raised in this section gives us more latitude in explaining how we might understand certain self-referential uses of ‘thug’ as acts of linguistic resistance. Now that I’ve brie y described how I am thinking of resistance for my present purposes, we can turn to self-referential uses of ‘thug’ and how they may count as acts of resistance.

7. Thug Life Instead of engaging in utterance-level refusal, which appears to leave the problematic association intact (for the moment, I’m leaving aside Discourse-level refusal), one could respond to racialized terms with acts of resistance. As I have noted in section 6, I understand resistance as an act intentionally performed by a constrained person that could, under an appropriate description, be recognized as an action aimed at reducing those constraints. One condition of constraint for Black people is, of course, that they are socially perceived as inherently criminal. One possible example of this type of resistance toward internal constraint may be found in the words of the late rapper Tupac Shakur. Tupac possessed a prominent visible marker in the form of a tattoo across his abdomen, “THUG LIFE .” He 27

once described the phrase’s signi cance as a new kind of “black power.”

Tupac remarks that to White

people, Black [males] are “thugs and niggers,” and that is what they’ll stay until they “set this shit right.” Tupac’s message seems to be that the term should be embraced and ultimately defeated through means of self-empowerment. At the same time, there is also a kind of de ance in embracing the expression. Tupac isn’t suggesting that Black males accept the negative judgments that accompany the particular uses of ‘thug’ we have been discussing. Rather, it is a turning of a racist narrative on its head, a refusal to play the part as it has been scripted. Tupac was often the target of harsh criticism by people like C. Delores Tucker, Rush Limbaugh, and Jay Leno. Tucker, Limbaugh, and others condemned what they saw as glorifying criminality and violence. But Tupac counters this charge in an interview in the Los Angeles Times: Let me say for the record, I am not a gangster and never have been. I’m not the thief who grabs your purse. I’m not the guy who jacks your car. I’m not down with people who steal and hurt others. I’m just a brother who ghts back. I’m not some violent closet psycho. I’ve got a job. I’m an 28

artist. p. 446

Read through the lens provided in this interview, I believe we can interpret Tupac’s use of ‘thug life,’ not as a glori cation of frivolous violence or as an acceptance of the negative perceptions of Black people in 29

general and Black males in particular but as a “ ghting back” against racist characterizations.

Note that what Tupac seems to have in mind is not a simple rejection of the White-generated caricature of Black males but rather a resistance-through-signifying, perhaps in ways similar to how some describe 30

certain Black minstrelsy performers.

The idea is that caricaturing the thug persona constructed to paint

Blacks in a negative light is a way of resisting it. Does Tupac’s strategy count as a form of resistance? I want to suggest that the way in which he adopts the term can be seen as primarily targeting internal constraints. We might envision the constraints in view to be psychological in nature. It seems reasonable to expect that, for instance, constant mischaracterizations of Black adolescents’ mischievous behavior as overly aggressive impact their self-conception. Further, the barrage of images of Black (and Brown) people, especially in the news media, coupled with a dearth of White faces, can also impact self-conceptions. If it is the case that the promulgation of Black criminality propaganda produces a limiting e ect on the Black psyche, then devising a strategy that counters those potentially damaging e ects is pertinent. If Tupac’s approach is to count as resistance, at least in the sense I’ve outlined in section two, it must satisfy the conditions listed. It is reasonable to conclude that condition (a) is satis ed. As we have already seen, external and internal constraints exist against Black males. It is also clear that Tupac’s use of ‘thug’ is intentional under his own description, thus satisfying condition (b). I see no reason to deny that there can be an appropriate description of Tupac’s preferred use of the expression that could reduce the constraints imposed on Black males, thus satisfying condition (c). Finally, it is plausible to assume that reasonable persons similarly situated and aware of the constraints raised against them would also perform similar actions to those of Tupac. Of course, the concept of “reasonable persons” presents a challenge since one can always ask, “reasonable to whom?” Although space does not allow us to address this concern fully, an initial response might be restricting the notion of reasonableness to what persons in the particular situation would conclude as the appropriate course of action. It is also true that situations of duress are not ideal conditions for clarity in decision-making. But it is important not to marginalize or dismiss the considerations of those in such conditions. Although I describe Tupac’s use as a form of resistance, you might also read it as an act of refusal. What he says appears to be consistent with refusal’s rejection of the salient Discourse’s system of classi cation and valuation. There is also the appearance of opting out. However, one thing that may make Pac’s example one of resistance rather than refusal is its self-aware adoption of the racialized term. This kind of self-aware adoption disrupts the valuation attached to ‘thug’ through direct engagement, one intended to confuse and disrupt one’s opponent. Presumably, acts of refusal, on the other hand, are acts of disengagement. Where p. 447

linguistic resistance seeks to change the value or meaning

of the existing word, linguistic refusal simply

chooses to speak a di erent language. It is unclear to me whether one could perform linguistic refusal by reappropriating the language in question, but if it is possible, it would need to be a form of disengagement, invested in its own bases for things like esteem, as opposed to an attempt to reform or reconstruct a hostile Discourse.

8. Conclusion We’ve seen that racialized terms like ‘thug’ and ‘terrorist’ pose danger because classi cation subjects its targets to punitive treatment. These classi cations often serve social orders structured to privilege some group interests and subordinate others. Individuals need not be consciously aware that these terms function this way; it is enough to simply use terms in a manner consistent with this function, undergirded by the relevant system of valuation.

I have also presented two broad strategies for dealing with racialized terms—refusal and resistance. I do not wish to recommend a preferred strategy; I leave this open, as e ectiveness can vary across di erent contexts. I hope I’ve provided a clear analysis of how such terms operate and some potentially fruitful guideposts for how social philosophers of language might begin to forge a more symbiotic relationship between social philosophy and the philosophy of language.

Notes

p. 448

1.

Baldwin (1998, 781).

2.

See “A Thug By Any Other Name” (Muller 2009).

3.

Muller (2009).

4.

Bourdieu (1999, 109).

5.

I should note that the concept of a ʻneutral counterpartʼ has been called into question. For instance, Lauren Ashwell denies that neutral counterparts are essential for identifying a term as a slur; whereas Geo Nunberg argues that the terms we identify as neutral carry evaluative connotations.

6.

Erlenbusch-Anderson (2018).

7.

Ashwell (2016).

8.

McWhorter (2015).

9.

Mack (2022).

10.

OʼDonnell (2020, 831).

11.

Perlstein (2012).

12.

OʼDonnell (2017, 28).

13.

Stanley (2015, 157).

14.

Charles Millsʼs (2017) characterization of White ignorance as a systematic unknowing might supply a general framework for this particular claim.

15.

Muhammad (2010).

16.

Eberhardt et al. (2004).

17.

Eberhardt et al. (2004, 877).

18.

Marcuse (1969).

19.

Houck and Gass (1999, 2).

20.

Gee (1991, 142).

21.

Gee (1991, 142).

22.

Gee (1991, 142).

23.

Farrar (2014).

24.

Farrar (2014).

25.

There are two kinds of implicatures that are recognized, conventional and conversational. The kind most relevant for our

purposes is conversational implicature. Conventional implicatures are, roughly, conventionalized features of expressions that arise due to convention but are not part of the assertoric content of the expression. Conversational implicatures, on the other hand, arise due to features of the conversational context. 26.

McGary and Lawson (1993, 48–49).

27.

Lazin (2003).

28.

Philips (1995).

29.

Side note: Admittedly, Tupacʼs use of ʻthugʼ was more complex, less straightforward, than represented in his interview response. He arguably had moments where he seemed to embrace, in either word or deed, at least some behavior o en associated with thugness. But that can be set aside for our current purposes since it has no bearing on the proposal we are considering.

30.

Taylor and Austen (2012).

References Ashwell, Lauren. 2016. “Gendered Slurs.” Social Theory and Practice 42 (2): 228–239. Google Scholar WorldCat Baldwin, James. 1998. James Baldwin: Collected Essays. New York: Library of America. Google Scholar Google Preview WorldCat COPAC Bourdieu, Pierre. 1999. Language and Symbolic Power. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Correll, Joshua, Bernadette Park, Charles M. Judd, and Bernd Wittenbrink. 2002. “The Police O icerʼs Dilemma: Using Ethnicity to Disambiguate Potentially Threatening Individuals.” Journal of Personality and Social Psychology 83 (6): 1314–1329. Google Scholar WorldCat Devine, Patricia G. 1989. “Stereotypes and Prejudice: Their Automatic and Controlled Components.” Journal of Personality and Social Psychology 56 (1): 5–18. Google Scholar WorldCat Eberhardt, Jennifer L., Phillip A. Go , Valerie J. Purdie, and Paul G. Davies. 2004. “Seeing Black: Race, Crime, and Visual Processing.” Journal of Personality and Social Psychology 87 (6): 876–893. Google Scholar WorldCat Erlenbusch-Anderson, Verena. 2018. Genealogies of Terrorism: Revolution, State Violence, Empire. New York: Columbia University Press. Google Scholar Google Preview WorldCat COPAC Farrar, Doug. 2014. “Richard Sherman says itʼs really disappointing to be called a thug.” Sports Illustrated, January 22. Accessed September 23, 2023. https://www.si.com/nfl/2014/01/22/richard-sherman-seattle-seahawks-super-bowl-xlviii. Google Scholar WorldCat Gee, James Paul. 1991. Social Linguistics and Literacies: Ideology in Discourses. London: Falmer Press, First Edition. Google Scholar Google Preview WorldCat COPAC Goddard, Angela, and Neil Carey. 2017. Discourse: The Basics. London: Routledge. Google Scholar Google Preview WorldCat COPAC p. 449

Greenwald, Anthony G., Brian A. Nosek, and Mahzarin R. Banaji. 2003. “Understanding and Using the Implicit Association Test: I. An Improved Scoring Algorithm.” Journal of Personality and Social Psychology 85 (2):197–216. Google Scholar WorldCat Haney-Lopez, Ian. 2015. Dog Whistle Politics: How Coded Racial Appeals Have Reinvented Racism and Wrecked the Middle Class. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Honig, Bonnie. 2021. A Feminist Theory of Refusal. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Hornsby, Jennifer, 2001. “Meaning and Uselessness: How to Think About Derogatory Words.” Midwest Studies in Philosophy 25 (1): 128–141. Google Scholar WorldCat Houck, Noël, and Susan Gass. 1999. Interlanguage Refusals. Berlin: De Gruyter. Google Scholar Google Preview WorldCat COPAC

Khoo, Justin. 2017. “Code Words in Political Discourse.” Philosophical Topics 45 (2): 33–64. Google Scholar WorldCat Lazin, Lauren, dir. 2003. Tupac: Resurrection. Produced by Paramount Pictures. Google Scholar Google Preview WorldCat COPAC Mack, Tom. 2022. “Thug Stalked and Throttled the Mother of His Children.” Leicestershire Live, May 16. https://www.leicestermercury.co.uk/news/leicester-news/thug-stalked-throttled-mother-children-7078355.amp. WorldCat Marcuse, Herbert. 1969. “An Essay on Liberation.” Marxists Internet Archive. Accessed May 18, 2022. https://www.marxists.org/reference/archive/marcuse/works/1969/essay-liberation.htm. WorldCat McGary, Howard, and Bill Lawson. 1993. Between Slavery and Freedom: Philosophy and American Slavery. Bloomington: Indiana University Press. Google Scholar Google Preview WorldCat COPAC McWhorter, John. 2015. “The Racially Charged Meaning behind the Word ʻThug.ʼ ” Interview with Melissa Block. All Things Considered. National Public Radio, April 30. Accessed September 24, 2023. https://www.npr.org/2015/04/30/403362626/theracially-charged-meaning-behind-the-word-thug. Google Scholar WorldCat Mendelberg, Tali. 2001. The Race Card: Campaign Strategy, Implicit Messages, and the Norm of Equality. Princeton, NJ: Princeton University Press. Google Scholar Google Preview WorldCat COPAC Mills, Charles. 2017. Black Rights/White Wrongs: The Critique of Racial Liberalism. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Moten, Fred. 2003. In the Break: The Aesthetics of the Black Radical Tradition. Minneapolis: University of Minnesota Press. Google Scholar Google Preview WorldCat COPAC Muhammad, Khalil Gibran. 2010. The Condemnation of Blackness: Race, Crime, and the Making of Modern Urban America. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Muller, Judy. 2009. “A Thug by Any Other Name.” Hu ington Post, February 6. Accessed October 31, 2022. https://www.hu post.com/entry/a-thug-by-any-other-name_b_155708. WorldCat OʼDonnell, Patrick. 2017. “Generics, Race, and Social Perspectives,” Inquiry: An Interdisciplinary Journal of Philosophy, published online, January 26, 2017. doi:10.1080/00207543.2016.1266801. Google Scholar WorldCat OʼDonnell, Patrick. 2020. “When Code Words Arenʼt Coded.” Social Theory and Practice 46 (4): 813–845. Google Scholar WorldCat Payne, B. Keith. 2006. “Weapon Bias: Split-Second Decisions and Unintended Stereotyping.” Current Directions in Psychological Science 15 (6): 287–291. Google Scholar WorldCat Payne, B. Keith, Alan J. Lambert, and Larry L. Jacoby. 2002. “Best Laid Plans: E ects of Goals on Accessibility Bias and Cognitive Control in Race-Based Misperceptions of Weapons.” Journal of Experimental Social Psychology 38 (4): 384–396. Google Scholar WorldCat

Perlstein, Rick. 2012. “Exclusive: Lee Atwaterʼs Infamous 1981 Interview on the Southern Strategy.” The Nation, November 13. Accessed October 31, 2022. https://www.thenation.com/article/archive/exclusive-lee-atwaters-infamous-1981-interviewsouthern-strategy/. Philips, Chuck. 1995. “Tupac Shakur: ʻI Am Not a Gangster.ʼ” Baltimore Sun, October 25. https://www.baltimoresun.com/la-metupac-qa-story.html. p. 450

Saul, Jennifer. 2018. “Dogwhistles, Political Manipulation, and Philosophy of Language.” In New Work on Speech Acts, edited by Daniel Fogal, Daniel Harris, and Matt Moss, 360–383. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Scheuerman, William. 2017. “What Is Political Resistance? An Exploration of the Word and Its Political Connotations.” Public Seminar, February 16. Accessed March 29, 2022. https://publicseminar.org/2017/02/what-is-political-resistance/. WorldCat Stanley, Jason. 2015. How Propaganda Works. Princeton, NJ: Princeton University Press. Google Scholar Google Preview WorldCat COPAC Stewart, Lindsey. 2021. The Politics of Black Joy: Zora Neale Hurston and Neo-Abolitionism. Evanston, IL: Northwestern University Press. Google Scholar Google Preview WorldCat COPAC Taylor, Yuval, and Jake Austen. 2012. Darkest America: Black Minstrelsy from Slavery to Hip-Hop. New York: W. W. Norton. Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

p. 451

20 Public Protest and Silencing  José Medina https://doi.org/10.1093/oxfordhb/9780192844118.013.43 Published: 22 May 2024

Pages 451–464

Abstract Sometimes protests cannot get o

the ground because they are preemptively silenced. Other times acts

of protest are produced, but they are not heard; and yet in other cases, protests are noticed but they receive no uptake or only a super cial and defective uptake. This chapter develops a critical examination of the speech act of protesting and of the forms of illocutionary silencing and of perlocutionary frustration or disablement that often neutralize acts of protest because of defective uptake or lack thereof. More speci cally, it will analyze the case of protesting racial violence and how so often these acts of protest fall on deaf ears and achieve no results because of speci c communicative dysfunctions and problems of reception and uptake.

Keywords: discursive power, epistemic activism, illocutionary force, illocutionary silencing, linguistic resistance, perlocutionary disablement, protest Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction In one of his postings on Twitter from May 28, 2020, Donald J. Trump referred to the demonstrators in Minneapolis protesting the police murder of George Floyd as “thugs” and said, “When the looting starts, 1

the shooting starts.” What does this demeaning and threatening response do to the communicative acts of street demonstrators? In what follows I will argue that there is a particular kind of silencing or communicative inhibition in this kind of uptake, in addition to the explicit intimidation and threat of violence contained in it: a silencing and inhibitory communicative move that subverts (and also recruits audiences to subvert) the communicative agency of the demonstrators and the illocutionary force of their communicative acts. I will identify di erent kinds of defective and distorting uptake to acts of public protestation, and in my analysis I will distinguish di erent kinds of silencing that result from such faulty and inhibitory responses. But before I get into the issues of defective uptake, let me o er a speech-act theoretic analysis of public protest in order to highlight the kinds of speech acts or illocutions involved in public protest and the kinds of uptake that they call for.

2. Public Protest as a Complex Illocution That Demands Uptake Public protests are ways of dissenting, criticizing, and demanding in public discourse. This short description already contains the three essential components and dimensions of public protest: the expressive dimension p. 452

of voicing dissent; the evaluative dimension

of criticizing; and the prescriptive dimension of issuing

demands. In their speech-act analysis of political dissent, Matthew Chrisman and Graham Hubbs (2018) underscore the evaluative and prescriptive dimensions of voicing dissent or protesting. I draw from their analysis, but I want to build on it and expand it by also underscoring the expressive dimension of public protest and by adding details about the evaluative and prescriptive dimensions that will be crucial to identify the kinds of uptake that public protests call for and the kinds of silencing that can occur when proper uptake is not given. In “Speaking and Listening to Acts of Political Dissent,” Matthew Chrisman and Graham Hubbs, focus their analysis on the normative question of whether (and according to which criteria) public acts of protest can be considered legitimate and worthy of public consideration; and, appropriately, their primary concern is the content of these communicative acts, what these acts are about, for only by examining what kind of contribution to public discourse these acts make can one determine what kind of due consideration they deserve. Although I will also address the normative issue of the audience’s prima facie obligation to listen and to give due consideration to acts of public protest, I want to start with an issue that is prior to that and more fundamental. Before addressing the issue of how the content of acts of protest should be communicatively engaged with and responded to, we have to elucidate how these acts can be recognized as acts of protests in the rst place, that is, as expressive, evaluative, and prescriptive acts of a particular kind. To begin with, it is very often the case that when we run into a public protest on the streets we can recognize it as a public protest even before (and independently of whether) we are able to identify the content of the protest, by reading signs carried by demonstrators or hearing their chants. We see people taking to the streets, we see people sitting in silence in a square, we see a gathering with some point or purpose, and we recognize it as a public protest. In doing so, we are recognizing the expressive power of assembly. When we recognize a street protest in the way I have described, we recognize a collective expressive act; we recognize that a group of people gather, form an assembly, to make a public statement of some sort, even if we do not know what content that statement may have or whether it has any de nitive content at all: maybe it is a silent protest, and demonstrators are not saying anything at all; maybe di erent demonstrators are saying di erent things But there is something bringing them (and their claims) together. The very act of gathering

is a collective expressive act, a way of saying “we are here together,” “we are performatively enacting a communal bond by gathering,” “we are here in protest.” At a minimum, what unites the protestors, what gives minimal unity and cohesiveness to their communicative act is (1) that they are assembling, and (2) that they are assembling in protest; that is, they are expressing publicly a dissenting communicative response to public life: “we are here in protest.” This expressive act of gathering to protest, which is prior to and independent of speci c contents that the protest may contain, is a speech act in its own right, an illocution: what I will designate as the illocution of voicing dissent, which is to be distinguished from the content-based illocutions of voicing criticisms and demands that appear in the evaluative and prescriptive dimensions of protest. p. 453

The expressive act of voicing dissent can be understood as the act of bonding publicly through dissent; and thus, it can be understood as containing two intimately related expressive acts: the community-forging and community-expressive act of bonding, and the expressive act of negative positionality—that is, the expression of protestation, of positioning themselves collectively against something, of signaling rejection or saying “No,” independently of the speci c claims and reasons that the protestation or rejection may contain for the individuals who endorse it. These are two key elements in protests as expressive acts: (1) they express community or communal bonds; and (2) they express dissent or a negative reaction to public life or public policy (even before, and independently of, the speci c criticisms and demands that can be articulated within the protest). First, an act of protest is the expressive articulation of a public—that is, the formation of a collective subject with expressive power, of a communicative plural subject. Acts of public protest are expressive of communal bonds, and at the same time, they are also formative of those bonds; that is, they forge social ties in the very act of being together and standing in solidarity with one another. Secondly, an act of protest is the expressive articulation of negative positionality, that is, of carving out a communicative space of repudiation and contestation, a space of standing against, a space for saying “No,” “Enough!,” “Stop!,” or “We will not take it anymore.” The last element of this complex expressive act of bonding publicly through dissent is the dissenting part, which is of course indexed to a particular content: something against which protestors position themselves, something they denounce, repudiate, or contest. But notice that one can recognize that people gathering in public are protesting even before the content of their protest becomes determinate or well-de ned, and before speci c claims can be ascribed to the protest. This is an important point to recognize to properly appreciate the expressive power of protest from the point of view of protestors, for they may come together in protest to express their unrest or discontent even before they articulate any speci c evaluative or prescriptive claim, or may come together in protest to express a sense of indignation even when they endorse very di erent evaluative or prescriptive claims to give content to their indignation. But it is also an important point to recognize to properly appreciate the expressive power of protest from the point of view of their audiences, who should be able to recognize something as a public protest even if they do not yet understand what the protest is about or what its speci c evaluative and prescriptive contents might be. The speci c contents of the dissent become the focus of other illocutions that are also essential components of public protest. These contents will be unpacked and elaborated in the evaluative and prescriptive acts typically contained in public protests. In its evaluative dimension, acts of public protest articulate a critique in public discourse. They contain the illocution of criticizing. This second, evaluative dimension is directly related to a third one that is often present as well in acts of public protest: a prescriptive dimension and the illocution of demanding. In their speech-act theoretic analysis of public dissent, Chrisman and Hubbs (2018) emphasize that dissenting political speech always contains an evaluative element and often, but not always “a corresponding prescriptive element: All verbal acts of dissent evaluate something as bad or wrong in some way, and most

p. 454

correspondingly demand change to rectify the badness

or wrongness in question” (174). To put it in

terms of my analysis of public protest as a complex illocution, Chrisman and Hubbs’s suggestion is that

protesting always involves the illocution of criticizing, and often also the illocution of demanding. I have also highlighted the importance of recognizing the expressive illocution of voicing dissent as distinct from (although closely related to) the evaluative illocution of criticizing and the prescriptive illocution of demanding. Chrisman and Hubbs (2018) go on to identify three constitutive felicity conditions of criticizing and demanding: sincerity in disapproval, good faith in making a demand, and being based on considerations of justice (174–176). If dissenting political speech meets these conditions, it yields felicitous public critiques and demands. If dissenters violate these conditions (they speak insincerely, in bad faith, or not guided by considerations of justice, but rather, by sel sh aspirations of personal gain or preferential treatment), then their political dissent “is a bad speech act because it is conducted in a way that is self-undermining and so is infelicitous” (175). Chrisman and Hubbs are interested in identifying abuses of dissenting political speech and in distinguishing between felicitous and infelicitous evaluations and prescriptions contained in that speech, so that we can determine whether or not we have a prima facie obligation to listen to public dissent and take it seriously: “If dissent is voiced in a sincere, good faith way and appeals to recognizable considerations of justice, that gives the community members to which the dissenters belong a reason to listen to and consider what the dissenters are saying” (176). Although I am interested in this normative issue in the pragmatics of public protest, I am more interested in the more fundamental issue of recognizing communicative acts of protest as such so that they can be heard at all as acts of protest and can be given due consideration and proper uptake (if felicitous). What is involved in properly recognizing the illocutions contained in protesting—voicing dissent, criticizing, demanding—and what is required in order to give minimal proper uptake to these communicative acts? Lack of proper recognition and failures of uptake can result in di erent kinds of silencing, which I will examine in section 3. But before we shift to the complex issue of silencing protest in its expressive, evaluative, and prescriptive dimensions, let me brie y elucidate how the illocutions of voicing dissent, criticizing, and demanding gure in acts of public protest and how they are interrelated. The expressive, evaluative, and prescriptive elements of public protest are intimately connected and build upon one another: articulating a public critique presupposes voicing dissent, and articulating a public demand is based on and presupposes (at least implicitly) a critical evaluation. But while the expressive is presupposed in the evaluative and the evaluative is presupposed in the prescriptive in the complex illocution of public protest, these presuppositions are unidirectional (from the expressive to the evaluative to the prescriptive) and do not go in the other direction: the evaluative does not presuppose the prescriptive, and the expressive does not presuppose the prescriptive or the evaluative (at least, not in its speci city—in p. 455

terms of speci c evaluative

contents—though it does presuppose the evaluative in a generic sense, that

is, as a generic negative evaluation of public life). As Chrisman and Hubbs (2018) point out, the prescriptive element of dissent is predicated on its evaluative element, which is always present in the act of voicing dissent; but not all acts of dissent contain a prescriptive element and involve the illocution of issuing a demand: dissenters can voice criticisms without necessarily specifying or agreeing on the demands that their criticisms should give rise to. For example, a Black Lives Matter protest can criticize racist practices and policies of the police without issuing any speci c demand about how to correct them; and protestors can speak in a uni ed, evaluative voice in their repudiation of racist police violence without agreeing on how the police should be reformed, defunded or abolished. Criticizing is in this sense prior to and independent of demanding. On the other hand, voicing dissent can also be said to be prior to and independent of demanding and even of criticizing, at least in a quali ed way. Voicing dissent does not have to be predicated on a particular critical evaluation on which the dissenters’ sense of injustice has to be based. Dissenters are not required to agree on their reasons and to speak in a uni ed, evaluative voice for their dissenting political speech to enter public discourse with su

cient unity and coherence. Think, for example, of a silent protest:

people are gathered to denounce something, and they express dissent with their silence, but not speci c critical evaluations. For example, a silent protest that responds to the killing of a person of color by the

police denounces racial violence, but while some protestors may base their public repudiation of racial violence in a critique of institutional racism, the critical evaluation of other protestors may focus on the cruelty of particular police o

cers. In voicing dissent, protesters are denouncing something with their

protest, but they may di er in their critical evaluations, and they may have di erent reasons on which they 2

base their dissent and di erent views of the source of the problem or wrong they denounce.

Giving proper uptake to public protests will require, at a minimum, recognizing the expressive, evaluative, and prescriptive communicative acts they contain and responding to them. In other words, giving proper uptake to a public protest will require recognizing it as a complex illocution that involves, at a minimum, the expressive act of voicing dissent, the evaluative act of criticizing, and often also the prescriptive act of demanding. The expressive dimension is particularly important for proper uptake, for, in order to get proper uptake o

the ground, public protests have to be recognized as such, as an expression of political agency, as

a way of voicing dissent. As we shall see in section 3, key forms of silencing public protest—illocutionary silencing and illocutionary ipping or discursive injustice—consist precisely in not recognizing properly its expressive dimension, not recognizing the communicative act of voicing dissent. But additional forms of silencing—locutionary and perlocutionary silencing—can occur also even if public protest is recognized as such when its evaluative and prescriptive dimensions are not properly recognized and their critical and prescriptive contents become distorted, diverted, or suppressed.

p. 456

3. Defective Uptake and Di erent Kinds of Silencing How are communicative acts of public protest silenced? In this section I will distinguish four di erent kinds of silencing of public protest, but the rst occupies a special place since it involves preemptive silencing before acts of protests even occur, whereas the other three types of silencing on which I will spend more time involve defective communicative reactions to acts of protests after they occur. First, would-be acts of protests can be prevented from occurring, and thus su er what I term pre-locutionary silencing. The act of protest is preemptively and pre-locutionarily silenced when the illocutions of protesting cannot occur because people are prevented from becoming protesters—they cannot gather, they cannot publicly dissent, criticize, and demand. This is what happens when there are explicit prohibitions against public protest, or when there are intimidating and silencing communicative climates that make it next to impossible and heroic to be able to gather to protest—e.g., the silencing of protests of segregationist policies during Jim Crow. There are also cases of near pre-locutionary silencing in which protests are allowed, but though they do occur, their illocutions (the forms of dissent, criticism, and demand they perform) are precluded from entering public life. Think, for example, of the creation of so-called free speech zones for protests after 9/11 by the George W. Bush administration. Here, a protest can occur, but in spaces (e.g., far-removed parking lots) where nobody but the protesters themselves will see it and hear it, so that, e ectively, the protest is preemptively silenced in all other public spaces and shared public life. Indeed, the designation of a special place as “a free speech zone” implicitly indicates that all other spaces are not spaces of free speech where the right to protest can be exercised. In the second place, when public protests occur, the communicative attempts to voice dissent, criticize, and issue demands may be silenced by the audience’s inability to recognize the illocutions performed in protest, and the audience’s inadequate treatment or lack of communicative response to such illocutionary acts. This way of neutralizing the illocutionary force of a speech act ts well with what Langton and Hornsby (1998) have described as the phenomenon of illocutionary silencing. Illocutionary silencing occurs when the illocutionary force of a speech act is nulli ed, and the speech act is thus rendered void, as if it had never taken place. As Luvell Anderson (2017: 144) puts it, illocutionary silencing is “a short-circuiting of illocution” produced by audiences refusing to give uptake. Langton and Hornsby describe it as “illocutionary disablement.” They illustrate this with the disabling of women’s capacity to refuse sexual

advances in a sexist culture in which their communicative acts of refusal are silenced by lack of uptake, and their saying “No!” is treated as pure noise, with no communicative force at all. But how are protests illocutionarily disabled or silenced? How do they become pure noise and lose the communicative force they were intended to have because of the audience’s defective uptake or lack thereof? p. 457

I began this chapter with a quote from Donald Trump calling street protestors “thugs” and referring to their activities as “looting,” and I alleged that this uptake was a way of silencing protest: it does so by neutralizing its illocutionary force, and therefore it should be counted as an example of illocutionary silencing. The particular kind of silencing protest instantiated by Trump’s characterization of protestors as thugs, rioters, and looters has a long history in the United States, but the prime example, which provides the perfect backdrop for Trump’s attempt at illocutionary silencing, can be seen in the defective uptake given in mainstream media to the 1992 Los Angeles uprising after the police o

cers who had brutally beaten Mr.

Rodney King were exonerated. This silencing has been brilliantly analyzed by Robert Gooding-Williams (2006) in “Look, A Negro!” Gooding-Williams points out that there were two dominant views of the media coverage of the L.A. uprising in 1992. According to what he calls “the conservative view,” protestors were depicted as opportunist rioters: “[T]he people on the streets were taken to embody an uncivilized chaos that needed to be stamped out in order to restore law and order. On this account, the ‘rioting’ had nothing to do with the King verdict but expressed a repressed opportunism just waiting for an excuse to out the law” (2006: 14). On the other hand, what Gooding-Williams calls “the liberal view” in the media coverage of the King uprising depicted protestors as out-of-need rioters: “the liberal view emphasized the social causes of the ‘riots,’ such as joblessness, poverty, and, more generally, socio-economic need” (14). As GoodingWilliams’s analysis emphasizes, by depicting the protestors as rioters (whether opportunistic or out-ofneed), both views in the media dissociated the uprising from the King verdict and made it di

cult for

publics to see it as a protest at all—that is, as an exercise of political agency, a way of voicing dissent; as he writes “[B]oth liberals and conservatives were refusing to see in it an expression of moral indignation” (14). Whether the “rioters” were depicted as “bearers of chaos” or as “looters,” they were not seen as protesting a repressive political order and legal system, as acting out of moral indignation and the belief that the harm in icted on Mr. King with impunity symbolized a larger social injustice that they could not tolerate any longer. As Gooding-Williams puts it, “It strains incredulity to deny, as did conservative and liberal pundits alike, that the L.A. uprising was not for many an act of political protest” (14). What this media coverage re ects is “a failure to regard the speech or actions of black people as manifesting thoughtful judgments about issues that concern all members of the political community” (14). In other words, what this media coverage does is to neutralize the illocutionary force of the complex illocution of protesting, so that publics become incapable of recognizing what the protesters say and do as voicing dissent, criticizing, or demanding. This is a paradigmatic example of the illocutionary silencing of protest. Just as the 2020 uprising was reminiscent of the L.A. uprising of 1992, so was Trump’s response reminiscent of the problematic uptake that the Rodney King protests had in mainstream media, and the silencing e ects are very similar. And note that this particular kind of illocutionary silencing of protestors in the United States had speci cally targeted people of color, depriving them of communicative and political agency. The p. 458

communicative mistreatment of protestors of color in this way also quali es as

what Quill Kukla calls

discursive injustice. As they put it, discursive injustice occurs “when members of a disadvantaged group face a systematic inability to produce a speci c kind of speech act that they are entitled to perform—and in particular when their attempts result in their actually producing a di erent kind of speech act that further compromises their social position and agency” (2014: 440). The last part of Kukla’s characterization suggests that, insofar as the depiction of protestors of color as rioters and looters amounts to a discursive injustice, it may involve more than illocutionary silencing: it may also involve illocutionary ipping. As Luvell Anderson explains, “illocutionary ipping” is the communicative phenomenon “in which a speaker sets out to perform one speech act—while being entitled to perform it in that context and use the conventionally appropriate words, tones, and gestures—but the kind of uptake given constitutes an entirely di erent act”

(2017: 144). Illocutionary ipping seems to be one of the communicative e ects that the activation of the rhetorical frame of “rioting and looting” can have on the audience of a protest. The description of protestors of color as rioters and looters can do more than neutralize the force of their communicative attempts and turn their dissenting political speech acts into pure noise. Such description can transform, rather than simply destroy, the expressive power of assembly and of the communicative attempt to protest. At least in some cases, by depicting protestors of color as rioters and looters, the communicative import of their taking to the streets and of their screams is not entirely destroyed, but it is, rather, transformed, so that audiences are directed to recognize what they say and do, not as acts of protest, but as communicative acts of an entirely di erent kind: threats or acts of political intimidation. In the third place, even when the communicative agency of protestors and the expressive power of their assembly are respected, and their acts of voicing dissent, criticizing, and demanding are recognized as such, silencing can still take place when the contents of those acts are radically distorted or subverted. This is what I will call locutionary silencing. In locutionary silencing, the illocution in question does take place unimpeded; but the content of the locution is lost or radically transformed in such a way that protestors’ voices are silenced by the defective uptake they receive, which makes it impossible for the audience to hear properly the content of what they say. A powerful example of locutionary silencing can be found in the “All lives matter” response to “Black lives matter.” Luvell Anderson (2017) has brilliantly analyzed what is involved semantically in this defensive and distorting response. As Anderson emphasizes, the obvious meaning intended by protesters chanting “Black Lives Matter” is captured by an inclusive reading of the phrase that can be glossed as follows: “There is an implicit ‘too’ attached to the end of the phrase so that it should read ‘Black lives matter, too!’ [… The central message] is that we as a society have historically treated Black lives as if they are valueless and expendable, and that this should no longer be the case” (Anderson 2017: 139). But when conservative pundits and counterprotesters retort with the phrase “All Lives Matter,” they are clearly distorting this message and imposing what Anderson calls an exclusive reading, which interprets “Black Lives Matter” as singling out Black lives as more valuable and proposing a devaluing of non-Black lives, almost as if saying “only Black lives matter” (139). As Anderson points out, p. 459

“Those who give an exclusive

reading to the slogan are belittling, downplaying, or ignoring the turbulent

history of how Black people have been treated in the United States” (Anderson 2017: 146). I could not agree more with Anderson’s analysis, but I take issue with his suggestion that there is a form of 3

illocutionary silencing involved here. The complex illocution of protesting and the expressive, evaluative, and prescriptive speech acts it can contain do take place; they are not entirely nulli ed (illocutionarily silenced) or transformed into something else (illocutionarily ipped). But the content of these acts is blocked from circulation; their proper interpretation is disabled. It is for this reason that I think it is more accurate to describe cases of defective uptake that does not respect the content of a protest as locutionary silencing. But it is true that the line between locutionary and illocutionary silencing may be hard to draw in particular cases and contexts, since the distortion of the content of a protest may cast the communicative 4

act of protesting in a di erent light, and even as something that no longer looks like a political protest. In other words, the illocutions of protest may not be una ected by the locutionary manipulation and silencing: 5

in particular, the evaluative illocution of criticizing may not disappear as such, but it may be downgraded as a form of defective (e.g., arbitrary, illegitimate, etc.) criticizing that does not carry evaluative force of the right kind. Finally, in the fourth place, there is also a kind of silencing that can occur in the communicative afterlife of a protest, so to speak. This is what I will call perlocutionary silencing, which takes place after the illocutions of voicing dissent, criticizing, and demanding have taken place and have received (at least minimal) proper uptake (hence no illocutionary silencing), and after their contents have been properly processed (hence no locutionary silencing). Perlocutionary silencing can be identi ed by tracing the communicative consequences and impact of acts of protests in terms of whether or not they were allowed to produce the

intended perlocutionary e ects. If the act of protest is cut o

from the communicative consequences it is

intended to have so that it cannot be given due consideration and cannot subsequently carry weight in public life, then the act of protest is being perlocutionarily silenced. The perlocutionary silencing of protest typically takes place when reparative measures and steps toward recti cation of the injustice denounced by protesters are systematically blocked, rendering the protest a vacuous, banal, or ine ective cry, and thus downgrading its illocutions. This kind of silencing is especially noticeable in the prescriptive illocution of demanding, which falls on deaf ears when the demands issued are not given due consideration and are not properly engaged with. Elsewhere I have introduced the concept of perlocutionary silencing to characterize one of the ways in which the grievances of incarcerated subjects in the American criminal justice system are silenced. In “Agential Epistemic Injustice and Collective Epistemic Resistance in the Criminal Justice System” (2021), I have argued that when the grievances of incarcerated subjects are registered, but in such a way that they are cut o

from their intended consequences, there is perlocutionary silencing. In that essay I o er a real-life

example of perlocutionary silencing drawn from a case study I co-authored with Matt Whitt. Medina and Whitt (2021) analyzed the defective handling of the grievances of detained subjects in the Durham County Detention Center, where grievances are processed by an automated system that gives the appearance of p. 460

proper recognition and uptake, but

e ectively short-circuits their communicative life by blocking them

from wider circulation, severely constraining the communicative impact that they could otherwise have— that is, their ability to produce perlocutionary consequences. Given the way the automated system is set up, the grievances have limited communicative scope by being handled by low-level administrators, without addressing structural problems and possible institutional reforms to correct those problems, ensuring that they do not reach the higher administration, have no communicative reach outside the jail, and cannot gure into the deliberations and policymaking of the county commissioner and any oversight committees. In the case of public protests, examples of perlocutionary silencing can be found by examining the traditional inability of the Black Lives Matter protests to produce political consequences and institutional changes, until recently. Until the 2020 uprising, the communicative impact of Black Lives Matters protests on American publics and institutions had been severely limited. This was not only because of the depictions of the protestors as “thugs” making noise (illocutionary silencing) or as insulting, threatening, and intimidating (illocutionary ipping), or because of the manipulations and distortions of the content of their protests (locutionary silencing). It was also by turning the protests into a media spectacle, inviting audiences and institutions to relate to them as spectators, without communicatively engaging them or addressing them as concerned parties. Only recently have mainstream American publics felt compelled to echo the demands of BLM protests. Only recently have institutions felt compelled to respond to these demands and at least consider serious institutional measures to address the injustices in the policing of people of color. Part of the explanation of how that defective communicative situation remained in place for 6

so long has to do with the spectacularization of the protests. Indeed, spectacularization is a mechanism of communicative disengagement—that is, of creating distance between the event (spectacle) and the audience (mere spectators), fostering disengaged and detached attitudes and making it di

cult if not

impossible for audiences to feel addressed as a concerned party and to be involved in working towards recti cation, toward reparative and preventive measures that properly address the demands issued in the protests. This is one of the reasons (a communicative reason) why Black Lives Matter protests had been taking place for many years—within a history of anti-racist violence protests that had been going on for many decades and yet they had not been able (until 2020) to move public opinion and to motivate institutions to consider changes and work toward structural reforms that address the vulnerabilities of people of color to police mistreatment in the United States. The perlocutionary silencing of BLM protests is ongoing, and some have argued that it is not only the media that has turned the street protests into a spectacle; even the participants themselves often contribute to this

spectacularization when they treat the protests as a mere photo opportunities. In a recent op-ed in The Washington Post (July 23, 2020), E. D. Mondainé, president of the Portland branch of the NAACP, criticizes White demonstrators who have transformed the street protests into a “white spectacle.” In particular, Mondainé says about the so-called “Naked Athena”—the White female protestor striking yoga poses nude p. 461

on the streets of Portland whose images have gone viral—that he does not see in her “a brave ally of the cause,” but rather, “something else: a bene ciary

of white privilege dancing vainly on a stage that was

originally created to raise up the voices of my oppressed brothers and sisters.” (1) Mondainé emphatically states, “I do not believe it is a time for spectacle,” because the spectacularization of street protests distracts audiences from the communicative and political aims of these protests, de ating them, and this is precisely what their detractors want: “The president [Trump] and his allies want spectacle, be it a naked yogi or the next shocking display of force. They need to distract the country by engaging our movement in empty battles where they have the advantage. … We cannot fall for their deception. We cannot settle for spectacles that endanger us all. This is a moment for serious action. … We welcome our white brothers and sisters in this struggle. In fact, we need them. But I must ask them to remain humbly attuned to the opportunity of this moment—and to re ect on whether any actions they take will truly help establish justice, or whether they are simply for show.” (2) Mondainé’s powerful words are a way of admonishing protesters, reminding them the dangers of perlocutionary silencing that the spectacularization of protests bring. As in locutionary silencing, in perlocutionary silencing, too, the illocutions of protest may not be una ected by perlocutionary manipulations: in particular, the prescriptive illocution of demanding may not disappear 7

as such, but it may be downgraded as a form of defective (e.g., arbitrary, illegitimate, etc.) demanding that does not carry normative force (at least not legitimate force). If the illocutionary silencing of protests typically disables the expressive illocution of voicing dissent, and if locutionary silencing typically subverts the evaluative illocution of criticizing and its evaluative contents, then the perlocutionary silencing of protests typically curtails and downgrades the prescriptive illocution of demanding insofar as the demands issued in a protest (implicitly or explicitly) get blocked and disabled. Note that there is a key di erence between the illocutionary silencing of protests, on the one hand, and the locutionary and perlocutionary silencing of protests, on the other: the former eliminates the illocution of protesting altogether; it is, so to speak, a total eclipse of the act of protestation, which either becomes depleted of communicative force altogether (illocutionary silencing proper that turns protests into mere noise), or becomes a di erent kind of speech act (illocutionary ipping that transforms legitimate protests into insults, threats, intimidations, etc.). By contrast, in locutionary and perlocutionary silencing, the illocutions of protest do not disappear completely, but they are partially eclipsed or downgraded, that is, their content becomes distorted, limited in scope, and de ated in its ability to address audiences and demand responses from them.

4. Concluding Remarks: Resisting the Silencing of Protests In section 2 above, I analyzed public protest as a complex illocution that contains expressive, evaluative, p. 462

and prescriptive speech acts: voicing dissent, criticizing, and

demanding. In section 3, I identi ed di erent

ways of silencing public protest: pre-locutionary silencing, which prevents would-be acts of protestation from occurring at all; illocutionary silencing, which nulli es and invalidates the speech acts of voicing dissent, criticizing, and demanding and sometimes transforms them into communicative acts of a di erent kind (illocutionary ipping); locutionary silencing, which distorts the evaluative and prescriptive contents of the complex illocution of protest; and perlocutionary silencing, which limits the capacity of the speech acts of protest to achieve perlocutionary e ects and thus de ates and downgrades these acts through distraction, diversions, and co-optations. This taxonomy of silencing points us in the direction of di erent forms that the task of resisting the silencing of protest can take. Resisting the silencing of protest should target di erent communicative

dysfunctions. The upshot of my analysis of illocutionary silencing is that a communicative strategy, such as the activation of the rhetorical frame of “rioting and looting,” which manipulates, nulli es, and/or subverts the communicative force of protests, should be considered as an attack on free speech and the right to protest; and that such strategy is an illegitimate, propagandistic way of closing audiences’ ears and hearts to dissenting voices that should be publicly censured and condemned. My analysis of locutionary and perlocutionary silencing further suggests that we need to denounce and ght distortions that block the evaluative and prescriptive contents of protests, and communicative strategies and dynamics in and around protests—such as spectacularization—that result in perlocutionary disablement and prevent protests from achieving their intended political aims. Resisting the silencing of protests often requires structural and institutional changes (such as the derogation of laws and city ordinances that outlaw protests or limit people’s capacities to voice dissent, criticize, and issue demands). But resisting the silencing of protests also involves critical re ection and vigilance with respect to the communicative attitudes, practices, and strategies that publics use when they confront protests. By focusing on proper uptake as a way of resisting silencing, my analysis calls attention to audience’s responsibilities. We need to become active listeners to the voices of protests and cultivate a sense of responsibility as concerned parties addressed by protests. We have a shared responsibility to listen to and respond to legitimate protests if we want to live in a community that feels concerned by the grievances of its members, and if we want to have institutions that feel compelled to rectify wrongs, repair damages, and prevent harms. In my view, being a responsible and active audience member of protests involves more than merely listening; it involves being compelled to engage with expressions of dissent and their evaluative and prescriptive contents, and being moved to consider reparative and preventive measures that address the grievances of public protests. This view may appear overly demanding for individual audience members, but it should be understood as laying out shared goals that audiences can collectively aspire to approximate as much as possible. Resisting the silencing of protests and working toward proper uptake should be thought of as collective goals that we aspire to achieve together. Audience’s responsibilities to properly recognize protests and to p. 463

give proper uptake to them are

not individualized responsibilities, but shared responsibilities that we

have as citizens and community members, as members of concerned publics. My analysis in this chapter underscores the importance of instilling these shared communicative responsibilities with respect to protests and holding individuals, communities, and institutions accountable for the uptake they give to protests.

Notes 1.

The complete tweet says: “These THUGS are dishonoring the memory of George Floyd, and I wonʼt let that happen. Just spoke to Governor Tim Walz and told him that the Military is with him all the way. Any di iculty and we will assume control but, when the looting starts, the shooting starts. Thank you!” https://twitter.com/realDonaldTrump/status/1266231100780744704.

2.

Think here also of the di erent protests that went into the Arab Spring and how, while the protestors were unified in denouncing an untenable political situation, their critical evaluations o en went in di erent directions: targeting the lack of democratic freedoms, targeting transnational, neoliberal economic policies, etc.

3.

Anderson claims: “The post-racial response to the slogan has at least two e ects. First, there is an obscuring of what the speaker intends to be the import of her speech. And second, there is a loss in the ability of certain speakers to both produce certain utterances and be interpreted correctly” (2017: 143). He goes on to explain the silencing e ects in terms of “illocutionary disablement” and “illocutionary flipping” (143–144).

4.

The exclusionary reading imposed on “Black lives matter” may make audiences perceive the original claim of the protest not as a claim of justice but as a plead for arbitrary privilege and preferential treatment. Given the third felicity condition

for dissenting political speech laid out by Chrisman and Hubbs (2018)—namely, that such speech must appeal to considerations of justice, the exclusionary reading of “Black lives matter” would make the political speech of their defenders infelicitous and abusive and, in fact, not legitimate dissenting speech at all, but rather, a manipulative attempt to earn privileges disguised as political protest. 5.

I am grateful to my colleague Megan Hyska for suggesting this conceptualization of the downgrading of illocutions as a form of silencing or near silencing.

6.

I have analyzed this in “Resisting Racist Propaganda: Distorted Visual Communication and Epistemic Activism” (2018).

7.

Megan Hyska suggested to me this conceptualization of the downgrading of illocutions.

References Anderson, Luvell (2017), “Epistemic Injustice and the Philosophy of Race,” in Ian James Kidd, José Medina, and Gaile Pohlhaus Jr., eds., The Routledge Handbook of Epistemic Injustice (New York: Routledge), 139–148. Google Scholar Google Preview WorldCat COPAC Chrisman, Matthew, and Hubbs, Graham (2018), “Speaking and Listening to Acts of Political Dissent,” in Casey Rebecca Johnson, ed., Voicing Dissent: The Ethics and Epistemology of Making Disagreement Public (New York: Routledge), 164–181. Google Scholar Google Preview WorldCat COPAC p. 464

Gooding-Williams, Robert (2006), Look, A Negro! Philosophical Essays on Race, Culture and Politics. New York: Routledge. Google Scholar Google Preview WorldCat COPAC Hornsby, Jennifer, and Langton, Rae (1998), “Free Speech and Illocution,” Legal Theory, 4, 21–37. Google Scholar WorldCat Medina, José (2018), “Resisting Racist Propaganda: Distorted Visual Communication and Epistemic Activism,” Southern Journal of Philosophy, 56, 50–75. Google Scholar WorldCat Medina, José (2021), “Agential Epistemic Injustice and Collective Epistemic Resistance in the Criminal Justice System,” Social Epistemology 35: 2, Special Issue: Epistemic Injustice and Collective Wrongdoing (2021), 185–196. Google Scholar WorldCat Medina, José, and Whitt, Matt (2021), “Epistemic Activism and the Politics of Credibility: Testimonial Injustice Inside/Outside a North Carolina Jail,” in Heidi Grasswick and Nancy McHugh, eds., Making the Case: Feminist and Critical Race Philosophers Engaging Case Studies. Albany: State University of New York Press, 293–324. Google Scholar Google Preview WorldCat COPAC Mondainé, E. D. (2020), “Portlandʼs Protests Were Supposed to Be about Black Lives. Now, Theyʼre White Spectacle,” op-ed, The Washington Post, July 23, 2020, 1–2. Google Scholar WorldCat

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

25 AI with Alien Content and Alien Metasemantics  Herman Cappelen, Josh Dever https://doi.org/10.1093/oxfordhb/9780192844118.013.47 Published: 22 May 2024

Pages 573–593

Abstract 1

AlphaGo plays chess and Go in a creative and novel way. It is natural for us to attribute contents to it, such as that it doesn’t view being several pawns behind, if it has more board space, as bad. The framework introduced in Cappelen and Dever (2021) provides a way of thinking about the semantics and the metasemantics of AI content: does AlphaGo entertain contents like this, and if so, in virtue of what does a given state of the program mean that particular content? One salient question Cappelen and Dever didn’t consider was the possibility of alien content. Alien content is content that is not or cannot be expressed by human beings. It’s highly plausible that AlphaGo, or any other sophisticated AI system, expresses alien contents. That this is so, moreover, is plausibly a metasemantic fact: a fact that has to do with how AI comes to entertain content in the rst place, one that will heed the vastly di erent etiology of AI and human content. This chapter explores the question of alien content in AI from a semantic and metasemantic perspective. It lays out the logical space of possible responses to the semantic and metasemantic questions alien content poses, considers whether and how we humans could communicate with entities who express alien content, and points out that getting clear about such questions might be important for more ‘applied’ issues in the philosophy of AI, such as existential risk and XAI.

Keywords: AI, semantics, metasemantics, content, explainable AI, existential risk Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. The Semantics and Metasemantics of AI: An Introduction In earlier work (Cappelen and Dever 2021), we took metasemantic views proposed for thinking about content in humans and tries to apply them to AI systems. Along the way, we encountered aspects of these anthropocentric metasemantic views that are ill-suited to AI systems. One response to that was to try to build more abstracted versions of those metasemantic features, so that some of the details that are highly speci c to humans get taken out and are replaced by tools that are more broadly applicable (that was the strategy in our 2021 book). In this chapter we want to look at another response. Maybe AIs have alien contents and we need a fundamentally di erent kind of metasemantics for AI systems than we do for humans. The aim is to explore a cluster of closely related topics: (i) What would it be for a system to have alien (non-human) contents, and how could we recognize these as contents? (ii) What would it be for a system to have alien (non-human) metasemantics, and how could we recognize it as a metasemantics? (iii) Could arti cial intelligences have alien contents and alien metasemantics? (iv) If AIs have alien contents and metasemantics, how do we communicate (and otherwise engage) with them? Much of what we have to say is exploratory. It is in large part an e ort to convince readers that it’s possible that AIs have both alien contents and alien metasemantics. To the extent that we have conclusions, they are the following: p. 574

Conclusion 1. AIs could have alien contents and alien metasemantics. Conclusion 2. Even if they do, we can nd ways to communicate with them. Conclusion 3. It is necessary to take these options seriously to think clearly about issues such as existential risk, the value alignment problem, and explainable AI. Our larger goal here (as it was in Cappelen and Dever 2021) is to illustrate how familiar work from philosophy of language is directly relevant to understanding central issues surrounding arti cial intelligence, our interaction with AIs, interpretable AI, and explainable AI. Re ection about AIs and their representational capacities for the last thirty years has proceeded in more-or-less complete isolation from developments in philosophy of language and metasemantics. A tacit assumption is that our understanding of AI’s representational and communicative abilities is best left to computer scientists and those who are 1

trained to build the relevant kind of software. The issues raised in this paper show that there’s signi cant potential for philosophical engagement. We brie y illustrate potential payo s in the nal section with a discussion of existential risk, the value alignment problem, and explainable AI.

2. Alien Contents in AI 2.1. Types of AI Content Some of the purposes to which we currently put, or hope soon to put, AI systems, involve communication with our current contents; but other applications will require contents that are di erent from the ones we have now. Let us illustrate both types before considering what to say about them.

AIs with familiar contents: When we build an image recognition/classi cation AI system, the goal is for its outputs to mean that the image is an image of a duck or of a goose. When we build a medical diagnostic AI system, the goal is that its outputs will mean that the tumor is malignant or benign. When we build a creditevaluation AI system, the goal is that its outputs will mean that the client is low risk or high risk. For these sorts of systems to play the roles we want them to play, their contents need to be our contents, because we want them to answer speci c questions that we have, and that we have formulated using our contents. As we stressed in Cappelen and Dever (2021), there are di

cult challenges in working out what about such AI

systems would make their outputs have meanings and have the speci c meanings that we need them to have. AIs with unfamiliar contents: Other applications of AI systems may require and give rise to contents that are not our contents. Consider an example. We nd that average lifespans have been decreasing in a community, and we aren’t able to work out why. So we feed enormous masses of data into an AI datamining system: individual medical records, hospital records, weather patterns, economic indicators, crime p. 575

reports and judicial rulings, social media activity, detailed tra

c records from GPS tracking, and so

on.

The data-mining system sifts the data performing linear regression and looking for signi cant patterns, and eventually identi es what it takes to be a crucial factor, pointing us to several vectors of that pattern (some hospital patients, a ock of migratory birds, a recent high-pressure weather system). It could be that the crucial factor is something nameable with our current contents (a known disease, a speci c parasite thriving under certain environmental conditions, etc.). However, another possibility is that the crucial factor identi ed by the AI system cannot be described in our language. A relatively innocuous version of this would be a disease we had never detected and for which we therefore had no name. Less innocuous versions might be causal factors that don’t fall nicely into our existing categories of “disease”, “violence”, and so on, so that we would even nd it hard to work out what sort of new thing we were looking for. AI systems thus confront us with the possibility of alien content. This is not the rst time theorists have considered that option. The process of radical translation/interpretation has always contained an element of alienation, though with the typical expectation that the contents, once uncovered, will prove to be familiar contents in alien garb. We return below to the connection between the tools of radical translation and the concerns about understanding alien AI contents. Kuhnian paradigm shifts also raise the spectre of alien contents—the post-relativistic notion of mass is perhaps not wholly expressible in the pre-relativistic language. We glance toward the connections with Kuhnian incommensurability when we turn to issues of alien metasemantics below. In the data-mining example, we at least know the logical category of the alien content. The AI system, by design, is expressing some categorizing predicate—the di

culty is just that (ex

hypothesi) our language contains no predicate, atomic or complex, matching the alien predicate in content. But the shared logical category at least gives us a place to start. Perhaps we can eventually learn the alien category by ostension of exemplars. Even if not—if, for example, no collection of facts about the extension of the alien category enable us to get onto its intension—there is a natural enough similarity metric here, letting us perhaps nd contents of our own that are similar enough to the alien content. However, we may not always have the supporting crutch of a shared logical category. Consider the possibility of a map-making AI system, whose output is map-like representations of various places and things. Already, we have taken away much of the familiar content-structuring infrastructure of our languages—the AI’s maps need not use the tools of reference and predication, of quanti cation and Boolean truth functions, of a broadly type-driven categorial semantics. And if the AI map-maker is given the exibility to redesign its own map-making methods, in a quest for better representation of whatever is worth representing about the mapped environment, the similarity of its maps to our language-like conventions for creating map representations can dwindle away. The AI system shifts to projection methods using functions less and less natural to us; represents spatial relations in a non-Euclidean geometry or in a sub-Hausdor

topology; con ates spatial, temporal, and sociopolitical relations in its sub-topological

organizational scheme; deploys map symbols holistically and interrelationally, and so on. At some point, it would be unsurprising if the AI’s maps were wholly incomprehensible p. 576

to us, and untranslatable into

anything we could recognize as a map or a meaningful representation. The possibility of alien content is independent of the structure of the vehicle. While a large language model (LLM) such as ChatGPT produces output that wears the syntactic form of English, that syntactic form is no guarantee of semantic form. As we stressed (in Cappelen and Dever 2021), the mere fact that an AI system outputs texts like “Lucie is high risk” does not mean, absent a controversial metasemantic argument, that the AI system’s output means that Lucie is high risk. The AI could be using our words to express its contents. But of course, the AI need not be using our syntax to express contents in a way guided by the syntactic structure at all. ChatGPT (if it means anything by its outputs) could be expressing contents without truth conditions, contents formally representable only as complex constraints on probability functions or other information measures, dynamic update rules on alien scoreboards, and so on. Again, there is no guarantee that anything “said” is something we could say or understand.

2.2. Communication (or Communication*) with Alien Content Users What should we do, then, were we to encounter an AI producing alien contents? Normally, when we encounter linguistic utterances we don’t understand, we have reason to engage in a process of translation, looking for a method of mapping bits of their language to bits of our language in a meaning-preserving way. But of course, translation is possible only when the same contents are available on the sending and the receiving ends, so that translation can match utterances with the same content. By hypothesis, that’s not the situation we are in with alien contents. A devotee of Quine and Davidson’s method of radical translation might object that this hypothesis cannot be realized. Perhaps we can always run the radical translation machinery and always get an output, so we can always translate. There are two reasons why we nd this an implausible idea, even from within the QuineDavidson framework. • First, despite the claims of Davidson (1973), even when the radical translation tools are ready to hand, they may produce no translation. We may recognize the alien activity as intelligent, guided by beliefs and desires—activity classifying objects into categories, for example. We might nevertheless nd that we have no contents matching the distinctions guiding the alien activity. We can nd that no categories of ours categorize as they do. • Second, and more importantly, the gears of radical translation might fail to engage with alien contents. Depending on its avor, radical translation requires in its object beliefs and desires, or rational action, or knowledge. AIs or other possessors of alien contents might have no such states. That’s perhaps the p. 577

expected situation for the kinds of AI systems we are familiar with: they have nothing like desires, and states only analogically like beliefs. They may be less possessors of knowledge and more tools for the expansion of our knowledge. One of the things that make AI systems in particular such an interesting case of alien content is that their natures are a mixture of the alien and the familiar. While they may not, for example, desire themselves, they are tools created by and serving satisfaction of our desires, and something like radical translation of them in light of their relations to us may be possible. We return to this theme below when we discuss meta-metasemantics.) If translation is impossible because there is nothing of ours to translate their contents into, what is to be done? How is communication with or the understanding of AI systems possible, given that communication and understanding also seem to require shared content? We’ll explore two options, which we will call bridging and integration.

2.3. Bridging Suppose an AI produces outputs with contents that we lack, so that we cannot simply translate its outputs into our language. We might, nevertheless, be able to build a bridge of understanding between us and the AI by nding suitable corresponding contents of ours. Some versions of this strategy are familiar. Some have suggested that successful communication requires only that audiences uptake contents similar to contents entertained or intended by speakers, without 2

requiring full identity. Frege nds it unlikely that two speakers will share the very same sense for a proper name, but thinks they can communicate as long as their senses are similar. (In the limiting case, perhaps shared denotation su

ces for similarity.) 3

Or bridges could be built using the mereology of content. A verbal description of the contents of a painting, for example, might share content with the painting by capturing some of what is represented in the painting, but unavoidably might also leave out some of the pictorial content because pictorial representation makes available other contents that simply can’t be produced with linguistic tools. Perhaps, similarly, the alien content of the AI is essentially richer than our content, adding some who-knows-what on top of our familiar contents. But if we can, when encountering the AI output, at least get onto the portion of its content that overlaps with ours, some form of communication will have been achieved. Or the mereology might run in the other direction. Perhaps we have the richer contents, and the AI system can only produce contents that are coarsenings or diminishings of ours. Again, a form of communication is available by associating a class, or a representative of a class, of our richer contents with the AI’s alien diminished content. But the really interesting cases are ones in which we aren’t so fortunate as to have a ready-made bridging relation like similarity or parthood. How do we build a bridge to alien contents, when it’s a new bridge that’s needed? Nothing of full generality can be expected here, but we can point to what we think are the crucial questions and tools. p. 578

It’s helpful to get started here to realize that the aliens are already among us, and that they are, in fact, us. Philosophers are adept at producing creative and radically varying pictures of the nature of content, so philosophers disagree wildly among themselves about what sort of contents they are producing. These philosophers can’t all be right (since they disagree), but we can still consider how communication would happen between communities properly modelled by di erent philosophical pictures of content. Consider two test cases: 4

First, consider the expressivist: Simplifying, the expressivists take utterances not to represent the world, but rather to express, to give literal voice to, various mental states. Perhaps we are all expressivist about “ouch”—its sole content is as a tool for the expression of pain. More thoroughgoing expressivists may view our moral language as expressive of conative or evaluative attitudes, or our modal vocabulary as expressive of global features of our doxastic state, or the vast swath of our descriptive vocabulary as expressive of our beliefs. The contents of the expressivists cannot be translated into the contents of the representationalist— the representationalist tra

cs in propositions, while the expressivist does not, so there is no common

ground between them. In some sense, the expressivist and the representationalist should each be licensed in simply ignoring the other, if taken on the other’s own terms, since by their own lights, the other is saying nothing they can comprehend. But there is a natural enough mapping of expressivist content to representationalist content. We can see that already with “ouch”. We say “ouch” to express our pain, but instead of expressing our pain, we can report on it, by saying, “I am in pain.” Anything that is expressed can similarly be reported on (“I have a con attitude toward murder,” etc.). The reportings don’t say the same thing as the expressions, but the relation between reporting and expressing can support a lot of our communicative practice. The representationalist

who takes the expressivists’ expressions as reportings, or the expressivist who takes the representationalists’ reportings as expressings, won’t get everything right, but they will have an 5

interpretational practice that will allow them to make considerable sense of the other. 6

Second, consider the Lewisian centered-worlds de se theorist: The fan of centered worlds takes contents to be properties, not propositions. When he says “I am hungry”, his utterance has the property of being hungry as content; when he says, “Brutus stabbed Caesar”, his utterance has the property of being B-related and Crelated to a pair of objects, the rst of which stabbed the other as content. Again, the fan of possible worlds content cannot translate the utterances of the centered-worlds speaker, since his contents are all propositions, not properties. In some sense, fans of centered worlds and fans of uncentered worlds should regard each other’s view proclamations as simply incomprehensible, by virtue of not having (by their own lights) the right sort of content. As in the expressivist case, there is a natural enough mapping between centered and uncentered content. Centered content does not have truth conditions (being properties rather than propositions), but there is a natural way for the fan of propositions to extract associated truth conditions. On Lewis’s view, what we do cognitively when we have an attitude whose content is a property is to self-ascribe that property. That selfp. 579

ascription

has accuracy conditions—one rightly or wrongly self-ascribes a property. So the uncentered

theorist can map the centered expression of property P by an agent A to the uncentered content that A is P via the correctness conditions of A’s self-ascribing P. The centered theorist can similarly extract a centered content from every uncentered content that phi, by mapping it to the property of occupying a phi world. Perhaps these mappings don’t allow for communication, in the strict sense, between expressivist and representationalist, or between centered and uncentered theorist, but they allow for some form of communication*—something that is communication-like enough to be recognized. How do we nd the right mapping to allow for communication*, especially as the contents to be mapped from get increasingly alien? One way to answer this question is to look to the theoretical role that content plays. One formal gadget or another (sets of worlds, sets of centered worlds, etc.) counts as a content because it does something for us. Because our fellow philosophers are much like us, we can look for familiar roles for (e.g.) expressivist or centered contents—explaining action, guiding assessments of rationality, allowing modi cation of the beliefs of others, and so on. A representationalist has one view on the sort of things best positioned to play those roles, and so when encountering an expressivist content, they can look for a content of their own that’s suited to play the role in the way the expressivist’s content, on the expressivist worldview, is. The sketch we just gave of the theoretical role of content is deeply anthropocentric. The challenge faced in extending this approach to alien contents of AI systems is to generalize and de-anthropocentrize our understanding of the theoretical roles that contents play. Only with a de-anthropocentized account can we get a clear picture of what kinds of contents can play the relevant role for AIs. This line of thought thus leads to an investigation of metasemantics—what features make things mean what they do—and the metasemantics suitable for AI systems. We turn to that investigation in section 3. Before doing that, we consider the second option to encountering alien contents: integration.

2.4. Integration Rather than seek a suitable content of our own to match with the AI’s alien contents, we could acquire the alien contents, integrating ourselves into their linguistic practice and making ourselves users of those contents. Consideration of simple cases might suggest that the process of integration is straightforward. When our diagnostic data-mining AI gets onto a new category, sorting objects in some way that was previously unavailable in our language, we can simply introduce a new predicate and stipulate that its intension matches the intension by which the AI is sorting. The integration move here is a deferential move, but it’s important to distinguish two forms of deference: • In semantic deference, the property of deference is part of the meaning of the deferential expression. “Gloobish”, in our language, means “has the feature that the AI system picks out when its red light ashes”. p. 580

•

In metasemantic deference, on the other hand, the property of deference is no part of the meaning of the deferential expression—rather, deference is a metasemantic feature that xes the nondeferential meaning. If the AI system is, in fact, picking out property P, then with metasemantic deference, our “gloobish” simply means that P, by way of our deference.

Semantic deference is easily achieved (as easily as any other stipulation). However, semantic deference is also too cheap for most communicative purposes. With semantic deference, we can stipulate that “iliadic” means whatever the Greek expression “μῆνιν ἄειδε θεὰ Πηληϊάδεω Ἀχιλῆος//Οὐλομένην” means. If we have no idea what it does mean, “iliadic” gives us only thin, arm’s-length access to the content, not genuinely integrating it into our language. It gives us nothing we didn’t already have via quotation. Metasemantic deference is what we want. That’s the tool that allows the alien contents to become contents of our language. Metasemantic deference, however, is much less obviously easily achieved. We need to know what kind of deferential relation is needed to make our words mean what the AI words mean. There are at least two worries here. First, our notion of deference may not apply to AIs. Perhaps to defer is, in part, to respect the target of deference, to think that the target knows more about the subject than we do. But AI systems may not be sensible objects of respect and may not have knowledge as we understand it. If so, perhaps we can’t defer to them. And second, our notion of deference may not be the right one. Perhaps human deference is a metasemantic ground of meaning in human languages—but why think that human deference grounds meaning acquisition between human and alien languages? Perhaps alien deference is needed, or perhaps the alien language is entirely deference-proof. Of course, deference isn’t essential here—it was only one possible route to integration. But the concern generalizes. What we want (to use the Lewisian [1969)] terminology) is to put ourselves in the actual language relation to the alien contents. But to do that, we need to know what it takes to stand in the actual language relation to that sort of content. That is, we need to know about the metasemantics of the alien language. If the alien contents are grounded (externalist-style) in causal interactions with other speakers of the alien language, perhaps integration is not too hard—we simply interact with the AI systems in the right way for a time. If the alien contents are grounded (conventionalist-style) in participating in conventions of trust and truthfulness, integration may be hard to impossible, since AI systems may not be capable of entering into the relevant sorts of conventions.

3. Alien Metasemantics There is something that makes our words and utterances mean what they do. There is much philosophical dispute about what that something is—perhaps intentions of speakers, perhaps conventions within a p. 581

community of language users,

perhaps teleofunctional facts, perhaps causal connections to the 7

environment, and so on. Whatever these metasemantic facts are that give our words meaning, their role is metaphysical, not epistemological. For the most part, guring out what someone’s utterance means is something we do without consulting the metasemantics—we understand someone’s utterance of “Aristotle was fond of dogs” without engaging in historical investigation of the chain of usages lying behind their utterance. However, when we encounter di

culties in interpretation, turning to the metasemantics to work 8

out content can be one of the tools open to us.

One of our major themes in Cappelen and Dever (2021) is that we ought to attend to metasemantic questions when considering questions about the content and intelligibility of AI systems. In that book, we weren’t primarily concerned with metasemantics as a tool for guring out what was meant. Rather, the emphasis was on attending to the metasemantics in order to convince ourselves that, or to evaluate whether, our AI systems are indeed producing meaningful outcomes. In addition, considering the metasemantics of AIs can help us determine whether they are similar to us in their content production in a way that would allow us to use our standard epistemic methods of determining content among ourselves (methods that, again, don’t typically run through the metasemantic determination) to the AIs as well. Once we are confronted with the possibility of alien contents in AI systems, the epistemological stakes are raised, and we will need all the tools we can nd to help achieve understanding—or, following the discussion of the previous section, whatever analog of understanding is available. We suggested in that discussion that attention to the metasemantics of AIs could be useful either to help us gure out what mapping relation between alien and domestic contents provided the best replacement for translation or to help determine how we needed to position ourselves relative to AIs to integrate their content production with ours. So what might such attention to the metasemantics reveal?

3.1. Metasemantic Contingency It’s helpful before discussing the metasemantics of AI to remind ourselves that metasemantic facts are contingent facts, which depend on details of the language and the language community for their truth. Disputes between internalists and externalists are often framed as if one of these two positions is a necessary truth capturing some essential fact about the nature of content and communication. But surely, this isn’t right. Classic arguments for metasemantic externalism—in, for example, Kripke (1980) and Burge (1979)—start with particular contingent facts about our languages. We speak a language whose content is xed externally in part because we judge that name-involving utterances depend on the person at the origin of the causal chain of name uses, rather than on the person satisfying the cluster of descriptions entertained by the speaker, or because we are disposed to defer to experts when they correct our referential assumptions. If we behaved di erently—disregarded the causal history, declined to defer—the externalist p. 582

arguments would fail, and an internalist metasemantics would t

our language. Surely, we could have 9

behaved di erently, and so surely, the metasemantics could have been otherwise.

3.2. Varieties of Metasemantic Variability Once we acknowledge that there can be variability in the metasemantics, we can identify possible dimensions of that variability. We might distinguish:

•  Lexical variation in metasemantics: perhaps names have their contents xed through one mechanism, sentential connectives through another, moral vocabulary through a third. •  Syntactic variation in metasemantics: perhaps names as used in atomic sentences rely on one metasemantics, while names in attitude contexts rely on another. •  Linguistic variation in metasemantics: perhaps meanings in English are xed in one way, while meanings in C++ are xed in another way. •  Speaker variation in metasemantics: perhaps demonstratives as used by some speakers have their referents xed by speaker intentions, while demonstratives as used by other speakers have their referents xed by discourse coherence relations. These various dimensions of metasemantic variability are independent—in principle, any combination of them could be manifested. Given our interest in looking at the metasemantics of AIs in order to gain traction on problems of alien content, we will focus primarily on interspeaker metasemantic variation, but the other 10

categories of metasemantic variation deserve further investigation as well.

3.3. Why AI-Metasemantics Will Be (to Some Extent) Alien As we stressed in Cappelen and Dever (2021), it’s not hard to see that an adequate metasemantics for AI content producers would likely need to be di erent from an adequate metasemantics for content producers like us. The details depend on one’s preferred metasemantics, of course, but to rehearse a few examples: • If our metasemantics has it that meanings of utterances are (à la Grice (1989)) determined by the communicative intentions of speakers, but AI systems lack intentions of any sort, then we need a di erent metasemantic story for those AI systems (perhaps appealing to the deferred communicative intentions of the AI programmers?) • If our metasemantics has it that meanings of utterances are (Davidson-style) determined by what best p. 583

rationally mediates between perceptual inputs and action

outputs for speakers, but AI systems lack

perceptual faculties or capacities for action, then we need a di erent metasemantic story for those AI systems (perhaps replacing perception with data input and action with data output?) Our strategy in 2021 was to nd appropriate anthropocentric abstractions of existing metasemantic theories, so that both AI and human content production could be seen as instances of a more generalized metasemantic theory. Thus we emphasized the potentially important role of externalist metasemantic theories in thinking about AI content attribution, and at the same time stressed that existing externalist frameworks had their speci c externalisms shaped by thinking exclusively about content generation for 11

humans.

If, however, the AI systems become di erent enough from us, there may be no possibility of

nding a single metasemantic story explaining the production of content in both us and the AI systems. This is a familiar lesson from the consideration of multiple realizability. There may be no single grounding story that explains both why mad pain is pain and why Martian pain is pain (Lewis 1980). So we may need to confront, in addition to the possibility of alien contents, the possibility of alien metasemantics. Consider an example: When the predictive language system ChatGPT produces textual output, it produces output that has (let us suppose) some content. What then determines the content of ChatGPT outputs? Given that ChatGPT outputs have the same syntactic and morphological structure as strings in our English, one option is to transfer the English-language metasemantics to ChatGPT outputs (assuming that is possible)—perhaps standing linguistic conventions x the meanings of both our sentences and the LLM’s sentences. Note, however, that this transfer move isn’t mandatory. We could instead consider a fundamentally alien metasemantics for ChatGPT. For example, ChatGPT’s neural net

weightings are determined by extensive training against massive text databanks provided by the Common Crawl archiving service. Given the role that the Common Crawl database plays in the training of ChatGPT, maybe the Common Crawl archiving algorithms play a metasemantic role for ChatGPT as well, so that what ChatGPT means by “snow is white” is determined (perhaps in some way we can’t fully comprehend) by the 12

way in which Common Crawl extracts and stores text content from various websites.

3.4. When Is an Alien Metasemantics Too Alien? If alien metasemantics gets too alien, we might wonder whether it is metasemantics at all. Suppose Alex says that their pet rock right now is saying that it’s snowing. Beth suggests that this just can’t possibly be right. There’s nothing about the pet rock that would make it say something. It doesn’t have any beliefs or intentions. It’s not a participant in any conventions, so a conventionalist metasemantics won’t make it be saying anything. It’s not in a causal chain of uses, so a Kripkean externalist metasemantics won’t make it be p. 584

saying anything. It’s not properly interpretable as saying that it’s snowing, so an

interpretationalist

metasemantics won’t make it be saying anything. And so on—none of our metasemantic stories make it say anything at all. Alex says: Well, of course not. You’re trying to use human metasemantic stories on a rock. That’s just not the right way to go about it. Rocks come to have content in very di erent ways than the ways people come to have contents. Look, do you see how there’s this cluster of quartz crystals there on the upper left of the rock? That’s what makes it the case that it’s saying that it’s snowing. That’s because the metasemantics for rocks is that having quartz crystals in various places is what makes it the case that it’s saying things. Alex’s response is unconvincing. Although we should be open to alien metasemantic determination, surely not just any old collection of properties can give rise to content. So the proponent of an alien metasemantics has an explanatory debt: they need to make sure that their alien metasemantic story really is plausible as a metasemantic story. A mere variant on the pet rock story won’t discharge that explanatory debt. How can the debt be paid, then? One approach starts by looking at the kinds of things contents are, and then guring out what it would take to ground those things. An analogy: take some economic event-type, like a recession. We can ask for a grounding story for a recession (think of this on analogy with “metasemantics of a recession”). We want to know in virtue of what a recession is happening. And there could be lots of stories here—high unemployment, stagnant wage growth, high interest rates, and so on. Furthermore, we could consider the possibility of an “alien metasemantics of recession”—we could imagine creatures who didn’t have banks or jobs, and for whom recessions weren’t matters of high interest rates or high unemployment rates. But we couldn’t just appeal to any old feature of these creatures—it’s not going to make sense that for them, a recession is a situation in which they all need haircuts. What we need to do is think about what kind of thing a recession is—some kind of slowdown in economic growth. Then we can look at these creatures and ask what makes for economic growth for them. They don’t have banks, because they don’t have currency. Instead, it’s all barter for them—but we could then imagine that changes in the bartering practices amount to a recession for them. We can see that the bartering changes work and the haircut changes don’t work by thinking about the nature of the thing we’re trying to ground. Alien metasemantics can be approached in the same vein. Suppose that content is just truth conditions. Then whatever our metasemantics is, it needs to be the kind of thing that can make sense of giving rise to truth conditions. Human metasemantic stories do explain that—maybe they involve tracking di erences in reacting to di erent kinds of situations, and extracting truth conditions from the di erences in the

situation kinds. But the distribution of quartz crystals in the rock just isn’t the kind of thing that’s going to be able to explain di erences in truth conditions, so it can’t be a good metasemantic story for a truthconditional account of content. That response to the “pet rock” challenge depends on a shared notion of content. With mild versions of p. 585

alien contents, there could still be enough sharing to get this response going.

If the AI is alien only

insofar as it has predicates that can’t be translated into open sentences of ours, but familiar insofar as it makes truth-conditional claims by applying (alien) predicates to familiar objects, we might be able to leverage the shared truth-conditionality to recognize the AI’s alien metasemantic machinery as properly metasemantic. But when the alien contents become more thoroughly alien, this response will fail. When the contents are alien, then, we can’t look to the nature of content to convince us that a putative alien metasemantic story is plausibly metasemantic. Where do we look, then? Our suggestion: just as uncertainty about alien contents can be tackled by looking to the metasemantics, uncertainty about the alien metasemantics can be tackled by looking to the meta-metasemantics.

3.5. Meta-metasemantic Guidance to Metasemantics Consider a hierarchy of explanation. We start with facts about content. The name “Aristotle” has the semantic feature of picking out a certain individual. We then ask: Why does “Aristotle” have that feature (rather than some other semantic value)? The metasemantics answers this question: “Aristotle” refers to Aristotle because Aristotle lies at the origin of the causal chain of usages terminating in the current use of “Aristotle”. Then we ask again: Why does the causal chain determine the content of “Aristotle”, rather than some other feature (such as the descriptions associated by the speaker with the name “Aristotle”)? The 13

meta-metasemantics answers this question.

The meta-metasemantic options are less thoroughly mapped

than the metasemantic options. Here is an example of how to think about a meta-metasemantic option: Informational-Epistemic Role Meta-metasemantics: It’s the causal chain that determines the meaning rather than the associated descriptions because we use names as tools for transmitting information and accumulating knowledge about distant objects, whose properties we track only imperfectly. A causal chain metasemantics gives contents to those names that let them play those informational and epistemic roles better than a descriptive metasemantics would. There are many other meta-metasemantic options, but for now, we’ll use this as an illustration of how it can guide us in understanding when something is a metasemantic mechanism.

3.6. An Illustration of Meta-metasemantic Guidance Suppose we encounter a sophisticated AI system that has been designed to help guide the global economy. Unfortunately, it is very obscure to us. We are told the following: Potential Alien Content + Alien Metasemantics: (i) The Economic AI outputs have as their contents p. 586

alien contents that are probability distributions on sets of strategies

for playing certain games. (ii)

These alien contents are plausible because we can see how they would be grounded in simulation games that the AI is playing both internally and with a collection of other game-playing AI systems, combined with facts about the way in which certain major stock-market indices typically react to outputs of the AI. Suppose we agree that these facts about simulation games plausibly ground a relation between the Economic AI and the probability distributions. This raises the question of why we should think of these

probability distributions on sets of strategies as contents—that is, why we should think of the grounding story as a metasemantic grounding story. In answering this, we appeal to the kind of meta-metasemantic considerations outlined above: Meta-metasemantic guidance: The point of contents is to allow for rational assessment of our action and of our internal transition from one content-bearing state to another. The Economic AI doesn’t perform human actions or engage in human-style belief updating, but based on its design, we know that it is to be assessed based on how well it guides our economy and on how well it internally con gures itself to spot opportunities for guiding our economy. The (alien) contents attributed on the basis of the Economic AI’s internal and external game-playing and of the stock-index reactions, the suggestion goes, make best sense of the AI as assessable on those grounds, so that attribution basis gives the right metasemantics for the AI. The point here is that looking to the point of contents—the role they play in our larger theorizing about language, action, cognition, and so on—gives us a sketch of a meta-metasemantics. The metasemantics can ground contents, rather than something else, because being so grounded makes sense of being able to play the right theoretical roles. Some plausible anthropocentric abstraction then allows us to take a humancentered meta-metasemantics and see how to adapt it to produce a meta-metasemantics, not thoroughly alien, that covers both us and the AI system.

3.7. Meta-metasemantics and Shared Ways of Life If a lion could talk, Wittgenstein tells us, we could not understand him (1953, 223). Its form of life, we take Wittgenstein to be suggesting, is too di erent from ours for there to be common ground for communication. We don’t know quite what a form of life is, but we can o er our own version of this thought: From Ways of Life to Meta-metasemantics: If the meta-metasemantics is going to guide access to alien metasemantics and alien contents, then the theoretical roles for contents in the AI need to be p. 587

similar enough to the theoretical roles for contents for

us. But the theoretical roles for contents

for us depend on a bunch of contingencies about the sorts of creatures we are and the sorts of things we do. That we have perceptual faculties, that we have interests in tracking speci c objects over time and space, that we trade information among ourselves, that our actions are explained by the quality of our evidential relations to the world, and so on—all of these things x the theoretical role of contents, and all of these things could have been di erent. We’re actually reasonably optimistic about hypothetical lion talk—lions seem a lot like us in many of the relevant ways. But there’s a range of cases here. Even more like us are our pets—they share much of our physiology, but also some of our sociology as well, unlike lions. Much less like us, for example, is the oceanic life form of Stanislaw Lem’s Solaris, which doesn’t obviously have anything at all similar to our goals or world-tracking mental states. The utility of looking to the meta-metasemantics will vary with the similarity of the alien content-bearer to us, and di erent dimensions of similarity and di erence can each bring their own challenges. The intriguing thing about AI systems, then, is that they have a unique combination of similarities to, and di erences from, us. In some ways, even the most sophisticated AI is much less like us than is a lion. The AI (plausibly) doesn’t have beliefs and desires, appetites and fears, or direct control over a physical body. But 14

in other ways, the AI is much more like us than even our pets are. It, like us, contains neural nets.

The AI

has epistemic resources that are closely analogous to what we have: they are trained on, e.g. visual inputs; they discriminate various properties from each other; they use these ‘perceptual’ capacities to classify

situations into those that, say, have stop signs in them and those that don’t. These discriminatory capacities 15

are familiar to us.

After all, the systems were made to do things that can be useful to us: for example,

distinguish ducks from not-ducks, or malignant tumors from benign ones, and classify people according to their credit worthiness. So at the core of the systems, we nd familiar abilities, based on mechanisms (neural nets) that we share with them. In short: we say that these systems are fundamentally di erent, but not alien. Hopefully, then, these similarities are enough to enable us to nd theoretical roles for AI content. More speci cally, hopefully we can use a familiar meta-metasemantics to make sense of an alien metasemantics, and then make use of that alien metasemantics to gure out how to map alien contents to our own contents or how to integrate ourselves into the alien contents of the AI. The de-anthropocentrizing may be di

cult.

Perhaps knowledge acquisition features in the theoretical role of content, but perhaps AI systems aren’t the kind of thing that can know. In our 2021, we suggested one way of performing the anthropocentric abstraction here: hold on to the role of knowledge, but shift from the language producers as knowing to the language consumers as knowing. Another option is to generalize our epistemic taxonomy, to nd some more general evidence-tracking state of which knowledge is one instance (the instance found in us) and some other state is another instance (the one found in the AIs).

p. 588

3.8. When Thereʼs No Common Meta-metasemantics: Meta-metametasemantics to the Rescue! Hope may eventually give out. Maybe the theoretical role of content in the end depends on categories that have no analog in AI systems. Suppose contents are those things that explain action and rational accessibility as evidence managers. Certain kinds of AIs have nothing analogous to action or to evidence management. What then? Of course, eventually we might become convinced that the entire ecosystem of content-based talk is inapplicable to AIs that di er enough from us. We’re not there, however, just because we can’t nd a common meta-metasemantic level. We can continue to ascend the meta-ladder. When appeal to a shared meta-metasemantics gives out, we can turn to the meta-meta-metasematics. We won’t examine this option in any detail here—it’s hard enough to think about what makes it the case that certain features make it the case that a speaker means particular things by what they say, without also trying to think about what makes it the case that those things make it the case that certain features make it the case that a speaker means particular things by what they say. We can perhaps get started by casting the “theoretical role” net more broadly—perhaps, just as we looked to the theoretical role of content in explaining action and rational assessment in order to get a grip on meta-metasemantics, we can look to the theoretical roles of action and rational assessment in order to get a grip on meta-meta-metasemantics. There is no in-principle limit to how far we might have to ascend the meta-ladder to nd a point of adequate commonality to begin a process of communicative bridging to su

ciently odd AIs. This can be a

dizzying thought. It’s natural to think that there must be a xed point somewhere—that what we want is communication, which requires shared content, and that if we can’t have that, there’s nothing to have; or, that what we want is knowledge, which requires something communication-like (but not necessarily communication in the sense of shared content), and that if we can’t have that, there’s nothing to have. What would it be like to engage with beings for whom there was no xed point at which the question of how to engage was set?

4. Concluding Applications: Explainable AI and Existential Risk The issues we’ve discussed here can easily seem abstract and of no relevance to the real and pressing practical, social, and political issues raised by the issues that AI presents us with. Nothing could be further from the truth. These are foundational issues with very direct practical implications. We provide two illustrations: Existential Risk and Explainable AI.

p. 589

4.1. Existential Risk There has been a great deal of literature on “existential risk” concerns about AI (see Bostrom 2014 for a popular recent discussion). These concerns often center on what is called The Alignment Problem: how to get the goals of AI systems su

ciently aligned with ours, and how we might design AIs to guarantee a

comforting level of alignment. We have preferred to focus on a collection of what we take to be underexplored semantic and metasemantic questions about AI systems, but the two topics intersect when we start thinking about alien content. Here is a simple version of the existential risk issue: we might want some version of Asimov’s First Law of Robotics ( rst presented in his 1942 story “Runaround”) incorporated into any AI systems we produce: Asimov’s First Law, AI version: No (AI system) may injure a human being, or by inaction allow a human being to come to harm Implementing exactly that law in an AI system requires shared content—if the AI’s semantic resources don’t include the concepts of injury or harm, then we won’t be able to tell it what to do. If an AI has alien contents, then we can’t share that content with the AI (because Asimov’s First Law is formulated using human contents). One immediate corollary is that settling the issue of alien versus non-alien content is a requirement on even thinking about the alignment issue. Another corollary of the above discussion is that the presence of alien contents doesn’t mean giving up on the First Law of Robotics. Suppose we’ve settled that the AI can’t share our concept of injury or risk. How do we recognize the alien concepts as being suitably related to harm and injury? One option we have suggested above is that we can look to the metasemantics or the meta-metasematics. We check to see if the alien concept is metasemantically determined by facts related to human ourishing; we consider what theoretical role the distinction between ourishing and non- ourishing lives plays for us. Pursued correctly, such discussions bring together questions of ethics, content, and meta-content into a single research project.

4.2. Explainable AI The aim of XAI is to create AI systems that are interpretable by us, that produce decisions that come with comprehensible explanations, that use concepts that we can understand, and that we can talk to in the way that we can engage with each other. The importance of this is highlighted by recent legislation that makes it a requirement that decisions made by (or with the help of) neural nets come with an explanation of the decision. Thus—and without getting too deep into the weeds of contemporary European Union data p. 590

16

protection law —the EU’s enshrines a ‘right to explanation’. One has the right to “meaningful 17

information about the logic [of the algorithm] involved.”

That means that those subject to neural net-

based decision-making are owed an explanation of how a decision a ecting them was arrived at. This is obviously a very tall order when we don’t even know whether an AI has the same kind of contents as we humans have. Most of the discussion of explainable AI proceeds on the assumption that we are dealing with a system that generates outputs with human-type contents. Even then the requirement that an explanation be available is hard to meet, but the challenge gets even harder once we recognize the real possibility that the relevant outputs could contain alien contents generated by alien metasemantic mechanisms. We then need procedures for recognizing these contents as contents, and we need ways to relate to those alien contents. This chapter has provided proposals for how to proceed, but it should be obvious that we have, at best, scratched the surface of some extremely hard questions, and that there’s much work to be done.

Notes

p. 591

1.

This was certainly true at the time of our original writing (several years ago), and it remains true now. Peruse, for example, Microso ʼs “Sparks of Artificial General Intelligence: Early Experiments with AGI” (Bubeck et al. 2023), or OpenAIʼs “GPT-4 Technical Report” (OpenAI 2023), and you will see no reference to philosophy. Influential work adjacent to computer science (Bender and Koller 2020, Piantadosi and Hill 2022), even on philosophical questions about the nature of language use and meaning, remains more or less disconnected from the philosophical tradition. What has changed is that increasingly philosophers are writing about AI, but as yet it seems our viewpoints are not being added to the conversation.

2.

See e.g. Carston (2001); Sperber and Wilson (1986); Bezuidenhout (2002).

3.

See e.g. Yablo (2014); Fine (2016); and the literature that it has spawned.

4.

See Gibbard (1990) and Schroeder (2008) for discussion and some references to the vast literature.

5.

These sorts of bridging by content-mapping strategies are related to, but, importantly, distinct from, the paraphrase strategies that get used in ideal language metametaphysics. When the mereological nihilist paraphrases the ordinary language “there is a chair” as “there are some simples arranged chairwise”, he finds a suitable analog for the non-nihilist content among the contents he accepts, much as we are suggesting that the representationalist finds a suitable analog for the expressivist content among her own contents, or (below) that we find suitable analogs to the alien AI contents among our own contents. However, the mereologist nihilist and non-nihilist share contents—the nihilist has the content “there is a chair”, but thinks that that content is false, and so wants to map onto another content (one also available to the nonnihilist) that she thinks is true. And the paraphrase has the stable goal of finding a paraphrase that matches the assertability conditions of the target sentence; whereas the kind of bridging maps we are considering can have highly variable success conditions. (Note that alien contents need not have assertability conditions that we can express any more than they have truth conditions that we can express, and may not have assertability conditions at all, because they may not be used in a practice of assertion.)

6.

Lewis (1979) is the locus classicus which spawned a large and ongoing literature.

7.

Weʼre terminologically following Robert Stalnaker here, who says that metasemantic questions are, “about what the facts are that give expressions their semantic values, or more generally, about what makes it the case that the language spoken by a particular individual or community has a particular descriptive semantics” (1997: 535). Stalnakerʼs paradigm of a metasemantic theory is Krikpeʼs causal theory of proper names.

8.

Of course, itʼs contentious what the right metasemantic view is, so di iculties in interpretation may simply spill over into di iculties in selecting a metasemantic framework. And even when we agree on the metasemantics, the method of metasemantic determination may not be scrutable by us.

9.

There are complicated questions here about whether our various possible linguistic behaviors would be constitutive of an internalist or externalist metasemantics or merely evidence for an internalist or externalist metasemantics. To think the first would plausibly be to assume a background internalism about the metasemantics—an internalist metametasemantics. We return to these higher-order questions below.

10.

Maybe we should never talk about variation in metasemantics, but instead look for a generalized or parameterized metasemantics that spans putative areas of variation? So, if weʼre initially driven to think that names have their contents determined in one way, and connectives have their contents determined in another way, we should, instead, take the data to suggest that there is a single metasemantic framework that tells us that (e.g.) being a name combined with having a particular causal history grounds one content, while being a connective combined with having a particular inferential role grounds another content. Itʼs not clear how much beyond terminological boundary-drawing is at issue here. In some sense, we could look for a single grounding story for absolutely everything, not just contents, that maps each “lowerlevel” collection of facts to its corresponding grounded “higher-level” fact. The question is just whether theoretically interesting generalizations about grounding in certain areas thereby get lost or obscured. Talk of multiple realizability can, in this context, be interpreted as recognition of important di erences in grounding generalizations within the same general sort of stu .

11.

AI systems have the potential to challenge, not only the way we think about external factors playing a role in content determination, but also the way we draw the line between internalist and externalist metasemantic framework. A conception of “internal duplicate” that is (perhaps) well-understood in the human case may become much more elusive when we extend it to the AI case. Are two copies of the same neural net implemented on di erent computers with (for example) di erent computational power internal duplicates? If an AI system regularly accesses other computational tools (database lookup tables, etc.), do those other tools count as “internal” or “external” to the AI? Questions of this sort are already complicated for us by the extended mind hypothesis, but they become particularly obscure when extended to AIs.

12.

Note that alien contents and alien metasemantics are orthogonal issues: An AI could have an alien metasemantics, but via those alien metasemantic mechanisms end up producing contents that are our familiar contents. Similarly, an AI could have alien contents produced by a familiar metasemantics. And, of course, an AI could have both—alien contents generated by an alien metasemantics.

13.

Meta-metasemantics can be connected to talk of grounding in two ways: (i) We can take the metametasemantics to ground the grounding claim provided by the metasematics: and (ii) we can take the meta-metasemantics to be further elaborating the grounding claim given by the metasemantics, by giving further conditions that are le implicit in the original grounding claim or by specifying enabling conditions (e.g. Dancy 2004) that allow the grounds to ground. So, roughly, within a grounding context the metametasemantic facts can be accommodated by moving either upward (to the grounding of the grounding) or outward (to a broader perspective on what is doing the grounding). We remain neutral here on the best way to connect our metametasemantic talk to the grounding literature.)

14.

It also includes features (e.g. the use of neurotransmitters and hormones) that artificial neural networks lack. So we donʼt describe a brain anywhere near exhaustively by calling it a neural net. A helpful discussion here is Buckner and Garson (2019), but the key point is that thereʼs at least an important structural similarity but also an important dissimilarity between human brains and neural nets, and that su ices to make our point.

15.

Plausibly, the point carries over to, for example, credit-assessing systems, or systems concerned with recidivism. Although they neednʼt take sensory inputs, it nevertheless processes the same sort of information about people we do when we assessing people.

16.

See Goodman and Flaxman (2017) for discussion

17.

Articles 13 and 14 of the GDPR, quoted in Goodman and Flaxman (2017: 6).

p. 592

References Bender, E. M., and Koller, A. (2020, July), Climbing towards NLU: On Meaning, Form, and Understanding in the Age of Data. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 5185–5198. Google Scholar Google Preview WorldCat COPAC Bezuidenhout, Anne (2002), ʻTruth-Conditional Pragmaticsʼ, Philosophical Perspectives 16:105–134. Google Scholar WorldCat Bostrom, Nick (2014), Superintelligence: Paths, Dangers, Strategies (Oxford, UK: Oxford University Press). Google Scholar Google Preview WorldCat COPAC Bubeck, Sébastien, Varun Chandrasekaran, Ronen Eldan, Johannes Gehrke, Eric Horvitz, Ece Kamar, Peter Lee et al. (2023) ʻSparks of Artificial General Intelligence: Early Experiments with GPT-4.ʼ arXiv preprint arXiv:2303.12712. Google Scholar Google Preview WorldCat COPAC Buckner, Cameron, and Garson, James (2019), ʻConnectionismʼ, Stanford Encyclopedia of Philosophy (Fall 2019 Edition), Edward N. Zalta (ed.), https://plato.stanford.edu/archives/fall2019/entries/connectionism/. Google Scholar Google Preview WorldCat COPAC Burge, Tyler (1979), ʻIndividualism and the Mentalʼ, Midwest Studies in Philosophy 4 (1): 73–122. Google Scholar WorldCat Cappelen, Herman, and Dever, Josh (2021), Making AI Intelligible (Oxford: Oxford University Press). Google Scholar Google Preview WorldCat COPAC Cappelen, Herman, and Dever, Josh (forthcoming), A Defense of Artificial Intelligences, OUP. Google Scholar Google Preview WorldCat COPAC Carston, R. (2001), ʻExplicature and Semanticsʼ, in S. Davis and B. Gillon, eds., Semantics: A Reader (Oxford: Oxford University Press), 817–846. Google Scholar Google Preview WorldCat COPAC Dancy, Jonathan (2004), Ethics without Principles (Oxford: Oxford University Press). Google Scholar Google Preview WorldCat COPAC Davidson, Donald (1973), ʻRadical Interpretationʼ, Dialectica 27 (1): 314–328. Google Scholar WorldCat Fine, Kit (2016), ʻAngellic Contentʼ, Journal of Philosophical Logic 45 (2): 199–226. Google Scholar WorldCat Gibbard, Allan (1990), Wise Choices, Apt Feelings: A Theory of Normative Judgment (Cambridge, MA: Harvard University Press). Google Scholar Google Preview WorldCat COPAC Grice, H. P. (1989), Studies in the Way of Words (Cambridge, MA: Harvard University Press). Google Scholar Google Preview WorldCat COPAC Kripke, Saul A. (1980), Naming and Necessity (Cambridge, MA: Harvard University Press). Google Scholar Google Preview WorldCat COPAC Lewis, David K. (1969), Convention: A Philosophical Study (Cambridge, MA: Harvard University Press). Google Scholar Google Preview WorldCat COPAC

p. 593

Lewis, David (1979), ʻAttitudes De Dicto and De Seʼ, Philosophical Review 88 (4): 513–543. Google Scholar WorldCat Lewis, David (1980), ʻMad Pain and Martian Painʼ, in Ned Block, ed., Readings in the Philosophy of Psychology (Cambridge, MA: Harvard University Press), 216–222. Google Scholar Google Preview WorldCat COPAC OpenAI (2023), “GPT-4 Technical Report”, https://arxiv.org/abs/2303.08774 WorldCat Piantadosi, S. T., & Hill, F. (2022). ʻMeaning without Reference in Large Language Modelsʼ. arXiv preprint arXiv:2208.02957. Google Scholar Google Preview WorldCat COPAC Schroeder, Mark (2008), Being For: Evaluating the Semantic Program of Expressivism (Oxford: Oxford University Press). Google Scholar Google Preview WorldCat COPAC Sperber, D., and D. Wilson (1986), Relevance (Oxford: Blackwell). Google Scholar Google Preview WorldCat COPAC Stalnaker, Robert C. (1997), ʻReference and Necessityʼ, in Bob Hale and Crispin Wright, eds., A Companion to the Philosophy of Language (Oxford: Blackwell), 902–919. Google Scholar Google Preview WorldCat COPAC Yablo, Stephen (2014), Aboutness (Princeton, NJ: Princeton University Press). Google Scholar Google Preview WorldCat COPAC Wittgenstein, Ludwig (1953), Philosophical Investigations (Oxford: Basil-Blackwell). Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

p. 594

26 Semantic Change in the Language of Technology  Daian Flórez https://doi.org/10.1093/oxfordhb/9780192844118.013.48 Published: 22 May 2024

Pages 594–612

Abstract problem of semantic change in science has been widely examined at the intersection between philosophy of language and philosophy of science. Philosophers have been concerned with the change of meaning of theoretical terms in scienti c theories, but the interesting question is whether we can identify a similar phenomenon in technological knowledge, since admitting that there are semantic changes in technology might violate the common sense intuition that technology is the totality of artifacts and as such, lacks semantic properties. However, if technology is as complex a phenomenon as science and there are also technological theories, it becomes reasonable to expect semantic changes in its domain. This view can be supported by the thesis of the incommensurability of theories Kuhn used to argue that there is no common language to enable the translation of terms ordinarily used in scienti c theories. Kuhn did not extend the scope of the incommensurability relation to the domain of technology, though he analyzed some technological cases to illustrate the nature of this semantic phenomenon. Controversial as it might look, this chapter shall show that the history of technology provides enough evidence for the thesis that there are profound semantic changes in technology or, in other words, that there is technological incommensurability. The chapter explains how this thesis works in technology domain by means of two historical cases: the substitution of the contact theory of the voltaic battery by the so-called chemical theory and the transit of the material theory of heat to thermodynamics.

Keywords: technological Incommensurability, semantic change, Kuhn, voltaic battery, chemical theory, material theory of heat, thermodynamics Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction Though philosophers have been concerned for some time with meaning change for theoretical terms in scienti c theories, the question of whether we can identify a similar phenomenon in the language of technology has drawn far less attention. I will argue, just the same, that meaning change permeates the 1

language of technology by invoking and defending a thesis of incommensurability.

As is well known, Kuhn argued that there is no common language to enable the translation of scienti c terms. Kuhn did not extend the scope of incommensurability to include the language of technology, although he did analyze some technological cases in order to examine the nature of this semantic phenomenon. In this chapter, I shall show that the history of technology provides strong evidence for the thesis that there are profound semantic changes in the language of technology, or, in other words, that there is technological incommensurability. I shall examine this thesis and illustrate how it works by means of two historical cases: the rst comes from the substitution of the contact theory of the voltaic battery by socalled chemical theory; the second comes from the transit of the material theory of heat to thermodynamics. Why does the question of whether there can be semantic change in technology matter? Several reasons render the examination of this problem important. The rst comes from a widely held image according to which technology is the totality of artifacts. The naive acceptance of such a view would leave us with no room to examine the semantic problems that, paradigmatically, take place in other areas of philosophy. Nevertheless, a careful consideration of technology as an activity shows that technological theories (whose formulation requires a language) are also a product of technology. Taking this for granted, it becomes necessary, along with analysis of the very structure of technological theories, to pay attention to the p. 595

languages used to formulate these theories and

to the problem of semantic change in the language of

technology. Furthermore, just as semantic change in science raises complex problems for philosophers of language (and of science), such as those related to the stability or change of meaning (or the relationship between the thesis of semantic change and scienti c progress), similar issues need to be investigated in the language of technological theories. In addition, as we will see, by considering the historical case of thermodynamics, it seems logically admissible that there is incommensurability—the impossibility of inter-translatability between the languages of competing theories—what amounts to the impossibility of translating the language of thermodynamics into the language of the preceding technological theory (the material theory of heat). All of this shows that incommensurability is a characteristic feature, not only of the languages of scienti c theories but also of the languages of technology, as well as of their possible interactions.

2. Inter-theoretic Incommensurability 2

Kuhn gave two versions (without substantial di erences) of the thesis of incommensurability. The rst (formulated in the sixties and known as the ‘semantic version’) appeared in The Structure of Scienti c 3

Revolutions (1962/1996; henceforth SScR). The second appeared in James Conant and John Haugland’s The Road since Structure (Kuhn 2000), a collection of Kuhn’s papers published after his death. This one is usually 4

called the “taxonomical version.” Here, I will discuss only the semantic version. Kuhn introduces the word ‘incommensurability’ in a section of SScR entitled “The Nature and Necessity of Scienti c Revolutions” to illustrate the nature and scope of scienti c revolutions from which revolutionary 5

changes (i.e., the substitution of one paradigm for another ) is a consequence. After explaining that scienti c revolutions “are here taken to be those non-cumulative developmental episodes in which an older paradigm is replaced in whole or in part by an incompatible new one” (Kuhn 1962/1996, 92), he anticipates, prima facie, four objections against the view that there are dramatic meaning changes in theoretical

language due to scienti c revolutions and responds to one of them (which I shall call the objection of intertheoretic reduction) by means of the thesis of incommensurability. Schematically, the objection runs as follows: (P1) If Newton’s theory can be derived from Einstein’s relativity theory (supposing the relative speeds of the studied bodies are lower than the speed of light), then there are no dramatic meaning changes in science, as required by scienti c revolutions. (P2) Newton’s theory can be derived from Einstein’s theory (supposing the relative speeds of the studied bodies are lower than the speed of light). (C) Therefore, there are no dramatic semantic changes in the language of science, as claimed by the supporters of the view that there are scienti c revolutions. p. 596

To refute (P1)’s antecedent, Kuhn argues that it is impossible to deduce Newton’s theory from Einstein’s theory, as we can see in the following: (S1) Suppose the theory of relativity includes statements: E1, E2 … En (these statements include variables and parameters like spatial position, time, and rest mass, among others) (S2) Suppose Newton’s theory includes statements: N1, N2 … Nn (which also include variables and parameters like spatial position, time, and rest mass, among others) (S3) We cannot derive N1 from E1 and N2 from E2, because the physical references are not identical. While Newtonian mass is conserved, Einsteinian mass is transformed. (Kuhn 1962/1996, 101–102) According to Kuhn, what prevents the inter-theoretic reduction from Einstein to Newton is the incommensurability of theories. The argument is simple: a conditio sine qua non for inter-theoretic reduction is the preservation of the meanings of the scienti c theories. Given that this does not obtain—due to incommensurability—inter-theoretic reduction becomes logically impossible; in other words, we cannot derive Newton’s theory from Einstein’s theory because they are incommensurable. This rst version of the thesis, introduced by Kuhn in SScR, is usually called the semantic version since incommensurability is tightly tied to changes in the meaning and referent of the terms of scienti c theories. Schematically: (D1) Theories T1 and T2 are incommensurable i the terms in the language of T1 are not semantically equivalent to terms in the language of T2. Based on this de nition, it becomes obvious that the thesis of incommensurability is of a linguistic nature. Under which conditions—necessary and su

cient—are two scienti c theories incommensurable? Given

that incommensurability is a dyadic relation, i.e., a relation between pairs of successive theories separated by a scienti c revolution, we can say that, strictly speaking, theories T1 and T2 are incommensurable i : (i) T1 and T2 are successive theories separated by a scienti c revolution. (ii) The terms in T1 are not semantically equivalent to terms in T2, hence, they are not intertranslatable. This version of the thesis is tied to the failure of full translation between theories. Kuhn understands by untranslatable the impossibility of applying a mechanism that substitutes words or sequences of words in a language for words or sequences of words in another language, warranting interchangeability salva 6

veritate. In other words: (iii) There is no common language (Lc) that renders possible a translation of the language L1 of T1 into the language L2 of T2.

p. 597

More clearly stated, translation is a mechanism that systematically substitutes words and expressions in one language by words and expressions in another. Its success or failure lies in (a) the preservation of meaning and reference of the words involved, and (b) their interchangeability salva veritate: (a) is a conditio sine qua non of (b); however, since there is preservation of neither the meaning nor the reference of the terms, due to incommensurability, the failure of translation is inevitable. (iv) As a consequence of (ii) and (iii), there is no a logical criterion, i.e., no algorithm of rational decision 7

that might help us to choose between T1 and T2.

(v) As a consequence of (i), (ii), and (iii), there is a change in the metaphysical, ontological and 8

methodological commitments of scienti c communities.

(vi) As a consequence of (i), (ii), and (iii), there is perceptual incommensurability, i.e., T1 and T2 are perceptually incommensurable i

9

supporters of T1 and supporters of T2 see the world di erently.

After a scienti c revolution, there is a change in world view or a perceptual change derived from theoretical commitments subscribed to by members of scienti c communities. This can be illustrated by taking into account the way in which two observers can simultaneously see the same object in di erent ways: “in order to appreciate the philosophical import of these observations it becomes crucial that both observers recognize that they are observing a single object in di erent ways” (Brown 1984, 75; my emphasis). One example that helps us to deploy the thesis that a single object can be observed in di erent ways comes from Gestalt change. Consider gures like the old-young-woman or the face-vase, which appear to a single observer in any of their two forms. Reversible gures, as they are also called, provide a clear instance of a situation in which a single observer sees two di erent things whilst the retinal stimulus is only one; e.g., the p. 598

famous gure of the duck-rabbit (Figure 26.1). I will see a duck or a rabbit according to the

angular

direction of my attention: if I watch from right to left, I will see a duck; if I watch from left to right, I will see a rabbit.

Figure 26.1. The duck-rabbit. Besides appealing to the reversible gures of Gestalt to support the thesis that there are perceptive di erences caused by revolutionary changes, Kuhn provides some historical examples (e.g., Lavoisier saw oxygen, where Priestley saw dephlogistized air) and suggests an imaginary meeting between Aristotle and Galileo to discuss the perception of oscillating bodies held by a thread, cord, or cable: in this situation, the former will “observe” an object in an impaired and tortuous fall toward its place of natural rest; the latter will “observe” a pendulum, which exempli es in a quite illustrative way a case of free fall, reducible to

regularities that can be mathematically expressed. At the same time, Kuhn provides several historical arguments to defend the thesis of incommensurability: the transit (i) from Aristotelian to Newtonian physics, (ii) from contact theory to the chemical theory of the voltaic pile, and (iii) from the phlogiston theory of combustion to Lavoisier’s chemistry. I will examine (i) in order to illustrate the profound semantic changes that occur after a scienti c revolution, and I will deal with (ii) below in order to establish how this historical case supports, particularly, what I call here the semantic thesis of technological incommensurability. Based on the rst historical episode, Kuhn argues that when we compare Aristotelian with Newtonian dynamics, we notice that although the term “movement” is used in both theories, it refers to quite di erent phenomena in each. By “movement” Aristotle’s theory refers to a general change in such a way that it includes di erent varieties of change (increase, decrease, growth, aging) that also include, besides change of place—typical of the local movement of objects in space—phenomena, such as the transformation of an acorn into an oak, or varieties, such as the passage of sickness to health, and vice versa. In addition, for Aristotle, local movement is a qualitative change that includes di erent subcategories of movement: growth or intensity alterations such as the heating of an iron bar, as well as the changes that take place by the transformation of properties, which can be appropriately classi ed under the concept of process. By 10

contrast, for Newton, “movement” refers only to a change of state.

The argument Kuhn gives to explain

these conceptual di erences shows that scienti c progress does not lie in the incorporation of a theory with less explanatory power into another with more explanatory power. In other words, progress does not lie in the fact that a theory like Aristotelian dynamics is subsumed by another, like Newtonian dynamics, since in Aristotelian physics the term “movement” designates a wider and more varied number of changes than its counterpart term in Newtonian physics. In addition, Kuhn contends that the meaning of a scienti c term is a holistic feature, i.e., the term’s meaning depends on the meanings of other terms. So, the meaning of “movement” is derived from postulates of Aristotelian cosmology, as well as from the way it relates to other terms in his physics, inter alia, “matter,” “position,” or “place.” Accordingly, if the world is nite along the lines of Aristotelian cosmology, changes in nature obey both qualitative variations and changes in spatial position (left–right, front–back, up–down) or regarding the objects’ natural place according to the well-known doctrine of p. 599

natural and nonnatural places of rest and of natural or violent movement.

This set of principles explains

why Aristotle believes that matter is not dispensable from his nite view of the world and why, as a partisan of the plenum, he compares a vacuum to the idea of a “squared circle” (Kuhn 2000, 19). All of this boils down to the conclusion that, since there is no semantic equivalence between “movement” in Aristotle’s dynamics and “movement” in Newton’s dynamics, both are incommensurable. 11

Once the thesis on the incommensurability of theories was formulated, several objections were raised.

The

gist of these arguments might convince us it is foolish to claim there is incommensurability, not only in the language of science, but also in the language of technology. However, there are good reasons to motivate technological incommensurability. The rst is that one of Kuhn’s own historical arguments for intertheoretic incommensurability comes precisely from the domain of technology—namely, the transit from the contact theory of the pile to chemical theory. The second is that the history of technology provides additional evidence for the thesis, as I show in comments on the transit (or substitution) of Black-Watt’s material theory of heat to thermodynamics.

3. Technological Incommensurability Maintaining incommensurability in the language of technology might provoke an even more passionate rejection than we usually get from proponents of incommensurability in the language of science, for a twofold reason. Firstly, technology is currently identi ed with the totality of artifacts that result from 12

technological actions and processes (Kline 2003)

in such a way that incommensurability as a semantic

thesis is odd to attribute to—or predicate of—objects like artifacts. Secondly, according to a generalized view, while the business of science is theorizing, the business of technology is the creation of artifacts. Still, it is not di

cult to show that theories are also extensively formulated in the language of technology. Before

considering how this can be, let us consider the argument of those who might resist acknowledging semantic changes in the language of technology (or technological incommensurability), which I propose to call the argument from common sense: (P1) If artifacts are the archetypical products of technology (and artifacts qua material objects lack semantic properties, then semantic change cannot be attributed to technology; i.e., there is no technological incommensurability. (P2) Artifacts are the archetypical products of technology. (C) Therefore, semantic change cannot be attributed to technology, i.e., there is no technological incommensurability. Obviously, this argument is awed since it identi es technology with the sum of artifacts that are the result 13

of a set of complex actions. p. 600

Although it is true that artifacts constitute our immediate encounter with

technology, technology is much more complex

than philosophers have considered. As a matter of fact,

technology implies conventional practices such as designing—running experiments, calculating, simulating—and technological achievements such as artifacts (including the design of conceptual artifacts like software programs, or programming languages). This argument mistakenly identi es science simpliciter with laws of scienti c theories, since, regardless of whether we count the formulation of laws and theories among the most noteworthy products—or results—of scienti c activity, science is much more. As a complex activity, science also involves certain practical activities (e.g., contrasting, measuring, calculating, experimenting; cf. Díez and Moulines 1999). The mistake identi able in the previous argument seems to be a fallacy of composition, in which a property of the parts is transferred to the whole, or an attribute of some members of a group is transferred to the group as a whole; for instance, if each part of this machine is perfect, then the machine (as a whole) is perfect. This example shows how the parts’ properties are illegitimately transferred to the whole to conclude that the machine is perfect. The commonsense perspective argument to deny semantic change in technology commits the same fallacy, since it transfers a particularity of one of the products of technology —namely, the impossibility of attributing semantic properties (like being incommensurable) to objects (in this case artifacts) to the whole—and through this move derives the conclusion that it is always erroneous to attribute incommensurability to technology. An additional argument for semantic change in the language of technology derives from the formulation of technological theories.

3.1. Argument from Technological Theories (P1) If there are technological theories, we can identify semantic changes similar to those that occur in scienti c theories. (P2) There are technological theories. [C] Therefore, we can identify semantic changes similar to those that occur in scienti c theories. To back up the previous contention, we can add that, for example, in the domain of structural engineering (a branch of civil engineering that deals with the structural design applicable to constructions like buildings, bridges, and retaining walls, among others) theories like the beam theory and the plate or metal sheets theory have been formulated. As is known, the structural engineer should pursue the design of safe structures. The success of such a task has been possible because physicists and engineers have successfully accomplished a detailed theoretical analysis of the behavior of those structures. One of the most renowned theories of structural analysis is the classic theory of beams, formulated in the eighteenth century by the mathematicians and physicists p. 601

14

Leonhard Euler and Daniel Bernoulli.

Beam theory plays an important role in structural analysis,

since

it provides the engineer with a simple tool to analyze very di erent structures. The theory is customarily used in the predesign stage and gives valuable information about the behavior of structures (cf. Bauchau and Craig 2009). By using beam theory, it becomes possible to calculate, not only the stress that a beam has to undergo, but also its displacements. These calculi are important, since, based on them, we can examine the resistance of 15

the materials used in the construction.

But beam theory’s use is not restricted to civil engineering, since

many parts of several machines—such as lever arms, axis, etc.—are beam-like structures. Some aeronautical structures, such as the wings and fuselage of a plane, can also be treated as beams of a thin wall 16

(cf. Bauchau and Craig, 2009).

Based on the foregoing, if a conditio sine qua non for semantic change in the language of technology is that there be theories in this domain, and since, notably, there are theories in some of the subareas of technology, such as engineering, the thesis that there are semantic changes or technological 17

incommensurability becomes, prima facie, admissible.

To show that the necessary and su

cient

conditions stipulated in (i)–(v) above are satis ed in the language of technology, we will consider two historical arguments: the transit from the contact theory of the voltaic pile to chemical theory and the transit from the material theory of heat to thermodynamics.

4. Transit of Contact Theory of Voltaic Pile to the Electrochemical Theory All of us have used, for quite varied purposes, electric batteries. Many readers probably know that an electric battery is a dispositive that transforms chemical energy into electric energy. And a few of you, familiar with the history of technology, may also know of its inventor, the Italian chemist and physicist Alessandro Volta. Kuhn expounds on this historical fact: The year 1800 is notable, among other things, for Volta’s discovery of the electric battery. That discovery was announced in a letter to Sir Joseph Banks, President of the Royal Society. It was intended for publication and was accompanied by the illustration reproduced here as gure I. For a modern audience there is something odd about it, though the oddity is seldom noticed, even by

historians. Looking at any one of the so-called “piles” (of coins) in the lower two-thirds of the diagram, one sees, reading upward from the bottom right, a piece of zinc, Z, then a piece of silver, A, then a piece of wet blotting paper, then a second piece of zinc, and so on. (Kuhn 2000, 20–21)

Figure 26.2. The voltaic battery is composed of a series of discs lying one on another in the following order: a zinc disc, one of silver, and onto these a piece of wet blotting paper (usually impregnated with brine), all of them p. 602

forming a pile. The cycle zinc, silver, wet blotting

paper repeats for an entire number of times (eight, to

be faithful to Volta’s original gure) precisely due to this disposition of its components; this dispositive takes the name “pile” that is still used in some Romance languages. Volta believed that the electric current produced by his battery was caused by the contact among the 18

materials it was made of. For this reason, this explanation is known as the “contact theory.”

However,

Volta’s contact theory was quickly rejected by many physicists, notably, by his coetaneous and fellow citizen Giovanni Fabbroni, who, after noticing that the zinc discs oxidized from being in contact with the wet paper, thought that oxidation—or corrosion—was the main cause of electricity. This suggested that corrosion was in fact an integral part of the battery’s capacity to produce an electric current, and it led to the substitution of Volta’s theory of the tension of contact for the electrochemical theory advanced by Fabbroni, Wollaston, Davy, M. de la Rive, and Antoine Becquerel, who thought that the battery’s electromotor power was the 19

result of a chemical reaction.

It comes as no surprise, as Kuhn suggests, that the explanatory di erences just summarized led to profound semantic divergences. Salient among them is the emergence of two terms that show why the contact theory and the electrochemical theory of the electric battery would be incommensurable—i.e., “battery’s element” and “electric resistance.” A unit element (or “battery’s element,” as it is also known) corresponds, in Volta’s theory of contact, to the following disposition: a zinc’s disc adjacent to a silver’s one (making sure the metals are in contact) and a piece of wet blotting paper (Kuhn 2002, 32). A battery element, considered from the current perspective, is made of a zinc’s disc, a silver’s disc, and between them the piece of wet blotting paper. This new p. 603

arrangement of elements, which corresponds to the modern view, implies the inversion of the current ow, besides making it external, since in Volta’s system it is internal (Kuhn 2002, 34). According to Kuhn: Far more important conceptually is the change in the current source e ected by the transition. For Volta the metallic interface was the essential element of the cell and necessarily the source of the current the cell produced. When the cell was turned inside out, the liquid and its two interfaces

with the metals provided its essentials, and the source of the current became the chemical e ects at these interfaces. When both viewpoints were brie y in the eld at once, the rst was known as the contact theory, the second as the chemical theory of the battery. (Kuhn 2000, 22–23) The transit from the contact theory of the Voltaic battery to the electrochemical theory not only changes the meaning of “unit element,” because the disposition of the elements in the pile change, but it also changes the meaning of “electric resistance.” According to Kuhn’s explanation: For a conducting material of a given cross section, it was measured by the shortest length the material could have without melting when connected across a given voltage. It is possible to measure resistance conceived in this way, but the results are not compatible with Ohm’s law. To get those results one must conceive the battery and circuit on a more hydrodynamic model. (Kuhn 2000, 24) We can notice that, along the lines of contact theory, electric resistance was measured by using the shortest length of a material to prevent the ow of the electric current or to avoid melting. According to Kuhn, any attempt at measuring resistance in this way nowadays would produce results incompatible with Ohm’s law, who understood the battery as a hydrodynamic model. In the electrochemical theory of the battery, the expression “electrical resistance” refers to the opposition to the ow of electric current through a conductive material. To measure resistance, electro-chemists took advantage of the studies and experiments by Georg Simon Ohm, which enabled him to establish—twentyseven years after inventing his electric battery—a relationship between electric tension and electrical resistance. Ohm discovered the scienti c principle on which grounds we can calculate resistance, and which embodies a way to approach this task that is radically di erent from the usual way to measure or calculate 20

electrical resistance by means of contact theory.

To electrochemists, the conductive’s resistance of a wire

(or any other conductive object) is directly proportional to its length (resistance increases with its greater length) and is inversely proportional to its transversal section (resistance decreases as its transversal section or thickness increases). The historical example just described allows us to infer that the theory of contact of the voltaic battery and the electrochemical theory of the voltaic battery are two genuinely incommensurable theories in the domain of technology, since the terms “unit element” (or “battery’s element”) and “electrical resistance” are not semantically equivalent, and therefore, are not inter-translatable. p. 604

This argument can be disputed on the grounds that both contact theory and the electrochemical theory of the battery are scienti c theories, not technological theories. We can respond to this reply arguing that since both theories were formulated with the purpose of nding out which was the causal agent of the electromotive force and in this way identify the principle that explained the working of this dispositive, we are entitled to consider them as genuine technological theories. In other words, technological theories have, like scienti c theories, explanatory capabilities, but, unlike the later, have a characteristic feature—namely, 21

that explanatory principles in technology are principles about the working of devices,

as we will see

shortly by considering the historical case of the explanation of the mechanism of the steam engine.

5. Transit from the Material Theory of Latent Heat to Thermodynamics The steam engine, considered the most emblematic achievement of the rst industrial revolution, was invented and patented by the mechanical engineer James Watt in 1769. Unlike the voltaic battery, which was designed to explain recurrence to the contact theory formulated by the inventor himself, philosophers and historians agree that the design of the steam engine did not necessitate science. The general belief on this matter suggests that the invention of the steam engine was the result of a complex process of experimentation and improvement of its preceding prototypes (the atmospheric machines), and for this reason its development was a result of the search for e “thermodynamic e

ciency, regardless of the fact that the notion of

ciency” had not been explicitly formulated yet. Taking into account that Sadi Carnot—

considered the father of thermodynamics—published, what could be considered the birth certi cate of thermodynamics, Ré exions sur la puissance motrice du feu et sur les machines propres à développer cette puissance, in 1824 ( fty- ve years after the machine was patented), and that almost a century later, Clausius and Lord Kelvin formulated the rst and second laws of thermodynamics. Based on these premises, some philosophers and historians hold that it was in fact the design of the machine what helped the development of the science, in particular, the formulation of thermodynamics: “Carnot discovered the pure science of thermodynamics as a result of his e orts to improve the e

ciency of

steam” (Feibleman 1983, 39). Precisely for this reason the meta-theoretic statement that “science owes more to the machine engine than what technology owes to science” (cf. Don Idhe 2013, 54) is a common place in literature or, in the words of Derry and Williams: The steam engine transforms thermal energy into mechanical energy, and its capacity obeys the laws of thermodynamics. These laws, however, were not formulated until the mid XIX century, and p. 605

hence, until then, the nature of heat was not

understood at all: such a circumstance reminds us

once more time that technology is not a synonym of applied science at all. (Derry and Williams 1977, 494–495) Although it is true that the design of the steam engine precedes by decades the scienti c theory that explains—in a systematic way—its functioning, the widely held belief that Watt invented the steam engine without using any theoretic resource is patently false. Watt was hired by the University of Glasgow in 1757 as an engineer and met there the celebrated chemist Joseph Black, recognized in the history of chemistry for his discovery of latent heat. Watt and Black engaged in a collaborative and friendly relationship until the latter moved to Edinburgh in 1766. Though Watt claimed he had reached the same conclusions as Black independently, Black not only had a tremendous in uence on his thinking but also advised him—in abundant correspondence—about the design and set up of many of the experiments Watt ran in the years previous to the invention of the condenser of the steam engine. Based on these historical records, we can ask ourselves which theoretic resources served as the basis for the design and invention of the steam engine. Watt was rmly committed to the material theory of heat. Though this is an even more archaic—or primitive —theory than the theory of the caloric itself, the material theory of heat stated that heat was a chemical substance (Miller 2009, 44). Such a theory also postulated latent heat. Black had introduced the terminology linked to this notion to explain the changes of status (e.g., liquid to gas). However, as Miller claims, Watt explored the scope of this idea beyond anything Black had done (cf. Miller 2009, 45). As a result of his investigations, he was able to formulate the widely known Watt’s law that stipulated that the addition of latent heat and sensible heat is a constant.

Watt had used this idea not only in his development of “expansive working” of steam engines but also, as we will see, it was important to his ideas about the chemical transformation of water into air. (Miller 2009, 48) Taking this into account, it is surprising that philosophers of technology, and even many historians, had produced the picture—very distorted—not only of the design of the machine as being the result of the sheer application of the trial-error method, in the total absence of theoretic resources, but also the depiction of Watt as not a man of science but as an ingenious mechanic: We become predisposed to regard Black as inhabiting the world of “theory” and Watt that of “practice.” We are encouraged to see Black as a cosmopolitan gure of the Scottish Enlightenment and Watt as a clever engineer hanging from his coattails, their relationship involving a “transfer of intellect to craft.” This is a mistake. (Miller 2009, 89) Controversial as they might be, Watt’s achievements in the domain of mechanical engineering ended up p. 606

overshadowing his contributions in the domain of chemistry that,

in a retrospective way, produced a

quite archaic chemistry that made us forget—as the engineer Andrew Ure rightly recognized—“that the steam engine was the most splendid success of the science built on the idea of caloric” (Miller 2009, 71). Now, suppose it is conceded that Watt was able to explain the functioning of the steam engine by recurring to a theory (i.e., the material theory of heat—and with it, the theory of latent heat). How does this support the thesis of incommensurability for the language of technology? As we have seen, incommensurability is a dyadic and inter-theoretic relationship—that is, a semantic relationship that holds between pairs of theories. The technological theories we are examining in the historical argument over the electric battery are Volta’s contact theory versus the electrochemical theory of the battery, and in this transition, there are signi cant and profound semantic changes. In the historical argument over the design of the steam engine the theories 22

in con ict are the material theory of heat by Black-Watt versus thermodynamics.

Comparing both

theories, it turns out that the meanings of the shared term “latent heat” are diametrically opposed, as we learn from Miller’s explanation of the theoretic underpinnings of the term: We are accustomed to thinking of Black as “the discoverer of latent heat,” and by this we usually mean the discoverer of what we understand by the term “latent heat.” But this is quite wrong. Our concept of latent heat is underwritten by a kinetic theory of heat; Black’s was based on a material (chemical) one in which latent heat was intrinsic to chemical reactions and was compounded with ordinary matter. (Miller 2009, 96) As Miller contends, while by the term “latent heat” nowadays we refer to the quantity of energy required to cause a change of state in matter, e.g., passing from the solid state to the liquid or gas state, Black and Watt understood latent heat as a chemical property inherent to ordinary matter. In fact, Black had interpreted “latent heat” as the cause of the changes of state of the matter: for example, water turned into steam due to additional heat that had to be provided and that “was not detectable by the thermometer” (the reason he called it “latent heat” or hidden heat), and which was combined with the matter of water to produce steam (cf. Miller 2009, 45). This is not the only reason for which the term “latent heat” in the material theory of heat is not semantically equivalent to the corresponding term in thermodynamics. While latent heat is—in light of

23

thermodynamics—a physical magnitude

that indicates the quantity of thermic energy that a mass has

either to lose or to gain in order to change its physical status; latent heat, in Black-Watt’s theory, is a chemical reaction. By the same token, although latent heat is the causal agent that explains the changes of status in both theories, their respective properties are radically di erent: while in the material theory of heat, latent heat is a property of the matter, in thermodynamics latent heat is energy that becomes transformed. It comes to no surprise that Black and Watt’s theory was called the material theory of heat. For both, heat has a material nature (not a dynamic one, as it occurs in current thermodynamics). As a matter of fact, Black reached this conclusion after considering the behavior of melting ice and the interpretations of the role of heat in those changes by Cleghorn: p. 607

Black’s most important discovery was the observation that melting ice absorbs heat without changing its temperature. Based on this fact, he reached the conclusion that ice needs latent heat for this modi cation of its physical condition. This was for him the substantial proof of the material nature of heat and in 1779 one of his students, William Cleghorn (1754-1783), gave the precise de nition of caloric. (Bozsaky 2010, 6) This sort of analysis could be rejected on the grounds that thermodynamics is not a technological theory, hence, appealing to this historical case seems inappropriate because it appears to support instead the idea that there is incommensurability between theories in the domain of science. However, we could respond by highlighting that, in fact, granted thermodynamics belongs to the quarters of physics, the material theory of heat defended by Watt—and his peculiar way of understanding the nature of latent heat—tried to explain essentially the functioning of the steam engine, which is a task of technology. If this is right, the example I just analyzed could illustrate a case of incommensurability between a technological theory, namely, the material theory of heat and a scienti c theory, i.e., thermodynamics which, all in all, was formulated after the e orts of prominent scientists to give a good explanation of the functioning of the external combustion engine. In addition, there is nothing that prevents us of reformulating condition (i)—stated above—that indicates that a conditio sine qua non to consider theories T1 and T2 like genuinely incommensurable theories lies in the fact that T1 and T2 have to be successive theories separated by a scienti c revolution. Along these lines, we are entitled to demand that T1 and T2 be genuine rivals, i.e., that are not trivially incommensurable, or that do not satisfy the following condition: Theories T1 and T2 are trivially incommensurable i

T1 and T2 are completely unrelated to each

other; that is, both their theoretical language as their testable language are semantically incompatible. For example, the catastrophism theory in geology and the value theory in Marx’s political economy would be trivially incommensurable. The triviality lies in the fact that they do not explain the same or they do not deal with the same facts. Since the material theory of heat and thermodynamics are not trivially incommensurable and the rest of the conditions for technological incommensurability stated in the foregoing are satisfactorily met, we can conclude that this case from the history of technology provides us with a genuine instance of semantic change in the language of technology. In this discussion, I have argued for technological incommensurability and I have dispelled the reluctance to apply incommensurability (understood as a semantic phenomenon) to artifacts. This line of reasoning comes from the mistaken identi cation of technology with artifacts, and by explaining the erroneous nature of such identi cation I hope to have given enough reasons to block this kind of reply. In addition, I have p. 608

shown that behind this type of argument hides the false belief that there are not and

cannot be

technological theories. Such a belief can be successfully refuted by showing, in the way I have proceeded, that there are technological theories, among them, precisely Volta’s contact theory that explains the

functioning of the battery or the material theory of heat that explains —though in a quite rudimentary way — the functioning of the steam engine.

Acknowledgments A preliminary version of this work appeared in: Cuadernos de Filosofía: ¿Hay inconmensurabilidad tecnológica? No 38 (119–141), 2019 ISSN 0716-9884 https://doi.org/10.29393/CF38-5DFHI10005. Chile. I want to thank Ernest Lepore for his careful reading of this work and his useful suggestions to improve it.

Notes

p. 609

1.

This thesis was advanced, independently, by Paul Feyerabend (1962) and Thomas Kuhn (1962). I will restrict myself to a couple of versions that can be identified in Kuhnʼs work.

2.

Some, such as Sankey (1993) and Falguera (1998), distinguish three or four phases in the formulation of the thesis of the incommensurability of theories. Nevertheless, I prefer parsimony, and so stick to a binary formulation of the thesis identifying just two versions: the semantic one (advanced in the sixties) and the taxonomic one (proposed in the eighties).

3.

The first formulation of the incommensurability thesis between theories separated by a scientific revolution appears in SScR, but can be found also in further works, e.g., the Post Scriptum (Kuhn 1969), The Essential Tension (Kuhn 1977) and “Postscript” (1969/1996).

4.

Schematically: Two theories T1 and T2 are incommensurable i their taxonomic structures are not homologous. Despite the fact that the first version of the incommensurability thesis is not radically faulty, the second one has the advantage of explaining the reasons for the semantic change (both at the level of meaning and at the level of reference) in incommensurable theories. The reason is that there cannot be an homology between taxonomical or set structures if there is incommensurability. Taxonomical categories are set terms that have at least two properties: (i) they are preceded by an indefinite article and (ii) they must satisfy the principle of not overlapping—namely, they can overlap only from genus to species but not from species to species. Kuhn explains this using the example “there are no dogs that are also cats.” It is not the case that two of these categories can have some common cases unless one of them subsumes fully and necessarily the other. Di erent taxonomical structures, e.g., those whose di erences depend on the subsumption and exclusion relationship are incommensurable if their di erences result in terms with fundamentally disparate meanings (cf. Kuhn 2000, 92).

5.

Several well-known critics of Kuhn believe he cannot give a satisfactory definition of the term “paradigm” because of its polysemy (cf. Masterman 1970). I consider that a functional definition for the purposes of this chapter is possible as follows: a paradigm can be understood as a hybrid entity that encompasses not only a set of scientific theories, laws, and models but also a set of ontological, conceptual, and methodological commitments (e.g., the classic mechanics of particles).

6.

Kuhn (2000) argues for the thesis of incommensurability as untranslatability by appealing to Quineʼs principle of the underdetermination of translation, according to which, when we translate from a language to another, there is no way to know which is the correct translation of an expression. This can be explained by Quineʼs (1960/2013) mental experiment in Word and Object under the following hypothetical situation: Suppose a linguist studies a recently discovered tribe whose language lacks a translation manual. The linguist can produce two manuals, the first one written in his language, and the second one written in the tribeʼs language. Though the words written in both manuals are compatible with the same sensory stimuli, they can still be mutually incompatible. For this reason, Quine holds that a radical translation requires the preservation or inscrutability of reference. Since this does not obtain (regardless of the nativeʼs observable behavior) the outcome is that multiple translations are mutually incompatible. Kuhnʼs view runs slightly di erently from Quineʼs. The former holds that a translation must preserve not only reference but also meaning. Given that the requirement for a successful translation is the preservation of the meaning and reference of the terms involved—and this is precisely what does not occur when we have incommensurability—the failure of translation is inevitable.

7.

This does not imply that we lack criteria to rationally evaluate and choose between rival theories. The point is that such

criteria are not rational decision algorithms (e.g., verification or falsification), but pragmatic criteria. For a more detailed analysis, see Kuhn (1998), “Objectivity, Value Judgment and Theory Choice.” 8.

Metaphysical and ontological commitments of scientists indicate the entities that the universe contains. Methodological commitments refer to the sort of rules and admissible practices for the members of a scientific community.

9.

An additional consequence of incommensurability is the collapse of communication. I will not be concerned with this problem; the reader can see The Essential Tension (Kuhn 1977).

10.

There are many works that explain the di erence between local movement, considered as a process, and the same phenomenon, considered as a state. See Blanché (1969); Koyré (1977, 1980); and García (1997).

11.

Popper (1975); Davidson (2001); Kitcher (1978); Putnam (1988).

12.

Kline mentions four uses of the term ʻtechnology.ʼ The first identifies technology with the totality of artifacts. See “USAGE 1: HARDWARE (OR ARTIFACTS) : Possible denotation: non-natural objects, of all kinds, manufactured by humans,” in Kline (2003,

210). 13.

I will not attempt to o er a non-controversial definition of ʻtechnologyʼ here. This problem is rich and complex, as I have explained in Florez et al. (2019).

14.

Obviously, this is not the only one available in this field. The reader might consider the more recent beam theory formulated by the Ukrainian engineer Timoshenko.

15.

As a matter of fact, beam theory belongs to a field devoted to the study of material resistance, which is of interest to civil engineers (especially structural engineers), as well as mechanical and industrial engineers.

16.

There are plenty of theories in technology. The philosopher Mario Bunge (1966) studied this field deeply. He deemed it possible to distinguish between “substantive technological theories” and “operative technological theories.” The di erence between them is that the former always have, as a theoretical substratum, scientific theories, whereas the latter share with science, at most, the scientific method. Flight theory, whose substratum is fluid dynamics, is an archetypical example of substantive technological theories; value theory, decision theory, and game theory are archetypal examples of operative technological theories (cf. Bunge 1966). We could refer to another theory from structural engineering, i.e., plank theory, formulated by the mechanical engineer Raymond Mindlin (1951) and the civil engineer Reissner (1945). Both devoted their best e orts to studying the deformation of planks and steel sheets and formulated a set of kinematic hypotheses on how a plank or a steel sheet deforms under flexion.

17.

It might be thought that my argument depends on establishing that engineering is a technology (or constitutes a branch of technology), which is controversial. However, I ask the reader to admit, for now, that engineering is a subarea of technology, as in Bunge (1980); Jarvie (1967); Feibleman (1983); and Skolimowski (1966).

18.

In a letter, Volta explains how his battery works: “On the electricity excited by the mere Contact of conducting substances of di erent Kinds. In a letter from Mr. Alexander Volta, F.R.S. Professor of Natural Philosophy in the University of Pavia to the Rt. Hon. Sir Joseph Banks, Bart. K.B. P.R.S. June 26, 1800.”

19.

Antoine Becquerel was the first to use an electrolytic process to separate the metal from the mineral in which it is encrusted. Electrolysis is a procedure that consists of passing an electric current through certain conductive substances (or electrolytes). Such a discovery turns him into the electrochemistryʼs father. He also, in 1829, discovered the battery of constant current, exploited years later by the British chemist John Frederic Daniell (1790–1845). These and other inventions and discoveries that were quite important for the growth of contemporary science made Becquerel an elected member of the Sciences Academy in 1829. Eight years later, he was honored with the prestigious chair of professor of Physics of the Muséum dʼHistoire Naturelle (in Paris), a position in which his son and grandson succeeded him.

20.

For a cable-like conducting, resistance is given by the following formula:

p. 610

R = p

l S

where ρ is the coe icient of proportionality or resistance of the material, ℓ is the cableʼs length, and S the area of its transversal section.

21.

Kröes (1992) undertook a study of characteristic features of theories in engineering that explain technological functions of a particular kind of artifact or of its materials. In his opinion, the domain of application of theories in engineering can be characterized because they refer to (a) designable technical artifacts; (b) can also contain basic principles over the design or construction of artifacts; and (c) use technical concepts.

22.

According to Miller, the 1764 work of Watt about heat “was undertaken as part of Blackʼs tradition of work. Black always made prominent reference to Wattʼs experiments in his lectures at Edinburgh University so far as we can rely on Robisonʼs edition of Blackʼs lectures” (Miller 2009: 94).

23.

In the international system of measurement it is measured in J/kg, and represented by the letter L.

References Bauchau, O. A., and Craig, J. I. (2009). “Euler-Bernoulli Beam Theory.” In O. A. Bauchau and J. I. Craig, eds., Structural Analysis, 173–222. Solid Mechanics and Its Applications, no. 163. Dordrecht: Springer Dordrecht. Google Scholar Google Preview WorldCat COPAC Blanché, Robert. (1969). La Méthode expérimentale et la philosophie de la physique. Paris: Armand Collin. Google Scholar Google Preview WorldCat COPAC p. 611

Bozsaky, David. (2010). “The Historical Development of Thermodynamics.” Acta Technica Jaurinensis 3 (1): 3–17. Google Scholar WorldCat Brown, Harold I. (1977). Perception, Theory and Commitment: The New Philosophy of Science. Chicago: Precedent Publishing. Bunge, Mario. (1966/1974). “Technology as Applied Science.” In Friedrich Rapp, ed., Contributions to a Philosophy of Technology: Studies in the Structure of Thinking in the Technological Sciences, 19–39. Dordrecht: D. Reidel. Google Scholar Google Preview WorldCat COPAC Davidson, Donald. (2001). “On the Very Idea of a Conceptual Scheme.” In Inquiries into Truth and Interpretation, 183–198. Oxford: Clarendon Press. Google Scholar Google Preview WorldCat COPAC Derry T. K., and Williams, Trevor, I. (1960). A Short History of Technology from the Earliest Time to A.D. 1900. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Díez, José A., and Moulines, Carlos Ulises. (1999). Fundamentos de Filosofía de la ciencia. Barcelona: Ariel. Google Scholar Google Preview WorldCat COPAC Falguera, José L. (1998). “Inconmensurabilidad y ontosemántica representacional.” Theoria. An International Journal for Theory, History and Foundations of Science 13 (31): 161–185. Google Scholar WorldCat Feibleman, James K. (1983). “Pure Science, Applied Science, and Technology: An Attempt at Definitions.” In En C. Mitcham and R. Mackey, eds., Philosophy and Technology: Readings in the Philosophical Problems of Technology, 33–41. New York: Free Press. Google Scholar Google Preview WorldCat COPAC Feyerabend, Paul K. (1962). Explanation, Reduction and Empiricism. Minneapolis: University of Minnesota Press. Google Scholar Google Preview WorldCat COPAC Florez, Daian T., García-Duque, Carlos Emilio, and Osorio, Juan Camilo (2019). “Is Technology (Still) Applied Science?” Technology in Society 59. https://doi.org/10.1016/j.techsoc.2019.101193. Google Scholar WorldCat García, Carlos Emilio. (1997). Evolución histórica del pensamiento científico. Manizales: Universidad de Manizales. Google Scholar Google Preview WorldCat COPAC Ihde, Don. (2013). “Technology and Science.” In J. K. B. Olsen, S. A. Pedersen, and V. F. Hendricks, eds., A Companion to the Philosophy of Technology, 51–60. Malden, MA: Wiley-Blackwell. Jarvie, Ian. (1967). “Technology and the Structure of Knowledge.” In C. Mitcham and R. Mackey, eds., Philosophy of Technology: Readings in the Philosophical Problems of Technology, 54–61. New York: Free Press. Google Scholar Google Preview WorldCat COPAC

Kitcher, Philip. (1978). “Theories, Theorists, and Theoretical Change.” Philosophical Review 87:519–547. Google Scholar WorldCat Koyré, Alexandre. (1939/1966). Études galiléennes. París: Éditions Herman. Google Scholar Google Preview WorldCat COPAC Koyré, Alexandre. (1973). Études dʼhistoire de la pensée scientifique. París: Éditions Gallimard. Google Scholar Google Preview WorldCat COPAC Kline, Stephen J. (2003). “What Is Technology?”. In Robert C. Schar and Val Dusek, eds., Philosophy of Technology: The Technological Condition; An Anthology. Malden, MA: Blackwell. (Original article published in Bulletin of Science, Technology Society). Google Scholar Google Preview WorldCat COPAC Kröes, Peter. (1992), “On the Role of Design in Engineering Theories: Pambourʼs Theory of the Steam Engine.” In Peter Kröes and Martijn Bakker, eds., Technological Development and Science in the Industrial Age, 69–98. Boston: Springer-Science+Business Media, B.V. Google Scholar Google Preview WorldCat COPAC Kuhn, Thomas S. (1962/1996). The Structure of Scientific Revolutions. Chicago: University of Chicago Press. Google Scholar Google Preview WorldCat COPAC Kuhn, Thomas S. (1969/1996). Postscript to The Structure of Scientific Revolutions, 174–210. Chicago: University of Chicago Press. Kuhn, Thomas S. (1977). The Essential Tension: Selected Studies in Scientific Tradition and Change. Chicago: University of Chicago Press. Google Scholar Google Preview WorldCat COPAC p. 612

Kuhn, Thomas S. (1998). “Objectivity, Value Judgment and Theory Choice.” In E. D. Klemke, Robert Hollinger, and David Wÿss Rudge, eds., Introductory Readings in the Philosophy of Science, 435–450. New York: Prometheus Books. Google Scholar Google Preview WorldCat COPAC Kuhn, Thomas S. (2000). The Road since Structure. Edited by James Conant and John Haugeland. Chicago: University of Chicago Press. Google Scholar Google Preview WorldCat COPAC Masterman, M. (1970). ʻThe Nature of a Paradigmʼ. In Imre Lakatos and Alan Musgrave, eds., The Criticism and the Growth of Knowledge, 50–89. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Miller, David Philip. (2009). James Watt, Chemist: Understanding the Origins of the Steam Age. Science and Culture in the Nineteenth Century. London: Pickering & Chatto. Google Scholar Google Preview WorldCat COPAC Popper, Karl. (1970). “Normal Science and its Dangers.” In I. Lakatos and A. Musgrave, eds., The Criticism and the Growth of Knowledge, 51–58. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Putnam, Hilary. (1981). Reason, Truth and History. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Quine, Willard Von Orman. (1960/2013). Word and Object. Cambridge, MA: MIT Press. Google Scholar Google Preview WorldCat COPAC Sankey, Howard. (1993). “Kuhnʼs Changing Concept of Incommensurability.” British Journal for the Philosophy of Science 44 (4):

759–774. Google Scholar

WorldCat

Skolimowski, Henryk. (1966). “The Structure of Thinking in Technology.” In C. Mitcham and R. Mackey, eds., Philosophy of Technology: Readings in the Philosophical Problems of Technology, 42–49. New York: Free Press. Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

27 On Retweeting  Eliot Michaelson, Jessica Pepp, Rachel Sterken

Author Notes

https://doi.org/10.1093/oxfordhb/9780192844118.013.49 Published: 22 May 2024

Pages 613–630

Abstract A small, but growing literature in philosophy is devoted to the understanding of a seemingly new communicative action that came with the internet, and with Twitter in particular: the retweet. The spur for this literature is a kind of puzzle in public discourse: on the one hand, there is a tendency to hold people responsible for their retweets, and to blame them for retweeting material considered o ensive or otherwise inappropriate. On the other hand, there is a widely shared, if not universally recognized feeling that, as the well-known disclaimer has it, “A retweet is not an endorsement.” But if a retweet is not an endorsement, what is it? And what is wrong with retweeting o ensive or misleading tweets? This chapter puts forward the view that bare, uncommented retweets are best understood as lacking any sort of default illocutionary force, in contrast to many other types of speech acts. And, whereas with most speech acts, it is massively di

cult to in uence the norms governing that type of

act, this may not be the case with retweeting. With retweeting, it is possible to consider ways that Twitter’s interface and code might be altered to revise the act of retweeting; likewise, it is possible to consider the likely impacts of such revisions on the norms surrounding retweeting. This raises a pair of interesting questions: (i) what should we want retweeting to be like, and (ii) how can we make progress in that direction?

Keywords: retweet, speech act, reshare, repost, assertion, locutionary act, illocutionary act, fake news, attention, conceptual engineering Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

During the town hall, the President … tried to separate himself from his recent retweet of a conspiracy theory from an account linked to QAnon, which baselessly claimed that former Vice

President Joe Biden orchestrated to have Seal Team Six killed to cover up the fake death of al Qaeda founder Osama bin Laden. … “I know nothing about it,” Trump claimed. “That was a retweet—that was an opinion of somebody. And that was a retweet. I’ll put it out there. People can decide for themselves.” … But [Samantha] Guthrie responded: “I don’t get that. You’re the President. You’re not like someone’s crazy uncle who can retweet whatever.” —Maegan Vazquez, “Trump Again Refuses to Denounce QAnon,” CNN Politics, October 15, 2020

1. Introduction A small, but growing literature in philosophy is devoted to the understanding of a seemingly new communicative action that came into being with the development of the internet, and Twitter in particular: the retweet. The spur for this literature is a puzzle in public discourse. On the one hand, there is a tendency to hold people responsible for their retweets, and to blame them for retweeting material considered o ensive or otherwise inappropriate. On the other hand, there is a widely shared, if not universally recognized, feeling that, as the well-known disclaimer has it, “A retweet is not an endorsement.” But if a p. 614

retweet is not an endorsement, what is it? And what is wrong with

retweeting o ensive or misleading

tweets? What sort of responsibility do people have for their retweets? The rst of these questions, What is a retweet, if not an endorsement? asks for retweeting to be placed under some broader category of speech action. One way to do this, as the question acknowledges, is by classifying retweeting as a form of endorsement. To do so is to see retweeting as similar or analogous to, for instance, saying “I agree” in response to someone else’s assertion, signing one’s name to a petition, or giving a thumbs-up sign when shown a photograph someone is suggesting to use for some purpose. E ectively, the claim that retweeting is a broad form of endorsement proposes to understand retweeting as having a default illocutionary force—prima facie at least, retweets are ‘assentives’ in the terminology of Bach and Harnish (1979, 43). We will review some compelling reasons against categorizing retweeting in this way, and pursue answers to a more general framing of this question, What type of speech action is retweeting? by considering which other kinds of actions are communicatively similar to retweeting. We won’t limit our answer to this question to existing categories of speech actions (e.g., what one might nd in Bach and Harnish 1979), and we will consider the possibility (raised, for example, by Marsili 2020) that retweets are an entirely new kind of speech act. Ultimately, however, we will suggest that retweeting has no default force and resists subsumption under illocutionary categories, old or new. This conclusion might seem disappointing with respect to the further aim of answering the normative questions asked above, What is wrong with retweeting o ensive or misleading tweets? and more generally, What responsibility do people bear for their retweets?. If we could place retweeting in a broader speech category with other actions whose associated responsibilities are more widely established, then that would be one way to make the case for a particular set of associated responsibilities. The disclaimer “A retweet is not an endorsement” presumes that endorsement brings with it certain responsibilities, and stipulates that these should not be applied to retweeting. But if we can identify no illocutionary category in which retweeting, by default, belongs, then judgments about any responsibility associated with such categories do not shed light on the question of responsibility for one’s retweets. There is, however, another way of approaching the e ort to categorize retweeting—one which opens the door to a rather di erent way of proceeding. On this alternative approach, instead of aiming to describe

what kind of speech act retweeting is in the here and now, we aim to describe what kind of speech act retweeting ought to be. Call this the ‘engineering’ project with respect to retweeting. While philosophers are used to distinguishing between the non-ideal question of what our norms regarding X are at present and the ideal question of what they ought to be like, this is not a distinction that has been applied speci cally to the case of speech acts. However, we believe that this distinction may prove especially important in the case of novel speech acts like retweeting. Of course, answers to these questions—What sort of speech act is retweeting?, What are the norms surrounding it?, and What ought those norms to be?—are clearly interrelated. If the best answer to the rst question is that retweeting is an endorsement, then the answer to the second should be fairly clear, and, supposing that our p. 615

present norms of

endorsement are in good working order, perhaps the third as well. As we’ll see below,

there is something distinctly interesting for philosophers about the phenomenon of retweeting: retweeting is a communicative act that we have explicitly engineered for ourselves, by way of the large and in uential social media company Twitter. Most of the speech acts with which we are familiar did not arise like this; rather, assertions, requests, inquisitives etc., presumably all developed out of repeated interactions between individuals, with the understanding of each of those individuals partially determining how subsequent uses of a particular grammatical construction or innovation were likely to be interpreted. The phenomenon of retweeting was created largely by the engineers who designed the platform, with a bit of code. Unlike our social evolution, that code can be tweaked in a relatively straightforward way—via the intervention of Twitter’s engineers. What this means, practically, is that we can seriously contemplate substantial, top-down revision of the way the act of retweeting functions. This would be almost inconceivable when it comes to more established 2

speech acts like asserting or commanding.

Here’s the plan for what comes next. In section 2, we will consider three proposals (including an earlier one of our own) that seek to categorize bare retweeting within extant illocutionary types. As we’ll explain, we take all of these proposals to be mistaken: retweeting is more basic than these proposals would have it, lacking any kind of default illocutionary force. In sections 3 and 4, we clarify what we take to be the interesting philosophical project of speech-act engineering, which is one of asking, not what speech acts we presently have, but what speech acts we ought to work to promote because they better serve our legitimate purposes (communicative, epistemic, moral, etc.). In section 3, we explore proposals that would aim to engineer retweeting so that it ts within a well-established illocutionary type. In section 4, we consider how we might engineer retweeting, even supposing that our descriptive account of retweeting is and should remain correct. Finally, section 5 concludes.

2. Retweeting as a Basic Act The question, What sort of speech act is retweeting? can be seen as a question about the illocutionary force of retweeting, in the traditional Austinian sense (Austin, 1975). The question might also be put as, What does 3

one do in retweeting a tweet?. It should be agreed, though, that the fact that an action is a (bare) retweet

does not in itself entail that the action is any particular sort of illocutionary act. This is analogous to the fact that an utterance of a declarative sentence like “Snow is white” can constitute a variety of di erent illocutionary acts in di erent conversational contexts. A bare retweet, like an utterance of “Snow is white,” is simply a basic communicative tool within a particular communicative system. In the case of “Snow is white,” the system is the spoken version of the natural language English. In the case of the bare retweet, the system is Twitter. (Obviously, these two communicative systems interact substantially.) p. 616

Even if it is clear that an act’s being a retweet does not entail that it has any particular illocutionary force, one might suppose that an act’s being a retweet would dispose it toward having a certain illocutionary force

—that a default illocutionary force will be attached to a bare retweet, even if that default can be overridden. This would be in the same way that an act’s being an utterance of “Snow is white” seems to dispose it to be an assertion, or an act’s being an utterance of “Close the door!” seems to dispose it to be a command. We can ask whether retweeting has a default, or dispositional, illocutionary force, in the way that utterances of sentences of various types arguably do. For reasons we are about to set out, we think it is di

cult to identify any such default force for retweeting.

Retweeting a particular kind of tweet lacks the sort of literal meaning or content that makes an act of uttering a declarative sentence well-suited to be an assertion, or an act of uttering an imperative sentence well-suited to be a command. The fact that it is di

cult to identify any such default force for retweeting is

problematic, because it makes it hard to tell what speech act people are performing when they retweet. It may even make it di

cult to perform distinctive speech acts by retweeting, especially in cases where

retweeters do not share rich informational backgrounds with their audiences. The good news, as mentioned above, is that retweeting—unlike uttering a sentence of English—is an act whose features can be programmed and reprogrammed without the tacit cooperation of an entire linguistic community. Retweeting can be ‘engineered’ so as to make acts of retweeting better suited to being acts with a particular illocutionary force (or forces), and less well-suited to being acts with other sorts of illocutionary forces. (This engineering will be the topic of sections 3 and 4.) In the remainder of this section, we will make the case that retweeting is not like other speech acts that seem to have a default illocutionary force. We start by considering three models of the default force for retweeting: endorsement, reporting, and 4

indication. These models yield the following analogies: A user, U, retweeting a tweet, T, by original poster, OP, is like: (1) U uttering “I endorse this,” where the demonstrative “this” refers to T. (endorsement) (2) U uttering “OP tweeted this,” where the demonstrative “this” refers to T. (reporting) (3) U uttering “Look at this,” where the demonstrative “this” refers to T. (where indicating = directing attention) Both Arielli (2018) and Marsili (2020) argue that the popular disclaimer “a retweet is not an endorsement” suggests that retweeting is not like (1). If retweeting were something like (1), it would seem contradictory to issue such a disclaimer. But, intuitively at least, this disclaimer does not seem contradictory. Further, if one is questioned about why one retweeted a certain tweet, it does not seem contradictory to reply, “I just thought people should see it—I don’t endorse it.” (Marsili 2020, 10464) Similar points could be made about any model of retweeting that treats the retweet as a way of expressing approval for the original tweet. If p. 617

retweeting were the same as “I approve of this” or “I like

this” or “Hurray for this!”, the popular

5

disclaimer would instead seem nonsensical. So retweeting is not akin to endorsing in any of these ways. Regarding (2), both Arielli and Marsili point out that to report what someone says, either by direct quotation or indirect discourse, is to make an assertion, which is true or false depending on whether that person spoke as reported, and is sincere or insincere depending on whether the reporter believes that that person spoke as reported. In contrast to this, a retweet is not true or false depending on whether the OP tweeted the tweet. A retweet shows that the OP tweeted the tweet and is not evaluable as true or false. If the OP did not tweet the tweet that a putative retweet appears to retweet, then the putative retweet is not a retweet at all (but, rather, some sort of fake). Likewise, if it makes sense to posit sincerity conditions for retweeting in itself at all (which we doubt), it seems most plausible that a retweet is sincere if and only if the particular act of retweeting is a genuine expression of the retweeter’s beliefs or feelings concerning the content or context of the original tweet and/or subsequent uptake. It surely does not turn on whether they believe that the OP tweeted the tweet. So retweeting is not like reporting on someone’s tweet by stating that they tweeted it.

The observation that a retweet shows a tweet, rather than reports on it, suggests that the right model for retweeting is indication or attention-direction. This is a model we have previously suggested for online resharing in general, retweeting included (Pepp, Michaelson, and Sterken 2019), which also nds some favor with Arielli and Marsili. It likens retweeting to pointing communicatively at the original tweet or, as suggested in (3), issuing the instruction, “Look at this!” There is much to commend this model over the others: it does not make the denial of endorsement contradictory, and it does not require retweets to be truth-apt or sincerity-apt in ways they clearly are not. Still, we are now of the opinion that it is not quite right. Both Arielli and Marsili argue that to retweet is not merely to indicate or direct attention to the original tweet, since a retweeter also reproduces the original tweet. This reproduction creates a new opportunity for people to see the tweet, thereby amplifying it or expanding its reach in a way that simple indication does not. We agree with this. But we think there is a further problem with treating retweeting on the model of (3). This problem, which is similar to the problem pointed out for (1), suggests that retweeting does not have the default force of directing attention to the original tweet. To see the problem, consider the following example: Suppose that @Nora, a Twitter user who is a huge fan of Roblox, retweets the post “Roblox rules!” by her favorite Roblox YouTuber star, @it’sakeila. Suppose further that (i) all of @Nora’s Twitter followers are also enthusiastic followers of @it’sakeila and so have seen the post, and (ii) that @Nora’s followers, being such dedicated fans that they keep in touch about all aspects of @it’sakeila, all know that all of them have seen the post, and know that they know that they’ve seen the post, and so on. That is, suppose that the tweet by @it’sakeila is common ground among @Nora’s followers. Now imagine that @Nora’s mother, who struggles to understand the ways of these young internet denizens, observes @Nora’s retweet and asks her daughter, “Why did you retweet that if you’ve all seen it already and know that each other have all seen it already?” @Nora rolls her eyes and replies (in the p. 618

exasperated tone

reserved for parents), “I wasn’t telling them to look at it! I was just showing that I’m

into it!” If retweeting had a default meaning along the lines of “Look at this,” @Nora’s reply would seem contradictory. But it does not. For comparison, imagine instead that @Nora is on the playground with her friends and suddenly grabs a sand pail and says, “Look at this!”. If her mother asks her, “Why did you tell them to look at that pail?”, it would be nonsensical for @Nora to reply, “I didn’t tell them to look at it, I was just showing that I’m into it!”. Given this, it seems incorrect to assimilate retweeting to a locutionary act like uttering, “Look at this,” or even a non-linguistic act of directing attention by pointing. One concern regarding this argument stems from the observation that one can say “Look at this/that” without intending to instruct someone to look at the thing referred to. For instance, it is common to react to a beautiful sight that one’s companions have already attended to and remarked on by saying, “Yeah, wow, look at that,” or the like. Given this, retweeting might well have a default meaning similar to uttering, 6

“Look at this/that,” and still be well suited to be used non-directively, simply to bond with others. This objection is more compelling if it is generally true that locutions whose default meaning is attentiondirecting are also easily understood as fostering bonding over already-attended items. If this is a robust, cross-linguistic pattern—something we are simply unsure about—then the @Nora case ts nicely into an account on which retweeting is like saying “Look at this” or pointing at something. If it is not, then the way in which retweeting seems to pattern with the English sentences “Look at this” or “look at that” may be more coincidental. It is also worth noting that physically pointing at things already jointly attended to and remarked on seems generally out of place. Suppose A, B, and C are out walking. A points at the sunset and says, “Look at that, how beautiful!” B says, “Oh, wow, what a great sunset!” It would be natural enough for C to chime in by saying, “Yeah, look at that.” But it would be downright strange for her to join in by silently pointing at the sunset, and even slightly strange for her to point at it while saying, “Yeah, look at that.” So, to the extent

that we think that retweeting isn’t merely just like saying “Look at this,” but also involves something like pointing to the relevant tweet, @Nora’s response here would be unanticipated by the present model. It seems fair to say that the use of the retweet, as in the @Nora case, calls (3) into question as a model of retweeting, but is not decisive against it. Even if (3) is not the right model, it may be feasible to treat retweeting as a more basic form of indicating; a form of indicating that does not carry the default prescriptive, attention-directing force of locutionary acts like uttering, “Look at this.” For instance, Marsili (2021, 10472–10474) analyzes retweeting as overtly showing or making manifest the original tweet. He uses this base analysis to build a general relevance-theoretic account of the communicative value of retweets. Marsili’s account is consistent with our suggestion here that retweeting does not have a default meaning or illocutionary force, since he understands indicating as something more basic than any type of illocution, something which gures into illocutionary acts via its interaction with patterns of relevance to particular listeners. In contrast, in our earlier work we took indicating to be a thicker, more prescriptive kind of activity—more in line with the model represented by (3). p. 619

At this point, we are going to leave aside this descriptive question regarding what the norms of retweeting are. For, as interesting a question as it is, we are more interested in the question of what the norms of retweeting could be—and what we might be able to do to push them in that direction.

3. Engineering the Retweet As we noted at the start, philosophers are sometimes interested not just, or not even primarily, in o ering descriptively adequate accounts of this or that phenomenon, but in o ering normative accounts of how those phenomena ought to be. Such projects are perhaps more familiar in ethics, focused as it is on how human beings ought to behave, and in political philosophy, focused as it is on how humans ought to behave, together, at various levels of idealization. More recently one strand of this project has come once again to prominence in the philosophy of language and mind, under the rubric of conceptual engineering. According to those engaged in the project of conceptual engineering, we should not only ask ourselves what characterizes a certain concept currently in circulation, WOMAN for instance, and why, but also what other revised or replacement concept(s) might serve us better than our present one, along various epistemic, ethical, and political dimensions. Disagreements persist about both how to weigh these various factors and just how far from our current practice it is legitimate to stray while retaining a claim to be modifying an extant concept, as opposed to introducing a new one, but we can safely ignore these for present purposes. With respect to retweeting, the concept is not really at issue. Rather, it is the background social norms and practices and, perhaps most saliently, the speech act itself. So, for instance, one might take it that the indication model of retweeting o ers an accurate account of our present norms of retweeting, but that the endorsement model would nonetheless be better for us in various ways if it were true. Rini (2017, E-55) recommends something along these lines: we should aim to make the endorsement model true, even though it very likely isn’t now. In contrast to the project of conceptual engineering, where we try to grapple with which concepts might serve us better or worse along various dimensions, the present project is one of speech-act engineering, a project where we ask what kinds of speech acts and accompanying social norms will serve us better or worse along various dimensions. Like conceptual engineering, speech-act engineering isn’t a new philosophical endeavor. Rather, it is a part of a broader, long-standing project of normative engineering, which is a project of asking which sorts of norms—be they political, ethical, epistemic, legal, aesthetic, etc.—we ought to adopt in order to promote goods of various kinds, or, perhaps, just the good in general. For example, Plato’s Republic contains an extended inquiry, not into what the norms surrounding family life were like in ancient

Athens, but rather, into what such norms ought to be like to help promote a more ideal political p. 620

arrangement.

Likewise, political philosophers tend to be interested not in what our norms of distribution

are now, but rather in which norms would serve to maximally promote justice or overall well-being under one or another set of idealizations. The interesting thing about speech-act engineering in the present context is that we are dealing with a very recent, and explicitly constructed, set of speech acts—something that holds out the possibility that these speech acts could be practically re-engineered in ways di

cult to conceive of when it comes to, e.g.,

assertions or commands. Twitter tweaks its algorithm all the time, and the notions of ‘liking’ and ‘retweeting’ are hardly immutable; these are acts that were introduced into the platform at a particular time, whose feel and e ects can be and have been altered via the adjustment of a few (thousand, or tens or hundreds of thousands of) lines of code, or via the introduction of external moderation of the platform. In other words, this is a place where philosophers have—at least in principle—a relatively clear picture of some of the mechanisms by which normative changes might be a ected. So let us ask: Given what we know about Twitter today, how might we hope to re-engineer the retweet? Rini o ers what we take to be some helpful initial suggestions here, starting with two concrete claims: (i) we ought to re-engineer retweeting to inhibit the spread of fake news on Twitter; and (ii) coming to a point where retweeting is seen as endorsement and carries accountability norms in accord with this will help us to realize this aim. Here is what Rini has to say on the matter: If we rmly established the norm that social media sharers are understood as conveying testimonial endorsement, then people would be less likely to share unveri ed stories, to avoid later being held responsible for errors. Alternatively, if we rmly established the norm that social media shares (without further comment) communicate no testimonial endorsement whatsoever, then people would be less likely to come to believe fake news on the basis of their friends’ transmissions. (E-55) Rini discounts the second alternative for being unrealistic, and advocates pursuit of the rst. She suggests that the establishment of an endorsement norm would allow us to deploy our ordinary ways of holding people to account for what they say—by applying to retweets the forms of criticism ordinarily suited to testimony. That, she reasons, should help to prevent the proliferation of fake news on the platform. The question we want to ask is: how might we hope to operationalize Rini’s proposal? That is, what might Twitter, Twitter users collectively, or the government regulatory apparatus do to establish a testimonial norm for retweeting? One very high-level suggestion would just be to advocate for a collective shift of consciousness; obviously, whether a bare retweet constitutes an endorsement depends in part on the norms surrounding bare retweeting, and those in turn depend partly on our attitudes towards tweeting and retweeting. Rini suggests one way that Twitter might help to promote the adoption of such a collective shift of consciousness: Twitter might provide infrastructure to track the testimonial reliability and reputation of particular users (2018: Ep. 621

57). This suggestion e ectively attempts to mirror, within Twitter, the

way that our pre-social-media

norms of endorsement function and are maintained: we track, both individually and collectively, the reliability of our interlocutors when they make assertions and endorse others’ claims. If Twitter were to mimic this o

ine behavior, this line of thought runs, then we can imagine that, over time, we might come

to treat (at least many) retweets as speech acts which are subject to evaluations in terms of reliability. Another natural option would be to suggest a ban on pure retweeting. All retweeting could require the retweeter to add a coherent comment. One would need an algorithm to prevent people from circumventing

this ban by simply mashing the keyboard or writing gibberish—but, presumably, at least some of that could be prevented. One worry, however, is that this would invite a sort of cat-and-mouse approach by subgroups looking to spread stories without taking full responsibility for their spread. So one might instead think that the thing to do is to require that all tweets be valenced—with something like a thumbs-up, thumbs-down, laughter, and perhaps a range of other reactions. Or perhaps one should be forced to commit to a take on how accurate the thing is that one is sharing. So a bare retweet would come with a tag like ‘very accurate’ or ‘unsure of the accuracy’. Granted, this sort of re-engineering would likely be put to some comical e ects when it comes to retweeting cute cat photos (very accurate!), but if our goal is to slow the spread of fake news, then perhaps that’s a price we should be willing to pay. How helpful should we expect any of these suggestions to be, if implemented? We’re rather skeptical that any would be of much help—though we are happy to grant that this is ultimately an empirical question, and 7

one that could fairly readily be tested. Our qualms stem mostly from the observation that, in our current, highly partisan atmosphere, fake news seems to have become largely tailored to those with a certain package of social and political views and a tendency towards skepticism when it comes to those whom they identify as of a di erent political identity. So criticism, either personal or institutional, of one’s endorsement of some fake content seems to us likely to have the opposite e ect of what we might have hoped: rather than prompting the targets of that content to tamp down their credence in it, it may well reinforce their credence and prompt them to decrease their trust in the ‘other side’, be that some individuals or rather Twitter itself. Again, controlled tests of Twitter-like platforms would likely tell us much more about the conditions under which tweaks like these might attain their desired e ects and when they might not. None of this is meant to constitute a direct argument against Rini’s proposal, or the more general idea that 8

we ought to try and make it the case that retweets are subject to the same norms as endorsements. Rather, our aim here is just to point out that projects in speech-act engineering, like other projects in normative engineering, are subject to an implementation challenge. The interesting thing about these projects when it comes to speech acts that arise in highly engineered environments like Twitter, is that we could plausibly answer questions like ‘Would doing X reduce the spread of Y?’ with some degree of accuracy without having to undertake massive social change. None of this is to claim that the answers to questions like this one are p. 622

likely to be either easy or

intuitive—just that they are somewhat more tractable than many that we as

philosophers have considered in the past.

4. A Broader View of the Speech-Act Engineering Project So far, we have followed Rini in accepting that the aim of re-engineering the retweet is to reduce the spread of fake news and disinformation. And while we certainly agree that this ought to be one important aim of 9

this project, we strongly suspect that there will be others as well. So let us step back and ask: what other problems are there that might be addressed by re-engineering the retweet? What bene ts are there to be had? And what norms would serve to minimize these problems while maximizing the relevant bene ts? In other words, what is the weighted overall aim of social media (Twitter, in this case), and what sorts of norms will best serve to promote that weighted overall aim? To ask this set of questions is, e ectively, to acknowledge another aspect of the scope and di

culty of the

project of engineering the retweet. To carry out that project, we will need to identify and weigh a range of di erent aims we might have for the use/practice of Twitter. We also need to determine which norms for retweeting are most conducive to each of these various aims and how those norms interact with each other. A comprehensive treatment of this kind is too much to attempt in this chapter. Indeed, as we pointed out in the previous section, determining which norms for retweeting are conducive to any given aim (for example,

whether an endorsement norm is conducive to the aim of limiting fake news) will involve substantial empirical work. So if we are con ned to our armchairs, we must in any event content ourselves with working hypotheses about what sorts of speech norms should be engineered. Here we will start with a still more modest plan. In section 4.1, we will consider a speech act engineering proposal that seems to be favored by some very di erent aims from those that might favor Rini’s. On this proposal, the norm for retweeting to be engineered ts with what we earlier called the indication account. We rejected this as a descriptive proposal about the illocutionary force of retweeting. By considering it now as an engineering proposal, we will explore one set of aims it might promote, which probably do not include the aim of limiting the spread of fake news. Still, these aims are worthwhile. This exercise will serve to emphasize that a broad view of the proper aims for social media is needed in order to carry out the speech 10

act engineering project successfully.

We will close, in section 4.2, by suggesting that in light of this requirement, there is much to recommend not working to attach speci c norms to retweeting as such. Instead, it might behoove us to work toward increased appreciation that retweeting itself is just a use of a exible communicative tool (like uttering “Snow is white”), which can itself be put to various further purposes. Using this tool without any evident p. 623

purpose is thus (at

least) an odd thing to do (like uttering “Snow is white” without any evident purpose).

As Savannah Guthrie pointed out to President Trump in the epigraph, it is something you might expect and tolerate from a ‘crazy uncle’, but which one ought not to tolerate from experts, politicians, and the like. More importantly, though, this outlook suggests that the norms for retweeting we should work toward will be local norms, tailored to di erent social networks, topics, and social positions of the retweeters. Accordingly, this will mean that responsibility for one’s retweets is a highly contextualized a air, depending not only on the fact of having performed the basic act of retweeting, but on which illocutionary acts are thereby performed in a given situation, and what the signi cance of these acts is given the topic of discussion and public roles of the participants.

4.1. Aims Promoted by Indication-Type Norms Twitter, we contend, can be productively viewed as a kind of social system, or game even, within which various actions are available (cf. Nguyen 2021). Certain aspects of this game are arguably rather negative, but thinking about Twitter as a collective, participatory activity provides a di erent lens on its currently prevailing norms, as well as on the ways in which those norms should be changed. It also prompts the question of what we all are collectively doing by being active on Twitter, which might not merely be the conjunction of what each of us individually are doing. (And we can ask the same about other social media platforms, and about social media in general.) One idea is that we are collectively structuring available perspectives on the internet. We need to create these perspectives because the internet is such a vast space of information. One way to interact with this sea of information is to try to nd some speci c information within it, for instance by using search engines or knowledge of relevant websites to tell you where to look. But another way to interact with the internet is simply to go online and see what is there, without looking for anything in particular. The di erence is analogous to the di erence between looking for a black cat in your visual eld as opposed to just taking in your visual eld without looking for anything in particular. But the internet does not have a ready analog for the visual eld—a limited space of information over which one may (voluntarily or involuntarily) direct one’s attention. In order to go online and see what is there, each of us needs some way of creating an online visual eld. Twitter, and social media more generally, is a common tool for this. We rely on other people—those we (mostly) choose to be connected to, and by extension those they are connected to—to structure our attention online. Collectively, then, one thing we are doing with social media is creating perspectives, or viewpoints, on the internet.

It is not so far-fetched to suppose that creating these perspectives on the internet is also an aim we should have for Twitter. The internet o ers its users the opportunity to gain a broader, less ltered awareness of political and cultural events, opinions and perspectives, works of art, and di erent ways of life than was p. 624

ever possible before. But

some narrowing and ltering is required for users to be able to realize this

opportunity. Twitter (among other social media platforms) has the potential to let users provide this narrowing and ltering for one another, without the need to appoint any authorities or rely solely on algorithms to do the job. It is plausible (but, again, ultimately an empirical matter) that establishing an indication-type norm for retweeting would promote this aim. If retweeting were universally seen as an act of making manifest one’s intention to make salient the retweeted content by replicating it (in the distinctive way enabled by Twitter), then the norm for retweeting might be something like this: retweet only if you wish to promote the visibility and accessibility of the original tweet within the Twitter perspectives of your audience. Establishment of this norm would, at least to some degree, free Twitter users from worries about accuracy and appropriateness that might otherwise stop some of them from retweeting tweets they nd worthy of visibility. The establishment of this norm would transform retweeting into a quick and easy ‘vote’ to keep a piece of content present in people’s online visual elds. Establishment of this norm would also free Twitter users from any felt demand to comment on retweeted content. To satisfy this norm, it is enough to want the tweet to be visible and accessible, and one’s reasons for wanting that are another matter. Indeed, commenting or otherwise making explicit one’s attitude to the retweeted content could interfere with the aim of structuring perspectives on the internet, since it could have the e ect of making the comment, rather 11

than the retweeted content, the primary or ‘at-issue’ content of the retweet.

It is worth emphasizing that the aim of creating perspectives on the internet is also the aim of creating what we might call ‘action-points’. In retweeting, one does not just promote the visibility of the original tweet, one also provides an opportunity to interact with it. This interaction can take many di erent forms, from further pure retweeting to liking to replying to commented retweeting. Thus, if the present hypothesis is correct, establishing an indication-type norm of retweeting would not only promote the aim of providing people with a manageable view of the internet, but also the aim of providing them with a manageable way to interact with the internet and further develop the available viewpoints and action-points. The goal of this section has not been to argue that we ought to promote an indication-type norm. Promotion of such a norm could have signi cant downsides, such as the legitimation of attempts to spread disinformation along the lines of Trump’s answer to Samantha Guthrie’s questioning (see the epigraph). What we wish to take away from the foregoing is that promotion of such a norm could also have upsides, and could help to ful ll some legitimate aims we might have for Twitter and other forms of social media. The broader lesson is that it seems inevitable that di erent legitimate aims will favor di erent norms for retweeting. Thus, the project of engineering norms for retweeting requires not only determining which norms are favored by certain aims, but also determining which aims should be given the most weight in driving the project, and how di erent aims should be balanced against one another when they inevitably come into con ict.

p. 625

4.2. Normativizing the Basic Act Approach With the complexity of the project of speech-act engineering in better view, it seems clear that one thing we might want is for retweeting to be governed by di erent norms in di erent situations. For instance, we might want the retweeting of certain types of content to be governed by di erent norms from the retweeting of other types of content. Perhaps it would be best for the retweeting of (apparent) news articles to have an endorsement or assertion norm, while the retweeting of opinion pieces or funny stories should have an indication-type norm. But from here, it is only a small step to noticing that we will probably want the norms to depend on the identity and status of the retweeter, as well. The president of the United States’ retweeting someone’s opinion gives that opinion a prominence a little known private individual’s retweet could never o er. It would probably be good if Guthrie were right that the norms of retweeting are violated when the president retweets something that he merely wishes to “put out there” but not endorse. Nor will the need for subtler norms stop with a dependence on the nature of the retweeted content, or the position and status of the retweeter. We will want di erent norms for those retweeting a content friendly to their known views than for those retweeting a content unfriendly to those views; di erent norms for those retweeting on a private Twitter than on a public one; di erent norms for retweeting a public gure’s tweet versus a little known person’s tweet. And so on. In short, we are going to want the norms for retweeting to be as variable and situation-dependent as the norms for uttering sentences. This suggests that the quest to choose an illocutionary category as the target one for retweeting, and to try to in uence people’s behavior so that retweeting in general will be viewed as part of that category, is probably misguided. What, then, should the normative engineering project for retweeting look like? It should focus, we think, on instituting speci c, local interventions or nudges designed to encourage the development of desirable norms for retweeting in speci c kinds of circumstances. As we have been at pains to emphasize, the questions of which nudges or other structural changes to platforms would encourage the development of which norms, and of which norms would lead to the achievement of di erent aims and goals for Twitter, are largely empirical questions. Nonetheless, we’ll brie y outline some concrete suggestions in the spirit of 12

illustrating the shape of the speech act engineering project we favor.

We o er this outline in a similar spirit to Rini’s proposal: as an illustration of the kind of thing that we think would be productive for social media companies to test. In contrast to Rini, however, our proposal targets speci c communicative circumstances with the aim of promoting certain norms for retweeting in those circumstances—as opposed to retweeting in general. So, for the sake of illustration, let’s consider the aim of reducing the spread of fake news on Twitter, while safeguarding the attention-structuring role of retweeting that we discussed in section 4.1. To do this, we p. 626

should develop interventions that are targeted at a

certain type of tweet: those that are apt to be treated

as news. This type of tweet has many sub-types. It will include tweets of article links from news websites, or websites that appear to be news websites. It will include tweets from journalists and media companies directly reporting what they present as news. It might include tweets from state and industry actors that simultaneously make and report news, as when Donald Trump tweeted on election night 2020, “I will be making a statement tonight. A big WIN!” With this tweet, Trump made news by announcing that he would make a statement and simultaneously reported that news. (He also falsely implied that it was already clear he would win the election.) Announcement tweets by people and institutions whose announcements constitute news of wide interest have a status similar to press releases or press brie ngs. Interventions could be targeted at retweets of any of these types of tweets via algorithms developed to recognize such retweets.

Having singled out news retweets, we would then want to categorize the retweeters according to in uence. Various dimensions of in uence could and should be recognized. Let’s just consider three: number of Twitter followers, status as a public gure, and status as a government o

cial or o

Number of followers is an easy metric to track. Public gure status is more di

ce.

cult, but Twitter could use

legal precedents from defamation actions to label certain users as public gures. Status as a government o

cial or o

ce should be fairly straightforward to identify. With these classi cations in hand, di erent

interventions could be targeted at public gures as compared with relative unknowns, at government o

cials as compared with private citizens, and at users meeting a series of di erent thresholds for follower

numbers. For instance, users with more than 1,000 followers (only 2.12% of Twitter users) might be required to have a label show up on all news tweets and retweets, either “[Twitter name] attests to the accuracy of this piece of news” or “[Twitter name] does not attest to the accuracy of this piece of news”. Users with over 20,000 followers (only 0.06% of Twitter users) and public gures might not be allowed to tweet or retweet news without the rst label on it, attesting to its accuracy. Finally, government o and o

cials

ces could be barred from tweeting or retweeting news at all, with the exception of news-making

tweets of the kind just described. This would prevent them from using the power and prominence of their o

ce to in uence journalism as it is now practiced. Twitter, and social media more broadly, is now an

important part of standard journalistic distribution practice, so preventing government actors from in uencing the spread of news on those platforms could be seen as a way of promoting a continued free press. The categorization and identi cation of tweeters/retweeters in this way might also help to address the issue of fake accounts and their role in the spread of fake news and structuring of public 13

attention/discourse.

A system like this would leave most Twitter users free to tweet and retweet news as they like, “just to put it out there”, and to help structure each other’s perspectives on the internet. But it would be made clearer that those with signi cant in uence are expected to commit one way or the other on accuracy, and those with p. 627

even more in uence must

stand behind the accuracy of news to which they provide a platform. The only

part of this system that is sure to achieve its aims is the barring of government o retweeting news (at least, from their o

cials from tweeting and

cial accounts). For in uential tweeters of other stripes, it could

happen that the required attestations would be viewed as pro forma and mostly be ignored by users, hence failing to create a norm of endorsement or a norm of accuracy-attestation around such retweets. But it is also possible that, at least, weaker norms would be created: perhaps it would come to be expected that one would choose the accuracy attestation only if one had fairly strong credence in the accuracy of the original tweet, for instance. And if such norms did develop, they might also trickle down, perhaps in yet weaker forms, to less in uential users in their retweeting practices. This is the sense in which a measure like the one suggested would be (largely) a nudge rather than a rule.

5. Conclusion We have surveyed extant accounts of retweeting’s default illocutionary category—and argued for their descriptive inadequacy. Instead, we suggested, the best descriptive account of bare retweets is that they are a basic type of act in the Twitter environment. As basic communicative acts, without a default illocutionary force, there are few, if any, generalizations to be drawn about what makes bare retweets appropriate. We take that to be an acceptable outcome, however, since it shifts attention to what strikes us as the more interesting question: What makes bare retweeting acceptable in this or that kind of context? We have suggested that there may, at present, be no one clear answer to this. Still, there is a closely related question that is of great interest: How ought we to re-engineer this communicative environment so that users are e ectively required, or at least collectively nudged, to accept certain norms of conduct in certain contexts? As a

rst step, we suggested that we should better di erentiate certain sorts of communicative contexts that will regularly arise on Twitter. In particular, we proposed to di erentiate retweets based on who is doing the retweeting and what they are retweeting, with di erent or stricter standards applied to better-known individuals or those occupying certain social roles or public o

ces when they retweet items of broad

epistemic importance. This, it seems to us, is likely to be a good starting point for improving our collective epistemic situation—assuming that we continue to rely on platforms like Twitter as an important aspect of our epistemic engagement with the world. Undoubtedly, more than this will need to be done, including implementing some of the further distinctions we mentioned in the previous section. And here, in contrast with many other parts of philosophy, we take there to be real scope for philosophers to put forward concrete suggestions that might well help to improve our future social media landscape.

p. 628

p. 629

Notes *

Though the publication date of this article is 2023, this article was written in 2019–2020. As a result, some facts about Twitter, tweeting, and retweeting have changed in ways not reflected by the language of the article—not least, Twitter is now called “X”, tweets are called “posts”, and retweets are called “reposts”. These changes do not significantly a ect the substance of this article, however.

2.

That said, we do believe that the new forms of interaction introduced by way of social media—the new ways of asserting or requesting—may eventually change, even quite drastically or fundamentally, even well-established speech acts like asserting or commanding. We will have to defer consideration of this issue to another occasion, however.

3.

In what follows we will o en use simply “retweet” instead of “bare retweet”. We will always mean bare retweets unless otherwise specified.

4.

In choosing these alternatives, we follow Arielli (2018) and Marsili (2020).

5.

In his discussion, Marsili focuses on a narrow notion of endorsement, according to which endorsing a tweet is expressing agreement with the content of the tweet. He argues that if retweeting meant something like “I agree with the content of this [original tweet]”, then it would seem contradictory to claim that one does not agree with a tweet that one retweeted; but it does not seem contradictory to do this. He also argues that if retweeting meant something like this, then most tweets would not be appropriate to retweet, since most tweets are not of the right form to be agreed with. Here we are suggesting that the first type of argument works equally well against analogies with various ways of expressing endorsement or approval more generally.

6.

Thanks to Rachel Fraser and audience members at the Jowett Society for pressing an objection along these lines.

7.

Some data might come from studying Twitterʼs current initiative to reduce misinformation online in the run-up to and a ermath of the US presidential election, which involves, among other measures, changing the retweet function so that whenever someone retweets a tweet, they are automatically prompted to make it a commented retweet rather than a pure retweet, by adding their own commentary. This is not as extreme a change as the one suggested above, which would force commented retweeting (it is still possible to do a pure retweet by leaving the comment field blank). But it would be interesting to study whether this change leads to less pure retweeting, and whether it leads to less retweeting in general of misinformation.

8.

Indeed, since we take Riniʼs proposal to be introduced primarily for the purpose of illustration, we hardly think that Rini would disagree with anything we have claimed here (E-58).

9.

Though we disagree on other important issues (cf. Pepp, Michaelson and Sterken 2022 and Habgoode-Coote 2022), we take this to be a point well made in Habgood-Coote (2019).

10.

Note: as above, the hypotheses we will suggest about the relation between norms for retweeting and aims for Twitter are just that, and are not worth much until they are tested empirically. The point of the discussion in section 4.1 is to illustrate a plausible scenario in which di erent, worthwhile aims for Twitter could favor di erent norms for retweeting.

11.

Indeed, this e ect is, in a certain way, built into the structure of commented retweets (i.e., ʻquote retweetsʼ), since they

create a new post that gets attention, acquiring likes and further retweets instead of the original. Because of this, many artists who promote their artwork in part by accumulating such statistics and monitoring comments on their own tweeted work view quote retweeting as damaging and rude. See, for example, Jacob Kastrenakes, 2020, ʻArtists Are Irked by Twitterʼs Change to Retweetsʼ, The Verge, October 15, https://www.theverge.com/2020/10/21/21527101/twitter-retweetchanges-artists-quote-retweet-qrt; and https://twitter.com/shiroganejpg/status/1113599683035897857?lang=en. 12.

It is worth pointing out that the need for empirical input into this project—to understand both the link between interventions and norm development, and the link between norm development and outcomes—means that it is unavoidably an interdisciplinary project, requiring collaboration with social and information scientists.

13.

None of this is to say that we think Twitter is at all likely to take up our suggestions.

References Ambrose, Alice (1952), ʻLinguistic Approaches to Philosophical Problemsʼ, Journal of Philosophy, 49 (9), 289–301. Google Scholar WorldCat Arielli, Emanuele (2018), ʻSharing as Speech Actʼ, Versus, 47 (2), 243–258. Google Scholar WorldCat Austin, J. L. (1975), How to Do Things with Words. Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Bach, Kent, and Harnish, Robert M. (1979), Linguistic Communication and Speech Acts. Cambridge, MA: MIT Press. Google Scholar Google Preview WorldCat COPAC Cappelen, Herman, and Plunkett, David (2020), ʻIntroduction: A Guided Tour of Conceptual Engineering and Conceptual Ethicsʼ, in Alexis Burgess, Herman Cappelen, and David Plunkett, eds., Conceptual Engineering and Conceptual Ethics. Oxford: Oxford University Press, 1–26. Google Scholar Google Preview WorldCat COPAC Eklund, Matti (2021), ʻConceptual Engineering in Philosophyʼ, in Justin Khoo and Rachel Sterken, eds., The Routledge Handbook of Social and Political Philosophy of Language. New York: Routledge, 15–30. Google Scholar Google Preview WorldCat COPAC Habgood-Coote, Joshua (2019), ʻStop Talking about Fake News!ʼ Inquiry, 62 (9–10), 1033–1065. Google Scholar WorldCat Habgood-Coote, Joshua (2022), ʻFake News, Conceptual Engineering, and Linguistic Resistance: Reply to Pepp, Michaelson and Sterken, and Brownʼ, Inquiry, 65 (4), 488–516. Google Scholar WorldCat Marsili, Neri. (2021), ʻRetweeting: Its Linguistic and Epistemic Valueʼ, Synthese, 198, 10457–10483. Google Scholar WorldCat Nguyen, C. Thi. (2021), ʻHow Twitter Gamifies Communicationʼ, in Jennifer Lackey, eds., Applied Communication. New York: Oxford University Press, 410–436. Google Scholar Google Preview WorldCat COPAC p. 630

Pepp, Jessica, Michaelson, Eliot, and Sterken, Rachel K. (2019). ʻWhatʼs New about Fake Newsʼ, Journal of Ethics and Social Philosophy, 16, 67. Google Scholar WorldCat Pepp, Jessica, Michaelson, Eliot, and Sterken, Rachel K. (2022). ʻWhy We Should Keep Talking about Fake Newsʼ, Inquiry, 65 (4), 471–487. Google Scholar WorldCat Rini, Regina (2017). ʻFake News and Partisan Epistemologyʼ, Kennedy Institute of Ethics Journal, 27(2), E–43. Google Scholar WorldCat

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

p. 631

28 Ideology and Intersectionality  Matthew McKeever https://doi.org/10.1093/oxfordhb/9780192844118.013.54 Published: 22 May 2024

Pages 631–652

Abstract Analytic philosophers increasingly make reference to the concept of ideology to think about how representational structures can lead to oppression. The aims of this chapter are twofold. First, it seeks to serve as an introduction to (some of) the best contemporary work on ideology in the analytic tradition and the theories of ideology it develops. Second, it proposes a novel challenge for any such theory. The challenge turns on the nature of intersectionality: it is hard to see how to render consistent the claim that ideology creates or sustains oppression with the claim that oppression is intersectional, without making substantial modi cations to extant theories.

Keywords: ideology, intersectionality, oppression, Sally Haslanger, Tommie Shelby Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

1. Introduction Much recent and interesting work in philosophy of language is concerned with how language is misused. Some of this work is, roughly, continuous with previous work that presupposed what we could call an ‘ideal speech situation’: a picture of communication according to which language was a vehicle for truth, and truths were slowly added to the common stock in a conversation by orderly utterances, following Grice’s conversational maxims (à la, respectively, Stalnaker 1978; Grice 1989; see Cappelen and Dever 2021 for this framing of the eld of contemporary philosophy of language). Thus, people like Sarah Jane Leslie, in addition to doing important work on the semantics of generics (2008), also published important and in uential work on how those peculiar semantics can be used to oppress (2017). One way of reading the work of Jason Stanley is as focusing on getting right what exactly semantic content, the content literally expressed by a given sentence in a given context, amounts to. Early Stanley (1997) was concerned with how much of a contribution compositional semantic value gives to what is said, or, again, to what extent the 1

acceptability judgements of certain subtle binding speak to the contextualism/minimalism debate (Stanley and Szabó 2000). Late Stanley (2015) seeks to explain how propaganda can bring about changes in what speakers accept without rationally persuading them of the truth, and he does so by joining together feminist philosophy of language (such as Langton and West (1999) and contemporary work in formal semantics (Murray 2014). Sally Haslanger’s hugely in uential work (some collected in Haslanger 2012) yokes together the externalist tradition in philosophy of language with novel theorizing about the nature of social ontology to propose her ‘revisionary’ or ‘ameliorative’ analyses. The thought here, roughly, is that even if certain words, such as ‘woman’, aren’t misused, they are nevertheless used in a way they shouldn’t be, because they are used in a way that obscures facts about the social world, such as that women p. 632

are often oppressed by

virtue of being women. An ameliorative analysis of ‘woman’ highlights that fact by making it part of the meaning of the expression that to be a woman is to be oppressed; for the ameliorative analysis, our everyday ‘woman’ is inadequate for our social and political ends. That’s by no means all there is to this recent turn to misuse. Slurs are a newly central topic in semantics. And while they are not obligatory reading for any semanticist, they are perhaps what e-type anaphora was a few decades ago: something with its own rich literature which one needed to have mastery of to do work in adjacent elds. In that literature (for an overview of which, see Jeshion 2021), we’re still trying to get right some of the same things we try to get right about anaphora: Is slurring a semantic phenomenon (Hom 2008), and if so, how; is it pragmatic or use-theoretic (Anderson and Lepore 2013)? Does it exploit some level of content that isn’t propositional content (such as presuppositional or expressive content, or implicature [Potts 2007], among others)? For another example, consider ‘woman’. A growing literature is doing semantics with an eye to moral and political complexities of sex and gender, and in that literature one sees, for example, people adopting a contextualist theory of the ‘woman’ (Saul 2012; Diaz‐Leon 2016 and references). In both of these cases, old frameworks are used to answer new questions. If we see, at least on this way of looking at things, a sort of continuity—a swap from sluicing to slurring— but the same basic toolbox being used to think about them, in some cases, we’ve introduced new theoretical tools from other traditions. Therefore, some people use the Nietzschean/Foucaultian idea of genealogy to think about how our concepts arise, and how their etiology might detract from their accuracy in depicting the world (Srinivasan 2019; Queloz 2021). The idea here is that our representational apparatus often arises as a result of, to use Srinivasan’s phrase, ‘alethically indi erent mechanisms’ (Srinivasan 2019: 132), such as the prevailing social or moral authorities of the society we inhabit. Learning that many of our concepts are contingent on where and when we happen to nd ourselves, we ought to reduce our con dence that they are successfully tracking reality. The topic of this chapter is another import: ‘ideology’. Arising from Marx and later thinkers downstream of him, ideologies are something like distorting ways of viewing and valuing the world, imposed on us,

perhaps, by short-circuiting our rational faculties or exploiting power (as when, in a totalitarian regime, all mass media are government owned and relay the government’s messages). Ideology is a central concept for many who work in what’s sometimes called ‘social and political philosophy of language’. Sally Haslanger (2011), for example, argues that well-studied linguistically and cognitively tricky features of generics help to inculcate ideologies. Jason Stanley (2015) argues that one of the bad things about propaganda, which he conceives of as a certain type of “claim[s]” or “argument[s]” (47), is 2

that it uses and entrenches ideology.

If either Haslanger or Stanley is right, and ideology is essential to explaining certain important linguistic p. 633

devices, it is imperative, for this part of philosophy of language, that

we have a viable theory of ideology.

My argument here is that we have no such viable theory therefore these in uential views in contemporary philosophy of language are on shaky foundations. Speci cally, I have two aims in this chapter. The rst is simply to present what I take to be the best, most recent, and most in uential work on the concept of ideology. The second is to suggest that, despite the merits of the various views, there are problems with these views of ideology. Some of these are rather unsophisticated—I suggest that it’s unclear that there’s a uni ed phenomenon that all thinkers are thinking about—but one problem I will raise is, hopefully, novel and interesting. That turns on the notion of intersectionality (Crenshaw [1989] being the locus classicus). Intersectionality is a theory of how oppression and discrimination is non-additive in a peculiar way: the oppression one faces by standing in multiple discriminated-against social roles (such as being both Black and a woman) is greater, or can’t be cashed out simply in terms of the oppression one faces as a Black person and as a woman. This is a problem for theorists of ideology because, as we’ll see, the one thing they agree on is that the function of ideology is to establish or maintain relations of oppression. A natural way to read that is that racist ideology establishes or maintains racist oppression; sexist ideology, sexist oppression, and so on. But if intersectionality is true, then it’s not the case that the oppression one su ers is simply a combination of racist oppression and sexist oppression. I argue that ideology theorists can’t capture this while holding that ideology functionally determines oppressive relations. Racist ideology and sexist ideology, even taken together, can’t capture the excess oppression that results from their intersection. I conclude by suggesting ways out of this argument, pointing out that it requires us to completely rethink our concept of ideology as relatively discrete collections of representational items.

2. Preface, and Two Arguments I begin by simultaneously describing some trends in the literature on ‘ideology’ while presenting some simple arguments to make us prima facie wary of the concept.

2.1. Argument 1: No Unity My topic is whether ‘ideology’ can play the role in contemporary social philosophy of language that people like Haslanger and Stanley think it can. An obvious enough way to begin assessing this is to see whether the word ‘ideology’ picks out anything uni ed in the world. So in this section, as a pre-theoretic introduction to the concept, I will simply list what the main theorists take to be examples of ideology in the search for unity. p. 634

The list:

(*) Anti-Black racism in the United States (Shelby 2003, 2014; Haslanger 2017; Stanley 2015: 59) (*) National socialism (Stanley 2015: 3) (*) The world view, when it comes to questions of science, of someone brought up in an anti-science cult (Stanley 2015: 199) (*) Analytic political philosophy (Mills 2005) (*) The belief that “seventh grade girls who wear crop-tops are cute” (Haslanger 2007) (*) The world view of someone who owned slaves in the antebellum South about the rightness of their living conditions (Stanley 2015: 192–195) (*) The belief, of someone whose identity involves the false belief that the French are untrustworthy, that the French are not trustworthy (Stanley 2015: 198–199) (*) Capitalist society, with material inequalities (Stanley 2015) This list is non-exhaustive, and I don’t have much of an argument here, but I would just point out that prima facie the claim that there is some uni ed phenomenon is questionable. And if there is no uni ed phenomenon under discussion, then it’s unclear whether ‘ideology’ ought to be an important part of our social and political philosophy of language toolbox. Of course, there are things we can say in response: it seems that racism is for many a paradigm example; and so, even if our thinkers disagree on some cases, it might be that there’s enough agreement on the central cases to say that there’s a common subject under discussion. We would use racism as if it were a sample, and say that anything su

ciently similar to that counts as ideology. I agree: this argument is weak,

and I only really expect it to lessen one’s belief in the usefulness of ideology a small amount. Moreover, as a reader of a previous version of this chapter pointed out, this form of argument threatens to massively overgenerate, since for pretty much anything interesting enough to draw philosophers’ attention, there will be di erent and competing conceptions of what the thing is. That di erent philosophers say very di erent things about ‘meaning’ isn’t a reason to ditch the concept of ‘meaning’. Agreed. So let’s move on, and hopefully we can nd something to anchor our use of ‘ideology’ in a way that concepts such as ‘semantic value’ or ‘implicature’ help anchor and sharpen (by di erentiation) uses of ‘meaning’.

2.2. Argument 2: No Baptism Baby logic tells us that terms have intensions and extensions. A term stands for a set of objects, an extension, which is determined by some perhaps descriptive condition, an intension. The argument from disunity just presented can be seen as challenging the idea that the putative extension of ‘ideology’ has the unity we’d expect of a term that carves the social world at its joints. Turning to a putative intension of the p. 635

term, we can

make a similar challenge. In particular, we can argue that the term hasn’t been given an

intension that could determine its extension. One way to make this point is by looking at the history of the use of ‘ideology’ (and its cognates in other languages). It is hard to hear the word ‘ideology’ without a whole intellectual tradition coming to mind, coming from Marx and Engels and reverberating through their twentieth-century followers, from French thinkers like Althusser to analytic Marxists like Jon Elster to social critics and theorists of all sorts (for a very helpful list, see Shelby 2003). If a term had a very clear original usage that we could use to steer our own contemporary use, that would be a good reason retain the concept. For example, there are surely some unclarities in Aristotle’s notion of

arete, but the term still, arguably, picks out a notion that is important to his moral philosophy and is one we can use, two-thousand-plus years later, in a completely di erent context, to think about moral theory. If it were the case that ‘ideology’ were similarly even roughly de ned in some canonical text, our faith in its usefulness should be higher, and we could then use that original use to try and sift through the opposing phenomena proposed as examples of ideology to see which were most faithful to it. But it wasn’t, a fact the literature is clear on. Here’s Shelby, whose groundbreaking work we’ll look at in detail later: Philosophical reconstruction is required here, not only to defend ideology critique against its detractors, but also because Marx does not o er us anything close to a precise de nition of “ideology.” In early works, especially the collaborative writings with Engels, his general conception of ideology must be gleaned from the few examples and remarks he o ers in the context of broader discussions of historical materialism. But these examples and brief comments do not provide us with an unequivocal general conception of ideology, much less a theory of the phenomenon. In later writings, Marx, again without explicitly de ning the notion, proceeds to analyze particular forms of ideological thought—demystifying their illusions, disclosing their distinctive social functions, and explaining their relation to the material conditions that he claims causes them to be produced and widely accepted … partly on the basis of these various examples, remarks, and particular analyses of Marx and Engels that I will reconstruct the concept of ideology (2003: 154–155; my emphases) In a recent overview article, Swanson (2021) concurs: “Su

ce it to say that there is likely not a unique,

determinate Marxian use of ‘ideology’ ” (332). Noting that the examples of ideology he will consider don’t refer “to ideologies under that mode of presentation”, he explains, “[so] I make connections explicit where appropriate” (336). Both Shelby and Swanson make two points: that there is no clear de nition, and that they’ll sometimes read ‘ideology’ into work that doesn’t explicitly mention it. This seems methodologically problematic. If a term were clearly de ned, then arguably, it would be okay to read it into the work of people who don’t use the term. For a just-so example: if Kripke had, for some strange reason, never read Frege, but nevertheless frequently talked about cognitive signi cance, it would be okay to discuss his views by using p. 636

the Fregean term Sinn, or

‘sense’. And this is because Sinn is relatively well-de ned (in this just-so story,

at least): it’s the more ne-grained aspect of meaning that accounts for cognitive signi cance. By contrast, Fregean ‘colouring’ is less well de ned—Frege only talks about it in a few places, and it’s hard to get clear about exactly what it means. For that reason, I claim, it would be less acceptable to gloss any of Kripke’s remarks about meaning using the word ‘colouring’. I think this generalizes: glossing others’ work with illde ned words is risky, bearing potentially large downsides and few possible bene ts. And I suggest theorists of ideology need to be more alive to that risk.

3. Cognitivist and Not-Cognitivist Theories of Ideology The foregoing is all by way of a preface. Both those arguments are extremely provisional, and were there a clear consensus about the nature of ideology in the recent literature, certainly, the second argument would be defanged, and most probably, we could get to grips with the disunity argument (maybe some of the examples above are mislabelled, something we can only see once we see what the concept amounts to). The body of this chapter is concerned with presenting two contemporary theories of ideology. Both are very in uential and, indeed, very useful ways of understanding the social world, but both, I believe, su er from the same problem.

I will borrow from Sally Haslanger in distinguishing between cognitivist and not-cognitivist theories of 3

ideology (2017: 7 ; see also 2012: 413). Cognitivist theories focus on belief and sets of belief as the bearers of ideology. Opposing positions, such as Haslanger’s own, are more practice based. I will treat them in turn. Having done so, we’ll be able to assess the suitability of the concept of ideology for work in applied philosophy of language.

3.1. Cognitivist Theories A fundamental theorist in the cognitivist line is Tommie Shelby, whose 2003 is a contemporary classic. His view is very clear, and fundamental to the recent discussion, so we’ll spend a decent amount of time on it. He is concerned, both in his 2003 and 2014 papers, with giving an account of racism as ideology. He begins with the core component notion of a form of consciousness. A form of consciousness is a subset of the beliefs held in common by a given group, in a given time and place, with these properties: a. The beliefs in the subset are widely shared by members in the relevant group; and within the group, and sometimes outside it, the beliefs are generally known to be widely held. p. 637

b.

The beliefs form, or are derived from, a prima facie coherent system of thought, which can be descriptive and/or normative.

c. The beliefs are a part of, or shape, the general outlook and self-conception of many in the relevant group. d. The beliefs have a signi cant impact on social action and social institutions (Shelby 2003: 158) As Shelby notes, we don’t want to call any set satisfying these conditions ideological; Shelby’s example is liberalism. So we need to nd a further thing, an I-property, such that form of consciousness + I-property = ideology. He goes on to helpfully canvass various possible I-properties, from across twentieth-century 4

thought (162–163). Space precludes quoting them all, but Shelby divides them, following Geuss (1981), into three dimensions: ideology has an epistemic dimension, a functional dimension, and a genetic dimension. It has certain epistemic aws; it performs a certain function role in society, and it has a particular etiology. Shelby aims to propose an I-property that will capture these three aspects. I will treat them in turn.

3.2. The Epistemic Component Ideologies have di erent epistemic properties than forms of consciousness, Shelby says. The beliefs that comprise an ideology have or lack some epistemic properties that forms of consciousness don’t. One might think, given that racism is a paradigm ideology, that this property is simply falsity. Importantly, for Shelby this is not so. The reason for this is interesting: item (d) in his list of forms of consciousness tells us that forms of consciousness, thus ideologies, direct behavior. But at-out false beliefs, Shelby thinks, don’t successfully direct behavior. If ideological beliefs were just false, we would not be able to coordinate our actions through them as e ectively as we do. Instead, ideologies often work, as Kai Nielsen reminds us, “by presenting and inculcating a false or slanted perspective that arranges the facts in a misleading way, or fails to mention certain facts, or places them in an inconspicuous context.” Rather than being simply false, then, ideologies are typically more or less distorting or biased in some way. (Shelby 2003: 165)

He gives an example of the widely held belief “that (most) black women who receive welfare support are poor because they are lazy, irresponsible, and promiscuous,” when ample research shows that those who receive welfare assistance are in fact often victims of impeded access to housing, education, and the job market, which disproportionately a ects Black people. The key idea is that if one is blind to or simply ignores these structural factors, some of the conduct of the black ghetto poor can seem to con rm the stereotypes of racist ideology and thus to justify resentment toward black welfare recipients. (166) p. 638

I take it that the idea is that there are two competing hypotheses about the undeniable fact that some Black people rely on welfare programs. One is backed up by sociologists; and the other is peddled by right-wing mass media. Both structural inequality (decades of mistreatment, bad schools, predatory nancial institutions, the prison complex) and individual properties (laziness or irresponsibility) are, in general, good ways of explaining many social phenomena, whereas—at the risk of stating the obvious—nonsense isn’t. We are, all of us, familiar with lazy people, irresponsible people, promiscuous people, and with the fact that such people’s lives are hampered by these qualities. In light of that, suggesting these properties as general explanations of Black poverty, especially when the media bashes one over the head with it, might seem reasonable. Of course, the laziness-based explanations aren’t good ones. But, as we have seen, Shelby thinks that ideologies needn’t peddle straight-out falsehoods, as opposed to biased or defective or distorted ways of thinking. But what, exactly, is the cognitive defect that isn’t necessarily falsehood? Shelby (2003) writes: There are many types of cognitive error that are typical of ideological thinking—inconsistency, oversimpli cation, exaggeration, half-truth, equivocation, circularity, neglect of pertinent facts, false dichotomy, obfuscation, misuse of “authoritative” sources, hasty generalization, and so forth. (166) This list should be familiar to readers of this volume. Many of these cognitive errors are well viewed as violations of Gricean principles, misleading in the technical sense, and a misuse of generics. It’s by now familiar that such language misuse can lead to bad social consequences of the kinds discussed elsewhere in this handbook. Shelby’s idea that ideology is distorted or partial is interesting but questionable. It seems that he is saying that the belief that Black women on welfare are lazy, and so on, isn’t false, but merely partial and distorted, and, indeed, that this non-falsity is an important component of an ideology if it is to guide action. Arguably, that’s more than we want to admit to the racist. Surely it is false, generally speaking, that Black women are on welfare because they are lazy (of course, some might be; many people of all races are lazy). It seems that Shelby needs something with rather unique properties to play the epistemic role. We (I have just argued) want it to have some negative epistemic status worse than being merely distorted or partial (we want to judge racists qua epistemic, not just moral, agents); but we also want it to have some positive status that enables it to guide action. But then, we’re looking for a ‘Goldilocks’ epistemic property that is bad, but 5

not too bad, and good, but not too good. Prima facie that seems a tricky ask.

Though plausible, then, I think there’s cause to wonder whether the epistemic dimension of Shelby’s theory is entirely satisfying. And there’s some reason to think that the requirements he explicitly seeks and that I suggest we want don’t t well together.

p. 639

3.3. The Genetic Component At the heart of Shelby’s theory of the genesis of ideology is the Marxist idea of false consciousness. Shelby tells us: To hold a belief with a false consciousness is to hold it while being ignorant of, or self-deceived about, the real motives for why one holds it: the individual who su ers from a false consciousness would like to believe that she accepts a given belief system (solely) because of the epistemic considerations in favor of it, but, as a matter of fact, she accepts it (primarily) because of the in uence of noncognitive motives that operate, as Marx was fond of saying, “behind her back,” that is, without her conscious awareness. 6

(2003, 170)

False consciousness serves several important explanatory ends for Shelby: rst, it helps explain why people accept ideologies. Imagine a working-class White person who believes that Black women welfare recipients are on welfare because they’re lazy, irresponsible, and promiscuous. Such a person will quite likely have dealt with some of the same structural forces that have held the Black woman down—intergenerational poverty, di

culty nding jobs, failing educational and health systems. One wonders, how can they not see

the similarities? The genetic component promises an answer: Given that ideologies su er from fundamental cognitive defects and yet are widely held, it might seem that ideology-critique assumes that most humans are quite credulous and perhaps even stupid. But if ideologies are held with a false consciousness, then this un attering and elitist view of ordinary people need not be assumed. It is a mundane fact about human beings that we are sometimes prompted to accept beliefs by motives that have little to do with a concern for truth or justi cation. Though presumably we do not do so consciously, we sometimes believe things because to do so would, say, bolster our self-esteem, give us consolation, lessen anxiety, reduce cognitive dissonance, increase our self-con dence, provide cathartic relief, give us hope, or silence a guilty conscience. When these and other noncognitive motives are psychologically operative, we easily fall into epistemic error. (171) That is to say, if one is in the grip of an ideology, one holds a set of beliefs with false consciousness. That means that the explanation for one’s holding that set of beliefs is to be found in various relatively universal ‘non-cognitive’ motives, such as those Shelby mentions. The genesis of ideological beliefs, then, is those non-cognitive motives, as opposed to one’s credulity or stupidity, and this whole story is the genetic component of a theory of ideology. The genetic component, then, manages to explain why epistemically problematic ideologies can be maintained. The White person, perhaps ignorant of the cause of their poverty, will attribute it to bad luck in their case and to inherent laziness in the case of their neighbour. p. 640

It also explains, Shelby tells us, why such ideologies are di

cult to give up. To use an analogy that isn’t his,

if you have a leak you think is caused by your faucet, but that is actually a problem with the water pressure in the pipes, you’ll be less likely to be able to x it because you’ll spend all your time watching faucet videos rather than pipe videos on DIY YouTube. So it is with your system of beliefs: if you don’t know that it goes wrong because it’s held to boost your self-esteem rather than track the truth, you’ll be less likely to x any errors in your system, because you’ll nd that any real-life correction that doesn’t also boost your selfesteem is unlikely to change what you think.

Shelby’s discussion is relatively programmatic, spanning a few pages of a very rich paper. There is certainly more to be said about how non-cognitive factors in uence what we believe. For example, sticking with just philosophy, we could study the impact of evolution on the beliefs we accept (e.g., Street 2006, and much else). We could study ‘genealogy’, the already-mentioned study of how concepts come into existence in response to historical and social forces, as opposed to being merely representations of reality (Queloz 2021), We could do both together (Srinivasan 2019). Outside philosophy, swathes of behavioral psychology tells us 7

how our social position, quirks of our brain, and how we reason systematically can mislead us.

All this is to say that the idea that there is a genetic component to ideology that serves to explain why people hold defective sets of beliefs seems prima facie plausible, and one susceptible to a more detailed treatment than Shelby or I present. Hopefully, what we’ve seen here is enough to show that, and that it will su

ce for

the purposes of this paper.

3.4. The Functional Component Shelby points out that even if we have an epistemically distorted set of beliefs engendered by false consciousness, we don’t have ideology: he points to Freud and Nietzsche on their explanations of religion, which debunk religious beliefs by giving them an etiology in sub-rational processes, but which, at least arguably, aren’t talking about ideologies. And indeed, the tradition helps here, as Shelby points out that it’s common among Marxists to take an ideology to be a constellation of beliefs that “helps to establish and/or stabilize … relations of subordination” (2003: 173) such as “labor exploitation, land and resource expropriation, imperial conquest and annexation, political disenfranchisement and marginalization, social repression and exclusion, expulsion and genocide” (173). It’s these relations of subordination that round out the concept of ideology. Ideologies cause things; they have a certain function in the generation of unjust social structures: “An ideological form of social consciousness contributes to establishing or stabilizing relations of oppression in virtue of its cognitive defect(s)” (Shelby 2003: 174). This will be very important going forward, as we’ll see, it’s the core feature shared by the di erent conceptions of ideology that we’ll consider, and the core feature my criticism will focus on.

p. 641

3.5. Assessing Shelby Shelby’s pioneering work is rightly in uential. Bringing the tools of critical theory descending from Marx into an idiom analytic philosophers can feel at home in is worthwhile and has shaped the subsequent 8

discussion of ideology in the analytic tradition. Nevertheless, his view isn’t without problems. I will focus on one particularly trenchant one Sally Haslanger has repeatedly voiced, before presenting my new problem after I’ve discussed her positive view of ideology. Haslanger (2017, 2021) has made the case that Shelby’s concept of the ideological is too cognitive in several places. For example, she asks, if ideologies are just sets of beliefs, which sets are the ideological ones? In de ning ideology in terms of shared beliefs, cognitivism seems to be committed to the idea that the content of the ideology is determined by the attitudes of the majority; ideology is just what most people believe, or believe together. But how do we identify the relevant ideological beliefs? Consider an oligarchy: suppose the ruling elite is invested in an explicit ideology and structures the society to embody it; the masses may then enact it but on the basis of a completely di erent set of beliefs or even multiple divergent sets of beliefs. (2017: 6)

The thought here is that a society’s ideology can be oligarchic, even though the majority of people don’t believe that oligarchy is the right way to structure political power. If one thinks ideologies are sets of beliefs, 9

then asking the question of which ideology a given society encapsulates may give more than one answer. That is surely right. Take Stanley’s example of the capitalist, meritocratic ideology. Many readers of this

paper, though fully involved in the capitalist system, nevertheless play their role in that system with, I take it, a certain cynicism and discomfort. In Haslanger’s (2021), she again points out that shared belief, as opposed to shared presupposition, is not the right thing to put at the center of ideology. Again, consider the Stanley example: we may not believe that economic distributions accord with merit, but we often need to presuppose it to understand political discourse, or make sense of university deans’ and sport coaches’ salaries. Or, again, we presuppose things about the police force when we watch Brooklyn Nine-Nine that we don’t actually believe. In Haslanger’s words: I am fully aware, for example, that according to the contemporary ideology in my social milieu, women are (supposed to be) deferential to male peers. I disagree with this, but also sometimes abide by it, sometimes use it to my advantage, sometimes explicitly challenge it, and often aunt it. My non-conforming actions may contribute to changing the sexist ideology in my immediate context (though backlash occurs!), but the public assumption of gendered deference—and sexist ideology more generally—remains broadly entrenched. In some social contexts, the majority may p. 642

not believe the ideology, although they act in accordance with it because it is the

(dominant?

Proper? Enforced?) framework of meanings and values that is used to guide social interaction … Ideology, then, does not seem to be aptly characterized as common belief, but as what individuals accept or presuppose for the purposes of interaction (2021: 17–8; my emphasis) This seems right: if sexism is an ideology, many of us presuppose (to make sense of a vast array of social phenomena) various tenets of it without believing them. Many of us unre ectively assume, on learning that a mixed-sex couple has just had a baby, that it’s the mom who’ll take work leave and, perhaps even, that it’s better, morally speaking, that she rather than the dad does so. But on re ection, we realize that that’s a value judgement we inherited from our society, and not one we want to actively assert. It’s worth wondering exactly how Shelby would respond to Haslanger’s point about the presupposedness of ideology. There are passages where he seems to grant ideological formations this common, taken-forgrantedness that we assume of presupposition (he talks of how ideological belief “frequently constitut[es] a part of so-called common sense”, of how nothing could be more “obvious” than racist ideology in a racist society). But prima facie, it seems like a challenge Shelby’s theory needs to rise to. And having presented it, it’s time to look at Haslanger’s own theory.

4. Not-Cognitivism about Ideology: Haslanger Sally Haslanger, for at least fteen years, has been working on her own theory of ideology. The aim here is to present it. I will concentrate on her most recent presentation, from her lecture 2021 ‘Ideology in Practice’ that has been published as a short book by the Marquette University Press. At the heart of Haslanger’s theory is her idea of a ‘cultural technē’. Of these she says, ‘Cultural technē’ is the general term for a set of cultural tools; one might think of it as ideology in the descriptive sense … I do not assume that all cultural technēs are ideological. Some organize us in good and just ways; those that (roughly) organize us in oppressive ways are ideological.

(2021, fn13) And again, she tells us that a framework of meanings and ideas, what I call a cultural technē, is ideological to the extent that it produces or perpetuates oppression. In other words, oppression occurs systemically and reproduces itself; ideology is a component in the system that contributes to the system’s reproduction. The speci c role of ideology, on this account, is to mask the real workings of the system; those in the grip of an ideology have false or distorted ideas. (43–4) p. 643

For this latter point she references Shelby, and ‘framework of meanings and values’ is a repeated refrain. Again, a cultural technē is a [cluster] of concepts, background assumptions, norms, heuristics, scripts, metaphors (and so on) that enable us to interpret and organize information and coordinate action, thought, and a ect. (63) 10

Taken together, Haslanger’s framework is notably di erent from Shelby’s.

We’ve already seen Haslanger

call into question the role of shared belief in accounting for ideology. She also says, “My account of ideology is sympathetic with the criticisms that ideology should not be evaluated primarily in terms of truth and falsehood” (67). After noting that her theory is broadly Althusserian, she (approvingly) says: Althusser is very explicit that ideology is not merely a set of ideas or beliefs. In fact, it is one of his main theses: “Thesis II: Ideology has a material existence.” (1917/2014, 258). He elaborates the thesis later: “I now return to this thesis: an ideology always exists in an apparatus, and its practice, or practices. This existence is material.” (33) This is all somewhat abstract; we’ll consider an example in a second. But the important thing to note is that Haslanger clearly disagrees with the epistemic component of Shelby’s theory of ideology (ideologies aren’t 11

composed of distorted beliefs),

but she clearly agrees with the functional component, that ideologies

produce or sustain oppression. This will be central to my positive argument. As an example, Haslanger talks about being a wife, and how that is constrained by one’s adoption of a particular cultural technē. One cannot intentionally become a wife without there being an institution of marriage situated within a broader frame of cultural meaning. Navigation of social life depends on the cultural technē, i.e., it depends on a sensitivity to social norms and shared meanings that are part of the “common ground.” Because our conceptual repertoire is inevitably limited, some forms of action are unintelligible. … Butler and Hacking have o ered a way to locate a kind of agency constrained by shared “concepts and languages of practical thought” —or what we have been calling a framework of meanings and values—that have social origins. The problem is not, rst and foremost, that we misrepresent the world; rather, we lack certain conceptual tools to make apt choices. (57–58)

I think we’ve gotten somewhere. For Shelby, the genetic problem of ideology was that it was produced by certain problematic epistemic mechanisms. For Haslanger, by contrast, it seems that the problem of ideology is fundamentally a problem of lack: we lack concepts that enable us to make sense and act on bits of the world. This is not all there is to Haslanger’s view. An important part of cultural technē is what Haslanger calls ‘schema’. I confess to being less certain about her schemas, so I’m moving into interpretation from p. 644

exegesis, but I think the way to read Haslanger here is as saying

that the schema in the cultural technē

are concepts for certain social actions. But they are strange concepts: they are, I think, concepts which are required in order to perform the action which the concept stands for. That’s a mouthful, so an example: there are some actions we can’t perform but by doing them under the description of a given concept. Consider our concept of a nut. It is part of the concept that nuts are food, but that is not a bald metaphysical truth about nuts; for something to be food is not merely for it to be edible (in a vegetarian society, animals aren’t food). These schema concepts are, I think, the most important part of Haslanger’s theory of ideology. Ideology opens up and forecloses possibilities of action by supplying or withholding the concepts we need to perform actions. It is this that gives Haslanger’s theory the materiality that, as an Althusserian, she requires, and enables her to avoid what she takes to be problems with cognitivist theories like Shelby’s. In a previous paper, she expresses her view about the practice-laden and materialist theory of ideology neatly when talking about racism. Racism is an ideology, and racism, on my view, is constituted by an interconnected web of unjust social practices that unjustly disadvantage certain groups, e.g., residential segregation, police brutality, biased hiring and wage inequity, educational disadvantage, etc. These are not random practices, but are connected by a racist technē. (2017: 13) Racism is a practice. With that in mind, I think we can answer the question that has puzzled all theorists of ideology: How can we think and act in ways that are contrary to our interests? For Shelby, the answer is that we are subject to distorted beliefs. For Haslanger, the answer is that the shared conceptual scheme we have renders visible oppressive courses of action while obscuring liberatory ones. Or again, the world is given to us already rich with meanings, as a Heideggerian Zuhand object, one which—to change idioms from phenomenology to cognitive science—comes with certain a ordances, which are themselves determined by the technē. It is hard to disagree with Haslanger’s analysis when presented with the sheer data of the foreclosed possibilities many face (the world one inhabits as a poor person with the prison system hanging over one is surely a di erent world than the one most readers of this chapter live in).

4.1. Assessing Haslangerʼs View I think there are several helpful lessons to be drawn from Haslanger’s discussion. The rst is positive. We wondered at the start of the chapter whether there was, or could be, unity to the topic of ‘ideology’: whether the di erent theorists weren’t going to be just talking past one another. But Haslanger, using the Geussian tripartite framework, is completely clear as to where her theory is similar and where it is di erent from the cognitivist: she di ers from the cognitivist on the epistemic, while agreeing about the functional component (see 2017: 2–3). p. 645

We get, then, by considering the two theories in detail and up close, a kernel that is shared between the theories in terms of function. And we could use that kernel, were we so minded, to make a general de nition:

Ideology. An x that produces and sustains oppression, where I don’t use ‘x’ solely to be over-fastidious, but to indicate that there is a slot for a variable element in this de nition, lled di erently by di erent theorists. We could then argue about the particular cases introduced here as we saw t, and, having our own Marxinspired conception, it wouldn’t matter much if we couldn’t nd much in the ipsissima verba in Marx’s work. Ideology, then, is on solid ground. Well, no. I want to nish by ruining the harmony I’ve created, by presenting a general complaint, arising from a core theoretical commitment of contemporary social philosophy, that challenges the functional component and, as far as I can see, every extant theory of ideology. Before doing that, let me refocus the discussion a bit. We’ve just gone into a couple of the most important contemporary theories of ideology. But notably, we’ve said little about language. But this is a book about language. That might seem like a lacuna. The point to recall is that for many of these theorists, ideology is closely connected with language. Recent years have focused our attention on the misuse of language, and undoubtedly, two of the most famous bits of language discussed are propaganda and generics. But they are, at least for Haslanger and Stanley, vehicles 12

or products of ideology.

So the success of their theories of language turns on the success of their theories 13

of ideology. Now I give reasons to doubt that those theories are successful.

5. Intersectionality: A General Problem for Ideology 14

A founding concept of contemporary social theory is intersectionality.

Coming from legal theorist Kimberlé

Crenshaw, intersectionality is the idea that oppression or disadvantage is not, to use her word, “singleaxis” (1989: 139). Crenshaw brought this out in her seminal “Demarginalizing the Intersection of Race and Sex” by considering several court cases. The actual details are, at least to this non-lawyer, a little subtle, so I’ll concentrate on the most famous case she discussed and the one that is often used to explain the concept. In DeGra enreid v. General Motors (413 F. Supp. 142 (E.D. Mo. 1976) May 4, 1976) a number of quali ed Black women complained that Black women were not hired for various jobs they were quali ed for. That is to say, they brought a class action complaint of discrimination on behalf of the class of Black women. At the time, p. 646

there was the law prohibited against discrimination based on sex and against discrimination based

on

race, but none that combined race and sex as a legal basis for a claim. And the court claimed that was where it had to stop: The prospect of the creation of new classes of protected minorities, governed only by the mathematical principles of permutation and combination, clearly raises the prospect of opening the hackneyed Pandora’s box. (DeGra enreid, 413 F Supp at 145; quoted in Crenshaw 142) That is to say, the court refused to recognize Black women as a target of unique discrimination that wasn’t merely the ‘sum of racism and sexism’ (Crenshaw 1989: 140). This means, Crenshaw argued, the law fails to make room for Black woman as the subject of discrimination that can’t be simply covered by the separate laws against sex and against race discrimination. As Crenshaw wrote in a slightly later paper:

Many of the experiences Black women face are not subsumed within the traditional boundaries of race or gender discrimination as these boundaries are currently understood, and that the intersection of racism and sexism factors into Black women’s lives in ways that cannot be captured wholly by looking at the race or gender dimensions of those experiences separate. (1990) As this later article, and Crenshaw’s work in general, shows, intersectionality is not just an arcane legal fact. It’s a conceptual or moral issue about the nature of oppression. There are various ways to try to phrase it, but one is that the oppression one faces as a Black woman is not simply the ‘sum’ (to use Crenshaw’s word) of the oppression one faces as a woman and as a Black person. Intersectionality is widely recognized as a core analytical tool in contemporary social theory, and its insights into the nature of discrimination are ones any theory in this domain ought to capture. But the theory of ideology is in this domain: the function of ideology, the one thing the cognitivists and notcognitivists agree on, is that it leads to unjust or oppressive social arrangements. So it should capture the intersectional nature of this injustice, in particular, that the oppression one su ers as being a Black woman is not simply the sum of that which one su ers from being Black and that which one su ers from being a woman. But—and this is the main claim of the chapter—I don’t see how extant theories can manage this. Consider a —partly ctionalized and just-so—story from the early 1980s: the use of the La er curve to justify lower tax rates. The idea is that lowering tax rates can serve to increase tax receipts, for reasons one can interpolate: taxes (it is claimed!) being an incentive against work, lowering them will lead to more work, and thus more tax revenue. Although the idea that paying taxes disincentivized working was never clearly empirically con rmed, it was used as a spur to the low tax policies of Reaganomics. p. 647

Here are some facts about the La er curve: it is plausibly distortional. It might not, strictly speaking, be false that there’s an inverse relation between taxation rates and tax receipt (i.e., there might be a grain of truth to it), but any relation must be extremely quali ed. It is plausibly introduced by false consciousness. Just as racists use debunked theories of race to give a veneer of legitimacy to their prejudice, so, too, those who didn’t want to pay taxes used the sheen of the supposed science of mathematical economics to justify low-tax policies. Moreover, this helps to keep in play relations of oppression. In particular—and this is the important fact—the former widespread belief in the truth of the La er curve plausibly intersectionally oppresses Black people, by virtue of the fact that it oppresses those with low incomes, and those categories —Black people and low-income people—intersect. If tax receipts fall, then, given that the army still needs its trillions, consumption taxes rise; these taxes are regressive, hitting disproportionately the poor, and Black people were and still are disproportionately poor. The La er curve, then, is a piece of ideology that oppressed Black people. But it seems incorrect to say that the supply-side economics of the Reagan period 15

was a racist ideology.

The La er curve, if it ideologically oppresses Black people, surely does so in a

di erent, more indirect way than the roughly contemporaneous ‘welfare queen’ ideology. But surely extant theories, de ning ideology as distortional or incomplete representational items, with a particular etiology and which sustain oppression, ought to count the La er curve as ideologically oppressing Black people. All the conditions are satis ed for these theorists to call it ideological, but if it is ideological, it’s surely not in virtue of the way it oppresses Black people (again, by contrast, ‘welfare queen’ ideology would rightly count as one that is ideological in virtue of how it oppresses Black people). Extant theories can’t distinguish La er curve ideology and the welfare queen ideology, and one diagnosis for this is that its conditions are too weak to capture the intersectionality of oppression. Since that’s surely a fundamental desideratum of a theory of ideology, extant theories are insu

cient.

Here’s another example. Let us take Crenshaw’s original case, which involves the oppression a Black woman encounters by virtue of being a Black woman. For the ideologist, this results from the presence in society of an ideology which has this oppressive function. But how? It’s not as if we can take the racist ideology, and take the sexist ideology, and work out—excuse the awkward phrasing—the contribution each makes, and then add them. No: the contribution is greater than what each ideology, taken separately, sums to. I don’t 16

see how the ‘ideology’ ideology has the tools to capture this fact.

The reader might be nonplussed. You might grant all the above and say that it’s simply a re ection of familiar granularity issues surrounding, in this case, the determination relation holding between ideology and oppressive outcomes. I agree that this sounds promising: but granularity issues are still issues that need to be solved, and it’s not at rst glance obvious what is the right thing for the intersectional ideology theorist to say. I will nish up by considering one plausible line of counterargument.

p. 648

6. A Tentative Conclusion: Holism about the Ideological How should the ideology theorist respond to this? How can they make room for intersectionality in the functional part of their theory? As far as I can see, there are two moves available here. One can sever the functional link between ideology and oppression. If one does that—and having presented the accounts, I hope this worry is tangible—one falls afoul of the lack of unity criticisms which started this chapter. The one thing I think we can con dently ascribe both to the cognitivist and Haslanger-esque theoreticians is the idea that ideology seems to engender or support oppressive social arrangements. But if I’m right, that can’t be straightforwardly true. To the extent that we have one, the mereology of sets of beliefs is simple and arithmetic: one belief and one belief is two beliefs. But ideological beliefs determine oppression, we’re told. And oppression isn’t arithmetic in this way. One can su er extra oppression by landing at the intersection of oppressed characteristics. This means that we can’t simply determine someone’s oppression (I pardon the somewhat awkward phrasing) by looking to their, say, racial oppression, and then looking to their gender oppression, and just ‘adding them up’. There’s an extra that comes in by virtue of the fact that they are both women and of color, say. That extra is something over and above, something that I don’t believe any theory of ideology we’ve canvassed will be able to capture. Crenshaw, interviewed by Vox, tells us: Intersectionality was a prism to bring to light dynamics within discrimination law that weren’t being appreciated by the courts. In particular, courts seem to think that race discrimination was what happened to all black people across gender and sex discrimination was what happened to all women, and if that is your framework, of course, what happens to black women and other women of color is going to be di

17

cult to see.

The big worry is that analytic Ideologie-Kritik, attempting to nd and dissolve oppression by looking at ideology, will like the courts overlook the intersectionality of oppression. But there’s a response available: holism about the ideological. In a sense, if you’re really on board with the intersectionality of oppression, and the link between ideology and oppression, then this should seem attractive. Ideological beliefs—to mimic Quine—come to the tribunal of oppression altogether, and we can only assess ideology in toto, as the sum and intersection of all the given ideologies oating about in society at a time. And that in turn is going to have consequences for philosophy of language. Theorizing about generics or about propaganda, or, in general, any form of language implicated in ideology, will have to somehow be done together. An ideological holism might lead to something in the vicinity of a semantic holism.

p. 649

This is nothing more than suggestive, and to fully make the case requires a paper of its own. For what it’s worth, at least to my moral sensibility, the view seems reasonable. If the mereology of oppression is as Crenshaw says, then why not, too, the mereology of content, given the links between the two? But one thing is clear: none of the paradigm ideologies with which we began—National socialism, racism, and so on— would count as ideologies in this new and strict sense, and the relation between particular linguistic forms and ideology must be considerably more complicated than has been considered so far. Ideology, if the argument of this paper is right, must be a holistic phenomenon, or it must be nothing.

Acknowledgements I thank Herman Cappelen, Daniel Vanello, and two readers. I especially thank the reader who marked up the manuscript as that improved the paper a lot. Finally, OUP’s copyeditor did a great job helping me clarify my meaning in a bunch of places.

Notes

p. 650

1.

Such as the putative literal truth of sentences like ʻevery teacher praised every studentʼ in a context where one is assessing how each teacher behaved with regard to their and only their class. The felt content is something like ʻevery teacher in the school praised every student in their classʼ, although it is controversial, to say the least, that this is the actual, literally expressed, content.

2.

See also Swanson (2019) for some other cases in which itʼs argued that language expresses ideology.

3.

ʻCognitivistʼ is Haslangerʼs word; ʻnot-cognitivistʼ is my ugly coinage to avoid ʻnon-cognitivistʼ given that itʼs already taken by metaethicists.

4.

Now is as good a time as any to point out that the literature on ideology within and without philosophy is absolutely vast, and I canʼt hope to cover it adequately. What I can hope to do is focus in on some salient and famous discussions, hopefully provide a preliminary map of the logical space, and advance a criticism that will apply to many views, albeit perhaps not all, including many not discussed here.

5.

See, however, Rahel Jaeggi on “the interpenetration of true and false” (2009: 67). Jaeggi is an important and influential ideology theorist whose work I donʼt discuss only for reasons of space.

6.

Engels is, arguably, slightly helpful here: ʻIdeology is a process accomplished by the so-called thinker consciously, indeed, but with a false consciousness. The real motives impelling him remain unknown to him, otherwise it would not be an ideological process at all. Hence he imagines false or apparent motives.ʼ Marx-Engels Correspondence 1893, Engels to Franz Mehring, London, July 14, 1893, Marxist Archive, https://www.marxists.org/archive/marx/works/1893/letters/93_07_14.htm.

7.

I hesitate to cite work here as an outsider, since I donʼt know what has been replicated, but the sort of thing I have in mind is the work made famous by Kahneman (2011) or by Haidt (2012).

8.

Space prevents me from considering the third big contemporary ideology theorist, Jason Stanley. I think this is acceptable as he admits that his theory is “influenced principally” by Haslanger and Shelby (Stanley 2015: 184). Although he doesnʼt use the Geussian framework, his view is clearly cognitive, ideology being marked by beliefs that are hard to rationally revise (epistemic component), and has a genetic component, the beliefs deriving from self-interest (see e.g., 2015: 200 for both these claims). His view also has a functional component, found throughout the book (see 203, and the talk of groups ʻhindered from acting in their own self-interestʼ for one manifestation of this). I believe my main criticism goes through against him, but wonʼt make the case here.

9.

Of course, one could just accept this, perhaps by saying that a society can house more than one ideology at a time. As my weaselly phrasing (ʻhouseʼ, ʻencapsulatesʼ) suggests, there are worthwhile questions about how we should think of the

relation between a given society and the ideology or ideologies that may be present in it. I have nothing informative to say about this, but I hope that for neither my expository nor critical purposes will this matter. 10.

For more on precursors to Haslangerʼs view, one could consult the work of Rahel Jaeggi (2006) and Robin Celikates (e.g., 2006). That work reveals that the position presented as Haslangerʼs can be found in disciplines such as anthropology. Ignorance and space preclude further discussion.

11.

Actually, thatʼs a bit quick. A reader pointed out that Haslanger might merely deny the necessity, for somethingʼs being ideology, that it be a set of beliefs with the characteristics Shelby lays out. She might be happy to say that itʼs su icient for something to be an ideology that it be a set of beliefs with those characteristics. That might be right; what seems, nevertheless, true is that for Haslanger the really interesting part of the theory of ideology, the thing she is most interested in, is notably not-cognitivist. I discuss in the rest of this section.

12.

What about Shelby? Itʼs less clear that he links language and ideology. One could argue that in his 2014, where he argues that racism is an ideology, he is thereby committing himself, at least weakly, to the claim that names of races (which may, of course, be empty names) are or can be components of ideologies.

13.

Some interesting extant criticisms of the use of ideology include Srinivasan (2016) and Sankaran (2020).

14.

For a helpful discussion of ways of understanding intersectionality, and how one can formalize these understanding using certain formal tools, see Bright et al. (2016).

15.

One can deny this! Or at least question it, as a reader did. One could hold that supply-side economics is a racist ideology, albeit one less blatant than the ʻwelfare queenʼ ideology. While Iʼm not sure about that, Iʼm tempted to just turn it into a dilemma. If one denies my premise that supply-side economics is not a racist ideology, then that personʼs concept of racist ideology is su iciently wide, I think, that earlier considerations about the unity of the concept start to reassert themselves.

16.

A reader alerts me to a body of work that might help here, one that goes metonymically by ʻkyriarchyʼ. As I understand it, the idea is that the atomist, additive framing that I assume (separate racist and sexist ideologies we can perform [perhaps very non-classical, if weʼre to account for intersectionality] mereological operations on) is a misrepresentation, and therefore, we need to instead consider oppressive social roles as inherently complex: as racist-sexist systems, where this canʼt be factored out into separate oppressive roles. If this is so, great! I think what Iʼm about to suggest fits very nicely with this, and o ers a vision for future work I hope to pursue that blends all of intersectionality, ideology, kyriarchy, and philosophy of language together.

17.

“The Intersectionality Wars”, Vox.com, Jane Coaston, May 28, 2019 https://www.vox.com/thehighlight/2019/5/20/18542843/intersectionality-conservatism-law-race-genderdiscrimination

p. 651

References Anderson, Luvell, and Lepore, Ernie (2013). ʻSlurring Wordsʼ. Noûs, 47/1, 25–48. Google Scholar WorldCat Bright, Liam Kofi, Malinsky, Daniel, and Thompson, Morgan (2016). ʻCausally Interpreting Intersectionality Theoryʼ. Philosophy of Science, 83/1, 60–81. Google Scholar WorldCat Cappelen, Herman, and Dever, Joshua (2021). ʻOn the Uselessness of the Distinction between Ideal and Non-Ideal Theory (at Least in the Philosophy of Language)ʼ, in Justin Khoo and Rachel Sterken, eds., The Routledge Handbook of Social and Political Philosophy of Language. New York: Routledge, 91–105. Google Scholar Google Preview WorldCat COPAC Celikates, R. (2006). ʻFrom Critical Social Theory to a Social Theory of Critique: On the Critique of Ideology a er the Pragmatic Turnʼ. Constellations, 13/1, 21–40. Google Scholar WorldCat Crenshaw, Kimberlé (1989). ʻDemarginalizing the Intersection of Race and Sex: A Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and Antiracist Politicsʼ. University of Chicago Legal Forum, 140, 139–167. Google Scholar WorldCat Crenshaw, Kimberlé (1990). ʻMapping the Margins: Intersectionality, Identity Politics, and Violence against Women of Colorʼ. Stanford Law Review, 43, 1241. Google Scholar WorldCat Diaz‐Leon, E. (2016). ʻWoman as a Politically Significant Term: A Solution to the Puzzleʼ. Hypatia, 31/2, 245–258. Google Scholar WorldCat Engels, Friedrich (1893). Engels to Franz Mehring, London, July 14, 1893, Marxist Archive, https://www.marxists.org/archive/marx/works/1893/letters/93_07_14.htm Google Scholar WorldCat Geuss, Raymond (1981). The Idea of a Critical Theory: Habermas and the Frankfurt School. Cambridge, UK: Cambridge University Press. Google Scholar Google Preview WorldCat COPAC Grice, H. P. (1989). Studies in the Way of Words. Cambridge, MA: Harvard University Press. Google Scholar Google Preview WorldCat COPAC Haidt, Jonathan (2012). The Righteous Mind: Why Good People Are Divided by Politics and Religion. New York: Pantheon. Google Scholar Google Preview WorldCat COPAC Haslanger, Sally (2007). ʻ“But Mom, Crop-Tops Are Cute!” Social Knowledge, Social Structure and Ideology Critiqueʼ. Philosophical Issues, 17/1, 70–91.

Haslanger, Sally (2011). ʻIdeology, Generics, and Common Groundʼ, in C. Witt, ed., Feminist Metaphysics: Explorations in the Ontology of Sex, Gender and the Self. New York: Springer, 179–207. Google Scholar Google Preview WorldCat COPAC Haslanger, Sally (2012). Resisting Reality: Social Construction and Social Critique. New York: Oxford University Press. Google Scholar Google Preview WorldCat COPAC

Haslanger, Sally (2017). ʻRacism, Ideology, and Social Movementsʼ. Res Philosophica, 94/1, 1–22. Google Scholar WorldCat Haslanger, Sally (2021). Ideology in Practice: What Does Ideology Do? Milwaukee, WI: MacQuarrie University Press. Google Scholar Google Preview WorldCat COPAC Hom, Christopher (2008). ʻThe Semantics of Racial Epithetsʼ. Journal of Philosophy, 105/8, 416–440. Google Scholar WorldCat Jaeggi, Rahel (2009). ʻRethinking Ideologyʼ, in Boudewijn Bruin and Christopher Zurn eds, New Waves in Political Philosophy. London: Palgrave Macmillan, 63–86. Google Scholar Google Preview WorldCat COPAC p. 652

Jeshion, Robin (2021). ʻVarieties of Pejorativesʼ, in Justin Khoo and Rachel Sterken, eds., The Routledge Handbook of Social and Political Philosophy of Language. New York: Routledge, 211–231. Google Scholar Google Preview WorldCat COPAC Kahneman, Daniel (2011). Thinking, Fast and Slow. New York: Farrar, Straus and Giroux. Google Scholar Google Preview WorldCat COPAC Langton, Rae, and West, Caroline (1999). ʻScorekeeping in a Pornographic Language Gameʼ. Australasian Journal of Philosophy, 77/3, 303–319. Google Scholar WorldCat Leslie, Sarah-Jane (2008). ʻGenerics: Cognition and Acquisitionʼ. Philosophical Review, 117/1, 1–47. Google Scholar WorldCat Leslie, Sarah-Jane (2017). ʻThe Original Sin of Cognition: Fear Prejudice, and Generalizationʼ. Journal of Philosophy, 114/8, 393– 421. Google Scholar WorldCat Mills, Charles W. (2005). ʻ“Ideal Theory” as Ideologyʼ. Hypatia, 20/3, 165–184. Google Scholar WorldCat Murray, Sarah E. (2014). ʻVarieties of Updateʼ. Semantics and Pragmatics, 7/2, 1–53. Google Scholar WorldCat Potts, Christopher (2007). ʻThe Expressive Dimensionʼ. Theoretical Linguistics, 33/2, 165–198. Google Scholar WorldCat Queloz, Matthieu (2021). The Practical Origins of Ideas: Genealogy as Conceptual Reverse-Engineering (Open Access). Oxford: Oxford University Press. Google Scholar Google Preview WorldCat COPAC Sankaran, Kirun (2020). ʻWhatʼs New in the New Ideology Critique?ʼ Philosophical Studies, 177/5, 1441–1462. Google Scholar WorldCat Saul, Jennifer (2012). ʻPolitically Significant Terms and Philosophy of Languageʼ, in Sharon Crasnow and Anita Superson, eds., Out from the Shadows: Analytical Feminist Contributions to Traditional Philosophy. New York: Oxford University Press, 195–217. Google Scholar Google Preview WorldCat COPAC Shelby, Tommie (2003). ʻIdeology, Racism, and Critical Social Theoryʼ. Philosophical Forum, 34/2, 153–188. Google Scholar WorldCat Shelby, Tommie (2014). ʻRacism, Moralism, and Social Criticisimʼ. Du Bois Review: Social Science Research on Race, 11/1, 57–74.

Google Scholar

WorldCat

Srinivasan, Amia (2016). ʻPhilosophy and Ideologyʼ. Theoria: Revista de Teoría, Historia y Fundamentos de la Ciencia, 31/3, 371– 380. Google Scholar WorldCat Srinivasan, Amia (2019). ʻGenealogy, Epistemology and Worldmakingʼ. Proceedings of the Aristotelian Society, 119/2, 127–156. Google Scholar WorldCat Stalnaker, Robert (1978). ʻAssertionʼ, in John Kimball ed., Syntax and Semantics. New York: Academic Press, 9, 315–332. Google Scholar Google Preview WorldCat COPAC Stanley, Jason (1997). ʻRigidity and Contentʼ, in Richard G. Heck, ed., Language, Thought, and Logic. Oxford: Oxford University Press, 131–56. Google Scholar Google Preview WorldCat COPAC Stanley, Jason (2015). How Propaganda Works. Princeton: Princeton University Press. Google Scholar Google Preview WorldCat COPAC Stanley, Jason, and Szabó, Zoltán Gendler (2000). ʻOn Quantifier Domain Restrictionʼ. Mind and Language, 15/2–3, 219–261. Google Scholar WorldCat Street, Sharon (2006). ʻA Darwinian Dilemma for Realist Theories of Valueʼ. Philosophical Studies, 127/1, 109–166. Google Scholar WorldCat Swanson, Eric (2021). ʻLanguage and Ideologyʼ, in Justin Khoo and Rachel Sterken, eds., The Routledge Handbook of Social and Political Philosophy of Language. New York: Routledge, 331–344. Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

p. 655

29 Shakespeare’s Proper Names  Tina Chanter, Andrew Cutrofello https://doi.org/10.1093/oxfordhb/9780192844118.013.57 Published: 22 May 2024

Pages 655–669

Abstract The article re ects on the contingency of proper names, and their fraught character in a range of Shakespeare’s plays. Di erentiating between Shakespeare the “man” and Shakespeare the “author,” the article asks what it means to be the bearer of a proper name, and analyses how names function as part of one’s social anatomy. It discusses Juliet’s e orts to disentangle Romeo from his name, and the di

culty of undoing names, except in privileged, prescribed performances such as marriage

ceremonies. Certain names, however, can and do migrate from one person to another, as when Richard can no longer name himself “King” in a play in which the accusation of traitor functions by way of unnaming. Confusion ensues when the name Dromio is shared by identical twins. Female characters often adopt other names, while disguising themselves in male clothing, as does Imogen in Cymbeline, a play which complicates the referentiality of the title king. Like many of Shakespeare’s plays, the drama has historical roots, so that names cut across the ction–non ction divide. Caliban’s name is an approximate anagram of cannibal, thus serving to mark Caliban as a native of the new world. The racially fraught attribution—or lack thereof—of names and titles is evidenced when Iago prefers to refer to Othello not as “General,” but as “the Moor.” The contentious and fractious manner is which names can function in ways that honor and dishonor those named or not named is at the heart of Shakespeare’s language.

Keywords: authorship, disguise, inheritance, names, race, Shakespeare, titles Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

“Shakespeare was not always Shakespeare,” claims Emma Smith (2019, 303), opening the chapter on The Tempest in her book This Is Shakespeare. Referring to the author, rather than to the individual, Smith is making a distinction between the divergent ways in which the proper name of Shakespeare has functioned

through the years. We might say that she is commenting on what Foucault has called the “author 1

function.” Smith points to a discourse that has arisen over time identifying Shakespeare with the epitome of genius, as the esteemed author of some thirty-seven plays, as well as sonnets and narrative poems. These writings by Shakespeare, as Smith explains, “were not always bundled up with English identity, they were not always set as school texts, they did not always appear in editions bristling with learned footnotes and scholarly exegesis” (2019, 303). If Shakespeare’s authorial name has come to be identi ed as a venerable institution through the years so that it signi es in a way that it did not when he was still a young playwright, not yet celebrated, it is now bound up with the mythical imaginary of what it means either to be English or to be a speaker or reader of the English language. Yet at the time in which Shakespeare was writing, his plays would have been less identi able with the name of a single author, seen rather as products of a collaboration between Shakespeare, the actors who played his characters, the scriveners of the quarto editions of the plays, those who added stage directions, and those who set the plays into print in its rst, authoritative, folio edition. As Brinda Charry (2015, 26) puts it, “The Folio … gave Shakespeare the status of the ‘author’ of his works in the modern sense. (The collaborative nature of playwriting and the fact that so much literary output in the time period was for the stage, meant that literary works were not always associated with a single authorial gure).” The long and culturally varied history of performances—theatrical, lmic, operatic, poetic—as well as the rich and variegated international history of appropriations and the rewriting of Shakespeare’s plays, continue to raise important questions about the collaborative nature of the function of Shakespearean p. 656

authorship. So, too, does the

history of Shakespeare’s editorial interventions, which reveals a series of

emendations to his texts, often bringing his language into line with the cultural assumptions of the editors, suggesting that the meaning assigned to the texts of Shakespeare’s plays has continued to be a collective endeavour, if more obliquely. These emendations tell their own tale about the changing reception of Shakespeare’s plays, allowing us to see their texts as unstable and subject to material and ideological change, thereby re ecting the sense in which, as Foucault highlighted, the function of authors is enmeshed with the prevailing discourses of a given historical period. Whatever authority might still attach to Shakespeare’s authorial name, the texts we assign to him bear the traces of a complex set of contestations. These contestations include the history of debates over whether Shakespeare (i.e., the man from Stratford) was, in fact, the author of the plays. It is tting that the name “Shakespeare” should have such a complicated history, for the plays we now attribute to its presumptive bearer repeatedly highlight the problematic character of names. Throughout his corpus, Shakespeare dramatizes philosophical problems about the way names attach to their bearers, and the di

culty (despite their contingency) of detaching

them. Shakespeare’s most famous dramatization of this di

culty is the balcony scene in Romeo and Juliet. Gazing 2

out her window, Juliet asks, “O Romeo, Romeo, wherefore art thou Romeo?” (2.1.75). Her question is 3

ambiguous: Juliet could be asking why Romeo is the person he is or why his name is “Romeo.” However, since Romeo is the person called “Romeo,” these two questions cannot be easily disentangled. Therein lies the play’s potential for both comedy and tragedy. Juliet attempts to disentangle them by asking Romeo to “[d]eny [his] father and refuse [his] name” (2.1.76). She promises, if he will not, to deny her father and refuse her name, provided he will “be but sworn [her] love” (2.1.77). The Chorus has referred to them as “star-crossed lovers,” but perhaps they can escape the fate that the rivalry between their houses imposes upon them. Juliet observes that it is not Romeo himself but only his name that is her enemy, and that the patronymic “Montague” is not located in any part of his anatomy. “Oh, be some other name!” she implores, a possibility that she takes to be grounded in the contingency, and therefore changeability, of his name: “That which we call a rose / By any other word would smell as sweet; / So Romeo would, were he not Romeo called, / Retain that dear perfection which he owes / Without that title” (2.1.84–89). She concludes by bidding him to “do ” (2.1.89) his name the way he would an article of clothing. She does not realize that he is within earshot and so can respond as the bearer of the very name she bids him to discard.

This puts Romeo in an impossible position. When he responds, he identi es himself as the person she has addressed while simultaneously denying that he is the bearer of the name she has used to address him. 4

Instead of saying, “I am and am not Romeo,” a dialetheic solution, he responds to her invitation to be 5

sworn her love by inviting her to call him “Love”: “Call me but ‘love,’ and I’ll be new-baptized: / Henceforth I never will be Romeo” (2.1.92–93). The order of these two statements suggests that he can lose 6

the name “Romeo” only by being renamed by her. When Juliet asks what man he is he replies that his name p. 657

is hateful to him and that if it were externalized in

writing he would tear it. Now recognizing him by his

voice, conventionally regarded as a more intimate form of expression than writing, she asks, “Art thou not Romeo, and a Montague?” (2.1.102). It would be tempting to hear “Art thou not Romeo?” as “Art thou ‘notRomeo’?”—i.e., as if “not-Romeo” were a negative name—but Romeo himself takes the negation in her query to be sentential, equivalent to “Is it not the case that you are Romeo, and a Montague?” This construal leaves the status of both names in suspense, allowing him to reiterate his o er to refuse them: “Neither, fair maid, if either thee dislike” (2.1.103). Instead of acknowledging that a kind of negative baptism has just successfully taken place, Romeo still awaits a new, positive baptism (“Call me but ‘love’ ”). Bypassing names altogether, Juliet now uses the familiar pronoun “thou”: “How camest thou hither, and wherefore?” This question echoes her initial question, “Romeo, Romeo! Wherefore art thou Romeo?,” but it changes the topic, for she now asks not “Why are you Romeo?” but “Why are you here?” For the rest of the play Romeo remains “Romeo, and a Montague,” a social fact that precipitates his fatal duel with Tybalt. He belatedly attempts to lose his name a second time when he asks Friar Laurence (who has the socially sanctioned power to baptize) to tell him in what vile part of his anatomy his name is lodged that he might “sack / The hateful mansion” (3.3.107–108). But Romeo’s name is not an excisable part of his anatomy, at least not of his physiological anatomy. It is a part of his social anatomy. Being the bearer of a name is a social fact, one that typically cannot be undone without the performance of a public ceremony such as 7

marriage. According to the patriarchal custom of the world of the play, Juliet would have become a Montague when she married Romeo (as Lady Capulet and Lady Montague took their respective husbands’ names). But since their marriage takes place in private (as well as o stage), none of its symbolic consequences can take hold until it is publicly proclaimed. Their exchange of engagement vows in the balcony scene is meta-dramatically witnessed by us, and their exchange of wedding vows in the marriage ceremony is metaphysically witnessed by the friar. But the latter—along with whatever acts of unnaming and renaming they involve—have to be acknowledged by the Duke of Verona and their families to take on the force of law. Tragically, this only happens after Romeo and Juliet have killed themselves. They are memorialized by the duke, in the play’s nal line, as “Juliet and her Romeo” (5.3.310), a metrical double of “Romeo and a Montague” that detaches his personal name from his patronymic while inverting the possessive relation between husband and wife. In Shakespeare’s day, men typically did not change their names when they married, but they did when they gained or lost a title, as happens throughout Richard II. Ernst Kantorowicz (1985, 26) calls this play “the tragedy of the King’s Two Bodies,” but it could just as easily be called “the tragedy of the King’s Two Names.” English law stipulated that a king had two bodies: a corporeal one subject to death and corruption, and an incorporeal one that never died. Each body had its own name, the incorporeal one bearing the name “King”: “King is a Name of Continuance, which shall always endure as the Head and Governor of the People (as the Law presumes) as long as the People continue … and in the Name the King never dies” (Justice p. 658

Brown, in Plowden, Reports,

cited in Kantorowicz 1985, 23; his ellipsis). Richard II is about the process by

which the ctional referent of the name “King”—the glorious incorporeal body that never dies—is imagined to pass from the corporeal body of Richard of Bordeaux to that of Henry Bolingbroke. Since the glorious body is referentially indistinguishable from the physical body associated with it, this process amounts to a re-assignation of the name “King.” In this sense, the play is about the unnaming of one man (who loses a title) and the renaming of another (who gains it).

Kantorowicz observes that Richard’s attitude toward kingship changes as he learns he has no power to stop the rebellion. At rst, he professes to believe that glorious angels will ght on his behalf. Later, he bids his name (i.e., “King”) to arm, as if it had the power he lacks. Finally, he acknowledges that the name “King” is just a name: A curious change in Richard’s attitude—as it were, a metamorphosis from “Realism” to “Nominalism”—now takes place. The Universal called “Kingship” begins to disintegrate; its transcendental “Reality,” its objective truth and god-like existence, so brilliant shortly before, pales into a nothing, a nomen. (Kantorowicz 1985, 29) This passage may overstate the reductive character of Richard’s nominalism. When he bids the name “King” to arm he represents it not as a mere name but as a glorious name: “Is not the king’s name forty thousand names? / Arm, arm, my name! A puny subject strikes / At thy great glory” (3.2.80–82). He is not necessarily mistaken in thinking that the name “King” is forty thousand names. His mistake lies in thinking that the name “King” is his name. He is mistaken in two ways: rst, in not realizing that it has already migrated, de facto if not yet de jure, from him to Henry Bolingbroke; second, in not realizing that the name “King” never truly referred to him, a mortal man whose proper name is “Richard.” He chides his followers not for having led him to believe that there is a universal called “Kingship,” but for having called him “King”: Throw away respect, Tradition, form, and ceremonious duty, For you have but mistook me all this while. I live with bread like you, feel want, Taste grief, need friends. Subjected thus, How can you say to me I am a king? (3.2.167–172) As the inde nite article indicates, Richard is now using the word “king” as a common noun rather than as a proper noun. He is not so much rejecting realism in favor of nominalism as treating kingship as a social kind. The point of the inde nite article is that “King” is not just a name; it is a title. Richard has come to realize how absurd it is to apply that title to a needy human being. Only a perfect divine being could deserve it (the biblical expression “King of kings” is used twice in Richard III, at 1.4.176 and 2.1.13). It is this change in attitude that explains his subsequent willingness to give up the crown (the material symbol of the king’s invisible body) even as he castigates Bolingbroke and his followers for presuming, respectively, to claim and bestow it. p. 659

Ironically, Richard himself has triggered Henry’s ambition by refusing to allow the Duke of Hereford to inherit the title Duke of Lancaster after the death of his father John of Gaunt. Henry returns from banishment ostensibly only to claim this title, but by suspending the right of succession, Richard has made his own title precarious. The importance of titles in the kingdom is underscored in the rst line of the play, when King Richard (as he is designated in the speech pre x) doubly addresses “Old John of Gaunt, timehonored Lancaster” (1.1.1) and goes on to explain why his “Cousin of Hereford” and “the Duke of Norfolk, Thomas Mowbray” stand before him: “Namely, to appeal each other of high treason” (1.1.27–29). The adverb “Namely” plays its usual role of speci cation here, but by subtly modifying the verb “appeal” (from Old French apeler, meaning “to call” or “to name”), it underscores the nature of the two men’s accusations. Each seeks, in e ect, to disentitle the other by calling him “traitor” (1.1.39, 144). When Bolingbroke does so he unwittingly directs his accusation against Mowbray back at himself through the use of a dangling modi er that foreshadows his later rebellion: “With a foul traitor’s name stu

I thy throat” (1.1.44).

Richard tries to prevent the two men from engaging in a trial by combat, but he is unable to do so because backing down would dishonor their names (1.1.167–168, 190–195). When the trial is about to begin, each of the two dukes is ceremoniously asked to state his name and cause. Mowbray is asked to do so “In God’s name and the King’s” (1.3.11), a formulation that indicates the symbolic foundation of the entire social system of names. It is this system that Richard undermines when he deprives Henry of the name “Lancaster.” When Richard departs for Ireland he leaves behind a power vacuum whose form is a general state of namelessness. This is sensed by Queen Isabel, who dreads an object without a name: “what it is that 8

is not yet known what, / I cannot name. ’Tis nameless woe, I wot” (2.2.39–40). She soon discovers what— or rather who—this nameless object is: Henry Hereford, who, coming back to England to seek the name of 9

Lancaster, will subsequently claim the name of King. When the assembled lords hail him “Henry, of that name the fourth!” he assumes both name and throne “[i]n God’s name” (4.1.106–7), wording that immediately provokes an ine ective protest from the Bishop of Carlisle. With Henry’s restoration of nominal order other titles have to be adjusted accordingly (such as the demotion of the Duke of Aumerle to the Earl of Rutland). Meanwhile, Richard experiences his loss of the name “King” as the loss of his personal name as well: “I have no name, no title – / No, not that name was given me at the font – / But ’tis usurped. Alack the heavy day, / That I have worn so many winters out / And know not now what name to call myself” (4.1.248–252). He eventually asserts the vanity of names, the fact that nominally unaccommodated man is no more but a poor, bare, forked animal: “But whate’er I am, / Nor I nor any man that but man is / With 10

nothing shall be pleased, till he be eased / With being nothing” (5.5.38–41).

Just as Richard III is a play about a king’s two names, so The Comedy of Errors is a play about a name’s two servile bodies. King Richard loses both his names, becoming nothing but a man, while the name “Dromio” changes from being a proper name to being a common name. The same thing happens to the name p. 660

“Antipholus,” but it is the commonness of the shared name of the propertyless servants that subverts the social—and nominal—hierarchy. The reunion of the Dromios and Antipholi at the end of the play tilts its homosocial axis away from the vertical master/servant relation to the horizontal relation of fraternity. It also shifts its heterosexual axis away from the unhappily hierarchical marriage of Adriana and Antipholus of Ephesus to the potentially more harmonious pairing of their siblings. It is not uncommon for a single name to have two referents, but it is uncommon, at least in our culture, for two identical twins to be given the same name. Sons are often given their fathers’ names, allowing for nominal continuity over time. When Henry IV dies in 2 Henry IV his son, now Henry V, purports to have inherited not only the title “King” but also the nickname “Harry”: “Yet weep that Harry’s dead, and so will I; / But Harry lives that shall convert those tears / By number into hours of happiness” (5.2.59–60). In The Comedy of Errors, nominal succession over time is horizontalized into nominal equality across space. We are to imagine each pair of twins not only as nominally identical but as physiognomically identical as well. The result is that everyone confuses them, including they themselves (chiastically) until each twin encounters his own double. When Antipholus of Syracuse asks Dromio of Ephesus what he has done with the money that he (Antipholus) has just given to Dromio of Syracuse, he addresses him as “Dromio”: “Come, Dromio, come, these jests are out of season” (1.2.68). (As a dutiful servant, Dromio does not address Antipholus as 11

“Antipholus”; he calls him “sir.” ) Antipholus thinks that the person he is addressing is named “Dromio,” and he is right about this. But his reason for thinking that “Dromio” is Dromio’s name fails to track the actual reason it is his name. This makes his apparent knowledge a classic instance of a Gettier problem— that is, a situation in which someone has a good reason for believing something to be true that is in fact true, but seems to lack knowledge because the good reason is not the actual reason the thing is true. The causal history linking the name “Dromio” to Dromio of Ephesus is not the causal history Antipholus purports to be tracking. If he knew that the person he is addressing had been living in Ephesus for years he would surely exclaim, “Dromio, Dromio, wherefore art thou ‘Dromio’?” The same epistemic problem would

arise if we were to take Antipholus’s use of the name “Dromio” to refer to Dromio of Ephesus not by virtue of the causal consequences of an original act of baptism but by virtue of his using it as an abbreviation of the descriptive phrase “the servant of Antipholus, son of Egeus,” for this phrase is not a de nite description. Other strategies—such as taking “Dromio” to refer to “my servant” when uttered by Antipholus of Syracuse—would fail for similar reasons. Each of the Dromios undergoes an identity crisis. When Dromio of Syracuse is guarding the door to Antipholus of Ephesus’s house, he identi es himself by voice to Dromio of Ephesus as Dromio. In response, Dromio of Ephesus charges him with stealing both his o

ce and his name, as if Dromio of Ephesus were a

king and Dromio of Syracuse, a usurper. Conversely, after Dromio of Syracuse is accosted by Nell (who mistakes him for Dromio of Ephesus), he asks Antipholus of Syracuse to reassure him of his identity: “Do 12

you know me, sir? Am I Dromio? Am I your man? Am I myself?” (3.2.73–74). p. 661

meet face to face, each receives his identity

When the two Dromios nally

back from the other: “Methinks you are my glass and not my

brother; / I see, by you, I am a sweet-faced youth” (5.1.419–420). Declining Dromio of Syracuse’s suggestion that they “draw cuts” (5.1.424) to decide who is the elder and therefore has precedence by right of primogeniture—the right to inherit a title—Dromio of Ephesus proposes to renounce primogeniture altogether: “We came into the world like brother and brother, / And now let’s go hand in hand, not one before another” (5.1.426–427). This plural “We,” which could be uttered by either of them, is not the royal “We” of a private individual with a personal name and a title, but the solidary “We” of two titleless subjects 13

with a personal name in common: “Dromios of the world, unite!”

Unlike the Dromios, the rebel Jack Cade in 2 Henry VI is happy to claim the prerogative of a title. To equal his opponent Sir Humphrey Sta ord, he knights himself: “Rise up, Sir John Mortimer” (4.2.109; with a pun on “Rise up”). Caius Martius, a member of the Roman nobility who disdains the people, is given the title “Coriolanus” by the Roman general Cominius (1.9.61–64). Later, he is stripped of this title by the Volscian 14

general Au dius (5.6.87–89).

In All’s Well That Ends Well, the King of France assumes that Bertram, the

Count of Roussillon, disdains Helena only because she lacks a title that the king has the power to bestow (2.3.115–116). While promising to raise Helena’s status to equal his, he warns Bertram not to mistake a virtuous name for virtue itself: If she be All that is virtuous, save what thou dislik’st— “A poor physician’s daughter”—thou dislik’st Of virtue for the name. But do not so: From lowest place when virtuous things proceed The place is digni ed by th’ doer’s deed. Where great additions swell ’s and virtue none, It is a dropsied honor. Good alone Is good without a name. (2.3.119–127) When Helena chooses to ee rather than accept her new status as the count’s wife she abandons her given name for the humble one of “Saint Jacques’ pilgrim” (3.4.4). Many of Shakespeare’s female characters take on new names, often while disguising themselves in male dress: Julia in Two Gentlemen of Verona, Portia and Nerissa in The Merchant of Venice, Rosalind and Celia in As You Like It, Viola in Twelfth Night, and Imogen in 15

Cymbeline.

As Marjorie Garber (2004, 521) points out, the cross-dressed Viola/Cesario (“Cesario” meaning

“little Caesar”) represents “another take on the Elizabethan dilemma, the ruler with two bodies, one ‘natural’ and female, the other ‘political’ and ungendered—which is to say, gendered as a universal male.” Garber notes that after the death of Queen Elizabeth “the cross-dressed woman disappears from Shakespeare’s plays, until the gure of Imogen in the late romance Cymbeline” (523).

Imogen identi es herself as “Fidele” when she is asked for her name by the banished Lord Belarius (and, p. 662

later, by the Roman general Caius Lucius). Belarius has kidnapped

and raised her two brothers,

bestowing new names on them as well as on himself. Throughout the complicated plot that Cymbeline partly dramatizes and partly narrates characters lose their names when they leave the British court and regain them when they return to it. In the nal scene a dozen characters reveal their true identities to the king, some through the revelation of names—as when “Morgan” identi es himself as Belarius, and explains that “Polydore” and “Cadwal” are, respectively, Guiderius and Arviragus—others through acts of confession. What matters is not so much their given or assumed names as “the place they have earned in the story,” to borrow a (slightly reworded) phrase of Enobarbus in Antony and Cleopatra. Cymbeline is a play in which characters can do

their names the way they do articles of clothing, and in which both prove to be unreliable

markers of personal identity. Cloten expects Guiderius/Polydore to tremble at the sound of his name, but when Guilderius/Polydore hears it he contemptuously declares, “Were it Toad, or Adder, Spider / ’Twould move me sooner” (4.2.90–91). Imogen misidenti es Cloten’s decapitated corpse because it is dressed in the clothes of her husband Posthumus Leonatus, and Posthumous misidenti es her when she is in the garb of a male page. The name “Imogen” is interesting for a textual reason as well. Despite the fact that it appears forty times in the Folio (Shakespeare 2017, 391), modern editors frequently emend it to “Innogen,” the name ascribed to the wife of Brut, the legendary descendant of Aeneas who was said to have founded Britain. To decide whether Imogen (the character in the play) has two names we would need to consider the way reference works not only in ction but in the textual apparatuses that represent ctions. A similar question could be raised about the divided reference of the name “Cymbeline.” Shakespeare derived his narrative in part from Geo rey of Monmouth’s Historia Regum Britanniae, which recounts the history of a British king whose Latin name was “Cunobelinus.” By basing his romance in history—or elevating history to the level of romance— Shakespeare creates a world in which the referential function of the name of his titular king cuts across the ctional–non ctional divide. Other questions about names bear on the relation between the worlds of Shakespeare’s plays and the world in which those plays exist. Editorial emendations to which a string of editors subjected the texts of Shakespeare’s plays in their time—and the history of the withdrawal of these editorial choices—re ect the sociocultural preoccupations and prejudices of each age. This shifting history tells us about the uses to which these texts were put rather than those which Shakespeare the author might have meant or intended. The question of language and naming is very much at issue in The Tempest, whose Prospero some readers 16

(beginning with Coleridge ) have construed as a semi-autobiographical gure, his magical powers symbolizing those of Shakespeare’s art. In his propensity for magic, the character of Prospero resembles— or appropriates—the witchery of Sycorax, mother of Caliban. The appellation that the presumptively white Miranda applies to the black-skinned Caliban proved unpalatable to the seventeenth-century sensibilities of John Dryden and William Devenant, who withdrew from her the lines Miranda addresses to Caliban, reassigning them to Prospero. Addressing Caliban, Miranda substitutes his name for a pejorative epithet p. 663

that subordinates him with a

vehemence that Dryden and Devenant found hard to reconcile with what

they took to be appropriate for Miranda’s character: Abhorrèd slave, Which any print of goodness wilt not take, Being capable of all ill. (1.2.350–352) Charry (2015, 28) explains this misattribution (which some later editions have corrected, although in others it remains uncorrected), by pointing out that Dryden and Devenant “could not stomach the fact that the gentle Miranda could utter such harsh words, even if they were addressed to Caliban. These lines were therefore assigned to Prospero till the twentieth century when editors and readers were willing and able to

view Miranda in more complex terms.” Caliban’s name, which Miranda declines to utter when she addresses him as “slave,” is an oral anagram of “cannibal.” Pointing to the alliterative play of the name that Caliban himself performs when, after encountering Stephano and Trinculo, he reverses the syllables of his own name, “’Ban, ’Ban, Ca-Caliban / Has a new master: get a new man” (2.2.174–175), Jonathan Goldberg (2004, 41) sees it “as an invitation to any audience who may have missed it to recognize the anagrammatic play involved in the name of the character.” When Miranda calls Caliban by the generic term “slave,” she at once asserts her ownership over him and consigns him to membership of a group of people undeserving of their own name (or rather, their given names—for the birth names of slaves were often substituted by names given them by slave owners). Miranda thus renders Caliban indistinguishable from those other subordinates with whom her words class him. She quali es her ownership of him with the word “abhorrèd,” alluding to a history of having repulsed him, of being repulsed by him, a history of revulsion, we might say, a history of revolt. Her words are heavy with the a ect of disgust. Her refusal to use his particular name, coupled with her disgust, is suggestive of what Rebecca Bach (2007, 157) identi es in Richard Hakluyt’s Principal Navigations (1566), a text that would have been available to Shakespeare, as an “essentially undi erentiated black African population,” depicted as “sexually excessive.” The “repeated nomination” of Caliban as monstrous—Goldberg (2004, 41) counts thirty-four occurrences of the term “monster” in The Tempest, the vast majority of which refer to Caliban— underscores his excessive nature, an excess that prompts Trinculo to cast doubt on his humanity. Is he “A man or a sh? Dead or alive?” (2.2.24). Caliban’s native language is dismissed by Miranda as that of a “savage” belonging to a “vile race,” as “gabble,” the “purposes” of his words claimed or imposed by Miranda (1.2.354–357). Required to substitute his native tongue for that of Prospero and Miranda, Caliban turns their language against them when he uses it to curse them. “You taught me language, and my pro t on’t / Is I know how to curse” (1.2.362–333), transforming the language of his oppressors into a site of resistance. Colonization is not just about appropriation of lands—of an island—but also about language and the imposition of ideology. As Charry p. 664

(2015, 43) says, “Prospero is the prototype of the colonizer who not only takes

possession of the land, but

also has the power to insist on his construction of events as the truth, while Caliban is the native who dares to talk back.” The purpose that Miranda divines in Caliban’s words is that from which Sylvia Wynter (2006, 109–127) distances herself in her essay “Beyond Miranda’s Meaning” in which she suggests that we look for the implicit meaning of The Tempest in a name that is absent from the play, beyond its imagination. This absence or lacuna around which Wynter reads The Tempest as circulating is its failure to include the name of any “physiognomically complementary” mate for Caliban. By interrogating the absence of any name for a black “mate,” Wynter is implicitly shifting the focus away from that of the explicit narrative of The Tempest, which invokes the attempted rape of Prospero’s daughter Miranda by Caliban, and toward Caliban’s symbolic extraction from a native context in which nding a physiognomic complementary mate would not have been a problem. Wynter’s implication is that the narrative scenario around which The Tempest navigates is a gure that is unnameable—unimaginable, unpalatable. That gure is the black sexualized woman—Caliban’s would-be mate—whose guration is banished from the white, colonial imagination. Black women, like black men, are described by Hakluyt as “libidinous,” but unlike black men’s, black women’s sexuality is not gurable by The Tempest, except in the shape of the “Hag” that seeded Caliban, Sycorax, who is consigned to death—returned to the natural world, to the world of things that speak no 17

language.

The historical background against which The Tempest was written, and to which many critics have taken it to allude, is one in which the expropriation of the “New World” was underway. In 1609, Sea Venture, one of nine ships to have set sail for Jamestown, Virginia, the rst English colony, had never arrived at its destination, sunk in a storm near the Bermudas. The documents of two of its survivors, both circa 1610, have

18

been taken as sources for The Tempest.

As Charry (2015, 32–33) says, “Caliban has been read as a ctional

representation of the New World native as perceived by Shakespeare.” If we accept that in the gure of Caliban is condensed all those nameless individuals, all those “abhorrèd slaves,” plucked from their native habitats, and deposited on the island of the “New World,” in which their names, history, and languages were obliterated, in which they no longer counted as individuals with proper names, then the absence of Caliban’s physiognomic mate would follow. In Shakespeare’s time and in his texts, Bach (2007, 23) suggests, “Some men had personal names that signi ed and some did not.” By pointing out that not all proper names functioned in the same way, Bach is taking account of the imaginary that di erentiates between those male characters deserving of proper names, and those like Caliban, whose nomination by Miranda as abhorrèd slave, subsumes his identity, eclipsing it by relegating him to a category rendered incapable of bearing traits that distinguish the particularity of individuals. For Bach, Shakespeare’s clowns and racially othered men constitute a class of “base men,” marked as sexually incontinent, whose ostensible failure to control their sexual desires aligned them with women, and opposed them to men worthy of the distinction of proper names. Bach distinguishes the “homosocial imaginary” that informs Shakespeare’s plays from the “heterosexual imaginary” that p. 665

arose

during the “long eighteenth century,” identifying as modern the tendency to construe all males as

inherently imbued with uncontrollable sexual urges, while (white) females were expected to be comparatively chaste and modest in their sexual desires. Challenging the imaginary fuelling the symbolic equation of Caliban’s name with cannibalism, Richard Halpern (1994, 262–292) overturns this symbolism, uncovering in its stead a dynamic of “white cannibalism” in The Tempest. Halpern accomplishes this interpretation through establishing the sense in which the speech of Caliban and that of Gonzalo mirror one another. In making this argument, Halpern draws attention to the way in which the name “Claribel” functions in the play, a name to which Gonzalo refers, a name that does not gure in the cast of characters. The only female name to designate a human character, rather than a spirit, within the cast of characters is that of Miranda. Halpern points to a site of slippage between the name “Claribel” and that of Dido, another name that might appear on the surface to be invoked merely in passing, but which, in its symbolic relationship to Claribel’s name, e ects a series of elisions for Halpern. Gonzalo introduces the name of Claribel just before the speech to which critics have referred to as his “Commonwealth” speech—the speech in which he evokes a utopic vision. As Halpern reads it, the confusion in Gonzalo’s speech between the names of Claribel and Dido—a confusion revealed in Gonzalo’s con ation of two place names, that of Tunis and Carthage—establishes and prepares for the “white cannibalism” that the Commonwealth speech performs. In identifying Claribel with Dido, and Tunis with The Aeneid’s ctionalised version of the ancient city of Carthage, Gonzalo at one and the same time calls up and erases the miscegenating relations between Claribel and her husband, the King of Tunis. Gonzalo refers to Alonso’s daughter, “the King’s fair daughter Claribel” who is married to the “King of Tunis” (2.1.66–67). Since Claribel is a “white European” and the King of Tunis is a “black African,” the marriage is “mixed or miscegenating” (Halpern 1994, 274). Gonzalo goes on to confuse Tunis with Carthage when he indicates that there has not been as “sweet” a marriage as that between Claribel and the King of Tunis since “widow Dido’s time” (2.1.72). Mocking what Halpern (1994, 275) identi es as Gonzalo’s “active forgetfulness,” Adrian and Antonio point out that by calling up Aeneas’s relationship to Dido, “the non-African queen of African Carthage,” Gonzalo has confused the ancient city Carthage, of which Virgil gives a “ ctionalized vision” in The Aeneid, with Tunis. In so doing, Halpern says that “Gonzalo both evokes and denies the miscegenous marriage of Claribel.” In confusing Tunis with Virgil’s ctional representation of Carthage, Gonzalo supplants “the material existence of a non-European society with a founding text of the Western tradition and, not incidentally, the great epic of Roman imperialism.” (Halpern 1994, p. 274). For, “Gonzalo’s forgetfulness performs an act of white cannibalism,” in e ect deleting Tunis and reinscribing it “within a humanist textual tradition” (274).

This deletion and re-inscription, this act of cultural cannibalism, is what anticipates the cultural consumption that Gonzalo goes on to reiterate as he describes his utopian vision. Gonzalo’s vision relies upon American Indians, even as it obscures any reference to the very racially othered bodies that it evokes p. 666

without naming, as if the utopia

it articulated came from nowhere—as if its provenance were wholly

European, wholly white. Gonzalo’s vision is of a semi-Edenic “New World” of purity and innocence, called up by his utopian evocation of “the ideal commonwealth he would create if he had ‘plantation of this isle’ ” (see 2.1.138). The word “plantation” marks a departure from Michel de Montaigne’s (1958) essay, “Of the Cannibals,” which Gonzalo’s speech otherwise follows closely; in doing so, Halpern (1994, 268) says, it 19

“unambiguously signi es a European colony.”

For Halpern, “Gonzalo’s ideal commonwealth doubly

e aces its references to the ‘New World.’ On the one hand it reinscribes them within a humanist genealogy. Yet insofar as it invokes the generic codes of utopia, it denies all lines of origin by posing as an autonomous act of philosophizing speculation” (275). Although Claribel is not included in Shakespeare’s cast of characters, falling outside the diegetic frame of The Tempest, her name carries considerable impetus. Just as the name of Helen propels the action of the Iliad from outside the frame of the action (though unlike Claribel, who is only invoked by Shakespeare in name, Helen does appear as a character within Homer’s narrative), so Claribel’s marriage to the King of Tunis, as Claire Conilleau (2010, 8) says, “is the original impetus for the other characters’ sea voyages across the Mediterranean and her father’s eventual shipwreck.” If Claribel’s story remains untold by Shakespeare, H.D. “pens a prequel to the Bard’s story,” writing the voiceless and invisible Claribel into writerly embodied existence, as she steps out “from the pages of Shakespeare’s play” to “travel to Stratford-upon-Avon to pay tribute to the poet, her creator—as does H.D. in 1945” (Conilleau 2010, 9). If H.D. recuperates Claribel’s story from the silence and obfuscation in which Shakespeare shrouded her, Julia Reinhard Lupton (2019, esp. 11) performs a di erent kind of retrieval for the name of Caliban. In her critical analysis of the arguments embedded in Caliban’s speech, Lupton relieves Caliban’s name of some of the reiterated monstrosity that accumulates in its cannibalistic overtones, inscribing it rather as his “creatureliness” by reading it within the tradition of black natural law. Taking up “the creaturely condition” (Lupton 2019, 8) of natural law theory, which was “alive in … Shakespeare’s age” (1), Lupton suggests that Caliban “intensi es” natural law theory “into a point of protest: he blackens it” (8). Lupton in ects Caliban’s words through the lenses of writers such as Cicero, on the one hand, and Frederick Douglass, on the other hand, reading Caliban as cultivating the “art of listening” that she sees as symptomatic of Caliban’s attunement to the natural world and to his place in the “trembling continuum of living beings.” (11). Distancing Caliban from what Halpern (1994, 282) appropriately identi es as the “most di

cult moment”

of The Tempest, when Caliban apparently shows no remorse for his attempted violation of Miranda, but merely retorts to Prospero: would that he’d “peopled” the isle “with Calibans” (1.2.349–350), multiplying his own name, Lupton echoes Halpern’s recontextualizing of Caliban’s callous retort. While Halpern understands Caliban’s words to be the reverse mirror image of the utopian strain of Gonzalo’s speech, which whitens Caliban’s cannibalism, Lupton sees Caliban as blackening natural law theory, identifying “a new and di erent order of self-rule” in Caliban’s self-attunement. Commenting on Caliban’s counsel “Be p. 667

not afeard: the isle is full of noises, / Sounds and sweet airs that give delight and hurt not” (3.2.128–129), Lupton (2019, 11) says “we glimpse spiritual techniques that give the ensouled creature inner strength and a new equilibrium along with heightened cognitive and a ective capacities.” Noting that “Caliban twice calls Prospero a tyrant,” Lupton (2019, 10) says, “ ‘Tyrant’ is a ghting word. It is also a natural law word. In the words of Cicero, ‘there can be no fellowship between us and tyrants.’ ” Lupton (2019, 10) goes on to recall that one of the dogs that Prospero unleashes against the rebels is named Tyrant: “Fury, Fury! There, Tyrant, there! Hark, Hark!” (4.1.249). The dogs embody Prospero’s own tyrannical fury and the

resurgence of his disavowed creatureliness … The image anticipates the canine tracking of runaway slaves. In Othello, another play in which names and titles are racially fraught, Roderigo disdains to call Othello either by his proper name (with one exception at 4.2.215) or by his title. Instead, he refers to him as “the thicklips” (1.1.64), “a lascivious Moor” (1.1.122), and “an extravagant and wheeling stranger” (1.1.132)— 20

while proudly identifying himself by name to Brabantio: “My name is Roderigo” (1.1.92). 21

implicitly echo Hakluyt’s description of Moors as “base” and “libidinous.”

These epithets

Iago uses both Othello’s name

and his title, but he far more frequently refers to him as “the Moor” as if this, rather than “General,” were his title. Just as Iago manages to strip Cassio of the title of lieutenant, so he verbally transfers Othello’s title to Desdemona (whose name signi es “the unfortunate”)—“Our general’s wife is now the general” (2.3.288–289)—insinuating that Othello’s “sexual interest” in his wife “debases and feminizes him” (Bach 2007, 159). As Bach elaborates, the disparaging overdetermination of Othello’s sexual passion is re ective of the pervasive di erentiation between dominant men and racially othered men to which she points. However, while Othello’s jealousy may be seen as pathological, Iago’s and Roderigo’s racial stereotyping is contested by other characters in the play, as it is by Othello himself when he asks the representatives of the world at large to “Speak of me as I am; nothing extenuate, / Nor set down aught in malice” (5.2.335–336).

Notes

p. 668

1.

“If I discover that Shakespeare was not born in the house we visit today, this is a modification that, obviously, will not alter the functioning of the authorʼs name. But if we proved that Shakespeare did not write those sonnets which pass for his, that would constitute a significant change and a ect the manner in which the authorʼs name functions. If we proved that Shakespeare wrote Baconʼs Organon by showing that the same author wrote both the works of Bacon and those of Shakespeare, that would be a third type of change that would entirely modify the functioning of the authorʼs name. The authorʼs name is not, therefore, just a proper name like the rest” (Foucault 1998, 210).

2.

All citations from Shakespeare are from the third edition of The Norton Shakespeare, (Shakespeare 2016), and given in standard Modern Library Association format. We omit the title of the play when the reference is clear.

3.

Jacques Derrida (1992, 426) downplays this ambiguity: Juliet “does not say to him: why are you called Romeo, why do you bear this name (like an article of clothing, an ornament, a detachable sign)? She says to him: why are you Romeo?”

4.

Cf. Troilusʼs perplexed statement: “This is and is not Cressid” (Troilus and Cressida, 5.2.146).

5.

Since Romeo uses this term as an imagined proper name, the capital L of the First Folio seems preferable to the lower case l of the First and Second Quartos. Earlier in the scene Mercutio addresses Romeo with a mocking series of alternative names: “Romeo! Humors! Madman! Passion! Lover!” (2.1.7).

6.

In King Lear, Edgar renames himself “Poor Tom” before declaring himself unnamed: “Edgar I nothing am” (2.3.21).

7.

Calling someone by a name that is not theirs is not always equivalent to renaming them. In King John, the Bastard (who has just had his title upgraded to Sir Richard Plantagenet) characterizes the former as a prerogative of power: “An if his name be George Iʼll call him Peter” (1.1.186).

8.

Cf. the Witchesʼ response to Macbethʼs question about what they are doing: “A deed without a name” (4.1.48).

9.

Cf. the Witchesʼ triple hailing of Macbeth as “Thane of Glamis,” “Thane of Cawdor,” and “King herea er” (1.3.49–51).

10.

Slavoj Žižek (2016, 211) argues that both Julietʼs and Richardʼs insistence on the contingency of their names is characteristic of hysterical subjectivity as theorized in Lacanian psychoanalysis. Such a subject is forever asking, “Why am I what youʼre saying that I am?” or “Why am I that name?” For Žižek (212–216), Richardʼs awareness of his lack of a name makes him a “desubjectivized subject,” but instead of identifying with this predicament as an analyst would, he first descends into a brief state of psychosis (maddened by the sound of music) before coming to terms with his new place in the symbolic order (that of a condemned prisoner).

11.

An excellent overview of Shakespeareʼs representation of appropriate and inappropriate forms of address can be found in Blake (2002, 271–283).

12.

Cf. the similarly self-questioning response of Christopher Sly when he is called “Lord” in the Induction to The Taming of the Shrew: “Am I lord? And have I such a lady? / Or do I dream? Or have I dreamed till now?” (Induction 2.66–67).

13.

In Julius Caesar, when Cinna the poet tries to save his life by distinguishing himself from Cinna the conspirator, the Fourth Plebeian responds, “It is no matter, his nameʼs Cinna. Pluck but his name out of his heart, and turn him going” (3.3.32–33).

14.

Shortly a er receiving his new title, Coriolanus forgets the name of the poor man who used him kindly when he lay in Corioles (1.9.80–89).

15.

In The Merry Wives of Windsor, Falsta briefly assumes the identity of the witch of Brentford.

16.

Coleridge refers to Prospero as “the very Shakespeare himself, as it were, of the tempest” (2016, 136).

17.

See Bach (2007, p. 157) on Hakluyt 157; and on Miranda and Prospero in The Tempest 1.2.

18.

Jourdain (1610) and Strachey (1625), both cited in Charry (2015).

19.

Halpern (1994, 271) specifies that although Gonzaloʼs ideal commonwealth “borrows its descriptive detail from Montaigne” it “also alludes in a more general way to Thomas Moreʼs Utopia,” but it deletes any reference to the “New World.” Rather than picturing an “Indian utopia,” Shakespeareʼs Gonzalo whitens the utopian vision, presenting it as an “Ovidian and Virgilian … Golden age” (Halpern 1994, 272).

20.

Declaring oneʼs name is a way of identifying oneself—e.g., “I am a knight, sir, and my name is Coleville of the Dale” (2 Henry IV, 4.2.3); “My name is Marina” (Pericles, 5.1.133)—but it can also function as an act of self-assertion: “My name is Harry Percy” (1 Henry IV, 5.4.59); “My name is Pistol called” (Henry V, 4.1.62); “This is I, / Hamlet the Dane” (Hamlet, 5.1.236–237); “My name is Edgar” (King Lear, 5.3.161); “My name is Caius Martius” (Coriolanus, 4.5.64). It can also be playful: “Make but my name thy love, and love that still, / And then thou lovest me, for my name is Will” (Sonnets, 136.14).

21.

Bach (2007, 157) quotes Hakluyt, vol. 2, 201–202.

p. 669

References Bach, Rebecca A. (2007), Shakespeare and Renaissance Literature before Heterosexuality, Palgrave Macmillan, New York. Google Scholar Google Preview WorldCat COPAC Blake, N. F. (2002), A Grammar of Shakespeareʼs Language, Palgrave, New York. Google Scholar Google Preview WorldCat COPAC Charry, Brinda (2015), The Tempest: Language and Writing, Bloomsbury, London. Coleridge, Samuel. T. (2016), Coleridge: Lectures on Shakespeare (1811–1819), ed. A. Roberts, Edinburgh University Press, Edinburgh. Google Scholar Google Preview WorldCat COPAC Conilleau, Claire (2010), “ʻWhatʼs in a Name?ʼ H.D.ʼs Re-Vision of Shakespeare,” Transatlantica, 1, 1–14. Google Scholar WorldCat Derrida, Jacques. (1992), “Aphorism Countertime,” in Acts of Literature, ed. Derek Attridge, Routledge, New York, 414–433. Google Scholar Google Preview WorldCat COPAC Foucault, Michel. (1998), “What Is an Author?,” trans. Josué V. Harari, in Aesthetics, Method, and Epistemology, vol. 2 of Essential Works of Foucault, 1954–1984, ed. James D. Faubion, The New Press, New York, 205–222. Garber, Marjorie. (2004), Shakespeare A er All, Anchor Books, New York. Google Scholar Google Preview WorldCat COPAC Goldberg, Jonathan (2004), Tempest in the Caribbean, University of Minnesota Press, Minneapolis. Google Scholar Google Preview WorldCat COPAC Halpern, Richard (1994), “ʻThe Picture of Nobodyʼ: White Cannibalism in The Tempest,” in The Production of English Renaissance Culture, eds. David L. Miller, Sharon OʼDair, and Harold Weber, Cornell University Press, Ithaca, NY, 262–292. Google Scholar Google Preview WorldCat COPAC Jourdain, Silvester (1610), A Discovery of the Bermudas, Otherwise Called the lie of Divels, London: n.p. Google Scholar Google Preview WorldCat COPAC Kantorowicz, Ernst H. (1985), The Kingʼs Two Bodies, Princeton University Press, Princeton, NJ. First published in 1957. Google Scholar Google Preview WorldCat COPAC Lupton, Julia R. (2019), “The Tempest and Black Natural Law,” Religions, 10/2, 91.

Montaigne, Michel de. (1958), “Of Cannibals,” in The Complete Essays of Montaigne, trans. Donald M Frame, Stanford University Press, Stanford, CA, 150–159. Google Scholar Google Preview WorldCat COPAC Shakespeare, William. (2016), The Norton Shakespeare, 3rd edn, eds. Stephen Greenblatt Walter Cohen, Suzanne Gossett, Jean E. Howard, Katharine Eisaman Maus, Gordon McMullan, W. W. Norton, New York. Google Scholar Google Preview WorldCat COPAC Shakespeare, William. (2017), Cymbeline, ed. V. Wayne, Bloomsbury, New York. Google Scholar Google Preview WorldCat COPAC Smith, Emma. (2019), This Is Shakespeare, Penguin Random House, New York.

Google Scholar

Google Preview

WorldCat

COPAC

Strachey, Wiliam (1625), “A True Repertory of the Wrack and Redemption of Sir Thomas Gates, Knight upon and from the Islands of the Bermudas,” in S. Purchas, Hakluytes Posthumus or Purchas His Pilgrimmes, London: n.p. Google Scholar Google Preview WorldCat COPAC Wynter, Sylvia (2006), “Beyond Mirandaʼs Meaning: Un/silencing ʻDemonic Groundʼ of Calibanʼs ʻWomanʼ,” in The Black Feminist Reader, eds. J. James and T. D. Sharpley-Whiting, Blackwell, Oxford, 109–127. Google Scholar Google Preview WorldCat COPAC Žižek, Slavoj (2016), Disparities, Bloomsbury, New York. Google Scholar Google Preview WorldCat COPAC

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Print ISBN: 9780192844118

Search in this book

CHAPTER

30 Art and Language  Noël Carroll https://doi.org/10.1093/oxfordhb/9780192844118.013.55 Published: 22 May 2024

Pages 670–688

Abstract In this chapter, four issues are addressed—namely (1) whether art is a language on the model of a natural language, (2) whether art can be de ned, in terms of necessary and su

cient conditions (3)

whether all aesthetic terms are non-condition-governed in a strict or a looser pragmatic sense, and (4) what is the nature of (art) critical communication? It is argued that art is not a language, that art might turn out to be a cluster concept, that some aesthetic terms may be condition-governed in the looser pragmatic sense, and that most critical communications are not primarily a matter of inducing experiences but are purpose driven.

Keywords: art theory, Tolstoy, Bell, Weitz, family resemblance, cluster concept, Gaut, condition governed terms, Sibley, critical communication, Isenberg Subject: Philosophy of Language, Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

Is art a language? The notion that art is a language is virtually a cliché. It is the stu

of edifying public

speeches. For example, President Dwight Eisenhower said: “Art is a universal language and through it, each nation makes its own unique contribution to the culture of mankind.” And yet it is very di

cult to take this claim seriously. Eisenhower claims that art is a universal language, but

that each nation makes its own cultural contribution. Does the art of each nation speak its own language? If so, how is art a universal language? Also, both the Germans and the Russians compose Romantic symphonies. Are there two kinds of musical artworks here? Indeed, what does it mean to say that “art is a language.” It doesn’t sound literally true, if what we have in mind is a language properly so called. That is, how could Art—the collection comprising literature, music, dance, theater, painting, sculpture, photography, moving pictures, etc.—constitute a single language, since

these practices share no common vocabulary. Perhaps the suggestion is that each art form has its own vocabulary. But many art forms have no vocabulary. Many visual art forms, such as photography, communicate by virtue of functioning as naturally recognizable prompts—that is, we can tell what they symbolize by just looking at them rather than by reading words arbitrarily associated with their referents. And some art forms without words, for example, much music, and many abstract paintings have no representational content whatsoever and, thus, do not symbolize anything either by natural generativity or association. So, it cannot be the case that each and every art form is a language. Of course, literature does use the language of its author’s choice. But that is not the language of literature, since it is also the same language found in ordinary business transactions. Is there a separate language of literature with its own semantics? Unlikely in general, although there may be some experimental cases. Nor is it clear that literature as such has a grammar, since, for example, a poet’s violations of linguistic conventions are not artistically wrong, even if they are syntactic errors. They are rated as successful or not in terms of whether they realize the artist’s purposes. p. 671

In short, art, or even the arts, are not, strictly speaking, languages. Perhaps they are metaphorically languages insofar as they are communicative. But even here there are problems, since some artists create only for themselves. They have no intention to communicate with others; they may hide away or even destroy their works, lest others see them. Indeed, as we shall discuss, one theory of expression explicitly denies that artworks are necessarily for audiences. So, art is not a language. Nevertheless, several recurring linguistic questions arise with reference to the arts, including, Can art be de ned? Are aesthetic terms condition-governed? and What does a critical communication transmit to an audience? In what follows, these three questions are addressed. Can “art” be de ned? The question of the de nition of art—often referred to as “the theory of art”—asks whether the concept of art can be characterized in terms of necessary and jointly su

cient conditions.

The de nition of “art” was not a concern of philosophers like Plato and Aristotle. For most of the ancients, art was a skill with respect to any practice or craft, such as the art of medicine or of statecraft or for Aristotle, poetry. Thus, a list of the arts in the classical world was wildly more diverse than the list of the arts that we consider commonplace today, comprising roughly, poetry, music, dance, painting, sculpture, photography, motion pictures ( lm, video, television, etc.), and architecture. The origin of that system is 1

not something that arrives on the scene until the eighteenth century. And once membership in that system became xed, an account of why certain practices belonged in the system—that is, a speci cation of why they were the art forms, properly so called—became, almost naturally, a matter of philosophical curiosity. An early, highly in uential response to this philosophical query was the Abbé Charles Batteux’s 1746 2

treatise The Fine Arts Reduced to a Single Principle. That principle was the imitation of the beautiful in nature. So, painting was an art form insofar as it rendered its objects—from landscapes to buildings to people— beautiful in images, while poetry did so in words. Theater represented actions beautifully and dance human movements. An art form belongs to the modern system of the arts if and only if it imitates the beautiful in nature; that is the single principle that makes a practice an art form under this early modern conception. But this account faced numerous challenges. Perhaps the most compelling was the rise to prominence of absolute or pure music—orchestral music sans words or program. Such music arguably imitated nothing, and yet was coming to be regarded as the art form to whose condition all the other art forms were said to aspire. Clearly, another account of what de ned art in the modern sense was needed. One candidate, abetted by the Romantic movement, was that art properly so called was the expression of emotion. Expression of the inner world—the poet’s feelings, for example—was the aim of art rather than the imitation of the outer world of nature. Expression theories of art proliferated in the nineteenth and

twentieth centuries. Probably, they still articulate what most people today take to be the de ning feature of art. Two of the most important expression theories of art were those of Leo Tolstoy and R. G. Collingwood. p. 672

For Tolstoy, art was the sincere communication of the self-same feeling experienced by the artist to her 3

audience. In contrast, for Collingwood, art was the clari cation of the artist’s emotion, capturing its unique timbre and thereby making external her inner life—not for an audience, but for the artist herself, to get in 4

touch with her own feelings.

Expression theories of art were undoubtedly more comprehensive than imitation theories. And yet, they, too, are not inclusive enough. Some works of art are not about feelings; some works of art are rigorously cognitive. For example, Jorge Luis Borges’s short story “Pierre Menard, Author of the Quixote” is a philosophical demonstration of the proposition that textual identity is not the same as artwork identity. No one’s feelings are being expressed. Similarly, many of Marcel Duchamp’s readymades had nothing to do with his emotions and more to do with his intellectual interrogation of the nature of art. Another reaction to the inadequacy of the imitation theory of art was the attempt to locate the essence of art in the form of the artwork. This probably was a result of the earlier association of art with beauty, where beauty, as in Kant, was connected to the form of purposiveness of the stimulus, rather than to the actual purpose of the artwork. For obvious reasons, this approach to de ning art is referred to as formalism. The 5

view was set forth most assertively in Clive Bell’s book Art.

According to Bell, something is art just in case it possesses signi cant form which, in turn, is identi ed in terms of the capacity of the work to engender an aesthetic emotion, a feeling of release from the concerns of everyday life. This theory of art confronts several problems, including the fact that its key concepts— signi cant form and aesthetic emotion—are de ned in a circular manner. But in addition, formalism is not inclusive enough. Duchamp’s philosophical meditations, such as the readymade Canine Grooming Comb, is an artwork in virtue of its cognitive adventurousness, which is foregrounded by its studied lack of any formal interest whatsoever. In the evolution from Batteux’s imitation theory to the expression theories and the formal theories, a signi cant shift in the object of the de nitions occurred. Whereas the objects of Batteux’s theory were art forms; the objects of the expression theories and the formal theories were individual artworks. This distinction is sometimes masked by the fact that both de nitional projects can be called de nitions of art. But this ambiguity invites equivocation, as we’ll see shortly. Once the enterprise of de ning art took o , proposals proliferated. But no sooner were de nitions constructed than they were challenged. This raised the question of whether or not there was something wrong with the very attempt to de ne art in terms of necessary and conjointly su

cient conditions. Several

philosophers, in uenced by Wittgenstein’s discussion of the concept of games in his Philosophical Investigations, argued that art could not be essentially de ned. Undoubtedly the best-known statement of this view—often called neo-Wittgensteinianism—was Morris 6

Weitz’s 1956 paper “The Role of Theory in Aesthetics.” In this in uential article, Weitz did three things: (1) he proposed an argument—called “the open concept argument”—to establish that art could not be de ned p. 673

in terms of necessary and conjointly su

cient conditions; (2) absent an essential de nition as

the means

to identify something as art, Weitz proposed an alternative method, which relied on family resemblances for determining the art status of works; and (3) he o ered an error theory, which explained what preceding philosophers were really up to in propounding their putative de nitions of art. Wittgenstein pointed out that many of the concepts that we employ in everyday life—like the concept game —are not de ned in virtue of necessary conditions that are jointly su

cient. In this respect, they are not

closed concepts, like the concept of a prime number. They are open concepts. Weitz argues that art is an

open concept. He does not do this on the grounds that all known attempts to de ne it have failed. Rather, he maintains that the attempt to de ne art itself would be self-contradictory. Why? Because art is a realm of practice that is permanently open to the possibility of innovation and expansion. This is not to say that all art explores new avenues of creativity, but only that “innovation” is always an option when it comes to art. The “rules of art” at time T1 are always open to being violated at time T2 by the genius-artist, thereby instituting a new “rule of art” at time T3. Given that the permanent possibility of expansion belongs to the concept of art, it would be self-contradictory to attempt to dictate necessary and jointly su

cient conditions, or even merely su

cient conditions, for art status.

But if artworks are not to be identi ed by essential de nitions, how are we going to pick out the artworks from the non-artworks? To this end, echoing Wittgenstein, Weitz recommends the family resemblance method. Family resemblances obtain across family members multidirectionally. That is, you may have your mother’s nose, your aunt’s chin, your grandmother’s generosity, your father’s temper, and your grandfather’s baldness. Family resemblances, in other words, are diversely distributed across a number of di erent people in terms of a variety of not necessarily converging dimensions. Rather than there being one necessary strand of resemblance, shared by each and every family member, there are strands of overlapping resemblances—your aunt’s chin, your grandfather’s baldness—criss-crossing your kinship group. Moreover, for Weitz following Wittgenstein, most of our concepts, including art, are applied like the concept of family resemblance. So, confronted with a new and unfamiliar work—like Borges’s “Pierre Menard, Author of the Quixote”— one would decide whether or not it is an artwork based on whether it convincingly resembles antecedently acknowledged artworks. For example, like Poe’s “The Tell-Tale Heart,” it is a ction; like Ring Lardner’s “Haircut,” it is wryly ironic; like Velasquez’s “Las Meninas,” it is re exive, and so on. As you survey the preexisting network of established artworks, you chart the accumulated resemblances to accepted paradigms until you feel that there are enough correspondences to declare the new work to be part of the family. Just as one determines whether a futon is a bed by looking to see how much it resembles already familiar beds, so one decides whether an avant-garde work is an artwork by looking to see and assessing the degree to which it resembles antecedent artworks. Inasmuch as traditional art theories aspired to do the impossible—to de ne an open concept—they were p. 674

fated to fail. But there was a certain regularity discernible in their

recurring defeats. Although they were

ostensibly striving to de ne the nature of all art, what they actually did was to fallaciously elevate the features of the art they favored above all other features to the status of the essence of art. Clive Bell was a champion of neo-impressionism and identi ed formal values as opposed to representational values as the quiddity of art; Tolstoy’s admiration of folk art’s infectious communal sincerity disposed him to elevate those qualities as the de ning features of genuine art. And so on. The continual error in traditional de nitions of art is that their proponents mistake their critical preferences for ontological discoveries. Of the three parts of Weitz’s position, the third part—the error theory—appears to be the only one that has survived criticism. Both the open concept argument and the family resemblance method have been subjected to intense and unfavorable scrutiny. The open concept argument would seem to ride on an equivocation. It presupposes that art is open to the permanent possibility of expansion and innovation. However, if that is true, it is true of the practice of art. Most individual artworks are not open to the permanent possibility of expansion and innovation, save in special cases, such as process artworks. The Mona Lisa is not open to innovation. If you attempt to “expand” it, you will be arrested. Its nature is xed.

Why is this a problem for Weitz’s argument? Because the theories of art that he intends to unhorse are ones that attempt to de ne what it is to be an artwork, not what practice counts as art (as opposed to sport). Thus, even if the practice of art is always in principle open to innovation, and, therefore, supposedly inhospitable to de nition, it does not follow that this presumption blocks de ning the concept of the artwork, since individual artworks are not typically open to the permanent possibility of change. That is, you cannot deduce the notion that the concept of a work of art is an open concept from the supposition that art as a practice is an open concept. That is just an equivocation. Furthermore, the claim that art (in the artwork sense) cannot be de ned in terms of necessary and jointly su

cient conditions because of the innovative nature of art has been shown to be false empirically by

de nitions of art like George Dickie’s Institutional Theory of Art, which, though set out in terms of necessary and jointly su

cient conditions can, nevertheless, accommodate the most radical artistic

innovation. The family resemblance method for identifying works as artworks has also come in for devastating criticism. The rst is the logical truism that everything is like everything else in some respect. Michelangelo’s David is like a sewer pipe in that both are material objects, both are metal, both are shaped, human artifacts, and so on; but the sewer pipe is not an artwork. You might think that the neo-Wittgensteinian can easily dodge this bullet by reminding the critic that he or she is not talking about just any resemblance; she or he is talking about family resemblances. But is that accurate? Putatively, tracking family resemblances is an alternative to identifying items by means of de nitions in terms of necessary and/or jointly su

cient conditions. But “family resemblances” properly so called are

not simply resemblances, come-what-you-will. They are family resemblances, which means that they are condition-governed. To resemble your cousin in a family way, you must belong to the same gene pool. If you p. 675

happen to have the same or a similar baritone voice to the one that Gregory Peck had, you don’t bear a family resemblance to him, unless he is a blood relation. Thus, the very notion of a family resemblance is logically at odds with Weitz’s project, since membership in the same gene pool is governed by necessary and jointly su

cient conditions.

As Maurice Mandelbaum pointed out, an authentic family resemblance depended on nonmanifest genetic properties, whereas the properties cited by philosophers of art were typically manifest properties, such as 7

imitation, beauty, expressivity, and form. Indeed, Weitz’s own recommendation that we look for resemblances between established artworks and candidates for art status suggests that he, too, was thinking in terms of manifest properties. But, Mandelbaum pointed out, that left open the possibility that a de nition of art might be framed with reference to nonmanifest properties, just as family resemblances depended on genetic features. Mandelbaum’s insight into the relevance of nonmanifest properties for the project of de ning was already 8

implicit in Arthur Danto’s seminal article “The Artworld.” In that essay, Danto, inspired by works like Warhol’s Brillo Box, conjectured that the features that de ne art are not something the eye can recognize. They are indiscernible. For example, whatever makes Warhol’s Brillo Box art, but Proctor and Gamble’s not art, is not something that you can see. Danto actually goes so far as to suggest one necessary condition for art status—namely, that the work be forged from a perspective informed by art theory and/or history, which are nonmanifest, generative properties of artworks. If Danto partially, albeit implicitly, de ned the artwork in terms of generative, nonmanifest historical properties, George Dickie, taking the hints from Mandelbaum and Danto, explicitly de ned the artwork sociologically—that is, in terms of social facts that are not manifest on the surface of the artwork.

In the inaugural statement of his Institutional Theory of Art, Dickie argued that the often neglected, but utterly crucial, factor in determining art status is the social context in which it is presented. According to Dickie, “A work of art in the descriptive sense is 1) an artifact 2) upon which some society or some sub9

group of a society has conferred the status of candidate for appreciation.”

For Dickie, this notion of a “conferral of status” was analogized to social rituals like marriage. But critics pointed out that marriage is a social institution with established rules. For instance, only certain o

cials

have the power to marry couples, and only certain couples can be married to each other. No one can preside over the marriage of a squirrel and a tarantula. But, in contrast, there don’t seem to be comparable rules governing the art world. Anything can be art, and it seems that art can be presented by anyone. For that reason, the art world does not appear to count as a genuine institution. Moreover, if it is an institution, how would Dickie propose to handle the existence of prehistoric art—art before there was an established art world? That is, institutional status cannot be a necessary condition for art insofar as there was art before there was an institution of art. And a further, frequent objection to Dickie’s Institutional Theory of Art charged that it was circular. Dickie maintained that someone had to present the work as a candidate for appreciation. But what kind of p. 676

appreciation? The obvious answer is art appreciation.

However, then the de nition is circular. Also, what

kind of person has to confer the status of candidate for appreciation? Someone acting on behalf of the 10

artworld? Circularity strikes again.

Danto himself has attempted a partial de nition of art. Following in Hegel’s footsteps, Danto maintains that something is an artwork only if (1) it is about something and (2) it is articulated in a form that is intended to 11

be adequate or appropriate to whatever it is about.

Danto tends to understand the “aboutness” condition

in terms of “having a meaning” and the notion of “form” in terms of “embodiment.” So, for Danto, an artwork is an embodied meaning—something that has a meaning that is “ eshed out” or “embodied” in a form appropriate to that meaning, as were the use of Corinthian columns in the decoration of Greek temples dedicated to oral goddesses. One problem with this analysis is that it appears to presuppose that it is a necessary condition for any artwork that it have a meaning. But aren’t there arts that are, so to speak, beneath meaning? Works, for example, that are simply beautiful—not about beauty, but just ravishing? Danto’s response to this objection is to say that for any example that his critics can come up with, he can identify a meaning for the work—either in terms of ritual meanings or artworld meta-discourses. Needless to say, such bravado has not silenced Danto’s critics. But Danto has a second, more persuasive response: although not every artwork has a meaning, it is always legitimate, if something is an artwork, to ask whether it has a meaning. It never makes sense to ask whether gardening tools have a meaning. Yet, if something is an artwork, it is at least a candidate for interpretation. The problem here, however, is that, if in certain cases, you need to know it is an artwork in order to know if it is a candidate for interpretation, then being a candidate for interpretation cannot be a necessary condition for being an artwork without courting circularity. A perhaps even more perplexing problem for Danto’s theory of art is that it doesn’t succeed in doing what Danto claims is the primary task of a de nition of art—to wit, to discriminate between artworks and real things. To see how this problem arises, consider one of Danto’s favorite examples, Warhol’s Brillo Box. On Danto’s account, Warhol’s Brillo Box is a work of art whereas Proctor and Gamble’s is a real thing. Yet, Danto’s theory of art as embodied meaning will not deliver that verdict. Proctor and Gamble’s carton is about something, namely Brillo, about which it claims is cleansing in forms appropriate to its meaning— red, white, and blue patriotic waves.

Nor is this only an isolated problem. If manufactured products are real things, then most of them nowadays are artworks à la Danto since they typically are designed in packaging that make a “statement” about what’s inside them. Danto’s theory of art, as stated, comprises only two necessary conditions. Perhaps the addition of further conditions will deal with this problem. But until then, the burden of proof belongs to Danto and his friends. The e orts of Dickie and Danto, especially, ended the neo-Wittgensteinian moratorium on attempting to de ne art. Monroe Beardsley, for example, proposed what was called an aesthetic de nition of art that de nes p. 677

the artwork as (1) an artifact (2) intended

12

to a ord some magnitude of aesthetic experience.

This

proposal, of course, confronts some of the problems that we encountered in Danto’s theory. So many commodities are designed to a ord some magnitude of aesthetic experience, but we do not regard all those cereal boxes as artworks. Nor will it help to stipulate that the primary intention in the creation of the artwork is to a ord aesthetic experience; the primary intention in the creation of Greek Orthodox icons was to facilitate spiritual communication, even though they could support aesthetic experiences on the part of nonbelievers. Moreover, the aesthetic theory of art is ill-suited to accommodate most of the anti-aesthetic art of the various avant-gardes of modernity. Like Beardsley, Jerrold Levinson emphasizes the intention the artist has in creating the artwork. For him, something is an artwork if and only if it is made by a person (1) who has a proprietary right over the work in question, and (2) who non-passingly intends the work for regard as a work of art—i.e., in one or more of the ways that artworks have been correctly regarded historically. Levinson argues that this de nition is not circular because we do know historically which works were counted as artworks and which ways of regarding them were the right ones. Because the intention introduced in the second condition of this 13

de nition is dependent on art history, Levinson calls his theory de ning art historically.

It is not clear why Levinson feels compelled to require that the artist have a proprietary right over the artwork. Did Frank Lloyd Wright have a proprietary right with respect to the Guggenheim Museum? But in any event, it is the art-regard condition that is the most important part of Levinson’s de nition. Levinson calls the de nition a historical one. But it really isn’t historical because it fails to take into account the fact that certain art-regards that were once correct are now obsolete. For example, at one point in history, appreciating the verisimilitude of pictorial representations was a correct art-regard. But I doubt that we would count as artworks most of the pictures taken on our cell phones during our vacations with the intention that they be appreciated for their perceptual accuracy. That would not only be historically anachronistic; it would leave us with a glut of (mostly undistinguished) artworks. Most of the theories of art canvassed thus far attempt to de ne the artwork directly in terms of art-making properties. Dominic Lopes de nes what it is to be an artwork indirectly, as something that is an instance of an art form, where he believes it will be possible to supply pro les of the relevant practices without de ning Art with a capital A. I am skeptical of this project because I maintain there may be artworks that do not belong to any extant artforms. Afterall, hasn’t that been the lesson of the proliferation of avant-gardes for the last one hundred and fty years? Lopes is aware of this objection, and he attempts to deal with it by means of two strategies: (1) by claiming that apparent outliers are really instances of Conceptual Art or (2) by claiming that they are forerunners of what will eventually become established art forms. The Conceptual Art gambit is ad hoc. Lopes’s category is not the established category of Conceptual Art as we know it. It is a catchall, question-begging collection stipulatively concocted of counterexamples to Lopes’s p. 678

theory masquerading as an actual historical

category of art. It is not the category of Conceptual Art as art

historians know it. It is an assembly of objects that beg the question.

Lopes’s second strategy also fails. He says that something may be art in virtue of belonging to a category that becomes established after other works of that sort are recognized as constituting said category. But this not only precludes the possibility of one-o

works of art; it also means that the initial work becomes a work

of art retrospectively. But doesn’t the work of art have art status at the moment of its inception? Wasn’t Picabia’s 1909 painting Caoutchouc a work of art as soon as the watercolors started drying? He didn’t have to wait for Kandinsky’s and Malevich’s abstractions to arrive on the scene. As proposed de nitions of art proliferated in the post-Dickie era, suspicions returned about the plausibility of the project. Berys Gaut, again returning to Wittgenstein for inspiration, resuscitated the notion of a “cluster concept,” arguing that it is adequate for classifying a candidate for art status as an artwork if it scores well in light of the following ten criteria: (1) it possesses positive aesthetic properties; (2) it expresses emotion; (3) it is intellectually challenging; (4) it is complex and coherent; (5) it has the capacity to express complex meanings; (6) it exhibits an individual point of view; (7) it is an original exercise of the imagination; (8) it is the product of skill; (9) it belongs to an established form of art; and (10) it is made with 14

the intention to be a work of art.

This is not an essential de nition of art. None of the properties on the list are necessary conditions for art status. Something that is art will possess at least one of these features. And as a work is recognized as having additional ones, it provides us with further reasons to categorize it as an artwork. For Gaut, a cluster concept is true of a concept, like art, just in cases it picks out features whose possession by the relevant works count necessarily toward its membership in that category. Moreover, Gaut does not think that the cluster concept approach as a nonde nitional approach to identifying artworks stands or falls with his list. He asserts that that even if there are counterexamples to his formulation, the cluster concept is still a viable approach. Perhaps needless to say, possible counterinstances are conceivable. One can readily imagines a master baker who for his twenty- fth anniversary confects a cake as gift for his wife that meets most of Gaut’s conditions. It is beautiful, expresses his unique and personal kind of love, has intellectually challenging and imaginative references to di erent events in their marriage, is complex and coherent in its design, is a product of skill, and is made by man who believes cooking is an art. But even if the baker intends his cake to be a work of art in the descriptive sense, it is unlikely that the editors at either Artforum or the Architectural Digest are likely to commission articles about it. Gaut’s list of properties may not only be overinclusive; it may also be too exclusive. For example, it may not accommodate Duchamp’s Fountain—at least in terms of certain interpretations of it—notably, the account that regards Duchamp as intending Fountain to be nothing more than a snarky retort to the essentially hypocritical, the self-avowed openness of the curators of the 1917 exhibition presented by the Society of Independent Artists. However, even if this account could be supported by compelling documentary p. 679

evidence, and Duchamp did not intend Fountain to be an artwork or to possess any of

the other features

Gaut lists, I suspect that art historians would still argue that it is not only an artwork, but a seminal one. These kinds of considerations, of course, need not be decisive. Logically, Gaut’s cluster concept approach may be repaired by either adding and/or subtracting the items on the list. And/or, the criteria on the list may be re ned in order to block misapplication. Or, Gaut may reject the preceding counterexamples on the grounds that they are misleadingly described. Furthermore, recall that Gaut is not committed to his version of the cluster concept approach as the nal version to be preferred. Nevertheless, even conceding this, at this point in the debate, the burden of proof weighs on Gaut. Are aesthetic terms condition-governed? The properties of a piece of choreography are describable both in aesthetic and non-aesthetic terms. The non-aesthetic terms include “The composition is made up of six dancers”; “The piece lasts for 20 minutes”; “The music that accompanies the dance is John Philip Souza’s

Stars and Stripes Forever”; “The dancers wear black tights and white T-shirts”; “There were three female and three male performers”; “Each dancer jumped three times per minute”, and so forth. Examples of aesthetic terms that might be applied to the same piece of choreography include “The dancers were graceful”; “The choreography was tightly knit and uni ed”; “The leaps were zesty”; “The relation of the music to the movement was imaginative”; “The scenery was beautiful,” and so on. The philosopher Frank Sibley was interested in working out the distinction between aesthetic terms and non-aesthetic terms and in arguing that the aesthetic terms, in contrast to nonaesthetic terms, are not 15

condition-governed.

For Sibley, a term is condition-governed if it can be de ned in terms that are necessary and conjointly su

cient. For example, “An equilateral triangle is a three-sided gure where all three sides are equal in

length.” However, this is not the only way in which a term can be condition-governed. Sibley regards the term “intelligent” as condition-governed, although it cannot be de ned in virtue of conditions that are necessary and conjointly su

cient. For example, if someone is a chess master, speaks seven languages, has

a PhD in physics from the Massachusetts Institute of Technology, wrote the de nitive study of Sumerian economics, and discovered the cure for cancer, you could not reasonably deny that she was intelligent and leave it at that. Even though none of the factors that I’ve just listed are necessary conditions for intelligence, and even though there are other features that could be added to the list and/or substituted for the terms on the list, if the features I’ve listed obtain and no countervailing considerations are put forward, we cannot rationally withhold the application “intelligent” to our chess-playing MIT graduate. Moreover, it is not that the features on the list are su

cient for the chess-playing MIT graduate to be

intelligent. Rather they are, all things being equal, enough to make counting as unreasonable the denial that she is intelligent. Sibley considers laziness to be similarly condition-governed. At a certain point, a list of mental and behavioral attributes would be such that were they to aptly describe someone, all things being equal, it would be unreasonable to deny that he is lazy. p. 680

So, there are two ways in which a term may be condition-governed. It may be strictly condition governed, as in the case of the term “equilateral triangle.” Or, a term may be more loosely, or “pragmatically,” condition-governed; that is, a person, place, or thing may have enough features of a certain kind and lack any pertinent canceling features such that we, all things being equal, cannot rationally withhold the application of the relevant term to it, even though no complete list of the pertinent features may ever become available. In this way, pragmatic condition-governance somewhat resembles the cluster concept idea. For Sibley, aesthetic terms are not condition-governed either in the strict sense or the pragmatic sense, with one exception. In certain cases, the nonaesthetic features of a work may preclude description of the work by means of various aesthetic terms. For example, if a painting is nonaesthetically characterizable as faint, pastel, monochromatic, and uniformly grey, it cannot be garish. So, even though no number of nonaesthetic features ever entails that a given object is describable by a particular aesthetic term, the possession of certain nonaesthetic properties may block the attribution of certain aesthetic terms to the 16

object. That is, aesthetic terms may be governed negatively by nonaesthetic conditions.

Lastly, Sibley stresses that the aesthetic terms he has in mind are descriptive terms, not evaluative terms. A picture could be accurately called pretty from the descriptive point of view, and yet not be a good picture from the evaluative point of view. Many of Thomas Kinkade’s paintings may be like this.

The aesthetic properties of an object depend on its nonaesthetic properties, though the applicability of certain nonaesthetic terms is never logically su

cient for applying a given aesthetic description to it.

So far, we have been relying on examples of aesthetic and nonaesthetic terms to set out Sibley’s argument. But what makes one term aesthetic and another term nonaesthetic? Sibley lays out the distinction in the 17

opening paragraph of his seminal essay “Aesthetic Concepts”:

The remarks we make about works of art are many kinds. In this paper I wish to distinguish between two groups. We say that a novel has a great number of characters and deals with life in a manufacturing town; that a painting uses pale colours, predominantly blues and greens and has kneeling gures in the foreground; that the theme in a fugue is inverted at such a point and that there is a stretto at the close; that the action of a play takes place in the span of one day and that there is a reconciliation in the fth act. Such remarks may be made by, and such features may be pointed out to, anyone with normal eyes ears, and intelligence. On the other hand, we also say that a poem is tightly knit or deeply moving; that a picture lacks balance, or has a certain serenity or repose or that the grouping of gures sets up an exciting tension; that the characters in the novel never really come to life, or that a certain episode strikes a false note. It would be natural enough to say that the making of judgments such as these requires the exercise of taste, perceptiveness, or sensitivity, of aesthetic discrimination or appreciation; one would not say this of my rst group. p. 681

Accordingly, when a word or expression is such that taste or perceptiveness is required in order to apply it, I shall call it an aesthetic term or expression, and I shall, correspondingly, speak of 18

aesthetic or taste concepts.

So, for Sibley the objects of nonaesthetic terms are properties that people of normal cognitive and perceptual abilities can identify just by employing those abilities in the usual fashion. Although Sibley does not use this language, they are the base properties on which aesthetic properties supervene. The apprehension of these aesthetic properties, however, requires a certain kind of sensitivity that Sibley calls taste. Consequently, in order to apply an aesthetic term correctly, a percipient must possess taste, which is just the capacity to notice emergent aesthetic properties. Although Sibley is recruiting a traditional idea—namely, that of taste—his is usage is not exactly the same 19

as that of a key gure in the tradition, like Hume.

Where taste, for Hume, involves noticing certain

properties in the object, it also involves feeling (pleasure or aversion) and evaluation (approbation or disapprobation); for Sibley, in contrast, taste is only a matter of sensitivity or qualitative discrimination—of having the capacity for noticing certain properties not perceived by normal folk enlisting only their ordinary cognitive and perceptual resources. Perhaps Sibley thinks that taste, as he construes it, must be involved in order to explain how it is possible for percipients to apply aesthetic terms if said terms are not conditiongoverned. For Sibley, aesthetic descriptions employing aesthetic terms have truth values. In this regard, they are like ordinary, nonaesthetic descriptions. What is distinctive about them is that they cannot be inferred on the basis of the applicability of any number of nonaesthetic terms. Here, Sibley’s position recalls, in part, Kant’s 20

analysis of “free beauty.”

Just as Kant denied that one could derive “x is beautiful,” by subsuming an

object under a concept, Sibley likewise denies that one can apply the aesthetic term “beauty” to an object on the grounds that it has satis ed a list of nonaesthetic descriptions. Indeed, Sibley believes this pertains not only to beauty, but to the whole range of aesthetic terms, including graceful, elegant, delicate, uni ed, tightly knit, garish, serene dynamic, tragic, somber, balanced, integrated, moving, tense, powerful, and so on. Sibley further believes that one can expand this list based on these examples, although that is a matter of 21

controversy.

Sibley appears to assume that an aesthetic term is one whose correct application requires taste. Taste, in turn, is understood as a capacity for insight above the range of the normal perceptual and intellectual capabilities. So, is Sibley presuming that wherever a term is applied on the basis of an insight that exceeds the norm, the term is not condition-governed? But that seems false. A trained mathematician may detect the implication of a battery of propositions beyond the ken of a normally numerate person, but that would not establish that his insight was not condition-governed. That is, the possession of a special ability or sensitivity above the norm does not imply that the insights issued in virtue of that special capacity for insight or sensitivity are not conditiongoverned. With respect to aesthetics, the musical term stretto is a condition-governed term (as Sibley apparently p. 682

22

concedes) but arguably requires a “musical ear” to be detected.

Thus,

the use of aesthetic sensitivity—

also known as taste—would appear to entail nothing general regarding the logical behavior of aesthetic terms. Furthermore, Sibley’s apparent identi cation of taste with the operation of perceptual and intellectual abilities above the norm also seems dubious. It would seem to be an incontrovertible fact that ordinary folk, exercising no more than their normal cognitive and perceptual resources, can easily and accurately describe this morning’s sunrise as stunning, last night’s thunderstorm as powerful, Raphael’s Assumption of the Virgin as beautiful, and on and on. … Is it that they can apply aesthetic terms without resorting to taste? But then taste does not appear required to describe things aesthetically. What, then, distinguishes aesthetic terms from nonaesthetic terms? It is not obvious, since there is some controversy, as was mentioned earlier, regarding the assumption that it is unproblematic to go on from an 23

initial list of so-called aesthetic terms to further candidates.

Moreover, given that ordinary folk using their normal perceptual and cognitive faculties can readily and correctly describe a well-cut suit as elegant when they see one, doesn’t that imply that elegance is condition-governed? Or worse: is elegant conditioned-governed when ordinary folk apply the term, yet not condition-governed when the aesthetically sensitive and discriminating taste masters do so? Or: are there nonaesthetic elegance and aesthetic elegance? These seem unhappy alternatives, suggesting that the aesthetic/nonaesthetic distinction in terms of taste may be incoherent. Furthermore, the claim that aesthetic terms are not condition-governed in the looser, “pragmatic” sense seems unconvincing. In the rst movement of Beethoven’s third symphony, the Eroica, the main theme 24

repeats at least thirty-seven times.

Wouldn’t someone who denied, upon hearing this description, that the

movement was uni ed be reacting irrationally? If our previous description of the chess-playing MIT graduate rendered it unreasonable to refuse to describe her as intelligent, then, surely, to deny the ascription of unity to the opening movement of Beethoven’s Eroica based on the aforesaid description (given that it is accurate and manifestly unchangeable, all things being equal) would be similarly unreasonable. That is, if the description of the MIT graduate as intelligent is condition-governed, it seems that the description of the rst movement of the Eroica as uni ed is equally condition-governed, at least in the looser sense of being condition-governed (which Sibley apparently accepts). Is it that Sibley thinks that the aesthetic case is not condition-governed because we never have a list complete enough to guarantee su

ciency? But that is more than the looser sense of condition-governance

—which he endorses—requires. On that conception, it would be nonsensical to deny that the repetition thirty-seven times of the main theme in the movement of a symphony, all things equal, warrants describing the movement as uni ed.

One consideration that Sibley raises on behalf of his claim that aesthetic claims are not condition-governed is that “a painting that has only the kind of features one would associate with vigor and energy but which even so fails to be vigorous and energetic need not have some other character, need not be instead say, p. 683

strident or chaotic. It may fail to

25

have any particular character at all.”

However, it is not clear that this

example shows that the terms in question are not condition-governed. The explanation of the painting’s characterlessness could be that the object does not possess the features in question to the right degree. Not wanting to go to work is a feature associated with laziness, but not wanting to go to work would not alone be enough to make to attribution of laziness to someone who felt that way something it would be outlandish to deny. It may be part of a package of things that lead us to attribute laziness to someone, but without the rest of the package, the authority of reason may not compel one’s interlocutors to agree. Thus, there may be nonaesthetic features associated with vigor and energy in an object but not enough of them to warrant describing the work as vigorous or energetic. We don’t regard chess mastery alone as enough to grant the non-deniability of the attribution of intelligence to someone; the chess master may be an idiot savant. Similarly, in order to apply an aesthetic term to an object, it may need more than one or two nonaesthetic conditions to warrant the attribution of a particular aesthetic character description that would be unreasonable to disavow. At one point, Sibley says that a reason to regard aesthetic terms as not condition- governed is that no matter how many nonaesthetic features are enumerated, “it is also possible for me to wonder whether in 26

spite of these features, it is really [in the relevant case] graceful and balanced.”

Possible? Maybe. But

rational? Is it really reasonable, all things being equal, to wonder if the movement is uni ed if its main theme is repeated no fewer than thirty-seven times? Or is Sibley shifting the standard of what it is required to be condition-governed from the looser or pragmatic standard to the stricter standard? Another consideration that Sibley advances in support of his thesis is this: “An object described very fully but in qualities characteristic of delicacy, may turn out on inspection to be not delicate at all, but anaemic or 27

insipid.”

However, this is a rather strange example for Sibley to propose. For “anaemic” and “insipid,”

sound like evaluative terms, whereas Sibley’s thesis supposedly only pertains to descriptive aesthetic terms. Thus, there is no tension between an object being described as “delicate” aesthetically, while, in addition, being evaluated as “insipid” in light of further considerations. Consequently, that the object is demeaned as insipid does not entail that the attribution of delicacy was not condition-governed. Furthermore, if a description, even a very fully developed one, were, upon inspection of the object in question, to fail to support the putative application of a certain aesthetic term, that may very likely be the result of the object being misdescribed nonaesthetically, and not because the term in question is not condition-governed. The fact that you cannot infer from just a verbal description of an object, even a very full one, does not show the object is not condition-governed; you still have to check to see that the description is accurate. It is perfectly consistent for the application of an aesthetic term to be conditionedgoverned, though the description upon which it is based be defective. p. 684

What does a critical communication transmit to an audience? One common understanding of the critical communication conceives of it as tripartite: There is a value judgment or verdict (V): “this picture or poem is good—.” There Is a particular statement or reason (R): “—because it has such-and-such a quality —.” And there is a general 28

statement or norm (N): “—and any work which has that quality is pro tanto good.” So, on this view, stated formally, a critical communication looks like this: (1) The artwork A has property P.

(2) Any artwork with property P is pro tanto good. (3) Therefore, the artwork A is pro tanto good. The problem with this model, however, is that putatively there are no universally good-making properties 29

of the sort assumed in the second premise. 30

always positive.

Consider unity, a property that Monroe Beardsley believed was

Yet the unity in the novel Jack Torrence is writing in Stephen King’s The Shining is hardly

good. It is monotonous. It is as dull as Jack is in the novel within King’s novel. Indeed, Arnold Isenberg writes: “There is not in all the world’s criticism a single purely descriptive statement concerning which one is prepared to state beforehand, “If it is true, I shall like that work so much 31

the better.”

One reason this seems persuasive is that for any apparently positive artistic property, say,

tragic terror, it may not combine with other properties of a particular work felicitously. It could render a romantic comedy incoherent. That is, typically the relevant properties in artworks come in combinations and those combinations may not mesh. But it is unimaginable that there are general rules that would cover all the possible combinations in all their varying degrees of salience. But if the preceding deductive model of the critical communication is unconvincing, what is the proper account? And speci cally, why are descriptive properties of the work put forward, if they are not to be taken as evidence for the application of a speci c norm such as the ones set forth in the second premise above? For Isenberg, the critic calls attention to certain properties of a work, say the symmetrical disposition around the cruci ed Christ in a painting, in order to induce the viewer to focus on that property in such a way that the viewer will experience the composition in the way the critic experiences it. That is, the critic tries to isolate the features of the work that account for the way she feels—positively or negatively—about the work with the aim of eliciting the same feeling in viewers. In this way, the viewer accepts the critic’s communication, not on the grounds of its being a generalization, but on the grounds of his or her experience. For example, this is the way in which Peter Schjeldahl addresses his readers with regard to Jacopo da Pontormo’s Portrait of a Halberdier: “Allow the hues and tones to surprise and absorb your gaze. You might p. 685

32

well swoon.”

If Schjeldahl is able to arouse the

same enthusiastic feeling in the reader that has gripped

him, the readers will likewise share his approval of the painting. Isenberg’s characterization of critical communication recalls Clive Bell’s proposal in Art of the way in which to conduct critical debates. If we disagree about whether an artwork contains signi cant form, what we will attempt to do is to get our interlocutor to see the work in the same way we do, by pointing to properties and relations in the work, by describing it metaphorically, by making comparisons and contrasts with other works—in other words, by trying in various ways to get others to experience the work as we do. We persuade others not by reasoning, as the common understanding we began with suggests, but by convincing them, so 33

to speak, experientially.

Undoubtedly, this is the form some aspects of critical communication take. But is it the whole story? Might there be a species of critical communication that proceeds by reasoning and argument rather than exclusively by “conversion experiences?” 34

Works have constitutive purposes—purposes that make them the very artworks that they are.

Christian

artworks portray the su ering of Christ in order to arouse a sense of gratitude in the faithful for the agony Jesus underwent for them. The art critic points to the way in which this purpose is realized or implemented or articulated in a work in terms of the way incidents are imagined but also in terms of the postures, shapes, colors, hues, vectors, and so forth, that the artist elects. For example, the critic emphasizes how Christ is slouched downward on the cross, his body language signaling pain and defeat.

On this view, the art critic shows—by observation, reasoning, and argument—why these very choices are appropriate or inappropriate, adequate or inadequate, to the achievement (or not) of the constitutive purpose (or purposes) of the work. In this, the critic need not rely upon generalizations of the sort that Isenberg rejects, but only on explanations about how speci c artistic choices are functioning in the particular work. Moreover, this is an alternative model of critical communication that Isenberg champions, 35

apparently to the exclusion of alternatives.

One objection to what we may call the purpose driven account is that it makes art criticism too much like the criticism of tools, like steak knives. Here the worry may be that the evaluation of tools is governed by generic norms such as “steak knives should be sharp.” However, the constitutive purposes of artworks are unique, and their realization is therefore customized to their particular goals and contexts. Thus, accounting for the success or failure of an artwork is more like accounting for a historical event than like evaluating a tool, and can, for that reason, be done without reference to general laws. Summary: In this article we have addressed four questions. First, we asked whether or not art is a language. We argued that it is not a language, if by that one means something with the structural and/or semantic features of a natural language. Perhaps all those who call art a language only mean that it is a vehicle of communication. But not all forms of communication are languages, properly so called. Next, we considered whether art could be de ned. After canvassing some of the best known failed attempts at de ning art, we concluded with a discussion of the idea that art might be a cluster concept, but we noted that the jury is still out on that proposal. p. 686

Our third question considered whether or not aesthetic terms are condition-governed. We were especially concerned with whether aesthetic terms are condition-governed in the loose sense that, at a certain point, the enumeration of a number of considerations makes the denial of the application of the term to an artwork unreasonable. We argued that some aesthetic terms are conditioned-governed in this looser sense. Uni ed was our primary example. Finally, we interrogated the nature of critical communication, understood as a species of evaluation. We examined the claim that it was not a matter of reasoning and argument, but of inducing a certain experience in the audience. In contrast to that proposal, we introduced the purpose-driven account of art criticism that (1) involves the exercise of reason and argument, (2) can incorporate the experiential account as a subroutine, and (3) does not invoke the kind of general rules that friends of the experiential account maintain preclude the portrayal of criticism as an a air of reasoning and argument. Further issues, of course, could be addressed under the rubric of “Art and Language,” such as the question of whether critical language is essentially metaphorical. But the four questions examined in this chapter seem to be the most central ones.

Notes 1.

Paul Oskar Kristeller, “The Modern System of the Ats,” in Essays on the History of Aesthetics, ed. Peter Kivy (Rochester, NY: University of Rochester Press, 1992).

2.

Charles Batteux, The Fine Arts Reduced to a Single Principle, trans. James O. Young (Oxford: Oxford University Press, 2015).

3.

Leo Tolstoy, What Is Art?, trans. Almyer Maude (Indianapolis, IN: Hackett, 1996).

4.

R. G. Collingwood, Principles of Art (Oxford: Clarendon Press, 1938).

5.

Clive Bell, Art (London: Chatto and Windus, 1914).

6.

Morris Weitz, “The Role of Theory in Aesthetics,” Journal of Aesthetics and Art Criticism 15 (1956): 27–35.

7.

Maurice Mandelbaum, “Family Resemblances and Generalizations Concerning the Arts,” American Philosophical Quarterly 2 (1965): 219–228.

8.

Arthur Danto, “The Artworld,” Journal of Philosophy 61, no. 19 (October 1964): 571–584.

9.

George Dickie, “Defining Art,” American Philosophical Quarterly 6 (1969): 254. In later versions of his position, Dickie speaks of an institution rather than a society. See George Dickie, Art and the Aesthetic (Ithaca, NY: Cornell University Press, 1974).

10.

One issue that does not seem to be a problem for Dickieʼs Institutional Theory of Art is that it does not seem to present any constraints on artistic innovation and creativity. In that regard, it is a counterexample to Weitzʼs open concept argument.

11.

Arthur Danto, “Art and Meaning,” in Theories of Art Today, ed. Noël Carroll (Madison: University of Wisconsin Press, 2000), 130–140.

12.

Monroe Beardsley, “An Aesthetic Definition of Art,” in What Is Art? ed. Hugh Curtler (New York: Haven, 1983).

13.

Jerrold Levinson, “Defining Art Historically,” British Journal of Aesthetics 19 (1979): 232–250.

14.

Berys Gaut, “Art as Cluster Concept,” Theories of Art Today, ed. Noël Carroll (Madison: University of Wisconsin Press, 2000), 25–44. Although Stephen Davies and Denis Dutton do not invoke the notion of cluster concepts, both have produced criteria comparable to Gautʼs. See Stephen Davies, “Non-Western Art and Artʼs Definition,” and Denis Dutton, “But They Donʼt Have Our Concept of Art,” in Carroll, Theories of Art Today, 199–240.

15.

See Frank Sibley, “Aesthetic Concepts,” Philosophical Review 68, no. 4 (1959): 421–450. The version of the article that I will be referring to here is the one reprinted in George Dickie and Richard J. Sclafani, eds., Aesthetics: A Critical Anthology (New York: St. Martinʼs Press, 1977).

16.

Sibley does agree that, in certain cases, aesthetic terms can be condition-governed by other aesthetic terms.

17.

Sibley, “Aesthetic Concepts,” in Dickie and Sclafani, Aesthetics, 816–837.

18.

Sibley, 816.

19.

See David Hume, “Of the Standard of Taste,” in David Hume: Selected Essays, ed. Stephen Copley and Andrew Edgar (Oxford: Oxford University Press, 1993), 133–154.

20.

Immanuel Kant, Critique of Judgment, trans. Werner S. Pluhar (Indianapolis, IN.: Hackett, 1987), 44, 54.

21.

See especially, Ted Cohen, “Aesthetic/Nonaesthetic and the Concept of Taste: A Critique of Sibleyʼs Position,” in Dickie and Sclafani, Aesthetics, 838–866. The issue of whether this list is obviously expandable in a straightforward manner is germane to the question of whether or not a distinction can really be made out between aesthetic versus nonaesthetic terms.

22.

See Peter Kivy, Speaking of Art (The Hague, Netherlands: Martinus Nijho , 1973), 38–39. I am especially indebted to Prof. Kivy in my discussion of the question of whether aesthetic terms are ungovernable.

23.

Cohen, “Aesthetic/Nonaesthetic.”

24.

Kivy, Speaking of Art, 49.

25.

Sibley, “Aesthetic Concepts,” in Dickie and Sclafani, Aesthetics, 822.

26.

Sibley, 822.

27.

Sibley, 821.

28.

Arnold Isenberg, “Critical Communication,” Philosophical Review 58, no. 4 (July 1949): 330.

29.

NB: the issue in this section di ers from the previous one in that the previous section was concerned with description,

p. 687

whereas this section focuses on evaluation.

p. 688

30.

Monroe Beardsley, Aesthetics: Problems in the Philosophy of Criticism, 2nd ed. (Indianapolis, IN: Hackett, 1981).

31.

Isenberg, “Critical Communication,” 338.

32.

Peter Schjeldahl, “Power Players: The Medici at the Met,” New Yorker, July 12 &19, 2021, 93.

33.

The move from a lack of rules to the arousal of the relevant experience in Isenberg and Bell may have a partial precedent in Humeʼs “Of the Standard of Taste” (in Copley and Edgar, David Hume: Selected Essays, 133–154). Hume claims that there are rules of art that are grounded in human nature that account systematically for our feelings in response to artworks; but that we havenʼt discovered these rules. So we cannot rely on them as a standard of correctness when ascribing certain e ects to the artistic stimulus or when we disagree with others about said e ects. Humeʼs solution to this “paradox” is to delegate the corroboration and/or adjudication of the appropriate experience to a court made up of Ideal Critics whose feelings identify the correct response. So, Hume deals with the problem of the lack of rules by reference to the feelings pronounced by the Ideal Critics; whereas, in contrast, Bell and Isenberg locate the pertinent feelings in the interaction of the audience members and the critics.

34.

See Jonathan Gilmore, “A Functional View of Artistic Evaluation,” Philosophical Studies 155, no. 2 (2011): 289–305; Noël Carroll, “Art Appreciation,” Journal of Aesthetic Education 50, no. 4 (Winter 2016): 1–14; Carroll, “Some Stabs at the Ontology of Dance,” Midwest Studies in Philosophy 44, no. 1 (2019): 70–80; Carroll, “Forget Taste,” Journal of Aesthetic Education (forthcoming). An early functional account of critical judgment can be found in “General Criteria in Aesthetic Theories” by Susan Feagin; this was a doctoral dissertation at the University of Wisconsin–Madison, 1976. Dr. Feaginʼs theory, however, di ers from the one recommended above in that it identifies the function of all artworks as the production of the same experience, whereas my view countenances a plurality of di erent constitutive purposes.

35.

This is not to maintain that the purpose-driven account of critical communication cannot incorporate Isenbergian experiential promptings as subroutines that establish certain argumentative premises. But at the same time, it does not appear that Isenbergian approaches can incorporate the purpose-driven method. On the question of whether there are further forms of critical communication or not, this chapter is, at this juncture, uncommitted.

The Oxford Handbook of Applied Philosophy of Language Ernie Lepore (ed.), Luvell Anderson (ed.) https://doi.org/10.1093/oxfordhb/9780192844118.001.0001 Published: 2024

Online ISBN: 9780191926853

Search in this book

END MATTER

Index  Published: May 2024

Subject: Philosophy Series: Oxford Handbooks Collection: Oxford Handbooks Online

Print ISBN: 9780192844118

p. 689

Index For the bene t of digital users, indexed terms that span two pages (e.g., 52–53) may, on occasion, appear on only one of those pages. Tables and gures are indicated by t and f following the page number @-ing 110–12 #MeToo 116–18 a priori justi cation and empirical presuppositions 17–18 AI 358 content  See content systems 573, 574–75, 576, 577, 579, 580, 581, 582–83, 587, 589–90 AAVE.  See African American English abortion 349–50, 359 rights 356, 361 acceptance in communication 46–47, 82, 94, 95, 109, 227, 228, 230–31, 296, 306, 395, 446, 594–95 accuracy-attestation 626–27 advertising 375–76, 558, 560–61, 563 aesthetic 2, 561–63, 676–77 defects 491–92 emotion 672 properties 680–81 terms 671, 679, 680, 681, 682, 683, 686 value 107, 178–79, 678, 679 a ective attitudes 135–36, 278–79 African American English 309 See also Black English identity 127, 149–50, 157, 438, 540, 542 See also Black racial identity privacy 349, 356, 360–61 agency communicative, of protestors 451, 457–59 constrained 643 discursive 304–5 impaired 220 political 455, 457–58 alien content users 576–77 See also content metasemantics 575, 583–85, 586–87, 589–90 alienation 63, 183 Alignment Problem 589 alt right 557 ambiguity 177, 184–85, 188, 282, 293, 323–24, 350, 444–45, 672 Americans with Disabilities Act 417 American Sign Language 415–16, 419–20, 421–23, 425, 426, 427 anagram 663 anaphora 271, 285, 632 anchoring 45–46 Anderson, Luvell 456, 457–59 androcentrism 320, 324, 339

androids 282, 284 Andromeda 65–66 angels 18, 658 anger 186, 195, 196, 205 expression of 142–43 anonymity 196–97, 351–52, 561–63 anti-ableist 338 anti-aesthetic 676–77 anti-black 549 anti-feminist 379 anti-racist 460 anti-semite 138 anti-sexist 377 anti-surveillance 355 Antipholus 659–61 Aphrodite 556 Apelles 556 Applied Philosophy of Language 1–2 p. 690

appropriation

165, 257, 301–2, 305, 306–7, 655–56, 663–64

Aristotelian kinds 31 Aristotle 403, 585, 598–99 armchair analysis 355 art as language 670, 671 de ning of 671–80 family resemblance method of identi cation 672–73, 674 articulation 163–66 See also slurs Arti cial Intelligence.  See AI artworld 675–76 Asimov’s First Law of Robotics 589 ASL.  See American Sign Language assent 83, 90, 95 assertion 59, 60, 62, 66, 68, 70–71, 79, 80, 81, 82–83, 84, 86–88, 89, 90–91, 93, 94, 95–96, 99, 106, 113– 14, 134, 214, 222–23, 224–25, 235, 353, 358–59, 390, 421–22, 435, 490, 615, 620–21, 625 asymmetry 66, 67, 73–74, 424–26, 427–28 at-issue/not-at-issue distinction 541–42, 543 See also content Athenaeus 556 Athens 32, 394 attention-direction 617 attestation 626–27 Atwater, Lee 439, 560, 561 audience.  See also general audience responsibility 462–63 audism 419–20, 421, 424–25, 426, 428 Augustinus, Aurelius 403 Austin, John Langshaw 7, 9, 14–15, 67, 68, 105, 211–12, 214–16, 217, 218, 219, 220, 221–22, 223–25, 226, 227, 228, 229–30, 232–33, 234, 246, 247, 248–49, 251, 257, 299, 356, 397, 435 authority 83, 105, 106, 107, 113–14, 116, 186–87, 219, 228–29, 245, 249–50, 253–54, 255, 257, 258–60, 303, 305–6, 310, 407, 435, 655–56, 683

autonomy 358, 360–61 Baldwin, James 434 Barthes, Roland 560 Batteux, Abbé Charles 671, 672 behavior aggressive 175–76, 187–88 See also passive aggression attacking 184–85 discriminatory 214 objectionable 184–86 of structures 600–1 sexual 222, 235 socially inculcated 9 verbal 9, 30, 119–20, 196, 200, 206 beliefs and desires 6 false 38, 64, 232–33, 256, 418, 607–8, 634, 637 ideological 642 mutual 82–83, 84, 85, 86, 87, 91–92, 93–94, 98, 99 See also common ground of AI 576–77, 587 relation 473–74 state 83, 85, 86, 87, 472–73, 474 true 24–25 Bell, Clive 674, 685 bene t 15, 44, 419–20 Bernoulli, Daniel 600–1 Biden, Joseph R. Jr. 224–25, 613 bidirectional dyad 407 bigotry 131, 132–33, 136–37, 213 bilingual 415–16 binary gender 367–68 bioethics 422, 423 bisexual 372–73 Bishop Berkeley 396 Black English.  See African American English Black, Joseph 605 Black Lives Matter 416–17, 454–55, 458–59, 460 blackening 666–67 blackmail 186 Blackness and criminality 182, 438, 440–41, 443, 445, 446 See also thug blocking in communication 87, 153, 217, 250, 253–54, 434–35, 459–60, 461, 462, 607–8 See also communication bondage-and-domination 234–35 bonding social 453, 618 boundaries of social groups 104, 107, 646 p. 691

Bourdieu, Pierre

435–36

Brandom, Robert 113–14 breakdowns in conversation 81, 87–89, 90–93, 94, 98, 99

bridging 550–51, 588 British National Corpus 201 burden of proof 117, 676, 679 Bush, George H. W. 438, 546, 548–49 Bush, George W. 183, 438, 456 Butler, Judith 257, 258, 299, 300, 643 Cabanel, Alexandre 554–56 Cambridge, University of 532 Camp, Elisabeth 539 cancelability.  See deniability cannibalism 663, 665–67 capitalism 60–62, 75 capitalist ideology 634, 641 Cappelen, Herman 11 care networks 381 caricature 34, 446 Carnap, Rudolf 9, 11, 29 Carnot, Sadi 604 CART.  See Communication Access Realtime Translation center of attention 491 centered worlds 578 challenging interpretations 179–80, 184–85, 187, 188 speech act of assertion 80, 84, 86, 87–88, 89, 90, 91, 95, 98 visibility/invisibility 324 charitable 36, 38, 39–40, 41, 43–44, 46 charity 7, 24, 25, 37, 38, 39–40, 46–47 ChatGPT 576, 583 Chicago, University of 351 Chomsky, Noam 9, 12, 194 Christian 183, 391–439, 524, 537–38, 552–53, 685 Christian identity movement 557 See also alt right chromosomes 266, 281 Cima, Giovanna Battista 524, 529, 530–31, 532, 537, 542 cis 297, 298, 306, 309, 338 cisgender 295, 359, 379, 382 cisnormative 366, 373 CMC.  See computer-mediated communication co-designating 138 co-directional 505 co-opt 557–58 co-referring expressions 436 See also neutral counterparts of slurs Cobb, Andrew Jackson 358 code of honor 228 software 614–15, 620 words 437–39, 440, 523–24, 538–39, 541–42, 543, 544, 548 See also dogwhistles; See also taboo words coercion 212, 259, 353 coextension 150, 151

cognates 194, 635 cognition 232–33, 275, 286–87, 373 cognitivist 636–42, 644, 646, 648 cognoscenti 42 coherence relations 502, 582 cohesiveness 452 colonization 440, 491, 663–64 command 105 See also speech acts commissive 230 See also speech acts common ground 79, 82, 87, 92, 94, 97–98, 152, 153, 578, 586, 617–18, 643 See also belief mutual; See also context common interest 86–87 common knowledge 86–87, 186–87, 424–25 commonness 659–60 communal 434, 452, 453, 674 communication acts of 250, 451, 452, 454, 455–56, 457–58, 459, 461–62, 531, 539, 541–42, 543, 544, 614–15, 627 coded  See code words computer-mediated 192, 196–97 conditions of 250, 256 interference in 250, 255 of meaning 524–26, 541 system of 615 Communication Access Realtime Translation 424, 425–26 community -expressive 453 of practice 229 See also speech community -speci c 106, 107–8, 109, 111, 112, 113, 114–15, 116 p. 692

competence linguistic 257, 425 compositionality 631–32 and map semantics 483–84, 507 of slurs 126, 127, 131, 132–33 conceivability.  See modal conceivability conditional 23, 130, 140–41, 143, 233, 268, 366–67, 584 conditionals 134 Conegliano, Cima da.  See Cima connectives 130, 131, 582 connotation 19–20, 333, 359, 368, 434, 437, 439–40, 441, 550, 560 conquest 491, 640 consciousness false 639, 640, 647 group 636, 637, 640 in Marx 61, 64 consent 37, 38, 61–62, 106, 230, 233, 254, 298, 358 consequences of passive aggression 176–77, 178–79, 186, 187 See also passive aggression conservative values 204–5, 206, 212, 213, 349–51, 356, 359, 457, 458–59 constatives 215, 216, 222–25 See also speech acts constellation 640

constructionist 63 content AI 574–76, 582, 583, 587 alien 573, 574–80, 581, 584–86, 587, 589–90 at-issue 158, 541–42 attributive 523, 529, 530–31, 536, 539, 543–44 cartographic 489–91, 493, 503, 504 centered 578–79 communicated 135 essentialism 23 expressivist 578, 579 incomprehensible 575–76, 578 map 470, 471–72, 473–75, 476–77, 479–81, 484, 485–86, 489–90, 500–1 not-at-issue 541–42, 543 o ensive 150–52, 153, 154, 155–56 propositional 138, 632 semantic 104–5, 108, 110, 113–14, 115, 126, 132–33, 151, 270, 279–80, 367, 369–70 singular 503, 523, 536, 537, 539, 540 structured 470, 474–75, 479, 481, 485–86 See also structured propositions contexts.  See also common ground apolitical 558 attitude 582 change 79, 82 defective 97–98 dependent 17, 202, 257 dominant 304 language 321 law enforcement 416–17 length-constrained 338 non-defective 94, 97 of utterance 84, 94, 97–98 contextualism 295–96, 632 Contrapoints.  See Wynn conventional implicature.  See implicatures conventional linguistic properties 127, 129, 130, 131 conventionally encoded meaning 154, 155–56, 161 conventions 73, 217–18, 219–20, 221, 226, 227, 234, 254–55, 266, 277, 286, 322–23, 415–16, 480–81, 491 community-level 278, 279–80 linguistic 127, 162, 266–67, 276–77, 583, 670 local 502, 524 social 215–16, 426 conversation breakdown in 81, 87–89, 90–93, 94, 98, 99 dynamics of 81, 86, 94 maxims of 334–35, 631–32 See also Grice participants of 83, 89–90, 92, 98 record 87, 89, 91–92 scoreboard 2, 82–83, 94–98 standing of participants 80, 81, 83, 88, 89–90, 94–96, 97–98, 99

conversational exercitives 260 See also speech acts conversational standing 80, 81, 83, 88, 94, 95, 96, 97, 98 cooperative principle 86, 87, 93 See also conversation, maxims of; See also Grice cosmopolitan gure 605 counterprotesters 458–59 p. 693

covert dogwhistles.

 See dogwhistles

creed 229 Crenshaw, Kimberlé 645–46 criminal justice 199 system 434–35, 459–60 critical communication 671, 684, 685, 686 critical de nitional facilitation 349 critical race theory 540, 542 critical theory 2, 641 critique of ideology 635, 639 political 108 public 454 cross-cultural di erences 197–98 cross-linguistic 268, 270–71, 272, 276 CRT.  See critical race theory Davidson, Donald 576, 582–83 Davidsonian charity 35–36 deaf/Deaf distinction 418–19 de-coding 514 de-contextualize 402 de-gendering 334–35 de-identi cation 358–59 decolonial feminist practices playfulness 303–4 decriminalization 356 defective contexts  See contexts de nitions analyses of 359–60 disputes over 354–55 facilitation 349, 361 prescription 360–61 Delegate Zero 72–73 demonstratives Chinese/Mandarin 164, 165 referent xing 582, 616 deniability in discourse 46, 151–52, 155, 542, 544, 554, 558, 683 denial direct 151, 155 explicit 154, 155 norm of 374–75 of conventional implicatures 155 denialism 44–45

denigrate 136 denotation and maps 495–96, 508, 512–13, 537–38, 577 derogation 126–27, 129–31, 132–33, 140–41, 144, 462 derogatory language/speech 1–2, 435, 436, 437 See also slurs desires 185–86, 222, 560, 664–65 and beliefs of AI 576–77, 587 Diagnostic and Statistical Manual for Mental Disorders 175 Dickie, George Institution Theory of Art 674–76 Di ee, Connie 415–16 Diogenes Laertius 402 disabilities.  See Americans with Disabilities Act disability status 378 disagreement conversation 11, 15–16, 70 online 109 disclaimers online 613–14, 616–17 disclose or deceive dilemma 368–69, 370 discourse about privacy 349, 351–54, 355–58, 361 coherence 582 dynamics of sexual assault 229–30 goals 82–83, 88, 93, 94, 97, 98 levels 299–300 -level refusals 442 nonscienti c, problems in 10 online 205 relations 516 sexual 254 discourse/Discourse distinction 442 discriminating 218–19, 682 discrimination against deaf people 419–20 See also audism gender 365, 382–83, 646 non-binary 378 racial 248–49, 633, 645–46, 648 See also slurs; See also thug discursive injustice 455, 457–58 discursive practices 112, 119 disempowered group 211 disenfranchisement 218–19, 640 p. 694

disinformation

229, 624

dissent 451–52, 453–55, 462 discrediting 115 expressing/voicing 453, 462 political speech 453–55, 457–58 dogwhistles 182, 183, 437–38, 543, 546–50, 551, 554, 557, 558, 560, 561, 563 dominant

de nition of "woman" 295–96 de nitions 296, 309–10 ideology 13, 23 See also ideology social understanding of rape 37 world 303–4, 307 Dotson, Kristie 80 double entendre 182 and norm- outing 188 doxa 423 doxastic eligibility 26–27 DSM.  See Diagnostic and Statistical Manual for Mental Disorders Du Bois, W.E.B. 160 Duchamp, Marcel 672, 678–79 Dukakis, Michael 438, 546, 548–49, 560 Dworkin, Andrea 212, 213, 245–47 Dworkin, Ronald 245–46 dyslexia 338 dysphoria 305 e ects of language.  See perlocutionary acts See extralinguistic e ects of language use Einstein, Albert 45, 136, 596 Eisenhower, Dwight D. 670 emojis 104–5, 119 entrenchment of binary gender 367–68 epicenes 334 semi- 334 epistemic injustice 80, 98, 421–22, 459–60 epistemic reasons 80, 83, 84, 86, 91 epistemological 581 epithets 204–5, 662–63, 667 See also slurs equal standing in conversation 89–94, 97, 98, 99 equality 349–51, 356, 361, 381 and rights of women 213, 356 critical agendas of 355 gender 297, 356–57 in communication 99 marriage 356–57 structural lack of 638 Erlenbusch-Anderson, Verena 436 Euler, Leonhard 516, 600–1 evils of pornography 235 See also harms of pornography See also power of pornography evolution linguistic 8 of grammar 118–19 of iconography 552–53 of taboo lexicons 204–5 exercitives 211, 219, 221–23, 224–25, 234, 248–49, 251, 258–59, 260, 305–6 See also speech acts existential generalization

and maps 483, 485–86 expectations about being silenced 88 about gender 256, 302, 304 about sex 257 about women 222, 235 development in infants 406 subverting of 302 thwarting of 220 expositives 225 See also speech acts expressives vs. reports 117 expressivism about slurs 126, 137–43 and AI content mapping 578–79 identity and non-identity theories 126, 133–35, 137, 143–44 extension null/empty of slurs 126–28, 131–33, 137, 139, 140, 143–44, 162 Extensional Mereology 468–69 externalism combinatorial 126 metasemantic 581–82, 583–84 extralinguistic e ects of language use 246, 247, 249, 250 f-bombs 142, 143 face-to-face contexts 200 See also contexts encounters 196–97 p. 695

Facebook

108, 112, 113, 116, 205

family resemblance method 673, 674 far-right 550–51 felicity conditions.  See speech acts feminine-gendered morphology 301, 303 feminism and trans rights 276–77 anti-pornography 212, 245, 246–47, 248–49, 257 post second wave 306–7 pro-pornography 212, 248–49, 255 purposes of 43 feminist philosophy 235 of language 80, 293, 355, 631–32 femininum 332–33 feminization 302, 379–80 Floyd, George 451, 558 force communicative 456, 461, 462 expressive 119 illocutionary 248, 249, 260, 451, 456–57, 614, 615–16, 618, 622, 627 See also illocutionary acts; See also illocutionary acts

normative 461 pejorative 441–42 pragmatic 109, 113–14, 118, 119–20 Foucualt, Michel 655–56 free speech 80 rights 213, 245–46, 249, 252, 462 zones 456 freedom from 245–46, 351–52, 360–61 and privacy 350 sexual 350–51, 356–57 to 245–46, 530 Frege, Gottlob 28, 29, 46, 161, 577, 635–36 Foundations of Arithmetic 27 Frege-Geach problem 137, 140–43 fruitful concepts 10, 46 functionalism about tabooed speech 194–95 gay 130, 212, 234–35, 548 Furie, Matt 557 Galileo 598 Gandhi, Mahatma 44, 45, 47 Gee, James 442 genealogy 632 gender-a

rmative care 305

gender-diverse communities 298–99, 302 gender-loaded grammars 338–39 gender-neutral language 314, 365, 369, 379, 381, 382–83 options for revising gender-speci c structures 313, 376–77 pronouns 272, 285, 286, 293, 296–97, 302, 320, 321, 364, 376–77 gender-neutrality and family leave 364–65 project 320–21, 324 gender-neutralizing language 297, 313–14 gender-nonconforming 298–99 gender-speci c framing 364–65 language 364–65, 366, 367, 370–71, 375, 377, 382–83 gender 267–72 See also binary gender and pronouns 265–67, 296–97, 364, 368–69, 372, 374 and indexicality 299–300 categories 266, 338, 372 discrimination 365, 382–83, 646 grammatical 266–67, 273–76 natural accounts of 266, 270–71, 273, 274 nonbinary 265, 267, 271, 272, 276–77, 284, 285–86, 296–97, 301, 303, 306–7, 308, 310, 325, 366–67, 368, 370–71, 373, 375–78 personal vs grammatical 266–67, 270–71, 272–73, 274–77, 278, 280–81, 283, 284, 286

transgender 265, 281, 296, 298–99, 369 See also trans vs. sex distinction 266 gendered identities 299–300, 304–5 linguistic features 299–300, 302, 308 meaning 299–313 pronouns, semantics of 267, 277, 283–84, 285 gendered morphology in German 326–27 in Italian 328 in Spanish 329–30, 330t p. 696

genderqueer

285–86, 296–97, 301, 303, 306–7, 320–21, 323, 324–25, 330, 331, 332, 333, 335, 336, 337–

38, 339–41, 342 general audience 2, 182, 391, 437, 438 generics 13, 294, 297, 313, 340–41, 364, 631–32, 638, 645, 648 Glasgow, University of 605 Gravity Constraint 505–6 Green, Marjorie Taylor 561–63 Grice, H. Paul communicative intentions 582 conventional implicature 154 See also implicatures conversational maxims 334–35, 631–32 See also conversations Cooperative Principle 86, 93 Gricean principles 87, 638 Guernica 391 Guthrie, Samantha 613, 622–23, 625 Hale, Matthew 37, 45 harm of failed sexual negotiation 298 of gendered language 294, 296, 298, 299, 311, 314 of hermeneutic injustice 301 of misgendering 294, 296–97, 364, 369, 375–76 of police violence 457 See also Rodney King of pornography on women 213, 235–36, 245–46 of silencing 99 of stereotypes 370 prevention 462, 589 state and institutional 371 to dead people 423–24, 426 Harris, Kamala 80 Haslanger, Sally 632, 636, 641, 642 Helsinki, University of Technology 409 hermeneutic injustice 301 hermeneutical gaps 301, 302, 303 hermeneutical resources 301–2 and schemes 301 hermeneutically challenging 117 Hesperus 18, 38 heteronormativity 369, 372–73, 382–83

heterosexuality 212, 257, 372–73, 376–77, 659–60 heuristics 46–47 hipster 162, 561–63 homophobia 368–69 in the work environment 369 homosexuality 136, 259–60 homosocial 659–60 Hornsby, Jennifer 157, 211–12, 225, 230–33, 234, 235–36, 246, 249, 252, 253–54, 255, 436, 456, 542 Horton, Willie 438, 546, 548–49, 558, 561 hostile attacks and passive aggression 176–77, 179–80, 181–82, 183, 184–85, 188 See also passive aggression discourse 446–47 semantic takeover 7–8 human-denoting nouns 270–71 humanist genealogy 665–66 humanitarian 391 Hume, David 5, 6, 681 See also is-ought problem humiliation 33 Hurston, Zora Neale 443 hybrid visual/linguistic dogwhistles 561–63 See also code words; See also dogwhistles hydrodynamic model 603 See also Alexander Volta; See also Thomas Kuhn hylomorphic perspective 31 hyperreal numbers 29 IAT.  See Implicit Association Test iconographic interpretation.  See interpretation iconography 523, 528, 535–36, 544, 551–53 p. 697

identity and AI mentasemantics 577 and Discourse 442, 499, 510–11 and expressivism 125, 126, 137–38, 141, 143–44 See also Frege-Geach problem and map semantics 485 community and speech acts 107 gender 126–27, 229, 295–96, 301, 310, 324, 367–68, 369, 373, 374, 375, 376–77, 378 See also gender; See also gender-nonconforming; See also gendered identities; See also genderqueer in Shakespeare 655, 660–62, 664–65 online 109–10, 114–15, 196–97, 625 personal 19, 20, 350, 661–62 political 621 revealing through code words 548 semantic, over time 7, 19–20 social recognition of 302 structural prejudice of 301 theories and slurs 125–26, 140–41, 142 ideology 2, 632 and political meaning 491 cognitivist and non-cognitivism theories of 636–42 de ning 635–36 distorting e ects of 7 dominant 13, 23

function of 633 Haslanger on 642–45 See also Haslanger holism about 648–49 white supremacist 537, 539 idiolect 272, 277, 279–80 idioms 12–13, 641, 644 Ikea 561–63 Iliad 666 illocutionary acts.  See also speech acts and Austin’s characterization of 216–17, 218, 223–25 and conventions 217, 218, 234 covert 303 institutional vs. non-institutional 248, 252–53 of disenfranchising 218–19 of gender ascriptions 303 of ranking inferiority 218–19 of refusal 254 of retweeting 615, 618, 623 of sentencing 249 See also exercitives; See also verdictives of silencing 456 See also silencing of subordination 211, 218, 219, 228, 234 of warning 248 illocutionary disablement 249, 250, 252, 254, 255 See also silencing illocutionary ipping 457–58, 459 imaginary, the heterosexual 664–65 homosocial 664–65 mythical 655 imagination and indigenous innovation 403 and reduced capacity for 444–45 impermissibility in conversation 260, 305–6 implicatures 129, 215, 443, 632 audience- 334–35 conventional 150–55, 157, 158, 161, 166 conversational 129, 132 vs. entailments 23 Implicit Association Test 275 in-group/community 136, 305 incommensurability Kuhnian 575, 594–95, 596, 597, 599, 600, 601, 603, 606, 607–8 See also Kuhn incomplete 197, 513, 647 inculcate behavior dispositions 9 ideologies 632 perspectives 637 inde nite 327, 499, 658 inde nitely 529

indeterminacy of linguistic judgments 277 of map content 494 referential 33 semantic 42, 46, 108, 277 index temporal  See map semantics indexicality and gender associations 299–300 See also gender indexicals Putman 16–17 singular content 503–4 in ammatory expressions.  See slurs Instagram 561–63, 562f institutional gender marking 374, 377 Institutional Theory of Art 674–76 institutional subordination 219–20 insult 177–78, 193, 201, 202–3, 204–5, 305–6, 439 intentions communicative 582 intercourse 37, 193–94, 236 internalized grammars 278 p. 698

interpretation and charitability 38, 39 and dogwhistles 438 disabling 459 functions in map semantics 469, 470, 475, 477–78, 479–81, 485, 509, 510, 512 See also projection semantics; See also tagging semantics iconographic 523–28, 529, 536, 537, 538–40, 543–44 See also semantics of iconographic interpretation innocuous 177–78, 179–81, 182–83, 184–85, 188 intensional 74 masculine 339–40 of map content 489, 493 of metaphor 275 Interpretation Act 323–24 interpreters.  See signed language interpreters intersectionality de nition of 633 Is-ought Problem 5–8 See also David Hume James, William 26–27, 435 Kant, Immanuel 28, 681 Khoo, Justin 437–38, 538–39, 541–42, 546 King, Rodney 457–58, 471–72 King of France 10, 661 Kripke, Saul 5, 17–18, 24, 581–82, 583–84, 635–36 Kripke-English 10 reference- xing 18 Kuhn, Thomas 594, 595, 596, 598–99, 601, 602

Kukla, Quill 297–98, 457–58 Lampedusa, Giuseppe di 5, 21 Lancaster Corpus of Abuse 201 Langton, Rae 2, 79, 80, 96, 211–12, 214, 218–19, 221–22, 225, 228–29, 230–33, 234–36, 246, 247, 248–49, 252–54, 255, 256, 257, 258–60, 456, 631–32 language performative function of 214 savage 663–64 language modality choices and inequities of 418–19, 427–28 defaults and bias 416, 417, 418, 425–26, 428 expression 418 rst acquired 425 signed 421–22 spoken 426 Leipzig, University of 332–33 lesbian 212, 366, 369, 372–73, 375–76 Lewis, David 2, 82, 308–9 centered-worlds 556 meaning eligibility 24 on naturalness 26 and reference magnetism 46 scoreboards 82, 96–97, 99 lexicon taboo words and evolution of 203–4 LGBTQ 352, 356–57 Lincoln Project 540, 542 linguistic encoding of derogation 129–30 See also slurs linguistic resistance 443–45 linguistic manifestations 435 locutionary acts 232, 247, 613 and Austin’s characterization of 216–17, 223–24, 246 and illocutionary disablement 250 attention-directing 618 retweeting 617–18 MacKinnon, Catherine 80, 212–13, 214, 222, 225, 245–47, 248, 249 Madagascar 19–20 male dominance and female submissiveness 245 manipulation and dogwhistles 546 See also dogwhistles and the politics of language 356 locutionary 459 See also locutionary acts map contents.  See content marital rape 37, 38 Marx, Karl 61, 63–64, 632, 635, 639, 641 Massachusetts, University of 198 material conditions vs. institution conditions 220

maxims, conversational 631–32 See also Cooperative Principle of manner 334–35 McGary, Howard 444–45 McGowan, Mary Kate 253–54, 255, 256, 260, 293, 294, 305–6 p. 699

meaning and locutionary acts 252–54 See also locutionary acts change and the di

culty of 33

embodied 20–22 Fregean colouring 635–36 Fregean sense (Sinn) 635–36 map vs. picture 499 of illocutionary acts 230 See also illocutionary acts meaning-infused words 20, 22 mediation 9 Medici, Marie de 559–60 Memphis, University of 415–16 Mendelberg, Tali 437–38, 541, 542, 548–50 Menninger, Colonel William 175, 184, 185–86 mens rea 254–55 mereology 468, 577, 648, 649 meta-tag groups 112 metaphor auditory 426 and interpretation 198 and slurs 437, 444 See also slurs; See also thug metasemantics.  See alien metasemantics Minneapolis 451, 558 misgendering 265, 278–79, 280–81, 282, 284, 286, 296–97, 298, 323, 324, 339, 369, 373, 375–76 See also harm of misgendering misinterpret 196, 254 misinterpretation of women’s speech 254 See also silencing mismatching gendered parts of language 313 misogynistic pornography 234, 235 Mitchell, W. J. T. 561 modal conceivability 137, 139–40 Muhammad, Khalil Gibran 440 muting derogatory content and negation 130 mutual belief.  See belief See also common ground mutual recognition 406, 407 MySpace 201 names and AI-metasemantics 582, 585 and gender 328, 339–40, 373–74, 375 and Indonesian tombois,C13P34 and maps 467–68, 495, 499, 501–2, 503 in Adamic ur-language 11 signed and cognitive load 425

naming playful practice in trans community 305 National Football League 443 National socialism 634, 649 natural kind term 16–17, 139 neo-classical growth theory 395 neoliberal language and double-speak 69 neural nets 589–90 weightings 583 neutral counterparts of slurs 125, 126–27, 128, 129–32, 136, 139, 140, 150–51, 158, 162, 278, 437 See also slurs Newton, Isaac 28, 38–39, 40, 45, 596, 598 NFL.  See National Football League non-cognitivism 141 non-conventional means and illocutionary acts 218 See also illocutionary acts; See also perlocutionary acts; See also speech acts non-echoic uses of slurs 130, 131, 132 See also slurs non-native language 205, 427 language modality  See language modality signers 426, 427 speaker 158, 195–96, 197, 198, 202 non-semantic di erences between pejorative and neutral language 140 word associations 339–40 nonbinary.  See gender noncognitive motives in Marx and false consciousness 639 p. 700

norms g-type 306 linguistic, of trans communities 310 of a person wearing people clothes 285–86 of conduct 627 of distribution 619–20 of endorsement 620–21 of etiquette 178–79, 185 of maps 505 of politeness 180 of professional silence 353 of retweeting 619, 625 of sexual behavior 222 s-type 306 not-at-issue content.  See at-issue/not-at-issue distinction nudging and norm development online 625, 626–27 tactic of visual artists 551 See also dogwhistles Obama, Barack 550, 560–61 obligations

and speech acts 221, 233, 299 See also speech acts of con dentiality 353 See also privacy prima facie and public protest 452, 454 o ense associational 279 generating 152, 158, 159–60, 196 not intending 178 potential 149–50, 154, 159 taking 129–30, 558 to cause 129–30, 150 o ensive content.  See content o ensive speech 196, 197, 199, 204–5 ontological mediation of social relations 65 oppression and ideology 162, 633, 640, 642, 643, 645, 647–48 and intersectionality 633, 642 and social organization 69, 75, 646 gender 307, 364, 368, 382–83, 648 mereology of 649 naming systems of 300 nature of 378 of women 230, 245 psychological 308 racial 633, 648 resisting and reducing 443–44 Orwell, George 5 ostension de ning by 17 xing referents by 18 learning by 575 online 111–12 Othello 667 Panofsky, Erwin 528, 552–53 parenthood and caregiving 381 passive aggression 175–77, 178–79, 180–81, 182, 183, 184–85, 186–88 consequences of 176–77, 178–79, 186, 187 See also passive aggression pejoratives content 150, 151–52, 154, 155 e ect 158, 166, 278 expression 131–32, 441 force 441–42 and Freges Puzzle 138 tone  See slurs people of color 360, 454–55, 457–58, 460 Pepe the Frog 536, 537, 540 performativity.  See speech acts See also Austin perlocutionary acts 216–17, 228–29 See also speech acts

and Austin’s characterization of 216–17 and silencing 96, 459–62 e ects of 218–19, 229, 246, 247, 257, 258, 259, 459, 461–62 perspectives and slurs 125, 156–57, 159–60, 161–62 perspective projection 506, 507 philosophy feminist 235, 293, 355, 631–32 philosophy of language and arti cial intelligence 574 and conceptual engineering 619 and metasemantics 574 applied 1–2, 265–66, 389, 394 as world making social practice 69 externalist tradition in 631–32 feminist 80, 293, 355, 631–32 social and political 436, 632, 633, 634 phonetics 12, 119, 149–50, 159, 299–300, 303, 309, 313, 325, 336–37, 424 Phryne 556 Plato 390, 391, 399, 402, 408, 671 p. 701

politeness.

 See norms of

politics of language 71, 355–56 of privacy discourse 358 of refusal 443 pornography and its authority 256–60 as speech acts 211, 212–29, 234–37, 246, 247–50, 251–55 portmanteaus and gender identity 304–5 possible worlds and alien content 578 See also alien metasemantics; See also content and map contents 472, 502–3, 506 semantics 73–74 power of the a ective stance 136–37 of Christ 183, 438 of hashtags 116 of language choice 415–16 languages expressive 119 of language modality  See language modality of law enforcement 416–17 of linguistic manifestations 435 See also Pierre Bourdieu of and to perform speech acts 228, 229 of pornography and pornographers 214, 222, 228–29, 249, 258, 259 of protest and assembly 452, 453, 457–59 social 97, 99, 107, 116, 259–60 See also conversational standing technologically enabled social 116, 118 to o end 158, 204 See also o ense; See also pejoratives; See also slurs

powerlessness 79, 181, 186, 188, 219–20, 250 pragmatic 2, 99, 104–5, 106, 107–8, 109, 111–12, 113–15, 116, 118, 119–20, 125, 143, 193–95, 202, 260, 278, 281, 296–97, 298–99, 513–14, 515, 632, 680, 683 pragmatically 108, 109, 166, 297–98 pragmatics 105–8, 113–14, 118, 195, 196–97, 297–98, 310, 311–12, 454, 490–91 predication 130, 155 predicativism 151 pregnancy and gender 359, 379, 380 presemantic words and lexical items 19–20, 22 presupposition 17–18, 45, 97, 129, 132, 152–53, 154, 155–56, 157, 268, 269, 270, 454–55, 641, 642 -carrying 268 projection 269, 270 triggers 152 privacy discourse 349, 351, 352, 353–54, 355, 356–58, 361 prescriptivism 353 projection semantics for maps 497, 506, 507–8, 509, 511, 514 See also map semantics projections, maps Gall-Peters 491–92, 492f and map semantics 493 Mercator 467, 470, 485–86, 491–92, 492f planar 494, 502–3, 505, 507 Robinson 470, 485–86 promise speech act 106, 220–21 See also speech acts pronouns.  See also gender; See also misgendering feminine 327, 339–40 gender-neutral 272, 296–97, 302, 320, 321, 364, 376–77 gendered 265–66, 267, 269, 270–71, 276–80, 281, 282–84, 285, 286, 297–98, 302, 314 H-series 267, 280–81, 284, 286 masculine 302, 337, 339–40 plural 327, 328, 329, 336 preferred 265, 267, 280–81, 282, 286, 297 S-series 267, 280–81, 284, 285 singular 271, 272, 286, 322, 324, 327 Putnam, Hilary 16–18, 24, 26 QAnon 561–63 quanti cation and presupposition projection 269 event 141 Quine, Willard Von Orman 25, 45, 576, 648 race-speci c interpretations 435, 439 pronouns 297 word use 436, 437 p. 702

racial resentment.

 See resentment

racialized groups 439 terms 435, 438, 439, 440, 443–44, 445, 447 racialization of terrorist 435, 436, 447 of thug 436, 441, 447 racism 140, 183, 354–55, 356, 547, 550, 557, 634, 636, 637, 644, 646 racist attitudes 133–34, 542, 547 ideology 633, 637, 642 See also ideology technē 644 radical interpretation 575 See also Davidson translation 575, 576–77 See also Quine Radnitzky, Emmanuel.  See Ray rape 37, 38, 41–42, 43, 45–46, 211, 212, 228, 236, 254–55, 256, 548–49, 664 Ray, Man 526, 528, 535–36, 537–38 Reaganomics 646 realism vs. nominalism 658 vs. social construction 658 reappropriation 257, 435, 446–47 recognition @-ing 116 and conversational standing 96 and silencing 454, 459–60 facial 358–59, 360 mutual 406, 407 political 443 use of gendered words 300–3 reference casual theory of 26 -determining 25 electromagnetism 26–27, 41 in ction 662 - xing 17–18 See also ostension Fregean 161 See also Frege; See also sense indexical 503–4 knowledge-maximization account of 26 magnetism 26, 27, 41, 46 -preserving 18 refusals sexual 253–54, 255 See also illocutionary acts of refusal rejection expressing dissent 453 of asserted content 81, 83, 87, 99 of racial caricature 446–47 Republican Party 229, 561–63 See also Lee Atwater resentment

racial 437–38, 439–40, 549–50, 637 retweeting 113–15, 613–14, 616–18, 620–21, 622, 624, 625, 626–27 See also tweeting) semantics and pragmatics of 114 Richard, Mark 23, 126, 133 Rubens, Peter Paul 559–60 Russell, Bertrand 44, 391, 474–75 Russell, Gillian 97 Saint Augustine.  See urelius Augustinus Saint Zenobius 531–32, 535–36 salience, raising to ones identity 301 prejudicial beliefs 279 race-speci c interpretive options 435, 439–40 racial attitudes 549–50 Sapho 403 Saul, Jennifer 182, 183, 225–26, 295, 296, 297–98, 311–12, 322, 437–38, 538–39, 540, 541, 542–43, 546, 547, 550 scoreboard conversational 2, 82–83, 94–98 See also conversation semantic content.  See content semantics of iconographic interpretation 528–36 of maps 469, 493, 496, 498, 499–500, 513, 515–16 See also map semantics of natural kind terms 16–17 See also Hilary Putnam of pronouns 267, 272, 277, 283–84, 285 of slurs 126, 132, 140 p. 703

semantic possibilities for amerliorists debunking accounts 42–45, 46 electromagnetism, CC2S11 46 eligibility 23–24, 45–46 factual entanglement 45 fruitful concepts 46 historisity and embodiment 20–22, 45 presupposition failure 45 semantic indeterminacy 46 understandable misstatements 46 semiotics 19–20, 299–300, 490–91, 515 sex discrimination 213, 245, 645–46, 648 refusing 250, 253–54, 255 See also silencing; See also speech acts of refusal unwanted 37, 198, 250, 253, 254, 255 sexual assault 38, 229–30, 234, 236, 254 See also rape sexual negotiation 298 Shakespeare, William 655–56, 662–63, 666 shaming in the signed language community 426 sharing online 19, 72, 108, 113–15, 116, 149–50, 265, 358, 365–66, 557, 584–85 Shelby, Tommie 635–36, 637, 638, 639, 640, 642, 643, 644

Sherman, Richard 443 Shils, Edward 351 Ship of Theseus 32 signed language interpreters 420–22 silencing 79–80, 81, 86, 87, 88–89, 90–92, 96, 98–99, 211, 230, 232–33, 235–36, 245–46, 248, 249, 250, 252, 253–54, 255–56, 257, 258–59, 260–61, 417, 451–52, 454, 455–63 singular content.  See content Skinner, B. F. 9 slur-containing sentences and truth value of 126, 133, 137 utterances and truth value of 129, 133, 136–37, 143–44 slurs and compositionality 131 and homophones 164 associational a ects of 279 conventional implicature accounts of 154–55 expressivist accounts of 155–57 non-content theories of 159–60, 163–66 o ensive e ect 152, 153, 155–56, 157, 158–60, 163–66 o ensive potential 150, 152, 155–56, 157, 158, 159–60, 163 pejorative e ects of 278 Pejorative Tone Account of 149–50, 163, 164–65 pragmatics of 632 presuppositional accounts of 152–54 Prohibitionist theory of 149–50, 159–60, 163–66 Projection Condition 130–31 semantics of 126, 132, 632 vs. bare expressives 139, 140, 142–43 vs. racialized terms 437, 441 Sochiapan Chinantec 270–71 social hierarchy 212–13, 559f, 559–60 social media 99, 104–5, 109, 110–11, 112, 114, 115, 118, 265, 310, 358–59, 360–61, 421–22, 574–75, 614–15, 620–21, 622, 623–24, 625, 626, 627 social ontology nature of 631–32 See also Sally Haslanger social power 97, 99, 107, 116, 259–60 See also conversational standing; See also power social projection 62 social practices 61–62, 68, 69, 71, 338–39, 341, 644 social relations and false belief 64 See also false belief arrangement of 61–62, 63–64 social systems 60–61 socialization 256, 285–86, 423–24, 426 Socrates 388, 389, 390, 391, 392, 393, 394, 395, 396, 397, 398, 399, 400, 401, 402, 403–4, 408, 556 Soros, George 561–63 speech act engineering 622, 625 speech act input/output type agent-neutral 106, 107, 111, 116–17 agent-relative 107, 111, 116–17

community-speci c 106, 107, 109, 116 p. 704

speech acts

70, 79, 96, 99, 105–6, 107–8, 109–11, 112, 113–18, 119–20, 129, 182, 211, 215, 219, 220, 221–

23, 225–26, 228, 230–31, 232–49, 250, 251–55, 305–6, 435, 438, 452, 454, 456, 457–58, 461, 614–15, 616, 619 felicity conditions 105, 216, 249–50, 260, 303, 454 illocutionary acts 211, 215–19, 221–22, 223, 224, 225, 227, 228, 230–34, 303, 456, 615, 618, 623 illocutionary disablement 249–50, 252, 254, 255, 256 locutionary acts 216–17, 223–24, 232, 247, 617–18 medium-dependent/independent distinction 105, 108 mis res 220, 227, 230, 232–33, 249–50 perlocutionary acts 216, 217 perlocutionary disablement 462 pragmatic structure of 104–5 uptake failure 248, 249–50, 253–54 speech sounds 425 Stalnaker, Robert 79, 80, 82, 83, 96, 471–73, 474, 631–32 standing in conversation  See equal standing Stanley, Jason 541–42, 631–33, 641, 645 Stanton, Elizabeth Cady 45 statecraft 671 stereotypes 88, 126–27, 235, 297, 302, 308, 320–21, 339–40, 355, 370, 371, 423–24, 426–27, 560–61, 637, 667 Stewart, Lindsay 443 Strawson, P. F. 9–10, 11 structuralism about language 194–95 See also Chomsky structured propositions and structured map contents 476–77 subordinating speech 97, 294 subordination 218 and pornography 212, 214 See also harm of pornography and power 259–60 and speech acts 219 See also illocutionary acts of subordination institutional 219–20 material 219–20, 221–22 of women 211, 212, 214, 221–22, 225, 245, 247, 248–49, 258–59 sulking 183 See also passive aggression Sundell, Timothy 11, 15 Supreme Court of India 354 Supreme Court of the United States 356, 359 swearing act of 196 and emotion 205 and masculinity 202 and ritual performance of power 202 and taboo 196–97, 203–5 cross-cultural di erences 197–98 pragmatics of 196–97

swearwords 196, 201, 206 See also taboo words taboo lexicon evolution of 203–5 taboo words.  See also slurs and ctional narratives 200 and swearing 195–97 cross-cultural di erences 197–98 structural vs. functional theorizing 194–95 tagging on social media 112–13 tagging semantics 510–11 See also map semantics Tarski, Alfred 10, 11 testimonial endorsement online 118, 620–21 See also #MeToo; See also retweeting testimony 117, 285, 421–22 Thomson, Judith Jarvis 351 TikTok 265 Tourette syndrome 198, 205 trans communities and joyous language use 307–11 and playful language use 304, 305, 306–7, 309, 314 transfeminism and gendered language 294 transgender.  See also gender-nonconforming authority 306 discursive agency 304–5 joy 298–99 preferred pronouns 265, 280–81, 282, 286, 297 See harm of misgendering; See misgendering transphobia 369 Trump, Donald J. 451, 457, 550, 552, 557, 561–63, 613, 622–23, 625–26 truth-conditional content 158 See also content truth-maker semantics 73–74 See also semantics turn-taking in communication 96 p. 705

tweeting

113–14, 115, 116, 616, 617–18, 624, 625–27 See also retweeting

Twitter 108, 115, 201, 422, 443, 451, 548, 613–14, 617–18, 620–21, 623–25, 626–27 uncooperative 86, 87–88, 93, 233 See also Cooperative Principle underprivileged romanticizing the 44 United States 37, 182, 201, 202–3, 204, 220, 229, 276–77, 323–24, 350–51, 353–54, 356–57, 358, 360, 394, 397–98, 440, 457–58, 473–74, 634 Civil War 558 Constitution 213, 214, 323–24 University, Shawnee State 265 unjusti ed a ect misrepresentations due to misplaced 135 uptake failure 248, 249–50, 253–54 utterance-level refusals 442, 443 See also illocutionary acts of refusal; See also refusals utterances

and AI metasemantics 582–83 performative vs. constantive 223, 224, 297–98 See also Austin; See also speech acts vagueness and privacy 351, 353–54 validity 28 veiled gaze 554–56 Veneziano, Domenico 532 Venus 17, 554–56 verdictives 211, 220, 221–23, 224, 225, 234, 248–49, 251, 259 See also illocutionary acts; See also speech acts violations of taboo 159–60 sexual 43, 116–17, 118 visual dogwhistles 554, 563 See dogwhistles Volta, Alexander 601, 602, 603 Watt, James 604, 605, 606, 607 Weitz, Morris 672–73 welfare 126–27, 551, 637, 638, 639, 647 West, Cornel 437 what is said 134, 136–37, 202, 427–28, 631–32 white supremacists 558 Williamson, Timothy 25, 26, 154, 157 Wittgenstein, Ludwig 105, 312, 586, 673, 678 wonder-working power 183, 438, 439 See also dogwhistles wordplay 419–20 workplace accessibility 402, 417, 588, 624 accommodation 381, 416, 506 worldview 249 Wynn, Natalie 305 X.  See Twitter Youngkin, Glenn 540, 542 YouTube 175 YouTuber 305, 617–18 Zapatistas 59, 60–62, 67, 69, 71, 73, 75