Mind, Language and Morality: Essays in Honor of Mark Platts 9780815385028, 9781351202596

Table of contents :
Cover
Title
Copyright
Contents
Foreword
1 Ways of Meaning and Knowing Moral Realities
2 Platts on Kant and Mandeville
3 Wrong Direction? A Criticism of Direction of Fit
4 Equality as a Foundation of Ethics
5 Inflation or Deflation of Rights?
6 The Debate on the Abuse of the Concept of Human Rights
7 Convergence or Divergence in the Evolution of (Criminal) Rights? A Case Study of the Multiple Incoherencies of the Presumption of Innocence
8 Wittgenstein on Rule Following: Some Themes and Some Reactions
9 Kantian Neuroscience and Radical Interpretation: Ways of Meaning in the Bayesian Brain
10 Reflections and Replies
a. “Barry Stroud on the Meaning and Transcendence of Values”
b. “Ralph Walker on an Attempted Alliance Within Moral Philosophy”
c. “Gustavo Ortiz-Millán on the Direction of the Mind”
d. “James Griffin on the Ethics of Equality”
e. “Juan Antonio Cruz Parcero, Larry Laudan and Rodolfo Vázquez on Candidates for Being Human Rights”
f. “Paul F. Snowdon’s Prolegomena to Wittgenstein and Rule-Following”
g. “Jim Hopkins on Theories of Meaning, Mind and the Brain”
11 Closing Comments: Philosophical Life
Contributors
Index

Citation preview

Mind, Language and Morality

Mark Platts is responsible for the first systematic presentation of truthconditional semantics and for turning a generation of philosophers on to the Davidsonian program. He is also a pioneer in discussions of moral realism, and has made important contributions to bioethics, the philosophy of human rights and moral responsibility. This book is a tribute to Platts’s pioneering work in these areas, featuring contributions from a number of leading scholars of his work from the US, UK and Mexico. It features replies to the individual essays from Platts, as well as a concluding chapter reflecting on his philosophical career from Oxford to Mexico City. Mind, Language and Morality will be of interest to philosophers across a wide range of areas, including ethics, moral psychology, philosophy of law and philosophy of language. Gustavo Ortiz-Millán is Research-professor of Philosophy at the Institute for Philosophical Research, National Autonomous University of Mexico (UNAM). Juan Antonio Cruz Parcero is Research-professor at the Institute for Philosophical Research, National Autonomous University of Mexico (UNAM).

Routledge Festschrifts in Philosophy

1 Mind, Language and Morality Essays in Honor of Mark Platts Edited by Gustavo Ortiz-Millán and Juan Antonio Cruz Parcero

Mind, Language and Morality Essays in Honor of Mark Platts Edited by Gustavo Ortiz-Millán and Juan Antonio Cruz Parcero

First published 2018 by Routledge 711 Third Avenue, New York, NY 10017 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2018 Taylor & Francis The right of the editors to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data A catalog record for this book has been requested ISBN: 978-0-8153-8502-8 (hbk) ISBN: 978-1-351-20259-6 (ebk) Typeset in Sabon by Apex CoVantage, LLC

Contents

Foreword

vii

GUSTAVO ORTIZ-MILLÁN AND JUAN ANTONIO CRUZ PARCERO

  1 Ways of Meaning and Knowing Moral Realities

1

BARRY STROUD

  2 Platts on Kant and Mandeville

13

RALPH WALKER

  3 Wrong Direction? A Criticism of Direction of Fit

26

GUSTAVO ORTIZ-MILLÁN

  4 Equality as a Foundation of Ethics

49

JAMES GRIFFIN

  5 Inflation or Deflation of Rights?

59

RODOLFO VÁZQUEZ

  6 The Debate on the Abuse of the Concept of Human Rights

68

JUAN ANTONIO CRUZ PARCERO

  7 Convergence or Divergence in the Evolution of (Criminal) Rights? A Case Study of the Multiple Incoherencies of the Presumption of Innocence

83

LARRY LAUDAN

  8 Wittgenstein on Rule Following: Some Themes and Some Reactions PAUL F. SNOWDON

97

vi Contents   9 Kantian Neuroscience and Radical Interpretation: Ways of Meaning in the Bayesian Brain

115

JIM HOPKINS

10 Reflections and Replies

148

MARK PLATTS

a. “Barry Stroud on the Meaning and Transcendence of Values”  149 b. “Ralph Walker on an Attempted Alliance Within Moral Philosophy”  153 c. “Gustavo Ortiz-Millán on the Direction of the Mind”  158 d. “James Griffin on the Ethics of Equality”  162 e. “Juan Antonio Cruz Parcero, Larry Laudan and Rodolfo Vázquez on Candidates for Being Human Rights”  166 f. “Paul F. Snowdon’s Prolegomena to Wittgenstein and Rule-Following” 174 g. “Jim Hopkins on Theories of Meaning, Mind and the Brain”  178 11 Closing Comments: Philosophical Life

183

MARK PLATTS

Contributors Index

191 193

Foreword

This book is a tribute to Mark Platts, a tribute paid by some of his colleagues, friends and old students from both sides of the Atlantic. In 1977, Platts came to Mexico for the first time, invited by the National Autonomous University of Mexico, and after spending seasons in the country as a visiting professor, he decided to accept an offer from this university to move to Mexico, as a research professor of the Institute for Philosophical Research, since 1984. At that moment, Platts was already recognized as part of a generation of Oxford’s young brilliant philosophers, many of whom were students of P. F. Strawson: Christopher Peacocke, Gareth Evans, Paul Snowdon, Colin McGinn, among others. Maybe a sample of Strawson’s stature as a philosopher and professor is that his students created their own philosophical theories, different from his theory; but despite their independence with regards to their professor’s philosophy, many of them shared the enthusiasm that, oddly, was felt at Oxford for an American philosopher: Donald Davidson. During the 1960s, several philosophers at Oxford (mainly David Wiggins, Barry Taylor, Gareth Evans and John McDowell) introduced what was known as the “Davidsonian program” in the philosophy of language. In 1969, Davidson has already taught the John Locke’s lectures at Oxford (with the topic “The very idea of a conceptual scheme”) and later, he became a visiting fellow at All Souls College. It is when Davidson’s influence over an entire generation of philosophers from Oxford was born, under the idea that it was possible to develop a semantic theory in terms of truth conditions, based upon Tarski’s work on the concept of truth in formalized languages. However, Oxford did not easily surrender to Davidson and many saw in Michael Dummett’s theory a response to the semantics of truth conditions. Davidson himself shares how he gave a seminar on truth theories, along with Dummett and “the seminar was apparently viewed by the Oxford community as a sort of gladiatorial contest”.1 Dummett was reading Davidson’s philosophy, from the implicit metaphysical content in the theory of truth conditions and was providing a realistic lecture of Davidson—with which, on the other hand, Davidson never really identified himself, given his reluctance to take part in the realism-anti-realism debate. The philosophical discussion held at Oxford during those years was well

viii Foreword captured in the book Truth and Meaning, edited by Gareth Evans and John McDowell in 1976. In that environment, Mark Platts was a key figure. Davidson’s papers were scattered in different journals and books, many of them were difficult to access, and an overview of what the so called “Davidsonian program” meant was missing. Platts published in 1979 Ways of Meaning: An Introduction to a Philosophy of Language, that is, an introduction to Davidson’s philosophy of language. It was the first systematic presentation of the Davidsonian semantics. When the book was published, Steven E. Böer, from Ohio State University, summarized his opinion in a review: For some time now, we have badly needed a clear, reasonably detailed, and methodologically sophisticated account of how the many pieces of the Davidson Program fit together and relate to broader philosophical concerns. Ways of Meaning meets this need admirably. That we are left hungry for more should be taken as a compliment to the quality of what we have been served.2 When, almost twenty years after the book saw the light for the first time, MIT Press published an expanded second edition, the language philosopher Peter Ludlow wrote the following comment: This is the book that turned on a generation of philosophers of ­language— turned them on to the Davidsonian program, that is. Clear, readable and to the point, Mark Platts made truth-conditional semantics intelligible to us. More than that, he surveyed a number of natural-language constructions, showing how they could be handled in such a framework, and thereby mapping out the landscape of what has since become a full-blown philosophical research program.3 Platts’s book was not only important due to its contribution to the philosophy of language in general, and the “Davidsonian program” in particular; it was so, because it preferred the realistic reading Dummett suggested, and that McDowell also adopted. Equally, that reading guided Platts, in the last part, to present the implications of this theory for moral philosophy. Platts defended moral realism, which would be articulated at the same time by philosophers such as McDowell, Wiggins, Putnam, among others, and that claims that the moral statements express propositions that can be true or false, and which truth depends on its correspondence with facts or moral properties independent from the mind. The last chapter of this book, “Moral Reality” has been reprinted in several anthologies and it was decisive in the development of moral realism, a theory of great popularity nowadays. Almost as an accompaniment to Ways of Meaning, Platts published in 1980, Reference, Truth and Reality: Essays on the Philosophy of Language, a collection of essays on the philosophy of language that bring together

Foreword  ix several essays on truth-conditional semantics. Platts included there his text “Moral Reality and the End of Desire”, a discussion on the nature of desires that would lead him to the elaborate a moral psychology more in tune with the moral realism he started to defend in his first book. That moral psychology will be fully developed in his book Moral Realities: An Essay in Philosophical Psychology (1990). A review of the book stated: This is an impressive book: one cannot but admire its subtlety of argument, its scope and the complex interrelations between its different parts. The sensitive account of desire is particularly welcome, and will, I am sure, have a beneficial influence in many areas of philosophy, while Platts’ capacity to explore a position in an imaginative and critical way is displayed to perfection in his discussion of Hume’s “master argument.”4 Perhaps, the most important part of the book is centered, precisely, in his analysis on the nature of desires and his criticism of the Humean theory of motivation. If there are moral realities, then it must be possible that belief or the perception of these facts motivate our desires and lead us to action. In order to provide such a theory of motivation, one should distinguish kinds of desires: unmotivated desires, like the ones Hume talks about, but also, motivated desires, like those emerging from cognitive states—such as the characterizations of the desirability of the object of our desire—but also from perceptual states. Platts’s moral realism had leaded him to a characteristic anti-Humean posture in moral psychology, according to which, desires are not always end points in processes of practical deliberation, beyond the influence of argumentative processes, but instead in many occasions they respond to evaluative beliefs. In both, Ways of Meaning and Moral Realities, there is a great number of important ideas for moral philosophy, such as the notion of direction of fit of desires and beliefs, which has had great influence on the discussions about the difference between beliefs and desires, precisely in the way Platts formulated it; but there are others that are still waiting for philosophical justice, for example, Platts’s notions of reasonable causation or that of protopractical deliberation, that is, the type of deliberation we get into when we think about the object of our desires and that does not necessarily lead us into action. When Platts decided to change his residency and move to Mexico, he had already been a professor and a researcher at Magdalen College, Oxford, and he was a professor at Birkbeck College, London (who appointed him lifetime honorary member upon his coming to Mexico). Once in Mexico, Platts developed an interest on applying many of his moral ideas into practical affairs, and the opportunity emerged during the AIDS epidemic in the mid 80’s. His book On the Use and Abuse of Morality: Ethics, AIDS, and Society—the first originally written in Spanish—offers a

x Foreword lucid analysis of the stigmatization attitudes towards AIDS patients, of the attitudes generated during an outbreak and of other moral issues. Platts then took his philosophical concern into practice, and participated in the committee that got universal access to the HIV/AIDS Programs in Mexico City, something that many other countries are far from reaching. This was a major accomplishment for the advancement in the battle against AIDS. Besides, in times when philosophers rarely get involved in practical matters such as fighting such an epidemic, Platts’s attitude was exemplary. Today, Platts continues as active as ever. Recently, he published the book To Be Responsible, a clear and thorough analysis of the concept of responsibility. He has also edited several anthologies on ethics, which collect some key texts for discussions in contemporary moral philosophy, something that also shows his commitment with teaching. Mark Platts has stood out in the establishment and development of analytical philosophy in Mexico; he has trained many professors and researchers in the country. In fact, we could say that his teaching task is comparable, in many ways, with those great teachers of the Spanish exile in Mexico— those who came after the Spanish Civil War—and who introduced topics, trained students and help with the spreading of fundamental books leaving a mark in the history of Mexican (and Latin American) philosophy. But Platts’s influence has crossed beyond our frontiers, and this is clear in many of the collaborations included in this volume. This book is the result of the tribute paid in Mark Platts’s honor at the Institute for Philosophical Research of the National Autonomous University of Mexico, which took place in August 29 to 31, 2012. It says a lot about Platts’s achievements that not only philosophers participated at the event, but also people who collaborated with him during his social battles related to the AIDS epidemic (unfortunately, not all the papers presented at that occasion could be reproduced here). Let this book be a testimonial of the appreciation that friends, colleagues and old students show to Mark Platts. Gustavo Ortiz-Millán Juan Antonio Cruz Parcero

Notes 1 Donald Davidson, “Intellectual Autobiography,” in The Philosophy of Donald Davidson, ed. L. E. Hahn (Chicago: Open Court, 1999), 53. 2 S. E. Böer, “Review of Ways of Meaning,” Linguistics and Philosophy, vol. 4, no. 1 (1980): 155. 3 Back cover of Ways of Meaning, 2nd expanded edition (Cambridge, MA/London: MIT Press, 1997). 4 A. D. M. Walker, review of Moral Realities, Philosophical Quarterly, vol. 43 (1993): 109.

1 Ways of Meaning and Knowing Moral Realities Barry Stroud

I would like to take up in rather general terms some of the important and distinctive features of Mark Platts’s explanation and defence of his views of morality. Not his own personal moral views, about which he warns us we are better off not knowing anything. I mean his views of the nature of morality and of the kind of knowledge and understanding human beings have of it. I find those views and his defence of them very congenial, and I am happy to have a chance to pay tribute to them and to their author here. I will eventually draw attention to one place in his reflections where Mark speaks of a certain resistance one might feel to the kind of view he has been defending. He considers a way of accommodating the source of that resistance, and so in that way overcoming it. I raise the question whether what he suggests is really a successful response to the resistance he has in mind. That turns out to be a more complicated matter than it looks, and I will try to explain why. It depends on what the resistance Mark identifies actually is, and from where it comes. That raises some large, difficult questions in turn. In his Ways of Meaning Mark’s aim was “to present and discuss [. . .] the most important recent contributions to the philosophy of language.”1 The most important recent contributions to that subject through the 1970s were Donald Davidson’s elaborations of the idea of a theory of meaning for a particular language. What came to be called a Davidsonic boom had been resounding through Oxford and beyond in those years, and Ways of Meaning tells us what it was all about. And it tells it, as it aspired to, “in a straightforward and accessible manner.” That book was the best and most thorough book-length treatment of that whole subject up to that time. It explained and developed the basic ideas of the new theory of meaning and investigated its implications as an account of a speaker’s understanding of the words of his language: what a speaker who knows English, or Spanish, knows simply by being a competent speaker of that language. Davidson’s idea was that a speaker of a language can be said to know the conditions under which each of the indicative sentences he understands would be true. The task of what he called a theory of meaning for a particular language was then to give a systematic description of the structure

2  Barry Stroud of the language in a way that would yield, for each sentence, a statement of what the truth-conditions of that sentence are. If all those statements about sentences of the language are derivable from the structure of the language itself, they would all express something that someone who understands that language can be said to know. A speaker of English, for instance, knows that the sentence “Ripe tomatoes are red” is true if and only if ripe tomatoes are red. It seems very hard to deny that a speaker of English does know such a statement to be true, and that his knowing it is involved in his understanding the English words in the sentence. But however undeniable that might be, it was regarded by many philosophers at the time as completely inadequate as an account of a speaker’s understanding of the meanings of his words. It was thought to offer no explanation of how a competent speaker knows what he knows, or how his mere knowledge of the truth of such biconditional statements “issues in” or is connected with the actual applications and responses he makes in real life with the words he is said to understand. Behind this objection was the thought that a speaker’s use of her words requires the grasp of certain “procedures” or “grounds” on the basis of which she applies or withholds her words on particular occasions. If some such “recognitional procedures” or capacities are required for a speaker to use and understand her words, it was thought that they must be an essential part of any account of the meanings of those words. And a theory of meaning in terms of truth-conditions alone says nothing about that aspect of a speaker’s competence. Mark did everything he could in that book—short of the “evangelism” to which he assures us he would never stoop—to disarm this line of resistance to the truth-conditional theory. And what he and others did was as much as I think should have been needed. The resistance came largely from those of a broadly verificationist persuasion, or those with a basically “epistemic” conception of truth. Mark accordingly called the kind of theory he defended a “realist” account of a speaker’s understanding, and a “realist” conception of meaning and truth. One can perhaps see why. But just how robustly metaphysical that talk of “realism” was meant to be is not easy to say. I think this is connected with the difficulty I will consider. Mark calls the theory he defends a form of “realism.” That is presumably meant to imply more than the evident fact that the conditions under which each of the sentences a speaker understands would be true can be stated in the very words used in the sentences the speaker is said to understand. That is an important feature of the theory, but there seems to be nothing distinctively “realist” about it. It means that anyone who understands the words used on the right-hand side of the relevant biconditional statement will know the conditions under which the sentence mentioned on the left-hand side is true. What enables a speaker to know such a thing is the speaker’s competence in the language used in the biconditional statement. That is the language the theory of meaning is expressed in. Someone who

Ways of Meaning and Knowing Moral Realities  3 does not understand that language could not learn from such biconditionals the truth-conditions of their mentioned sentences. This shows that a theory of meaning of this kind is not meant to provide a way into a language for someone who does not already speak and understand a language; it does not explain new words or concepts to someone who does not already have them. A theory of this kind could state the truth-conditions of sentences of one language in the words of another language. But only someone who already understands that other language could learn from that theory of meaning what competent speakers of the first language know and understand. This means that speakers must rely on the understanding they already have of some language to know what a theory of meaning for that language or for any other language says. A theory of this kind is perhaps more appropriately called “realist” in contrast with the “anti-realism” or “non-realism” of verificationist or “justificationist” views. Unlike those views, a theory of meaning of the kind Mark defends takes no account of any alleged “recognitional procedures” speakers make use of in applying their words or in determining the actual truth-values of the sentences they understand. The theory is in that sense also what Mark calls “austere”; it offers no explanation of speakers’ application of their words and no “basis” or “ground” of their knowledge of the meanings of words. The main point in the label “realism” is that the theory assumes that sentences can be true (or false) independently of our capacity, or incapacity, to recognise them as true or false [. . .] that we can know what it is for a sentence to be true or false [. . .] even if it is beyond our capacities to recognise whether those truth-conditions apply or not. (WM, 5–6) I think there is no question that this is the way Donald Davidson understood the implications of the kind of theory he had in mind. As far as I know, Mark Platts was one of the first philosophers to publicly and explicitly take up the question of the extent to which a theory of meaning and understanding of this kind is applicable or can be extended to the domain of evaluative language. Mark had in mind specifically morality, and he certainly raised the question before Donald Davidson himself had begun to discuss the issue publicly. David Wiggins and John McDowell were already working in nearby territory. I believe they both had an influence on Mark’s thinking at that time, perhaps even in the direction I am going to question.2 In the early pages of Ways of Meaning Mark expresses some doubts about the plausibility of extending the application of what he calls his “realistic reading [. . .] of the truth-predicate” to “an ethical sentence” (WM, 12). But in the last chapter of that book, “Moral Reality,” he begins to go into the question directly. It cannot be said that he carries the investigation of that question very far in that chapter. But he insists that to apply the

4  Barry Stroud truth-conditional theory of meaning and understanding to moral sentences we must honour what both “utilitarianism” and “intuitionism” would preserve. Both those theories are “realist” in seeing ethical judgements as either true or false, and intuitionism is also “austerely realistic” in rejecting “justificationist” demands for “explanatory” accounts of conditions or procedures of application of moral terms, or for reductions of their contents to non-moral bases. The point to be preserved in any case is that “A speaker can know, have a grasp of, the truth-conditions of a moral sentence even if those truthconditions are beyond his (present) recognitional capacities” (WM, 245). When this kind of account is generalized, it would mean that: By the process of careful attention to the world, we can improve our moral beliefs about the world, make them more approximately true; by the same process, we can improve our practical understanding, our sensitivity to the presence of instances of the moral concepts that figure in these beliefs [. . .] Our moral language, like all the realistic part of that language, transcends our present practical comprehensions in trying to grapple with an independent, indefinitely complex reality. (WM, 247) This is a good expression of what the successful application of the truthconditional theory of meaning and understanding to moral language would give us. I think it captures the ways we actually think of moral assessment and helps explain whatever sense we have of discernible moral progress. Mark pursues this apparently intelligible goal in his Moral Realities.3 In that book he explains how a proper defence of the view depends on the correct understanding of what he calls, broadly speaking, “desires”—those “active” or “practical” attitudes that are essential to all intentional action. This leads into an intricate discussion of such attitudes, always with an eye on this central question of motivation and reasons. For our purposes here we can focus on one particular set of attitudes he draws attention to: so-called “motivated” or “reason-following” desires, as opposed to other so-called “unmotivated” or “reason-producing” desires. These latter are desires or wants we just find ourselves with which seem in their very presence to give us some reason to do something that we did not have reason to do before. “Motivated” or “reason following” desires, on the other hand, are present when an agent is moved or can be said to “want” to act in a certain way, not because he just feels like it or because it is a way of satisfying some desire he independently happens to have, but because he recognizes the desirability or value in acting in that way—either in its outcome or in the performance itself. The agent sees and takes something he recognizes to be true of the action or its outcome as reason to do it, and is so motivated by that recognition. For “desires” or attitudes of this distinctive kind, as Mark puts it: “The agent does what he does because he thinks that something—say, honesty, or

Ways of Meaning and Knowing Moral Realities  5 scientific truth—is a value, or is of value, or matters, or is desirable, or is worthy of desire” (MR, 76). The point is that, in such cases the agent acts, or is moved to act, as he does because he thinks or believes something about the action or object in question; he sees some feature he takes the action or object to have as reason to act in that way, and he is moved to act in the way he sees there is reason to act. The agent’s thought or belief in such a case could be called an “evaluative” thought about something. But an agent’s having such an “evaluative” thought or belief does not imply that the value she sees in the action or object is, or is thought to be, dependent on her or anyone else’s having that “evaluative” thought or belief, or “valuing” the thing in the way she does. The reason the agent sees in favour of doing such-and-such is a reason she takes to hold of the action she considers, and to hold of it whether or not she or anyone else actually thinks it holds. This is, after all, a familiar aspect of propositional thought in general, about virtually any subject-matter. In the thought that ripe tomatoes are red, for instance, what one thinks is something that holds, or fails to hold, and is thought to hold or fail to hold, quite independently of whether anyone, including the present thinker, actually thinks that ripe tomatoes are red. If the same is true of evaluative thought, then whatever value- or desirabilitycharacteristic an agent takes an action to have is something thought to be true of or present in that action independently of whether anyone, including the agent, actually thinks that characteristic is present, or even takes that characteristic to be a desirability-characteristic at all. Mark sums up his view of socalled “motivated” or “reason-following” valuings or desirings in this way: in these cases valuing can count as a case of being presented with some value-property or “quality in objects” which is there anyway and which is at least worthy of being valued independently of actual human desirings and valuings; [. . .] an agent’s desiring the thing concerned can be at least a reasonable response to the value independently had by the thing. (MR, 69) Having ably explained and defended this view, Mark then considers a certain objection he thinks many would make against it: the charge of what he calls “transcendentalism.” Here we approach the source of my doubts. Mark puts the objection like this: In view of what has been said in defence of the idea that [. . .] such valuing can count as a cognitive response to some value-property or value-feature which is independently there in objects without the mind, the thought might be invited that this account holds the existence of such v­ alue-properties or value-features [. . .] to be completely independent of human valuings and desirings. And that might be judged rather hard to swallow. (MR, 98)

6  Barry Stroud This is the charge or reaction Mark calls “transcendentalism.”4 Whatever it is called, it is an expression of resistance to the idea that the properties we ascribe to objects or actions in our evaluative judgements about them are independent of, or transcend, actual human valuings and desirings. Mark suggests that might that be thought hard to swallow because it appears to be incompatible with what many regard as an obvious truth: that “actual human valuings and desirings are the source of all value” (MR, 95). I find this last idea (“the source of all value”) a very slippery thought. We can begin to get a firmer grip on it by asking whether the thought is obviously true, or even true, in a way that threatens the idea of something’s having a certain value independently of anyone’s valuing or desiring it. There seems to be one way in which that slippery thought represents no threat to that form of independence, as we have just seen. Ripe tomatoes are red (or not) independently of whether anyone thinks they are red, so someone who thinks that ripe tomatoes are red thinks something that is true (or false) independently of whether he or anyone else thinks it. Of course, if a certain person thinks that ripe tomatoes are red, then someone does actually think it, but what that person then thinks does not itself imply that he thinks it or that anyone else does. Similarly, it seems, someone who thinks that slavery is wrong or deplorable can think it is wrong or deplorable independently of whether he or anyone actually thinks it is wrong, even if everyone (except maybe the slaves) actually thinks it is a wonderful idea. Of course, someone who condemns slavery does in fact make that particular evaluation of slavery, but what he thereby thinks about slavery does not imply that he or anyone else makes that or any other judgement about it. Mark gives other examples of valuings that are understood to hold independently of whether anyone actually makes such valuings or even would make them. How, then, does the complaint of “transcendentalism” represent a threat or challenge to the idea that the value-properties we ascribe to objects in evaluative judgements can hold of them independently of all actual human valuings or desirings? What exactly makes that thought of independence hard to swallow, if it is? Mark does not try to answer that question directly. His first step in trying to “free” his view from what he calls “its appearance of (a certain kind of) ‘transcendentalism’ ” is to consider the well-known view of Hume that “Vice and virtue [. . .] may be compar’d to sounds, colours, heat and cold, which according to modern philosophy, are not qualities in objects, but perceptions in the mind.”5 The introduction of this view at this point is somewhat puzzling. Of course, Mark does not recommend the full Humean doctrine of values as something like so-called “secondary” qualities. That is precisely the kind of view he opposes. For his particular purpose here he means to endorse only that aspect of the view that says that so-called “secondary qualities” are dispositional properties. They are in effect dispositions objects have to produce certain responses in certain kinds of conscious subjects under certain specified conditions. Let us not ask for the moment what those responses might

Ways of Meaning and Knowing Moral Realities  7 be, or under what conditions they would be produced. The point so far is that the ascription of such qualities to something takes a conditional form: the object would produce such-and-such effects if such-and-such conditions were fulfilled. Qualities of this kind are of course just as much qualities of the objects they are truly ascribed to as are all the categorical or non-conditional qualities objects also possess. So this might look like a way of saving the idea that objects can have evaluative properties independently of whether anyone actually responds to them in any of the relevant ways. For objects to have evaluative properties in this way it would be sufficient for the objects simply to be disposed to produce valuings of certain kinds in certain kinds of subjects, even if none of those subjects ever actually makes any of the relevant valuings. Objects’ having value-properties would be in that sense independent of any actual human valuings or desirings in just the way something can have the dispositional property of being soluble in water independently of whether it is ever actually put in water or dissolves. Is this a way of defending the view that valuing is a “cognitive response to some value-property [. . .] which is independently there” against the charge of “transcendentalism”? The answer obviously depends on what that charge really is, or what its source is. I think there is a way of understanding the charge of “transcendentalism” in which the answer is No; this move does not overcome the objection. Someone who objects to Mark’s view on grounds of its alleged “transcendentalism” could grant that the objects that affect us in evaluative ways have dispositional properties. We could be disposed to respond to objects in certain evaluative ways independently of our actually making any such evaluative judgements or “valuings,” even if “actual human valuings and desirings are the source of all value.” Mark’s view is not just that we can be disposed to make evaluative judgements even if no one ever in fact actually makes those judgements. His view is that the evaluative judgements we do or would make can be “cognitive responses” to some value-properties we think an object possesses independently of anyone’s either making or being disposed to make evaluative judgements about it. I take this to be what those who make the charge of “transcendentalism” object to. The objection is not that objects could not have dispositions to produce certain kinds of evaluative responses without someone’s actually making any of the evaluative judgements in question. The objection is presumably directed against the very conception of evaluative judgements or “valuings” of the kind Mark says we are disposed to make. I think Mark is right to understand evaluative judgements or valuings in that way. What, then, is the objectionable “transcendentalism” or transcendence that the objection complains is present in that conception of evaluative judgement? This is not a simple question. When Hume said, “Vice and virtue [. . .] may be compar’d to sounds, colours, heat and cold, which according to modern philosophy, are not qualities in objects, but perceptions in the mind,” he

8  Barry Stroud was invoking the “modern” view that the sounds, colours, heat and cold that we believe in can be understood only in terms of certain effects things in the world have on conscious subjects. The idea was that without such effects on perceivers, there is nothing in the independent world for sounds, colours, and so on to be. In drawing the parallel with values, Hume meant that there is nothing in the world, nothing in “objects as they really stand in nature, without addition or diminution,”6 for evaluative terms like “vice” and “virtue,” “good” and “bad,” and so on, to stand for or mean. Hume thought all such terms can be understood only as involving reference to certain kinds of responses conscious subjects would have to objects that affect them. Any other way of trying to understand those terms would involve something evaluative that is completely independent of, or transcends, all such possible responses. And what could that be? The objection does not seek an answer. The point is that this can be only a rhetorical question. Suppose we ask why those “modern” philosophers thought there is nothing in the independent world for terms for colours, sounds, and so on to pick out. Why did they think there could be nothing in the world independent of all possible human responses for sounds and colours and so on to be? I think that is actually a very difficult question that has still not been adequately answered. I suspect certain metaphysical preconceptions were at work; fixed ideas about the only kinds of things there could possibly be in reality. We need not go into that now. But in the case of values, when we ask “Why do Hume and so many other excellent philosophers think there is simply nothing in the world independent of all possible human responses for values or value-properties to be?” there is thought to be a convincing answer to the question. And acceptance of a certain version of that answer is what I think is the source of the charge of “transcendentalism” against views of evaluative attitudes of the kind Mark defends. The answer to the question that supports subjectivism about values has two parts. The first, which is the more convincing part, is that evaluative judgements or “valuings” must be understood as “active” or “productive” attitudes that play an essential role in human action. In order to act in one way rather than another an agent must prefer or assess or “value” that way of acting more highly than others. If a potential agent were completely indifferent among all available possibilities, with no preference for or evaluation in favour of any option over any others, there would be no intentional action, and so in that sense no agency. So it looks as if agents’ attitudes can be understood as evaluative only in relation to their potential role in action. There could be nothing in the world for evaluative or value-properties to be that was not connected in some way or other with possible intentional actions. That is the first step, and it seems convincing. It is in the second part of the answer, which is meant to explain exactly how evaluative attitudes are to be understood as connected with action, that there is room for resistance or disagreement. Not all versions of this second part of the story are equally convincing. For Hume and countless other

Ways of Meaning and Knowing Moral Realities  9 philosophers, agents who act intentionally must be understood to have certain kinds of feelings or passions or sentiments. That is because only something like feelings, passions, or sentiments were thought to be capable of bringing about actions. Reason, or the discovery of truth alone, without any “addition” of anything else, as Hume put it, is “utterly impotent in this particular.”7 Belief in the truth of something, all on its own, could never bring about action. Because feelings or sentiments must be present for action, all “active” evaluative attitudes or “valuings” must involve the presence of feelings or sentiments. Such feelings or sentiments will be the evaluative attitudes of a potential agent only in so far as they “give” to objects or possible actions that an agent considers the only value or value-properties they can have for her. So in this view, no object or action would have more value for an agent than anything else unless there were feelings or sentiments inclining her in one direction rather than another. Without some such feelings, she would find no more reason or inclination to do one thing rather than another thing, even, as we say, doing nothing. I think this idea of desires as feelings or sentiments that move one to act is what gives force to what I called the “slippery thought” that “actual human valuings and desirings are the source of all value.” The idea is that it is only our actually having certain desires and sentiments that ever gives us any reason to do the things we do. Those desires and “active” feelings are in that sense the “source” of all motivation, and so the source of whatever value we can find objects or actions to have. Mark suggests that his view that valuing is a “cognitive response to some value-property [. . .] which is independently there in objects without the mind” is “straightforwardly incompatible with the suggestion [. . .] that actual human valuings and desirings are the source of all value” (MR, 95). I think that is right, if those valuings and desirings are understood as requiring certain feelings or sentiments. The two different views of evaluation or human motivation are incompatible. Attachment to the idea that feelings or sentiments are required for all evaluative intentional action is the real source of the charge of “transcendentalism” against more purely “cognitive” views. It is widely believed to be obvious that there is simply no room in the world for independent values or value-properties, and no need for them, since evaluation must be connected with action, and desires are feelings or sentiments that move us to action. If that were the only way to understand desires and acting for reasons, that “slippery thought” about “the source of all values” would perhaps have to be accepted. But that is precisely what the defence of so-called “motivated” or “reason-following” desires denies. That is the heart of Mark’s view; it is where the real dispute with those who complain of his “transcendentalism” is to be settled. That is a huge question. We can perhaps take a few beginning steps towards defence of a more “cognitive” view if we ask, as a matter of observed fact, whether it ever happens that someone is led to act in a certain way by recognizing something as a reason to act in that way. For example,

10  Barry Stroud does it ever happen that someone is moved to act in a certain way by seeing an unattended child struggling in the sea, or by learning that a friend of his is in some distress? We all know people, I hope, who would immediately act in such circumstances. It happens. Such people go to help, it seems, because they see the impending danger for the child or the distress of the friend as reason to do what they do. They recognize something they take to be true of the other person as reason to act in a certain way, and they act accordingly. And, it seems, those people would acknowledge that the child’s being in danger is reason to act in that way whatever anyone might think or feel about it or whether others would be moved to act as they do. It would be to no avail at this point to insist that if those people act to help it is only because they want to act in those ways. Of course they want to; they want to because they see the danger or the distress as reason to help. They have a so-called “motivated” or “reason-following” want or desire to do what they do. This does not mean that everyone who sees that the child is in danger or that the friend is in distress is moved to act in the same way. Seeing that a child is in danger or that a friend is in distress is never alone sufficient to bring about an action to help. Whether such an action is forthcoming depends in part on who the person is who sees the child in danger, or what kind of person learns of the distress of his friend. But for those who do regard what they find to be true of the other person as reason to act, and who act accordingly, nothing more is needed to explain why they do what they do. They see what they take to be a reason, and straightaway they act. Here, I think, we can expect the complaint of a kind of “transcendentalism” or “transcendence-ism” to reappear. It will be objected that there is nothing in the facts described so far that explains why or how the helpful agents are moved to act as they do. It will be said that “recognizing a reason” is not enough, since there is no such fact in the world, it will be said, as one thing’s being a reason to act in a certain way independently of the desires and interests of the agents concerned. It will be felt that more is needed to explain how the agent comes to take something like danger or distress for others as a reason to act in a certain way. It will be felt not to be enough to say that the agent takes the danger to the child as a “reason” because he regards a child’s death as something “bad,” “horrible” or “one of the worst things that can happen.” Those are, admittedly, evaluative terms, and there is felt to remain a question of how an agent’s understanding and applying those evaluative terms in the way he does actually issues in the actions he performs when he thinks of things in those ways. There is a feeling that it has not yet been explained what leads the agent to act as he does, or what he is relying on. I simply observe, by way of conclusion, that here we seem to be confronted with a charge of unexplained mystery, obfuscation or “transcendentalism” of the very kind raised against purely truth-conditional theories of meaning for a language. The complaint was that such theories give no account of how a speaker comes to know the things he knows by being a

Ways of Meaning and Knowing Moral Realities  11 speaker of that language. Theories of meaning in terms of truth-conditions alone do not explain how the understanding of the language they ascribe to speakers issues in speakers’ actual use of their terms. Those theories were said to give no account of the procedures or conditions on the basis of which speakers apply an expression to an item or a situation, or what speakers rely on in applying one expression in one situation and a different expression in another. And to that extent they were thought to leave speakers’ and agents’ knowledge and understanding unexplained. I think there is a short and correct, but unfortunately almost universally unsatisfying, reply to this demand for an explanation of speakers’ and agents’ competence. Speakers come to know what they know about the meanings of the expressions of their language by learning that language and becoming experts in saying and responding to things expressed in its terms. They learn how to converse intelligibly and to assess reasons for saying and doing the things they do and say. All this does not come to them overnight, and it does not come easily, although it does come, on the whole, painlessly and without much effort. That is because we acquire the capacities we do only over time and in a culture surrounded by others who already have those capacities. We learn from others, and learn to act accordingly. We who learn English come to know that the English sentence “Ripe tomatoes are red” is true if and only if ripe tomatoes are red. Of course, we could not know that without coming to understand the words “ripe,” “tomatoes,” “red” and so on. What we have learned about English is something we can express in the very words we have come to understand. For many of us, those are the only words we have. And in learning to act and to speak as we do, we learn to recognize and to follow good reasons for doing or saying such-and-such. We come to understand what many people at least regard as good reasons to do or say things, and we eventually become sufficiently competent with the idea to make independent critical assessments ourselves of whether what are thought to be good reasons really are good reasons after all. As speakers of English we know that the sentence “Someone’s being in danger of drowning is reason to help” is true if and only if someone’s being in danger of drowning is reason to help. To have the linguistic knowledge expressed in that biconditional statement we must have shown ourselves competent with the words that appear in it. There is no independent access to the understanding we have of those evaluative words beyond our having come to understand them as they are used in our culture. In understanding them we use them appropriately in acting and responding in all the ways we do. But what we have learned in learning the meanings of those words is something we can express only in some of the very words we have come to understand. We understand ourselves and others as prepared to act in ways we can see there is reason to act, and as sometimes acting accordingly. There is no prospect of reducing all those evaluative ways of thinking of ourselves to something fully non-evaluative, so we have no way of saying

12  Barry Stroud how we understand those ways of speaking and acting except by using some of those very evaluative words of which we are masters. There is no neutral, non-evaluative way of expressing what we know by being masters of evaluative thought and language. There is no way of expressing what we know in understanding the words of our language except by using some of those very words we have come to understand.

Notes 1 Mark Platts, Ways of Meaning (London: Routledge, 1979), xi (page numbers with WM in parentheses in the text refer to this book). 2 Both Wiggins and McDowell were developing versions of what Wiggins first called “a sensible subjectivism” and then, after further reflection, “a sensible subjectivism?” See “A Sensible Subjectivism?” in Needs, Values, Truth, ed. D. Wiggins (Oxford: Oxford University Press, 1998). 3 Mark Platts, Moral Realities (London: Routledge, 1991) (page numbers in parentheses with MR in the text refer to this book). 4 For what it is worth, I would suggest that if this complaint must have a name, “transcendence,” or even “transcendence-ism” would be better. “Transcendentalism” already means something else. 5 D. Hume, A Treatise of Human Nature, ed. L. A. Selby-Bigge (Oxford: Oxford University Press, 1958), 469. 6 D. Hume, Enquiries Concerning Human Understanding and Concerning the Principles of Morals, ed. L. A. Selby-Bigge (Oxford: Oxford University Press, 1966), 290. 7 Hume, Treatise, 457.

2 Platts on Kant and Mandeville Ralph Walker

Mark Platts defended moral realism at a time when that was quite unfashionable.1 His account of morality, and of how it impinges on our lives, is subtle and suggestive in a variety of ways. It is grounded in a careful examination of motivation. Amongst other things, he repudiates the common view that all forms of motivation are sub-species of desire, and argues instead that we can be motivated by moral facts, and by evaluative facts more generally. He considers and assesses the views of many people, but it is what he has to say about Kant and Mandeville with which I shall be concerned. Mandeville he treats at some length in his book Moral Realities, as one of morality’s critics. I think he is right to disagree with Mandeville, but Mandeville is less of a threat to his position than he believes. I also think that Kant’s position is closer to his own than may at first appear. In many respects Kant is on his side, and is a valuable ally for him, though Platts is much more subtle than Kant is on the phenomenology of moral choice. Platts sees Kant’s ethics as an ethics of rules, and rejects it on that basis; I shall argue that it is wrong to think of it in that way, even though Kant sometimes thinks of it in that way himself. Above all, I shall argue that both Platts and Kant are committed to a morality of objective imperatives. There are differences between them, but the differences are not as great as they seem. Platts regards morality as objective: objective in the sense that it obtains in its own right, independently of what anyone thinks or feels about it. Now the question of what an objective morality would be like has recently been discussed by Derek Parfit, and Parfit sees the requirement of objectivity differently.2 He thinks an objective morality must be grounded in what he calls “object-given” reasons, reasons which not only obtain in their own right, independently of what anyone thinks or feels about them, but are also such that the motivation of agents has no necessary dependence upon them. Thus for Parfit the objective requirements of morality cannot have the characteristic that Kant assigns them, of necessarily motivating anyone who is aware of them.3 However, Parfit has been vigorously attacked by Simon Blackburn. Blackburn says he cannot give an adequate account of moral motivation, because he cannot explain why these objective reasons “should matter to

14  Ralph Walker anyone.” No doubt they could be acted on by those who felt like acting on them, but they would be quite without interest to those who are differently inclined.4 Now Blackburn is not a moral realist at all: he does not think that moral requirements obtain in their own right, independently of what anyone thinks or feels about them.5 But one of his reasons for rejecting moral realism is that so many moral realists ignore the imperative force of morality. They claim, as Parfit does, that moral facts hold independently of us, but they leave it puzzling why they should so affect our attitudes and our actions. That seems entirely inadequate to morality as we ordinarily conceive of it. Blackburn must be right to say some objective entity, whether it is called a reason or a norm or whatever, cannot be what ethical theory needs unless it can link directly with the motivation of agents. The need for the link is something Platts has always recognised. Kant recognises it too, in holding that moral requirements necessarily motivate everyone who is aware of them at all. In recent years moral philosophers have talked a lot about reasons, and about what things are rational, or reasonable. Words like “reason” and “rational” can easily fudge the question of whether this link obtains or not. People who talk about rational requirements, or things that it would be reasonable to want, are usually implying that these are things that everyone ought to want, because everyone ought to want to be rational or reasonable: to want otherwise is to be irrational. But then we need an account of this “ought.” As Platts says, “the claim of irrationality is in general nothing but bluff.”6 Often the thought seems just to be that intelligent decent people would want these things if they reflected carefully. Rawls is unusual in that he explicitly recognises that the standards of intelligence and decency are relative to a particular range of places and times.7 But that of course means that what is rational or reasonable by these standards cannot claim real objectivity; it is not independent of what people think or feel about it. The compelling character of morality is captured by the idea that moral judgements are prescriptive, in the sense that anyone who assents to them is motivated by them: motivated to do what is good or right, and to avoid what is wrong. R. M. Hare thought moral judgements were very strongly prescriptive indeed: he held that one cannot sincerely assent to a moral judgement without acting on it in the appropriate circumstances. That of course leaves no room for weakness of will.8 Most people who think moral judgements are prescriptive do not go as far as Hare, and hold only that anyone who assents to a moral judgement is thereby aware of being motivated to act accordingly—but not necessarily so strongly motivated that they will actually perform the act. Countervailing motivations may be more compelling. Hare thought that because moral judgements are prescriptive, they cannot be objective in the sense of being true or false independently of what anyone thinks or feels about them. On the contrary, they must be expressions of attitude, much as Blackburn and arguably Hume take them to be.

Platts on Kant and Mandeville  15 But that leaves open the possibility that both Platts and Kant adopt. They both hold that there are moral facts, to which true moral judgements correspond, and which differ from ordinary facts in that one cannot be aware of them without being at least to some extent motivated by them. They use different terminology to express this, but they both take moral judgements to be prescriptive—not in Hare’s very strong sense, but in the weaker sense that sincere assent to a moral judgement involves being motivated to act in that way, though one may not actually do so. Platts writes about “the potential motivation internal to perception of moral value.”9 He also holds that moral judgements are true or false, in that they correspond or fail to correspond to a moral reality which is itself objective—objective in the sense of being independent of what anyone ­ thinks or feels about it. This moral reality is a reality of moral facts, but unlike Parfit’s “moral facts” these are motivating facts: facts which by their nature motivate whoever is aware of them. Platts holds that there are also motivating facts of other kinds—aesthetic ones for instance—but I shall not discuss those, and shall concentrate on moral facts here.10 Moral facts are genuine facts, being independent of our beliefs or feelings about them, but to be aware of them is to have a motivation to act accordingly, without the need for any intermediate desire, want or inclination. A motivation, though not of course a motivation that one will inevitably act on. Kant puts it rather differently, in terms of a moral law rather than moral facts, but his moral law is both objective and prescriptive, independent in its own right, given to our consciousness as binding. It is thus a distinctive feature of reality, different in kind from substances, properties, electric fields, quarks and the like. It exists independently of what we, or anyone else, feels or thinks about it; but it has the capacity to influence us, and all rational beings, by motivating us, though again, not in such a way as to make the relevant action inevitable. Because this law determines specific moral truths, e.g., that I should pay back what I owe, they inherit its prescriptive character.11 Since both Platts and Kant regard morality as objective and prescriptive, Platts as well as Kant must be committed to holding that an action has a special kind of value if it is done because the agent sees that the situation morally requires it. Kant puts this by saying that only such actions have “genuine moral worth.” That is rather a provocative way of putting it, because it can hide the fact that Kant places great importance on ensuring that people’s actions are at least in conformity with what morality requires, even if not done because morality requires it. The whole of that part of moral philosophy that he calls “the theory of right” is concerned with this.12 Platts too must accept that there is a distinction to be drawn between actions done in response to the demands of morality, and actions which although motivated entirely differently are nevertheless in line with those demands.13 Otherwise there would be no room for morality to guide our conduct, and its prescriptivity would be without effect. Kant is sometimes criticised for

16  Ralph Walker going too far here by saying that an action has “genuine moral worth” only if it is done for the sake of duty alone, with no other concomitant motive. But he did not mean that, as Barbara Herman and others have shown.14 It must be possible to act in response to the demands of morality, while at the same time having other, non-moral motivations for the action. I am not sure how far we can take Platts’s suggestion that the awareness of the moral facts may lead the agent to “hunt for other motivating considerations to back up” the moral motivation;15 but he must be right to see the moral motive as operating much of the time as a second-order motive, telling us which of our first-order motives we should act on in the current circumstances. Someone may have strong feelings of loyalty to her friends, and may often act on these feelings without conscious reflection of whether it is right to do so; but it may also be true of her that if a circumstance arose in which being loyal would require serious injustice, the feelings of loyalty would be overridden by the demands of justice. It seems to me that this is how Kant sees the moral law as functioning to guide our conduct much of the time. He is not so very different from Aristotle on this. For Aristotle phronésis determines when we should act or not act in accordance with our inclinations— phronésis being the practical wisdom that discerns moral truth.16 Kant does not picture the ideal human being as one who lacks ordinary desires and inclinations, but as one who has them trained in such a way that they will be acted on when and only when the moral motive, in its second-order role, does not oppose them—though it is true that this more human side of Kant is much clearer in works like the Anthropologie than in the Grundlegung. I imagine that Platts would think in much the same way. This relates to Platts’s concern about Mandeville. Mandeville finds the origin of morality in the attempts by “skilful politicians” to exercise control over other people, by exploiting the basic human attributes of self-love and self-liking. What he means by “self-liking” is the desire to think well of oneself and to be thought well of by others.17 Platts does not take very seriously the idea (which may or may not have been intended by Mandeville) that there is nothing to morality but a social construction designed by rulers in their own interest and sustained by the incorrigible vanity of the rest of us, with the help of our capacity for endless self-deception. Nor does he have much more sympathy for the better written, but less coherent, attacks on morality by Nietzsche. But he sees Mandeville as posing a problem that does need to be taken seriously. For if all human behaviour were ultimately driven by self-love and self-liking, or indeed any other combination of basically selfish human instincts, then acting from a moral motive—Kant’s sense of duty, or Platts’s awareness of moral facts—would be impossible for us.18 If morality could never motivate our actions, then it could hardly be the case that there is an objective morality that necessarily motivates everyone who is aware of it. It seems to be for this reason that Platts attaches particular importance to the fact that the moral motive can function as a secondorder motive: even if all our first-order motives have a non-moral character,

Platts on Kant and Mandeville  17 the moral motive can still enable us to choose the right first-order motive on which to act. But even so, he remains concerned that people are very unlikely to act in this way, so that morality is in danger of slipping beyond the reach of ordinary human beings.19 Why should it be unlikely? The worry is quite unjustified. In the first place, there is no good reason to accept Mandeville’s pessimistic moral psychology, or anything like it. People are capable of doing things because they see that the moral facts require them to do so, and indeed much of Platts’s work has helped to demonstrate this. It is true, of course, that we often deceive ourselves about our motives.20 But that is no reason for thinking that we always do. In the second place, however, even if some non-moral motivation were always involved in prompting action, morality could still play its part as a second-order motive. It is only because he has an implausible picture of how the second-order motive works that Platts thinks it unlikely to do this in practice. As he sees it, the second-order motive prompts the agent to “hunt for other [. . .] motivating considerations recognition of which would in the circumstances be sufficient to produce a first-level desire” which would lead to the morally appropriate action.21 That might indeed be difficult, and susceptible of going wrong in various ways. On the Aristotelian view of things, however, it is much more straightforward and natural. The morally-minded person may have strong first-order desires to do the wrong thing on some occasion, but behind them lies a propensity to subordinate such first-order desires to second-order scrutiny. This second-order scrutiny is what enables her to see that on this occasion this first-order desire is not to be acted on. Equally it enables her to see that on another occasion her first-order desire can properly be acted on, or that it is morally important that it should be acted on. There is nothing implausible in this; it fits how many people do act, much of the time. Mandeville may be correct to think that when we do the right thing, it is often partly because we want to be thought well of by ourselves and by others; but this is just the kind of first-order motive that may be controlled, and often is controlled, by a sensitivity to the moral facts, so that one acts on it whenever and only when the demands of morality allow it. The desire to live up to a certain image of oneself may lead us astray at times, but it can also be recruited by phronésis, or by its Kantian near-equivalent, the sense of duty. So Kant and Aristotle can help Platts against Mandeville, and we have seen that Platts shares with Kant the view that there is an objective morality which is prescriptive in character. But Kant sees it as taking the form of a moral law, which is encapsulated in the Categorical Imperative. Platts, on the other hand, does not think morality can be encapsulated in a general and all-inclusive way under a single Moral Law. It is not clear to me that his difference from Kant here is really as great as it seems, but it is certainly clear that he pays a great deal more attention than Kant does to the complexities of the particular moral situations that one meets in real life. In attaching the importance he does to the specificity of particular situations, he can

18  Ralph Walker seem to be anticipating the moral particularism of Jonathan Dancy.22 He accepts that two cases that were exactly similar qualitatively would have to be judged in exactly the same way morally (as Dancy also does), but is acutely aware of the importance of sensitivity in moral judgement and of the danger of simplistic rules. His view is that the detailed non-moral facts of the case give rise to a moral fact about it, a fact that is prescriptive in nature but which may itself be highly specific to the case concerned. Clearly moral situations can be very complex, in a way that makes moral decisions intensely difficult. It is also evident that some moral theorists, including Kant at times, have not taken this seriously enough, and sought to use oversimple rules or principles to cut elaborate knots. This is most obviously true in Kant’s unfortunate essay “On a Supposed Right to Lie from Benevolent Motives,” where he denies that lying can ever be right in any circumstances.23 But he did not always understand himself in this way, and there is no good reason why he should have done. Kant’s Categorical Imperative requires that we should promote a state of affairs which he calls the highest good: a condition in which virtue and happiness are jointly and fairly maximized.24 This “highest good” is to be promoted by treating all rational beings as ends in themselves and not privileging one’s own case or anyone else’s. His formal requirement, to act in such a way that the maxim of your action could become through your will a universal law, means that what is right for you in the particular circumstances in which you find yourself must be right for everyone else in the same circumstances. But there is room for considerable detail in what counts as “the same circumstances,” and Kant only occasionally remembers to call our attention to this. He is often thought to condemn, quite generally, the making of a promise one does not intend to keep, but in fact he is careful to say that what he condemns is doing this in a certain kind of case—where one believes oneself in really serious difficulty, like being badly short of money.25 He is often thought to condemn suicide too, but in the Metaphysics of Morals he raises as a “casuistical question”—meaning, a difficult one that needs to be thought about—whether “a great king who died recently” did wrong when he carried poison with him in going into battle, in order to avoid the harm to the state that would result if he were captured.26 Since the king in question was Frederick the Great, whom Kant admired immensely, there can be little doubt of his own view in this instance. Thus Kant has no need to deny, and sometimes clearly affirms, that knowing what is right in particular cases requires great sensitivity to their details. So rules like “never lie” and “never kill yourself” can only be rough rules, and may be misleading in hard cases. Some of the time at least, Kant is quite clear about this; and to that extent his position would seem to be not so distant from that of Platts. But he thinks of moral sensitivity as primarily an awareness of the non-moral facts, together with an ability to see which of them are morally relevant and which of them are not. Platts would disagree with that, I think. For he considers it important that there be a special moral

Platts on Kant and Mandeville  19 intuition in each case, as Aristotle also does, when he speaks of moral apprehension as a kind of perception.27 For Kant, in contrast, what it requires is really the capacity to tell whether this or that feature of the situation bears on the promotion of perfection and happiness, with no discrimination in favour of one individual (such as oneself) or against another. On the other hand, Kant does fully recognise that it can feel as though we have direct moral insights about different kinds of case, because very few of us have before our minds an articulated version of the moral law, any more than we have clearly before our minds a codification of the principles of inductive inference. What we really have is an inchoate knowledge of these things, which we apply in our understanding of particular cases; we can get clearer about what the moral law requires by thinking about what our moral attitudes imply. But here again the difference may be less than it seems. For Platts, a special intuition is needed in each case; but it would be difficult to hold that each intuition is entirely special, to the extent of having no discernible relation to any other. Moral intuition is often thought of as analogous to perception. When we perceive physical states of affairs, we perceive them individually, and we do not appear to need anything like an overall law to guide us in this. But if our perceptual claims are called into doubt, we do need some way of supporting them or correcting them. If moral intuition is to be like perception, the intuitionist will likewise want to be able to support or correct her initial impressions. Otherwise she would have no standard of objectivity. We should be on very weak epistemic grounds in claiming objective accuracy for our perceptual reports if we had no way of checking them against one another. Those who claim objectivity for their moral intuitions need some similar support if their claim is to be plausible. If a doubt is raised about whether a perception of something is really objective, rather than an illusion, we have to see whether it fits in with the most coherent and unitary account of things. That means whether it fits with the various things we know or believe about the external world and how it works. If we think we perceive a man carrying his head under his arm, we will question whether we have got things right. I think Kant would have no real disagreement with a moral intuitionist who held something like this about moral perception. He would point out that if each moral intuition could only be considered on its own, in isolation from all the rest, there would be no good ground for thinking it represented an objective reality; but if there were a coherence among moral intuitions the case would be stronger, and the more coherence the stronger the case. Now coherence can only be detected if cases can be compared with one another and brought under generalisations of some sort. If each maximally specific type of case had to be treated as entirely distinct from all others, we could not get very far. What enables us to find coherence, or the lack of it, is the grouping of such maximally specific types of case into higher types, as our expectations for where people wear their heads are grouped into a type; to bring them

20  Ralph Walker under generalisations, and—ideally—those generalisations under higher generalisations still. Kant would say, of course, that he has set out the highest generalisation that morality requires, in his formulation of the Categorical Imperative. He would also say that it is not just a generalisation: it is a synthetic a priori truth. That may seem incompatible with our reaching it by a sort of inductive process, generalising from our intuitive assessments of individual cases. But it is not really so. What has to be a priori is the apprehension of moral truth, which cannot be given by experience. Ordinary moral thinking reflects our inchoate grasp of this a priori moral truth, and it is this that reflects the special intuitions we have about particular cases. By reflecting on what we feel about particular cases we come to be able to appreciate more and more clearly the a priori truth in its most general form, a truth that we have previously been able to appreciate only in a dim and unsystematic way. Kant’s case for saying that the Moral Law takes the form he ascribes to it is based on the claim that this is what ordinary moral thinking really amounts to, when people reflect on it properly and render their moral thinking as coherent as possible.28 So he entirely agrees that our moral awareness starts in practice from the intuition that this situation is wrong, and that that person is good. His position requires it. The intuitionist might still object that the search for generalisation is mistaken, pointing out that the argument I have just given is an argument about our grounds for thinking moral intuitions to be objectively correct. They might just be correct, without grounds. In that case each type of nonmoral situation that has any moral implications at all would give rise, quite separately, to its own prescriptive moral fact (to the effect that we should do this, abstain from that, in these circumstances). But this would make it impossible to see how we could build up moral experience—something to which Platts rightly attaches great importance. The building up of moral experience depends on the fact that new types of case will resemble old ones in useful ways; it depends on there being generalisations of some sort linking different types of case. It may be said that sensitivity increases with practice, as our capacity to taste, say, the tannins in a wine may be enhanced with practice. But the example does not work. For with the wine, one is not learning a way of finding sensory input that was not previously present, but learning to discriminate effectively between elements that one had previously been aware of mixed together. Kant will accept the parallel of that for the moral case: we learn to distinguish the descriptive elements in a complex moral situation, and to see which are particularly relevant. But one can do this only by learning to separate out clearly the non-moral facts and see how they relate to one another; and then by evaluating them correctly, with one’s skill in such evaluation increasing as one builds up experience of previous applications of the moral law in difficult situations. Intuitionists need not agree with Kant that there is one single moral law, to be applied in every circumstance. But they do have at least to recognise that the moral evaluation of one highly specific type of situation cannot

Platts on Kant and Mandeville  21 be wholly distinct from the evaluation of all others. Our minds must have some capacity for telling which cases are morally similar to which; just as, in ordinary perception, they must have some capacity for telling which cases are similar in respect of colour or of shape. This capacity for seeing similarities amongst things, and thus for seeking unity amidst the diversity of experience, is fundamental to Reason as Kant thinks of it, since (as he says) “without it we should have no reason at all, and without reason no coherent employment of the understanding.”29 That is why he calls the moral law “rational,” and why he says that practical and theoretical reason are the same at root.30 Nothing mysterious or threatening is implied by the use of the word “reason” here. Kant’s own formulations of the Categorical Imperative give us guidance on how to see moral similarities between cases. The initial formulation— “Act only on that maxim through which you can at the same time will that it should become a universal law”31—says that we should be prepared to treat like cases alike; his first strengthening of it is to the effect that treating like cases alike at least means recognising that there is no moral distinction between oneself and others. If I treat someone in a certain way, then, I should accept being treated in that way myself in the same situation. His second and more explicit strengthening is in saying that one ought always to treat everyone, including oneself, as an end and not just as a means. As it stands this is rather a vague idea, but at least it rules out saying that I am not really in a different situation from someone else just because that other person is a slave, who exists to fulfil the needs of the owners. In due course Kant fills it out more fully, in the form of the requirement to promote one’s own perfection and the happiness of others, the union of perfection and happiness being the ultimate end prescribed for us by the moral law.32 The intuitionist might well disagree with Kant over these things, and might well have more than one standard for confirming or correcting the individual intuitions one has about particular cases; perhaps different standards for what can naturally be thought of as different areas of moral concern. Thus there might be a standard for justice and a different standard for loyalty. This would certainly be a possible account of things. Kant may be wrong in thinking that there is a single moral principle to be sought; of course he is far from the only philosopher to have thought that, but that may only show the propensity of philosophers to oversimplify. Of course, if there were several different equally fundamental moral principles, they could conflict, and we should have moral dilemmas that were irresoluble in principle. Platts welcomes this, and thinks we should abandon the “false hope” for a unitary overriding law.33 Here I am more inclined towards Kant, and for a Kantian reason. Moral reality could be pluralistic in the way Platts envisages, but reason—in that perpetual search for systematic unity which drives all our investigations—should at least make us hesitant to believe that it is. Beyond this, though, it is hard to argue, unless one can come up with a single all-embracing formula that proves satisfactory.

22  Ralph Walker Such conflicting principles would still be imperatives. Since the moral facts are prescriptive facts, they motivate action and command it. So they are effectively legislative in character, even if there is more than one law— perhaps a law of justice and a law of loyalty, which may potentially conflict. They are also categorical imperatives in Kant’s sense, just as much as Kant’s unitary moral law would be. They are categorical because they tell us what we ought to do in the relevant circumstances; what we ought to do just because those circumstances obtain. They differ from hypothetical imperatives, which tell us what we should do if we are to achieve some objective that we may or may not have. Categorical imperatives apply to us, and motivate us, in their own right, not as providing us with the means to some further end we may have.34 If Platts is right, morality is a matter of categorical imperatives, several distinct categorical imperatives, and that means there is room for conflicts between them that are irresoluble in principle. He may be right that there are several. But there cannot be indefinitely many, if we are to have plausible grounds for thinking them to be objective. Some people have thought it odd that a categorical imperative should exist in its own right. Mackie said such a thing would be “queer,” but Platts rightly observes that “the world is a queer place” and that this is no argument.35 In any case it is just not true that an objective moral law—or moral laws—would be unlike anything else. Kant says that there is “only a single categorical imperative,”36 but he must mean there is only a single categorical imperative in morals, for the principles of theoretical reasoning are categorical imperatives too. The basic principles of logical and inductive reasoning are objective: their rightness and their reliability are not the products of human thinking, nor do they simply reflect attitudes that we happen to have. I shall return to them. There is also at least one other categorical imperative of practical reasoning, the one Kant formulates as “Whoever wills the end, wills also the means which are indispensably necessary and in his power.”37 Kant sees this as the principle that lies behind hypothetical imperatives, imperatives that tell us what we should do if we are to achieve a given goal. But it is not a hypothetical imperative itself. It is a basic principle of practical reasoning. Though Kant calls it “analytic,” he does not mean that it is part of what it is to want a particular end that one should will the necessary means; on the contrary, he repeatedly says that this is required if one is guided by reason. One might not be guided by reason. Indeed it is intrinsic to Kant’s idea of an imperative not only that imperatives tell us the rational course to follow, but also that they do not force us to adopt it: whether we do or not is up to us.38 One might want the end, and see that the necessary means are in one’s power, and yet do nothing about it. This might be because there is something else one wants to do instead. I might very much want to finish writing a book, and be well aware that this requires persistent hard work, yet allow myself to be distracted by other things. It may seem less obvious that the principles of logical and inductive reasoning are categorical imperatives, or imperatives at all. It is true that they

Platts on Kant and Mandeville  23 do not motivate us to do things, if we do not count as doing things when we accept the conclusions of inferences, but that is only to say that they belong to theoretical rather than to practical reason. They do motivate us to accept conclusions. And “motivate” here is the right word, for they do not compel us to accept the conclusions that are validly drawn from premises we accept, any more than the moral law compels us to perform the right actions. Setting aside simple mistakes, which are possible in any sphere, one can deliberately reject a logical or inductive conclusion. This is most obviously the case where one has some other motivation for not accepting that conclusion. There are many cases of mothers refusing to believe that their sons are dead, despite all the evidence. Calculations which give unwelcome results are often rejected with the vague thought that there must be something wrong somewhere. Kierkegaard and William James have no difficulty in accepting certain religious beliefs in full awareness that they go against the evidence, and people have accepted some religious doctrines on the grounds that they are contradictory.39 Such people are refusing to follow a rational imperative, which they see as a motivation they reject. In doing so they are of course being irrational, but they may not mind that. It is not that they lack the motive to be rational, but that they have other motives that countervail. Most of us spend a good deal of the time deceiving ourselves, refusing to take seriously the evidences of our own inadequacies. Here again we are refusing to follow a rational imperative of which we are at some level aware; and the experience is a very common one. Some people would say that even if these theoretical principles are imperatives they are not categorical imperatives, because they only affect us if we want the truth. That is also wrong. People like Kierkegaard reject reason precisely because they want to get to the truth, and believe reason will not deliver it; but the demands of reason remain, even if Kierkegaard allows them to be overruled in certain cases. Moreover, the idea that theoretical reason will get us to the truth has nothing more to back it up than theoretical reason itself—which tells us that this is what we ought to think. Logicians confidently tell us that logical principles always take us from truths to truths. But our only ground for thinking that they may not, on some occasions, take us from truths to falsehoods lies in our acceptance of the principles themselves. We can, it is true, test our deductive principles by seeing whether they give us the right results in experimental practice, but the inductive support they can gain in that way itself rests on our principles of inductive reasoning—and inductive evidence is naturally limited, especially with complex proofs. Could it not be the case that Gödel’s theorem is a valid deduction from true premises, yet actually false? How do we know? So there is nothing strange or unfamiliar about the idea of a categorical imperative, existing objectively as part of the furniture of the universe. And even if there were, it would be hard to find any reason why such a thing should not exist.

24  Ralph Walker Actually to prove that there is an objective moral truth may not be possible, and wisely Platts—so far as I know—makes no real attempt to show that there is. Consilience amongst people’s judgements may provide the best reason for accepting it, and Platts does an excellent job in showing the weakness of the various arguments that attempt to undermine this consilience or affirm some kind of moral relativism. What Kant says is that “all human insight is at an end as soon as we have arrived at basic powers or basic faculties,” and that no proof of the objectivity of the moral law is possible, because a proof must start from somewhere and there could be no more ultimate starting point.40 “Nothing is left but defence”—defence against those arguments that would seek to show that there cannot be an objective moral law.41 For the principles of logic and inductive inference, too, this is the most that can be said: we cannot establish their objective validity except in a circular way, and all we can do is to defend them (or try to) against anyone who questions it. So far as morality is concerned, this is something Platts has very effectively achieved.

Notes 1 Mark Platts, Ways of Meaning (London: Routledge & Kegan Paul, 1979; 2nd ed., Cambridge, MA: MIT Press, 1997), ch. 10; “Moral Reality and the End of Desire,” in Reference, Truth and Reality, ed. Mark Platts (London: Routledge & Kegan Paul, 1980); Mark Platts, Moral Realities: An Essay in Philosophical Psychology (London: Routledge, 1991). 2 Derek Parfit, On What Matters (Oxford: Oxford University Press, 2011), 2 vols. 3 Kant, Grundlegung zur Metaphysik der Sitten, Ak. IV:408. Here and in what follows references to Kant are to the edition published by the Deutsche Akademie der Wissenschaften (Berlin, 1902–), by volume and page. There is an exception: References to the Kritik der reinen Vernunft are in the standard A/B form, referring to the pagination of the first/second editions. 4 Simon Blackburn, Review of Parfit’s On What Matters, www.phil.cam.ac.uk/ ~swb24/reviews/Parfitfinal.htm. I am not sure this is entirely fair to Parfit, but his conception of an object-given reason, and his related conception of normativity, is puzzling. It does not help that he declares his fundamental concept of a reason to be primitive and unanalysable: Parfit, On What Matters, I: 31. 5 Simon Blackburn, Spreading the Word (Oxford: Oxford University Press, 1984), ch. 6; and Ruling Passions (Oxford: Oxford University Press, 1998). 6 Platts, Moral Realities, 171. 7 John Rawls, “Justice as Fairness: Political Not Metaphysical,” Philosophy and Public Affairs, vol. 14 (1985); reprinted in his Collected Papers (Cambridge, MA: Harvard University Press, 1999). 8 R. M. Hare, The Language of Morals (Oxford: Oxford University Press, 1952), and Freedom and Reason (Oxford: Oxford University Press, 1963), ch. 5. 9 Platts, “Moral Reality and the End of Desire,” 81. 10 Platts, Moral Realities, ch. 5. 11 See for instance Kant, Kritik der praktischen Vernunft, esp. Ak. V:19, 44–5, 47. The moral law is a law of pure practical reason, and by pure practical reason we cognise its obligatory character as an objective law. Cf. also the discussions of Achtung and Interesse in Grundlegung zur Metaphysik der Sitten, Ak. IV:401n, 460. 12 Kant, Metaphysik der Sitten, Rechtslehre, Ak. VI. See also A. W. Wood, Kantian Ethics (Cambridge: Cambridge University Press, 2008), ch. 2.

Platts on Kant and Mandeville  25 3 He does not seem always to recognise this: thus Moral Realities, 213–6, 223. 1 14 Barbara Herman, “On the Value of Acting from the Motive of Duty,” Philosophical Review, vol. 90 (1981), reprinted in her The Practice of Moral Judgment (Cambridge, MA: Harvard University Press, 1996). 15 Platts, Moral Realities, 214. 16 Aristotle, Nicomachean Ethics, II:6, 1106b36f. 17 Bernard Mandeville, The Fable of the Bees, ed. Philip Harth on the basis of the versions of 1714–25 (Harmondsworth, Pelican, 1970). 18 Platts, Moral Realities, 196, 223. 19 Platts, Moral Realities, 216, 223. 20 As Kant recognizes: Grundlegung, Abs. 2, Ak. IV:407. 21 Platts, Moral Realities, 216. 22 Jonathan Dancy, Moral Reasons (Oxford: Blackwell, 1993), ch. 4; “The Particularist’s Progress,” in Moral Particularism, eds. B. Hooker and M. Little (Oxford: Oxford University Press, 2000). 23 “Über ein vermeintes Recht . . .,” Ak. VIII:425ff. 24 Kritik der praktischen Vernunft, Ak. V:109ff.; “Über den Gemeinspruch . . .,” Ak. VIII:279. The centrality of this idea for Kant’s ethics is disputed by Wood, Kantian Ethics, ch. 15. This is not the place to go into details, but see Paul Guyer, “Ends of Reason and Ends of Nature,” in his Kant’s System of Nature and Freedom (New York: Oxford University Press, 2005). 25 Grundlegung, Ak. IV:402, 422. 26 Metaphysik der Sitten, Tugendlehre, Ak. VI:423. 27 Nicomachean Ethics, VI:11, 1143b5. 28 Grundlegung, Ak IV:403ff., 445. 29 Kritik der reinen Vernunft, A 651/B 679. 30 Kritik der praktischen Vernunft, Ak. V:121. 31 Grundlegung, Ak. IV:421. 32 Metaphysik der Sitten, Tugendlehre, Einleitung, Ak. VI:385ff. 33 Ways of Meaning, 246. 34 Kant, Grundlegung, Ak. IV:414f. 35 J. L. Mackie, Ethics: Inventing Right and Wrong (Harmondsworth: Penguin, 1977), 38 ff.; Platts, “Moral Reality and the End of Desire,” 72. 36 Grundlegung, Ak. IV:421. 37 Grundlegung, Ak. IV:417. 38 Grundlegung, Ak. IV:413. 39 Søren Kierkegaard, Concluding Unscientific Postscript, trans. D. F. Swenson (Princeton: Princeton University Press, 1941); William James, “The Will to Believe,” from his book of the same title (New York: Longmans, 1897). 40 Kritik der praktischen Vernunft, Ak. V:46. 41 Grundlegung, Ak. IV:459.

3 Wrong Direction? A Criticism of Direction of Fit Gustavo Ortiz-Millán

1. Introduction Many contemporary philosophers have drawn on the metaphor of direction of fit to explain the difference between beliefs and desires. The basic difference between these two states, they claim, lies in the way they relate to the world: while beliefs try to fit the world, the world has to fit our desires, or in John Searle’s terminology, while beliefs have a “mind-to-world” direction, desires have a “world-to-mind” one.1 The distinction between these two mental states is important because it has applications in several fields: in theories of motivation it has been used to explain processes of practical rationality and the mechanisms of motivation that lead to intentional actions. For instance, it has traditionally been claimed that desires and beliefs are necessary elements in the motivation of actions given their different directions of fit; any action has to be explained as the product of a belief-desire pair, where the desire sets the goal of the action and the belief plays an instrumental role in achieving it. However, this distinction is not only important in discussions about practical rationality, but also in many discussions where cognitivist and non-cognitivist positions are at stake, or in debates about the ethics of beliefs, the analysis of knowledge and the doctrine of double effect.2 In other fields, such as the philosophy of language, the idea of direction of fit has been used, not so much to distinguish between mental states, but to differentiate speech acts: statements, descriptions and assertions have a “word-to-world” direction, since they are supposed to match the world, whereas orders, commands, requests and promises have a “world-to-word” one, “are supposed to bring about changes in the world, so that the world matches the propositional content of the speech act.”3 When used within the theory of speech acts, it also serves as the basis for social ontology.4 However, I am concerned here only with the distinction between beliefs and desires, and I am going to restrict my discussion to this issue—what I say about beliefs and desires might be extended to other debates, but this is something I am not going to explore here. In this paper I shall argue that the distinction between beliefs and desires in terms of their direction of fit carries several epistemological and

A Criticism of Direction of Fit  27 psychological commitments, resulting in a very theoretically loaded distinction, which may be difficult to accept from different theories in those fields where it is important to distinguish beliefs and desires. A distinction as central as this one should be as clean and neutral as possible, so that it can be used beyond the context of some particular theory—or in any case, anyone embracing it should be aware of the commitments it carries. Additionally, I shall argue that this distinction is not exhaustive: it is usually presented as a distinction between all kinds of cognitive and conative states (of which beliefs and desires, respectively, are taken to be the paradigmatic cases), but it fails to fulfill this promise, leaving aside a number of mental states that are usually grouped on one side or the other of this distinction. The idea of direction of fit is one among other ways of distinguishing between beliefs and desires, since these two mental states are different in ways other than their relation to the world. Although other criteria have been proposed, I shall argue for a different way of distinguishing them, one that, instead of emphasizing the relation of these mental states to the world, emphasizes the norms that govern these states (that govern, for instance, the possession, change and revision, of these states). Instead of what is a mostly descriptive distinction in the hands of some philosophers, I want to stress the normative differences of beliefs and desires. This proposal would give us not only a criterion for distinguishing these states, but also one that would help us telling these apart from other states that respond to different norms. Talk of these two states, under the so called “belief-desire model,”5 has tended to subsume other mental states, with more cognitive or conative features, under the broad categories of beliefs and desires, but we should be more careful when it comes to bunch states such as intentions, emotions, imagination and so on, on one side of this dichotomy or the other. Reductions of states such as these ones to the belief-desire pair tend to leave aside some of their essential features. This is what Michael Bratman has called the “strategy of extension,” that is, the strategy of extending talk about beliefs and desires to accounts of intentions, emotions or some other intentional states.6 However, my account does not go as far as for providing criteria for distinguishing between beliefs, desires and these other states, although, in principle, it is possible to do it in terms of the different norms that govern them. But in this point, as in probably some others, my approach is tentative and incomplete. In this paper, I start off putting forward some classic formulations of the metaphor of direction of fit (those of Anscombe and Platts) and I shall argue that the use of this notion commits us to a certain theory of truth. I question whether a distinction between beliefs and desires should commit us from the outset to a certain (or for that matter to any) theory of truth. In section 3, I present some doubts about the commitment this distinction has to a certain theory of motivation; once again, I wonder whether such a distinction should also commit us to some particular theory of motivation. In section 4, I argue that this metaphor does not help us to distinguish all

28  Gustavo Ortiz-Millán forms of beliefs and desires, but only certain forms of these mental states, showing us the limitations of the notion of direction of fit. Finally, in the last section I sketch a different way of distinguishing these states; I try to provide elements for what we might call a normative distinction of beliefs and desires, a distinction that makes no use of the metaphor of direction of fit, but anyhow takes the normative element already present in some of the interpretations of the metaphor. My basic claim is that, since normativity is an essential feature of intentionality, a fundamental difference between beliefs and desires lies in the different norms that govern these states. My contention is that the different values that govern beliefs and desires, the truth and the good, respectively, must play a central role in the explanation of the norms governing these states. A distinction made in these terms also results in a less compromised distinction than that of direction of fit— although probably not as neutral as one would like it to be. Additionally, although I may not develop this topic as much as I would like to do it here, it helps us distinguishing beliefs and desires from other intentional states.

2. A Metaphor’s Commitments G.E.M. Anscombe distinguished between beliefs and desires (or more precisely, beliefs and intentions) in terms of their direction to the world, although she does not use the phrase “direction of fit.” In a famous passage of her book Intention, she writes: Let us consider a man going round a town with a shopping list in his hand. Now it is clear that the relation of this list to the things he actually buys is one and the same whether his wife gave him the list or it is his own list; and that there is a different relation when a list is made by a detective following him about. [. . .] What then is the [. . .] relation to what happens, in the order and the intention, which is not shared by the record? It is precisely this: if the list and the things the man actually buys do not agree, and if this and this alone constitutes a mistake, then the mistake is not in the list but in the man’s performance (if his wife were to say: “Look, it says butter and you have bought margarine,” he would hardly reply: “What a mistake! We must put that right” and alter the word on the list to “margarine”); whereas if the detective’s record and what the man actually buys do not agree, then the mistake is in the record.7 Anscombe focuses on the place where the mistake lies. If we believe that p and it is not the case that p, then we have the wrong belief; whereas if we intend to do p and we do not do p, it is the action that is in fault, not the intention. Mark Platts takes up Anscombe’s distinction,8 using this time the notion of direction of fit and applying it to beliefs and desires: Miss Anscombe, in her work on intention, has drawn a broad distinction between two kinds of mental state, factual belief being the

A Criticism of Direction of Fit  29 prime exemplar of one kind and desire a prime exemplar of the other. The distinction is in terms of the direction of fit of mental states with the world. Beliefs aim at the true, and their being true is their fitting the world; falsity is a decisive failing in a belief, and false beliefs should be discarded; beliefs should be changed to fit with the world, not vice versa. Desires aim at realisation, and their realisation is the world fitting with them; the fact that the indicative content of a desire is not realised in the world is not yet a failing in the desire, and not yet any reason to discard the desire; the world, crudely, should be changed to fit with our desires, not vice versa.9 According to this passage, beliefs “aim at the true,” where this means fitting the world: if a belief does not fit the world, it is the belief that is at fault, not the world, and false beliefs should be discarded; if a belief does not fit the world, one can repair the situation by changing the false belief with a true one. On the other hand, desires are not susceptible of being true or false, they are not truth-apt, but they aim at their realization or satisfaction. If a desire can’t get no satisfaction, one cannot repair the situation just by changing desires—although that is what we, in our fickle will, often do. The fact that the indicative content of a desire fails to be realized is not a fault of the desire, but, so to speak, it is a fault of the world, and it is not a reason to change our desires, as Platts says, “the world, crudely, should be changed to fit with our desires, not vice versa.”10 Although it is not at all clear why the world should be changed to fit our desires—in many cases it is probably better for the world not to have our desires satisfied. Immediately after the above characterization, Platts admits one of the limitations of the metaphor: desires do not seem to be pure cases of a world-tomind direction of fit, since “all desires appear to involve elements of belief.” I take it that Platts means that desires are constituted (at least in part) by beliefs, i.e., beliefs are not just enabling conditions of desires. In this sense, desires are constituted by cognitive elements and by what we take to be true. This would imply that when we desire something, we have some beliefs about the object of desire that are involved in the desire. We cannot desire something about which we have no beliefs. Otherwise, this would probably bring us to a conception of desires as some kind of instincts or drives emerging in a completely atomistic and non-inferential way, separate from the rest of our mental life, of what we take to be true, and as irredeemably non-rational (the kind of conception of desires that we find in Hume, who held a corpuscular theory of desires, where we are told that desires have no influence of beliefs).11 However, we might agree that desires are not atoms that emerge out of the blue, in a completely non-inferential way, involving no other elements of our mental life, impacts from outside the realm of thought coming from our own bodily functions, but that they emerge in a background of other intentional states and involving beliefs. Desires involve at least beliefs about particular matters of fact regarding the object of desire, but also general truths and norms. The process of forming desires responds

30  Gustavo Ortiz-Millán not to an atomistic idea, but to one involving, for instance, acquiring habits of response to various objects of desire in diverse circumstances. Whereas beliefs may be taken as irreducible elements of our mental life, desires seem to be composite attitudes involving our beliefs. If all this is true, then we have broken the symmetry with which the distinction is usually presented; there is an asymmetry between beliefs and desires if desires involve elements of beliefs in a way in which beliefs usually do not involve elements of desire. Now, it may be a form of anthropomorphism to say, as Platts and many others do, that beliefs “aim” at the true and desires “aim” at their realization. Neither beliefs nor desires “aim” at anything: they are mental states with no intentionality—or at least not in that sense of the word, i.e., having intentions. This is why this claim must be interpreted as a metaphor. There are several aspects that may be criticized in Platts’s characterization, for instance, he does not tell us exactly why we should discard false beliefs, how should we understand the relationship of beliefs and truth or why the lack of realization of a desire is not a reason to discard the desire. However, these are minor faults of a characterization that is formulated in very broad lines; and for my present purposes it will be much more interesting to flesh out more precisely the key terms used here so metaphorically: “direction” and “fit.” With regard to the term “direction,” let me emphasize the fact that in Platts’s characterization, the direction of fit of beliefs and desires with the world is given, so to speak, in “vertical” terms. Beliefs are distinguished by their relationship with truth, and truth is defined as fitness to the world; desires are characterized by the direction they have with respect to their realization in the world, this has to fit their content and not vice versa. By this I mean that the direction that defines and distinguishes both of these states is the one they have in relation to the world, and not, for instance, the one they have to each other or to other mental states—that would be the level at which relations of rationality take place, and that we can think of as horizontal relations (but in this case, then, should we rather talk about a mind-to-mind direction of fit?).12 The term “fit” breeds even more doubts, given its epistemological and metaphysical connotations, especially when we talk about “fitting the world.” The characterization of beliefs in terms of their fitting the world seems to imply a commitment of this distinction to a correspondence theory of truth or to a certain form of representationalism, if we understand “fitting the world” as correspondence with facts (states of affairs, event, the world. . .) or as representing the world the way it really is. The fact that the characterizations of the notion of direction of fit tend to prioritize, in a vertical way, the relationship of beliefs and desires with the world, rather than horizontally, with other beliefs or mental states, is an indication of this. This correspondentism is even clearer in Philip Pettit’s description of what direction of fit is: The different roles of belief and desire are reflected nicely in the fact that what it means for a belief to be successful [. . .] is that the mind comes

A Criticism of Direction of Fit  31 to correspond to the world; whereas what it means for a desire to be successful [. . .] is that the world [. . .] comes to correspond to the mind. Belief has a world-to-mind direction of fit, desire a mind-to-world direction of fit.13 To say that beliefs aim at, or are directed towards, the truth is one thing, but to commit oneself to a certain definition of truth while distinguishing and characterizing beliefs is something else. Truth certainly depends on how the world is, but there does not seem to be any reason, in order to distinguish between beliefs and desires, for committing oneself to a certain definition of truth as correspondence to the world. With the notion of direction of fit defined in these terms, one comes to buy a distinction between beliefs and desires, and additionally gets an implicit definition of truth for the same price—call it a true philosophical bargain. There does not seem to be any reason of why a characterization of the distinguishing feature of beliefs has to commit us to a certain theory of truth. I do not see any argument in favor of this procedure. Probably not many people would question that it is on the right track to define beliefs in terms of their truth-directedness (whichever way we define truth), since it would be very hard to define the concept of belief without making any reference to the concept of truth, but why, in order to define beliefs, should we commit ourselves to a certain theory of truth? It might be argued that we could understand the notion of fitting in a different way and that the connection between direction of fit and a correspondence theory of truth is not essential. Perhaps we could understand fitting in terms of coherence, where “fit” is understood, horizontally, as coherence with other beliefs, for instance. But then we would have lost a good deal of the force of the metaphor: a belief would be fitting other beliefs and that would be what would make it true, not the world. Truth would be coherence among beliefs. But then the direction would have changed from the world to other beliefs. Should direction change correspondingly in the case of desires in order to keep the symmetry? At what exactly, if not the world, would they be directed? If we do not keep their world-directedness, but we shift to a coherentist sense, then we would have several options: we could say that desires fit other desires—for instance, desires fit higher-order desires, say, some second-order desires, although we might then be condemned to some kind of infinite-regress problem. On the other hand, one might think that desires would be fitting our beliefs. However, if this were their direction of fit, then, again, we would have lost a good part of their difference with respect to beliefs, which also adjust to other beliefs—and which, by the way, sometimes also adjust to our desires, like in cases of self-deception and wishful thinking; but this adjustment would only be causal (and casual), and what really matters here is how they rationally fit with those other mental states. In any case, we would have lost the supposed symmetry of the distinction at the price of some level of obscurity, or even more, we would have lost for good the notion of direction of fit, which seems tailor-made for correspondentism, and not for any other theory of truth.

32  Gustavo Ortiz-Millán Someone might claim that even if the notion is formulated in terms of correspondence, we could get rid of the idea of correspondence and then just say that the concept of truth that is implicit in the distinction gets closer to the Aristotelian concept of truth, according to which to say of what it is that it is and of what it is not that it is not is true, while to say of what it is that it is not and of what it is not that it is is false. Thus, we would have left aside the idea of correspondence. I take it that this move would amount to adopt some sort of minimalist or deflationary theory of truth, according to which there is no such thing as the “nature of truth,” and truth is not any kind of substantial property. But again, either in this case or in the one that favors some sort of coherentism, wouldn’t we be tying our distinction to a certain theory of truth without providing any argument for doing so? Insofar as I can see, there is no argument for the strategy of committing our distinction to a certain theory of truth, whichever it is. The question here is whether we can define and distinguish beliefs in terms of truth without committing ourselves to some particular view about it. Are these reasons enough to abandon the distinction between beliefs and desires in terms of their different direction of fit? Most likely not: none of these comments constitutes any conclusive argument against this notion. Many realists or those persuaded by correspondentism may not be uncomfortable at all with the implications of this kind of characterization.14 However, if we want a distinction that may go beyond the scope of a particular theory of truth, this should be a clear and neutral one and not commit us to more than what we need. If we cannot get such bases, then the reach of our distinction may be rather limited.

3. Direction of Fit and the Humean Theory of Motivation The metaphor of direction of fit does not only commit us from the very beginning to a certain theory of truth, but also to a certain theory of motivation: the Humean theory of motivation. This may also make us wonder whether a criterion for distinguishing between beliefs and desires should commit us from the outset to a certain theory of motivation. When one tries to use this notion in theories of motivation other than the Humean, then this may result in more problems for the theory than the ones it tries to face, as I shall try to argue. According to the Humean theory, actions are always the product of two distinct kinds of mental states: conative states, of which desires are a paradigmatic case, providing the essential and necessary motivational force for intentional action, and cognitive states, beliefs being their most representative example, which get into the picture of motivation only when standing in an appropriate relation to a desire. Desires must always be present in any process of practical deliberation, because they alone provide the motivational force for acting, and they are in fact the starting point in any such process.15 Beliefs have only an instrumental role in motivation; they have

A Criticism of Direction of Fit  33 nothing to do with determining the objects of desires, says Hume. A belief must always be conjoined with a desire in order to produce an intentional action: its role in motivation is to guide and to provide the necessary information to satisfy desires. Practical rationality, according to the Humean theory, is instrumental rationality. We can already find a primitive formulation of the idea of direction of fit in Hume, even though he did not use this terminology and neither talked about beliefs and desires, but rather about reason and passions—somehow antecedents of those notions. According to Hume, reason is a representation or a copy of the world, it aims at representing the world, at discovering “the real existence or the relations of objects,”16 and by itself it is motivationally inert. On the other hand, a “passion is an original existence, or, if you will, modification of existence, and contains not any representative quality, which renders it a copy of any other existence or modification” (2.3.3.5). Passions are neither representations nor copies of the world, and in that sense they are neither true nor false, but they are original existences that provide the motivational force that drives us to action; being modifications of existence, passions aim at modifying the world. Actions, being also modifications of existence themselves (changes in the world or events, we might say), reveal the presence of states with such a direction. This Humean thesis has been nicely synthesized by Platts, an anti-Humean, who tells us: “Any action (other than intentional omissions) is, and must be meant to be, a change in the world; but any such intended change requires the obtaining of a mental state with a ‘direction of fit’ of the kind distinctive of desire.”17 In this way, the Humean theory claims that the explanation of action always requires the reference to beliefs and desires, given their different directions of fit. Anti-Humeans (sometimes also called rationalists or motivational cognitivists) about practical reason, such as Platts, McDowell or Dancy,18 have traditionally challenged the Humean claim that the explanation of action always requires the reference to desires. These theories claim that some beliefs might be intrinsically motivational and enough for explaining certain actions; or in Kantian terms, that pure reason by itself can be practical in its issues, without the intervention of desires or “empirical inclinations.” That is, they challenge the claim that we always need the belief-desire pair to explain intentional actions: sometimes, a belief is more than enough. Some others, such as Humberstone,19 take the somewhat different strategy of claiming that some or all desires might be beliefs. The realization that I ought to do X, that Y would be the virtuous thing to do or some rational considerations, are capable, by themselves, to motivate an agent insofar as he is rational. In this way, anti-Humeans typically embrace a form of internalism about motivation, that is, the view that the acknowledgment of moral truths or moral judgments by themselves can be motivationally effective reasons for action. This is what Humeans deny: that something other than states with a world-to-mind direction of fit can motivate us to act.

34  Gustavo Ortiz-Millán However, instead of questioning the Humean distinction between beliefs and desires in terms of direction of fit, many anti-Humeans have rather reformulated the distinction, allowing for the possibility of motivating beliefs. Since what they want to argue is that beliefs by themselves are able to motivate desires and, ultimately, intentional actions, they usually claim that some beliefs (usually moral, and sometimes, prudential beliefs) not only have a mind-to-world direction, but also a world-to-mind one, that is, they have a double direction of fit. J.E.J. Altham has called this kind of states “besires,”20 i.e., beliefs that in addition to their natural cognitive role, also have the motivating power of desires—and they are at the heart of some motivational cognitivist positions.21 Of course, if they have both directions, then there are two ways for them to fail, since they can fail if they fail to fit the world, but also if the world fails to fit them. However, this move seems to leave us with several problems: 1) We might wonder whether, by introducing the idea of states with double direction of fit, the very distinction does not lose its force. What is the point of distinguishing between beliefs and desires in terms of their different direction of fit, if there is a whole lot of beliefs that do not fit this distinction and that possess the defining characteristics of desires? (Or it may fit the distinction if the directions are distinct properties of the same state of besire.) 2) The problem of having to account for the distinction between beliefs with double direction from those with single direction of fit. If it is only moral and prudential beliefs that have this feature, then some story would have to be told about the difference between these and other beliefs with evaluative or normative content; if it is normative beliefs more generally, then we need some other story about the difference between these and other non-normative beliefs, e.g., purely factual beliefs. (Wouldn’t this, by the way, rest on some distinction between facts and values that many of the same anti-Humeans deny?) 3) The problem of making desires somehow irrelevant or trivial in explanations of actions where beliefs with double direction of fit are involved. Some anti-Humeans claim that in some cases beliefs (i.e,. besires) are more than enough for the explanation of action, but some others continue talking about a commitment to ascribe desires even in cases where beliefs provide the motivational force; but what is the point, in these cases, of insisting to ascribe desires when beliefs with double direction are doing the motivational work?22 Besires have the motivational powers of desires (powers that beliefs with single direction apparently lack), which make desires somehow unnecessary or irrelevant in the explanation of some actions. 4) Some philosophers have taken the somewhat different, but radical, strategy of not getting rid of desires, but of reducing some or all of them to beliefs. Desires are just forms of beliefs, we are told. In some

A Criticism of Direction of Fit  35 versions of this theory, desiring is equated, for instance, with believing that something is desirable, and it is possible to construe desires as beliefs of a certain sort—beliefs about what is desirable or valuable, or beliefs about normative truths. The basic idea is that we can identify “Desiring X to be the case” with “Believing that [something like] it would be good if X were the case.” This “desire-as-belief” position, as it sometimes has been called, is held among others by Humberstone,23 for whom desires can be taken as species of beliefs and it is possible to paraphrase a desire statement in terms of a desiderative belief, with the clause “it is desirable that.” Very roughly, Humberstone shows that introducing an operator ‘D’ to represent this kind of clauses allows us to paraphrase a desire for α as the belief that ‘Dα’ (‘BDα’), that is, desiring α can be paraphrased as believing that α is desirable. This proposal has been criticized in several ways,24 but in relation to our issue, desires then would have the same direction of fit of beliefs—which would question the distinction in terms of direction of fit. 5) There are other problems with mental states having two directions of fit. Arto Laitinen has argued against such states, claiming that in cases of discrepancy between a besire and the world, both should change. Since there are two ways for besires to fail, if they fail in both ways, then both one’s besire should change and the world should change. But this, in its turn, would bring about another discrepancy, “and so both should change again, ad infinitum.”25 So, the best thing we can do is to drop the notion of a state with two directions of fit. These are some of the problems that seem to follow when trying to adjust the notion of direction of fit to motivational models other than the Humean theory. None of these may be an insurmountable problem and anti-Humeans have proposed a number of ways for facing them, but my contention is that they would not have to face these problems from the very outset if they provided a different way of distinguishing between beliefs and desires than that given first by the Humean. The use of the notion of direction of fit inclines the balance from the very beginning towards the Humean side. Here I want to introduce another way of presenting the direction of fit distinction, since it has been used to argue against the plausibility of antiHumean accounts of besires. Aware of the problems that arise with the talk of besires, Michael Smith denies that states with double direction of fit may even be possible. In the context of his defense of a Humean theory of motivation, Smith gives us a different characterization of the notion of direction of fit. He tells us that the difference between beliefs and desires in terms of direction of fit can be seen to amount to a difference in the functional roles of belief and desire. Very roughly, and simplifying somewhat, it amounts, inter alia, to a difference in the counterfactual dependence of a belief that p and a

36  Gustavo Ortiz-Millán desire that p on a perception with the content that not p: a belief that p tends to go out of existence in the presence of a perception with the content that not p, whereas a desire that p tends to endure, disposing the subject in that state to bring it about that p. Thus, we might say, attributions of beliefs and desires require that different kinds of counterfactuals are true of the subject to whom they are attributed. We might say that this is what a difference in their direction of fit is.26 The difference between beliefs and desires, then, comes down to their different counterfactual dependence on perception: whereas a belief that p tends to disappear in the presence of a perception with the content that not p, a desire that p tends to endure, and even disposes the subject to bring about p. Even if the content of the desire does not get realized, the desire tends to endure. Given this characterization, then it turns out that there cannot be a state with both directions of fit, as Smith says: A state with both directions of fit would therefore have to be such that, both, in the presence of such a perception it tends to go out of existence, and, in the presence of such a perception, it tends to endure, leading the subject who has it to bring it about that p. Taken quite literally, then, the idea that there may be a state having both directions of fit is just plain incoherent.27 As a mind-to-world state, it would have to go out of existence in the presence of a perception that not p, but as a world-to-mind state, it would tend to endure in front of such a perception. This is incoherent, and that is why there cannot be such a thing as a state with double direction of fit. Thus, Smith blocks the way for an anti-Humean argument for the possibility of motivating beliefs or besires given in the context of the notion of direction of fit, obliging the anti-Humean to look for a different way of arguing for this theory. It is not my purpose here to endorse Smith’s characterization of direction of fit, but to show how this notion can be used against the anti-Humean, particularly if he wants to continue playing with the same cards first put forward by the Humean. Smith’s characterization may be good at pointing out some problems already present in the struggle of anti-Humeans to provide a theory of motivation within the conceptual framework of the Humean theory and the notion of direction of fit. However, Smith’s characterization is controversial and reveals us other problems; some of them are specific of Smith’s formulation, some others go deeper into the general notion of direction of fit. Let’s see some of these problems.

4. A Rather Limited Distinction? Smith claims that the difference between beliefs and desires lies in the different counterfactual dependence they have on perception; a belief tends to

A Criticism of Direction of Fit  37 go out of existence in the presence of a perception with content contrary to that of the belief, whereas a desire tends to endure. However, we should analyze more carefully the relationship of these two states with perception. In the presence of some perception with content contrary to that of my desire, it is not always characteristic of desires that they tend to endure: in the face of such a perception, many of our desires tend to go out of existence. Endurance would be a characteristic of desires in the situation of somebody strong-willed or persistent enough to retain his desire in the face of some perception with content contrary to that of his desire; many people would rather change or completely drop the desire in the presence of a perception with content contrary to that of the desire—as Marcel Proust says in Remembrance of Things Past: “We do not succeed in changing things according to our desire, but gradually our desire changes,”28 and sometimes, he might have added, suddenly, our desire just disappears. In their criticism of Smith’s characterization, Sobel and Copp present us with the case of a woman who says she roots for the 49ers, but it has been noticed that the team she says she roots for changes frequently, in response to the successes or failures of the team. So she may desire that the 49ers do well, but if she sees that they are losing, this tends to drive this desire out of existence.29 On the other hand, sometimes instrumental desires also go out of existence with no intervening perception: I desire to get in my car to go meet a friend, until I remember that we are meeting at home and then my desire to get in my car disappears. It is also proper to ask what kind of states are perceptions. Discussing Smith’s characterization, Humberstone sees two options. The first one is that we understand perception as belief-like states, with the same direction of fit that beliefs have; but then the distinction between beliefs and desires would amount to how these two states relate to belief.30 The second option is that we understand “perception” as a state other than belief, as “mere appearances and seemings,” but then in this case the account is false. Many beliefs endure even in front of some perception with content contrary to that of my belief. This may be a common case to those familiar with the notion of illusion in epistemic contexts: someone might have the perception that the two lines in the Müller-Lyer illusion are different without believing that they are different, given that she might take her perception to be illusory. We know that some appearances may be misleading.31 There are some other cases of beliefs that do not tend to go out of existence in the presence of a perception with the content that not p: two cases that have already been used by Sobel and Copp32 in their discussion of Smith are (i) the case of stubborn beliefs, that is, cases where people continue holding a certain belief even after they have perceptual evidence against their belief, such as someone’s belief that it is not going to rain, even though he sees big storm dark clouds approaching. Cases of self-deception may also be cited: a woman may believe that her husband is faithful, even though it is evident that he is not; she may acknowledge that the appearances are

38  Gustavo Ortiz-Millán against her, but continue to be convinced about her claim. It may be argued that these are not counterexamples to Smith’s claim, since these are examples of imperfectly rational beliefs, but anyhow they show that beliefs do not always go out of existence in front of a perception with content contrary to that of the belief. However, let us grant that Smith is talking about tendencies and not about some kind of beliefs’ universal laws. But we may find a stronger case in (ii) beliefs in necessary truths. This is where the criticism of Smith’s theory seems to lead us to a more general criticism of the notion of direction of fit. The way in which Smith makes beliefs depend on perception refers us to the issue of whether all our beliefs depend on perception in the way he claims they do. It seems that it is only factual or empirical beliefs that are characterized within this distinction: this is something Smith seems to share with Platts, who talks only about factual beliefs as the “prime exemplars” of the kind belief. As it is formulated by both of them, the metaphor of the direction of fit, that purports to tell us the difference between desires and beliefs in general, seems to be limited to just one kind of beliefs. This distinction puts aside some other belief-like states or states we also call “beliefs,” such as a priori beliefs, i.e., beliefs we get independently of any empirical experience: beliefs about necessary truths, mathematical or logical beliefs. No appearance or perception can tend to undermine this kind of beliefs. Cognitive states other than factual beliefs may also have troubles passing Smith’s perception test, such as those beliefs whose content does not let them be empirically tested through perception: here I am thinking about beliefs whose content is not formulated in the indicative mode, such as the case of subjunctive conditionals; probability judgments may also be left aside—although I am aware that all this may be debatable, and it is not my intention to get into this discussion here. But we may also add doubts, states of supposing, considering, imagining, dreaming, among others people usually put together under the label “cognitive.”33 These states do not fit the world as factual beliefs do. This is also where the general notion of direction of fit, with its commitment to correspondentism, seems to get in trouble, since some of these beliefs do not seem to correspond with the world as factual beliefs do. If all this is correct, then the notion of direction of fit does not help us to distinguish the whole set of attitudes we call “beliefs,” but just a certain set. It is controversial that some of these beliefs be directed to truth in the way factual beliefs expressed in the indicative mode are. We must take into account that an extension of what distinguishes beliefs of other cognitive states may result in an illegitimate move. Something similar happens with desires, if we take them as the “prime exemplars” of a whole kind of mental states with similar direction of fit. As I have already mentioned, in this strategy of extension, intentions may be seen as having the same direction desires have, but it is controversial that intentions may be reduced to desires. However, with this I have only tried to point towards one of the limitations of the notion of direction of fit, as we find it in some of its classical formulations. If

A Criticism of Direction of Fit  39 this is supposed to be an exhaustive distinction between beliefs and desires, it is certainly not.

5. Sketching a Different Distinction Something else worth noticing of Smith’s characterization is that it is a purely descriptive interpretation of the difference between beliefs and desires: it tells us about the different functional roles that certain beliefs and desires actually play. They can be distinguished in descriptive terms: beliefs have such counterfactual dependence on perception, and desires, such other; they have different functional roles as well. To be sure, the difference between beliefs and desires can be made in these terms, but something that these characterizations miss from at least those of Anscombe and Platts is the normative interpretation of the distinction. For the latter, beliefs and desires are subject to different normative requirements: beliefs are at fault when they fail to track truth and they should be changed to fit the world; desires, on the other hand, aim at realization, but in this case it is the world that should be changed to fit our desires. However, none of these authors tells us much about the relation between beliefs and truth or desires and realization. “Beliefs aim at truth,” they say, but without telling us much about how this should be understood; and even less about how we should understand the claim that “desires aim at realization” or that the world should be changed to fit our desires. However, the relationship of beliefs with truth is not just something that happens: it is basically a normative relation. I shall argue that there is a similar normative relation in the case of desires, but not in their relation to realization, but to the good. However, notice that I am not trying, as Zangwill and others do, to give a normative interpretation of the notion of direction of fit. What I am trying to do, in any case, is to sketch an alternative view to that of direction of fit, one that stresses the normative differences between beliefs and desires, rather than their relation with the world. My contention is that such an account would not have the epistemological and psychological commitments that the notion of direction of fit has. How are we to understand the claim that “beliefs aim at truth”? There may be different ways of understanding this claim, but here I adhere to the interpretation that sees it as a normative claim.34 A concept is normative if it involves some kind of evaluation or some standard of correction; the right way to give an account of the nature of a normative property is as the property that makes a certain sort of mental state, belief in this case, correct or appropriate. We might say that it is in the nature of beliefs to be ruled by norms: by adjusting to these norms, they are correct or appropriate. The norm that defines when a belief is correct is the principle that a belief is correct if and only if the belief’s content is true. The concept of belief is inseparably linked to “truth” and “falsehood,” “appropriate” and “inappropriate” and other normative terms. To hold a belief is to be subject to

40  Gustavo Ortiz-Millán certain norms, the most important of which is that related to truth: truth and falsehood are a dimension of evaluation of beliefs. This norm applies to beliefs as such and not because they belong to a broader category—such as “cognitive.” It is an essential and distinctive feature of beliefs that they are subject to this norm: an intentional mental state is a belief if and only if its content is able to be assessed in terms of truth and falsehood. We assess our beliefs for their relation with truth (although not exclusively: truth and falsehood are one dimension, but not the only one, given that there are some other related dimensions of assessment, such as consistency, transitivity, justification, and so on), but we could say that this is the most significant element of evaluation of our beliefs and the one from which others depend. Other epistemic norms that apply to beliefs can be explained by this norm— the norm of consistency to which beliefs are subject, for instance, can be explained just because beliefs are subject to the previous requirement of truth. This feature of beliefs of being normatively linked to truth in such a way is characteristic of beliefs to the point of making it their distinguishing feature: beliefs may be distinguished from other intentional states that are not assessed by these standards. It is clear that we do not assess desires in terms of truth and falsehood—and this may count as a reason for resisting accounts that try to reduce desires to forms of beliefs. Desires do not aim at truth (unless we take them to aim at the truth of the proposition, as when I desire that something I believe is true, which is something different), and this is also why they are not subject to other normative criteria such as that of consistency. We do not require our desires to be consistent with each other in the same way as we do with our beliefs. This, it seems to me, is what was meant by Richard Wollheim, who, while criticizing the notion of direction of fit, pointed out to a different distinction between these two states when he claimed that what we require of our beliefs we require of them distributively and collectively, but what we require for our desires we require them only distributively. [. . .] Each desire offers us, as it were, a view through a keyhole, but there is no reason that there should emerge from these views a coherent picture of what lies on the far side of the gate or that, should it be thrown open, there would be revealed a garden that we could conceivably enter.35 The requirement of consistency that applies to beliefs is what makes that such a coherent picture emerges from them; it does not emerge from desires, since we can have inconsistent desires without this posing any problem to the agent holding those desires. The requirement of consistency imposes some restrictions to any rational believer. If someone recognizes that what she has believed up to that point is false, the belief is in fault and she is in a turning point for changing her mind. If the agent is rational, she thereby

A Criticism of Direction of Fit  41 would abandon the false belief and adopt the new belief that she takes to be true; there is some psychological pressure that obliges her to suspend judgment, in which case she judges that hypothesis h is true or that h is false, and remains in doubt. Each one of her beliefs is taken to be true and, when adding new beliefs to this corpus, the agent should be concerned with not importing falsities or errors in a way that takes into account the truth of the beliefs she already holds. When adding new beliefs to our corpus of beliefs, this corpus is regarded as true and as consistent, so when importing new beliefs, the rational agent should do so in a manner that preserves the consistency of that corpus. She should not introduce as true information that contradicts elements in her set of beliefs. This, evidently, does not happen when she adopts new desires, since these attitudes are not subject to this requirement of consistency. Someone can have two conflicting desires without this putting any psychological pressure on her to necessarily suspend judgment. This happens all the time with desires, we have lots of inconsistent desires and, most of the time, this does not seem to bother us; for instance, right now I want to be in Cancún, but at the same time I want to be in New York City (conflicts of desires, however, may get us into practical problems when we are pressed to act on only one desire and we have to decide which one we should act on). So, their being subject to the standards of truth and falsehood are also what distinguishes beliefs from desires, but also from other cognitive intentional states, such as guessing, supposing and some others—here I am taking beliefs as “full beliefs,” that is, beliefs that express certainty about the truth or falsity of a proposition. Now, as I said before, there is no need to flesh out our concept of truth and commit ourselves to a certain theory of truth to make sense of the idea that beliefs are distinguished by their relation with truth. The fact that beliefs aim at the truth is independent of how truth is defined. By the same token, there is no need to talk about beliefs as states distinguished by their fitting the world. This approach may have the advantage of moving away from the commitment to correspondentism as well as a purely representational image of beliefs, as the one we get from Hume. According to the formulations of the notion of direction of fit that we have analyzed, the essential characteristic of desires is that they aim at their realization or satisfaction (understanding these as the matching of the world to the content of the desire). But neither “realization” nor “satisfaction” is a normative concept, so there seems to be an asymmetry in this respect. An unsatisfied or an unrealized desire is not, properly speaking, a mistaken, incorrect or wrong desire, and, as we already know, this is not a reason for dropping it. If a desire does not get its goal, if it is not satisfied, we do not say that the desire is at fault36—but neither is the world: there does not seem to be anything wrong with the world if it does not fit the content of the desire, and it is not clear why the world should be changed to fit our desires. As John Milliken tells us, “It does not seem correct to say that the world ought to match up to our desires. For one thing, it is questionable to

42  Gustavo Ortiz-Millán say that the world is subject to norms of any kind. For this reason, we cannot really mean to say that the world ought to conform to our desires.”37 We cannot really mean that it ought to be the case that the world conforms to our desires. To be sure, an account of desires such as Hume’s cannot be understood normatively: Hume famously thought that desires were some kind of causal forces that push us to action and that there were no rational constraints to them, that they are non-inferential, or that they were “loose” with regard to reason.38 Most contemporary philosophers, Humeans and anti-Humeans alike, have abandoned this “misconception” for one that sees desires as intentional states, and many would accept the claim that intentional mental states are normative. This means that desires, being intentional states, are also characterized by this normativity, and normative terms such as “correct,” “incorrect,” “rational” and “irrational” apply to them. If we accept this picture, how do we make sense of the idea that desires, being intentional states, are regulated by certain norms and that these norms distinguish them from other intentional and normative states, such as beliefs? I want to adopt here the ancient idea that, just like beliefs are directed towards truth, desires are directed towards the good. Just like everything we believe, we believe it because we take it as true (even if it is actually false), everything we desire, we desire it because we take it as good or it appears as good to us (even if it is objectively bad). Beliefs take place sub specie veri, desires, sub specie boni. This idea goes at least back to Augustine and to the scholastics, for whom nihil appetimus, nisi sub ratione boni; nihil aversamur, nisi sub ratione mali (we desire nothing except under the form of the good; nothing is avoided except under the form of the bad). This idea is also already in Anscombe, who claims that “truth is the object of judgment and good the object of wanting; it does not follow from this either that everything judged must be true, or that everything wanted must be good.”39 Dennis Stampe puts it in the following way: The difference between beliefs and desires lies somehow in the differing ways in which they determine the rationality of action. To locate the difference, we must look back, after all, to the representational character of the two states: for while the belief and the desire that p have the same propositional content and represent the same state of affairs, there is a difference in the way it is represented in the two states of mind. In belief it is represented as obtaining, whereas in desire, it is represented as a state of affairs the obtaining of which would be good.40 A desire is a mental state whose intentional object is conceived as or appears to be good, it involves the evaluation that the object desired is good, valuable, or worth desiring.41 Just like beliefs aim at the truth, desires have a teleological character, and they aim at the good. And just like in the case of beliefs, we do not have to commit ourselves to a strong characterization or

A Criticism of Direction of Fit  43 definition of what “good” means, with just a cautionary warning: “good” should not be understood subjectively, because then the idea that desires aim at the good would be trivialized, if we understand the good as whatever we desire, and there would be no way in which desires could be wrong under that picture. Subjectivism tells us that there is no independent normative notion that guides the appropriateness of the objects of our desires; subjectivism is at odds with this old formula. However, this is compatible with claiming that desiring is conceiving something to be good from a particular evaluative perspective: we conceive it as good or it appears to us as good, even if it is objectively bad—however we conceive this objectivity. Although we might see something as worth pursuing, we might be wrong. Some people may make the bad their good, and in order to make their actions intelligible, we should see them as taking their goals as worth pursuing: the desire of the thief to steal a wallet takes place under the appearance that it is good—even if it is not and the thief himself might even know that it is not. The thief sees something good in the act of pick pocketing, he gets some money, and this is why he performs the act; most of us do not see anything good in such an act and that is why we do not desire to act in such a way. We can desire what we know is not good because we see something worth pursuing in it. Under such an understanding, we might say that some of our desires are wrong. It might be claimed that someone may not see anything good in some of his desires, such as the junkie who would rather not have a desire to have the crack for which he craves. But even in that case, the junkie may see something good in the act of taking the drug, since he may fear the abstinence syndrome that usually comes if he does not get it. He does see something good in acting on the desire he wishes he did not have. This view does not imply that desires should be identified with beliefs, i.e., beliefs about what is good, desirable or valuable, as the “desire-as-belief” thesis claims. If desires were subspecies of beliefs, then they would have the same roles in theoretical and practical reasoning and the same kind of norms would apply to them, but this is not the case. Truth imposes certain norms on beliefs that do not work in the case of desires. As I claimed before, the idea that beliefs aim the truth carries with it some other associated norms that the idea that desires aiming at the good does not.

6. Coda In this paper I have argued that there are reasons for thinking that the notion of direction of fit results in a theoretically loaded distinction, one that carries different commitments to epistemological and psychological theories. This makes the notion of direction of fit hard to accept from different theories in the fields where it is important to distinguish beliefs and desires. Instead, I propose to distinguish these two attitudes in purely normative terms, given the central character that this feature has in our mental life. In this way, we

44  Gustavo Ortiz-Millán can save the idea that beliefs aim at the truth and desires at the good from the picture originally drawn by Anscombe, and I argue for a purely normative interpretation of the distinctions of these concepts. Now, do we get a less compromised way of distinguishing between beliefs and desires than that of direction of fit? Our normative distinction does not need to rest on any substantive view of the nature of truth: truth, however it is defined, is the value that governs our beliefs, and that imposes normative constraints to our holding, changing and revising our beliefs. This idea is compatible with different views of what truth is. A normative interpretation of the idea that beliefs aim at the truth is independent of how truth is defined. This distinction is also not committed from the very beginning to some Humean theory of motivation, and for the same reason, does not result in the problems the anti-Humean has to face when trying to apply this notion to an explanation of motivation. If the idea of direction of fit disappears, then it also seems to disappear the controversial notion of states with double direction of fit. It is also not committed to any substantive theory about the nature of the good, insofar as there is an independent normative notion that guides the appropriateness of the objects of our desires, but this independent normative notion may be conceived in different ways. One of my purposes in this paper has been to sketch an alternative way of distinguishing between beliefs and desires to that of direction of fit. There may be many more ways to distinguish between beliefs and desires, this is just one among others. To be sure, the distinction I have presented can be supplemented since, as I have already said, there are some other norms that govern these mental states that can help us making some other, finer, distinctions. These norms not only help us distinguishing between beliefs and desires, but also between these two and some other intentional states. It would just be a matter of fleshing out the idea that intentional states are normative in terms of the different norms that govern them and that help us distinguishing them. But this is something I would have to do somewhere else.42

Notes 1 John R. Searle, Intentionality (New York: Cambridge University Press, 1983), 7. 2 Lloyd Humberstone, “Direction of Fit,” Mind, vol. 101 (1992): 61–3. 3 Searle, Intentionality, 7; see also J. L. Austin, “How to Talk: Some Simple Ways,” in Philosophical Papers, ed. J. O. Urmson and G. J. Warnock, 3rd ed. (Oxford: Oxford University Press, 1979), 134–53 (originally published in Proceedings of the Aristotelian Society, 1952–3). Austin was the first one using the terminology of direction of fit in his theory of speech acts. 4 See John R. Searle, Making the Social World. The Structure of Human Civilization (Oxford: Oxford University Press, 2010). 5 Talbot Brewer argues that the notion of direction of fit, along with the idea that actions have to be explained as products of belief-desire pairs, are parts of the same package of dogmas, which ought to be abandoned. See his “Three Dogmas of Desire,” in Values and Virtues: Aristotelianism in Contemporary Ethics, ed. T. Chappell (Oxford: Clarendon Press, 2006), 257–84.

A Criticism of Direction of Fit  45 6 Bratman analyzes the case of intentions, and claims that these resist their reduction to belief-desire talk. Unlike desire, intention is a conduct-controlling pro-attitude, a commitment to action, he claims (see his Intention, Plans, and Practical Reason [Cambridge, MA: Harvard University Press 1987]). Also for Gilbert Harman, “intentions are not merely implicit in beliefs and desires but are distinctive attitudes that play a special role in thought and action” (Change of View [Cambridge, MA: MIT Press, 1986], 95). Emotions and some other intentional states have been subject to this strategy of extension as well; see for instance the way in which Searle explains emotions such as regret, fear, hope, pride or shame in terms of beliefs and desires (Intentionality, 29–36; see also Joel Marks, “A Theory of Emotions,” Philosophical Studies 42 [1982]: 227–42). Peter Goldie, among others, criticized these belief-desire accounts of emotions, claiming that these approaches leave feeling out of the picture, “but then, because it is glaringly obvious that feelings are, to say the least, relevant to emotional experience and cannot be left out of the picture, they then have to be added on, at the end of the story so to speak, for example as comprising awareness of one’s bodily condition” (Goldie, The Emotions. A Philosophical Exploration [Oxford: Oxford University Press, 2000], 4). He called these approaches the “add-on view” of emotions. 7 G.E.M. Anscombe, Intention (Ithaca: Cornell University Press, 1957), 56, §32. 8 Kim Frost has argued that Anscombe is not a direction-of-fit theorist, not only because she does not use the concept, but also because she was not trying to distinguish between kinds of mental states, but between a kind of mental state, a belief or a judgment, to a kind of event, a performance (“On the Very Idea of Direction of Fit,” Philosophical Review 123 [2014]: 429–84). We should be aware of this alternative reading of Anscombe, but for the present discussion, I just follow Platts’s reading, since many philosophers making this distinction have followed this kind of reading and I want to focus on their take on the notion. 9 Mark Platts, Ways of Meaning (London: Routledge and Kegan Paul, 1979), 256–7. 10 Searle actually says that “it is, so to speak, the fault of the world if it fails to match the intention or the desire” (Intentionality, 7–8). But it not clear why the world is at fault if it fails to match the content of our desires. 11 Even though at some point Hume has to accept that desires can have influence of beliefs when he claims that desires (“passions,” he says) can be unreasonable when they are founded on the supposition of the existence of objects that do not exist (David Hume, A Treatise of Human Nature [1739–40], eds. D. F. Norton and M. J. Norton [New York: Oxford University Press, 2000], 2.3.3.7). 12 I take here the distinction horizontal/vertical from Nick Zangwill (“Direction of Fit and Normative Functionalism,” Philosophical Studies 91 [1998]: 173–203). The verticality of the directions of fit of these states with the world is even clearer in the work of Searle, who even represents the direction of fit of desires with an upward arrow (↑) and that of beliefs with a downward one (↓) (Searle, Rationality in Action [Cambridge, MA: MIT Press, 2001], 48). 13 Philip Pettit, “Desire,” in Routledge Encyclopedia of Philosophy, ed. Edward Craig (Oxford: Routledge, 1990), 30. 14 Searle, for instance, takes seriously the commitments of the notion of direction of fit with correspondentism (and realism) and then goes on to “defend the idea that truth is a matter of correspondence to facts” (The Construction of Social Reality [New York: The Free Press, 1995], 199). 15 Although we may distinguish here between basic and instrumental desires, i.e., the desires we have for the means to our ends. Instrumental desires have both directions, but basic desires might not.

46  Gustavo Ortiz-Millán 16 Hume, Treatise, 1.3.2.2. 17 Mark Platts, Moral Realities (London-New York: Routledge, 1991), 40, his emphasis. 18 Platts, Moral Realities; John McDowell, “Are Moral Requirements Hypothetical Imperatives?” Proceedings of the Aristotelian Society, supp. vol. LII (1978): 13–29; or Jonathan Dancy, Moral Reasons (Oxford: Blackwell, 1993). 19 Lloyd Humberstone, “Wanting as Believing,” Canadian Journal of Philosophy, vol. 17 (1987): 49–61. 20 J. E. J. Altham, “The Legacy of Emotivism,” in Fact, Science and Morality, ed. C. Wright and G. Macdonald (Oxford: Blackwell, 1986), 275–88. 21 See H. Jacobson-Horowitz, “Motivational Cognitivism and the Argument from Direction of Fit,” Philosophical Studies, vol. 127 (2006): 561–80. 22 This would be the case, for instance, with John McDowell’s “consequential desires.” He claims that “the commitment to ascribe [. . .] a desire is simply consequential on our taking [an agent] to act as he does for the reason we cite; the desire does not function as an independent extra component in a full specification of his reason, hitherto omitted by an understandable ellipsis of the obvious, but strictly necessary in order to show how it is that the reason can motivate him. Properly understood, his belief does that on its own” (“Are Moral Requirements Hypothetical Imperatives?” 15, his emphasis). What is the point then of ascribing these consequential desires in these cases? But this could be extended to any other case where normative beliefs provide the motivational force. This, it seems to me, is also the case with Thomas Nagel’s “motivated desires” (The Possibility of Altruism [Princeton: Princeton University Press, 1970], chap. V). Platts (Moral Realities, chap. 2) subscribes Nagel’s distinction and makes a further distinction between instrumental desires (desires for a means to something else which is desirable) and independently justifiable desires (where something is characterized as desirable independently of one’s having that or some other desire). 23 Humberstone, “Wanting as Believing,” 49–61. 24 David Lewis, for instance, argued that this thesis is incompatible with standard formal theories of Bayesian decision. If the thesis of desire-as-belief were right, this would trivialize these theories—such as Richard Jeffrey’s decision theory. This would also imply that classical decision theory is not motivationally neutral, because it does not tolerate non-Humean interpretations (“Desire as Belief,” Mind 97 [1988]: 323–32; see also John Collins, “Belief, Desire, and Revision,” Mind 97 [1988]: 333–42). Arlo-Costa, Collins and Levi defend Lewis’s position, but they get a stronger conclusion: they make the claim independent of Jeffrey’s theory, and argue that the desire-as-belief thesis is incompatible even with a weaker system of decision theory that does not presuppose Jeffrey’s theory (H. Arlo-Costa, J. Collins, and I. Levi, “Desire-as-Belief Implies Opinionation or Indifference,” Analysis 55 [1995]: 2–5). A reply to the critics of the desire-as-belief thesis can be found in Price (“Defending Desire-as-Belief,” Mind 98 [1989]: 119–27). Similar conclusions would be valid for the field of belief revision, since in this kind of account, when new information is added to a set of attitudes, beliefs evolve in one way and (instrumental) desires in another, and this rules out the identification of desires with beliefs. A hybrid of beliefs and desires, trying to go both ways would be torn apart. 25 Arto Laitinen, “Against Representations with Two Directions of Fit,” Phenomenology and the Cognitive Sciences, vol. 13 (2014): 179. 26 Michael Smith, The Moral Problem (Oxford: Blackwell, 1994), 115. 27 Smith, The Moral Problem, 118. 28 Marcel Proust, Remembrance of Things Past, vol. III, The Fugitive, trans. Scott Monkrieff (New York: Vintage, 1982), 460.

A Criticism of Direction of Fit  47 29 David Sobel and David Copp, “Against Direction of Fit Accounts of Belief and Desire,” Analysis, vol. 61 (2001): 48. 30 Humberstone, “Direction of Fit,” 63–5. 31 Humberstone proposes a different way of understanding the metaphor of direction of fit, in terms of background conditional intentions: whether a mental state is a belief or a desire is determined by higher-order intentions regarding that state (“Direction of Fit,” 73–5). Some mental state that p is a desire that p only if the agent has a background conditional intention that p, that is, an intention that is conditional on her having the mental state. On the other hand, some mental state that p is a belief that p only if the agent has a conditional intention to stop having this mental state if p is false. Desires are states we intend to satisfy, beliefs, states we intend to leave if they are false. However, Gregory has rightly subjected Humberstone’s characterization to a similar objection to the one he raised against Smith: if intentions have the same direction of fit as desires, then this has to be explained. If we apply Humberstone’s account of direction of fit to intentions, then this starts a vicious regress and “then we never really get an explanation of how the original state has the direction of fit that it does,” so the account must be false. If, on the other hand, we do not use Humberstone’s account, we would have to explain the direction of fit of intentions with some independent account, but then his account is incomplete, “and it is not clear why that other account could not itself be used to make sense of the direction of fit of beliefs and desires in the first place.” So we should probably just abandon Humberstone’s way of saving the notion of direction of fit. See Alex Gregory, “Changing Direction on Direction of Fit,” Ethical Theory and Moral Practice, vol. 15 (2012): 603–14. 32 Sobel and Copp, “Against Direction of Fit Accounts of Belief and Desire,” 47–8. 33 For this line of criticism, see Velleman, who argues that imaginings, unlike beliefs, are not states which we have reason to constrain in the face of the world. Also, there is an ambiguity in the term “cognitive,” which may have several different meanings. See David Velleman, “The Guise of the Good,” in The Possibility of Practical Reason (New York: Oxford University Press, 2000), 99–122. 34 In my interpretation of this claim I partially follow Wedgwood’s normativist account in “The Aim of Belief,” Philosophical Perspectives, vol. 16 (2002): 267–97. Even though I realize normativism in philosophy of mind is debatable, I will not attempt to defend it here against its critics. For some of the arguments for and against it, see K. Glüer and Å. Wikforss, “The Normativity of Meaning and Content,” in The Stanford Encyclopedia of Philosophy (Spring 2016 Edition), ed. E. N. Zalta, https://plato.stanford.edu/archives/spr2016/entries/ meaning-normativity/. 35 Richard Wollheim, The Thread of Life (New Haven and London: Yale University Press, 1984), 53. 36 Although there may be exceptions, such as when someone desires something that they should have known they could not achieve or when I desire X as a means to Y, when I should have known that X would not cause Y. 37 John Milliken, “In a Fitter Direction: Moving Beyond the Direction of Fit Picture of Belief and Desire,” Ethical Theory and Moral Practice, vol. 11 (2008): 568. 38 This is the standard reading of Hume, but instrumental desires may not conform to this view. In any case, for a criticism of this misconception of desires as causal forces with no rational constraints, see Platts, Moral Realities, and also his “Moral Reality and the End of Desire,” in Reference, Truth, and Reality, ed. M. Platts (London: Routledge and Kegan Paul, 1980), 69–82. 39 Anscombe Intention, 76, §40.

48  Gustavo Ortiz-Millán 40 Dennis Stampe, “The Authority of Desire,” The Philosophical Review, vol. 96 (1987): 355, his emphasis. 41 Michael Stocker criticized this perspective arguing against the (up to that point) overwhelming consensus that “only the good attracts.” Stocker defended a view in which there is a gap between motivation and evaluation to account for the case of desiring the bad. See his “Desiring the Bad: An Essay in Moral Psychology,” The Journal of Philosophy, vol. 76 (1979): 738–53. David Velleman (“The Guise of the Good”) adopted a similar view in which judging that something is good cannot be identified with motivational attitudes such as desires. However, Sergio Tenenbaum has criticized this separation between evaluation and motivation, arguing that even in cases of akrasia, accidie or perverse motivation, one’s desires for something can be identified with conceiving it to be good. It would be impossible to develop a whole defense of this idea here. I follow Tenenbaum in his general outline. See his Appearances of the Good (New York: Cambridge University Press, 2007). 42 I presented earlier versions of this article at conferences of the Portuguese Society for Analytic Philosophy in Lisbon, the Colombian Philosophical Association in Cartagena, the Berkeley Social Ontology Group at the University of California, Berkeley, and the Institute of Philosophical Research, National Autonomous University of Mexico (UNAM), I thank the audiences in all those places for their comments and discussions on the topics of this article. I am particularly indebted to Veli Mitova, Mark Platts, John Searle, Walter Sinnott-Armstrong and Moisés Vaca for their very helpful comments. This work was supported by the DGAPA (UNAM) under a PASPA grant.

4 Equality as a Foundation of Ethics James Griffin

1. The Changing Role of Equality In the twelfth century the corpus of Roman law was recovered and a critically ordered edition of some of the enormous mass of cannon law texts appeared in the Decretum of Gracian (ca. A.D. 1140). For the first time the Medieval monks had access to two great bodies of law, covering worlds both sacred and profane. A term with our modern sense of a “right” emerged in the twelfth or thirteenth centuries, probably first in Bologna in the work of the canonists, scholars who were glossing and to some extent harmonizing cannon and Roman law.1 At that time, the use of the Latin word “ius” expanded from meaning what is just to include also our modern sense of a “right”—that is, an entitlement that an individual person has to do or have something. In the curious poverty debates following the death of Francis of Assisi, one finds the transition from the form of words that it is a natural law (ius) that all things are held in common, and so a person in mortal need who takes from a person in surplus does not steal, to the newly emergent form of words that a person in need has a right (ius) to take from a person in surplus and so does not steal. Pico della Mirandola, an early Renaissance philosopher who studied cannon law in Bologna in 1477, gave a highly influential account of the link between human freedom and the dignity of human status, which went like this: God fixed the nature of all things in Creation, except for man; man alone was left free to determine his own essence; in this he is God-like. This freedom constitutes, as it is put in the title of Pico’s best-known work, “the dignity of man.” The appearance of natural rights was a large step in the journey of the idea of equality to the centre of the ethical stage. Natural rights are protections of our normative agency,2 which all human beings have equally, so if natural rights become central, so too does equality. Of course, the role of equality in ethics started growing long before the late Middle Ages. The idea of equality was prominent in both the Old Testament and the New Testament; for example, all human beings, Genesis tells us, are made in God’s image.3 But Christianity, having given human beings equality, then drastically qualified it in Augustine’s doctrine of election and Aquinas’s doctrine that slavery

50  James Griffin is the consequence of sin.4 Aristotle’s well-known principle of equality in Book III of Nicomachean Ethics is that injustice arises from treating equals unequally and unequals equally. But, as Aristotle’s own beliefs reveal, his principle is compatible with a society’s treating free men as vastly superior to women, slaves and barbarians.5 The first group in the West to champion equal treatment of all human beings was the Stoics; whereas Aristotle saw women as having inferior capacities to men and barbarians as being childlike, and so on, the Stoics saw human beings in general as having a kind of rational agency that confers upon them a dignity that demands respect. This Stoic conception of equality, though much like our own modern one, did not achieve the centrality in ethics that our modern one eventually did. Perhaps this was because the extraordinary success the Christians met with in proselytizing their religion in the Mediterranean world overshadowed Stoic ethical teaching. With the late Middle Ages, however, equality began to be seen to fill not just an important role in ethics but a foundational one. To regard other persons with equal consideration or equal respect or whatever characterization seems best became, for many, the moral point of view itself. The idea of equality often appeared at two points in this thought: once as fact (we are born equal; that is, normal human beings are born with the potential for normative agency) and again as obligation (so we must be treated accordingly—that is, equally). Equality requires not only embracing these abstract egalitarian principles but also exercising strong scepticism about the ethical relevance of certain facts, for example, that barbarians are childlike. Somewhere in our journey from the late Middle Ages to today, we passed a point at which we could reasonably claim to have entered the Age of Equality.

2. Problems for the Age of Equality What is the nature of the “ethics” that these changes brought about? We cannot immediately tell. We have to work out where practical reason takes us. Is it an ethics of specific laws of behaviour, such as the Ten Commandments? Does it take us to a Christian absorption of Aristotle’s more systematic ethics, such as Aquinas made? To an ethics of an ultimate decision procedure, such as maximizing well-being (as with utilitarians) or determining what respect for autonomous agency demands (as with Kant)? This equality-centered model of ethics, though, suffers badly from incompleteness. Even today, we have not worked out to what it amounts. None of us has a satisfactory grasp on what the principle of equal respect says. Does it require universal impartiality? Virtually none of us live, or think that we should live, completely impartially. We are partial to family and friends and to certain institutions and causes. This may be because we have greater moral obligations to family, friends and certain causes. Or perhaps it is not at base a matter of morality at all but of human capacity; we just cannot care as much about strangers as we do about family and friends, and

Equality as a Foundation of Ethics  51 “ought” implies “can.” And we have no real grasp of what exactly it is that should be made equal, and whether if everyone were comfortably above a certain level of well-being the moral demand of equality would cease. We shall not know what moral norms can be derived from equal respect unless we manage to put more content into this troublesomely elusive idea. It is to the ideas of equality and impartiality, many believe, that reason carries us; there is, it seems, no reason why I should be treated as more important than you, or my child as more valuable than yours. The fundamental principle in ethics, it seems, is “each to count for one and nobody for more than one” or “respect equally the autonomous agency in all persons” or some such egalitarian norm. Still, the location of the notion of equality in a deep place in ethics left problems. If equality is understood as the moral point of view, it encompasses all of morality. If it is taken to be a principle of justice or fairness, then it may encompass only a part of morality. What is more, many different sorts of principles of equality have moral weight. How are these different principles related? I shall return to this ill-understood matter shortly. Our present-day Western ethics is largely what the Enlightenment has bequeathed to us. I do not mean that there have been no developments in ethics for the last three hundred years, but that most of the central features of Enlightenment ethics are with us still. What is more, the egalitarianism of the Enlightenment remains, installed in present-day ethics at a deep, organizing level, though the unclarities of equality that I mentioned earlier also remain. Most philosophers today believe, or rather take for granted, that philosophy’s job is to produce a proper, i.e., systematic, ethics: for example, consequentialism, Kant’s ethics, virtue ethics, and so on. Equality of civil rights, or of human rights or of well-being, or of opportunity, or of income, or of resources are all intelligible ideas, but none of them is abstract enough to be plausible as the, or a, foundational principle of equality. Take, for instance, equality of human rights. This could not be the foundational principle, if, as I think, human rights cover part of the moral domain but not the whole of it. They cover what is needed to ensure that normal human beings can function as normative agents. When it comes to resources, normative agency can be ensured at a fairly low level, considerably lower than the level at which most of us in the developed world now live. However, most of us take the foundational principle of equality as also requiring certain equalities above that minimum. A foundational principle of equality is one thing, a principle of human rights quite another. When trying to express the foundational principle of equality, writers reach for abstract expressions such as equality of respect or of consideration or of treatment. But there are problems here too. These ideas are commonly thought to require impartiality between people. But it is almost universally accepted that a flourishing life consists of such things as relations of love and friendship, commitments to worthwhile causes, and so on, and that many of these typically require deep partiality.6 If we Homo sapiens cannot

52  James Griffin do a certain thing, in the sense of “can” that is meant in the principle that “ought” implies “can” whatever that turns out to be, then it is not the case that we ought to do it. Such actions, as I have argued elsewhere, do not even enter the domain of moral obligation.7 Is the foundational principle of equality a principle of equal treatment? No, because I do not have to treat strangers equally to my own children. Is it a principle of equal concern? No, because concern has to do with giving attention, caring, and even motivation and action to help, and I do not have to be as much concerned in this way with strangers as with my own children. Is it that we all matter equally? If this is merely equal concern, as just explained, then again no. Nonetheless, there is also a sense in which we do indeed all matter equally. Although I favour my children over yours, I do not think that my children are more valuable than yours. We are all born equal; for instance, all normal new-born Homo sapiens are potential normative agents. We are equal in that crucial dignity. What implications does this undeniable equal status have for ethics? Most of us would say that it requires perhaps among other things, respect for a person’s normative agency. For example, it prohibits slavery, arbitrary imprisonment, brain washing, torture and many other forms of degrading treatment or treatment as an inferior. But the norm, do not deny a person normative agency, though immensely important, has only limited consequences for ethics. It constitutes no challenge to my partiality to those to whom I stand in certain close relations. It does not require, for instance, equality of welfare or resources or life prospects. It requires equal human rights, but that is not the foundational equality for which we are looking. The partiality that lies outside ethics includes partiality to one’s children. But this does not allow the enormous partiality that parents in the developed world now commonly lavish on them. One can have very much more concern for the poverty-stricken children of the Third World than most of us now have without at all sacrificing one’s valuable relations with one’s own children. So we do not escape the question: what is a better balance between our concern for own children and for strangers than the one we have so far struck?

3. Equality: Miscellaneous Principles but No Master Principle I am not suggesting that there are no principles of equality. There is a miscellany of them, each of narrower scope than the hypothesized foundational principle that we have just been trying, but failing, to formulate. This is T. M. Scanlon’s conclusion.8 He decides that the elimination of inequalities may be morally required for five reasons: (first) in order to “relieve suffering or severe deprivation” (that is, to narrow or eliminate “the gap between rich and poor [. . . as] a way of reducing the suffering of some without causing others to suffer a similar fate”); (second) to “prevent stigmatizing differences in status” (that is, to narrow or eliminate the gap in order to produce a society in which all regard one another as equals);

Equality as a Foundation of Ethics  53 (third) to “avoid unacceptable forms of power or domination” (that is, to prevent an unacceptable degree of control over the lives of others); (fourth) to “preserve the equality of starting places which is required by procedural fairness”; and (fifth) “procedural fairness sometimes supports a case for equality of outcomes.”9 One can see the case for considering these five, anyway, to be principles of equality. Scanlon thinks that there is, in addition, a fundamental principle of equality: A . . . formal notion of equal consideration, as stated for example in the principle that comparable claims of each person deserve equal respect and should be given equal weight. This is an important principle. Its general acceptance represents an important moral advance, and it provides a fruitful—even essential—starting point for moral argument.10 This formal notion of equality is like what Isaiah Berlin had in mind when he proposed that the “irreducible minimum of the ideal of equality” is captured by the formula “every man to count for one and no one to count for more than one”—a formula, he says, not uniquely connected with any one philosophical system.11 But nonetheless, it is connected historically with one particular philosophical system, namely utilitarianism, where it is conjoined to a summum bonum—pleasure, happiness, utility, well-being. “Maximize well-being, counting each person’s well-being equally” is a relatively clear direction. But the formal principle alone without utilitarian or other substantive evaluative addition—for example, the meagre instructions “count persons equally” or “show equal respect for persons”—is close to empty. There is a challenge to the very existence of this formal principle of equality that comes from the principle that “ought” implies “can.” The formal principle is inconsistent with the forms of partiality that fall outside ethics; there is no moral requirement, as we saw earlier, to treat all persons equally.12 If we try to avoid this inconsistency by raising the abstraction of the value occupying the variable place in the phrase “equality of x,” the threat changes to one of emptiness: for example, the emptiness of the expressions “equality of respect” or “equality of consideration.” What behaviour constitutes treating someone with equal respect? No one knows. One reply to that question might be this. Let us accept that we are not required to treat all persons equally (for example, our own children and complete strangers). I say that we arrive at this conclusion for reasons that lie outside of morality, namely in empirical considerations about the extent of human motivational capacities.13 But most philosophers who agree with my conclusion do not agree with my premises. The true justification of these inequalities, many of them say, comes entirely from within a fundamentally egalitarian ethics: there may be reasons of equality for allowing certain inequalities. The world is better off, all things considered, they think, if parents are allowed to be guided by their natural attachments to their own children. But

54  James Griffin how could we know? There are many different degrees of bias that parents might show. There are many different degrees and kinds of improvement that First World parents might make in the lives of the millions of starving children in the Third World. A parent can reliably enough know that lavishing less on one’s own children and instead increasing one’s donations to Oxfam can produce great benefits at relatively modest cost. But the calculations needed for us to conclude that the world is better off if parents are allowed to follow their natural attachments to their children are beyond us, at least to a degree of probability on which we would be willing to act. I have touched on this earlier.14 Not only human motivation has its limits; human capacity for understanding does too.

4. Discrimination Nowadays much of our concern about promoting equality is expressed in terms of prohibiting discrimination. Some groups of people are treated as inferior: a certain race or gender or class or caste. They are in fact equal in morally relevant ways, we believe, so they should be treated equally. However, just as not all inequality is morally objectionable, not all discrimination is either. A critic with fine discrimination is exactly what we want. Even discrimination that harms its object is not therefore morally objectionable, as in the case of an author whose reputation is much reduced by a critic with fine discrimination. Discrimination can even be damaging to a person’s life prospects without being morally objectionable, as when a top university turns one down because one is of inferior ability. Finally, treating people as superior—awards and honours—can be entirely welcome discrimination. What forms of discrimination, then, are morally objectionable? Suppose men have the vote but not women. Treating men and women differently is not in itself objectionable. But having a vote—that is, having a say in the laws to which one will be subject—is a human right. Human rights are rights that we have simply in virtue of being human, which in this morally relevant sense women are. Denying them the vote is objectionable because it violates their autonomy and denies them this civic equality. Again, suppose that two top executives of a multi-national firm, equally competent and with equal responsibilities, both paid very well indeed, are paid different amounts only because one of them is the boss’s child. There is no violation of a human right here, I should say, and at this high level of wealth we are unlikely to get greatly upset by this particular case of inequality. But there may still be something wrong about it—namely, unfairness. Suppose a society does not allow same-sex couples to marry. This seems to me a major violation of their liberty.15 Liberty, as I think we should use the term here, is a matter of not being stopped from pursuing the life that seems to one worthwhile. Being able to form a stable, socially recognized relation to a person one loves, within which one may also choose to raise

Equality as a Foundation of Ethics  55 children, is at the heart of what, for many, is the worthwhile life. Preventing same-sex couples from living this form of life is what makes it an especially gross violation of their liberty. This is by no means the only objection to it. It may also be demeaning to the same-sex couples involved. Another objection is failure of rationality. Different-sex couples have a human right to liberty because of their status as normative agents. But samesex couples are normative agents too. To treat them differently in this case is not only unfair but also irrational. These successful objections to discrimination are based on unfairness, illiberality, insult, and irrationality. Are they also based on inequality? Suppose a society made housing for its members more equal as a way of reducing the harmfully low esteem in which the poor in that society are held. Equality is relevant to morality when, for example, it has this sort of causal connection to something valenced in itself: for instance, harm to the poor. Equality, however, is not in itself a substantive value; it is a state of the world and in itself neither good nor bad. It becomes relevant to morality only by having the right sort of connection to something else that is substantively valuable. For example, human beings are characteristically normative agents; it is that status—normative agency—that is a substantive value. It constitutes, as the United Nations puts it, “the inherent dignity of the human person,” and this dignity-bestowing status must be protected.16

5. Implications for Normative Ethics Consequentialists use Bentham’s formula “Each to count for one, nobody for more than one” as foundational. They also adopt a criterion of moral right and wrong: maximizing or satisficing and so on. The most plausible form of consequentialism, to my mind, is indirect; it applies the foundational consequentialist principle not directly to the choice of action but to the choice of lower level principles that are then used to decide action. This system of ethics makes huge demands on the calculation of consequences. Many of the calculations are possible, but many of them—and essential ones at the heart of an indirect consequentialist system—cannot be done to a degree of probability on which we should be willing to act. Besides that, consequentialism requires a level of impartiality that normal motivation often cannot manage. Consequentialism fails certain elementary feasibility tests. We are so used to Newtonizing ethics that elements of system often seep into normative thought without our noticing. Suppose we are thinking about a principle of equal well-being. We come to realize, let us say, that a society can achieve equality of well-being by levelling down as well as by levelling up. Every society, no matter how poor, can therefore easily satisfy the principle; it can simply level down. But we also have some such principle as “Don’t gratuitously make people worse off.” Our various principles, many of us think, are like forces in physics. Several may be present in one instance, and then we must, as in physics, determine their vector.

56  James Griffin Having noticed it or not, we have just installed scientific models of system in normative ethics. We accepted that the principle of equality applicable here says only “Make equal”—an obligation with an extraordinarily thin specification. Influenced by our Newtonian ideal, we think that, as physicists break down mechanics into its elementary forces, we moral philosophers should break down ethics into its elementary principles, such as “make equal.” However, the specification of this supposed principle of equality is so thin that we ought to wonder whether there is enough to it to qualify as a moral principle. The principle could not plausibly say merely “Make everything equal.” Suppose it says “Equalize well-being.” A person’s level of well-being can matter hugely, which is why well-being is so central to ethics. The principle “Equalize well-being” has to include elements of this motivational sort for it even to be a moral principle, for it even to acquire normative status. The first specification I gave of the principle was so very thin, so without connections to human motivation, that it lacks action-guiding force. How then could the principle be subject to levelling down? Why was the belief that it could be levelled down arrived at so easily? I suspect that it was because it fits the sort of paradigm of system provided by the natural sciences. In physics, for example, the smaller the component into which one can analyze one’s data the better. But in normative ethics one does not have anything that will serve as a unit in this vectoring exercise unless it qualifies as normative.

6. A Wider Problem: Indeterminacy of Sense So far in this paper I have been considering the indeterminacy of sense of the supposed foundational principle of equality. Indeterminacy of sense is a common phenomenon in ethics—and in thought generally. If asked what human rights we have, we can answer with a widely agreed list of them, although perhaps with a few disputed cases. We have human rights to life, liberty, autonomy, basic well-being, privacy, etc. But lists of names, which is what the United Nations gives us, are by no means enough to support a satisfactory discourse of human rights. What, for example, is the right to life a right to?17 This least challenged of all human rights is also, I think, one of the most indeterminate. In the seventeenth and eighteenth centuries the right to life was conceived largely negatively, as a right not to be deprived of life without due process—that is, a right not to be murdered. But the scope of the right to life seems irresistibly to expand. If living at liberty is of great value (to take another undisputed human right), then living, as well as living in that way, is valuable, and that seems to justify a wider negative right than just prohibition of murder—say, a prohibition of gratuitously endangering other people’s lives or of destroying a person’s sanity. What is more, it would seem to justify some positive rights. If you are drowning, and all that I have to do to save you is to toss you the lifepreserver beside me, and I wantonly disregard your plight, do I not violate

Equality as a Foundation of Ethics  57 your right to life? Does the right not include a positive right to rescue, if the cost to the rescuer is not great? And if it includes a right to be tossed a life preserver if one were drowning, would it not include a right to food if one were starving or to medicine if one were dangerously ill? And if it includes those, does it not also include a right to conditions, such as clean water and female literacy, the absence of which drastically shortens a child’s life? This ballooning of the content of the right to life is not just a theoretical possibility; it is exactly what has happened. The putative right has grown from a right against the arbitrary termination of the life of someone already born (murder), to a right against other forms of termination of life (abortion), to a right against the prevention of the formation of life (contraception, sterilization), to a right to basic welfare provision, to a right to a fully flourishing life. That last extension, no doubt, goes too far. But what is, and what is not, the content of the right to life? We need, but have not got, considerably more than just the name of the right. We ought not to be satisfied with where we are now. We now have no agreed criteria for the existence of a human right, nor for determining the content of a human right, nor for rationally resolving conflicts involving human rights. It is not that we have no criteria for the existence of a human right; it is just that in unusually many cases they are not enough to allow us to use the term in exactly the jobs it was introduced to perform. Terms certainly do not need highly determinate senses in order to be useful to thought. So why not get on with our highly indeterminate term “human right”? Of course, we should; we have no alternative. And think of how much good was accomplished during the twentieth century with the idea of “human rights” despite its considerable indeterminacy of sense. But that a term can still be of some use in thought despite a certain degree of indeterminacy of sense does not mean that it can provide the services that we need to have—and perhaps mistakenly assume that we have. Think now of the indeterminacy of sense of the supposed foundational idea of “equality.” It is far greater than that of the term “human rights.” The indeterminacy of foundational “equality” is so great that it has distorted the whole of ethics; it has fostered the belief that ethics will take the form of a theory of Newtonian abstraction. For all we know, the foundational principle of equality is empty and there is no such principle at the base of ethics. We are in the Age of Equality but cannot say what this foundational equality is.

Notes 1 Here I borrow a large discussion from my book On Human Rights (hereafter OHR) (Oxford: Oxford University Press, 2009), esp. ch. 2. 2 See OHR, ch. 2. 3 Genesis I:27; in the N.T. see Galatians 3:26–9. 4 See Stanley I. Benn, “Equality: Moral and Social,” in Encyclopedia of Philosophy, vol. 3, 2nd ed., ed. Donald M. Bonchert (Detroit: Thomson Gale, 2006), 329.

58  James Griffin 5 See Aristotle, Politics 1255a, 1254b, 1252b. 6 See my book Well-Being: Its Meaning, Measurement, and Moral Importance (Oxford: Clarendon Press, 1986), pt. I. 7 See my “ ‘Ought’ Implies ‘can’,” Lindley Lecture, University of Kansas, 2010. 8 Thomas M. Scanlon, “The Diversity of Objections to Inequality,” The Lindley Lecture, University of Kansas, 1996, repr. in his The Difficulty of Tolerance: Essays in Political Philosophy (Cambridge: Cambridge University Press, 2003), esp. 203–7. 9 Scanlon thinks that reasons (1) and (3) rest on “powerful moral ideas that are not fundamentally egalitarian,” that reason (4) is “only weakly egalitarian,” and that reasons (2) and (5) are “the clearest expressions of egalitarianism.” Nonetheless, I think, and Scanlon may too, that even in reasons (2) and (5) equality is only a necessary condition of realizing another value. For example, in (2) equality is sometimes necessary to bring about some other social value: namely, a certain spirit in society—a sense of confidence, security, or camaraderie. The deeper value at work here is more fraternité than egalité. And in (5) it is more fairness than equality of outcomes. 10 Scanlon, “The Diversity of Objections to Inequality,” 202. 11 Isaiah Berlin, “Equality,” in Concepts and Categories: Philosophical Essays (London: Hogarth Press, 1978), 81. 12 In the penultimate paragraph of section 2. 13 Well-Being, chap. XI. 14 Well-Being, chap. XI. 15 See OHR, chap. 9, sect. 3. 16 United Nations, Universal Declaration of Human Rights. Preamble. 17 I discuss this right more fully in OHR, chap. 12, sects. 1–3.

5 Inflation or Deflation of Rights? Rodolfo Vázquez

One answer to this question can be read in Mark Platts’s “The Language of Rights and Human Rights.”1 Platts adopts a quotation from James Griffin, taken from his book, On Human Rights: “There are strong inflationary pressures on the term [‘human right’] which have brought about its debasement [. . .], and they are still at work.”2 This “inflation” of the list of human rights seems to be a negative feature that responds to a sort of basic logical law that can be expressed in the terms of the Spanish philosopher of law, Francisco Laporta: “the more the roster of human rights is expanded, the less force each right will have as claims; and the shorter the list of rights, the greater will be the moral or legal force adequately accorded them”.3 This “inflation” of human rights has operated at the institutional level and it is a fact that the nominal list of rights, especially in the constitutional sphere, has increased exponentially. The recent Mexican constitutional reform4 establishes in its Article 1 the so-called constitutional block, that is to say the legal systematization of all materially constitutional norms that are not necessarily contained in the Constitution, like, for example, the rights recognized in international conventions and which form part of the Mexican legal order. I quote the first paragraph of the Article 1 in force: In Mexico all individuals shall enjoy the human rights recognized in this Constitution and in the international conventions to which the Mexican State is party, as well as the guarantees for their protection, the exercise of which may not be restricted nor suspended, save in such cases and under the conditions stipulated by this Constitution. I do not believe that this type of quantitative inflation of rights is something negative in itself, nor do I believe that it reduces their moral and legal force. It is not to this kind of inflation of rights, more procedural than substantial, that Platts and Griffin refer, but to that kind of inflation of rights whose contents tend towards a difficult and even impossible satisfaction; neither, do I believe, are they advocating a thinning down of rights oblivious to the vital interests and needs of individuals and social collectives. At the constitutional level, there is a tendency to compare, for instance, the more than

60  Rodolfo Vázquez 400 articles of the Constitutions of Brazil, Ecuador or Bolivia with the seven articles and some twenty amendments of the Constitution of the United States. It is well known that the austerity of the latter is accompanied by a “very hostile practice as regards social, cultural or economic rights in general”.5 Here we would be in the presence of a clear deflation of rights. But we have also seen that the impossibility of complying with an interminable list of rights (in Brazil there has even been a debate on the inclusion of the right to happiness in constitutional normativity) discredits and devalues the very constitutions of which they form part, giving rise to pejorative qualifications such as “poetical Constitutions”. At the international level, too, we often meet with rights enshrined in a normativity that is difficult to fulfill, such as the often mentioned “right to periodic paid vacations” which was set forth in the Universal Declaration of Human Rights of 1948. But then, beyond the mere matter of quantity, what are the bottom or top lines for reduction or expansion of the list of human rights? What criteria should be used for the purpose? What cannot be denied is that any attempt at thinning down raises the suspicions that a conservative agenda is at work, while any attempt at widening them raises those of populist inclinations. It is very common in the Anglo-Saxon literature, as I have already said, to present objections to the fattening of rights, especially when regarding social rights; it is equally common in developing or emerging countries to feel comfortable with the enlargement of constitutional charters, especially if social rights are at stake. I think that the criteria proposed by Griffin and Platts, for all their differences, which I shall come back to in a moment, situate both at a point—I’m not sure exactly where to place it on the geographical map—at a considerable distance from the conceptions of “the South”, since for them not all injustices are to be treated as a matter of the violation of rights. At the same time they are quite far from the conceptions of “the East” (for which, read North), since rights are not something to be determined exclusively by judges or the legislative authorities, in a sort of no-holds-barred positivism; rather, the moral critique of law plays an important part in their determination. Regarding the “eastern” conceptions, Platts cites the words of Herbert Hart, without doubt one of the most lucid exponents of contemporary legal positivism: A concept of legal rights limited to those cases where the law . . . respects the choice of individuals would be too narrow. For there is a form of moral criticism of law which . . . is inspired by regard for the needs of individuals for certain fundamental freedoms and protections or benefits. Criticism of the law for its failure to provide for such individual needs is distinct from, and sometimes at war with, the criticism with which Bentham was perhaps too exclusively concerned, that the law often fails to maximize aggregate utility.6 Having said the above, I would like to center my attention on the problem of the thinning down of rights, and there are two reasons for this. The first

Inflation or Deflation of Rights?  61 is that, despite the danger of an overwhelming expansion bringing legal discourse into a kind of “populist” discredit on both the constitutional and international planes, I do believe that excess is preferable to deficit, pleonasm to austerity. No few of the conquests that have been achieved in terms of a better legal guarantee of human rights, especially social rights, have begun by entrenching themselves in constitutional principles or international treaties as a basis for constructing the necessary framework of secondary laws for their legal protection. The problem of thinning down presents, I believe, far greater risks, since it implies leaving unprotected personal or collective social rights, exposing disadvantages or vulnerabilities that could hardly be compensated for from an extra-institutional sphere of action. The second reason is that the center of attention of Platts’s concerns, in the chapter of his book referred to, involves the criteria proposed by Griffin which—in his opinion, and I agree with him—thin down unjustifiably the extent of the rights of individuals. I quote Platts: the restrictions on the talk allowed in such cases by [Griffin’s] “trinist” theory [“persons trapped in non-functioning bodies,” to use one of the examples mentioned by Griffin] seem to go against that agreement, representing unacceptable “deflation” of the discourse upon human rights through exaggerated strengthening of the conditions presumed necessary for their possession.7 I shall approach the problem of the thinning down of rights from two possible points of view; one institutional, in the field of what has recently come to be called the global justice debate, a debate that confronts cosmopolitan theses with pluralist ones; and another, individual or moral, of course with deep social implications, but which has to do with the conditions that individuals must satisfy and which, as Platts says, are “conditions presumed necessary” for the possession of human rights.

Minimalism of Rights and Global Justice There seems to be consensus that efforts should not be spared to respect, protect, guarantee and promote civil and political rights, those called “first generation” rights; but we have many reservations about doing the same where the so-called social rights are concerned; and it is precisely towards the latter that critiques are aimed with the purpose of deflating rights on the institutional level. The advocates of deflation maintain that, where social rights are concerned, in the best of cases there is room only for political justification, but in no sense a moral one. Throughout the 1990s, in the framework of the democratic states, a robust expansion of rights was witnessed on the global scale, giving rise to considerable optimism. Legal protection of rights was manifested in special courts or tribunals, as exemplified by, on the one hand, the international tribunals set up for Rwanda and the former Yugoslavia, or the coming into

62  Rodolfo Vázquez force of the Rome Statute that created the International Criminal Court and, on the other, independent initiatives like the Pinochet case. This optimistic mood, to some extent complacent as regards the virtues of liberal democracy, was shared by illustrious political philosophers. It is enough to recall that, in the early 1990s John Rawls published Political Liberalism and Jürgen Habermas Between Facts and Norms. “Both authors”, as Cristina Lafont states, “articulated interesting solutions to the problem of how to reconcile demands of justice with respect for pluralism in modern democratic societies: the idea of an ‘overlapping consensus’, in the case of Rawls, and the ideal of ‘deliberative democracy’ in that of Habermas”.8 Was not the United Nations Charter itself a clear example of the feasibility of an overlapping consensus? And on the other hand, what could prevent the spread of Rawlsian constructivism and Habermasian deliberationism on the world scale or disavow such solid justifications for the already irreversible expansion of democracy? From the late 1990s, as we all know, the process of globalization began to present its inherent fissures: the escalation of terrorism was symbolized tragically in the destruction of the Twin Towers, and the most reactionary forces of so-called post-secular society began to present multiculturalism in its darkest colors. All of this put paid to the so long heralded cosmopolitan justice and global democracy; for it was in this context that the possibility of global justice for a pluralist world society began to seem questionable. And it was precisely on that point—the priority of justice over the pluralism of conceptions of the good, where we might have expected support from two of the most outstanding defenders of universal justice—that the response has been most frankly disappointing. For the conception expressed by Rawls in his The Law of Peoples regarding the overlapping consensus in the international context severely limits aspirations for justice. According to Rawls, the constitutional rights of the liberal democracies express “liberal aspirations” that, as such, it would be illegitimate to impose on other societies once the fact of global pluralism has been recognized: the standards of international human rights must be different and much less demanding than the standards of constitutional rights accorded to liberal citizens. It is not just that Rawls decides to establish a bottom line of “justificatory minimalism”—of the kind that allowed, as I have said, a consensus on the rights consecrated in the United Nations Charter—but that he recedes into substantive minimalism, that is to say, a thinning down of the contents and the scope of the rights recognized in the Charter itself. Rawls excludes, for example, the right to full equality and non-discrimination on grounds of sex, race or religion (a. 1 and 2), but also—and this is what is most serious—some first-generation rights: freedom of expression and association (a. 19 and 20), as well as political rights to democratic participation (a.21), with the purpose of rendering dialogue viable between the kinds of societies that he calls “decent liberal” and “decent hierarchical”.9

Inflation or Deflation of Rights?  63 Much the same can be said of Habermas. For this author, economic questions must be disconnected from obligations of justice incumbent upon the international community and be interpreted as political aspirations that, as such, reflect different evaluative orientations and, whose realization must therefore depend on the compromises negotiated between the opposed values and interests of the different transnational powers.10 Hence his proposal of a “multilevel system” with different political units on the supranational, transnational and national planes.11 Given that questions of distributive equity do not correspond to world organization on the supranational, but rather the transnational, level—that supposes political objectives related to the economy which are the object of “negotiated compromises”—the function of protecting human rights at the supranational level must be limited to the necessary minimum, which means, only and exclusively, the negative duty to prevent massive violations of human rights arising from armed conflicts, such as “ethnic cleansing” and genocide. On the level of global domestic policy there are, strictly speaking, no obligations of justice, but only “political goals to be aspired to”. Thus, for instance, the eradication of extreme poverty might be one of those goals, or it might not be. What remains clear for both these authors—whether on account of the conditions that would make an intercultural dialogue viable, or else for the sake of greater efficacy as regards redistributive decisions on the basis of negotiated compromises—is that it is political reasons and not moral ones that must predominate in the discourse of human rights in the sphere of global justice. Furthermore, not only do both authors regard social rights as political goals to be aspired to that cannot possibly generate correlative obligations; they also consider that even civil and political rights themselves must be substantially curtailed, thus leaving individuals in a state of evident defenselessness as regards the domination of powerful or simply majority community groups. Among jurists it is difficult to break the paradigm of positivism and lead them to understand that social rights are as justiciable as civil and political rights. It should not surprise us, therefore, that in the recent reform of the Mexican constitution—that can justly be seen as framed within a progressive context—no reference is made in Art. 29 to any social right. This article incorporates a list of rights that under no circumstances can be restricted or suspended: clauses engraved in stone so to speak. The article reads as follows: The exercise of the following rights [and principles] shall not be restricted or suspended: the [individual’s] right to non-discrimination, to recognition as a person before the law, right to life, to physical, mental

64  Rodolfo Vázquez and moral integrity, to the protection of his/her family, name, nationality; political rights; freedoms of thought, conscience and to profess any religious belief whatsoever; the principle of legality and retroactivity; the prohibition of the death penalty; the prohibition of slavery and servitude; the prohibition of forced disappearance and torture; the legal guarantees indispensable for the protection of such rights. It cannot be denied that the cited paragraph speaks of rights that are strictly personal, and also of civil and political rights, legal and judicial guarantees; but why does the list not include expressly (rather than indirectly through the right to non-discrimination) certain social rights—to health, education, social security, food—in terms of the safeguarding of a minimum to which every human being ought to have access? As regards this minimum, a consideration of urgency operates which warrants judges to intervene to prevent such rights being violated, or in ordering measures to be adopted to ensure its satisfaction, and thus avoiding serious damage to the person. And of course such minima in terms of basic needs or primary goods are as urgent as the safeguarding of any strictly personal right, and should be enforceable, not only in one cultural space in particular, but universally. As “universally” as is demanded by the satisfaction of basic needs of any individual. Platts would reject a political conception of human rights à la Rawls or à la Habermas. His proposal for the justification of rights is not political but moral. On this point, Griffin and Platts agree: rights must be made extensive to every individual, whether or not they belong to the Western tradition. Griffin finds support for this point in the work of Charles Taylor, Joshua Cohen, Amartya Sen and in the analyses carried out by these authors regarding Islam, Buddhism, Confucianism and Hinduism.12 But whereas Griffin puts the accent on what he calls “individual personhood” or “normative agency”,13 Platts stresses acknowledgment of the needs, interests or vulnerabilities of individuals. Let us pause for a moment to consider this distinction and these differences as compared with Griffin.

Rights and Needs Platts offers a definition of “human rights” that seeks to reconcile the “southern” and “northern” (or what in the published text he calls the “eastern”) positions. The definition is the following: There is a human right to enjoy φ when there is a sufficiently strong moral case for special legal recognition and protection of the human interest in having φ, such recognition and protection being practically feasible whether they would be innovations or are already present.14 The definition is formally impeccable, and inevitably conciliatory, but as Platts puts forward as a possible objection of Griffin: the notion would be “nearly criterionless”. Let us move on, then, to the criteria. For Griffin, as

Inflation or Deflation of Rights?  65 I have already mentioned, the criterion corresponds to “normative agency”. Let us make clear from the outset that there is no disagreement between Platts and Griffin as regards the relevance of “normative agency” in the sphere of individual morality. Perhaps Platts would prefer to speak, in order to make ourselves more readily understood, of the principle of autonomy, by which: “we must leave it up to rational, competent, agents to take the important decisions for their own lives according to their own values, wishes and preferences, free of coercion, manipulation or interferences”, as he defined it in his Sobre usos y abusos de la moral;15 or as he expressed it in his more intuitive and laconic version: “leave me in peace”. Who could cast doubt on Platts’s liberal credentials? But the criterion of normative agency or personal autonomy for determining the extension of rights seems to Platts excessively restrictive. The attempt to make of autonomy a kind of supreme value, as Griffin’s thesis for justifying human rights seems to suggest, is for Platts not a perfecting of the liberal tradition but its abandonment. For Platts, “there are human rights corresponding to categorical, serious, urgent, basic and non-substitutable human needs, whether or not these rights are recognized in society”.16 This recognition of rights, not just legal but also moral rights, corresponding to needs such as these implies not only negative duties but also positive ones on the part of the state. I can see no other way of reading the following paragraph by Platts: Far from favoring some type of laissez-faire social system, the due valuation of individuals’ autonomy presupposes a corresponding valuation of social policies on nutrition, education and health, directed at the creation of the—morally obligatory—indispensable conditions for that autonomy and its full exercise.17 It is true that in Sobre usos y abusos de la moral Platts had already devoted a complete chapter to the subject of rights and needs, and I have quoted him from that chapter. I believe that the reading of Griffin’s book, On Human Rights, led Platts to give a further turn of the screw to the subject of autonomy and needs, and to criticize Griffin’s thesis, as excessively deflationary of human rights, insufficient to deal with limit situations that were only insinuated in Sobre usos y abusos de la moral. One might say that the case of fetuses, or anencephalic babies or patients with irreversible coma enter into the group of “easy cases”, but what can one say of infants with spina bifida, of mentally defective individuals, of those who suffer from advanced dementia, paraplegia, or those crippled by polio or “trapped in non-­functioning bodies” to use Griffin’s expression? For this author, human rights should not be made extensive to these cases. But then, wonders Platts, with reason: quite what are we to say of persons trapped in non-functioning bodies? That they only have 50% of the usual human rights? That they have all such rights but that those rights only carry 50% of their usual weight

66  Rodolfo Vázquez in their case? Or even that they have no human rights at all since they simply fail to satisfy the conjunctive condition of normative agency corresponding to Griffin’s account of the idea’s components?18 For Platts, I repeat, the criterion of normative agency or that of personal autonomy is insufficient. The criterion of needs is more radical because, in the end, “the due evaluation of autonomy brings with it the evaluation of the necessary conditions for its full exercise”,19 what Dan Brock has called “broad functional capacities”.20 It is true that, despite such deflation of rights, Griffin does not deflate the moral obligations with respect to limit cases. Rather, such obligations have greater weight than the corresponding legal obligations, since not all the important moral questions, as I said at the beginning of this essay, have to become matters of law. It would be difficult not to agree with Griffin on this point if we accept the relevance of the liberal distinction between private and public life. Nonetheless, the cases we are dealing with—those of the most vulnerable, limited in the exercise of their normative agency and their personhood—cannot be confined to the private sphere, but require a determined activism on the part of third parties and the state in a sort of coactively justified paternalism. In the legal sphere, and in relation with the debate on abortion and euthanasia, Jorge Carpizo proposed a distinction, which has proved felicitous as a means for dealing with limit cases: “legal rights” are one thing; “legal interests protected by law” are another. Hence, for example, between the legal rights of women and the existence of the fetus—which still has not developed the basic bio-psychic conditions for survival—there is an asymmetry that fully justifies the protection of the former over the latter. The fetus must be protected constitutionally, as an interest that deserves protection, but it is not a subject of rights. The distinction proposed by Carpizo would, at least, work for “straightforward cases”. Could we say the same for those individuals who are in the process of development or whose agency is severely diminished, but have already developed the “basic bio-psychic conditions”? Is it enough to say that our moral obligations to these people are more powerful even than the obligations correlative to rights? Is Griffin appealing to a sort of partial altruism? I believe that this is not sufficient and that such cases are not resolved in the sphere of benevolence but in that of justice and the legal safeguarding of human rights. In view of their urgent nature, the recourse to basic needs seems to me, in agreement with Platts, the most reasonable path for establishing rights that would be adequate for seeking the satisfaction of such needs. That the latter should not be hedged in by the need for freedom or agency and should extend, for instance, to hedonistic needs, taking seriously the principle by which we assign value to pleasure and anti-value to pain for limits in cases where some bio-psychic development has occurred; and that this broadening of the criteria of needs might bring with it an “inflation” of rights in order to take such vulnerable individuals into account, must be seen as a

Inflation or Deflation of Rights?  67 moral and legal progress of possible (not impossible) satisfaction, both in courts of law and the legislative authority. The same might be said of the criterion of needs in order to give a possible response to the problem of global justice. If this is the case, I believe that we must give a cordial welcome to a moderate inflation of rights and be on the alert against a restrictive and senseless deflation.

Notes 1 Mark Platts, “The Language of Rights and Human Rights,” Philosophy, vol. 85, no. 3 (2010): 319–40. 2 James Griffin, On Human Rights (Oxford: Oxford University Press, 2008), 92. 3 Francisco Laporta, “Sobre el concepto de derechos humanos,” Doxa, vol. 4 (1987): 23. 4 Published in the Diario Oficial de la Federación, June 10, 2011. 5 See Christian Courtis and Roberto Gargarella, El nuevo constitucionalismo latinoamericano: promesas e interrogantes (Santiago de Chile: CEPAL, 2009), 32. 6 “Bentham on Legal Rights,” in Oxford Essays in Jurisprudence, ed. A. W. B. Simpson (Oxford: Oxford University Press, 1973), 200. 7 Hart, “Bentham on Legal Rights,” 166. 8 Cristina Lafont, “Justicia global en una sociedad pluralista,” Isonomía 31 (October 2009): 108. 9 See John Rawls, The Law of Peoples (Cambridge, MA: Harvard University Press, 2001), second part. 10 Cristina Lafont, “Justicia global en una sociedad pluralista,” 111. 11 See Jürgen Habermas, “A Political Constitution for the Pluralist World Society?” in Between Naturalism and Religion (Cambridge: Polity, 2008), 312–52. 12 See Griffin, On Human Rights, 25–7. 13 Griffin, On Human Rights, 44 ff. 14 Platts, “The Language of Rights and Human Rights,” 323. 15 Mark Platts, Sobre usos y abusos de la moral (Mexico City: UNAM-Paidós, 1999), 50–1. 16 Platts, Sobre usos y abusos de la moral, 100. 17 Platts, Sobre usos y abusos de la moral, 103. 18 Platts, “The Language of Rights and Human Rights,” 328. 19 Platts, Sobre usos y abusos de la moral, 103. 20 Dan Brock, Life and Death Decisions in Health Policy (Cambridge: Cambridge University Press, 1993), 282.

6 The Debate on the Abuse of the Concept of Human Rights Juan Antonio Cruz Parcero

On several occasions Mark Platts has concerned himself with conceptual problems related with the subject of rights, and particularly human rights. My points of agreement with him are many, but here I shall try to deal in some depth with certain aspects of his theses that I find problematical. In his most recent book, Ser responsable. Exploraciones filosóficas (“Being Responsible: Philosophical Explorations,” published in Spanish 2012),1 he condenses, but also restructures some of the ideas he had presented in other works, particularly Sobre usos y abusos de la moral, (“On Uses and Abuses of Morality,” published in Spanish 1999). The axis of Platts’s conceptual preoccupations lies in his attempt to distinguish between, on the one hand, “genuine rights”—that is rights sensu stricto—and, on the other, “supposed rights”—i.e., claims to objects that can be highly desirable for individuals but whose status as rights is debatable. In his most recent book he presents two approaches or models for speaking about rights and one of these he calls the conception of the South and the other the conception of the North.2 The Southern conception would be one that blurs the distinction between what would be desirable and what is in truth a human right. The Northern conception apparently takes more seriously the distinction between what is merely desirable and the idea that rights are—and must be—paradigmatically legal rights, which implies that any claim to rights must be balanced by a correlative obligation supported by institutions that exist for the purpose of resolving cases of conflicts over them.3 What principally worries Platts is what he calls the “pull towards the south”;4 a similar worry is expressed by other authors concerned with the subject of rights. Platts takes up once again this concern of James Griffin, who notes in his excellent book On Human Rights the existence of “strong inflationary pressures on the term [human rights] which have brought about its debasement,”5 and warns of “a great temptation to assert that anything to which one wants to have a legal guarantee is a human right.”6 Griffin admits that these “features of the discourse of human rights are responsible both for great good and great bad,” but he underlines the bad, this “being the ballooning of the discourse itself.”7 Platts warns immediately afterwards

Abuse of the Concept of Human Rights  69 that this Southern inflationary tendency may even be found among cautious legal theorists, and gives the example of a recognized jurist, the Mexican constitutionalist Jorge Carpizo, who speaks unconcernedly of a “right to revolution.” Platts proposes a way to reconcile the two conceptions, since the Southern conception is correct in setting in relief the need for the moral or ethical evaluation of the laws that specify rights;8 while the Northern conception has the virtue of demanding respect for the distinction between “mere desirabilities” and rights. The way proposed by Platts consists in upholding that “there is a human right to enjoy φ when there is a sufficiently strong moral case showing that the human interest in having φ merits legal recognition and protection, such recognition and protection being practically feasible.”9 The first question I shall discuss in this paper is the thesis that the proliferation of rights is a problem; I shall try to understand this phenomenon and see its dimensions, since much of the perspective adopted by Platts—and other authors such as Griffin—depends on seeing in this phenomenon a serious danger that must be avoided or counteracted. I shall then set out to analyze Platts’s proposal for overcoming the problem, and I shall emphasize what seems an interesting way—if I may be permitted the expression—to administer the problems set before us by the phenomenon of proliferation.

1. The Proliferation of Human Rights What do we mean when we warn of the danger of inflation of the concept of human rights (or of rights in general)? What kind of danger is it? Griffin speaks of “goods” and “bads” regarding the proliferation of rights. What are these goods and what are the bads? To what does he refer when he describes the mere inflation of the discourse of rights as a “bad”? This affirmation seems strange when he himself has said that he is not against the proliferation of rights: I should stress again that what moves me is not the wish to reverse the proliferation of rights. I have no views about how many human rights there are. Nor, given the different levels of abstraction in their formulation, do I know how to enumerate them. We speak of “proliferation,” in a pejorative sense, only because we suspect that some of the declared rights are not true rights, What moves me is the wish to end the damaging indeterminateness of sense of the term “human right.”10 I am not sure whether the danger Griffin sees in the indeterminateness of sense of the notion of “human right” is the “bad” of which he speaks, but I suspect that the indetermination of the concept is a different question from the problem of the proliferation of rights, although clearly there may be a relation between the two. Of course, this concern does not arise out of mere analytical niggling, but responds to certain actual consequences.

70  Juan Antonio Cruz Parcero Platts refers to these when he recalls a discussion that took place in a forum on human rights and disability in Michoacán, Mexico. The discussion was marked by polarized stances; while public officials embraced a Northern conception (loath to admit any possibility not sanctioned by prevailing legislation), members of the public manifested a frivolously Southern conception (making “free and easy” use of the language and notions of human rights).11 What concerns Platts is a conceptual problem that goes beyond that of communication between people whose conceptions are different; he suggests, rather, that both conceptions are informed by mistaken notions: those of the South because they cast what is merely desirable in terms of rights and those of the North—at least this is my understanding of Platts— for being excessively cautious about accepting the existence of rights not regulated by law. To sum up, the danger seems to consist just as much in the legalistic attitude of the North as in the frivolity of the South. I would like to clarify what there is of good and bad in the proliferation of rights, or in other terms, in the inflationary phenomenon of the discourse on rights. I shall thus pause for a moment to try to arrive at a more precise notion of this phenomenon, for perhaps some misunderstandings arise when we take for granted that everybody understands the phenomenon in the same way. 1.1. The Concern Over the Proliferation of Rights Both Platts and Griffin share what is a rather common concern among specialists in human rights: a widespread complaint against the proliferation of alleged rights, which is said to be the cause of a devaluation of the very discourse of human rights, since this ends up being used to defend merely utopian aspirations. Since the signing of the two international pacts on human rights, (the International Covenant on Civil and Political Rights and, especially, that on Economic, Social and Cultural Rights, both adopted by the UN General Assembly in 1966), voices began to be heard warning of the danger of the proliferation of rights. The objection against social rights formulated in 1967 by Maurice Cranston is, for instance, very well known.12 The critics single out—as Platts himself does—certain items from the Universal Declaration of Human Rights (1948) itself, or from other tracts or pacts, whose status as genuine rights is dubious. For Platts those of dubious validity include the right to free choice of employment, to protection against unemployment, to just and favorable remuneration, to rest and leisure, to regular holidays with pay, to food, clothing, housing and medical care, to education, to enjoyment of the arts and to a share in scientific advancement and its benefits.13 Other authors include in their critiques rights recognized by the Human Rights Courts such as the right to sleep soundly at night (Hatton case),14 or more general ones such as the right to strike, the right to revolution, the rights of animals, the right to happiness; but also the rights of children, the reproductive rights of women (including the right to

Abuse of the Concept of Human Rights  71 abortion), and the rights of indigenous peoples to use, and be educated in, their native languages, etc. For some authors the phenomenon of the inflation or proliferation of human rights arises when the attempt is made to extend the protection recognized in the principal conventions on human rights to new and different areas—such as armed conflicts, humanitarian intervention, social rights, labor rights.15 One of the partisans of reducing drastically the list of human rights is Rawls,16 whose starting point is the distinction between the constitutional or fundamental rights that operate in the domestic sphere and the human rights that ought to prevail in the international arena; the latter should be few and especially important, so that they can be recognized by states that do not subscribe to doctrines of a liberal kind. The notion of human rights should thus take in only those rights to which non-liberal (hierarchical) nation states can subscribe. Rawls leaves out of his list of human rights many of those based on autonomy, such as freedom of expression and of the press. As some authors have pointed out, it would seem that Rawls is attempting “an unstable compromise between liberalism and communitarism,” but as part of this compromise he is prepared to accept serious restrictions and intromissions of a paternalist type against some individuals and groups, particularly women.17 It should be noted that critiques of the proliferation of rights sometimes reflect something more than a concern for distinguishing between genuine rights and the merely desirable. The problem is not just a conceptual one; concerns of a political or ideological nature also tend to insinuate themselves here. Why should health, education, a certain degree of protection against unemployment, food, housing and other such needs have to be seen as mere aspirations, as utopian, and not as genuine rights? Why would the stipulation of what constitutes a true human right have to depend on non-liberal nation states? Why should the so-called social rights tend to be seen as a problem? What is it that gives the oldest formulations on rights their higher status than more recent ones? Attempts to present conceptual or technical-legal arguments with the aim of delineating radical differences between many of these suspect rights and the so-called civil and political rights have failed; it is not that no differences exist, but that those that exist are insufficient as arguments for not recognizing the status as rights—indeed as human rights—of many of those allegedly false rights.18 The tendency of the Northern vision to emphasize the legal character of human rights is often no more than a way of upholding the status quo. Platts’s observation that the wish to distinguish between the desirable and what can really be claimed is what underlies the Northern concept does not therefore seem to me true (perhaps it is true in the case of Bentham, but not where Kelsen is concerned). The Kelsenian definition, for example, widely adopted in Europe and Latin America during much of the twentieth century, consists in a fallacy of petitio principii that tells us that the rights that can

72  Juan Antonio Cruz Parcero be claimed or guaranteed legally are those that are guaranteed or protected legally; in other words, a genuine right would be one in respect of which the legislator has already taken measures for protection by legal means. Therefore, what has to be regarded as a right and what does not is a matter to be defined by legislative practice. Now, in the legislative field, many determinations upon rights boil down to a problem of public expenditure, and hence the wish to curtail the expansion of rights could be a simple matter of fiscal responsibility. Indeed, the neoliberal argument against the proliferation of rights is characterized exclusively by concerns of an economic nature, since the more rights there are the more expenditure is required—which, of course, implies that taxes have to be raised. However, the fact that rights may well turn out to be expensive is one thing; whether they are economically feasible is another. And what we tend to find is that not all rights that are seen to be expensive are rejected as such, but only those that have been identified doctrinally as “supposed” rights.19 Hence only those rights that combine high cost with lack of prestige end up being characterized as unfeasible; hence one may doubt whether the idea of feasibility is being taken seriously by the Northern vision. It is, however, only fair to add that the reasons for recognizing rights also sometimes reflect complex political or ideological motivations. The adoption of the two International Agreements on rights to which I have referred are a good example of how political-ideological considerations ended up playing a preponderant role. 1.2. What Does the Proliferation of Rights Consist in? If we analyze the concerns of different authors regarding the proliferation of rights, we find four different aspects or phases: (i) a process of recognition and declaration of abstract rights; (ii) a process of concretion or specification of rights that allegedly derive from other more abstract rights; (iii) a process of application of rights by judges and courts; and finally, we can also speak of (iv) a process of appropriation by citizens or the general public of the language of rights. 1.2.1.  The Recognition of Abstract Rights Following the end of the Second World War a Universal Declaration of Human Rights was signed (1948); this was followed by the two important International Covenants referred to above, which came into effect in 1966 and were to a considerable extent a reflection of the ideological dispute between the two political blocs in the framework of the Cold War. These first lists of human rights enabled a certain international consensus to be reached, but henceforth were to generate the first suspicions that something was not quite in order regarding certain specific rights. At present all the members of the United Nations have signed at least one of the most

Abuse of the Concept of Human Rights  73 important treaties on human rights and 80 percent have signed more than four of the most important treaties.20 The first declarations on rights included principally abstract rights; however, despite their abstraction, not all of them have been accorded the status of “universal” rights, if we understand by this term rights that are attributable to all individuals by the sole fact of being human beings. This idea of “universality” generated many problems, for there were certain rights that—despite enjoying a broad consensus—did not satisfy that characteristic of universality. Hence the problem of what is to be understood by universality of rights has been present ever since, and attempts have been made to reconstruct this idea in various ways which all present problems of their own. 1.2.2.  The Specification of Rights Shortly after the promulgation of the Universal Declaration of Human Rights and the abovementioned Conventions, a process of specification began, making provision for different circumstances and contexts. The international bodies were not slow to realize that declarations and conventions had to be brought down to earth in order to provide solutions for particular situations and contexts that required specific actions by international organizations. Thus began a process of expansion that we might better describe as one of concretion or specification on the basis of the more abstract recognized rights. The defense of those rights required attention to specific circumstances and problems affecting parts or sectors of humanity. A start was made on attending to problems arising from colonial domination, discrimination against specific groups (especially women, children, older adults, the disabled, etc.), specific problems such as genocide, torture, armed conflicts, the trade in human beings, intra-family violence, exclusion and marginalization of indigenous or aboriginal groups, etc. This specification process thus generated an expansion in rights (and duties of states), but an expansion that seemed to be necessary in order to advance from what was desirable in the abstract to what was necessary in practice in order to protect those human beings most prone to suffer violations of their rights. During this phase, theories of rights began to be rethought. Some important contributions arose from these reflections. I think the distinctions made a century ago by W. N. Hohfeld21 have proved of enormous utility for understanding this type of process. Hohfeld’s idea was basically that every abstract right can be broken down into more concrete rights, and that such concretions would require the use of more precise notions, since speaking of rights in the abstract was often of little use in practice and generated much confusion. Hohfeld proposed four types of legal relations in order to explain how to break down an abstract right. His ideas, on which I cannot comment more at this moment, have been widely accepted (although with nuances and reformulations). If what concerns us about rights is how to

74  Juan Antonio Cruz Parcero apply them and how to resolve conflicts, abstract formulations are of little or no use; we shall thus have to seek to give greater specificity to their contents and establish specific claims, freedoms, competencies and immunities. To pass from abstract formulations to more concrete ones was seen as also a theoretical necessity. H.L.A. Hart22 warned that in the process of translation or specification it was essential to say what relations or specifications had to be seen as essential—the “core of a right”—since otherwise one might lose the sense of what is a specification of a basic right. The idea was thus that every abstract right needed specifications, but that these specifications cannot be merely contingent and arbitrary. After an initial phase of concretion of rights, came a phase of recognition of new abstract rights (although this phenomenon is more subject to contention if we consider that it was hardly possible for any new right to arise that was not already implicitly formulated or derivative from one or other of the abstract rights recognized initially in the Universal Declaration). For instance, a right that is now accepted is the right to water in some form or other, but this right—although it is set forth as a very abstract right that would take on different aspects (water for drinking, water for irrigation, etc.)—could itself be seen as a concretion of some aspects of the right to life, the right to health, the right to decent food, etc. In this phase, what to my mind are more or less abstract formulations and which are highly contestable also arise; such are, for example, the rights of animals, the right to happiness, etc. 1.2.3.  The Justiciability of Rights The third phase is that of “justiciability” or judicial applicability of human rights. The boom in human rights generated the possibility of creating regional instruments for their adequate protection through the creation of International Commissions and Courts of Human Rights. Human rights claims and the judicial application of such claims (or their application via international non-judicial bodies) to concrete cases of alleged violations of rights generated a different way of concreting rights. New rights began to proliferate. What characterized this process was to a considerable extent the need for a means of balancing different rights when they come into conflict in a specific case, or of weighting rights against other kinds of social or national interests. Hence a large body of specifications have seen the light of day—many of course challenged or subject to doubt—that serve to give greater definition to the scope of more abstract rights in specific cases, but which may also lead to the conceiving of better abstract formulations. This process of application has led to the formulation of specific rights that may seem strange or absurd if taken out of context. We may also, of course, demur to rulings in particular actions, but such differences of legal opinion are in any case natural. It should moreover be pointed out that, along

Abuse of the Concept of Human Rights  75 with such new “rights,” certain rules and principles have come into being as regards the application of rights in general and concerning the priority of some over others in specific cases of conflict. Thus it is relatively easy to predict what courts or human rights organizations will do when cases are presented bearing similarity to others that have already been adjudicated; as commonly happens in domestic jurisdictions, the rules or principles established by such institutions are applied. These three characteristics of the phenomenon of proliferation can be regarded as essential to the very process of defense and application of rights. We might say they are the consequence of taking rights seriously. What is, then, the real extent of the danger inherent in this expansion or proliferation of rights that is so worrying to some theorists? From Cranston—who first objected that their expansion supposed a translation of rights to the sphere of the merely desirable—to the ultra-minimalist Rawls—whose proposed list of human rights contemplated only those whose violation falls within the jurisdiction of the International Criminal Court or might justify action by the United Nations Security Council, including intervention by the international community in a sovereign state—it would seem that proliferation is a very serious problem that has to be resolved by means of some kind of radical pruning. Even Griffin—despite warning that proliferation has something good to offer—seems to be in agreement with the idea that the “ballooning” affecting the concept of human rights represents in itself a danger. If the worry consists in the fact that, in this process of expansion or proliferation of rights, some claims to rights that are not valid are treated as if they were, it seems to me that the concern is genuine. Nevertheless, the dimension and degree of seriousness of this evil depends much on the evaluation we make of the process of proliferation of rights in general. We cannot evade this point. If we simply set off from the idea that the less rights we have the greater value will accrue to each—on the lines of the traditional wisdom enshrined in proverbial phrases like de lo bueno poco (“good things come in small packages” or “you can have too much of a good thing”) or el que mucho abarca poco aprieta (“grasp all, lose all” or “don’t bite off more than you can chew”)—I think we are committing a gross simplification. Whoever thinks in this way is either eluding the effort of making a proper diagnosis of the phenomenon, or is perhaps captive to a facile attribution of the disappointing progress in the human rights field (a certain regression has even been noted) to the proliferation in the discourse referring to them: incontestable evidence, supposedly, of an inadmissible mental laxness. My own view is that the history of human rights shows their increasing importance and that this importance is reflected in the phenomenon of proliferation. Hence it is false to believe that the more rights we recognize the less importance they will have; as I have already stated, their proliferation is a necessary consequence of taking rights seriously. I must, however, recognize that this explanation of the phenomenon of the proliferation of rights

76  Juan Antonio Cruz Parcero is quite independent of the matter of whether the rights we are confronted with are true or false. As John Tasioulas puts it: the international regime of human rights is not morally self-validating; instead its legitimacy depends upon conformity with independent moral standards, including genuine human rights. So the intended short-cut through law and political practice rapidly leads back to our original problem. Although philosophers disagree about whether or how the problem can be solved, many of them agree on the general character of any adequate solution. According to this standard picture, as I shall call it, which human rights exist is a moral question to be distinguished from the predominantly institutional question of the extent to which they are recognized, respected or enforced.23 Nonetheless, although it is important to keep this distinction in mind, in practice moral and institutional aspects are not completely independent of each other; the contents of rights are also conditioned by specifications dependent upon institutional and contextual criteria. For human rights are not obtained from a mere exercise of reflection on human needs in the abstract but, rather, their contents are in part determined by specific historical circumstances. In other words, the formulations we make of human rights—both of certain general rights and their concrete specifications— depend upon their historicity. The right to a fair trial, for example, depends on the institutions of administration of justice that we possess (and to which we can reasonably aspire at a particular stage of history). As Griffin remarks, there is a difference between the (more abstract) right to freedom of expression and the specific right to freedom of the press;24 the latter might not actually be necessary in a society in which the press does not exist, not because it has been prohibited, but because it has still not arisen. Historically, societies have existed—and some still do—in which mass communications media (press, radio, TV, etc.) have not yet been conceived, without this implying any kind of violation. To speak of a “right” to press freedom implies taking into account the features of the type of society and the needs of individuals in that society. In any process of legislation or judicial application there is always the risk of making a bad justification of certain claims that are formulated in terms of rights, of making bad interpretations or applications of rights, of extending their scope unjustifiably to cases that are not analogous, of wishing to cover any problem or conflict where legal voids or loopholes exist (or where the moral solution is controversial) via the postulation of an equally controversial human right. So what I do fully share is a desire to see this historical process of specification and proliferation of human rights accompanied by a process of reflection and review of the recognized lists of rights, along with a review of sentences adopted by the courts in relation to human rights. This is not a process that should be under the control of legislatures

Abuse of the Concept of Human Rights  77 or some particular court, but a broader process on which I shall go into greater detail below. 1.1.4.  The Popular Appropriation of the Language of Rights The final facet of the phenomenon of the proliferation of human rights is one that particularly worries Platts. What he calls the pull to the South would be a consequence—one of the various possible ones—of the appropriation of the discourse of rights by the general public and, specifically, by the social and political organizations that fight to defend particular or collective interests: a struggle and defense that takes place via the appropriation of the language of human rights. Certainly, this appropriation by the general public, in which claims are cast in terms of rights, tends to blur the line between the desirable and feasible: everything desirable can be demanded in terms of human rights. But this is only one aspect of a phenomenon that is of much greater complexity and importance, for it is through the popular appropriation of the discourse of human rights that societies become aware, detect problems, demand solutions and responsibilities of the state, etc. It is of course important that society and social actors do not end up distorting the discourse of human rights, for a sensible understanding of these is necessary in turn for channeling effective social action.

2. Platts’s Proposal: Another Perspective for Dealing with the Problem of the Proliferation of Rights A very suggestive point made by Platts is his denial that there is any reason why an “account of human rights need [. . .] take the form of a theory yielding a relatively straightforward way of telling which those rights are.”25 Against the strategy of Griffin that is based on grounding “genuine rights” on the idea of agency, Platts proposes a more pluralist idea that recognizes basic needs, certain human interests and certain vulnerabilities or dependencies of human beings that may serve as a foundation for human rights. The idea of controlling the inflation of human rights by way of reducing the number of those we regard as their bearers (which is equivalent, Platts holds, to a limitation of the relevant characteristics of human beings), does not seem quite to respond to the problems that concern us when we analyze the phenomenon of the proliferation of human rights. It seems, as Platts avers, that a proposal like that of Griffin “seems to border on overkill” when it denies to certain human beings all rights with the purpose of preventing the proliferation of some pseudo-rights.26 First, because it would seem that if we consider agency as the only criterion we will end up leaving outside the idea of entitlement to rights human beings who normally—and in the terms of the existing international discourse on human rights—we would consider entitlement holders to certain rights. Second, because there

78  Juan Antonio Cruz Parcero are other problems involving the phenomenon of inflation of rights that this measure fails to resolve, for what is at stake is not the problem of who is to be considered an agent but, in every case, a question of how to exercise and protect such agency and of defining the limits to its exercise and protection. I am totally in agreement with Platts in that there are other important moral reasons that do not rest on questions relative to agency or autonomy, reasons that we do not have to remove from the justification of human rights. And this does not mean that I disagree with Griffin in his considerations on the relevance of agency, only that unlike him I can accept that there are other relevant moral values that do not depend on the individual’s agency. But, returning to the subject of “inflation” and Platts’s proposal, I find it interesting to reflect on the following passage: Acceptance of the claim that some particular right exists always incurs a potential cost, be this a literal, economic one or be it one to be understood in terms of a reduction of liberty, autonomy or some other valued thing; so it is that, once faced with a claim as to rights, some ­counter-claim as to supposed rights relating, directly or indirectly, to non-­incurrence of that potential cost can itself often reasonably be made. But acceptance of this counter-claim would itself incur a potential cost too, possibly over and above that involved in any denial of the original claim as to the existence of some particular right (assuming, that is, the original claim to have had some at least minimally reasonable grounding); and so the interchange thus initiated need not stop here. In cases, however, in which both claimant and counter-claimant ground their pretensions in moral considerations, the possibility can arise that at some point they come to some reasonable agreement as to the comparative moral weighting of their respective moral grounds. . .27 Following Wiggins28 on this point, it seems to me that what Platts is suggesting is that the affirmation of the existence of rights—genuine rights—has to be seen always as part of a network of “relativeness” linking rights to other interests (the needs referred to by Wiggins) which is underdetermined; that is, that we have to make this relativeness clear at any moment when we wish to make a justified application of rights. The subject of control of inflation seems to lead us along a quite different path from that of Griffin. The search for a “reasonable agreement”29 between the rights in conflict (en pugna) or between the rights and the costs implied by their protection brings us onto the arena of what Platts calls a “kind of morally-grounded dialogue: a dialogue [. . .] in which all participants may be one and the same person.”30 It is within this moral dialogue that we have to demand that rights proliferate; it is within this dialogue that we have to offer moral grounds (and also other types of reasoning, consequentialist arguments, for instance, regarding the feasibility of rights, arguments regarding the kinds of legal protection that may be offered, etc.)31 whenever we wish to deny some particular claim the

Abuse of the Concept of Human Rights  79 status of a right. It is within this dialogue too that we may accept concretions or specifications of more abstract rights already accepted, and even accept new rights. Such a dialogue would have to colonize the juridical and politico-economic fields just so as to give pride of place to the idea that human rights cannot depend upon the capricious notions of a legislator; there must, however, be a practical discourse that is not content merely to proclaim a world where these beautiful ideas exist, but that seeks to translate them into the actions and omissions required by the rights themselves; for this purpose the aspect of feasibility underlined by Platts is important. Feasibility itself must be a central topic of discussion in this moral dialogue. And to this end, many different kinds of elements must be considered for an adequate evaluation of whether something is feasible or not. But the challenge is complex: on the one hand, the notion of feasibility must be related with the affairs of the world, i.e., to a large extent it is an empirical category; on the other, however, it needs to be treated as a normative category, for the notion of feasibility does not necessarily have to be circumscribed to what is or is not possible today; the feasible is also a part of what we can imagine and construct for the future. Without ignoring the danger of postulating as a right what is merely desirable, one ought to recognize that there are human rights in respect of which we have a duty to imagine and construct conditions of possibility, even if we accept that at the present moment their realization is not possible to an acceptable degree. The problem with the distinction that Platts uses between the merely desirable and genuine rights is that the latter are in themselves desirable and what converts them into genuine rights is apparently, in Platts’s judgment, the fact that they can be translated into justiciable rights (and not only the degree that they actually enjoy legal recognition, if I understand him correctly), and it is at this point that the idea seems to me problematic. When Platts speaks to us of legal protection he seems at times to be referring to what is generally known as the justiciability of rights, that is to say that they can be made subject to legal claim. But his thesis speaks only of “recognition and special legal protection,” which is rather vague. As far as the former way of understanding legal protection is concerned, if in general terms one can accept that human rights have to be justiciable, that does not mean that all of them have to be so. If by justiciability of a right we understand its propensity to be legally claimed, I think that this ties us to a limited conception regarding how rights can be realized and protected in an effective way. While not wishing to dispute the idea that rights ought to be justiciable, I would like to point out that the efficient and effective protection of rights does not necessarily depend on judges and courts of law. There are many ways to protect human rights effectively that fall into the category of economic and public policy nature measures; there are, besides, many forms of social and educational organization that contribute to the realization of and respect for rights; many of these things have no reason to be necessarily linked to the work of the courts, although, as I have set forth

80  Juan Antonio Cruz Parcero elsewhere,32 good institutional design entails the justiciability of rights as one recommendable (if not precisely necessary) elements of a good design of public policies. Let us also remember that an important aspect of rights is that they serve to justify the creation and imposition of duties (and not vice versa); hence the establishment of fully specified rights—which to a large extent have to depend on the reasonableness and feasibility of compliance with the said duty—cannot be something that conditions the existence of a right, as those who subscribe to something like the Northern conception tend to imagine. Rights often serve to justify a series of measures aimed at creating the conditions of possibility for the imposition of duties upon specific subjects or institutions (even for first creating or transforming institutions in line with such purposes). At what point, then, does one have to judge that a genuine right has come into existence? I think the answer cannot consist in holding that the function of moral reasons points towards justifying the justiciability of the interests involved, as Platts seems to hold at some moments. Nonetheless, if we hold to the idea of a broader legal protection—one that is not limited to justiciability but to any measure that might work and contribute to the protection of rights—then I would see no problem with the thesis. Let us now return to the idea of the morally grounded dialogue. The idea seems to me very similar to that of Dworkin when he seeks to explain law as a social practice.33 For Dworkin, to know what law is, or rather, what is the law applicable to the case in point depends on a series of interpretative operations where, in the final analysis, moral arguments may be decisive for choosing between different options to resolve a particular case. What “law” is depends, in the final instance, on the best answer that can be offered in the light of current legal material, and on a series of theories and values that serve to justify our conception of right. To know what law is applicable in a particular instance is the outcome of a process of argumentation that seeks to justify decisions. Similarly, we might say that the idea that there are genuine rights is a notion that we have to construct and justify at the moment of applying declarations of rights, treaties, constitutions, jurisprudence, etc. Genuine rights are those rights that appear as justified, not only in the abstract, but especially when we apply them to resolve disputes, to legislate in particular areas, or to design policies that prioritize some rights over others. The subject of rights, it must be understood, moves in this moral dialogue to which Platts alludes, a dialogue—even, why not, a dispute or ­controversy—where we move generally from abstract rights to more specific rights, where some rights limit others, where often we have to assume the costs implied by certain measures for protecting rights. We move in a reasoning on the basis both of principles and rules, not only legal but also moral. For example, does the right to life include the right of a patient in a terminal situation to die? And they also set off from conflicts between laws and from the question of how one arrives at a reasonable weighting among them; for example: does the right to the free expression of ideas include the

Abuse of the Concept of Human Rights  81 right to offend others in their religious or moral beliefs? Or whether reproductive freedom includes the right to use techniques of fertilization in vitro that imply the necessary destruction of some human embryos? What is important in the idea of Platts that I have underlined is that the strategy for controlling the proliferation of rights cannot be reduced to a theoretical stipulation, but that it has to set off from a plurality of values that are articulated reasonably in a moral dialogue. In my view it is a central labor of philosophers and jurists to enter in depth into the functioning of this dialogue, in order to understand better how it works and how it would have to function in order to seek for an adequate justification of human rights and to be able to question false rights.

Notes 1 Quotations are taken from the as yet unpublished original in English; page references are to the published Spanish version. 2 The published essay refers to the “concepción del Este,” but it can be assumed that Platts was merely trying to keep the arguments out of the sphere of geopolitics. 3 Mark Platts, Ser responsable. Exploraciones filosóficas (Mexico City: UNAM, 2012), 147–8. 4 Platts, Ser responsable, 149. 5 James Griffin, On Human Rights (Oxford: Oxford University Press, 2008), 92. 6 Griffin, On Human Rights, 92; Platts, Ser responsable, 150. 7 Platts, Ser responsable, 150; Griffin, On Human Rights, 92. 8 Platts, Ser responsable, 152. 9 Platts, Ser responsable, 153. 10 Griffin, On Human Rights, 93. 11 Platts, Ser responsable, 149. 12 For Cranston the problem arose from the Universal Declaration of Human Rights (1948) itself, which introduced a series of economic and social rights, of a different nature from the traditionally accepted civil and political rights. His argument against such social and economic rights is that they cannot be made into positive rights, that is to say, legal rights. For him, civil and political rights can be guaranteed by means of simple legislation, whereas social and economic rights cannot (Maurice Cranston, “Human Rights, Real and Supposed,” in Political Theory and the Rights of Man, ed. D. D. Raphael [Indiana University Press, Bloomington, 1967]; reprinted in The Philosophy of Human Rights, ed. Patrick Hayden [St. Paul: Paragon House, 2001], 163–73). This conception now seems somewhat naive, but at that time it was a dominant notion that arose, of course, from the belief that rights, as they were traditionally understood, could only be correlated with negative obligations; social and economic rights necessarily entailed positive obligations. 13 Platts, Ser responsable, 147. 14 Hatton and Others vs. The United Kingdom, ruling no. 36022/97 of the European Court of Human Rights at Strasbourg July 8, 2003. In this case the Court decided that the regulations of Heathrow airport would have to be modified due to the excessive annoyance caused to people living in the vicinity by noise, especially at night. 15 George Letsas, A Theory of Interpretation of the European Convention on Human Rights (Oxford: Oxford University Press, 2007), 4.

82  Juan Antonio Cruz Parcero 16 John Rawls, The Law of Peoples (Cambridge, MA: Harvard University Press, 1999). 17 See William J. Talbott, Which Rights Should Be Universal? (Oxford: Oxford University Press, 2005), 11–12. 18 The very definition of what is a right is a subject that has been much debated and in which, precisely, ideological aspects have been intermingled in the guise of legal-theoretical arguments (or legal-technical arguments). Such is the case of those who hold that the “first generation” rights (civil and political liberties) are the only genuine rights, arguing alleged structural differences of these rights as against the so-called economic, social and cultural rights. It is, for instance, common to appeal to the false idea that the former are correlative of negative obligations and the latter of positive obligations. Those who, on the other hand, defend a technical-legal definition—I am thinking of definitions inspired by Bentham or Kelsen that are shared by many legal positivists and a majority of jurists (those of the North to whom Platts refers)—are party to a stipulative reduction, which tends to be of little help in resolving these controversies. 19 See Stephen Holmes and Cass Sunstein, The Cost of Rights. Why Liberty Depends on Taxes (London-New York: Norton, 1999). 20 James Nickel, Making Sense of Human Rights, 2nd ed. (Singapore: Blackwell Publishing, 2007), 92. 21 Wesley Newcomb Hohfeld, Fundamental Legal Conceptions as Applied in Judicial Reasoning (New Haven: Yale University Press, 1913). 22 H.L.A. Hart, Essays on Bentham: Jurisprudence and Political Theory (Oxford: Clarendon Press, 1982). 23 John Tasioulas, “The Moral Reality of Human Rights,” in Freedom from Poverty as a Human Right: Who Owes What to the Very Poor? ed. Thomas Pogge (Oxford: UNESCO-Oxford University Press (2007), 75. 24 Griffin, On Human Rights, 37 ff. 25 Platts, Ser responsable, 170. 26 Platts, Ser responsable, 172. 27 Platts, Ser responsable, 173–4. 28 David Wiggins, “Claims of Need,” in Needs, Values, Truth (Oxford: Blackwell, 1987), 1–57. 29 I understand the expression “reasonable agreement” not necessarily in the sense of parties coming to an agreement as regards the relative weight of their claims; even when this is not possible in a fully legalistic sense, “agreement” can take place in the sense of an ability to reflect on and discuss rationally such claims. 30 Platts, Ser responsable, 174. 31 See Talbott, Which Rights Should Be Universal? 32 In Juan Antonio Cruz Parcero, “Los derechos sociales y sus garantías: un esquema para repensar su justiciabilidad,” in Los derechos sociales en el Estado Constitucional, eds. J. Espinoza de los Monteros and J. Ordóñez (Valencia: Tirant lo Blanch, 2013). 33 Ronald Dworkin, Law’s Empire (Cambridge, MA: Harvard University Press, 1986).

7 Convergence or Divergence in the Evolution of (Criminal) Rights? A Case Study of the Multiple Incoherencies of the Presumption of Innocence* Larry Laudan No legal principle of criminal law and procedure has generated more interest and debate than the rule that one is presumed innocent until proven guilty in a court of law. Quintard-Morenas1

Introduction The paper begins with a confession: as time goes on, I become increasingly cynical about many of the rights now usually touted as the universal, core rights of every criminal defendant. Such doubts don’t spring from a general skepticism about legal rights, whether those of criminal defendants or ordinary citizens; they are rooted, rather, in a strong suspicion that many of the most widely acknowledged rights for defendants are ill-defined, lacking a clear rationale, and incapable of justifying the many public policies (often in the form of rules of evidence and procedure) that are, usually mistakenly, identified as corollaries of these rights. I have chosen to examine here the right to a presumption of innocence (hereafter: PI), not because it is more than usually incoherent, but because it is almost universally acclaimed as one of the bedrock rights, potent enough to allow us to derive from it all manner of rules of evidence and procedure for criminal trials. The first part of the paper seeks to show that—despite the near-universal obeisance paid to the PI—there is no widely shared view of exactly what the PI requires in terms of rules and procedures. Different jurisdictions and different constitutional courts construe the PI in vastly different (and frequently mutually inconsistent) ways. Having illustrated some of its multiple ambiguities of interpretation and of application, I will argue in the second part of this paper that the PI is not only deeply ambiguous in terms of the practices that it justifies or excludes but fundamentally incoherent in its most familiar formulations. We will see that, depending on how one interprets this “presumption” (which is technically not a presumption at all), it is either (i) demonstrably false, (ii) a mere tautology, (iii) impossible to comply with in its demand for a belief in innocence prior to conviction,

84  Larry Laudan or (iv) a rule laying down restrictions on criminal justice that no jurisdiction in the world either does, or should, obey.

Part I: Existing Divergences Among Legal Jurisdictions and Scholars About What the PI Does and Does Not Entail, Justify or Exclude It is a commonplace among legal scholars concerned with comparative and international law that the “developed” countries of the world, along with many of the “developing” ones, share a strong and rapidly converging consensus about the most important rights of criminal defendants. This process of “harmonization,” as it is often called, is felt to be particularly powerful in Europe but is also said to extend significantly beyond that continent’s courtrooms. The existence of such supposed commonalities in a large variety of legal systems is taken by many authorities as a sign that the world’s legal systems are converging upon something like a common or shared code of procedures and rules of evidence, at least where the most fundamental rights are concerned.2 The features of the PI to be reviewed in this section will challenge the plausibility of the argument that we can soon expect a general harmonization of the rights of defendants in legal systems around the world.3 I will argue that we are much further away from convergence about those rights than this idyllic, consensual picture suggests. We shall see that almost every legal system has its own construal of what the PI means; that is, each has its own conception of the concept in question, which often diverges—sometimes sharply—from the conception operative, under the same name, in a neighboring jurisdiction. Because that is so, it is crucial to understand that there is no unanimity about the content of the presumption of innocence, (nor, for that matter about the right to silence, or the right to a demanding standard of proof), despite the fact that legal systems around the world proudly proclaim their reverence for such concepts. The arguments to follow will show that, while (to mention only a few) the English, the French, the Spanish and the Americans all claim that the presumption of innocence (or the right to silence or to a demanding standard, etc.) is a key element in their respective systems of criminal justice, their agreement amounts to little more than advocacy of the same string of words, a happy coincidence that disguises some profound disagreements about what the concept of the presumption of innocence requires and what its implementation entails. Accordingly, this section will a) document some of the more striking differences in the construal of the PI in a handful of legal systems and b) explore some of the philosophical consequences of the continuing and deep disagreements among nations about how to interpret the core rights of criminal defendants. A. Challenging the Illusion of Widespread Agreement about Rights. Nearly all Western legal systems incorporate what each chooses to call “the

Evolution of (Criminal) Rights  85 presumption of innocence.”4 In many countries, such a presumption is written explicitly into the national constitution; in others, it is articulated in legislation or by constitutional jurisprudence. More than including a PI, the legal literature (academic as well as jurisprudential) of most countries insists that the PI is a, if not the, bedrock principle for structuring many of the features of the criminal justice system.5 It is frequently said, for instance, that the PI justifies and requires a demanding standard of proof, necessitates the imposition of the burden of proof exclusively on the state, undergirds the right to bail, protects the silence of the defendant, and (as we shall see shortly) shapes in crucial ways all sorts of other rules of evidence and procedure. In the next few pages, I will summarize a few of the striking differences in the ways in which distinct jurisdictions construe the meaning of the PI. 1)  Pre-Trial Practices In virtually all criminal jurisdictions, the PI is thought to apply during the trial itself (though, as we shall see, usually not in the same sense). There are, however, many situations that occur before trial when some jurisdictions believe that the PI is operative and others flatly deny its pertinence. Here are some of them: Bail Granting In most civil law jurisdictions (as well as in England and Wales), the decision about granting bail cannot take into account the strength of the existing case against the defendant. The basis for refusing to acknowledge the pertinence of such information in such systems is the claim that the PI must be honored until full proof of guilt has been given. Since a bail hearing cannot possibly adjudicate guilt fully, so the argument continues, the judge in a bail hearing must not even attempt to guess whether the bail seeker is innocent or guilty. In such a situation, and given the PI, the judge at a bail hearing must presume the innocence of all those who come before him, in which case bail cannot be denied simply because the case against a defendant looks powerful. (Civil and English courts are reluctantly willing to deny bail under certain other circumstances but not because the case against the defendant makes his guilt likely.) In the US and Canada, by contrast, bail can be and often is denied when, among other things, the case against the accused already assembled by the prosecutor seems quite powerful to the judge presiding over the bail hearing.6 The North American practice of allowing the weight of evidence against the defendant to be appraised in a bail hearing, and utilized as a decision rule for granting or denying bail, is perceived by civil law jurists as tantamount to a preliminary finding of guilt when the defendant has not even been properly tried. Not surprisingly, bail denial occurs more frequently in US and Canadian courts than in most civil law ones.7

86  Larry Laudan The differences between various jurisdictions about the pre-trial applicability of the presumption of innocence go well beyond what sorts of factors can play a role in decisions about the granting of bail. For instance, the US Supreme Court has held that the PI is a doctrine governing trial procedure and only trial procedure; as such, it has no relevance whatever to determining the rights of a defendant prior to trial.8 By contrast, in many jurisdictions, the PI enjoys a much larger scope and is “broadly concerned to protect the dignity of the individual against the power of the state by prescribing that individuals should be treated as innocent throughout all stages of criminal proceedings.”9 Here are a few examples of the more ample construal of the PI: Identification Procedures It is common, when police are investigating a crime with eyewitnesses, to stage a line-up in which the latter are asked to pick out the perpetrator from a group of five or more people who generally fit the witnesses’ description. In the last few years, several states in the US have changed their procedures from this traditional design, ostensibly out of deference to the PI.10 Specifically, in those jurisdictions, the eyewitness must now be told before the line-up begins that the perpetrator may not be among those selected for the line-up. (Earlier practice had the police saying nothing to witnesses about whether they did or did not believe that the perpetrator was among those in the line-up.) Critics of the traditional procedure hold that it is “biased” and that such bias violates both the spirit and the letter of the PI.11 Most jurisdictions in the US and abroad do not (yet) believe that the PI entails such a disclaimer. Pretrial Publicity In many common law jurisdictions, the arrest of a suspect for a serious crime is likely to be accompanied by a release of his mug shot to the press and often a so-called “perp walk” where the newly accused is paraded before the cameras, frequently in handcuffs or other forms of constraint. Such public and humiliating displays of a suspected offender are completely off limits in most civil law jurisdictions.12 In Switzerland and Germany, the press is even instructed not to publish the names of criminal defendants until and unless they have been convicted.13 Italy and France forbid the publication of photographs of defendants in handcuffs. Once again, the PI is the engine explicitly driving such decisions. In those countries that prohibit such publicity, the argument is that the presumption of innocence applies not only to the legal system itself but also to the media, even if enforcing compliance with this policy in effect muzzles the freedom of the press. In the US, the right to freedom of the press is usually thought to trump any privacy or presumption of innocence rights that the defendant may possess.

Evolution of (Criminal) Rights  87 2)  Trial Procedures If many common law countries tend to exhibit a rather emaciated conception of the role of the PI prior to trial (in comparison with civil ones), the tables are turned for the trial itself. Among other things, the former issue a stern instruction to jurors at the outset of trial that they must assume the defendant is innocent until and unless the prosecution has established his guilt beyond a reasonable doubt. Failure to give this instruction can be grounds for appellate reversal of a conviction. In most civil law countries, even those that involve a lay or mixed jury, this direction figures much less prominently. Adverse Inferences From Silence For centuries, US courts—like those in most parts of the world—permitted jurors or judges to draw adverse inferences from defendant’s silence, where they seemed appropriate. In the 1970s, however, the Supreme Court ruled that the defendant’s exercise of his right to silence is grounded on the PI.14 Roughly, the argument is this: the presumption of innocence makes it clear that the defendant need prove nothing to win an acquittal. The burden of proof falls entirely on the prosecution. Accordingly, the defendant’s decision to maintain his silence cannot be taken as an indicator of his guilt. The presumption of innocence, it is said, requires that the jury ignore defendant’s failure to produce certain kinds of evidence (including his own testimony) since the PI makes it clear that defendant has no probatory obligations. In virtually all civil law countries (and recently in England and Wales), after a decades-long but now abandoned experiment with precluding inferences from silence, however, adverse inferences from defendant’s silence— whether during police interrogation or during the trial itself—are permitted, even encouraged.15 In such jurisdictions, there is no perception of a linkage between the PI and the defendant’s right to be shielded from adverse inferences from his silence. Courtroom Ecology and Defendant Demeanor Of late, and particularly in the US, the PI has come into play in unexpected ways. Appellate courts there have held that the presumption of innocence requires that during a criminal trial the layout of the courtroom itself and the apparel of the defendant must do nothing to suggest that he is guilty. Hence, he wears street clothes rather than, as was the earlier practice, prison garb. He must not be visibly shackled nor should he be obliged to occupy a “dock” or isolating enclosure of any kind. In 2005, the Supreme Court insisted that “[v]isible shackling undermines the presumption of innocence and the related fairness of the fact-finding process.”16 (In 1983, the US Court of Appeals for the First Circuit claimed that the presence of a dock

88  Larry Laudan was intrinsically prejudicial to the rights of the accused person. The court argued that a dock was a “brand of incarceration” and insisted that such arrangements were “inconsistent with the presumption of innocence.”17 Such niceties on the part of US courts are in marked contrast to most jurisdictions where both the defendant’s apparel and his location in the courtroom can identify him as a marked man. Tait gives a succinct summary of the situation: In the United States, the accused generally appears free and unfettered in a privileged position at the bar table. In Ireland and the Australian Capital Territory, the accused sits just behind the bar table, also unconstrained. Meanwhile, in France, Germany, England and Wales, and most parts of Australia, defendants sit in a designated enclosure, and an increasing number of courts are being built with glass-enclosed docks. In Spain and Italy (as well as numerous jurisdictions that do not have juries), the accused may be located behind metal bars in the courtroom.18 Once again, we see vivid evidence that a practice that one jurisdiction deems a gross violation of the PI another jurisdiction finds wholly compatible with it. Trial Publicity We have already seen that many jurisdictions discourage (and some even criminalize) pre-trial publicity about a defendant’s arrest on the grounds that it violates his right to a presumption of innocence, while other systems permit, even encourage, such publicity. The reporting of the events during the trial itself is likewise bound up in complicated ways with various versions of the PI. In general, trials are public events, not held in secret star chambers. That notwithstanding, many countries (both civil and common law) impose tight restrictions on what can be said publicly about a trial and the defendant while it is in progress. Such restrictions are almost always bound up closely with that legal system’s understanding of the content of the PI. Other countries happily allow press coverage and even the public broadcast of a trial, seeing no tension between such practices and the PI. The Burden of Proof and Affirmative Defenses It might be thought, despite all the mentioned divergences in the conceptions of the PI that are to be found in different jurisdictions, that there is still a shared, minimal hard core version of the PI. Specifically, one might suggest that virtually all legal jurisdictions share the view that the burden of proof in a criminal trial falls squarely on the prosecution and that the defendant is not required to prove anything. The fact is that this hard-core commitment is respected in almost no jurisdiction (except in US federal courts and approximately half the state courts in the US).

Evolution of (Criminal) Rights  89 This is due to the existence of “affirmative” or “excuse” defenses, and the bizarre manner in which they are handled in most civil jurisdictions as well as in the half of the American states. In the latter, if a defendant claims to have acted in self-defense or under provocation or while insane (or in a wide variety of other innocence-establishing situations), then it usually falls to him to prove—generally to the preponderance of the evidence but sometimes to an even tougher standard—that he is innocent of the crime.19 In sum, when an affirmative defense is in play, the probatory burden generally shifts unequivocally to the defendant, whose innocence the trial is supposedly presuming. In such circumstances, the defendant faces a de facto presumption of guilt, which he must defeat if he is to win an acquittal. Through the ages, proponents of the PI have concocted various fairy tales to reconcile the existence of such burden-shifting practices—which are patent violations of even the most minimal conception of the PI—with the PI itself. They usually involve falling back on such obiter dicta as “he who affirms must prove,” or “the defendant is best placed to prove his state of mind when the prima facie crime was committed” or the patently false claim that “it is impossible to prove a negative.” All these rationales are, of course, fallacies and wholly inconsistent with the PI’s core thesis that the burden of proof is the prosecutor’s alone. (While the latter thesis is not sufficient to capture the core sense of the PI, it surely ought to be a necessary condition.) One final observation is in order before moving on to a more conceptual analysis of the PI. In virtually every known explicit version of the PI, that presumption is defeated by a conviction and ceases to apply to the defendant in question. Recently, however, there have been moves afoot to extend the privileges of the PI even beyond the point of its defeat. For instance, in 2005, the US Supreme Court found that the PI applied to sentencing hearings held after the defendant was convicted. Specifically, in Deck v. Missouri (where the defendant was convicted of multiple capital crimes), the Court claimed that it violated the convicted prisoner’s right to a presumption of innocence to be shackled or dressed in prison garb at the sentencing hearing.20 It appears that there are no limits to how far the PI can be stretched by judges keen to show their robust endorsement of the doctrine. In certain jurisdictions, it is policy that the PI demands that a convicted felon who has appealed his conviction (and whose entitlement to the protections of the PI have been technically defeated by the act of his conviction) is nonetheless eligible for bail during appellate review. In such circumstances, the PI comes close to providing a cradle-to-grave protection for felons, even convicted ones. B.  Some Implications of the Ambiguity Question Clearly, there are nearly as many conceptions of “the” presumption of innocence as there are legal jurisdictions. More to the point, each jurisdiction grounds its version of that presumption on arguments that other

90  Larry Laudan jurisdictions either ignore or, oftentimes, explicitly repudiate. Thus, US law is unambiguous that the PI does not entail automatic release on bail while many other countries insist that the PI entails pre-trial freedom unless very specific PI-defeaters apply in the case in question (for instance, if the applicant for bail has a prior history of becoming a fugitive when granted bail). At the other extreme, US law insists that the PI guarantees a defendant’s right not only to silence but to an instruction to the trier of fact to draw no adverse inferences from that silence; many other jurisdictions (e.g., England and Wales as well as most civil law countries) hold that the drawing of plausible inferences from defendant’s silence is entirely consistent with a robust version of PI, as they understand it. Given such discrepancies, it is worth asking whether we are dealing here with what Gallie called an “essentially contested concept.”21 We encounter such a phenomenon (i) when there are several rival (and mutually incompatible) conceptions meant to capture a given concept, (ii) where proponents of those rivals are all persuaded that their own version captures the legitimate meaning of the concept, and (iii) it is wholly unclear that there is any definitive argument that would move the advocates of rival conceptions to converge on any one of them. It seems clear that the presumption of innocence comes into this category. (If more space were available it could similarly be shown that the right to silence, the right to a demanding standard of proof, and the right to exclusion of illegally seized evidence are open to drastically different interpretations in different jurisdictions.) If this diagnosis is correct, we have no reason to expect convergence let alone closure on the definitive meaning of various key concepts of the rights of defendants. Is reconciliation of the various versions of the PI nonetheless a viable option, at least in theory? One has to be dubious about that since (i) various constitutional courts have already considered such alternatives and repeatedly rejected them and (ii) the very different ways in which civil law courts and common law courts respectively conceive the trial process and how common law courts do creates moral and epistemic tensions that appear to preclude rational adjudication. Under such circumstances, it is not easy to conjure up an argument that would compel either common law or civil jurisprudence to absorb or assimilate the other’s conception of the reach and/or the limits of the PI. To that degree at least, the presumption of innocence appears to qualify as an essentially contested concept. If it is, then the much-heralded international convergence of the rights of criminal defendants seems more a verbal mirage than a present reality or even a plausible eventuality.

Part II: A More Drastic Dissection of the Presumption of Innocence Thus far, I have been nibbling gently around the edges of the problems confronting the presumption of innocence. The aim of the present section is

Evolution of (Criminal) Rights  91 to move directly to a critique of the hard core of the PI. Having suggested that there is no universally or even broadly shared sense of the scope and limits of the PI and that multiple versions of it abound (none of which is obviously superior to its rivals), I now want to move the critique one stage further. I will argue that the presumption of innocence is not only essentially contested but that, even worse, it is intrinsically incoherent. We begin with the most familiar rendering in English of the core meaning of the PI: to wit, (1) A person is innocent until proven guilty. “Innocent” here might mean that the person did not commit the crime. This, indeed, is the ordinary meaning of “innocent” and almost certainly how lay jurors—instructed to assume the defendant innocent until proof of his guilt—would interpret it. But that cannot be what the PI means to say since, if a person did not commit the crime, then he is materially or factually innocent, whatever the verdict at trial. Likewise, if the person really committed the crime, then he is (in the lay sense of the term) guilty, whatever verdict ensues and the PI, on this gloss, asserts as fact what is a fiction if defendant really committed the crime. So, in the factual sense of innocence, (1) is false on its face. If John is innocent, nothing whatever that happens at trial can alter that brute fact. If, on the other hand, John committed the crime, then it is patently false to say that he is innocent until and unless his guilt is proven. There is a quite different gloss on (1), if one construes “innocent” not in the material or factual sense but in the probative sense. If we take “innocent” in as something like “guilt not proven,” then (1) ceases to exhibit the problems just described. But whereas the first reading of (1) was false in suggesting that a materially innocent person could give up that status if convicted, this second way of reading (1) turns it into an empty tautology, for it now is tantamount to the assertion “a person’s guilt is not proven until he is proven guilty.” We have thus replaced falsehood by vacuity. We obviously need to look further afield to come up with a plausible version of the PI that is neither false nor redundant. What then is (1) trying to say? One obvious gloss on (1) is that “innocent” here refers neither to a putative fact about the world nor to a probative sense of “innocence” but rather to the appropriate doxastic state of the triers of fact (and perhaps of other actors in the legal proceedings as well). On this interpretation, the PI is an injunction to the jury to believe firmly that defendant is innocent until and unless the evidence eventually persuades them to the prevailing standard that he is guilty. We could express this succinctly as (2) You should believe that defendant did not commit the crime until the standard of proof has been met.

92  Larry Laudan With this shift from facts or probative states to jurors’ beliefs, (2) presents a host of new problems. The principal one is that rational people do not, indeed cannot, adopt a certain belief simply because someone tells them they should believe it. A juror at the beginning of trial (when she typically receives the instruction about the PI) cannot simply will herself to adopt a belief about a matter of fact (defendant’s innocence or guilt) for which she has no evidence whatever. The idea that, being told by the judge that she must believe defendant did not commit the crime until it is convincingly proven otherwise, is a sufficient ground to adopt that belief makes a nonsense of human belief practices. Absent any evidence, whether exculpatory or inculpatory, a rational juror will have no basis either for believing defendant committed the crime or that he did not.22 Agnosticism about guilt and innocence is the only reasonable doxastic posture open to a juror who has yet to hear or see any evidence. By wholly ignoring this fact, (2) enjoins jurors to do something that most humans cannot do, to wit, to believe a contingent empirical claim absent any evidence for it whatever. That notwithstanding, typical jury instructions in the US enjoin jurors to believe that defendant is innocent until it has been proven otherwise. Still, (2) points us in a rather more promising direction than (1) did. What we really want jurors going into a trial to understand is that the trial is fundamentally a probative activity. The prosecutor must either prove convincingly that defendant committed the crime or the jury must return a not-guilty verdict. Further, we want them to attribute no inculpatory probative weight to the fact that defendant has been charged with a crime. One natural way to get those points across would be to characterize the PI along these lines: (3) A conviction requires that defendant’s guilt be proved by the prosecutor beyond a reasonable doubt; any level of proof less than that requires a finding of not-guilty. Defendant does not have to prove anything. This formulation of the PI has the virtues of being neither false nor tautologous (as versions of (1) are), and of not asking jurors to believe an empirical proposition for which they have no evidence (as (2) does). It is, however, targeted specifically at the trier(s) of fact. Since the PI is widely supposed to have a compass that extends well beyond jury deliberations, we have to say that (3) may be an adequate instruction for the jury but that it is woefully incomplete otherwise. So, many would argue, (3) needs to be supplemented by a version of the PI that goes well beyond a mere jury instruction, couched purely in probative terms. One possible formulation might be along these lines: (4) From the moment of arrest, defendant is to be treated as if he were innocent, enjoying the rights autonomy of ordinary citizens not suspected of having committed a crime.

Evolution of (Criminal) Rights  93 This appears to embody the idea that apparently drives much thinking, especially in civil law countries, about how defendants are to be treated by the authorities. It explains, for instance, why those jurisdictions are very reluctant to deny bail and why they clamp down firmly on publicity about the defendant, both pre-trial and during the trial. Even in common law jurisdictions, (4) seems to undergird reasoning about various matters (among them, shackling defendants, eliminating the dock and prohibiting inferences from silence). The trouble with (4) is that, as formulated, it is much too ambitious. If taken seriously, every system of criminal justice in the world would have to change its ways drastically. Consider just a few examples of what can legally befall the person who is suspected of a crime: he can be arrested, photographed, searched, forced to participate in a line-up, subjected to hostile interrogation, see his home invaded and searched by police with a warrant. He can have his personal property seized, and be required to submit samples of his prints, his voice, his DNA and his handwriting. His phone can be tapped and conversations with associates recorded. He can, in many legal systems, be jailed pending trial. To put it mildly, this is not how we expect the legal system to treat those persons it genuinely believes to be innocent. Precisely because all these acts involve restricting the rights, the liberty, the property or the privacy of a suspect, most legal systems put in place checks on these intrusions. Specifically, the authorities have to be able to show probable cause to believe that defendant is guilty before any of these acts can be legal. The point is that we treat those suspected of having committed a crime very differently from the way in which we treat nonsuspects. Two centuries ago, Jeremy Bentham nicely summed up the conundrum posed by a broad and liberal interpretation of the PI: In all cases of penal procedure, the declared supposition is, that the party accused is innocent; and for this supposition, mighty is the laud bestowed upon one another by judges and law-writers. This supposition is at once contrary to fact, and belied by their own practice. [. . .] The defendant is not in fact treated as if he were innocent, and it would be absurd to deal by him as if he were. The state he is in is a dubious one, betwixt non-delinquency and delinquency: supposing him nondelinquent, the[n] immediately should that procedure against him drop; everything that follows is oppression and injustice.23 That, in turn, means that (4) cannot be right. In general, virtually all broad versions of the PI scandalously ignore the fact that once one acquires the status of a warranted suspect, many rights and liberties that one formerly enjoyed vanish. While it is true that an unconvicted suspect cannot be punished in the technical sense, his status as a suspect puts him on a very different legal footing from that enjoyed by those who are not suspects.

94  Larry Laudan At the risk of belaboring the obvious, these are not the ways in which a civilized society handles citizens whom the authorities genuinely believe to be innocent. On the contrary, this is how a liberal society treats those who are thought likely to be criminals. Are we really supposed to believe that a judge who authorizes an arrest for probable cause still believes the defendant is innocent? Or that a prosecutor who brings a defendant to trial presumes his innocence? If that were true, then he should not be bringing the person to trial at all. Do the police who enter the house of a suspect with a court order, seeking inculpatory evidence against the suspect, believe him to be innocent or at least treat him as if he were? Of course not. They seek a search warrant because they have reason to believe the suspect is guilty and seek the warrant in order to gain access to evidence that may demonstrate that guilt. The obvious fact of the matter is that a properly conducted police inquiry requires the identification of suspects and the accumulation of evidence against them. That would be next to impossible if police, prosecutors and judges were obliged to believe or act as if they generally believed that those whom they investigated or charged with a crime were likely to be innocent.

Conclusion While it seems not so difficult to fashion a coherent version of the PI—along the lines of (3) above—that can serve as a guide for the trier(s) of fact, we are on much less solid ground when it comes to figuring out how to formulate a viable statement of the PI that applies to everyone else. Judges, prosecutors and the police—the PI notwithstanding—need not and do assume a suspect is innocent until he is convicted at trial. Arguably, neither need the media nor the general public. The authorities, for their part, are routinely and statutorily involved in making determinations of probable cause. Most of them lead to the conclusion that the suspect in question is probably guilty and should be treated accordingly in the sense of surrendering certain of his liberties in the interests of the investigation of a likely crime. The shadow cast by the presumption of innocence is scarcely to be seen in this phase of the operation of the criminal justice system. As for the implications of the PI for the press and the general public, different countries have very different sensibilities that appear to elude any one-size-fits-all remedy. In short, it is deeply misleading to speak of “the” presumption of innocence. There is no such thing. In its stead is a confusing diversity of presumptions of innocence, no one of which is either compelling or free from incoherence.

Notes * It is a pleasure to be invited to contribute a short essay to this Festschrift for my good friend Mark Platts. The challenge has been to find common ground between his professional interests and my own. While Mark has made his mark as a moral theorist (which I patently am not), I decided to take my cue from his impressive facility for bringing philosophical tools to bear on practical moral questions about public policy. Being simultaneously an exercise in political

Evolution of (Criminal) Rights  95 morality and in applied epistemology, the criminal law seems to be fertile terrain for such explorations. My aim here is to focus on one key component of the criminal law—the much-heralded presumption of innocence—and to show that, whether viewed epistemically or morally, it is a welter of confusions and contradictions. 1 Francois Quintard-Morenas, “The Presumption of Innocence in the French and Anglo-American Legal Traditions,” The American Journal of Comparative Law, vol. 58, no. 1 (2010): 107. 2 The first officially approved proposal that the PI should become an internationally recognized right is found in Article 11 of the Universal Declaration of Human Rights, approved by the United Nations General Assembly in 1948. 3 See especially John Jackson and Sarah Summers, The Internationalisation of Criminal Evidence: Beyond the Common Law and Civil Law Traditions (Cambridge: Cambridge University Press, 2012) and Mariolina Eliantonio, “The Future of National Procedural Law in Europe: Harmonisation vs. Judge-made Standards,” Electronic Journal of Comparative Law, vol. 13, no. 3 (2009): 1ff; and Diane Amann, “Harmonic Convergence? Constitutional Criminal Procedure in an International Context,” Indiana Law Journal, vol. 75, no. 3 (2000): 809–73. See also Craig M. Bradley, “The Convergence of the Continental and the Common Law Model of Criminal Procedure,” Criminal Law Forum, vol. 7, no. 2 (1996): 471. 4 Ingraham writes: “There is virtually no nation today that does not claim to abide by the “presumption of innocence” in the administration of its criminal justice system.” Barton L. Ingraham, “The Right of Silence, the Presumption of Innocence, the Burden of Proof, and a Modest Proposal: A Reply to O’Reilly,” The Journal of Criminal Law & Criminology, vol. 86, no. 2 (1996): 564. 5 According to LexisNexis, the US Supreme Court, in the half century beginning in 1960, discussed the presumption of innocence in no less than 107 distinct cases. 6 Caroline Davidson gives a succinct summary of the current situation: “Presumption of innocence notwithstanding, some domestic courts appear to allow the strength of a case against an accused to be factored into the provisional release decision. In the United States federal system, in certain serious cases where there is a presumption of risk of flight or danger, “the weight of the evidence” is one of the factors a judge is to take into account in deciding release. Similarly, in Canada, the strength of the prosecution’s case can be relevant to the release inquiry. By contrast . . . in the United Kingdom, the likelihood of conviction has been deemed irrelevant to the bail inquiry since 1976” (“No Shortcuts on Human Rights: Bail and the International Criminal Trial,” American University Law Review, vol. 60, no. 1 (2010): 1–18). 7 Even in Canada, there are prominent voices insisting (as one Supreme Court justice puts it) that “when bail is denied to an individual who is merely accused of a criminal offence, the presumption of innocence is necessarily infringed” R. v. Hall, [2002] 3 S.C.R. 309, 2002 SCC 64, P 48 (Can.) (J. Iacobucci, dissenting). 8 In 1979, the Court insisted: “The principle that there is a presumption of innocence in favor of the accused is the undoubted law, axiomatic and elementary, and its enforcement lies at the foundation of the administration of our criminal law” Coffin v. United States, 156 U.S. 432, 453 (1895). But it has no application to a determination of the rights of a pretrial detainee during confinement before his trial has even begun” Bell v. Wolfish, 441 U.S. 520, 533 (U.S. 1979). 9 See Jackson and Summers, The Internationalisation of Criminal Evidence, 205. 10 For a fascinating discussion of the basis for these new identification procedures, see Steven E. Clark, “Blackstone and the Balance of Eyewitness Identification Evidence,” Albany Law Review, vol. 76, no. 2 (2010–11). 11 For a critique of the notion of “bias” used here, see Clark, “Blackstone and the Balance of Eyewitness Identification Evidence.”

96  Larry Laudan 12 As early as 1862, Austria prohibited the print media from publishing “statements regarding the value of evidence [in a trial] and the assumed outcome of the proceedings” (quoted by Giorgio Resta, “Trying Cases in the Media: A Comparative Overview,” Law and Contemporary Problems, vol. 74, no. 4 (2008): 42–3. 13 In Austria, the press cannot even mention the names of crime victims if such information might be “prejudicial.” 14 For a much fuller account of how the US Supreme Court has tried to forge a link between the “right to silence” and the PI, see Larry Laudan and Erik Lillquist, “The Sounds of Silence,” University of Texas Law, Public Law Research Paper no. 215 (2012), https://ssrn.com/abstract=2037575 or http://dx.doi.org/10.2139/ ssrn.2037575. 15 To be clear, civil law courts recognize defendant’s right to silence, but they do not view that right as precluding adverse inferences from refusal to testify. 16 Deck v. Missouri, 544 U.S. 622, 630 (U.S. 2005). 17 The Court went on to say that “Because confinement in the prisoner dock is unnecessary to accomplish any important state interest, and may well dilute the presumption of innocence, the Massachusetts prisoner dock must be considered, as a general matter, to be an unconstitutional practice” Young v. Callahan, 700 F.2d 32, 34 (1st Cir. Mass. 1983) 18 David Tate, “Symposium on Comparative Jury Systems: Glass Cages in the Dock? Presenting the Defendant to the Jury,” The Chicago Kent Law Review, vol. 86, no. 2 (2011): 468. 19 Lest one think that these are rare cases, it is worth pointing out that in the US, fully a quarter of the cases of violent crime that go to trial involve an affirmative defense. 20 A fuller discussion of this vexing question can be found in Tara J. Mondelli, “Deck v. Missouri: Assessing the Shackling of Defendants During the Penalty Phase of Trials,” Widener Law Journal, vol. 15, no. 3 (2006): 785–813. 21 W. B. Gallie, “Essentially Contested Concepts,” Proceedings of the Aristotelian Society, vol. 56 (1956): 167–98. 22 Indeed, any savvy juror will know from the beginning of the trial that various parties—the police, the prosecutor, a grand jury and an arraigning judge—have already concluded that the defendant committed the crime. Under such circumstances, it must be virtually impossible to believe firmly that the defendant is innocent. 23 Jeremy Bentham, “Principles of Judicial Procedure (1829),” in The Works of Jeremy Bentham, vol. II, ed. John Bowring (Edinburgh-London: William Tait/ Simpkin Marshall, 1843), 169.

8 Wittgenstein on Rule Following Some Themes and Some Reactions Paul F. Snowdon

Mark Platts and I were students reading PPE at University College, Oxford, in the late 1960s. For some time after that we were teachers and colleagues in England, before Mark moved to Mexico. Mark was the most entertaining of friends, and his work then and since has struck me as ahead of its time, original and fascinating. I intend this paper to be an expression of my admiration for him. I have chosen not to discuss his own work here, but, instead, to explore some Wittgensteinian issues which have links with one of the major themes of Mark’s work, that of the nature of meaning. The Wittgensteinian ideas that I shall explore come from the sections currently called the Rule-Following sections in the Philosophical Investigations, that is (approximately) sections 138 to 242. My hope is that these somewhat tentative attempts to engage with Wittgenstein will at least help others who are engaged in trying to make sense of Wittgenstein’s thought.

1. Wittgenstein in the 1960s The dominant attitude to Wittgenstein at the time we were students, as I see it, was that the later Wittgenstein had refuted a long-standing approach to meaning, the approach which equated understanding language with the presence of suitably caused images in those who understand, and that he had launched an attack on a long-standing conception of experience using the Private Language Argument. But no one was sure quite what this attack was, nor whether it was sound. A vast amount of attention was devoted in the late 1960s and early 1970s to an inconclusive debate about this issue. There was also considerable scepticism about other central aspects of Wittgenstein’s approach. For example his employment of the notion of criteria had been vigorously attacked in a series of papers by Putnam and others. This sceptical attitude was also to some extent a consequence of the fact that the philosopher taken to most embody the Wittgensteinian approach at that time was Norman Malcolm, and it was not hard to locate in his work ideas that struck people as dubious and extreme.1 Another source at that time of a move away from Wittgenstein was the emergence in the 1960s of a general conviction that psychological concepts are causal concepts, inspired

98  Paul F. Snowdon by philosophers such as Grice, Davidson and C. B. Martin. This conviction was generated initially by a debate about the notions of action and action explanation, in which the causalists were regarded as having triumphed.2 Such causalism was plainly alien to Wittgenstein’s system. This situation was radically altered in the early 1980s by the publication of Kripke’s book Wittgenstein on Rules and Private Languages.3 It caused the rule following sections to be viewed as the most important part of the Philosophical Investigations, and it regenerated an interest in Wittgenstein’s thought. It has occasioned an intense debate about rule following and Wittgenstein, which has continued ever since. But this debate has to a large extent moved away from Wittgenstein’s original discussion and developed a life of its own, involving, as it has, many philosophers of considerable talent. I want myself in this paper to return to Wittgenstein’s own discussion and to consider his arguments and views, rather than to engage with that more recent debate. I do so, not because I think that this more recent debate is uninteresting or misconceived, but because I believe that there is still much to learn from Wittgenstein’s original text. I have selected a number of aspects of Wittgenstein’s text, most from the rule-following passage, but some from other parts of the Investigations, and tried to interpret them and determine what we can learn from them. In the rule-following passage the bits of text that I have selected come between 138 and 210, and do not include the later passages in which Wittgenstein stresses the idea that rule following is social or a matter of “custom.” I am not, then, offering an interpretation of these passages as a whole, but simply singling out some themes and reacting to them. I remain unsure what Wittgenstein’s central proposal is. I start, though, with the background, that is to say, with how we should conceive of Wittgenstein’s general approach to philosophy, which is, I think, of some importance in relation to these passages.4

2. The Background When people propose interpretations of the rule-following arguments, their proposals are shaped by the general reading of Wittgenstein they accept. Now, this is, clearly, a legitimate attitude. So I shall start the development of my own partial reading by asking: how should we understand Wittgenstein’s general approach to philosophy? I want to propose a way of reading him that seems plausible to me (and which is by no means novel).5 This way summarises the approach that I have found most helpful whenever it seems to me I have understood what Wittgenstein is saying, or when I get, as I feel, close to understanding it. The central idea is that Wittgenstein is first and foremost a negative philosopher. This means that his primary aim is to establish conclusions of the form Not (P), or, perhaps, do not believe that P. Usually he is saying something like; “Here is a conception that we should reject.” We might put it like this;

Wittgenstein on Rule Following  99 Wittgenstein is often saying; “Don’t get worried about this sort of thing, but above all do not think this (viz. P) about it.” It may be that the term “negative” has itself a rather negative tone, and another, perhaps more positive, way of expressing the same point is to say that Wittgenstein is engaged in demythologising. He is uprooting myths. Now, frequently the myths he is opposing are associated, by him, with great thinkers. There is, for example, the Augustinian myth about language. On other occasions he attacks Russell, Freud, G. E. Moore (in On Certainty), Frege, Fraser, Kohler, and, of course, himself. Often too, Wittgenstein couples rejection of the myths with endorsement of a similar conception, or a conception that can be expressed in a similar way. His idea is that what the great figure thought was wrong, but in, or behind, their erroneous conception there is something correct. Now, whether this characterisation should be accepted can only be determined in the course of a general close reading of the text. I do not expect a reader of this paper simply to accept this interpretation on the basis of my say-so, but I believe that it is helpful to keep it in mind when you engage with the text. There are two main reasons to view him this way. The first is that it is the obvious way to read him given the manifest content of his discussions. Here are two examples. In sec. 1 of the Investigations Wittgenstein gives a quote from Augustine expressing what has been called “the Augustinian picture of language.” In the next sections, up to about 46, Wittgenstein is picking away at that conception, suggesting that we should not accept it. The dominant impression is one of rejecting some important ideas. Again, in 243, Wittgenstein introduces the famous question about whether someone can introduce their own private terms for their sensations. In the next sections (to about 315) Wittgenstein is clearly attacking a conception of private experience according to which the answer to this question would be “yes.” He aims to uproot that conception by arguing (amongst other things) that actually the correct answer is “no.” The discussion is primarily negative, to such an extent that most readers would be hard pressed by the end of it to say what positive conception of sensations and our way of talking about them Wittgenstein favours. The second reason for regarding Wittgenstein as a negative philosopher is that he advances a conception of (good) philosophy which seems to represent it as highly negative. Consider his remark “The philosopher’s treatment of a question is like the treatment of an illness.”6 This is one of his most famous metaphors. What does it mean? The obvious suggestion is that it represents the goal of proper philosophy as eliminating something bad, an “illness,” health here not representing having a better theory, but rather solely being clear of the confusion, that is of the bad health. The strong suggestion is that returning to health is not improving one’s theory of the world, nor is it making a (true) addition to one’s cognitive understanding, but is rather eliminating the activity which amounts to being ill. Consider

100  Paul F. Snowdon also 309: “What is your aim in philosophy?—To show the fly the way out of the fly bottle.” This compares bad philosophy to being trapped like a fly and the aim is to release the fly from the trap. Clearly, the idea of a trap here is that of accepting a bad view, and the aim is to release one from that view. Here we again have the idea that good philosophy acts to free you from traps, simply releasing you. (We do, of course, in considering these remarks of Wittgenstein face one major difficulty posed by his work, that of determining the significance of his omnipresent metaphors and similes.) However, these remarks about Wittgenstein’s view of philosophy need some modification if they are to be true to his actual practice. First, he does offer more positive accounts of what he is doing. Thus, in 415 he describes himself as supplying remarks on the natural history of human beings. It is hard to know what he means by that, but the remarks he is offering are presumably positive, and to some extent theoretical. Wittgenstein claims they would be disputed by no one. However, the fact that a claim would not be disputed when advanced does not mean that it was already appreciated, so its proposal and acceptance amounts to a cognitive advance. One might also say, with some confidence, that no claim, at least of the kind that Wittgenstein is advancing, are such that no one will dispute them. Second, he clearly offers general hypotheses for consideration. Thus, for example, he links meaning to use, he makes conjectures how sensation words gain their meaning, he stresses the role of context in judgement and so on. Another example of what is clearly positive theorising is his idea that language is what makes us do bad philosophy. So Wittgenstein invites us to consider and presumably accept some manifestly theoretical proposals. Further, although he compares good philosophy to curing a condition, it in fact works by persuading you of some unobvious truth, e.g., that something is a consequence of a claim, or that there are difficulties with a claim, which amounts to a type of proposition being added to one’s theory. There is clearly a level in Wittgenstein’s practice where unobvious and so (in one sense) theoretical claims are being proposed. I am suggesting, then, that there are at least two levels in Wittgenstein’s actual practice of philosophy where there are propositions and claims advanced for assessment and possibly belief. He advances what are clearly general theories about certain things. Second in trying to achieve his negative purpose he advances arguments containing (unobvious and new) claims that are also offered as true. What is the relevance of these descriptions of Wittgenstein’s general approach to the rule-following passage? Amongst philosophers who accept that Wittgenstein’s later philosophy is basically negative there is some tendency to accept an argument which might be formulated as follows: (i) Wittgenstein does not mean to alter or criticise any of our ordinary thinking, and (ii) if Wittgenstein had accepted claim P then he would be altering ordinary thinking; therefore Wittgenstein cannot be proposing that P. This type of argument is used to support what is

Wittgenstein on Rule Following  101 sometimes called a quietist or deflationary reading of the rule-following discussion.7 Now, there are, as I see it, at least three reasons for not placing too much reliance on this style of argument. The first, already brought out, is that despite his own characterisation of what he is doing as non-theoretical, leaving everything as it, merely a treatments of an illness, etc., it has to be acknowledged that in practice Wittgenstein proposes unobvious new claims (theories of greater or lesser generality) for our consideration. He is simply not totally “quiet.” Second, if Wittgenstein characterises himself as leaving everything as it was, what that amounts to entirely depends on Wittgenstein’s own view of how everything was. When I declare on looking over my room that nothing has been changed that clearly depends on my sense of how it was before. So, Wittgenstein might think that what he is saying amounts to changing nothing, but it is quite possible that he has completely failed to realise how “things” were. Wittgenstein’s assurance that he is changing nothing is most certainly not self-certifying. Third, when Wittgenstein says that philosophy leaves everything as it was, he has to exclude philosophy itself. The practice of (bad) philosophy and the convictions generated by it are to be excised. The question then arises as to what Wittgenstein thinks the extent of old bad philosophy is, since he certainly does not aim to leave it as it has been. Indeed, once raised this represents a very hard question for Wittgenstein, since how does he know how widespread bad philosophy (on his conception) has been? It remains quite possible for him to think that it has been widespread, and so, maybe, many relatively common or even quite usual opinions are the products of bad philosophy. The fact that the extent of so-called “philosophy” has no really discernible boundaries means we have no idea what Wittgenstein thinks it is legitimate to challenge and what it is not legitimate to challenge.8 Although I have tried to place some brakes on the line of thought just considered, I do think it is right to say that the nature of Wittgenstein’s approach in the rule-following sections is certainly influenced by his negative approach to philosophy. It comes out in the following way. Wittgenstein’s main endeavour is to criticise and eliminate misunderstandings of understanding. Positively, what remains is primarily the assertion that there is understanding. There is no very systematic attempt on Wittgenstein’s part to say to what understanding actually amounts, an absence that surely reflects the basically negative conception of philosophy that he held.

3. The Initial Problem: Use and Understanding Wittgenstein starts with a question in 138; “But can’t the meaning of a word that I understand fit the sense of a sentence that I understand?” He then adds “Of course, if the meaning is the use we make of the word, it makes no sense to speak of such ‘fitting’.” So the question that Wittgenstein is posing starts from the idea of meaning as use; that idea frames the discussion (at this stage, and indeed throughout). But what does this initial question or

102  Paul F. Snowdon problem mean? What is meant by “fitting”? What is supposed to be fitting what? What might lead someone to talk of fitting in this way? Now, my own inclination here is to acknowledge that talk of fitting is rather mysterious and unclear, but not to be too worried about that. We can let the question function simply as a lead in to an argument that develops and subsequently becomes clearer.9 The problem is supposed to be that since meaning is use a formula cannot fit the meaning. It is, surely, somewhat hard to understand this problem. We can perfectly well say that a definition does (or does not) fit how someone (seems to) use a term. So there is no obvious worry here. (Here is an analogous way of talking. If I say—he does not cut his lawn with that mower—and he does use that mower then my statement does not fit his use of that mower. Fitting a use is something that it seems a statement can do.) But in 139 the question becomes this: when I hear a word I understand it, but if meaning is use how can that be? Thus Wittgenstein says; When someone says that word ‘cube’ to me, for example, I know what it means. But can the whole use of the word come before my mind, when I understand it in this way?. . . . Can what we grasp in a flash accord with a use, fit or fail to fit it? And how can what is present to us in an instant, what comes before our mind in an instant, fit a use? Now, this seems to be a question about what Grice calls occasion understanding, that is, the understanding that someone achieves on an occasion when hearing an expression. They there and then understand it. Wittgenstein seems to have the same worry as before; how can use be present in an instance? On hearing this question one is surely inclined to say: the use itself cannot be present, since the use is a historically extended group of occurrences, but what the use is can be given in an instance. So, in the absence of something more, there seems to be no problem. The idea of meaning as use has not raised difficulties. A way of taking what is going on here is this; in the early sections of the Investigations Wittgenstein has presented an account of meaning as amounting to use. This might be taken as saying when there is meaning. There is meaning when there is the right sort of use. But the other side of meaning is that individual audiences grasp the meaning of what is said on an occasion. So there is simply the question: can the general conception of meaning as use yield an acceptable notion of understanding on an occasion? Relative to this interpretation the so called rule following sections are an investigation of occasion understanding within the framework of a general account of meaning as use. Now, I have suggested that it is not totally obvious why there is a problem here flowing from the idea of meaning as use. But it also is an option to oppose the existence of the whole problem by querying whether meaning is use. If it is not then one assumption of the passage is dropped. However, it

Wittgenstein on Rule Following  103 needs stressing that even without that background assumption there would then remain the issue—what is understanding? Wittgenstein immediately considers the suggestion that a picture (or something like a picture) comes before the mind. As far as the Investigations is concerned this is, I think, really the first place where he engages with what is called the imagistic theory. (Elsewhere, of course, he had considered this idea earlier.) Now, Wittgenstein responds by pointing out that if a picture comes before the mind on hearing a word, that picture does not actually fix a use. He then considers the suggestion that another image or picture can fix the use that the first image yields. It is obvious that this second image does not fix a use (or what he calls a method of projection) of the image any more than the first image itself fixed a use. With these profound remarks and examples (which I shall not spell out here) Wittgenstein effectively makes us see that images are not the key to occasion understanding. What he also conveys is that in fact, given how we react to pictures, they can certainly suggest uses of things to us. Think, for example, of the way that pictures on, e.g., hand dryers, such as picture of hands in the right place, convey to us how to use them. We do learn from the pictures how to use the dryer. So the idea that understanding consists in the reception of pictures is not, from the point of view of meaning as use, totally absurd. Wittgenstein’s idea though is that pictures can convey uses to us, but that is a matter of how we react or respond to the pictures. The picture in itself does not fix a use. What needs stressing so far is that all that has really happened is that the model of understanding (in this case, occasion understanding) as grounded in pictures before the mind has received profound criticism. The criticism is based on the assumed idea that meaning is use.

4. Some General Themes I want next to list some of the sub-questions and themes that arise from, or in, Wittgenstein’s pursuit of this initial question. My hope is that reading the passage with this list in mind aids comprehension of it. The first theme is, quite simply, what do such terms as “understanding” or “knowing how to go on” or “grasping a rule” actually stand for? Or putting it less linguistically, what is it to understand or know how to go on or to grasp a rule? There is surely no doubt that the dialogue gets its momentum from attempts to answer this sort of constitutive question. Or perhaps, it is better to say that the dialogue gets its momentum by considering attempts to answer such constitutive questions. There is, however, an issue that arises from this list. It is clear that Wittgenstein focusses on some rather different cases. One type of case is understanding a word (or a sentence). But, second, he is also interested in learning what we might call a series, such as the alphabet, or the decimal numerals from 1 to 9. Note that these are in effect simply lists. There is nothing recursive about them.

104  Paul F. Snowdon The third case is the recursive case, learning the decimal system, or learning how to carry out an operation, for example, addition. Another case is that of realising there is a pattern in a series, which recurs. This is recursive pattern recognition. Here also belongs the idea of knowing a rule. Rules can be reapplied, and their application recurs. Now, these are rather different cases, and one might wonder why Wittgenstein groups them together. In particular, what has grasping rules to do with understanding words? How might the idea that understanding a word amounts to grasping a rule be developed? One approach might be that to understand a word is to grasp a rule for making judgements about sentences which contain it. Roughly, to understand the word “red” is to know a rule of the form; judge that it is true that something “is red” if condition C obtains. But there is a problem with this idea. The problem is that in order to employ the rule one must be able to make judgements which, at some stage, are not rule governed. Thus, to follow the proposed rule one must be able to judge that C obtains. The moral of this observation is that fundamentally judgement cannot be ruled governed.10 But if judgement cannot be rule governed fundamentally, then understanding words cannot, it seems, be represented in terms of rule mastery. A second central theme is this: to what extent do the psychological terms we are using in relation to this sort of condition pick out or stand for an episode in consciousness? Wittgenstein seems to allow that there are such things as events within consciousness, but what have such events to do with understanding or knowing how to go on (say)? Wittgenstein’s view is clearly that these terms do not stand for episodes in consciousness. One might indeed say that he is claiming that nothing in consciousness can count as the presence of understanding, nor of grasp of a rule. Indeed one could say that his view is that concepts cannot be present in consciousness. We cannot encounter or grasp a concept in consciousness. My own attitude, not to be properly argued for here, is that this is correct, and that we are still thinking out its implications. One possible and very serious consequence being that Frege’s postulated basic psychological act of grasping a concept or a sense is something for which no coherent model exists. A third theme is this: sometimes we can tell or know that we know how to go on. This is an example of self-knowledge, knowledge by the subject that the subject, him- or her-self, knows how to go on. The question is; what gives the subject this self-knowledge? Now, I want to suggest that these last two themes are linked. If the cognitive state was or consisted in an episode in consciousness then we could, presumably, know we had it by observing, or scanning, or directing our attention to, our consciousness. We could recognise what was there. But if that is not true then how do we know? It is clear, I think, that Wittgenstein was obsessed by this sort of problem. I am inclined to say that he is right to regard such self-knowledge as raising a major problem, but that he fails to offer a proper solution to the problem.

Wittgenstein on Rule Following  105 A fourth theme is that of when and how we as observers can tell or know that the state or condition is present in another. This is the question of other knowledge about such states. A fifth and final theme is evaluating the suggestion that these cognitive terms stand for mechanisms or brain states or dispositions—or something like that, something of that category. As I read Wittgenstein’s engagement with this idea or type of idea, he is opposed to it. A crucial question is whether he assembles enough to persuade us to turn against such a natural idea. Without any great confidence I am offering these five themes as (amongst) the major recurring aspects of Wittgenstein’s discussion. We really only understand Wittgenstein when we know how he proposes to answer these questions.

5. The Cognitive Role of Images Wittgenstein’s discussion of the imagistic model of understanding both here and in his other writings is destructive philosophical criticism of the highest order, and I have suggested that it is totally credible. There is, though, I want to suggest, a tendency to over-generalise this conclusion. There is some tendency, I believe, once Wittgenstein’s opposition to ideational theories of understanding is accepted, to think that images are of no cognitive interest, but this obviously does not follow from the falsity of ideational theories of linguistic understanding. The point is that if we accept that it is a mistake to think of understanding as having images of the things meant in one’s mind triggered by the words, that in no way amounts to a full assessment of the cognitive role of images. Moreover, the existence of images and a general capacity to imagine makes it likely that they have a significant cognitive utility for us. But what is that cognitive utility? One obvious cognitive role of images is as elements which aid problem solving. For example, if asked the route from X to Y, I can construct a description by reading off the sequence of turns by imagining doing the walk. In the course of that it will dawn on me that, say, the first turn is left. I regard the image sequence as a reliable guide to the journey. In fact, I think of the imagery as an output of a reliable memory mechanism. We can put it this way: we regard the imagining of the route as representing in a reliable way what the route is. And we cognitively respond to imagining the route by deriving a shareable description (expressible in complex ways) that can be conveyed to another. Second, it seems to me that we regard imagination as reliable in more active abstract problem solving. Thus, if I am asked to solve a problem along the following lines I would employ and rely on imagination. The question is how many cubes are created by taking a large cube and cutting it vertically, then horizontally, and then vertically but at right angles to the first cut. Most of us would imagine the described process and deliver our verdict by scrutinising the imagined outcome. But, third, it seems to me that imagery is central to what we call occurrent thinking. In occurrent thinking we engage with a problem

106  Paul F. Snowdon and formulate, to ourselves, aspects of the issue, as a means of forming a view. And the way we do this is, I suggest, by imagining speech. We employ what psychologists call phonic imagery.11 Now, this role is central to our cognitive lives, but it reflects two fundamental facts. The first is that we can control our imagery.12 We “run” the imagining process ourselves. Of course, we have no awareness of how we control the flow of images. There is no intermediary item that we knowingly employ to control the images. We simply, and as we might say, directly control the images. In fact, the centrality in our intellectual lives of occurrent thinking with its dependence on controlled imagery implies that the ability to have imagery and to control our imagery are amongst the most important psychological capacities we have.13 Second, the process is one of imaging speech, and its cognitive role for us in thinking out a problem depends on our prior understanding of the imagined speech.14 Crucially, in solving a problem about rose growing, say, what we imagine is not roses or red things, it is rather speech about roses or red things. This brings out why such imaging which is central to our cognitive lives cannot play an explanatory role in linguistic understanding. The function of such imaging presupposes and depends on an understood language. The important point is that dislodging images from the role assigned to them in understanding does not imply that they do not have extremely important roles in our cognitive lives.

6. Induction, Learning and Working out a Series Having dismissed the imagistic model of understanding, Wittgenstein introduces a type of example that figures prominently in his discussion. A character here called B is ordered to write down a series of signs (numerals) according to a certain rule. In a sense, then, this is the point that rules enter the discussion. Initially the example is simply writing down the ascending numbers in decimal notation, starting, in 143, with the sequence of numerals from 0 to 9, and then, in sec. 145, with the recursive application of these numerals in the decimal system. Wittgenstein primarily focuses on two things. He focuses on the process of teaching, and he also focuses on how to characterise the state of understanding that is created by successful teaching. Now, in 143 Wittgenstein points out that whether the learner learns depends on how he or she reacts to the inputs provided for them by the teacher. This may result in no learning or in mistaken learning or in proper understanding. (Wittgenstein developed a similar point in relation to the upshot from ostension, or ostensive definition, in 28.) In 145 he points out that we cannot state a limit of how much to test before we conclude that the learner has understood or grasped the rule or system. It is obvious that both these points are correct. However, it seems to me that Wittgenstein does not articulate in any full way what significance these points have. I want to develop them somewhat. But before I do there is one feature of Wittgenstein’s discussion that deserves highlighting.

Wittgenstein on Rule Following  107 In 144 Wittgenstein seems struck by the idea that saying that teaching may go wrong seems to be something that we can do without investigation. (In a language that Wittgenstein does not use here, it could be called “a priori.”) His response is to describe what saying this amounts to as “put[ting] that picture before him, and his acceptance of the picture consists in his now being inclined to regard a given case differently: that is, to compare it with this rather than that set of pictures.” It seems here as if Wittgenstein is offering a non-factual interpretation of modal language; such utterances affect how we look at things, what comparisons we are inclined to make. This attitude to modality, roughly that modal utterances are not properly true or false, surfaces at different times in the Investigations. My purpose is not to engage with it, but rather to note it. How, then, should we think of Wittgenstein’s remarks about the upshot of teaching? I want to propose what seems to me to be a very natural way of thinking about the phenomena that Wittgenstein is focusing on. I do not know whether this natural way of thinking is consistent with what Wittgenstein himself is intending to establish here, or, more generally, in the rule following section. Probably it is not. Let us start with the teachers determining whether the pupil has got it, and think of ourselves as the teacher. The natural way to think of the testing is that we are sampling what can only be limited data and making an inductive inference from that limited data. As we might say, we have tested the learner in various cases, with a successful outcome, and so we conclude that he has got the rule. This is simply like concluding something general from a restricted range of cases. Of course there is no number of cases that we can test that guarantee the result conclusively. It may be that next time the subject goes wrong, but in this such testing is like standard testing. In fact, the form of testing in our case can be more complex. The learner can be asked to explain the system to us in general terms. Their response to this request gives us more evidence as to whether successful learning has occurred. This general way of thinking about the status of testing in such cases seems perfectly acceptable. Let us now think about the learners, and the teaching of them. Now Wittgenstein stresses that the learning process need not work. This is again clearly true, but there is a natural way to think of this too. When a child is being taught we can think of ourselves as aiming to generate in the child a cognitive condition C, say, knowing the alphabet, or knowing the structure of the decimal notation. What we then do is to affect the child in a limited set of ways, which we can call W. These ways, included in W, are inputs into the system that we can think of the child as, that we assume will cause it to move into cognitive state C. In other words, we assume that the child arrives at the learning situation with the disposition to acquire C given input W. So learning depends on the presence in the learner of certain dispositions. What the disposition amounts to is the presence of a generalising faculty, which instils in the learner a cognitive condition which is more general

108  Paul F. Snowdon than the simple inputs. Now, obviously, the child in front of you need not have that disposition. In that case the intended learning will not take place. Instead the child might arrive with a disposition to respond to such inputs by acquiring what we might think of as a faulty understanding, or with dispositions that result in a response that is too chaotic to count as even misunderstanding. But normally we employ inputs in teaching that have been found to trigger the right output in the normal range of human learners. We are creatures who have experimented with teaching over the ages, and observed and handed on what has seemed successful. We have here a way of thinking of development. Normal humans are born with a nature that grounds dispositions to acquire dispositions, to acquire further dispositions, until they arrived disposed to learn, say, the decimal notation. And after that, having had the right inputs, given their dispositions to learn, they do learn and thereby acquire new dispositions to learn further things. But the whole process is contingent on the match at each stage between input and disposition. All this seems correct and illuminating to me, and it fits well with Wittgenstein’s remarks about our teaching and learning processes being dependent on shared forms of life. Now, I have spoken of the child as coming disposed to generalise beyond the limited samples or inputs. In secs. 209 to 213 Wittgenstein himself rather brilliantly constructs a dialogue around a view that seems close to the idea endorsed. It remains unclear to me quite what Wittgenstein is trying to reveal about it. Thus, in 209 Wittgenstein says: “But then doesn’t our understanding reach beyond all the examples?”— A very queer expression, and a natural one!—But is that all? Isn’t there a deeper explanation; or mustn’t at least the understanding of the explanation be deeper?—Well, have I myself a deeper understanding? Have I got more than I give in the explanation?—But then, whence the feeling that I have got more? As I read this Wittgenstein is trying to oppose the conviction that the understanding of the learner reaches beyond the examples. His line of thought seems to be that the learner does not pick up more because when I consider myself I cannot find that I have more than I give in the explanation. If this is Wittgenstein’s point it is not obviously correct. It would seem quite fair to say that when I present the examples to guide the learner I do not thereby present my general understanding. In saying this I am not committing myself to any particular picture of what my general understanding is, just that it amounts to something general. In 210 Wittgenstein continues the dialogue. “But do you really explain to the other person what you yourself understand? Don’t you get him to guess the essential thing? [. . .]—Any explanation which I can give myself I give to him too.”

Wittgenstein on Rule Following  109 Now, in one respect the words that Wittgenstein gives his interlocutor here are misleading. The idea that what the learner picks up goes beyond the examples should not be represented as the learner guessing. That is no more correct than describing someone who is learning something about, say, salt in general on the basis of certain samples should be described as guessing about salt. Rather the learner is engaged in generalising. However, Wittgenstein does have a genuine point in spite of this. There is nothing I can lay before myself that as it were encapsulates my own understanding that I cannot lay before the learner. Thus, maybe I can articulate a formula to myself, but then I can present the same formula to the other. And if I cannot articulate such a formula then there is nothing I can give myself that I cannot give the learner. However, the conclusion from this should not be that learners of language do not go beyond the examples and evidence, but, rather, that the condition of understanding is not itself presentable to another or to oneself in any such way. On this reading what Wittgenstein brings out is not that people acquiring understanding do not go beyond the evidence presented to them, but that the person with understanding cannot give to themselves more than they can give the learner. I suspect that Wittgenstein would not like what I am proposing as a natural response to some of his observations about teaching and testing for learning. However, the fact that some of his reflections, which we have just considered, and which grow out of his dialogue about teaching, are in fact assimilable within that natural approach lends, it seems to me, some support to it. Wittgenstein opposes some other ideas about understanding in a more direct way and I want to discuss two moves in his argument. In 149 Wittgenstein is opposing the quite natural idea that knowing the ABC should be thought of as a disposition, which explains the so called manifestations on occasions of that knowledge. Against this idea Wittgenstein brings the point that to attribute a disposition requires fulfilment of two criteria, one being what it does, but the other being what aspect of a things construction grounds the disposition. His point is that when we talk about knowing the ABC we completely lack the latter information. This argument invites the following response. We seem to attribute dispositions to things without having any knowledge of structure. People said that sugar is water-soluble without knowing anything about the structure of sugar (or indeed water). Indeed, part of the point of talking of dispositions is to convey the powers of things where we are ignorant of the composition of the things themselves. I read another passage as opposed to the same suggestion. In 153 Wittgenstein is opposing an idea which he formulates as amounting to the idea that understanding is something “hidden behind those coarser and therefore more readily visible accompaniments.” Now, that way of speaking fits the disposition idea. Against this Wittgenstein raises what seems to be an epistemological problem. Wittgenstein says; “And how can the process of understanding have been hidden, when I said ‘Now I understand’ because

110  Paul F. Snowdon I understood?!” In the next sentence Wittgenstein insinuates that the idea that the state of understanding is “hidden” is problematic. At least two points suggest themselves here. The first is that Wittgenstein’s own puzzles about our self knowledge of understanding count against simply saying that we know we understand simply because we understand. How we know we are in that state is quite unclear. Maybe it is like knowing we have a disposition. The second is that if there is some sense in which a disposition is per se hidden, there is no obvious reason to deny that the presence of understanding is not in the same way hidden. Whatever that notion of being hidden is, it is simply not obvious that aspects of ourselves are not hidden from us. Clearly that dispositions are hidden does not mean that we cannot easily get to know about them. My response then is to suggest that Wittgenstein does not produce any strong evidence that the standing condition of, for example, knowing the ABC is not a disposition, or something disposition like. Nor does he provide any evidence that the theory of understanding should not attribute a central role to the notion of a disposition.

7. Circumstantialism and Reading I want finally to engage with two related developments in Wittgenstein’s discussion. The first is his employment of the notion of circumstances in thinking about understanding, and the second is his investigation of reading between 156 and 178. The notion of circumstances is brought to our attention in 154 with these words. “If there is to be anything behind the utterance of the formula it is particular circumstances which justify me in saying I can go on.” What is the significance of this proposal? It is hard to interpret this remark. For one thing, Wittgenstein is stressing circumstances in relation to justification, whereas the dominant issue has been to what understanding is or amounts. For another, it is far from clear what Wittgenstein means by “circumstances.” But for all its unclarity and strangeness of focus, Wittgenstein seems to be introducing here a notion, that of circumstances, that is of considerable importance to him. Further, when Wittgenstein says (in 155) “This will become clearer if we interpolate the consideration of another word, namely ‘reading’,” it is hard not to understand Wittgenstein as suggesting that the role of circumstances in an account of understanding will become clearer. If so, this is more evidence that Wittgenstein attached considerable importance to “circumstances,” since he is about to clarify that notion over twenty sections. By “reading” Wittgenstein simply means pronouncing correctly what is written. It is not required that the reader understands what is said. If that had been built in then reading would simply be a case of understanding, and so hardly something that was sufficiently independent to shed light on understanding. Of course, by making this separation Wittgenstein invites the question: why suppose that understanding resembles reading? Why suppose that a consideration of reading illuminates understanding?

Wittgenstein on Rule Following  111 But taking “reading” this way Wittgenstein is dealing with two separate cases with similar structures. As Wittgenstein sees it, in both cases a subject, let us call him or her, S, does something, say pronounces a word correctly, or runs through a formula in their head. The latter is an inner event, whereas the former is an outer event. We can call the performance X. Now in each case the question arises as to whether the performance of X can be categorised as an instance of G. G in one case represents reading, and in the other properly understanding a sequence. Wittgenstein’s basic question is: what is it to be a G? His idea is that simply pronouncing the right sounds or simply formulating the right formula might not be a case of G. Maybe S just guessed the right sounds to make, and maybe S does not properly understand the formula. It seems to me that Wittgenstein is proposing that we think of both cases along the same lines. Namely, what makes X into a genuine G is a matter of the circumstances in which X occurred. This seems to me to be a sufficiently important element in Wittgenstein’s thinking to coin the name “circumstantialism” to apply to his approach. I rather feel that it is an underappreciated part of Wittgenstein’s thought. There are though two interpretative issues that I need to raise. The first, already hinted at, is whether Wittgenstein intended the invocation of circumstances to relate to the epistemology of attributions of condition G, be it understanding or reading, or whether he thought that the notion of circumstances should be brought into an account of to what G-ing is or amounts. The second question is what Wittgenstein understood by circumstances? The answer I propose to the first question is that Wittgenstein intends the notion of circumstances to figure in an understanding of what G-ing is and it is not intended to relate solely to the epistemology of judgements. One reason for saying this is in the reading sections; it is quite clear the Wittgenstein is considering attempts to say what reading is. He looks at proposals that identify it with a certain inner psychological occurrence, or with a certain mechanism, or with some general causal link. The notion of circumstances intrudes itself in this debate, which is not solely epistemological. Second, it is hard to escape the impression that for Wittgenstein there is no sharp contrast between clarifying the epistemology of the notion and clarifying the nature of for what it stands. This emerges in his tendency to clarify the nature of the feature in terms of its being the feature that we ascribe in certain language games, where the description of the language game includes epistemological features of attributions. These two reasons support our understanding Wittgenstein as regarding circumstances as central to clarifying the feature we are considering. What then does Wittgenstein mean by circumstances? It seems that he is contrasting the notion of circumstances with such things as conscious processes, and mechanisms, and causal links. Wittgenstein seems to be saying; do not link G-ing to such things, rather bring in (other types of) circumstances. Moreover when he gives examples he seems to bring in whether the X-ing occurred after normal instruction or not. (See for example 162.) He also stresses that which circumstances are relevant itself varies in different

112  Paul F. Snowdon circumstances. So, I believe, the natural way to interpret Wittgenstein’s circumstantialism is that the relevant circumstances are relatively external to the X-ing and do not involve links or mechanisms or special experiences. In fact, this is the best we can do since Wittgenstein does not really indicate what circumstances are relevant. Nor does he attempt any explanation as to why certain features belong amongst the relevant circumstances, whereas others do not. I have space to make only three points about what I am calling Wittgenstein’s circumstantialism. The first is that there are categories or notions which can usefully be treated along circumstantialist lines. Thus, one who says “I do” counts as consenting to marry someone only because they perform that speech act in certain circumstances. The second and rather dogmatic remark is that this approach is inconsistent with a strong intuition that the phenomena that Wittgenstein is dealing with are not ones for which a circumstantialist account is plausible. We strongly tend to think that whether someone is reading is a matter of what links the sight and the sound. And we strongly tend to think that whether someone has understood the rule depends on what happens within the subject as the expression of the rule occurs. The third point starts from a very interesting remark by Wittgenstein about reading. In 157 Wittgenstein contrasts what he calls a reading machine with what he calls a living-reading-machine, that is, a human reader. He remarks that with the former we can say that the machine started to read when certain mechanical links were installed, whereas he says that with people our judgements that they are reading are independent of mechanisms. I want to suggest that Wittgenstein misses the important point that this contrast reveals. The need to make a mechanism linking the visual capacity of the machine with its vocal production reveals that whether there is reading depends on whether there is some suitable causal link between input and output. In the artificial case the maker needs to make the mechanism. However, in the human case the object has a pre-existing nature which will yield such a link given appropriate training. Human trainers do not need to make or insert the mechanism, since the human learner already contains the complex physical basis for the emergence of such links. Fundamentally, though, both cases must involve such links and mechanisms. Reflection on the contrast that Wittgenstein notes should lead us in the opposite direction to that taken by Wittgenstein himself. And I think it is clear that the same, or a similar point, point would go for grasping rules and for understanding.

8. Conclusion I have tried to highlight some aspects of Wittgenstein’s discussion of understanding in the rule-following sections, and to provide a critical response to them. In what I have investigated he unearths some real problems about

Wittgenstein on Rule Following  113 understanding and about the epistemology of understanding, and also some real problems with some approaches to the phenomena. But in other respects he sets himself against ideas that seem correct or plausible, and so engages in that very common philosophical exercise of throwing out the baby with the bath water.15

Notes 1 I do not intend by this remark about Malcolm to convey the idea that his thought is devoid of interest or is of low quality, which would not be true at all. 2 In the Oxford of that time the causalist approach to action explanation was presented very effectively in famous lectures by David Pears. Listening to Pears attack on the non-causalist theory very strongly gave the impression that one was witnessing the stealing of sweets from blind children. 3 Saul Kripke, Wittgenstein on Rules and Private Language (Oxford: Blackwell, 1982). 4 I find writing about Wittgenstein especially hard. I am trying to avoid two types of approach. One approach is to basically comment on every section—which really requires a book. The other is to present an interpretation and simply relate some sections to it. I prefer to critically engage with more sections than the latter approach tends to, but to be less comprehensive than the former approach. 5 I have written at greater length about this question in P. F. Snowdon, “Private Experience and Sense Data,” in The Oxford Handbook of Wittgenstein, eds. O. Kuusela and M. McGinn (Oxford: Oxford University Press, 2011), 402–28. 6 Ludwig Wittgenstein, Philosophical Investigations (Oxford: Blackwell, 2001), section 255. 7 Here are, perhaps, two examples of more or less this structure of argument. John McDowell opposes an interpretation of these passages that Wright proposed, which, according to McDowell, committed Wittgenstein to what McDowell calls “idealism,” by saying: “We may well hesitate to attribute such a doctrine to the philosopher who wrote: ‘If one tried to advance theses in philosophy, it would never be possible to debate them, because everyone would agree to them’ (PI sec 128)” (John McDowell, “Wittgenstein on Following a Rule,” reprinted in J. McDowell Mind, Value and Reality, London: Harvard University Press, 1998). Again, William Child says that “the deflationary interpretation of Wittgenstein’s discussion of rules is . . . in many ways more faithful to his philosophy than the constructivist reading.” In fairness to Child it needs remarking that he himself is not entirely happy with the “quietist” reading. See William Child, Wittgenstein (London: Routledge, 2011). 8 An aspect of Wittgenstein’s attitude to allowing the philosophy can be challenged but nothing else is his thinking about religion. In relation to that it seems that Wittgenstein is opposed to what we might call philosophical criticisms of religion. But why is that? Wittgenstein must think that religious faith cannot be regarded as having a philosophical origin, since if it does then it is not beyond criticism. However, even if we accept that, Wittgenstein’s slogan that philosophy leaves everything as it is, does not imply that faith is beyond criticism, unless we have to regard any criticism of faith as amounting to philosophy, which is not allowed to alter anything. And we might very well wonder why we should describe any general criticism of faith as “philosophy.” Why would the criticism of religion amount to philosophy, but, say, the criticism of astrology, which, I assume, Wittgenstein would not rule out in advance, not amount to that? Sorting out these issues would take us too far away from rule-following, but they illustrate just how obscure in many central ways Wittgenstein’s attitude to philosophy is.

114  Paul F. Snowdon 9 I have assumed in the text that 138 is the start of Wittgenstein’s discussion of rule-following. But Rupert Read has pointed out to me that in 136 and 137 Wittgenstein is already talking of “fitting” and in the light of that it is natural to read 138 as a continuation of a discussion already started. I wish here to remain uncommitted on that question. 10 Herein lies the problem with the idea espoused by different philosophers, notably Descartes, that one ensures that one is right if one judges in accordance with a truth guaranteeing rule. Following the rule might guarantee truth, but what guarantees that one is actually following the rule? 11 See R. Passingham, What Is Special About the Human Brain (Oxford: Oxford University Press, 2008), chap. 8. It needs pointing out that not all occurrent thinking involves phonic imagery, for example, the thinking of the deaf will not involve that. Their occurrent thinking will have relate to other imagined modes of communication. 12 I do not mean to imply here that this ability emerged with Homo sapiens. We possess the ability and it is of considerable importance for us, but other and earlier creatures may have possessed it. 13 It needs remarking that the ability to control our images can be lost. Uncontrollable flashbacks suffered by people with post-traumatic stress seem to be examples. (I owe this example to Katherine Snowdon.) Other possible examples are musical hallucinations, in which subjects “hear” tunes which they cannot control. In normal circumstances, though, we possess control. 14 There is at this point a question left open that some may wish to ask: when we think by imagining speech, do we do so by imagining ourselves speaking or do we do so by imagining ourselves hearing speech? This is a genuine issue, but the more important thesis is that imagery has a central role in thought (whatever is imagined precisely). 15 I am very grateful to the many students whose questions and reactions in some recent seminars on Wittgenstein that I have given have very much influenced my thinking. An audience at the University of East Anglia gave very helpful feedback. I am also very grateful to the audience at the conference in Mexico for Mark Platts at which I presented the talk that became this paper for their comments, which have altered its content significantly. I also wish to express my gratitude to the conference organizers for inviting me to be part of such a stimulating occasion.

9 Kantian Neuroscience and Radical Interpretation Ways of Meaning in the Bayesian Brain Jim Hopkins Mark’s first book, Ways of Meaning, played an important (if non-evangelical) role in disseminating Davidsonian ideas in the philosophy of language. In his preface to the latest edition, he comments on “talk of the death of the kind of philosophy of language presented in the first edition of this book.” As he says, philosophical desires “to initiate funeral rites” are commonly exaggerated and premature. Still this raises the question as to how— fluctuations in philosophical fashion apart—we should now evaluate the philosophical program to which the book contributed.

1. Radical Interpretation In answering this we should bear in mind the full ambition of that program: to explicate the human abilities of understanding and meaning. Members of our species are capable of co-ordinating their activities, for good as for ill, to an extent that is unique among vertebrate animals. This is evident in the fact that almost everything we do, or make or use—the food we eat, the clothes we wear, the books we read, the tools we use, devices by which we communicate and share experience, the science and technology that plays an ever-increasing role in our lives—involves tacit or explicit co-ordination with very many others, in intersecting networks of co-operation or competition that span the globe. All this pivots on the way each of us can produce behaviour that is meaningful both to ourselves and to others. This enables us to pursue goals by moving in ways that are perceived as actions informed by desires, beliefs and intentions, and to communicate these and other states of mind by making sounds (or producing other interpretable signs) perceived as meaningful utterances. In thus knowing and communicating our thoughts we establish a socially shared consciousness which that gives us power over life and death never before concentrated in a single species. All these accomplishments— including the most precise and abstract reaches of science and mathematics— are sustained by the interpretive abilities that underpin meaning and understanding, and would wither in the instant should these fail. Davidson rightly regarded explicating these abilities as an important philosophical task. He sought to cast light on them by constructing a fully

116  Jim Hopkins explicit theory that would enable an imaginary scientific investigator— a “radical interpreter”—to explain partly similar data (the production of observable behaviour) by construing it in a partly similar interpretive framework (explaining the behaviour as action, including speech, caused by the beliefs and desires of its agents). In this, as is familiar, Davidson was extending Quine’s project of explicating meaning via an imaginary scientific radical translator. And Quine can be seen as following the later Wittgenstein, who had also sought to investigate our ability to understand meaning and motive, and particularly the way it sustained a social extension of individual intentionality. His final investigations also were conducted via the figure of an imaginary explorer—the first radical interpreter—whose task, like that set for Quine’s translator, was to understand peoples whose languages and forms of life he had never encountered before, but including, for Wittgenstein’s particular purposes, a group who spoke only in monologue.1

2. Kant, Helmholtz and Unconscious Prediction In seeking to understand perceptual and cognitive abilities via the construction of a scientific theory, Quine and Davidson perforce located themselves in another, and Kantian, tradition. This stemmed from the great nineteenth century scientist and mathematician Hermann von Helmholtz,2 and had previously influenced both Frege and Wittgenstein. Among other things Helmholtz was one of the first formulators of the principle of the conservation of energy. His research in sensory physiology—supplemented by a range of personal inventions such as the opthalmascope—expanded the science of optics and revolutionized the treatment of malfunctions of the eye. More recently his work has been formative both for the empirical psychology of perception and current developments in neuroscience. His relevance to analytical philosophy, however, remains to be fully acknowledged. Helmholtz and Kant Tim Lenoir describes Helmholtz’s relation to philosophy, and to Kant in particular, as follows: Although he considered himself a physicist, Hermann Helmholtz devoted more pages of his published work to physiological psychology and philosophical problems related to spatial perception than to any other subject. [. . .] Helmholtz pursued fairly consistently over many years a line of investigation centered around the questions, How do our mental representations of external objects get constructed? And how do those representations relate to the world of external objects? And as regards Kant Helmholtz considered his work in epistemology [. . .] as updating Kant’s views in light of new developments in experimental physiology. [. . .]

Ways of Meaning in the Bayesian Brain  117 Helmholtz thus considered himself to be more consistently Kantian than Kant had been himself. In his Critique of Pure Reason Kant had advanced the first detailed model of the human mind that did not assume, but rather sought to explain, basic aspects of consciousness and perceptual experience. According to Kant, we have a manifest image of ourselves as self-conscious subjects of experience and agents of action. We take our experience to be internal to our minds, but to disclose a world of physical objects and events external to our minds, and in spatio-temporal and causal relations with one another, including our own bodies and organs of perception. Our actions in moving among these objects and events are guided by our experience in perceiving them, and this arises from their effects on our sensory organs as we do so. Philosophers had long had reason to know that experience was not simply produced by effects on the sensory organs. Descartes had stressed that full conscious and apparently perceptual experience regularly arose in dreams, and hence without the participation of the sensory organs themselves. Accordingly Kant recognized that we must understand experience— and the whole conscious image of ourselves of which it is part—as somehow created by internal activity of our minds, and in a way that was prior to the sensory impingements that we understood as conscious experiences of the objects we perceived. Since this creative activity apparently preceded the conscious experience it produced, it had to be unconscious. Kant called the ­consciousness-creating activity synthesis and argued that it was governed by the basic concepts—e.g., of self, experience, time, space, substance, object, event, cause—in whose terms we experience the world. Kant’s notion of unconscious processes that created perceptual consciousness remained in place as (most) later investigators came to acknowledge that the organ of synthesis must in fact be the brain: so that, as Freud was to put it, “consciousness is the subjective side of one part of the physical process in the nervous system.”3 Helmholtz himself had definite ideas on this topic. As Lenoir says, he criticized the Kantianism of his time for setting up two worlds, an objective physical world and a subjective world of intuition, somehow causally related to one another but existing as independent, parallel worlds relying on some unexplained pre-­ established harmony between perceptions and the real world as the basis for the objective reference of knowledge claims. Helmholtz, believing it was necessary to escape the subjectivity of idealist positions asked: But how is it that we escape from the world of the sensations of our own nervous system into the world of real things?4 Helmholtz’s question—how do we escape from the world of the sensations of our own nervous system into the world of real things?—arises in both scientific and philosophical forms. As noted above, we are familiar with dreams as creating a kind of virtual reality: an emotionally charged and

118  Jim Hopkins presence-embodying image of ourselves as (seemingly) engaged in interaction with the world. In waking life, by contrast, this same internally generated image is so constrained by sensory impingements as to yield the largely veridical perceptual experiences by which we govern our movements and actions. But given the internality of its generation, how is the accuracy of this image secured? We can take this as a scientific question about the brain and nervous system, enclosed as they are within the skull and skin, and so physically remote from the objects and events we take them to represent. But a scientific answer to this question—an account of how the brain in fact achieves representational accuracy in perception—should also help to address the philosophical question as to how a conscious subject can know that she is not a brain housed in a vat, or embedded in a Matrix. So it is clear that Helmholtz’s enquiries, although scientific in their aim, also relate to questions that would be regarded as philosophical even in the narrowest confines of the analytic tradition. Helmholtz was deeply concerned with the relation between science and epistemology. Indeed, like Quine after him, he took epistemology to be the domain in which the differing disciplines could most fruitfully inform one another. Thus he says the problems considered fundamental to all science were those of the theory of knowledge: “What is true in our sense-perceptions and thought?” and “In what way do our ideas correspond to reality?” Philosophy and the natural sciences attack these questions from the opposite directions, but they are the common problems of both.5 And he notes that Kant’s purpose was to examine the source of our knowledge and its justification, something that will always remain the task of philosophy and that no generation can shirk with impunity.6 Helmholtz did not shirk this philosophical task. Rather he brought science to bear on it, in what he hoped would be a continuing collaboration between the disciplines. And his account of mental representations and their relation to the physical world combined scientific and justificatory thought in an extraordinarily prescient way. Synthesis and Prediction In explaining his notion of synthesis, Kant had stressed that the concepts that enable us to form our conscious image of the world are used both normatively and predictively in relation to it. As he says By synthesis, in its most general sense, I understand the act of putting different representations together, and of grasping what is manifold in

Ways of Meaning in the Bayesian Brain  119 them in one knowledge. [. . .] If we enquire what new character relation to an object confers upon our representations, what new dignity they thereby acquire, we find that it results only in subjecting the representations to a rule, and in necessitating us to connect them in one specific manner.7 The concepts by which we understand an object of experience require— so that our use of these concepts tacitly predicts—that experiences unfold “in one specific manner.” So, as Kant sought to bring out in the Analogies of Experience,8 in construing experiences via concepts of substances we expect, or tacitly predict, that such experiences will represent substances as enduring in space and time. Likewise (and as Hume had previously stressed) in construing experiences as of causes, we tacitly predict experiences of their effects. Again, in construing experience as of a material object like a house, or yet again as of our own bodies, we expect (predict) one perceptual sequence if we walk around a house one way, another if another. Perception guides movement, and so varies with it. Spatial Determination, Perception and Action Helmholtz saw the relation of movement to perception in a further and deeper way. Scientific observation indicated that the nervous system actively controls the collection of sensory data via bodily movements and changes that it initiates. Hence, as he stressed, all “sensations of the external senses” were “preceded by some sort of innervation” so that they were all “spatially determined.” This determination, moreover, was not solely governed by adjustments of the eyes or other organs of perception. Rather it was also brought about by intentional actions and other purposive movements, whose sensory feedback served to regulate and synchronize them. Unconscious Predictive Models Helmholtz therefore reconceived perception and action as co-ordinate parts of a single underlying process—one that related movement and perception in a predictive way, so as to achieve the active control of sensory data that was required for us to thrive. The “spatial determination” of sensory data for this purpose entailed the existence of powerful internal models of the objects and events we perceive, manipulate and navigate; and these models, as Helmholtz saw, would have to be at least partly constructed via the same active collection of sensory data as they were to facilitate. Helmholtz thought that such a construction of internal models should be regarded as a psychological process, and one that began early in life. So he hypothesized “a field of mental operations which has seldom been entered by scientific explorers,” the drawing of Induktionsschluesses,9 or inductive conclusions.10 He repeatedly stressed that these were “not conscious activities, but unconscious ones,” and according to his account they were more

120  Jim Hopkins or less continuously engaged with the data collected by the nervous system. Since these processes involved the use of models or hypotheses, they might now be described as a form of tacit model-based abduction.11 As philosophers of post-Chomskian cognitive science were later to do, Helmholtz argued that such inferential processes should be seen as required for many kinds of knowing-how that could not be understood as knowing-that.12 These unconscious abductive inferences were to be described in psychological or mental terms because their conscious conclusions—judgments of perception, belief or knowledge—were already described in this way. Judgments of perception, such as “I am seeing and touching a table” were based on data arising in (internal and external) sensory systems, and made use of concepts in synthesizing this data. The objects and events so perceived and conceived (the table, the positions and movements of the arm and hand in touching it, etc.) were causes of the data on which they were based. So the unknown concept-applying models (encompassing the seeing eyes and the extended arm, hand and fingers, as well as the table, its surface, the space between eye and table, eye and hand, etc.) could be compared to those by which scientists drew conclusions about causes of perceptual data that were beyond their pre-theoretical perceptual ken. As Helmholtz argued An astronomer, for example, comes to real conscious conclusions of this sort, when he computes the positions of the stars in space, their distances, etc., from the perspective images he has had of them at various times and as they are seen from different parts of the orbit of the earth. His conclusions are based on a conscious knowledge of the laws of optics. Still it may be permissible to speak of the psychic acts of ordinary perception as unconscious conclusions, thereby making a distinction of some sort between them and the common so-called conscious conclusions. And while it is true that there has been, and probably always well be, a measure of doubt as to the similarity of the psychic activity in the two cases, there can be no doubt as to the similarity between the results.13 In the tacit instances Helmholtz was considering the inferential process yielded particular predictive conclusions about the causes of sensory data. In his technical writings, as Lenoir explains, Helmholtz explicated such inferences as answerable to complex physiological analogues of theoretical simplicity and statistical accuracy.14 In his popular expositions, by contrast, he described them in statements that were lucid and compact, but far-­reaching in their implications. Thus, Each movement we make by which we alter the appearance of objects should be thought of as an experiment designed to test whether we have understood correctly the invariant relations of the phenomena before us, that is, their existence in definite spatial relations.15

Ways of Meaning in the Bayesian Brain  121 This was a radical claim, and Helmholtz intended it to hold not only for conscious perceptual experience but also at multiple levels of motor and sensory processing in the nervous system. As we shall see below, it can be also be said, with appropriate qualification, to hold for the guidance of animal movement generally, and over the whole domain of life.

3. Hypothetical Synthesis and Computational Neuroscience Perceptual synthesis is accomplished by the use of concepts, and these are realized in the brain. How can we understand the brain as using concepts to model, predict and explain experience—“the appearance of objects”—in the way Helmholtz envisaged? To see more about this let us use his example of astronomical thinking, and take the models he envisages for unconscious abductive inference as analogous to scientific models of the solar system, such as those propounded by Ptolemaic and Copernican astronomy.16 Such models can be more or less abstract and general than the particular hypotheses they may be used to generate, where the latter might explain and predict, say, data consisting in actual observations of the relative positions of celestial bodies (Helmoltz’s “perspective images”) in the night sky. Also such models may themselves be generated by models that are yet more general and abstract, as in the case of Copernican models and Newton’s or Einstein’s theory of gravitation. In such cases the confirmation or disconfirmation of a particular data-explaining hypothesis may be transferred to a series of more general models and hypotheses, with increasing distance, and increasing indirectness of evidential impact, from the data originally involved. Taken this way the view we are considering would be that we tacitly use our concepts to create models of ourselves and our situation in the world, and these in turn to generate particular hypotheses that explain the appearances of the objects and events we perceive, both externally and internally to ourselves. For these we can use “H,” “d,” and “e” to stand for “hypothesis” “data” and “evidence” respectively, and restrict ourselves to explanations in which the H’s specify internally or externally perceivable causes of the basic sensory data that—on this account—they unify, predict and explain. Going over some points about this will also be useful for relating Helmholtz’s version of Kantian synthesis to the radical interpretation of Wittgenstein, Quine and Davidson. For, as noted above, the latter all proceed via the notion of hypotheses used to explain data as in the practice of science. Causal Explanations as Consisting of Working Hypotheses Confirmed or Disconfirmed by Data or Evidence We characteristically frame casual explanations by proposing causal hypotheses to explain data—perhaps just to understand the data better, or perhaps because we find the data surprising, in the sense that we cannot

122  Jim Hopkins see, and want to understand, why the data are as they are. Such hypotheses explain and unify data by integrating them into larger but hypothetical causal ­patterns—these, of course, being patterns produced by the causal mechanisms or processes hypothesized for explanation. In this way the hypotheses explain why the data are as they are, and so render the data unsurprising, in the sense that they enable us to see that the data are as we should expect them to be, given the causal mechanisms or processes hypothesized to explain them. To say this, however, is to say that such hypotheses predict the data they serve to explain. A degree of prediction is inherent in showing that the data are as we should expect them to be, given the hypothesis—for in this, the data are represented as expectable in a way they were not before we brought the hypothesis to bear on them, and again in a way they would not be, given the negation of that hypothesis. In consequence we can regard an hypothesis as explanatory in the sense with which we are concerned just if the probability of the data given the hypothesis is greater than the probability of the data given the negation of the hypothesis, which we can abbreviate P(d given H) > P(d given not-H). Bayesian Predictive Confirmation This, however, means that data so explained become evidence by which the hypothesis can be confirmed or disconfirmed. For by Bayes’ theorem17 the credibility (probability) we assign to an hypothesis given data that it predicts [P(H given d)] should be greater than the credibility or probability [P(H)] we assign it prior to the explanatory prediction. But for data to increase the credibility (probability) we assign to hypotheses is for the data to constitute evidence that tends to confirm those hypotheses, and comparably for disconfirmation. So in Bayesian terms, and taking data now as evidence, we can regard e as confirming H just if P(H given e) > P(H), which will be so just if H predicts e in the sense that P(e given H) > P(e given not-H). Predictive Confirmation and the Brain Helmholtz took the concepts, models, hypotheses and data used in tacit abductive inference to be embodied in the nervous system. The data originated in the sensory systems that Kant described in terms of inner and outer sense, as these were affected by causes internal or external to the body. So for example in the case of reading by sight the data might arise in the response of neural cells in the retina of the eye to what Quine describes as “certain patterns of irradiation in certain frequencies” reflected from letters on a printed page, or again generated from a computer screen. These were what Helmholtz called “the sensations of the nervous system,” from which our epistemic problem (or that of the brain) was to “escape” by making trustworthy inferences about “the world of real things.”

Ways of Meaning in the Bayesian Brain  123 In the case of reading the “real things” would include the page, letters, words and sentences upon which the eyes were focussed, and also the action of reading in which the reader would take herself as engaged, the book she was reading, the author she took to have written it, the real or imaginary things she took the author to be writing about and much of the rest of the world, as represented by her concepts. As Frege noted, with “a few syllables” an author can evoke “an incalculable number of thoughts.” So in reading, or even glancing at a page, we are cognitively prepared to reckon almost anything as among the causes that might be cited in explaining the exact words (“epistemology,” “photon,” “vampire” “galaxy”) upon which our eyes might momentarily fix as we read. The data Kant described in terms of outer and inner sense are now often discussed in terms of exteroception, proprioception, and interoception.18 Roughly, exteroception gathers data from causes outside the body: from those of seeing, hearing, smelling, tasting, touching, tensions on the skin, etc. By contrast, interoception collects data from the interior of the body, and so from homeostatic functioning and the arousal of emotions; and proprioception collects data from the deployment of muscles, and so from movements, posture, etc. The data from these sources are processed in a way that integrates them into a single conscious image of the self in the world—the image available to us in the first-person perspective of current conscious experience—and the skin, which bounds the body and is the source of a multiplicity of data, plays an important role in delineating the boundaries assigned to the self. Helmholtz exemplified his claim that exteroceptive data are collected by “innervation” and movement by citing the movements caused by the muscles of the eyes.19 Although it may seem to us (and many philosophers insist) that when we look we simply take in the visual scene before us, our experiences in visual perception are based on data collected by unconscious but fast and systematic movements of the eyes. These concentrate light reflected by things (e.g., particular letters in words on a page) in a small part of the visible field—corresponding to a thumbnail seen with arm extended, or seven or eight letters of print on a page we are reading—on the most sensitive part of the retina (the fovea), for about a fifth of a second at a time, during which time, among other things, we unconsciously enlarge or contract the pupils of our eyes in accord with the emotional significance of what we expect or find. The resulting rapidly changing unconscious mosaic of thumbnail-sized “retinal snapshots,” each succeeding others in a fraction of a second, makes up a central part of the stream of visual sensory data. Aspects of the nature of these data are illustrated in Figures 9.1 and 9.2.20 Such conceptually governed and active but involuntary collection of sensory data—the kind in which the reader of this paragraph is now engaged— accords with Kant’s idea that prior to conscious experience sensory data are synthesized via concepts, as also with Helmholtz’s particular understanding

Figure 9.1 Conceptual guidance of data-collection by involuntary saccades in viewing a picture. Source: Wikimedia Commons.

Figure 9.2 Saccades and fixation points in reading. Source: Wikimedia Commons.

Ways of Meaning in the Bayesian Brain  125 of synthesis by analogy with theoretical unification. For in using concepts as hypotheses to collect, explain and unify fragmentary sensory data, the brain does seem, as Kant says, to “grasp what is manifold in [the data] in one knowledge.” This “one knowledge” is the interpretation of the data implicit in perceptual experience, which represents—not the fluctuating internal data, but—their significant distal causes. Conceptual Hypotheses and Sensory Data in Vision Helmholtz describes this in terms of the concepts used in perception by saying The idea of a body in space, of a table for instance, involves a quantity of separate observations. It comprises the whole series of images which this table would present to me in looking at it from different sides and at different distances; besides the whole series of tactile impressions that would be obtained by touching the surface at various places in succession. Such an idea of a single individual body is, therefore, in fact a conception (Begriff ) which grasps and includes an infinite number of single, successive apperceptions that can all be [unconsciously] deduced from it; just as the species “table” includes all individual tables and expresses their common peculiarities. Hence, as he says, The idea of a single individual table I carry in my mind is correct and exact provided [that via unconscious abductive reasoning] I can deduce from it correctly the precise sensations I shall have when my eye and my hand are brought into this or that definite relation with respect to the table [. . .] the idea of a thing is correct for him who knows how to determine correctly from it in advance [via unconscious abductive reasoning] what sense-impressions he will get from the thing when he places himself in definite external relations to it.21 Experiences as Hypothetical Representations of the Causes of Sensory Data The interaction between sensory data, Kant/Helmholtz synthesis and selfascriptive judgments of conscious perceptual experience is well illustrated in Dennett’s description of eye-tracking experiments in Consciousness Explained.22 When your eyes dart about in saccades the muscular contractions that cause the eyeballs to rotate are ballistic actions [. . .] whose trajectories at lift-off will determine where they will hit ground zero at a new target [. . .] if you are reading text on a computer screen your eyes will leap

126  Jim Hopkins along a few words with each saccade, farther and faster the better a reader you are [. . .] a computer equipped with an eye-tracker can detect and analyse the lift-off in the first milliseconds of a saccade, calculate where ground zero will be, and, before the saccade is over, erase the word on the screen at ground zero and replace it with a different word of the same length. What do you see? Just the new word, and with no sense at all of something having been changed. As you peruse the text on the screen, it seems to you for all the world as if the words were carved in marble, but to another person reading the same text over your shoulder (and saccading to a different drummer) the screen is aquiver with changes. The effect is overpowering. When I first encountered an eye-tracker experiment, and saw how oblivious subjects were (apparently) to the changes flickering on the screen, I asked if I could be a subject. I wanted to see for myself. [. . .] While I waited for the experiment to start I read the text on the screen. [. . .] “Why don’t you turn it on?” I asked. “It is on” they replied. Here two viewers of the same screen—the “oblivious” subject and another watching, say, over his shoulder—see (or experience) the screen in strikingly different ways. Their different visual experiences systematically reflect the differing visual data collected by the saccadic movements of their eyes, and so can be understood in Helmholtz’s terms as tacitly formed abductive hypotheses that explain and unify those perceptual data by reference to the objects or events that cause them. The oblivious subject, whose sensory data are so controlled by the eye-tracking mechanism as to be explained on the hypothesis that he is reading an unchanging text, sees such a text “as if the words were carved in marble.” The other, naturally collecting the fuller data available to her, rightly sees the screen as “aquiver with changes.” The differing conscious perceptual experiences here are just as we would expect them to be, on the assumption that their contents were determined by tacit abductive hypotheses—about successive letters, words and ­meanings— of the kind Helmholtz envisaged. Contrasting such conscious experiences with the data they synthesize, we can see that it is not the data but the hypotheses realized as perceptual experience that provide the continuity and order (the carving in marble) that we perceive the world to have. By contrast, the manifold unconscious data these hypotheses explain (Helmholtz’s “sensations of the nervous system”) fluctuate with a rapidity we scarcely imagine. Again, we can appreciate the hypothetical character of experience, and also the way it is structured top-down by our concepts, by reflecting that both “The words seemed carved in marble” and “The screen was aquiver with changes” could be reports of what was experienced in a dream. In such a case the dreamer’s (shut) eyes might well have been saccading, as they characteristically do in REM sleep. But there would have been no incoming data

Ways of Meaning in the Bayesian Brain  127 from the retina—no bottom-up visual impingements—for these saccades to collect. The brain apparently produces hypothesis-realizing experiences in both waking and dreaming consciousness—with the former constrained by sensory data but the latter not.23 Perception as Conscious Experience of Causes Taking the example above as characteristic, we can say that the brain operates tacitly and abductively to transform sensory data into conscious experience of their probable causes. Helmholtz expressed this by saying that in perceptual experience “the concept and cause combine,” so that “our impression, consequently, seems to us a pure image of the external state of affairs, reflecting only that condition and depending solely on it.”24 And in recent years—as indicated by the title of Karl Friston and colleagues: “Perceptions as Hypotheses, Saccades as Experiments”—25Helmholtz’s ideas have become central to a rapidly growing research program in neuroscience. Expositions of this program by Andy Clark and others26 have begun to attract attention from philosophers and cognitive scientists as well as the neuroscientists to whom it has become familiar. The beginnings of this program are often traced (e.g., by Friston)27 to work by Geoffrey Hinton and his colleagues, which transformed connectionism28 into the computational neuroscience of the Bayesian brain.29 Thus Dayan, Hinton, Neal and Zemel, began their seminal paper “The Helmholtz Machine,”30 by saying Following Helmholtz, we view the human perceptual system as a statistical inference engine whose function is to infer the probable causes of sensory input. We show that a device of this kind can learn how to perform these inferences without requiring a teacher. Here, in accord with then-recent findings about primate visual cortices, the authors envisaged the “statistical inference engine” of the brain as working via “a complex hierarchical generative model of its inputs”31—that is, a hierarchical model that served to predict sensory inputs—realized in the cerebral cortex. In the Helmholtz machine the layers of cortical processing were represented by layers of neuron-like processing units, connected hierarchically by two distinct sets of synaptic weights, alterable by built-in learning algorithms. The bottom layer corresponded to “sensory” input, and weights at higher layers could be adjusted to embody hypothetical representations of causes of input, as abstracted by pattern-recognizing algorithmic learning. This meant that, like the cortex, the Helmholtz machine could work both top-down (via one set of synaptic connections) and bottom-up (via the other). From the top “conceptual” layer down it acted as a “generative model” implementing inferences predicting input. From the bottom

128  Jim Hopkins “sensory” layer up it acted as a “recognition model,” learning the generative weights and recognizing inputs by inferring their most probable causes. This was done via an important conceptual innovation—that of assuming that the Helmholtz machine, and by analogy the brain, proceeded by minimizing the information-theoretic free energy of the probabilities with which it was concerned.32 Free Energy as Prediction Error As used for the Helmholtz machine, free energy is a quantity associated with a predictive model of the causes of data. This quantity is minimized by maximizing the power of the model to predict the data, and so by minimizing the unpredictability of the data, and therewith errors in predicting it.33 Crucially, this could be done from within the computational system of which both model and data were part, and by tractable computations that approximated Bayesian inference about causes external to it. So apparently the brain could perform the kind of abductive inference Helmholtz had envisaged, by using a generative model such as seemed realized in the cortex, and implementing a mathematically describable process of minimizing free energy. In light of this Karl Friston and his colleagues were able to frame an account of the brain as minimizing free energy that fit with many sources of data from neuroscience,34 as well as from perceptual and Freudian psychology.35 Howhy, Friston and Roepstorff have recently described this approach as follows: There is growing support of the idea that the brain is an inference machine, or hypothesis tester, which approaches sensory data using principles similar to those that govern the interrogation of scientific data. [. . .] Rather than trying to work backwards from sensory effects to environmental causes, neuronal computational systems work with models, or as we shall say hypotheses, that predict what the sensory input should be, if it were really caused by certain environmental events. The hypothesis that generates the best predictions then determines perceptual content.36 The predictive hypotheses are generated from the top (concept-realizing) level of the cortical hierarchy, and propagated down—so as to fix hypotheses at a series of lower (and hence successively more sub-personal and less integrated) layers, often working at successively faster temporal scales—until it reaches, and determines the collection of sensory input, at the physiologically separated loci of input for the various sensory systems. (This was illustrated in Figure 9.1, in the conceptual direction of fast collection of visual data in saccades.) From the top down, each level works both to predict activity at the level below, and to suppress the activity it succeeds in predicting. At the same time, each level apart from the highest also works bottom-up, to send

Ways of Meaning in the Bayesian Brain  129 unpredicted and so unsuppressed activity to the level above. This restricts input to higher levels to data that are so far u ­ npredicted—and this input works to alter the higher-level hypotheses so as to predict it, and so to bring about its own suppression. In this way the model constantly works on input that it is failing to predict, improving its hypotheses (at every level of neural processing) to reduce its ongoing errors. For such predictive brains, or for persons with brains of this kind, things would be as Helmholtz described. Each movement they made could be construed as an experiment in which they altered sensory data so as to test their conceptual hypotheses, and thereby how well they had understood the causes, including the bodies of which they are part, with which they interact. This locates the brain within the same epistemic horizons as we find ourselves—as part of a world of causes whose nature requires to be inferred from those of their effects that we encounter as sensory data. And while there could be no eliminating the possibility of divergence between our (or our brain’s) best hypotheses and the reality they serve to capture, this could be minimized through abductive practice. Prediction as Realizing Both Perception and Action In the preceding section, we used the idea that “the hypothesis that generates the best predictions then determines perceptual content” to describe the content of the experiences of different observers in eye-tracking experiments. But on Friston’s account the brain works not only to understand the world, but also to change it. The model also generates predictions—and further predictive representations including desires and intentions—about the sensory consequences of motor activity that, if it were initiated, would lessen error in the circumstances as modeled. In this case the brain can minimize error by acting to make its own predictions come true, and it does so, straightway appropriate predictions about the (proprioceptive) consequences of muscular/skeletal activity are generated, by engaging motor reflexes to correct those that are unfulfilled. In this way the brain’s “best predictions” assume the role of sensory goals that motor systems are driven to attain.37 To put the idea in Davidsonian terms, the brain makes the sentences that describe the agent’s desires true, by making true those that (would) describe the sensory consequences of acting on those desires. We saw such work on sensory consequences in the first diagram above, in the way saccades—and hence the ocular reflexes by which they ­implemented—are ultimately directed by viewers’ desires to gather information the experimenters have requested. Friston accordingly describes such movements together with intentional action as involving “active inference.” The predictive role of desires and intentions is signaled (and partly implemented) in commonsense psychology, by our mode of describing them in terms of the conditions of satisfaction or fulfillment that, if acted on, they will work to bring about. As Davidson noted, this is a way of describing

130  Jim Hopkins causes in terms of (predicted) effects. So here we can see this as amplifying causal theories of action (such as those of Fodor and Millikan), according to which desires are constituted by representations of the situations they operate to cause. On the present account beliefs, desires and intentions perform this task by setting sensory goals that determine the trajectory of motor activity, often towards what we regard as the conscious experience of satisfying particular desires.

5. Helmholtz’s Influence on Frege and Wittgenstein Helmholtz seems to have influenced both Frege and Wittgenstein. Consider Frege’s introduction of concepts in the Begriffschrift: Even most animals, through their ability to move about, have an influence on their sense-impressions: They can flee some, seek others. And they can even effect changes in things. Now humans have this ability to a much greater degree; but nevertheless, the course of our ideas would still not gain its full freedom from this ability alone [. . .] without symbols we would scarcely lift ourselves to conceptual thinking. Thus in applying the same symbol to different but similar things, we actually no longer symbolize the individual thing, but [. . .] the concept. This concept is first gained by symbolizing it; for since it is, in itself, imperceptible, it requires a perceptible representative in order to appear to us.38 Frege’s final sentence above anticipates Wittgenstein’s claim in Tractatus 3.1 that “In a proposition a thought finds an expression that can be perceived by the senses.” But he initially focuses on the action- and movement-guiding role of sensory experience sketched just prior (“flee some [sense-­impressions], seek others”). So like Helmholtz, and perhaps following his example, Frege took concepts to be linguistically enhanced extensions of natural representations that guide movement via sensory consequences they make it possible to predict. Thus, it seems plausible to suppose that when Frege speaks of the way “our ideas gain full freedom,” he is extending Helmholtz’s claim that “Having learned how to read these symbols we are enabled by their help to adjust our actions so as to bring about the desired result, that is, so that the expected sensations will arise.”39 As the term “symbols” suggests, Helmholtz’s influence is also apparent in Wittgenstein’s picture theory. The Viennese physicist Ludwig Boltzmann discussed scientific models in a number of his publications, including an Encyclopaedia Britannica article in 1903.40 These discussions make clear that Maxwell and Helmholtz were regarded as leading figures among a whole series of scientists, including Boltzmann himself, who developed the notion of a model or Bild for the exposition of physical theory.41 We can see something of Helmhotlz’s account in his discussion of the idea of a table quoted earlier. He says of the idea and the table that

Ways of Meaning in the Bayesian Brain  131 The one [the idea] is the mental symbol of the other [the table]. The kind of symbol was not chosen by me arbitrarily, but was forced on me by the nature of my organ of sense and of my mind. This is what distinguishes the sign-language of our ideas from the arbitrary phonetic signs and alphabetic characters that we use in speaking and writing.42 And he later puts this in terms that more explicitly look forward to the conception of the Tractatus: The excitation of the nerves in the brain and the ideas in our consciousness can be considered images of processes in the external world insofar as the former parallel the latter, that is, insofar as they represent the similarity of objects by a similarity of signs and thus represent a lawful order by a lawful order.43 Helmholtz’s ideas on this topic were developed further by his student Heinrich Hertz, whom Wittgenstein cites at Tractatus 4.04. In his Principles of Mechanics Hertz had written We form for ourselves pictures or symbols of external objects; and we make them in such a way that the necessary consequents of the pictures in thought are always the pictures of the necessary consequents in nature of the things pictured.44 Wittgenstein’s source for Hertz, in turn, was most probably Boltzmann himself,45 whose works Wittgenstein had in his library, with whom he had intended to study, and who in his expositions introduced another example that would be elaborated in the Tractatus: Nobody surely ever doubted what Hertz emphasizes in his book, namely that our thoughts are mere pictures of objects (or better, signs for them), which at most have some sort of affinity with them but never coincide with them but are related to them as letters to spoken sounds or written notes to musical sounds.46

6. Helmholtz and Analytical Philosophy As we have seen, Helmholtz had integrated Kantian epistemology with a scientific understanding of the nervous system by taking unconscious abductive inference to operate on the actual data collected by the sensory systems, and in such a way as to yield conscious experience of the “world of things” that caused these data. Yet despite Helmholtz’s influence on Frege and Wittgenstein, the tradition of analytical philosophy that stemmed from Frege was to forego this progress entirely. As Russell reported in his Our Knowledge of the External World as a Field for Scientific Method in Philosophy,47

132  Jim Hopkins the basic notions to be used by the new “logical-analytic method in philosophy” of which “the first complete example is to be found in the writings of Frege” were those of logic on the one hand, and sense-data on the other. These sense-data were “what is actually given in sense” and so despite being constitutents of “private spaces” or “private worlds,” were “the hardest of hard data” upon which empirical knowledge could be based. Frege likewise espoused “an inner world distinct from the outer world [. . .] of sense-impressions [. . .] of the contents of consciousness.” He indicated the primary epistemic status of these inner contents by saying I cannot doubt that I have a visual impression of green but it is not so certain that I see a lime-leaf. So, contrary to widespread views, we find certainty in the inner world while doubt never altogether leaves us in our excursions into the outer world. And he and Russell were of course agreed that this conception might seem to render knowledge of the external world impossible. Thus Frege again: If we call what happens in our consciousness idea, then we really experience only ideas but not their causes. And if the scientist wants to avoid all mere hypothesis, then only ideas are left for him, everything resolves into ideas, the light-rays, nerve-fibres and ganglion-cells from which he started. So he finally undermines the foundations of his own construction.48 This adhered to an understanding of perceptual experience that went back to Descartes, and that had apparently been refuted twice, first by Kant and second by Helmholtz and other sensory physiologists. As Kant had argued, the actual data should be seen as prior to conscious experience, and not as given but rather as actively synthesized by the use of concepts. As Helmholtz had argued, the conscious experiences thus synthesized had the epistemic status (not of data but) of hypothetical abductive conclusions about the probable causes of the data. Moreover, as Helmholtz had shown in his work on illusions, these conclusions were liable to systematic error, explicable in terms of the physiological processes by which they were produced. All this indicated that the experiences that Russell and Frege regarded as sense-data “in our consciousness” were not in fact “given in sense,” and did not have the status of “the hardest of hard data” (or even of data) for the basic epistemic workings of the mind or brain. Rather their traditional empiricist status as person-level conscious data depended upon their integration with the subpersonal neural mechanisms whose working Helmholtz had begun to describe. So when Helmholtz put the matter in traditional empiricist terms, he described “our knowledge of the actual world” as resting “upon experience, with constant verifications of its accuracy that we

Ways of Meaning in the Bayesian Brain  133 perform with every movement of our body.”49 In light of this there should have been no question of thinking of such knowledge as ultimately based on data other than those actually collected by the sensory systems; nor of considering, as Frege did, that for the scientist seeking data as opposed to mere hypotheses, “everything [. . .] the light rays, nerve-fibres, ganglion cells” resolves into “what happens in our consciousness.” But as things went, Russell and Frege’s picture of sense-data—the picture, as Wittgenstein would later say, that “forces itself on us at every turn”—set the stage for decades of work in analytical philosophy,50 and for decades of work on the part of Wittgenstein himself. As noted earlier, he finally rejected the conception—and, unlike most philosophers, the picture of the internality of consciousness of which it was a part—via the account of correctness for the application of concepts that he developed through his own use of radical interpretation.51 Over the same period analytical philosophy also ignored Helmholtz’s account of abductive inference in the nervous system. Philosophers began giving serious consideration to such an idea only after Chomsky reintroduced it in describing language-learning as requiring “what from a formal point of view is a deep and abstract theory [. . .] [the] concepts and principles of which are only remotely related to experience by long and intricate chains of unconscious inferential steps.”52 And finally—over a century after Helmholtz’s work, and as he said, “now that we have stopped dreaming of reducing science to sense-data”—Quine independently introduced the kind of epistemology that Helmholtz had practiced, together with some of the ideas Helmholtz had advanced.

7. Quine, Davidson and Naturalized Epistemology According to Quine’s53 account Epistemology still goes on, but [. . .] as a chapter of psychology and hence of natural science. It studies a natural phenomenon, viz, a physical human subject. [. . .] This human subject is accorded a certain experimentally controlled input [. . .] certain patterns of irradiation in certain frequencies, for example, and in the fullness of time the subject delivers as output a description of the three-dimensional world and its history. We study the relation between input and output in order to see how evidence relates to theory, and in what ways one’s theory of nature transcends any available evidence. [. . .] We are studying how the human subject of our study posits bodies and projects his physics from his data, and we appreciate that our position in the world is just like his.

134  Jim Hopkins And this is epistemologically significant because Our talk of external things, our very notion of things, is just a conceptual apparatus that helps us foresee and control the triggering of our sensory receptors in the light of previous triggering of sensory receptors. The triggering, first and last, is all that we have to go on. Quine is surely right that sensory input (and as opposed to the philosophical phantasm of sense-data) can be taken to provide a basis for experience, belief and knowledge. (This is also illustrated by the use of sensory input in “brain-in-vat” scenarios that seek to render sense-data skepticism consistent with acknowledging the metaphysical dependence of conscious experience on the physical brain.) Like Quine, Davidson held that there is “a causal bridge that involves the sense organs” between “external events and ordinary beliefs”; and also that it was an error to try to cross this bridge “with sense data, uninterpreted givens, or unwritable sentences [experiential protocols] as its impossible spans.”54 But he was willing “to count myself a naturalized epistemologist” only in the sense that epistemology should be third-person and “externalized,”55 as in his project of radical interpretation. Davidson regarded thought as holistic because the application of one concept to an object of belief or perception involved the application of many others, and externalistic because in basic cases the contents of thoughts were determined by the objects and events that caused them. So when he compared his own views with “what we now know about neurons, neural nets, and the processing of information (so-called) in the brain,” he emphasized how three central and partly a priori features of his account—holism, normativity and externalism—might be thought to distinguish it from “serious” science.56 But since the neuroscience with which we are concerned describes the processing of (Shannon) information via concepts that determine the contents of perceptual experiences by specifying the objects and events that cause them, it has these features as well. We can see this in the way the hypothesis-generating conceptual level that dominates cortical processing always meets the disparate sensory data it is gathering with a current hypothesis tacitly involving the whole range of objects and events conceived as possible objects of perception. A recent study of ongoing cortical conceptualization during the perception of a range of familiar objects and events coheres with such a view. Huth, Nashimoto, Vu and Gallant57 mapped a series of short films to 1705 hierarchical categories of the objects and events that appeared in them, as taken from the WordNet lexicon. They then mapped the categories to the cortical activity that arose as viewers recognized (categorized) those objects and events while watching the films. Under statistical analysis the categories could be seen to correspond to locations in an apparently holistic representation—a continuous multidimensional semantic space—spread smoothly over the cortical surface. Although far short of a precise tracking of the cortical activities

Ways of Meaning in the Bayesian Brain  135 involved, this has a prima facie claim to be taken as an empirical illustration of Davidsonian holism and externalism, as realized in the processing of sensory data.

8. Epistemology, Evolution and Interpretation Philosophers have often criticized Quine for giving up the attempt to evaluate claims about these topics, and seeking instead merely to describe them. But if we take his suggestion in terms of the program sketched above this will not be so. Rather, the abductive inferences by which a human subject “posits bodies and projects his physics from his data,” as studied in the example above, should be evaluable by the standards of Bayesian epistemology.58 Still, and in the same naturalistic perspective, we should also assume that the overall function of the brain’s modeling of the world is to promote reproductive success; and this would entail that its activity might further that end, as opposed to approximating the real causes of sensory data. Thus, we might expect a high degree of veridicality for the basic perceptual judgments Helmholtz and Davidson emphasize, and that figure in the study above. But we should also expect the kind of systematic susceptibility to perceptual illusion Helmholtz found for situations with which evolution had not prepared us to cope accurately.59 Likewise, and turning to the naturalized epistemology of interpretation, we might expect a high degree of accuracy in the cognitive understanding of language, upon which so much of the rest of our thinking depends. But as regards the understanding of persons and their actions, we might also expect something like the kinds of self- and group-serving biases we manifest in understanding ourselves as opposed to others, or again in understanding ingroups with whom we co-operate, as opposed to outgroups with whom our ingroups compete.60

9. Radical Interpretation in the Perspective of Neuroscience Davidson’s project was “not to describe how we actually interpret, but to speculate about what it is about thought and language that makes them interpretable.”61 At most, he argued, his account might provide “a model [or ‘satisfactory description’] of an interpreter’s linguistic competence.”62 This, again, was “not about the details of the inner workings of some part of the brain.” Still he regarded it as trivially true that “if the theory does correctly describe the competence of an interpreter, some mechanism in the interpreter must correspond to the theory.”63 Many influenced by Davidson, from Gareth Evans onwards,64 have tried to describe the corresponding mechanisms in more detail, and in sub-­personal terms that admit neural realization. The idea that the corresponding neural systems themselves employ models and hypotheses should encourage us to take this further, since it raises the question how far the models and

136  Jim Hopkins hypotheses tacitly employed by the brain might correspond to those Davidson explicitly provided. Commonsense Psychology and Truth-Conditional Semantics In Ways of Meaning, Platts had argued, in accord with John McDowell, that a theory of truth “will only be part of an overall theory of understanding” and one whose contribution required to be assessed via ascriptions of attitudes like desire and belief.65 This corresponds to the two main components of Davidson’s final model: a (Bayesian) decision theory, cast in terms of sentences, yielding degrees of belief and strengths of desire from preferences revealed in choice; and a theory of truth for interpreting the sentences ordered by the theory. In Davidson’s minimalistic and non-question-begging “official” procedure for radical interpretation, the interpreter systematically compared the interpretee’s choices as to which of a pair of uninterpreted sentences the interpretee would prefer to be true. The ordering of preferences over sentence enabled her to assign subjective probabilities and utilities to them, and so to take them as expressions of belief and desire. She could assign semantic content to these by finding “what episodes and situations cause [the interpretee] to prefer the truth of one sentence to another.” This assumed both rationality (e.g., transitive choices) and first-person authority (choices that correlated with non-verbal action) on the part of the interpretee. In the simplest and most basic case, for example, a sentence construed as expressing a desire on which an interpretee would act would be one that would be made true by her acting. Such a sentence would be assigned relatively high utility and low probability immediately prior to action, and relatively high probability and low utility afterwards. The corresponding changes in preferences would be mapped to the events that caused them, that is, to the changes in the interpretee’s nervous system, voluntary movements and the targets of these in the environment that constituted the satisfaction of her desire. Davidson later said that this “technically rather byzantine” procedure could be brought “closer to psychological reality” by integration with everyday interpretive understanding.66 In this we could regard ourselves as cross-checking hypotheses about the truth-conditions of an interpretee’s utterances, taken as authoritative expressions of belief and desire, against hypotheses as to the beliefs and desires explaining her non-verbal acts (those that “speak louder than words”). Given an interpretee’s co-operation (required also for the “official” procedure), this would enable us to test our understanding of the meanings of her sentences against our understanding of the beliefs and desires upon which she acted as often, and in as varying circumstances, as we chose to arrange, and thereby also to confirm the extent of the first-person authority upon which we were relying. In this way we might attain the kind of precision in understanding one another’s language that we seem to enjoy.

Ways of Meaning in the Bayesian Brain  137 Interpretation and Neuroscience So far we have considered Davidsonian hypotheses about the truth-conditions of sentences on the one hand, and ascriptions of desire and belief in the explanation of action on the other. It seems reasonable to take these (as well as many other accounts) as related by the claim that the use of concepts in framing models and hypotheses about the causes of sensory data is central to the working of the brain. Previously we briefly considered a study describing the brain as forming (holistic) conceptual representations of objects and events as they are perceived. Similar studies could refine our understanding of the mental and neural representation of words and sentences. And roughly, insofar as we use the same concepts of things in synthesizing perceptual experiences of them as in forming desires and beliefs about them, then accounts of language that link words with concepts should also link them with their referents (and vice-versa), and the contents of desires and belief about these should coincide with the referential truth-conditions of the sentences we use to ascribe them. Davidson also emphasized that interpretation was successful just when a hearer understood a speaker’s utterance in the way the speaker had desired that it be understood. He took this to show that a semantic theory construed as a “portable interpretation machine”—that is, “a machine which, when fed an arbitrary utterance (and certain parameters provided by the circumstances of the utterance) produced an interpretation”—was insufficient to yield the required interpretive matching.67 This emphasis on interpreters’ co-ordination with interpretees’ desires distinguishes Davidson’s account from many versions of truth-conditional semantics that have followed in his wake,68 and much work in the philosophy of language has been devoted to understanding how such matching might be achieved.69 On the present account the matching is produced by the hearer’s tacit abductive modeling of the speaker and her desires in speaking—but working in such a way that the hearer fulfils the speaker’s desires without explicitly interpreting them.70 All this relates interpretation, radical and everyday, to the interactive abductive understanding—of one another’s desires, beliefs and other motives—that underpins human group co-operation. Helmholtz/Bayes neuroscience casts light on this as well. We saw that action (including speech) minimizes prediction error by making predictions about the sensory consequences of action come true—as in speech a Davidsonian speaker makes her proprioceptive (articulatory muscular) predictions about the consequences of vocalization come true by uttering a sentence, and thereby starts to make her desire in speaking (and thereby sentences describing it) come true as well, as she determines by the further sensory consequences of hearing herself saying what she wanted to say, seeing her hearer start to understand to understand or misunderstand what she intended to convey, and so on. For all actions an agent’s brain must model a series of causes including (i) the motives that are the internal causes of forthcoming bodily movements (in this case the agent’s desires in speaking), (ii) the bodily movements (in this case

138  Jim Hopkins of speaking) that these are to cause (iii) the perceivable changes in the causes of the sensory input (the sounds of her own voice, the visible signs of uptake on the part of her hearer) these bodily movements are to cause (the speakers experience of the satisfaction of her desires in speaking); and (iv) the changes these produce in the internal causes with which action originates (the pacification of the speaker’s desires by the experience of their satisfaction). This means that an agent’s brain must predict changes in both internal and external sensory data from internal causes, and changes in both internal and external causes from sensory data. For speech and other actions aimed at causing changes in others, the external causes must include the movements and actions (e.g., those of understanding) of the others to be changed. This requires internal models of the motivational causes of others’ movements which Bayesians sometimes call “models of inverse planning.”71 The representation of commonsense psychology in these models is continuous with Davidson’s, and like his must be extended to include irrationality. Friston and his colleagues have described how such models can be framed via the brain’s re-using of the predictive capacities required for intentional action.72 As such neuroscientific “inverse models” are framed via the use of concepts, they combine elements of different philosophical approaches (simulative and theoretical) in a single account. Here again the explanatory framework of neuroscience suggests paths for progress in the philosophy of mind and language.

10. Life and Death in Philosophy In the perspective sketched in the preceding section, there is no question of the death of the philosophical program of which Ways of Meaning is a part. Rather the philosophy practiced in that book, like the work of Quine and Davidson themselves, should be seen as returning analytic philosophy to the Kantian and scientific roots from which it first drew nourishment, by renewing those that were severed by its founder’s adherence to the notion of sensedata. And if this program, and analytical philosophy with it, are increasingly integrated with neuroscience, this is surely all to our good. In this case the fate of our theories will be, as Einstein said, the fairest for which mortal theories can hope: to live on as part of those that succeed them. Appendix: on an Objection to Helmholtz Philosophers frequently criticize Helmoltz and those who take themselves to have learned from him for misusing the notion of prediction. The criticism is that to speak of tacit prediction (or again of the “predictive brain,” “predictive coding,” “predictive processing,” etc.) is to misuse the notion of prediction. This notion, the criticism goes, can rightly be applied only to the conscious activities of human beings, such as the scientists Helmholtz describes as predicting the positions of celestial bodies. It cannot rightly be

Ways of Meaning in the Bayesian Brain  139 extended to representation per se, nor to hypothetical unconscious activity of the mind, nor again to the brain, the visual system or any other subpersonal thing or process. Since this criticism might be directed against the argument herein, as well as many other claims in recent scientific discourse, it may be worth disarming it. This can be done by specifying a notion of predictive representation that is free from confusion, and that we can take to underpin the uses in question. Ruth Millikan’s notion of biosemantic representation, which can be brought into the neuroscience and philosophy of mind and language we are considering, serves this purpose well. Biological Representation As is familiar, Millikan takes representations in living things to be concrete particulars—characteristically events in behaviour or physiology— produced by certain physiological (producer) mechanisms, and used by other (user) mechanisms. These have been selected for co-operating in discharging some further evolutionary function of the users, often in the governance of behaviour, but which differs from case to case. The items so produced and used qualify as representations because their role in the overall process requires that they bear a particular kind of mapping—such as might be expressed in a rule of representation like “Socrates” stands for Socrates or “man” is satisfied by men—to a range of other items, those that, intuitively, they represent. So in this account representation is explicated as a kind of rule-governed mapping. Representations must map in particular ways to the things they represent, and it is in light of such mappings—e.g., those that were discovered via von Frisch’s “radical interpretation” of the dances of honeybees— that we regard the former as representing the latter. We can readily apply this idea to non-biological entities that we take to use representations, such as computers or robots. We might design a mobile robot that used a video camera to produce representations of the space and objects around it, where these were to be used to enable the robot to navigate the objects, say by motors that drove wheels on which it was mounted. For this we would require the patterns produced by the camera to map to the objects to be navigated, and in the particular ways that would enable the motor mechanisms to use the patterns in successful navigation. The particular representational patterns produced by the camera, that is, would be required to map to motoric mechanisms in ways that enabled the function for which this representation producing and using system was designed. This is in turn would distinguish representations (outputs from the ­representation-producing camera) that were accurate or correct from those that were not, as an account of representation seems required to do. But in this case the status of the patterns as representations would depend upon our having designed them to fulfil a particular role, and hence upon

140  Jim Hopkins our thoughts and intentions as expressed in the speech and writing that ­co-ordinated the project. This would mean that no light would be cast on these more basic forms of representation. In Millikan’s account, by contrast, the representational roles of states of mind like thoughts and intentions are also explained. Representation is seen as a natural biological phenomenon, such as we expect to find in our own and other animal brains. Taking a still broader perspective, we can see Millikanian representation-producing and using devices as having evolved to subserve the most basic processes of life, including the replication, expression and modification of genetic material by producer and user mechanisms in individual cells. These representational processes ramify throughout the whole functional mesh of communicating cellular systems that make up a complex multicellular organism.73 Such natural forms of representation— and particularly those realized in the nervous system, as in perception, thought and action—provide the foundations of the further representational practices established via learning and in culture, and including the production of representational artefacts like written texts, abacuses and computers, and so on. Representation in Single-Celled Organisms Millikan often uses the example of the dances of honeybees, the parameters of which map to the locations of pollen, water and other resources to be exploited by the hive. Still more basic examples are found in the simplest of living things, organisms that consist of a single cell. Each of these is a distinct individual, with a body bounded by a protective membrane housing the genetic material and the mechanisms of replication, expression and modification that implement the vital processes by which it maintains and reproduces itself. For survival and reproduction these cells, like all other organisms, must extract and incorporate nutrients from the environment, and also avoid toxins and other dangers that might impinge from there. Thus consider the e coli bacteria that swim in our guts. These are far simpler than the cells of our bodies (about a thousandth of their size, as they lack the mechanisms of differentiation and co-operation required for multicellular life). Their bodies are continually battered by random molecular movements in the fluids around them, so they must move non-randomly to ensure nutrition and safety. They do this by activating tiny “flagellator motors,” directed by “sensors”—transmembrane chemoreceptors—that use current chemical gradients in the environment to predict future gradients expected on current motion. These “sensors” release “signalling proteins” that drive the motors and the cell forward towards nutrients, but pause for toxins, letting the cell “tumble” out of dangerous paths.74 As their designation implies, the motor-directing “signalling proteins” are representations. Their production maps to, and so represents, the nutrients and toxins ahead. Accuracy in this is required for e coli reproductive success, which the organism accomplishes by its simpler version of the

Ways of Meaning in the Bayesian Brain  141 mechanisms that replicate and use DNA in ourselves. These enable the cell to duplicate its genome at something like 2000 neucleotides per second, with one “mistake” per billion.75 This duplication, in turn, can be regarded as proceeding via representations of the kind we have been considering, with the producer and user mechanisms those that manipulate the genetic material inside the cell, and the “mistakes” misrepresentations produced in this process—tokens that fail to map to what they are supposed to represent in such a way as to enable the mechanisms that use them to perform their replicative functions. Considered in this light, the representation-guided movement of e coli indicates that the conception Helmholtz framed to understand human perception and movement applies far more widely. All living organisms must be “spatially determined” in Helmholtz’s sense, because all must do as e coli do—they must extract vital resources from the environment while avoiding dangerous impingements. Very many do this via systems of representation selected for directing their movements to these ends. Where this is so the movements can be thought of as experiments testing (often, as in the case of e coli, by life or death) how accurately these representations depict their environments. As Popper remarked, we differ from many other animals in co-operating to produce scientific hypotheses—explicit and shared predictive representations—that can die in our stead. Predictive Representation So is it a philosophical mistake to describe e coli chemoreceptors as predicting chemical gradients of nutrients and toxins? This would seem analogous to saying that the brain, or the cortex or representations in these, predict the sensory consequences of movements of the eye, or the movements of intentional actions. If a reductio of this view of the brain is possible, it might be thought provided by such examples as e coli, whose mechanisms of representation are far removed from those of human thought. But the notion of prediction seems to be used in the same way in all these different cases—it is used to characterize the working of biological representations that are required to map to future events or states of affairs, often in complex and highly co-coordinated ways. In this the function of these representations (like descriptions of weather on the face of a barometer) is reasonably described as predictive. The guidance of biological movement is almost always effected by representations that are predictive in this sense— as these range from e coli protein signals through honeybee dances to the concepts that direct human saccades and intentional actions. Neither the coherence nor the usefulness of this concept should be denied.

Notes 1 On radical interpretation in Wittgenstein, see my “Wittgenstein, Davidson, and Radical Interpretation,” in The Library of Living Philosophers: Donald Davidson, ed. F. Hahn (Chicago: Open Court, 1999). Many others have also treated

142  Jim Hopkins interpretation as a matter of scientific hypothesis. Turing, who pioneered the use of Bayesian methods in cryptography, said that “There is a remarkably close parallel between the problems of the physicist and the cryptographer. The system on which a message is enciphered corresponds to the laws of the universe, the intercepted messages to the evidence available, and the keys for a day or a message to important constants which have to be determined” (quoted in J. Leiber, Invitation to Cognitive Science [Oxford: Blackwell, 1991]). 2 Lydia Patton, “Hermann von Helmholtz,” The Stanford Encyclopedia of Philosophy (Winter 2012 Edition), ed. Edward N. Zalta, http://plato.stanford.edu/ archives/win2012/entries/hermann-helmholtz/. 3 Sigmund Freud, “Project for a Scientific Psychology,” in The Standard Edition of the Complete Psychological Works of Sigmund Freud, vol. I, ed. J. Strachey (London: Hogarth Press, 1895/1950), 311. 4 T. Lenoir, “Operationalizing Kant: Manifolds, Models, and Mathematics in Helmholtz’s Theories of Perception,” in The Kantian Legacy in 19th Century Science, eds. M. Friedman and A. Nordman (Cambridge, MA: MIT Press, 2006), 142. 5 H. Helmholtz, “Recent Progress in the Theory of Vision” [1868], in The Selected Writings of Hermann von Helmholtz, ed. R. Karl (Middletown: Wesleyan University Press, 1971), 368. 6 H. Helmholtz, “Über das Sehen des Menschen” [1855], Vorträge und Reden von Hermann Helmholtz, vol. 1, 5th ed. (Braunschweig: F. Vieweg, 1903), 88. 7 I. Kant, The Critique of Pure Reason, trans. N. Kemp Smith (London: Palgrave Macmillan, 1963), 224. 8 See Kant’s Critique of Pure Reason, as cited above, 308–27. 9 I am grateful to Leonardo Nascimento for help with Helmholtz’s language and concepts here. If we consider Helmholtz’s explication of tacit induktionsschluesses by reference to the explicit practices of scientists, as appears later in the text, he seems to have in mind a tacit version of the process Pierce was later to describe as abduction. His own phrase, however, also indicates the conclusion-like nature of the tacit reasoning he seeks to describe: perceptual experience has the character of a conclusion as to the causes of the data synthesized in it. 10 See James Hawthorne, “Inductive Logic,” in The Stanford Encyclopedia of Philosophy (Winter 2012 Edition), ed. Edward N. Zalta, http://plato.stanford.edu/ archives/win2012/entries/logic-inductive/. 11 Igor Douven, “Abduction,” in The Stanford Encyclopedia of Philosophy (Spring 2011 Edition), ed. Edward N. Zalta, http://plato.stanford.edu/archives/spr2011/ entries/abduction/. 12 See Helmholtz “Recent Progress,” 217 ff; and for a post-Chomskian comparison Jerry Fodor, “The Appeal to Tacit Knowledge in Psychological Explanation,” The Journal of Philosophy, vol. 65 (1968): 627–40. 13 Helmholtz, Treatise on Physiological Optics [1867], trans. J. P. C. Southall (Rochester: Optical Society of America, 1924), vol. III, 23, http://poseidon. sunyopt.edu/BackusLab/Helmholtz/. 14 As Lenoir explains in his “The Eye as Mathematician: Clinical Practice, Instrumentation, and Helmholtz’s Construction of an Empiricist Theory of Vision,” in Hermann von Helmholtz and the Foundations of Nineteenth-Century Science, ed. David Cahan, (Berkeley, CA: University of California Press, 1993, 109–53), Helmholtz used both the principle of least action, and a least-squares account of accuracy, in describing the working of the eye. See section 5 and following. 15 Helmholtz, “The Facts of Perception” [1878], in The Selected Writings of Hermann von Helmholtz, ed. R. Karl (Middletown: Wesleyan University Press, 1971), 384. His stress on experiment appears also in the Treatise, 45–6.

Ways of Meaning in the Bayesian Brain  143 16 See Roman Frigg and Stephan Hartmann, “Models in Science,” The Stanford Encyclopedia of Philosophy (Fall 2012 Edition), ed. Edward N. Zalta, http:// plato.stanford.edu/archives/fall2012/entries/models-science/. As these authors note models can be either models of phenomena or models of data, and a given scientific theory, or set of hypotheses, can function as both. I will not attempt to disentangle these issues in the text. 17 See James Joyce, “Bayes’ Theorem,” in The Stanford Encyclopedia of Phi losophy (Fall 2008 Edition), ed. Edward N. Zalta, http://plato.stanford.edu/ archives/fall2008/entries/bayes-theorem/, as well as Douven, previously cited on abduction. 18 These terms were introduced by Charles Sherrington in his The Integrative Action of the Nervous System (New Haven: Yale University Press, 1906). 19 Thus, see the discussion in his “Recent Progress in the Theory of Vision,” 216: “A movement of the eye which causes the retinal image to shift its place upon the retina. . . . In this way we learn to recognize such changes as belonging to the special phenomena which we call changes in space.” As Lenoir explains (“The Eye as Mathematician,” 147), this passage is part of Helmholtz’s argument that “the mind is in effect performing a series of experiments with the eye, testing the hypothesis of an object.” (Helmholtz debated the nature of the movements involved with Hering, and they remain a matter of discussion.) See O. Coubard, “Saccadic and Vergence Eye Movements: A Review of Motor and Premotor Commands,” European Journal of Neuroscience, vol. 38, no. 10 (2013): 3384–97. Helmholtz applied the same ideas to touch, e.g., in “The Facts of Perception,” 377. 20 Diagrams gratefully reproduced from Wikimedia Commons. The first is from Alfred Yarbus, Eye Movements and Vision (New York: Plenum, 1967), reproduced in “Eye Movement in Reading,” Wikipedia, https://en.wikipedia.org/wiki/ Eye_movement_in_reading. The second appears in “Eye tracking,” Wikipedia, http://en.wikipedia.org/wiki/Eye_tracking. 21 Helmhotlz, Treatise on Physiological Optics, vol. III, 23. 22 Daniel Dennett, Consciousness Explained (New York: Penguin Books, 1993), 360–1, Kindle. It is worth noting that the multiple drafts in Dennett’s account partly correspond to the multiple levels of cortical processing in the Helmholtz/ Bayes account. 23 Dream experience, unlike waking experience, seems constrained mainly by the arousal of memory, and this may prove very important for the understanding of dreams, symptoms, hallucinations and mental disorder generally. This is discussed in my “Extending the Royal Road: from Dreams to Neuroscience,” in Philosophy, Science, and Psychoanalysis, eds. S. Boag, L. Brakel and V. Talvitie (London: Karnac, 2015). 24 Treatise on Physiological Optics, vol. III, 29–30. This mode of conscious representation seems responsible for the claim often advanced in philosophical accounts of perception, that we directly perceive the objects we see. But then direct perception is consistent with (indeed requires) the claim to which it is often opposed, namely that experience represents what we see. 25 Frontiers in Psychology, vol. 3, no. 151 (2012), doi.org/10.3389/fpsyg.2012. 00151. 26 A. Clark, “Whatever Next? Predictive Brains, Situated Agents, and the Future of Cognitive Science,” Behavioral and Brain Sciences, no. 36 (2013): 181–253, https://doi.org/10.1017/S0140525X12000477. See also the further comments and replies in “Forethought as an Evolutionary Doorway to Emotions and Consciousness,” in Frontiers in Theoretical and Philosophical Psychology, www. frontiersin.org/Theoretical_and_Philosophical_Psychology/researchtopics/ Forethought_as_an_evolutionary/1031.

144  Jim Hopkins 27 For Friston’s account, and some useful exposition, see “The History of the Future of the Bayesian Brain,” Neuroimage, vol. 62, no. 2 (2012). 28 James Garson, “Connectionism,” in The Stanford Encyclopedia of Philosophy (Winter 2012 Edition), ed. Edward N. Zalta, http://plato.stanford.edu/archives/ win2012/entries/connectionism/. 29 K. Doya, S. Ishi, A. Pouget, and R. Rao, Bayesian Brain: Probabilistic Approaches to Neural Coding (Cambridge, MA: MIT Press, 2007). 30 P. Dayan, G. Hinton, R. Neal, and R. Zemel, “The Helmholtz Machine,” Neural Computation vol. 7 (1995): 1022–37. 31 P. Dyan and G. Hinton, “Varieties of Helmholtz Machine,” Neural Networks, vol. 9, no. 8 (1996): 17. 32 As Friston describes this: “Hinton’s terribly important contribution was to cast the generally intractable problem of Bayesian inference in terms of optimization. The insight here was that the same problems that Richard Feynman (1972) had solved in statistical physics, using path integral formulations and variational calculus could be applied to the problem of Bayesian inference, namely, how to evaluate the evidence for a model. This is where free energy minimization comes in, the sense that minimizing free energy is equivalent to (approximately) maximizing the evidence for a model. Note again the underlying role of optimization, which here finessed a difficult but fundamental problem in Bayesian inference.” These issues are discussed in “The Helmholtz Machine” and “Varieties of Helmholtz Machine,” and in a more accessible treatment in chapter 10 of P. Dyan and L. Abbott, Theoretical Neuroscience (Cambridge, MA: MIT Press, 2001). 33 Friston has developed a number of technical formulations relating to these topics. See his website at www.fil.ion.ucl.ac.uk/~karl/. 34 This is discussed in a number of papers on Friston’s website noted above. 35 In Friston’s account emotions and other subcortical sources of internal input that are not currently being modelled remain in a state of Bayesian suppression, and so as sources of prediction error seeking higher-level (and ultimately conscious) representation. In this, Friston’s account of free energy is Freud’s, as described in R. L. Carhart-Harris and K. J. Friston, “Free Energy and Freud: An Update,” in From the Couch to the Lab. Trends in Psychodynamic Neuroscience, eds. A. Fotopoulou, D. Pfaff and M. A. Conway (New York: Oxford University Press, 2012), 219–29. See also my “Psychoanalysis Representation and Neuroscience: the Freudian Unconscious and the Bayesian Brain,” in the same book (230–65). 36 J. Hohwy, A. Roepstorff, and K. Friston, “Predictive Coding Explains Binocular Rivalry,” Cognition, vol. 108, no. 3 (2008): 687–701. 37 On this general account, see R. Adams, S. Shipp, and K. Friston, “Predictions Not Commands: Active Inference in the Motor System,” Brain Structure and Function, vol. 218, no. 3 (2003): 611–43. For an application of sensorimotor prediction to speech, see G. Hickok, “Computational Neuroanatomy of Speech Production,” Nature Reviews Neuroscience, no. 13 (2012): 135–45. 38 Gottlob Frege, “On the Scientific Justification of a Conceptual Notation” [1879], Conceptual Notation and Related Articles, ed. T. W. Bynum (Oxford: Clarendon Press, 1972), 83–4. 39 Helmholtz, Treatise, 19. 40 L. Boltzmann, “Model,” in Encyclopaedia Britannica, 10th ed. (London: The Times Printing House, 1903), vol. 30, 788–91. 41 On the transition from Helmholtz to Wittgenstein, see L. Patton, “Signs, Toy Models, and the a Priori: From Helmholtz to Wittgenstein,” Studies in History and Philosophy of Science, vol. 40 (2009): 281–9; M. Heidelberger, “From Helmholtz Philosophy of Science to Hertz’s Picture-Theory,” in Hertz: Classical Physicist, Modern Philosopher, ed. D. Baird, R. I. G. Hughes and A. Nordmann (Dordrecht: Kluwer, 1998); P. Barker, “Hertz and Wittgenstein,” Studies

Ways of Meaning in the Bayesian Brain  145 in History and Philosophy of Science, vol. 1, no. 3 (1980): 243–56; A. Wilson, “Hertz, Boltzmann, and Wittgenstein Reconsidered,” Studies in History and Philosophy of Science, vol. 20, no. 2 (1989): 245–63, and H. Visser, “Boltzmann and Wittgenstein or How Pictures Became Linguistic,” Synthese, vol. 119 (1999): 135–56. 42 Helmholtz, Treatise, 23. 43 Helmholtz, “Recent Progress in the Theory of Vision,” 186. 44 H. Hertz, The Principles of Mechanics [1894], ed. R. S. Cohen, trans. D. E. Jones and J. T. Walley (New York: Dover, 1956), 1. 45 On the role of Boltzmann, see particularly Wilson, “Hertz, Boltzmann, and Wittgenstein Reconsidered,” and Visser, “Boltzmann and Wittgenstein or How Pictures Became Linguistic.” 46 Boltzmann, “Lectures on the Principles of Mechanics” [1897], in Theoreti cal Physics and Philosophical Problems: Selected Writings, ed. B. McGuinness (Dordrecht: Reidel, 1974), 225. 47 Our Knowledge of the External World as a Field for Scientific Method in Philosophy (London: Open Court, 1914), 68, 71–2, 90 and 89. 48 The quoted phrases and sentences are from Frege “The Thought: A Logical Inquiry” [1917], Mind, vol. 65, no. 59 (1956): from pages 299, 306, 304, and 299. 49 Helmholtz, “Recent Progress in the Theory of Vision,” 222. 50 Michael Huemer, “Sense-Data,” in The Stanford Encyclopedia of Philosophy (Spring 2011 Edition), ed. Edward N. Zalta, http://plato.stanford.edu/archives/ spr2011/entries/sense-data/. 51 On this, see my “Rules, Privacy and Physicalism,” Wittgenstein and the Philosophy of Mind, eds. J. Ellis and D. Guevara (Oxford: Oxford University Press, 2012), 107–44. I have argued elsewhere that the “picture” of consciousness Wittgenstein rejected is a product of evolution, precipitated by the brain’s use of metaphorical processing in representing the mind as a kind of container within the body, and one that misrepresents the genuine physical internality, subjectivity and privacy enjoyed by conscious processes as they are realized in the brain. See my “The Problem of Consciousness and the Innerness of the Mind,” in Perspectives on Perception, eds. M. M. McCabe and M. Textor (Frankfurt: Ontos, 2007), 19–46. 52 Noam Chomsky, Aspects of the Theory of Syntax (Cambridge, MA: MIT Press, 1965), 57–8. 53 Quotations from W. V. O. Quine, “Epistemology Naturalized,” Ontological Relativity and Other Essays (New York: Columbia University Press, 1969), 82–83, and Theories and Things (Cambridge, MA: Harvard University Press, 1981), 1. 54 Donald Davidson, “Empirical Content,” in Subjective, Intersubjective, Objective (Oxford: Oxford University Press, 2001), 175. 55 See Davidson “Epistemology Externalized.” 56 See Davidson, “Could There Be a Science of Rationality?” in Problems of Rationality (Oxford: Oxford University Press, 2004) 124–5. 57 A. G. Huth, S. Nishimoto, A. Vu, and J. Gallant, “A Continuous Semantic Space Describes the Representation of Thousands of Object and Action Categories Across the Human Brain,” Neuron, vol. 76, no. 6 (2012): 1210–24, http:// dx.doi.org/10.1016/j.neuron.2012.10.014. This url also links to a video about the research, and the category-to-cortex mappings themselves can be accessed via color coding at http://gallantlab.org/semanticmovies. 58 William Talbott, “Bayesian Epistemology,” in The Stanford Encyclopedia of Philosophy (Summer 2011 Edition), ed. Edward N. Zalta, http://plato.stanford. edu/archives/sum2011/entries/epistemology-bayesian/. Many standard arguments about Bayesian epistemology are altered by holding (as here) that conscious beliefs are formed by subpersonal abductive reasoning, the parameters

146  Jim Hopkins governing which have been set by evolution. This may leave intact the idea that decisions will register preferences, but only as adjusted for non-rational factors such as impingements or error signals from emotions not consciously modeled; but not that cognition or the evaluation of evidence will be rational at the personal level. 59 For some interesting recent discussions of illusions about the body and self in Bayesian terms, see J. Hohwy and B. Paton, “Explaining Away the Body: Experiences of Supernaturally Caused Touch and Touch on Non-Hand Objects within the Rubber Hand Illusion,” Plos One, vol. 5, no. 2 (2010): e9416, https://doi.org/10.1371/journal.pone.0009416, and M. Apps and M. Tsakiris, “The Free-Energy Self: A Predictive Coding Account of Self-Recognition,” Neuroscience and Biobehavioral Reviews, vol. 41 (2014): 85–97, doi: 10.1016/j. neubiorev.2013.01.029. 60 Some of these are discussed in evolutionary and psychoanalytic terms in my “Emotion, Evolution and Conflict,” in Psychoanalytic Knowledge and the Nature of Mind, eds. M. Cheung Chung and C. Feltham (London: Palgrave Macmillan, 2003), 132–56, and “Conscience and Conflict: Darwin, Freud, and the Origins of Human Aggression,” in Emotion, Evolution, and Rationality, eds. D. Evans and P. Cruse (Oxford: Oxford University Press, 2004), 225–48. 61 Davidson, “Could There Be a Science of Rationality?,” 127. 62 Donald Davidson, “A Nice Derangement of Epitaphs,” in The Essential Davidson, eds. E. Lepore and K. Ludwig (Oxford: Oxford University Press, 2006), 256. 63 Davidson, “A Nice Derangement of Epitaphs,” 256. 64 Gareth Evans, “Semantic Theory and Tacit Knowledge,” in Wittgenstein: To Follow a Rule, eds. S. H. Holtzman and C. M. Leich (London: Routledge, 1981), 118–37. 65 Mark Platts, Ways of Meaning: An Introduction to Philosophy of Language, 2nd ed. (Cambridge, MA: MIT Press, 1997), 59 and 62. 66 “Technically rather byzantine” is from Davidson, “Could There Be a Science of Rationality?”; “closer to psychological reality” from Davidson, “Reply to Jim Hopkins,” The Philosophy of Donald Davidson, 286. The method for bringing closer is described in my text “Wittgenstein, Davidson, and Radical Interpretation,” in the same volume, 276–80. 67 The quoted phrases are from Davidson, “A Nice Derangement of Epitaphs,” 256 and 264. The argument about matching interpretation with speakers’ desires is continued in Davidson, “The Social Aspect of Language,” in The Philosophy of Michael Dummett, eds. B. McGuinnes and G. Oliveri (Dordrecht: Kluwer, 1994), 1–16. 68 For an account of this program as it has evolved since Ways of Meaning, see E. Lepore and K. Ludwig (eds.), Donald Davidson’s Truth-theoretic Semantics (New York: Oxford University Press, 2007). For an extension into Chomskian linguistics and cognitive science, see R. Larson and G. Segal, Knowledge of Meaning: An Introduction to Semantic Theory (Cambridge, MA: MIT Press, 1995). 69 See, for example, J. Barba, “Formal Semantics in the Age of Pragmatics,” Linguistics and Philosophy, vol. 30, no. 6 (2007): 637–68, https://doi.org/10.1007/ s10988-008-9031-4, and E. Borg, Pursuing Meaning (Oxford: Oxford University Press, 2012). 70 There is neuroscientific support for accounts of this kind in the case of irony. See N. Sportono, E. Koun, J. Prado, J.-B. Van der Henst, and I. A. Noveck, “Neural Evidence that Utterance-Processing Entails Mentalizing. The Case of Irony,” Neuroimage, vol. 63, no. 1 (2009): 25–39. 71 See C. L. Baker, R. Saxe, and J. B. Tenenbaum, “Action Understanding as Inverse Planning,” Cognition, vol. 113, no. 3 (2009): 329–49, doi:10.1016/j. cognition.2009.07.005.

Ways of Meaning in the Bayesian Brain  147 72 See K. Friston, J. Mattout, and J. Kilner, “Action Understanding and Active Inference,” Biological Cybernetics, vol. 104 (2011): 137–60, doi: 10.1007/ s00422–011–0424-z, and the previous articles cited there. 73 On this, see J. A. Shapiro, Evolution: A View from the 21st Century (Upper Saddle River: FT Press Science, 2011), who describes cellular working and intercellular communication in representational terms drawn from symbolic information processing; and the specifically Millikanian but far-reaching argument in N. Shea, “New Thinking, Innateness and Inherited Representation,” Philosophical Transactions of the Royal Society B, vol. 367, no. 1599 (2012): 2234–44, and “Representation in the Genome and in Other Inheritance Systems,” Biology and Philosophy, vol. 22 (2007): 313–31. 74 This is described in interesting detail at An Overview of E. coli Chemotaxis, available at the web site of the Biology Department of the University of Utah: http://chemotaxis.biology.utah.edu/Parkinson_Lab/projects/ecolichemotaxis/ ecolichemotaxis.html. 75 For this, see Shapiro, Evolution: A View from the 21st Century, 10.

10 Reflections and Replies Mark Platts

First things first: I want to express my enormous gratitude to all those who have contributed to this volume: friends, colleagues and former students whose theses I supervised, some all three. The generosity with time and effort involved, for example, in coming far to participate in the homage which gave rise to this book and to join there with others who shared their uncharacteristic willingness to shed their usual concern for truth when talking on that occasion of me as philosopher, teacher of philosophy or even just person, is something that seems to me quite extraordinary. Any compliments then paid I could surely return at least threefold, but I think I know those involved well enough to be able to say that they expect better, or worse, from me here: namely, my reactions free of dissimulation to what they have written. So here goes . . .

Barry Stroud on the Meaning and Transcendence of Values

In an age in which the academic paradigm is apparently the PowerPoint presentation, Barry Stroud reminds us of another way of going about things: be it in lectures, articles or books, a notable sensitivity to nuance in the expression of a philosophical position, extreme care and patience in the presentation and untangling of philosophical arguments, and a quite extraordinary feeling for both the philosophically problematic and the problematic in philosophy itself—all in splendid isolation from the whims of fashion. Much the same consideration, in at least two senses of the expression, exhibited elsewhere when working on the views of philosophers as diverse—to take but the perhaps most obvious examples—as Hume (of course!), Wittgenstein, Quine and Davidson is brought to bear here upon the efforts of an indefinitely slighter figure. I especially admire the way in which Barry’s paper here draws out and emphasizes the continuities between one approach grounded in the philosophy of language to issues of value and another grounded in what I called “philosophical psychology” focusing upon the notions of reasons for action (including of course reasons for speech acts), of desires and of evaluations; at least in my own case, the one approach led to the other. I am also heartened by the fact that we seem to be largely in agreement upon the concluding point of my discussion of the theory of value in Moral Realities, which I am happy to follow Barry in calling here that of the “transcendence” of values: namely, in his words, that “the evaluative judgments we do or would make can be “cognitive responses” to some value-properties we think an object possesses independently of anyone’s either making or being disposed to make evaluative judgments about it.” Relative to that agreement most other differences between our views are I think incidental, nor do I either need or wish to dispute here Barry’s attempt to say more than I have on the question of how speakers’ and agents’ knowledge and understanding are to be explained in a way compatible with that point. There is, however, one interesting, albeit perhaps minor, point on which I think we differ, though to reach that I should first try to clear the ground of some possible misunderstandings of my earlier argumentative strategies.

150  Mark Platts Firstly, I did not wish to claim or even to suggest that the subjectivist’s thought that “actual human valuings and desirings are the source of all value” is correct or even clear enough for its compatibilities and incompatibilities to be confidently determined; still, as Strawson might have it, “though darkling, one has some inkling—some notion of what sort of thing is being talked about”1 and I thought that enough to see how the subjectivist’s thought is false as a description of our actual thinking on, say, moral values in view of the thought’s neglect of the element of “transcendence” or “necessity” within that thinking. Secondly, my acceptance of John McDowell’s account of secondary qualities such as colours2 was occasioned, not by any great conviction as to the truth of the positive claim within that account to the effect that secondary qualities are properties the ascription of which to objects is true, when it is true, in virtue of objects’ dispositions to present certain sorts of perceptual appearances to subjects, but rather by the conviction of the falsity of the quite distinct claim, rejected too in McDowell’s account, that secondary qualities are not qualities “in objects,” are not qualities that exist “outside of the mind,” their “source” is exclusively within our minds: my point then and there being that even conceding the whole of McDowell’s account, any reasonable way of construing value-properties by analogy with (such) secondary qualities remains false to our actual understanding of our thoughts of value—the element of transcendence (or “necessity”) is still neglected. Though, thirdly, if I ever thought that any of that was enough to render pellucid my initial inkling as to the content of the subjectivist’s thought regarding the only source of value I was certainly wrong. Now to approach the point of difference between Barry’s account and mine. In considering human desires Barry contents himself with a presumed distinction between, on the one hand, “so-called ‘motivated’ or ­‘reason-following’,” ones and “other so-called ‘unmotivated’ or ‘reason-­ producing’ ” ones on the other. Perhaps that is so only for the purposes of his paper here, but I think it possibly indicative of doubts about my presumably finer-grained taxonomy of desire in Moral Realities, not least because Barry’s distinction seems to disregard the very claim of Stephen Schiffer’s3 to the effect that his “reason-producing” desires are only a proper subset of Thomas Nagel’s “unmotivated” desires4 which had led me to distinguish two kinds of “unmotivated” desires. Firstly, there were Schiffer’s “reasonproducing” desires which, amongst other features, were held to “have an essential phenomenological character, being either disagreeable in feeling when not satisfied, agreeable or pleasurable in feeling upon satisfaction or both.”5 And, secondly, there was a kind of “unmotivated” desires characterized in (logically) negative terms and in particular through the claim that for such desires there is no possibility of desirability characterizations of their objects, of characterizations which bring to an agreed end questioning of why agents want them, of what they see in them, as opposed to characterizations that serve merely to communicate the objects of the desires.

Reflections and Replies  151 Now, in his recent book Engagement and Metaphysical Dissatisfaction: Modality and Value6 Barry at one point says this: The idea of feelings or passions as forces that we simply find ourselves with and that push us into action is encouraged by the traditional rhetoric of a contrast between “reason” and “passion” or “thought” and “feeling.” The gap can seem absolute if feelings are regarded as something like sensations with a distinctive felt character. [. . .] But [. . .] [m]any passions, feelings, and sentiments involve evaluative and other propositional attitudes. [. . .] Even to feel thirsty or to “feel like having a drink” is to have a positive evaluative attitude toward having a drink. Sometimes when you feel that way your throat might feel dry and at other times you might have a slight headache, but neither of those sensations is a feeling of thirst or a desire for a drink.7 But, however it was with Schiffer, if anyone were to think this to tell against the pertinent part of my characterization of “reason-producing” desires, and so perhaps even against their existence, they would have misunderstood my talk of the “phenomenological.” Already, in an earlier excursion into this terrain, I had warned against the assimilation of desires to feelings like pains,8 and I was later to offer sustained criticism of attempts to construe desires as felt impulses or forces that move us to act9 (I also said there explicitly that appetitive desires by no means exhaust the exemplars of “reasonproducing” desires).10 It is rather, I suspect, my attempt to characterize a second kind of “unmotivated” desires which gave rise to the interesting, albeit perhaps minor, difference between Barry’s views and mine: the kind of negatively characterized desires most inviting the thought that the agents concerned just want their objects, merely wanting them with no grounding in “evaluations.”11 Perhaps, that is, it was the belief in the possibility of such desires which led Barry to invoke in the mentioned book Warren Quinn’s striking example of someone with a strong but apparently unintelligible urge to turn on radios. The man is said to see nothing in turning on radios and not to favor or evaluate positively in any way their being on. [. . .] With that evaluative element missing, we cannot understand him as intentionally turning on the radio.12 Now, I have no need to deny the counsel always to search for the type of understanding proportioned only through achievement of some desirability characterization of the object of an agent’s desire and of the content of his intention when he acts upon it; but “always search for” does not imply “is always there to be found.” Nor need I protest too much if Barry or anyone else wishes to stipulate that talk of an agent’s “wanting” something or of

152  Mark Platts his acting “intentionally” accordingly is to be in place only when some such desirability characterization is at least potentially available; but the absence of that stipulation represents no threat to our identification of the objects of agents’ desires in the cases now under consideration if we have once achieved an at least quite general understanding of their communicative speech acts. Still, in the present context—though this may amount to a further difference between Barry and myself—I doubt that the prolongation of this endgame is greatly worth the candle. Much as I like the idea of those who lust after chastity or have capricious desires for other monkish virtues or even literally hunger and thirst after righteousness, at least in my diminishing circles such people are rarely to be met with.

Notes 1 P. F. Strawson, “Freedom and Resentment,” in Freedom and Resentment and Other Essays (London: Methuen, 1974), 1. 2 John McDowell, “Values and Secondary Qualities,” in Morality and Objectivity, ed. Ted Honderich (London: Routledge and Kegan Paul, 1985), 110–29. 3 Stephen Schiffer, “The Paradox of Desire,” American Philosophical Quarterly, vol. 13 (1976): 195–212. 4 Thomas Nagel, The Possibility of Altruism (Oxford: Oxford University Press, 1970), 29–32. 5 Mark Platts, Moral Realities (London and New York: Routledge, 1991), 61–2. 6 Barry Stroud, Engagement and Metaphysical Dissatisfaction: Modality and Value (Oxford: Oxford University Press, 2011). 7 Stroud, Engagement and Metaphysical Dissatisfaction: Modality and Value, 110. 8 Mark Platts, “Moral Reality and the End of Desire,” in Reference, Truth and Reality, ed. M. Platts (London and Boston: Routledge and Kegan Paul, 1980), 76. 9 In the first chapter of Moral Realities. 10 Platts, Moral Realities, 73. 11 Cf. Bernard Williams, Ethics and the Limits of Philosophy (London: Fontana/ Collins, 1985), 59 and n. 9 at 210. 12 Stroud, Engagement and Metaphysical Dissatisfaction: Modality and Value, 109; Warren Quinn’s example is in his “Putting Rationality in its Place,” in his Morality and Action (Cambridge: Cambridge University Press, 1993), 236 ff.

Ralph Walker on an Attempted Alliance Within Moral Philosophy

Employing characteristically good sense in his exploration here of some of the seeming possibilities and interconnections found within his partial mapping of contemporary moral philosophy and its historical roots, Ralph Walker avoids both the dangerous plurality of “isms” and the vacuous dropping of philosophers’ names at the same time as showing comparably good judgment as regards which issues are worth discussing—and which are not. To some it might seem, at least at first sight, as if his attempt to diminish the apparent distance between—please forgive the following words— Kant’s views in ethics and mine turns too much on Ralph’s willingness to set aside, in a way more fanatical card-carrying Kantians are likely to condemn as an impertinence, some of the perhaps less fortunate thoughts of Kant, including that of his ethics as “an ethic of rules,” his denial that “lying can ever be right in any circumstances” and his claim that there is “only a single categorical imperative,” as well as Kant’s own willingness to set aside at times both his “more human side” which precisely “does not picture the ideal human being as one who lacks ordinary desires and inclinations” together with his recognition “that knowing what is right in particular cases requires great sensitivity to their details.” But I think that would seriously underestimate the interest of what Ralph Walker attempts to do in his contribution here, not least because a side effect of his pulling it off would be the partial amelioration of the deplorably lingering situation within moral philosophy in which, as he puts it elsewhere when talking of moral philosophy in Oxford in his student days, “the disagreements [go] too deep for good discussion” and in which one group regards another’s views as “morally reprehensible.”1 Still, notwithstanding the suggestive nature of Ralph’s closing remarks as to the common ground between anything amounting to a sensible Kantian view of ethics and mine as regards what moral philosophy can reasonably be expected to try to do—in terms, that is, of attempting principally to exhibit the defective character of arguments purporting to establish there to be no (possibility of) objective truth within moral thought and discourse—I nonetheless suspect, unsurprisingly enough, that what Ralph actually achieves here falls short of his assumed aim. It would, however, require considerably more space than I can reasonably take here

154  Mark Platts to justify my obstinate attachment to certain continuing differences with the kind of Kantian view of ethics Ralph defends here, and so I shall have largely to limit myself on this occasion to pointing out those continuing differences: what I shall offer, that is, will largely be not so much a reasoned defense of my ethical views when they conflict with those Ralph maintains in his contribution here but rather a (further) mapping out of the points at which I think conflicts continue to arise. Mandeville is my chosen point of departure here, and not just because this will give me the opportunity to try to explain a little more clearly than I did at the end of Moral Realities just why I examined some of his opinions there. Mandeville was interested in what might be called “the use” of morality, not of course as an exercise designed to clarify strictly speaking semantic aspects of moral discourse, but rather as an investigation of its “pragmatic” aspects, understood in terms of just what moral discourse is used for and of quite how it is used for it. In this context I tried in my earlier discussion to explore quite how moral discourse might be expected to be used if once we put aside the perhaps misleading impression Mandeville sometimes seems to give that flattery is the be all and end all of this part of his story: to explore, that is, some of the different ways in which moral discourse might be used, some of the different kinds of “speech acts” we might expect to find within the employment of moral discourse by ordinary—viz. nonpolitical—people.2 In so doing I wanted to exemplify the point—though this should certainly have been made clearer—that any such account of how moral discourse is presumed to be used will be fully intelligible only within the context of certain more or less specific, albeit perhaps tacit, assumptions about what people are at least in general like. And while it is true that Mandeville’s own assumptions on that matter, to the effect that such human behavior as cannot be explained within the terms of the operations of the innate instinct of self-love can be explained within the terms of the operations of the distinct innate instinct of self-liking, doubtless illustrate Philippa Foot’s warning that “general theories about the springs of action are traps for philosophers,”3 the point I wanted to exemplify remains standing. Now, Ralph Walker apparently finds clear and clearly acceptable talk of morality’s “prescriptivity” and of its “imperative force.” It is true that he distances himself from Hare’s thought that moral judgments are “very strongly prescriptive indeed,” that “one cannot sincerely assent to a moral judgment without acting on it in the appropriate circumstances.” But at this point Magdalen, the Oxford College where I best came to know Ralph and to appreciate amongst other virtues his splendidly unsentimental and all the more genuine concern for students’ wellbeing, comes to play a dual role. Firstly, it was that delightful person (and frequently underrated philosopher) Patrick Gardiner who first, as far as I know, showed the need for the exact kind of distancing Ralph accepts;4 but secondly, it was I think I remember another Fellow of Magdalen Geoffrey Warnock (the person who first tried to teach me about Kant’s moral philosophy) who held that even with that

Reflections and Replies  155 distancing the “prescriptivity” thesis notably overlooks the considerable variety of kinds of speech-acts occurring within ordinary moral discourse at least if the thesis is to be understood in terms of commonplace talk of “prescribing.”5 And here I would just add that so understood the thesis carries with it a remarkably simplistic assumption as to what people are at least in general like, to the effect that at least in this area of their lives they are quintessentially “prescribers”: an assumption which has perhaps even less to be said for it than the assumption that they are essentially “proscribers” or even just “persuaders.” If however the thesis is instead to be understood as invoking some “technical” use of talk of “prescribing,” further difficulties seem to me to arise over and above the consequent obligation to explain the presumed use involved and to explain it in terms which respect the distinctions, which Ralph Walker seems at times to ride roughshod over, between the “evaluative,” the “prescriptive” and the “normative”— or “whatever”!—and which thereby incorporate the recognition that, for example, the “evaluative,” unlike the “prescriptive,” can coincide more or less unproblematically with the “descriptive” when all the expressions concerned are understood in terms of distinct uses of language and so not of distinct “kinds” of literal meaning. The most important of the (other) difficulties within this context is that of showing just how the use in question of “prescriptive”—now more or less a term of art—can render compatible the presumed objectivity of moral talk as so used with its presumed link with the motivation of agents.6 The expression “objective imperatives” seems merely to take for granted that this difficulty has been overcome; but talk of “imperatives,” “commands,” “laws,” “rules” or even just “promises” seems to function so differently in cases of self-“imposition,” for example, that its invocation here would seem to explain little or nothing. And until we are given the requisite account of the more or less “technical” use in question of talk of “prescribing,” we shall remain in the dark as to the content of the assumptions of the “prescriptivity” thesis as regards what people are at least in general like in this area of their lives (and so as to the plausibility or otherwise of those assumptions). My account of the matter was at least meant to be quite another (though I suspect Ralph’s efforts to minimize differences are kindly enough motivated, at least in part, by the desire to save me from myself). Perhaps to a certain extent under the influence of Austin’s wish that in aesthetics we “forget for a while about the beautiful and get down instead to the dainty and the dumpy,”7 my account tried to focus above all upon our use of fairly specific expressions such as, say, “inconsiderate,” “cruel,” “humiliating” and “snobbish,” and even upon our talk of someone’s being, for example, “in danger of drowning” or simply “about to drown.” At least some judgements involving the use of such expressions can express objective reasons for actions: can, while being straightforwardly assessable at least in principle as true or false, yet quite reasonably elicit, or simply make manifest the presence of, desires in agents; there is nothing here, at least not in theoretical

156  Mark Platts terms, any more problematic in a case in which the desire elicited or made manifest is that of helping another who is about to drown than in the case in which the desire in question is that of avoiding one’s own death by drowning when that risk has arisen. And just as lemming-like suicidal urges do not count against the possibility of the latter kind of case, so motivated or even just wanton hopes for another’s death fail to count against the possibility of the former kind. What I should, though, have made clearer was this: while in moral and most other kinds of cases acceptance of what might be called “a full and explicit desirability characterization” of his action can indeed be attributed to the agent concerned—a characterization with evaluative force and “content”—that attribution, just like the attribution of the corresponding desire itself, can be simply consequent upon his acting as he does; the agent’s motivation for so acting, his reason for action, can be fully expressed on occasion in terms apparently free of “evaluative content,” in terms like, say, “he was about to drown” or “he was in pain.” (Even talk like that of his having been in “danger” of drowning can introduce at times a species of misleading “evaluativism.”) I did indeed say, what Ralph quotes me as having said, that in the characterization of cases in which the appropriate (moral) actions are not even attempted talk of “irrationality” is “in general nothing but bluff”; but that is quite compatible with the thought that the idea of a reasonable response to, for example, the kinds of reasons just now mentioned is indeed indispensable for any more or less articulated understanding of our natural responses to things happening around us (and not just our natural “active” responses either). The absence of any such response can bring us into the area where “rhetoric”—construed in a non-pejorative way and in terms free of any supposedly clear-cut distinction between “evaluative” and “descriptive” expressions—is in place: perhaps in an attempt to heighten perceptual attentiveness to the particular case at hand—without necessarily any reliance upon some supposedly clear-cut distinction between “the moral” and “the non-moral”—or to prompt more vivid appreciation of some more general matter of value. (I think I remember Cora Diamond mentioning to Putnam and myself a story of Dickens’s in which the most victimized of the galley slaves on a boat explains why he risked his own life to save that of his greatest slave-driving tormentor in terms of his tormentor’s being “a fellow passenger to death”; or think if you will upon Virginia Woolf’s comment upon our “journey” that “perhaps—who knows?—we might talk by the way.”) Now, Ralph Walker at one point says this: Kant puts [his point. . .] in terms of a moral law rather than moral facts, but his moral law is both objective and prescriptive, independent in its own right, given to our consciousness as binding. [. . .] It exists independently of what we, or anyone else, feels or thinks about it; but it has the capacity to influence us, and all rational beings, by motivating us. [. . .]

Reflections and Replies  157 Because this law determines specific moral truths, e.g., that I should pay back what I owe, they inherit its prescriptive character. All emphases other than the first have been added by me so as to highlight further continuing, and I think important, differences between the kind of Kantian view of ethics Ralph wishes to defend in his contribution here and the kind of ethical view I should wish to defend: for, even though Ralph at one point seems to try to diminish the import of Kant’s calling the moral law “rational” by apparently taking it as little (or no) more than Kant’s insistence here upon the need to satisfy the conditions for the exercise of our conceptual capacities, and despite Ralph’s attempt to exploit the (surely correct) point that even a priori truths can be arrived at by experience, by inductive procedures, I take any position in ethics reasonably labelled “Kantian” to hold “the moral law” to be a condition of “rationality” potentially “given to our consciousness” in an a priori way, and to be a condition which once met potentially enables us to “determine” by pure reflection “specific moral truths.” But as far as I can see the arguments designed to illustrate such a position fail time and again, and while some might wish to say that such merely human failings cannot demonstrate the impossibility of sound arguments illustrating the Kantian position, I fear I must echo in relation to the hoped for possibility of such an argument a remark of Quine’s in another context: I’ll believe it when I see it. And while I might I suppose have registered my differences with Ralph on this point by saying that I think there is much less scope than he does for deployment in ethical questions of the powers of reason a priori to detect truths and to draw out their consequences, I take it to be clear enough that those differences are not primarily matters of degree.

Notes Ralph Walker, “Parting Reflexions,” Oxford Philosophy, vol. 3 (2011): 5. 1 2 Mark Platts, Moral Realities (London and New York: Routledge, 1991), esp. 201–3. 3 Philippa Foot, “Nietzsche: The Revaluation of Values,” in Virtues and Vices (Oxford: Blackwell, 1978), 94. 4 Patrick Gardiner, “On Assenting to a Moral Principle,” Proceedings of the Aristotelian Society (1954–5). 5 In G. J. Warnock, Contemporary Moral Philosophy (London-New York: Macmillan/St Martin’s Press, 1967). 6 See, e.g., Warnock, Contemporary Moral Philosophy, and Bernard Williams, Ethics and the Limits of Philosophy (London: Fontana/Collins, 1985), especially at 124–6. 7 J. L. Austin, “A Plea for Excuses,” in Philosophical Papers (Oxford: Clarendon Press, 1961), 131.

Gustavo Ortiz-Millán on the Direction of the Mind

Besides contributing the interesting and instructive paper I am about to discuss, Gustavo Ortiz Millán bore the lion’s share of the organization of the academic participation both in the Conference that gave rise to this book and in the book itself. His patience and unshowy efficiency when faced with the hopelessness in the use of more modern forms of communication shared by me with some of the participants was simply remarkable; I am most grateful to him. Gustavo offers us in his contribution here an examination of a collection— I am reluctant to call them “a family”—of ideas expressed in the terminology of “direction of fit.” Perhaps a little ironically, the size and at least in general notable quality of the writing of other philosophers he draws upon comes to represent, at least for me at this moment, something of a problem, since he considers and subjects to criticism a fair number of ideas I have no wish to defend and generally have never been drawn to: by way of examples, attempted “reductions” of intentions to belief-desire pairs, “correspondence” theories of truth, the supposed notion of “besires,” “functionalism” as a doctrine about our commonplace concepts of beliefs and desires and “purely descriptive” interpretations of the difference—or differences—between beliefs and desires in that same everyday sense. And so, churlish as it might seem of me, I shall leave other (living) philosophers to defend themselves if they so wish and shall restrict myself here to reflection upon some of the things Gustavo says about some of the things I have said. To begin with, a few related, perhaps seemingly minor or even unnecessary, clarificatory points. First, when offering a gloss upon talk of how “beliefs aim at the true” I said the following: “falsity is a decisive failing in a belief, and false beliefs should be discarded; beliefs should be changed to fit with the world, not vice versa.”1 But talk of a failing in a belief is not tantamount to that of that’s being the fault of the belief concerned any more than my saying that false beliefs should be discarded amounts to an invitation to such beliefs to end their own lives! And when offering a gloss upon talk of how “desires aim at realization” I said that “the world, crudely, should be changed to fit our desires, not vice versa” the little word “crudely” was there for a purpose: to take account amongst others of the possibility expressed,

Reflections and Replies  159 for instance, by Descartes in his Discourse on the Method of adopting the maxim “to try always to conquer myself rather than fortune and to alter my desires rather than change the order of the world”:2 a maxim which is no invitation to unrealized desires to commit suicide. Of course neither beliefs nor desires “have intentions,” only those with beliefs and desires do. That those with beliefs and desires are the proper subject of the matter can also be appreciated clearly enough—I am sorry if I am labouring the obvious to some—by a little reflection upon the expression “direction.” Think on the differences, for example, between talk of the cardinal “directions” north, south, east, and west and that of the “directions” of left, right, forwards, backwards, up, and down: perhaps helped, if necessary, by remembering the case of the contestant in a British quiz show who, when asked “What way does the profile of the Queen always face on a coin?,” replied “Is it north?”.3 And then think of the differences between both those ways of talking and those involved in talking, on the one hand, of the “directions” for getting from A to B and, on the other, of the directions for using a computer or a combine harvester: perhaps helped, if necessary, by remembering the title of Descartes’s early, uncompleted work Rules for the Direction of the Mind and by thinking too of how, were we so to wish, we might so to say separate the “directions” for believers as believers purporting to increase the likelihood of their coming by the “right” beliefs from the “directions” for desirers as desirers purporting to increase the likelihood of their coming by the “right” desires, even though we know of course that the believers are the desirers and vice versa. That such “directions” are—please note the phrase—directed at believers and desirers rather than at beliefs and desires themselves is surely just as evident as the fact that Francis Bacon’s worry concerning how “men should love lies . . . [not] for advantage, as with the merchant; but for the lie’s sake”4 has to do with the fault of men as liars—when fault is in place—rather than with any “fault” of the lies themselves. Now, although I’m unenthusiastic about Gustavo’s use of terminology, at this point I suppose it clear enough that the account I gave of belief was not “mostly descriptive” but, rather, “normative,” being supposed to be a consequence of taking seriously the “distinguishing feature” of beliefs that “truth and falsehood are a dimension of [their] evaluation,” indeed “the most significant element” of it: my “directions” or recommendations were not in play there. Nor am I much concerned by the claim that this way of construing “direction of fit” seems to bring with it commitment to some Aristotelian account of truth: quite the contrary in fact since I share Strawson’s fondness5 for E. M. Forster’s maxim “only connect” at both the philosophical level and the moral and personal ones. Though I might just add that if more is wanted upon “the relationship of beliefs and truth” far from worse could be done than to turn attention to the connected matter of the role of beliefs within reasons for action (without, however, presuming some bridge to have been built to a Humean account of the latter notion).

160  Mark Platts An occasion for greater worry arises, however, I think, in connection with Gustavo’s apparently holding that, just as all beliefs are to be governed by the norm of the true, so desires—all desires—are to be governed by the norm of the good. He explicitly recognizes that that would be “at odds with” what he calls “subjectivism” about desires, with the ideas that “good [is] whatever we desire” and that there is “no way in which desires [can] be wrong” since “there is no independent normative notion that guides the appropriateness of the objects of our desires.” Here I can be brief since within the ancient tradition concerning the nature of desires in which Gustavo places himself is also to be found Barry Stroud in virtue of his contribution to this volume and whose views I have already called into question. Putting the point at its simplest, I think both of them overlook the possibility of the kind of “subjective” desires I characterized as “just wanting,”6 of which Bacon’s “love . . . for the lie’s sake” might well prove to be an instance. (It was in fact the full recognition of the possibility of desires of that (logically speaking) negative kind which led me to abandon the attempt7 to equate desire with acceptance of some “full and explicit desirability characterization”8 of the object of desire.) This means of course that what Gustavo calls the “normativity” of desire is, so to say, less entrenched than that of belief. That consequence is, it seems to me, welcome enough if our subject of study is human life as it is;9 the worry now, however, is that in these cases of “just wanting” the nature of their “direction of fit” seems as deeply problematic as the very possibility of anything like a uniform, unequivocal account of desire itself once they are admitted. That is: correction of the mistake shared by Barry and Gustavo comes to represent, too, a perhaps more serious difficulty for the account of desire I gave and which Gustavo was indeed concerned to criticise.

Notes 1 Mark Platts, Ways of Meaning: An Introduction to a Philosophy of Language (London: Routledge and Kegan Paul, 1979), 256–7. 2 R. Descartes, Discourse on the Method, trans. Elizabeth S. Haldane and G.R.T Ross (Cambridge: Cambridge University Press, 1970), vol. I, 96–7 (the third maxim of part III). 3 Quoted in Private Eye 1274 (29 October–11 November, 2010), 13. 4 Francis Bacon, “Of Truth,” in Essays (London: Dent, 1906), app. 3–5, at p. 3 (an essay best remembered by many of us for its opening sentence—“What is truth?, said jesting Pilate, and would not stay for an answer”—later quoted in perhaps the most emblematic opening paragraph of a paper in mid-twentieth century English philosophy). 5 See his Skepticism and Naturalism: Some Varieties (London: Methuen, 1985), 22. 6 In Mark Platts, Moral Realities (Routledge, London and New York, 1991), 61–8. 7 In Mark Platts, “The Object of Desire,” Crítica, vol. 7 (1985): 3–27. 8 For this expression, see now my remarks upon Ralph Walker’s contribution to this volume. 9 Think, for example, of the implausibility of trying to model upon Mandeville’s remark about some of those who disregard the “normativity” of belief—“Some

Reflections and Replies  161 impose on the World, and would be thought to believe what they really don’t; but much the greater number impose upon themselves, not considering nor thoroughly apprehending what it is to believe”—another suggesting that those who disregard the “normativity” of desire don’t really know what it is to want something. (Mandeville’s remark is quoted in Moral Realities, 195.)

James Griffin on the Ethics of Equality

When Mario Molina was awarded the Nobel Prize for Chemistry in 1995, a colleague of his observed that that showed that nice guys can indeed win; I remember thinking much the same when learning that Jim Griffin had been elected to the White’s Professorship of Moral Philosophy (a chair, though, that for many of us will always be Austin’s just as Paul Snowdon’s Grote Professorship will always be Ayer’s). Instructively enough, Jim first worked on the notion of necessary truth and on the early philosophy of Wittgenstein; that goes some way to accounting for how it is that his subsequent work in ethics is characterized by a feeling, unusual in this area, for the philosophically important details of arguments. And that, together with his sheer good sense and decency, explains how it is that his books achieve the rare ­distinction— a distinction particularly rare, Strawson once remarked to me, among books on moral philosophy—of actually giving pleasure in the reading. On this occasion Jim Griffin is concerned with the issue of equality and in particular with the very idea that there is some “foundational principle of equality” to be uncovered by philosophers as they set about the “job” of producing “a proper, i.e., systematic, ethics,” a system of normative ethics modeled upon “the paradigm of system provided by the natural sciences,” a “Newtonized” ethics; that concern is developed in Jim’s contribution here largely within the context of the views defended within his well-known book on human rights.1 Since I have explained elsewhere2 some of my doubts concerning those views, I can be brief here in indicating my reasons for not taking over that context. Invoking the key notion grounding his theory of human rights, Jim tells us here, for example, that all human beings “have equally” what he calls “normative agency.” But that is simply false: many human beings do not “have” normative agency, some having irrecuperably lost it and others neither having had it in the past nor going to have it in the future. And more generally, if human beings—all human beings—are to be “equally” bearers of human rights, we need to embrace the point that human beings can be more, or less, than agents, their bodies more, or less, than instruments of agency, and their minds more, or less, than harbours for their practical deliberations; once embracing that we can be led to abandon an account of human rights cast just in terms of normative agency in

Reflections and Replies  163 favour of some more pluralistic or disjunctive account of those rights which recognizes the relevance for them of certain (other) basic human needs, of certain human interests, and even perhaps of certain human vulnerabilities or dependencies, together with ordinary human responses to them. In so doing the account is at least closer to one characteristic egalitarian thought aiming to capture our common humanity, even though within the context of that common humanity such rights control the harm otherwise caused by the inequalities of, amongst other things, misfortune. Now, even though he has earlier distinguished the point in a possible thought where equality appears as fact from that where it appears as obligation, Jim subsequently says this: Equality, however, is not in itself a substantive value; it is a state of the world and in itself neither good nor bad. It becomes relevant to morality only by having the right sort of connection to something else that is substantially valuable. I find that, however, just as question-begging about any possible (“substantive”) value of equality as Kant’s critical remark about Rousseau to the effect that “inequality among men” is a “rich source of much that is evil, but also of everything that is good.”3 The point here is not only that such views have paved the way for so many of the largely mythical claims4 about the supposedly beneficial consequences for all of us of inequality—consequences supposedly attainable by no other means—which have come to shape so much contemporary economic and political “thinking,” nor just that, in the light of Jim Griffin’s own repeated emphasis upon the need for ethics to take into account more or less general truths about what human beings are as far as we know like, those views are at least in great danger of blithely passing over matters of condescension, contempt, degradation and—above all—humiliation. The main point, rather, is that such views, apparently presuming there to be no way of selecting among possible “means” to some adopted “end,”5 simply disregard how moral claims can “arise from”6 the possession of certain characteristics, thereby presuming additionally that “whether a certain consideration is relevant to a moral issue is an evaluative question” to be answered if at all by committing oneself to a certain kind of moral principle or outlook. The idea that our common humanity, being “a state of the world and in itself neither good nor bad” and so able to become “relevant to morality only by having the right sort of connection to something else that is substantially valuable,” is thereby revealed to be “not in itself a substantive value” seems to me on a par with another, mentioned here in my comments upon Ralph Walker’s contribution, to the effect that, say, “he was about to drown” or “he was in pain” cannot fully express on occasion an agent’s (moral) motivation for acting. Still, there are reasons for suspecting Jim’s primary purpose in his contribution to this volume to be that of curbing certain pretensions commonly

164  Mark Platts found amongst philosophers as regards moral thought and practice. While recognizing there to be “a miscellany” of principles of equality, he is concerned to reject the view that there is a “foundational principle” of it, understood initially it seems in terms of a principle covering the whole of “the moral domain” and deemed perhaps to constitute “the moral point of view itself.” From that starting point Jim is led, I think, to reject any idea that moral thought and practice needs to be “systematized” or “organized” in terms of some “foundational” theory understood now as one which philosophy itself is peculiarly empowered to provide—that is its “job”—albeit under the influence of “the sort of paradigm . . . provided by the natural sciences,” as in the “Newtonizing” thought that just as “physicists break down mechanics into its elementary forces, we moral philosophers should break down ethics into its elementary principles, such as ‘make equal’.” I quite agree that there is little or nothing to be said in favour of, and a very great deal more to be said against, the very idea that moral thought and practice stands in need of some “systematic grounding theory” which philosophers are peculiarly equipped to come up with, let alone when that idea is complemented by the suggestion that “scientific models of system” should be “installed” in “normative ethics.” But I cannot see why an egalitarian concern has to take on board any, let alone all, of those (additional) ideas or suggestions, as opposed to resting content in the case of for instance a society which does not allow same-sex couples to marry with the thought that that society’s stance is “irrational” precisely because it disregards the equality—in more or less any terms—of such couples with different-sex couples. And, indeed, it is not altogether clear to me just who the egalitarian thinkers are whom Jim thinks to have taken on board the (additional) ideas or suggestions in question (other than perhaps T. M. Scanlon). Moreover, a far more general worry arises about Jim’s views, one at least strikingly similar to that expressed at greater length in my comments here upon the contributions of Juan Antonio Cruz and Rodolfo Vázquez: at least on my understanding of “criterion,” continuing insistence upon a supposed need for “agreed criteria for the existence of a human right” (or for “determining the content” of any such right or for “rationally resolving conflicts involving” such rights) or for giving a sufficiently “determinate sense” to the idea of equality so that it “can provide the services we need to have” is nothing more than a hangover from the presumption of a need for some “systematic grounding philosophical theory” of moral thought and practice, a presumption which, be it cast in “foundational” terms or not, I can, as I have said, see no good reason to embrace.

Notes 1 James Griffin, On Human Rights (Oxford: Oxford University Press, 2008). 2 Mark Platts, “The Languages of Rights and of Human Rights,” Philosophy, vol. 85 (2010): 319–40.

Reflections and Replies  165 3 Quoted in Ralf Dahrendorf, “On the Origin of Social Inequality,” in Philosophy, Politics and Society, eds. Peter Laslett and W. G. Runciman (Oxford: Basil Blackwell, 1962), 88 and again at 108–9. 4 For a splendidly incisive and accessible criticism of those largely mythical claims, see Joseph E. Stiglitz, The Price of Inequality (New York and London: Norton, 2012). 5 For more on this presumption, see Platts, “Whisky y pequeñas bondades,” Diánoia, vol. 38 (1992): 203–15. 6 This expression and the following quotation are from Bernard Williams, “The Idea of Equality,” in Laslett and Runciman, Philosophy, Politics and Society, 112 and 113, respectively; cp. my account of the (Humean) relation of “giving rise to” between mental states in Moral Realities (London and New York: Routledge, 1991), 119–33.

Juan Antonio Cruz Parcero, Larry Laudan and Rodolfo Vázquez on Candidates for Being Human Rights

That philosophical discussion seems, even more than other kinds, to be characterized by misunderstanding of what others have written or said is perhaps unsurprising enough given the difficulty of philosophy itself, but quite another matter is when such discussion is warped by the posturing desire to take the moral high ground through wilful misrepresentation of others’ views. I well remember, for example, when after I had begun my first (public) consideration of the subject of human rights by insisting upon the distinction between the claim that it would be desirable, even extremely desirable, that everybody enjoy something—say, good health—and the claim that everybody has a right to it, a colleague informed me that I was obviously against human rights! I am indeed opposed to calling gut reactions “moral opinions,” to blurring the distinction between “by accident” and “by mistake,” and to using “lion” as if it were coextensive with “feline,” but that (perhaps) fastidiousness on my part takes on a different tone when faced with sloppy usage of an expression like “human rights” precisely because I am convinced of the importance of the rights in question. I am therefore all the more appreciative of the care with which Juan Antonio Cruz and Rodolfo Vázquez examine my views upon those rights in their contributions here. It may well be they consider those views to be too narrowly analytical in focus, too exclusively directed towards “conceptual” problems and worries in Juan Antonio’s terms, thereby failing to address a considerable number of related matters (though I am sure they realize that that is in itself no criticism of the contents of the views so expressed—indeed, that the fact that some are left hungry for more can on occasion be, as one reviewer1 of my first book was kind enough to say, a tribute to the quality of what has already been served). And it is indeed clearly the case that I have said little or nothing in my publications on this subject about, for instance, the historical development of the use of “human rights” or about its relations to legal thought and to political philosophy or even to the (fortunately increasing) interest in matters of “global justice.” It is also true that detailed examples of how discussion and reasoning—moral, legal or political—might reasonably go on as regards disputed cases of possible human rights are somewhat

Reflections and Replies  167 thin on the ground in my writings on the subject. Still, this second omission might perhaps point to a seemingly slight, maybe just seemingly terminological, difference between my views and those I suspect to be shared by Juan Antonio and Rodolfo as to quite what can reasonably be expected of a philosopher here—and not just because of my conviction that there is philosophical virtue in knowing when not to try to say any more on a subject—or even perhaps as to quite what can reasonably be expected of anyone else here. At one point in his contribution here Rodolfo Vázquez speculates as to why, within the terms of a recent constitutional reform in Mexico, the list of rights specified does not explicitly include certain “social rights” like “health, education, social security, alimentation,” rights which he holds to be as “universal” as those required by the satisfaction of the “basic needs” of human beings. Now, suppose within the context of some general commitment to the value of respect for autonomy the possibility is being contemplated of allowing a “mixed” social practice of the professional activities relating, on the one hand, to health, and, on the other, to education; and suppose too there to be good reasons for thinking in both cases that the existence of, so to say, a “private” sector—private hospitals, clinics, laboratories, medical practice, kindergartens, schools, universities and tutoring—would be significantly detrimental to the quality of the services provided within the corresponding “public” sector. Then, even assuming the commitment to the value of respect for autonomy to be constant across both spheres of professional activity concerned, be it in terms of the possible autonomy of those realizing their professional activities or of those receiving their services, and even disregarding any possible differences between those spheres as regards the probable consequences as regards the perpetuation or even augmentation of existing social inequalities within other areas of human concern if their corresponding private sectors be left unfettered, I think an important difference between them would remain: for it seems to me that even then much less damage to the quality of public services would be required in the case of health than would be required in the case of education for there to be a reasonable possibility of considering at least some degree of restriction upon the activities to be permitted in the private sphere. Now, perhaps Rodolfo would agree with me on that and was led only by brevity to pass over the suggested difference (a difference which perhaps exemplifies Juan Antonio’s thoughts at the end of his longer contribution here as to the considerations involved in the move from some abstractly characterized right to some more specific interpretation of it).2 But be that as it may, the difference I suspect there to be here between their views and mine is that I think them both to presume the need for some criterion serving to distinguish the case of health from that of education in the requisite way if the suggestion I have just now made is to prove defensible. (I fear Rodolfo fails to recognize the hollow tone with which I quoted James Griffin’s talk of the supposed threat that a notion be left “nearly criterionless.”)3

168  Mark Platts The kind of “morally-grounded dialogue” which I proposed to be necessary in the attempt at showing there to be “a sufficiently strong moral case” for special legal recognition and protection of some human interest is unlikely to involve with great frequency “criteria” as I understand that much abused expression (think of philosophers’ wielding of talk of “criteria of identity” or even of “criteria of personal identity”). More than being “a way of telling” when, for example, there is a genuine human right involved in the matter, that kind of dialogue seems to me most likely to be no more than “the only reasonable way of trying to find out” whether that is so or not—and is none the worse for that; in terms perhaps more congenial to Juan Antonio and Rodolfo, such a dialogue is much more likely to turn on matters of “principles” than of “rules,”4 on issues of evaluative counsel rather than those of practical instructions—and is none the worse for that. That of course relates to my continuing rejection of any conception of moral or ethical thought as “a calculus of action” (a rejection which once led a marvellous colleague, Ian McFetridge, to characterize my own view as “It doesn’t matter what you do so long as you feel bad about it afterwards”); it also relates to my conviction that, whatever else moral thought and practice might need, they do not need anything resembling “a philosophical theory.” I never cease to be amazed by the ease with which moral philosophers who would never think of themselves as being “Kantians,” indeed would even think of themselves as being “anti-Kantians,” nonetheless take on board what seems to me a clearly Kantian agenda, a clearly Kantian view of what moral philosophy should try to do, right down to the proffering of arguments supposed to influence any rational person concerned with moral problems and designed to favour certain “philosophical criteria” for solving them. And this is as good a moment as any to say that as far as I know I have never held or said anything that might suggest that the problems with talk about human rights arise from its now having been taken over by “common” or “ordinary” people or even (the mythical) “simple folk”5 (in contrast, I suppose, to “we sophisticated philosophical thinkers”). I hesitate to attribute the Kantian agenda to either Rodolfo or Juan Antonio, and it might well be that in what I have just now been saying I am guilty of a certain obstinacy in construing their talk of “criteria” as if it had some far more determinate sense than they meant it to have; if that is so, then there’s irony here since I think just the same is true of, for instance, Juan Antonio’s criticisms of things I have said about “legal” recognition and protection of some specific human interest. I also think he labours somewhat needlessly the question as to quite what my concern for—in Griffin’s terms’—the “strong inflationary pressures” upon the expression “human rights” is really a concern for. I may have been a little careless at times and I have no wish to speak for Griffin, but all I think is needed here are a few, simple enough, distinctions: it is one thing to hold that talk of human rights is now both more frequent and more widespread, another to hold claims positing the existence of distinct human rights have proliferated, yet another

Reflections and Replies  169 to hold a proliferation of those rights themselves to have occurred, but another to hold that the expression “human rights” is “indeterminate” (or “nearly criterionless”). (I’m afraid I can make nothing of talk of the concept of human rights having been “inflated” except that, like those of marriage and the family, its importance has been overrated.) And now my concern can surely be seen to have been simple enough: claims positing the existence of distinct human rights are at times no better than instances of “special pleading”—“unfair or one-sided argument aiming rather at victory than at truth”—and claims which at times, by disregarding the distinction between questions about rights and questions about mere desirabilities, manage to overlook not just issues arising from practical conflicts between putative, abstractly characterized, rights but also the obvious enough impossibility of guaranteeing and protecting the content of the supposed rights in any foreseeable circumstance (good health, a good night’s sleep, happiness, etc, etc.). (The word “mere” was of course present so as to make my distinction between questions about rights and questions about (mere) desirabilities exclusive; without it, the distinction would be like that between “lion” and “feline.”) And that concern of mine was, I take it, pretty much the very same one Juan Antonio himself expresses when he talks of the “distortion” of discourse about human rights! Though never having been formally taught philosophy of law, I tried in the distant past to pick up what I could, first and specially, from Hart—not just a hero of mine since my undergraduate days but quite simply the most wholly admirable man I have ever known—and from Ronald Dworkin, and far more recently from Rodolfo Vázquez and Juan Antonio Cruz; but most recently of all I was lucky enough, though for all too short a time, to have as a colleague in Mexico Larry Laudan, at that time moving away from working primarily within the philosophy of science towards the philosophy of law and even political philosophy. Even more striking than Larry’s capacity for hard work and deep, clear thought was his generosity with the results of his own labours (and in my case at least his judgement about how much of his learning he should try to share). Sadly, their having run up against some of the less lovely features of life in Mexico means that my friendship with him and Rachel must now largely be conducted from a distance (a common feature of life as time goes by, it seems), but at least I am aware of how much philosophy in Mexico missed out by their decision to return to the United States. In the contribution he was kind enough to send for this volume Larry offers a specimen of one of the kinds of things he does so well: calling into fierce questioning some sacred cow, in this case the supposed right to a presumption of innocence on the part of every criminal defendant (hereafter, following Larry’s own practice, “the PI”), which the Universal Declaration of Human Rights of 1948 proposed should become an internationally recognized right. He begins his questioning by showing how different jurisdictions and different constitutional courts have construed this PI “in vastly

170  Mark Platts different (and frequently mutually inconsistent) ways” as regards what it requires in terms of rules and procedures, and then goes on to consider the possibility that we are dealing here with what W. B. Gallie called “an essentially contested concept”;6 Larry holds that “the presumption of innocence comes into this category,” as I think he also thinks do the related ones of “the right to silence, the right to a demanding standard of proof, and the right to exclusion of illegally-seized evidence,” and says that if that is so then “we have no reason to expect convergence let alone closure on the definitive meaning of various key concepts of the rights of defendants.” But here a first worry arises. Larry’s version of Gallie’s account of what it is for a concept to be essentially contested is peculiarly truncated. Gallie himself gave seven characteristics of that: (i) the concept in question is “appraisive” of some given activity; (ii) the activity concerned is of an “internally complex character”; (iii) there are rival orderings in terms of importance of the component parts or features of the activity; (iv) the activity admits of considerable modification in the light of changing, and perhaps unforeseen, circumstances; (v) within disputes each party recognizes that its own use of the concept involved is contested by those of other parties, and each party has at least some appreciation of the “different criteria” in the light of which the other parties claim to be applying the concept in question; (vi) the concept derives from an “original exemplar” whose authority is acknowledged by all the contestant users of the concept; and (vii) it is plausible that the discussions about “the correct use” of the concept will enable the original exemplar’s achievement “to be sustained and/or developed in optimum fashion.” What Larry offers us is that “(i) . . . there are several rival (and mutually incompatible) conceptions meant to capture a given concept, (ii) . . . proponents of those rivals are all persuaded that their own version captures the legitimate meaning of the concept, and (iii) it is wholly unclear that there is any definitive argument that would move the advocates of rival conceptions to converge on any one of them.” That, however, seems simply to pass over what might be called the “tentatively conciliatory” conditions (iii), (iv) and (v) mentioned by Gallie, and to pass over too the possibility, which I have argued for elsewhere,7 that Gallie’s condition (vii) should be understood in terms which, while admitting of a reasonable possibility on occasion of agreement as to the comparative orderings in application of the concept in question to actual cases and to specific, imaginable possibilities, do not require any idea of “the one true conception” of the concept (and so activity) in question, let alone that of some “definitive argument” that would move (or “compel”) the advocates of differing conceptions to converge upon that “one true conception.” (This has to do, by the way, with my repeated emphasis8 upon the possibility that upon occasion all participants in a morally grounded dialogue about human rights may be found within one and the same person and with my consequent reluctance to follow Juan Antonio Cruz in his celebration of the possibilities for “disputes” therein.)

Reflections and Replies  171 Perhaps it is his awareness of those differences between Gallie’s account of “essentially contested concepts” and his own that leads Larry to try to show that the PI “is intrinsically incoherent.” His argument takes the form of considering various “renderings” of “the core meaning of the PI” so as to show, for example, that while one is false and another a tautology, yet another “enjoins jurors to do something that most humans cannot do, to wit, to believe a contingent empirical claim absent any evidence for it whatever.” Still, he shows a certain sympathy for a rendering of the PI along the following lines designed to make jurors going into a trial “understand that the trial is fundamentally a probative activity”: 3) A conviction requires that the defendant’s guilt be proved by the prosecutor beyond a reasonable doubt; any proof less than that requires a finding of not-guilty. Defendant does not have to prove anything. It seems, however, that the problem with that is that it is precisely targeted at “the trier(s) of fact” such as jurors whereas “the PI is widely supposed to have a compass that extends well beyond jury deliberations,” and so Larry is led to consider the possibility of supplementing it with 4) From the moment of arrest, defendant is to be treated as if he were innocent, enjoying the rights of autonomy of ordinary citizens not suspected of having committed a crime. But, perhaps thinking that to be on a par with, say, Prince Harry’s desire to be “treated like any other soldier” despite the fact that his presence in the front-line immediately puts those around him at greater risk and anyway results in his having, unlike any other soldier, body-guards on the battlefield, Larry can amuse us through the “treated as if”: Are we really supposed to believe that a judge who authorizes an arrest for probable cause still believes the defendant is innocent? Or that a prosecutor who brings a defendant to trial presumes his innocence? If that were true, he should not be bringing the person to trial at all. Do the police who enter the house of a suspect with a court order, seeking inculpatory evidence against the suspect, believe him to be innocent or at least treat him as if he were? Of course not. They seek a search warrant because they have reason to believe the suspect is guilty and seek the warrant in order to gain access to evidence that may demonstrate that guilt. Still, even in what I take to be Larry’s first, brief and exploratory, incursion into this terrain it would have been good to hear a little more on a few, obvious enough, points. First, at least if the PI is taken to be an “essentially contested concept” in Gallie’s sense, then there is nothing in the least bit surprising or disconcerting

172  Mark Platts in the fact that its “rendering” adequate for the cases of jurors or other “trier(s) of fact” may need modification for the cases of others involved in other parts of the investigative and legal processes, in contexts other than the trial itself—and certainly nothing sufficient to show that 3) itself is no more “free from incoherence” than other “renderings” of the PI are taken to be. This kind of “contextual modification” is on Gallie’s account more or less what is to be expected and represents no sort of pernicious “relativism,”9 any more than the modification in standards of proof as between criminal and civil proceedings does. Think, by analogy, of the differences as regards what is ethically required in terms of respect for confidentiality, not just within the differing professional activities of lawyers, doctors, journalists and priests, but even within the exercise of differing kinds of medical specialties and within the context of differing kinds of information and of differing kinds of individual or societal responses to such information; and while, doubtless, the law, unlike ethics, has good reasons for passing over at least some of those complexities, that leaves the main point here quite untouched. And even if, moreover, it were to prove impossible to come up with any adequate “renderings” for those involved in investigative and legal processes other than the triers of fact, that would not straightforwardly eliminate all interest and importance in 3) itself. But it is anyway far from clear that that does prove to be impossible if once contextual “sensitivity” is accorded its due place. Take for example the investigating police. There is a novel by Agatha Christie—Mrs McGinty’s Dead, I think—in which after a defendant’s conviction of murder the chief investigating officer seeks out Poirot’s help because, despite the procedurally flawless nature of the trial itself and despite the perfectly admissible way in which the evidence leading to the conviction was come by, the officer in question has come, for reasons which even if they had been admissible would not have weighed with the jury, to doubt the conviction. The officer believed, we might naturally say, that the evidence presented in court clearly pointed to or sufficiently “probabilified” the defendant’s guilt but nonetheless did not believe him to be guilty; and while the former kind of (“relational”) belief might be more than enough to prompt a search for further evidence of guilt that does not make it equivalent to a straightforward belief in that guilt itself (any more than the evidentially justified decision to bring the defendant to trial itself requires this latter kind of belief). (Talk of the police seeking a search warrant because they have reason to believe the “suspect” to be guilty seems to me ambiguous on this point.) And even if, long before evidence “demonstrating” guilt has come to light or even when investigating officers come to believe themselves to have it, they are straightforwardly convinced of the person’s guilt, what an appropriate “rendering” of the PI requires seems far less problematic than, say, the attempt “to believe a contingent empirical proposition absent any evidence for it whatever”: namely, as a first shot, in so far as it is humanly possible they should remain open to the possibility of error by remaining open to the possibility of, for instance,

Reflections and Replies  173 fresh, exculpatory evidence, and should in that sense remain ready to put aside their conviction as to the person’s guilt (aided perhaps by the thought that in another sense it is not for them to convict anyone). Finally, it is perhaps a little unclear just what would be consequent upon accepting Larry’s conclusion that there is no such thing as “the” presumption of innocence, that “in its stead is a confusing diversity of presumptions of innocence, no one of which is either compelling or free from incoherence.” Quite what rationale could then be given, for example, of “the right to a demanding standard of proof,” a right which Larry seems I think to recognize (and seems at least close to relating to at least one of the “diversity of presumptions of innocence,” 3))? Or is that supposed right, and others too, to be simply discarded as well?

Notes Steven E. Boër, in Linguistics and Philosophy, vol. 4 (1980): 155. 1 2 Cp. my editor’s introduction to La ética a través de su historia (Mexico City: UNAM, 1988), 10. 3 James Griffin, On Human Rights (Oxford: Oxford University Press, 2008), 14, n. 3. 4 See, e.g., the first chapter of Manuel Atienza and Juan Ruiz Manero, Las piezas del derecho (Barcelona: Ariel, 1996). 5 See P. F. Strawson, Skepticism and Naturalism: Some Varieties (London: Methuen, 1985), 56, and John Berger, A Fortunate Man (Cambridge: Granta, 1989), 110. 6 W. B. Gallie, “Essentially Contested Concepts,” Proceedings of the Aristotelian Society 56 (1955–6): 167–98. 7 In Mark Platts, Moral Realities (London and New York: Routledge, 1991), 159. 8 See, e.g., Mark Platts, “The Languages of Rights and of Human Rights,” Philosophy, vol. 85 (2010): 340. 9 I still think I was right to say, in Ways of Meaning (London: Routledge & Kegan Paul, 1979), 249, n. 3, that “different conceptions of one thing” implies realism, not relativism.

Paul F. Snowdon’s Prolegomena to Wittgenstein and Rule-Following

I gather that in his student days other Balliol students, typically scornful of the mere idea of attending tutorials, would nonetheless gather around when Bernard Williams was talking of classical Greek thought. University College students, at least in the late 1960s, were somewhat different in their attitude towards tutorials, not so much I fear for being “ever so ‘umble” as for appreciating their great fortune in having as their philosophy tutors Strawson, Mackie and McDowell; even so, I suspect I learnt most from my occasional conversations with Paul Snowdon (then just one year above me). Strawson himself has written of how were already manifest from the very beginning of Paul’s student days “the clarity and care, the balance and judiciousness which characterize all his work, together with a kind of majestic comprehensiveness which is rare at any age and in any context.”1 I would emphasize two of the expressions there. Firstly, “majestic,” since I remember Strawson remarking to me that Paul’s ability to exhibit the majesty of philosophy was unequalled by any other student Strawson had taught. (That “majesty” has some kinship, albeit distant, with Strawson’s term of greatest praise, “elegance”; Avishai Margalit recounts Strawson’s explaining that elegance in philosophy “involves both manner and matter, namely having good taste in choosing problems, exercising ingenious simplicity in tackling difficult problems, while presenting them in seamless graceful style.”)2 And, secondly, “judiciousness.” So many philosophers seem to share, at least in these “professional” times, a conception of philosophy in which the fate of at least nearly all philosophical views is to receive some fatal blow, be it in the shape of some inescapable counter-example or in that of a decisive argumentative refutation. But what usually happens is I think less dramatic and perhaps less pleasing to those enamoured of the endings of classic Western movies: namely, the gradual accumulation of difficulties which finally result in the comparably gradual abandonment of the views in question, at least by most philosophers at that particular time. In this way one key component in reasonable evaluation of a philosophical proposal is usually that very kind of judiciousness, understood in terms of good sense or judgement as to the weight or importance of considerations counting in favour of or against the proposal within the context of similar sense (or “taste”)

Reflections and Replies  175 as to which questions are worth pursuing, which continues to characterize Paul’s work. This will almost certainly disappoint those fixated upon a literally superficial notion of the dramatic or passionate “style” deemed to determine (their own) “presence” as well as those pining for “Enthusiasm” in Hume’s none too favourable sense;3 it might even disappoint those simply willing to throw the baby out with the bath water, as opposed to those duly appreciative of, for instance, Paul’s separation of rejection of an imagistic theory of understanding from rejection of all cognitive utility in the general capacity to imagine. But those are not Paul’s problems: from the outset he knew what he was trying to do and how it might best be achieved if it can be achieved at all. On this occasion Paul brings those virtues to bear upon the apparently much discussed matter of Wittgenstein’s examination in his Investigations4 of rule following. I say “apparently much discussed” because I share Paul’s suspicion that recent contributions have to a large extent moved away from Wittgenstein’s original examination, perhaps most notably when in those contributions arguments or theses are claimed to be present in Wittgenstein’s text when they clearly are not, and perhaps most deplorably when accompanied by the kind of hand-waving and ritual parrot-cries of “forms of life” or “language games” which rightly so outraged Bernard Williams. The return to Wittgenstein’s own discussion does not of course mean that the views of previous thinkers which substantially determine the strategic structure of that discussion should also be passed over. In particular, I continue to think it impossible to make much sense of at least the first 242 section of the Investigations except by seeing them as largely an examination of certain central ideas in Frege’s philosophy of language—and not just because of the “hitherto unknown kind of madness”—5and now think that by seeing them in that way Paul’s diagnosis that “Wittgenstein couples rejection of the myths with endorsement of a similar conception, or a conception that can be expressed in a similar way” is borne out: for far from rejecting the whole of that philosophy of language, I think that what Wittgenstein was perhaps principally concerned to do was to eradicate from it the kind of “psychologism” present within Frege’s own expression of it, maybe most notably in the presumption of some “basic psychological act of grasping a concept or a sense.” I also share Paul Snowdon’s lack of enthusiasm both for the ways in which Wittgenstein’s occasional remarks about “philosophy” are often bandied about and accorded some supposedly decisive interpretative weight within more recent contributions as well as for many of those remark themselves. But despite his careful separation of the view of Wittgenstein as being “first and foremost a negative philosopher” from that of his being essentially a “quietist” or “deflationary” one in the light of the point that the “fact that the extent of so-called ‘philosophy’ has no really discernible boundaries means that we have no idea what Wittgenstein thinks it is legitimate to challenge and what . . . not” when he says, for example, that philosophy

176  Mark Platts “leaves everything as it is,”6 my own preference, at least in seminars on the Investigations, has been to adjourn sine die consideration of Wittgenstein’s remarks about “philosophy” or at least to see how far we can get while leaving them aside: a preference explained in part by the fear that someone will trumpet a supposedly alternative “revolutionary” view of Wittgenstein and his philosophy. There are a couple of points of more detail concerning things which Paul perhaps surprisingly does not explicitly say to be commented upon, as well as a third, more general, matter. Firstly, after emphasizing the somewhat different natures of various particular cases of supposed rule-following which Wittgenstein at one point groups together, Paul goes on to ask quite what grasping a rule has to do with the particular case of understanding words and comes to the conclusion that if judgement cannot be fundamentally rule governed—as indeed it cannot—then “understanding words cannot be represented in terms of rule mastery.” The only, slight enough, puzzle here is why Paul does not mention as an instance of that rejected idea the thought that understanding a word is “fundamentally” knowing a verbal definition of it, which definition is a rule determining the word’s correct use: a thought tidily related of course to Augustine’s “picture of the essence of human language”7 and to be rejected for the same reason Wittgenstein gives in section 32 for rejecting that picture. Secondly, when remarking that the ability to control our images can be lost, Paul gives the example of “uncontrollable flashbacks suffered by people with post-traumatic stress” as well as that of “musical hallucinations.” But, at risk of exhibiting the abnormal character of my thoughts as to what is “normal,” I see no need for such, admittedly correct, exotic examples: having saved the cat from jumping out the fifteenth storey window, it seems to me completely normal or natural if one is haunted by hateful images of what would otherwise have been the outcome, nor do I think there anything unduly concerning about an individual’s “compulsively” continuing in his head the argument that has just driven him from another’s company. Finally, starting from a remark of Wittgenstein’s in section 154—“If there is to be anything ‘behind the utterance of the formula’ it is particular circumstances, which justify me in saying I can go on when the formula occurs to me”—Paul is led to coin the expression “circumstancialism” for a “sufficiently important” and “underappreciated” part of Wittgenstein’s thought. Bringing together a case in which a person pronounces a word correctly with that in which a person runs through a formula in his head as both being “performances” of kind X, the question is raised of whether the performance of X can be categorized as being an instance of the comparably dual kind G, which in the one case represents reading (with no requirement of understanding) and in the other properly understanding a sequence. Here Paul says this: Wittgenstein’s basic question is: what is it to be a G? [. . .] It seems to me that Wittgenstein is proposing that we think of both cases along the

Reflections and Replies  177 same lines. Namely, what makes X into a genuine G is a matter of the circumstances in which X occurred. And Paul goes on to argue “that Wittgenstein intends the notion of circumstances to figure in an understanding of what G-ing is and . . . is not intended to relate solely to the epistemology of judgements.” Now, Paul certainly does not need me to tell him that his discussion here of “circumstancialism” is crying out for more to be said. For the moment, just a couple of questions: Is that “circumstancialism” meant, for example, to be something other than the “contextualism”8 which is surely often enough at least talked about in connection with the Investigations? And, rather than being a matter of either what G-ing is or the epistemology of judgements about G-ing, might it not concern and be designed to illuminate that of utterer’s (or “occasion”) meaning when judgments about G-ing are expressed? Of course, the “alternatives” here are hardly clearly that, so—as so often with careful philosophy—progress may require a step backwards.

Notes 1 P. F. Strawson, “Reply to Paul Snowdon,” in The Philosophy of P.F. Strawson, ed. Lewis Edwin Hahn (Chicago and Lasalle, IL: Open Court, 1998), 311. 2 See Professor Margalit’s contribution to A Tribute to Sir Peter Strawson, eds. David Clary, Ralph Walker et al. (Oxford: Magdalen College, 2008), 6. 3 See David Hume, “Of Superstition and Enthusiasm,” in Essays: Moral, Political and Literary (Oxford: Oxford University Press, 1963), 76, where Hume identifies “the true sources of Enthusiasm” as “Hope, pride, presumption, a warm imagination, together with ignorance.” 4 Ludwig Wittgenstein, Philosophical Investigations, trans. G. E. M. Anscombe (Oxford: Basil Blackwell, 1963). 5 The phrase is of course Frege’s. See his Grundgesetze del Arithmetik, vol. I, in The Frege Reader, ed. Michael Beaney (Oxford: Blackwell, 1997), 203. 6 Wittgenstein, Philosophical Investigations, section 124 when talking of “the actual use of language” (though the point is immediately extended to cover “mathematics” too). 7 Wittgenstein, Philosophical Investigations, section 1. 8 A “contextualism” which seems to require for its utility some suitable idea of “relevance” (an expression which has received almost comically bad treatment at the hands of many linguists and philosophers).

Jim Hopkins on Theories of Meaning, Mind and the Brain

In his formidable and fascinating contribution here Jim Hopkins wisely puts aside at the outset the matter of fluctuations in philosophical fashion for his purpose of considering quite how the philosophical programme of which Ways of Meaning was a part might now be reasonably evaluated: though it has immediately to be stressed that the interest of Jim’s contribution to this volume goes way beyond anything suggested by his characteristically generous way of describing his purpose. That interest represents, however, a quite insuperable problem for me: to do anything like justice to the scope and subtlety of Jim’s discussion would require a reply of at least comparable length and detail. By placing the comparatively neglected figure of the nineteenth-century scientist and mathematician Hermann von Helmholtz1 within the Kantian tradition, Jim is able to pay special attention to the philosophical dimension of his work, and not just in terms of his influence upon Frege and Wittgenstein; rather, that attention, focusing most notably upon the ways in which Helmholtz tried to relate perception and movement, enables Jim to connect Helmholtz’s own version of Kantian synthesis to the idea of radical interpretation as deployed by Davidson (following Quine). And the point of doing this is to exhibit one major way in which “the tradition of analytical philosophy” was to come to be distorted, a way which the work of Quine and Davidson (and the later Wittgenstein) was subsequently to come to amend: Helmholtz had integrated Kantian epistemology with a scientific understanding of the nervous system by taking unconscious abductive inference to operate on the actual data collected by sensory systems, and in such a way as to yield conscious experience of the “world of things” that caused them. Yet [. . .] the tradition of analytical philosophy that stemmed from Frege was to forego this progress entirely. [. . .] [T]he basic notions to be used by the new “logical-analytical method in philosophy” [. . .] were those of logic on the one hand, and sense-data on the other. These sense-data were “what is actually given in sense” and so despite being constituents of “private spaces” or “private worlds,” were “the hardest of hard data” upon which empirical knowledge could be based.

Reflections and Replies  179 Russell and Frege’s picture of sense-data set the stage for decades of work in analytical philosophy, so passing by the earlier lessons Jim is concerned to emphasize: As Kant had argued, the actual data should be seen as prior to conscious experience, and not as given but rather as actively synthesized by the use of concepts. As Helmholtz had argued, the conscious experiences thus synthesized had the epistemic status (not of data but) of hypothetical abductive conclusions about the probable causes of the data. And so the correction of the distortion within the analytical tradition might now be seen like this: Davidson, like Quine before him, held there to be “a causal bridge that involves the sense organs” between “external events and ordinary beliefs,”2 it being an error to try to span this bridge with sense-data or uninterpreted givens.3 Within this perspective, patiently drawn out by Jim, and “fluctuations in philosophical fashion apart,” there is “no question of the death of the philosophical program of which Ways of Meaning is a part”: And if this program, and analytical philosophy with it, are increasingly integrated with neuroscience, this is surely all to our good. In this case the fate of our theories will be, as Einstein said, the fairest for which mortal theories can hope: to live on as part of those that succeed them. There is indeed far more to be learnt from Jim Hopkins’s contribution here than such a summary outline might suggest; but an essential part of the learning process being doubt and difference I shall now—tentatively—move into that mode. I finished writing Ways of Meaning in the Summer of 1977, but Davidson was to live—marvellously—a great many more years and continued working and developing his ideas right up until the very end. One possibility, of admittedly little moment, consequent upon this is that my own attitude towards some of the views expressed in Davidson’s later writings is somewhat more agnostic than Jim’s, an agnosticism echoing that perhaps detectable in my book as regards certain “Quinean” elements influential from the outset in Davidson’s theorizing: think, say, of Davidson’s attachment to his notion of “holding true”4 or of Quine’s view, apparently endorsed by Jim, that the “triggering [of our sensory receptors], first and last, is all that we have to go on”.5 Another consequent possibility seems to me, however, of far greater interest. In his attempt to relate “interpretation, radical and everyday, to the interactive abductive u ­ nderstanding—of one another’s desires, beliefs, and other motives—that underpins human group co-operation” Jim takes himself, I think, to being faithfully deploying Davidson’s later writings, but his attempt to connect themes in those writings with “Helmholtz-Bayes neuroscience” in order to show how “the explanatory framework of neuroscience suggests paths for progress in the

180  Mark Platts philosophy of mind and language” seems not just—unsurprisingly and unworryingly—to go well beyond anything to which Davidson had wished to commit himself but also to come at least close to distorting the tone or force of one of Davidson’s remarks. Jim represents Davidson as holding it to be “trivially true” that “if the theory [of an interpreter’s linguistic competence] does correctly describe the competence of an interpreter, some mechanism in the interpreter must correspond to the theory,” but what Davidson actually said is worth quoting I think at greater length: [C]laims about what would constitute a satisfactory theory are not [. . .] claims about the propositional knowledge of an interpreter, nor are they claims about the details of the inner workings of some part of the brain. They are rather claims about what must be said to give a satisfactory description of the competence of the interpreter. We cannot describe what an interpreter can do except by appeal to a recursive theory of a certain sort. It does not add anything to this thesis to say that if the theory does correctly describe the competence of an interpreter, some mechanism in the interpreter must correspond to the theory.6 Amongst other points, that seems to me far more dismissive than Jim allows of the idea that there is any philosophical interest in the (un)additional saying mentioned. The appearance of nits being peremptorily harvested here could be reduced, I hope, by reflection upon the relations between the preceding and a much more general worry concerning Jim Hopkins’s stance in his contribution to this volume—even though some might think it too general to be much worth considering. Jim quotes in an apparently approving way Helmholtz’s remarking that [T]he problems considered fundamental to all science were those of the theory of knowledge: What is true in our sense-perceptions and thoughts? And In what way do our ideas correspond to reality? Philosophy and the natural sciences attack these questions from the opposite directions, but they are the common problems of both. But that, surely both metaphorical and elusive, remarking leaves it quite unclear quite to what the hoped for “continuing collaboration between the disciplines” of natural science and philosophy is to amount, much as Jim’s thought of a future in which Davidson’s programme, and analytical philosophy more generally, is to be “increasingly integrated with neuroscience” seems, by leaving the nature of that “integration” indeterminate, silent upon the question as to the need for it. Certainly, the absence of any “scientistic” or “reductionist” component in, for example, John McDowell’s notably Kantian and Wittgensteinian treatment of the philosophical issues here in his Mind and World7 seems to do nothing to threaten its (considerable)

Reflections and Replies  181 interest. I am well aware, however, that in saying that I may be—mistakenly as well as ungratefully—inviting the thought that Jim himself believes some Helmholtzian “integration” to be in some way an obligatory aim of thought in this area. The “scientism” I’ve just now alluded to—which Strawson thought to have been more of a temptation for analytical philosophy in the US than in Britain—8is to be understood of course as a distortion and exaggeration of the scope and powers of what are now (as in Helmholtz’s remark quoted earlier) called “natural sciences,” and so is precisely not to be understood in terms of what Strawson, following Collingwood, held to be “the older and correct sense” of the term “science”: “a body of systematic or orderly thinking about a determinate subject-matter.”9 And finally, without wishing to anticipate future relations between work in linguistic science (in either sense) and that in the philosophy of language, I might just immodestly add that I still think it a virtue of Ways of Meaning that so much space was dedicated there to specific questions about the semantic structures to be found in certain natural language constructions. In using such constructions—­ competently or otherwise—it is doubtless true that “mechanisms” are in play in speakers’ brains; but the questions and answers about the semantic structures present in those constructions still seem to me to be properly a part of science only in “the older and correct sense” of that expression and so to be properly, if not quite securely, resistant to the sirenic allures of “scientism” or “reductionism.”

Notes 1 Comparatively neglected by philosophers that is: contrast, for example, R. L. Gregory, who describes von Helmholtz as “the greatest figure in the experimental study of vision” (Eye and Brain, [London: Weidenfeld and Nicolson, 1966], 90). 2 Donald Davidson, “Empirical Content,” in Subjective, Intersubjective, Objective (Oxford: Clarendon Press, 2001), 175. 3 Note that this in no way commits Davidson to the view, often attributed to him by Richard Rorty (see, e.g., the interview with Rorty in Theoria, vol. 1 (1993): 113–22, at 117), that this helps to liberate us from the notion of “experience”; as Davidson himself puts it, “ ‘Experience’ and ‘perception’ are perfectly good words for whatever it is that goes on in our minds when we look around us, smell, touch, hear, and taste” (op cit., p. xvi). And compare too Strawson on “a disorder that affects much, though by no means all of philosophical thinking today”; “an inclination to minimize or reduce or even, in extreme cases, to deny the reality of what I shall unashamedly refer to as inner or subjective experience” (“Reply to Hilary Putnam,” in The Philosophy of P.F. Strawson, ed. Lewis Edwin Hahn [Chicago and Lasalle, Ill.: Open Court, 1998], 292). 4 See Davidson, “Radical Interpretation” and “Belief and the Basis of Meaning,” in Inquiries into Truth and Interpretation (Oxford: Clarendon Press, 1984). 5 W.V.O. Quine, “Things and Their Place in Theories,” in Theories and Things (Cambridge, MA: Harvard University Press, 1981), 1, emphases mine. 6 Davidson, “A Nice Derangement of Epitaphs,” in The Essential Davidson, eds. E. Lepore and K. Ludwig (Oxford: Clarendon Press, 2006), 256. For some worries about just what on this account a theory of interpretation or meaning is supposed

182  Mark Platts to be doing, see my “What Should a Theory of Meaning Do?” Crítica, vol. 20 (1988): 43–67. 7 John McDowell, Mind and World (Cambridge, MA: Harvard University Press, 1994). 8 In an interview for the Sino-British Summer School in 1992, published in A Tribute to Sir Peter Strawson (Oxford: Magdalen College, 2008), 112. 9 Strawson, “Reply to Simon Blackburn,” in The Philosophy of P. F. Strawson, ed. Lewis Edwin Hahn, 172.

11 Closing Comments Philosophical Life* Mark Platts

Almost 40 years ago a student of mine in England was kind enough to give me a book he had come across in a second-hand bookshop in London, Norman Clark’s An Introduction to Kant’s Philosophy (London: Methuen, 1925). The gift was explained in large part by the book’s opening sentence: “It may seem strange that one perhaps more readily associated with the National Sporting Club [a club for gamblers, especially upon horse-races] than academic pursuits, should have written on Kantian philosophy.” Were I going to offer today a “philosophical autobiography,” or even just some more general autobiographical remarks, that might have been close to a good starting-point here too, but to the immense relief of all present I shall do no such thing. Aside from the “asquito”—(literally, a slightly nauseous feeling)—provoked in me by the very idea of my trying to do any such thing—you will surely appreciate how both my literary style and emotional life have been shaped by a recent Governor of Jalisco—along with my reluctance to follow in the footsteps of a young Carlos Monsiváis in seeking out so as to destroy all copies of an autobiography almost as precocious as that of Justin Bieber, I have always had—with one caveat—a marked preference for biographies rather than their auto-cousins. A competent biographer saves me at least from the often laborious task of reading between the lines, of trying to decypher quite what lurks within the caution or conceit (schoolmasters, observed Evelyn Waugh’s Dr. Fagan, must temper discretion with deceit). A bland and friendly example: is Strawson’s almost jarring remark in these politically correct times as to how his life was enriched by “the company of clever and beautiful women” best taken as a conscious echo of Hume’s, perhaps uncharacteristically, observing that in the context of the “particular pleasure” he “took . . . in the company of modest women” he had no reason to be “displeased” in the reception he “met with from them”? (Not that much labour is always necessary: understanding the former Knightsbridge Professor in Cambridge C. D. Broad’s albeit not so guarded remark in 1954 as to how he had “derived more happiness . . . than from any one other source” from his “power to make friends with the kind of young men” whom he liked and admired, “despite

184  Mark Platts great disparity in age,”1 requires nothing more than passing knowledge of the treatment then handed out to (male) homosexuals in England (think of Alan Turing).) Still, I mentioned a caveat to my general preference for the biographical, and it is an important one though not of much relevance here: when considering the life of a serious thinker, his or her thought must be adequately represented. That is not of course an impossible task for another to pull off—Nicola Lacey’s recent biographical study of Hart, doctor honoris causa of this University, at least comes close to it—any more than it is guaranteed that a thinker will himself or herself manage it (Popper comes to mind); the desideratum may nonetheless amount to a prima facie reason in favour of the auto-cousins in the cases of serious thinkers and I should have liked, if time had permitted, to consider in some detail two of the most interesting examples of this kind to have emerged from what for want of a better expression I shall call my little island (I allude to John Stuart Mill and R. G. Collingwood, both serious thinkers indeed, by contrast to those who can aspire at most to the category of underlabourers’ assistants’ mates, let alone those of us unable to see further even by standing on the shoulders of a giant who is standing on the shoulders of a giant who is standing. . .). What I shall dare to offer here are some brief, rather general, observations about the Contemporary Philosophical Scene; in doing so I shall draw heavily upon what I shall persist in calling “expert testimony.” In his engaging account of his life and opinions, Paul Grice said this: I have little doubt that it is the general beastliness of human nature which is in the main responsible for the fact that philosophy, despite its supposedly exalted nature, has exhibited some tendency to become yet another of the jungles in which human beings seem so much at home, with the result that beneath the cloak of enlightenment is hidden the dagger of diminution by disparagement.2 I would be one of the last to call into doubt “the beastliness of human nature” although my preferred way of registering it is in an echo of Broad’s bleak confession “I am not the kind of person whom I like”—as yet another (subsequent) Cambridge professor responded to a (then) Oxford colleague, we don’t have to be smug—but although Grice goes on to mention some far more specific matters which have contributed to philosophical activity’s present sorry state in most parts of the world, I think he fails to pay attention to one of the more important, well expressed by another Knightsbridge Professor (and one who followed in Grice’s footsteps to Berkeley), Bernard Williams, when reflecting upon the shape that the “professionalization” of the humanities in general is in danger of coming to take (if it has not yet already done so): “at the present time, the study of the humanities runs a risk of sliding from professional seriousness, through professionalization,

Philosophical Life  185 to a finally disenchanted careerism.” Philosophy returns to centre-stage, though, when Williams speculates upon “one of the reasons” why that might be so: If you do not really believe in the existence of truth, what is the passion for truthfulness a passion for? . . . If the passion for truthfulness is merely controlled and stilled without being satisfied, it will kill the activities it is supposed to support.3 Still, the utterly essential insistence upon truth’s being the central aim of the exercise—and so precisely not “careerism” so often only to be achieved by wielding “the dagger of diminution by disparagement”—is not I think sufficient much to reduce the (now probably realized) risk diagnosed by Williams; for that we shall at the very least have to find ways of taking seriously once more two (further) ideals. The first can be appreciated by reflection upon a remark once made by the great Mexican literary figure José Emilio Pacheco: “Poetry is a private art, not a show, a tournament or a horse-race; no poet can beat another by a nose or win in the final round.”4 Transposing Pacheco’s idea to philosophy, we might come to see again how much that is of value within a philosophical life is, albeit comparatively, yet essentially private, in the cases both of research and teaching (at least in my preferred forms of the latter, tutorials and direction of theses): scarcely any others’ eyes at all, let alone applause. (That is part of the reason why, I confess, I feel “asquito”—once more— when even in the more “public” of academic activities some participant turns around to make sure that (the rest of) the audience is truly savouring the brilliance of, say, his questions or (more likely) objections, and why too I was delighted to learn recently that the splendid British stage actor Sir Derek Jacobi loathes curtain-calls and applause). If some feel inclined to protest that that privacy has fortunately been undermined by current emphasis upon the formation of groups of researchers, I might, if the day be a sunny one, mention the arguments of one great philosopher—Strawson, of course—against the value of such formation in the particular case, given the particular nature, of philosophy, while on a day of more British weather I might mention the matter of what has actually happened in many places within philosophy as a result of this emphasis upon “group activity.” And here I should call upon Grice once more: (W)hen I visit an unfamiliar university and (as occasionally happens) I am introduced to, “Mr Puddle, our man in Political Philosophy” (or in “Nineteenth-century Continental Philosophy,” or “Aesthetics,” as the case may be), I am immediately confident that either Mr Puddle is being under-described and in consequence maligned, or else Mr Puddle is not really good at his stuff. Philosophy, like virtue, is entire.5

186  Mark Platts And that, I shall add, is part not just of the extraordinary difficulty of philosophy but of its majesty too, at least when it is well done. And then I might ask: is it mere chance that, as “group activity” in philosophy has increased almost without limit, so too even any awareness of the thought expressed by Grice, and made manifest of course throughout the history of the discipline (“This is Professor Kant, he works in ethics”), has been buried under the resulting fragments and rubble? (A quite distinct though not unknown problem is that raised by the fact that “group activity” can seem now a pre-requisite of interdisciplinary studies, doubtless perfectly desirable in themselves of course but which can give rise to the risk in the case of philosophical participation that, to take an obvious enough example, philosophy of science is little by little transformed into what are usually called “Science Studies”—doubtless most worthy of respect in themselves, but not I fear when masquerading as philosophy of science.) (I hope, incidentally, that that is enough to put paid to any idea that I think of myself as having ceased to be a philosopher of language so as to become instead a moral philosopher (or “our man in Political Philosophy”), even though a question might still arise as to why the main focus of my published work changed from one area of philosophy to another (or to others). Well, at first I was interested in exploring the possibility of extending truthconditional semantics so as to cover at least much of moral discourse since it seemed to me that moral philosophy in particular had suffered through philosophers’ reliance upon erroneous philosophies of language; but then a significant rôle came also to be played I think by my struggles to come to terms with the relations between our use of language and our understanding of it. Despite worries as to how much space was left for any substantial account of understanding once the at least generally unreflective character of our use of language was recognized, I became convinced that any veridical account of defensible moral talk had to be true to what I thought of as the “phenomenology” of such talk, and so true to the difficulty, for example, of practical decision in many cases in which salient and substantial moral considerations are in play. (And that despite the dangers of “psychologism” and of playing at being an amateur scientific psychologist.))† Respect for privacy does not in itself preclude all forms of competitiveness, of course, and some have ideas about them which seem to me to relate them in truly deplorable ways: I think, for example, of perhaps the most competitive of the great contemporary philosophers I have been lucky enough to try to learn from and who apparently thought of “his” students as his private property to be shielded from all alien influence. Strawson had it healthier again—he wrote that he had “no expectation of, or wish for, disciples”—and I remember with pleasure when Broad, in a magazine interview, expressed his amusement at the sight of his Cambridge colleagues “sycophantically dancing to the pipings of Wittgenstein’s flute” (I hope my memory has that right after 40 years or so). And I hope it unnecessary to add that nothing said here is designed to call into question the value of entering

Philosophical Life  187 the one competition that all—Plato, Aristotle, Descartes, Hume, Kant, Wittgenstein, the lot—are doomed to lose, the competition with the truth in philosophy, nor to qualify in any way my admiration for those—perhaps Stephen Jay Gould is the most obvious example in our times outside of philosophy, and Ayer within it—who managed to combine excellence in their “private” work with comparable distinction in the area of popularization. I mentioned the need to take seriously two ideals apart from truth if in the case of philosophy the risk, probably now realized, diagnosed by Williams is to be reduced. As many of you will remember, Wittgenstein’s Russian language teacher Fania Pascal tells of how, when recovering in a clinic after the removal of her tonsils, she was visited by him, only for him to respond, apparently disgusted, when she said, croaking, “I feel just like a dog that has been run over,” “You don’t know what a dog that has been run over feels like.” Though some have speculated that this was a joke designed to cheer the patient up, I prefer to take it at face value and to draw two morals from it. The first, which applies just as well even if it were indeed a joke on Wittgenstein’s part and which seems to me confirmed by more or less everything I have ever heard about him as a person, is that there is every reason to be grateful for never having known him. The second, again confirmed by more or less everything I know of him, is that he was so to say, a philosopher through and through (perhaps, in Grice’s terms, “entire”). By this I do not mean he was incapable of talking about anything other than philosophy: even in one of his genius that would most likely have been tiring and tiresome and might even have come close to disregard for the profound difficulty of even approximating to truth in philosophy. Rather, I mean that in Wittgenstein’s case, as I hope in those of far lesser mortals, there is need to employ some secularized notion of vocation, free not just of connotations of God’s calling some of us, of there due solemnity, but even of any suggestion that social workers such as probation officers are under consideration. In the case of philosophy, the pertinent notion might begin to be shaped in terms of the (internalized) kind of professional seriousness Williams touched upon in characterizing the first, essential stage in the study of humanities, a seriousness which might more tendentiously be related to the conviction, for example, as to the importance, and the profound difficulty, of even approximating to the truth in it (and perhaps too to the majesty of any such approximation), as well of course, though some may think sad to say, to the fitness or aptitude for the corresponding endeavour. (No favour is done to any student, friend or even institution by forgetting that last condition, notwithstanding the considerable prestige and public recognition that might accrue in time to some of those lacking such fitness or aptitude: hence in part, I suspect, Williams’s invocation of the “disenchanted.”) If some such ideal of philosophical vocation can still now be recovered, then use might also arise for another duly secularized notion, this one remotely akin to that of an Epiphany, available in characterizing the occasional discovery that a philosopher is what one is to be or at least is what

188  Mark Platts one is seriously to try to be, perhaps long before one has much idea of to what that would really amount, as well as some notion of pleasure or even fun—remember that the secular is all that is in play here—which can be an essential part of philosophical seriousness. (In ever so many cases you can detect good philosophers by and in their sense of humour, and not just as something which enables them to live with their “suffering” and “torment”: hence perhaps Wittgenstein’s remark about the possibility—which he himself perhaps mercifully did not attempt to realize—of a work of philosophy’s consisting entirely of jokes. But please remember: there is such a thing as joyless laughter.) In this way we might even come to understand the possibility—this one exemplified by some of those present today—of being a philosopher through and through without being a pain in the neck to others (be they philosophers or not). And we should anyway remember that the philosophically fraudulent, the charlatans and the “purely professional” tend to provoke more unexpectedly located pains. On that exalted note I am almost done, though I might just add that the most likely response to what I have said here today—apart from that of ignoring it altogether—is also the one that I suppose will most irritate me: the even lazier version of certain politicians’ already lazy rejoinder to any criticism at all of what they are doing: TINA, or, “There Is No Alternative”; and that, because there is, even if some thought and a little imagination are needed to uncover, or recover, it. And now it only remains to thank those involved in bringing this Conference about—most notably (as usual) Laura Manríquez and Gustavo Ortiz Millán—as well of course as the other friends, colleagues and former students who have sacrificed (or squandered) their time by participating during these days—especially those who have travelled far to be here (and who have over the years taught me far more about how to do philosophy than they realize, though far less than they would have done had I been less unapt). I should have preferred a far less formal setting to get together with Linda and José Luis, but you take what you can get. A thousand or two thanks to all of you for being here. So now, remembering the dislike I share with Derek Jacobi, I close with one final quotation from his favourite playwright (those who don’t already know its location will surely be helped out by José Luis): “The rest is silence.”

Notes * This text, here published with minor modifications and one addition, was read by Mark Platts in the closing session of the Conference in August of 2012 on which this book is based; both the archaeologist Linda Manzanilla and the theatre director José Luis Ibañez also participated in that closing session. 1 C.D. Broad, “Autobiography,” in The Philosophy of C.D. Broad, ed. Paul Arthur Schilpp (New York: Tudor, 1959), 68. 2 Paul Grice, “Reply to Richards,” in Philosophical Grounds of Rationality, eds. R. Grandy and R. Warner (Oxford: Clarendon Press, 1986), 63.

Philosophical Life  189 3 Bernard Williams, Truth and Truthfulness (Princeton and Oxford: Princeton University Press, 2002), 3. 4 “Lamenta José Emilio Pacheco que la sección de cultura sea el sótano de la de espectáculos,” La Jornada, June 22, 2008. 5 Grice, “Reply to Richards,” 64. † This paragraph has been added in 2017, at the insistence of the editors of this volume, to the text originally written and delivered in 2012.

Contributors

Juan Antonio Cruz Parcero is research professor at the Institute for Philosophical Research, National Autonomous University of Mexico. He is the author of El concepto de derecho subjetivo (1999) and El lenguaje de los derechos (2007). James Griffin is White’s Professor of Moral Philosophy Emeritus, Fellow of Corpus Christi College Emeritus, and Honorary Fellow of Keble College, in the University of Oxford. He is the author of Wittgenstein’s Logical Atomism (1964), Well-Being: Its Meaning, Measurement and Moral Importance (1986), Value Judgement: Improving Our Ethical Beliefs (1998), On Human Rights (2008) and What can Philosophy Contribute to Ethics? (2016). Jim Hopkins is reader emeritus in philosophy at King’s College, London. He is currently an Honorary Professor in the Psychoanalysis Unit of the Research Department of Clinical and Health Psychology at UCL. He was Kohut Visiting Professor of Social Thought at the University of Chicago in 2008. Larry Laudan teaches at the University of Texas at Austin. He is the author of several books on the philosophy of science and law. His most recent books are Truth, Error and Criminal Law: An Essay in Legal Epistemology (2006) and The Law’s Flaws: Rethinking Trials and Errors? (2016). Gustavo Ortiz-Millán is research professor at the Institute for Philosophical Research, National Autonomous University of Mexico. He works on moral psychology and applied ethics. He is the author of La moralidad del aborto (2009), and has also edited several books. Mark Platts is research professor at the Institute for Philosophical Research, National Autonomous University of Mexico (UNAM). He is the author of Ways of Meaning (1979), Moral Realities (1991), Sobre usos y abusos de la moral (1999), and Ser responsible (2012). He has edited many books on the philosophy of language and ethics. Paul F. Snowdon is Grote Professor of the Philosophy of Mind and Logic at University College London. His research covers metaphysics, especially

192 Contributors the nature of persons and selves, the philosophy of mind, especially perception, and also the history of twentieth century Anglo-American philosophy. He is the author of Persons, Animals, Ourselves (2014). Barry Stroud is the Willis S. and Marion Slusser Professor Emeritus of Philosophy at the University of California, Berkeley. He is the author of Hume (1977), The Significance of Philosophical Scepticism (1984), The Quest for Reality (1999), Engagement and Metaphysical Dissatisfaction (2013), and three volumes of collected essays. Rodolfo Vázquez is emeritus professor of law at the Autonomous Technological Institute of Mexico (ITAM). He is the author of several books, some of the most recent ones are: Las fronteras morales del derecho (2009), Consenso socialdemócrata y constitucionalismo (2012) and Derechos Humanos. Una lectura liberal igualitaria (2015). Ralph Walker is an Emeritus Fellow of Magdalen College, Oxford University. He has written extensively on the history of philosophy, Kant, epistemology and metaphysics. He is the author of several books, among them Kant (1978), The Coherence Theory of Truth (1988), and Kant and the Moral Law (1998).

Index

abductive inference 119 – 29, 132 – 7, 178 – 9 agency, normative 49 – 52, 55, 64 – 6,  162 Altham, J. E. J. 34 Anscombe, G. E. M. 27 – 8, 38, 42, 44 Aquinas, Saint Thomas 49 – 50 Aristotle 16 – 17, 19, 50, 187 Atienza, M. 173 Augustine, Saint 42, 49, 99, 176 austerity 60 – 1 Austin, J. L. 44, 155, 162 autonomy 53 – 6, 62, 65 – 6, 71, 78, 92, 167, 171 Ayer, A. J. 162, 187 Bacon, F. 159 – 60 Bayes’ Theorem 122, 127 – 8, 135 – 8,  179 Bentham, J. 55, 60, 71, 82, 93 Berger, J. 173 Berlin, I. 53 “besires” 34 – 6,  158 Bieber, J. 183 biosemantic representation 139 – 41 Blackburn, S. 13 – 14 Boër, S. E. viii, 173 Boltzmann, L. 130 – 1 Bratman, M. 27 Broad, C. D. 183 – 4, 186 Brock, D. 66 Carpizo, J. 66, 69 causal explanation 98, 121 – 2 Chomsky, N. 120, 133 Christie, A. 172 Clark, A. 127 Clark, N. 183 Cohen, J. 64

Collingwood, R. G. 181, 184 Cranston, M. 70, 75 criteria 64 – 6, 76, 97, 109, 164, 167 – 8,  170 Cruz Parcero, J. A. 164, 166 – 70 Dancy, J. 18, 33 Davidson, D. vii – viii, 1, 3, 98, 115 – 16, 121, 129, 133 – 8, 149, 178 – 80 Davidsonian Programme see truth-conditional semantics Dennett, D. 125 Descartes, R. 117, 132, 159, 187 desirability characterizations (full and explicit) ix, 4 – 5, 150 – 2, 156, 160 desires see reasons for actions and desires Diamond, C. 156 Dickens, C. 156, 174 direction of fit ix, 26 – 48, 158 – 60 discrimination 19, 20, 54 – 5, 62 – 4, 73 dreaming 127, 133 Dummett, M. vii – viii Dworkin, R. 80, 169 Einstein, A. 121, 138, 179 equality 49 – 57, 162 – 4; as ethical foundation 50 – 2, 53, 57, 162, 164; and human rights 49, 51 – 2, 54 – 5, 56 – 7, 162 – 4; and impartiality 50 – 2, 53 – 4, 55; and natural rights 49; principles of 50 – 3, 56 – 7 essentially contested concepts 90 – 1, 170 – 1 “evaluativism” 156, 163 Evans, G. vii – viii, 135 experience 20 – 1, 23, 97, 99, 112, 115, 117 – 19, 121, 123, 125 – 7, 129 – 34, 137 – 8, 157, 178 – 9

194 Index Foot, P. 154 Forster, E. M. 159 Francis of Assisi 49 Fraser, J. G. 99 freedom of the press 71, 76, 86 Frege, G. 99, 104, 116, 123, 130 – 3, 175, 178 – 9 Freud, S. 99, 117, 128 Gallie, W. B. 90, 170 – 2 Gardiner, P. L. 154 Gödel’s Theorem 23 Goldie, P. 45 Gould, S. J. 187 Gregory, A. 47 Grice, H. P. 98, 102, 184 – 7 Griffin, J. 59 – 61, 64 – 6, 68 – 70, 75 – 8, 162 – 4, 167 – 8 Habermas, J. 62 – 4 Hare, R. M. 14 – 15, 154 Harman, G. 45 Harry, Prince 171 Hart, H. L. A. 60, 74, 169, 184 von Helmholtz, H. 116 – 23, 125 – 33, 135, 137 – 8, 141, 178 – 81 Hertz, H. 131 HIV/AIDS x Hohfeld, W. N. 73 “holding true” 15, 179 – 80 Holmes, S. 82 Howhy, J., Roepstroff, A., and Friston, K. 128 human nature 154 – 6, 160, 163, 184 human rights: ballooning of 57, 68; “deflation” of 59 – 61, 65 – 7, 78 – 81; and feasibility 55, 72, 78 – 9, 169; “inflation” of 59 – 61, 66 – 7, 68 – 71, 77 – 8, 168 – 9; justification of 56 – 7, 60 – 2, 64 – 6, 77 – 8, 80 – 1, 162 – 4, 168; and needs 60, 64 – 7, 71, 76 – 8, 163, 167; social 60 – 1, 63 – 4, 70 – 1,  167 Humberstone, L. 33, 35, 37 Hume, D. ix, 6 – 9, 14, 29, 32, 41 – 2, 47, 119, 149, 175, 183, 187 humiliation 163 Ibañez, J. L. 188 images 97, 103, 105 – 6, 120 – 1, 125, 176 impartiality 50 – 5 interests, human 59, 64, 77 – 8, 163

irrationality 14, 22 – 3, 42, 55, 136, 138, 156 – 7, 164, 168 Jacobi, Sir D. 185, 188 Jacobson-Horowitz, H. 46 James, W. 23 Kant, Immanuel 13 – 24, 50 – 1, 116 – 19, 153 – 63, 183, 186 – 7; on consciousness and perceptual experience 15, 117, 122 – 3, 125, 132, 156 – 7, 179; on moral law 15 – 24, 156 – 7; on moral motivation 13 – 17, 23, 156; on moral rules 13, 18, 153; on moral worth 15 – 16; on synthesis 117 – 19, 121, 125 Kelsen, H. 71, 82 Kierkegaard, S. 23 Köhler, W. 99 Kripke, S. 98 Lacey, N. 184 Lafont, C. 62 Laitinen, A. 35 Laporta, F. 59 Laudan, L. 169 – 73 Laudan, R. 169 Lenoir, T. 116 – 17, 120, 143 Lewis, D. 46 liberalism see autonomy Ludlow, P. viii Mackie, J. L. 22, 174 Malcolm, N. 97 de Mandeville, B. 13, 16 – 17, 154, 160 – 1 Manríquez, L. 188 Manzanilla, L. 188 Margalit, A. 174 Martin, C. B. 98 McDowell, J. vii – viii, 3, 33, 46, 113, 136, 150, 174, 180 McFetridge, I. 168 McGinn, C. vii Mill, J. S. 184 Millikan, R. 130, 139 – 40 Milliken, J. 41 Mirandola, G. Pico della 49 misfortune 163 Molina, M. 162 Monsiváis, C.183 Moore, G. E. 99 moral intuition 4, 18 – 21

Index  195 moral motivation ix, 4, 9, 13 – 17, 33 – 4, 155, 163 moral motivation and action ix, 15 – 17, 33 – 4 moral phenomenology 13, 15, 186 moral realism viii – ix, 3, 13 – 14 moral rules 13, 18, 80, 153, 168 moral worth 15 – 16 motivation see reasons for actions and desires Nagel, T. 46, 150 natural rights 49 needs, human 59 – 60, 64 – 7, 71, 76 – 8, 163, 167 Newton, I. 55 – 7, 121 Nickel, J. 82 Nietzsche, F. 16 objectivity/subjectivity 13, 15 – 17, 19, 22 – 4, 43, 117, 150, 153, 155 Ortiz-Millán, G. 158 – 60, 188 Pacheco, J. E. 185 Parfit, D. 13 – 15 partiality see impartiality Pascal, F. 187 Peacocke, C. vii Pettit, P. 30 phronésis 16 – 17 PI see presumption of innocence Plato 187 Platts, Mark: on contemporary philosophical life 184 – 8; on “deflation” of human rights 61, 65 – 6; on “desirability characterizations (full and explicit)” ix, 4 – 5, 150 – 2, 156, 160; on direction of fit ix, 27 – 33, 38 – 9, 158 – 60; on essentially contested concepts 171; on HIV/AIDS ix – x; on human nature 184; on human rights and feasibility 78 – 9; on human rights and needs 60, 64 – 6, 77 – 8, 163, 167; on “inflation” of human rights 59, 61, 69 – 70, 77 – 8, 168 – 9; on irrationality 14, 156; on justification of human rights 61, 64 – 6, 78 – 81; on moral intuitions 19 – 21; on moral motivation ix, 13 – 17, 156; on moral phenomenology 13, 15, 186; on moral realism viii – ix, 3, 13 – 14; on moral rules 13, 18,

153, 168; on moral worth 15 – 16; on objectivity/subjectivity 13, 15 – 17, 19, 22 – 4, 153, 155; on possession of human rights 61, 77 – 8, 162 – 3; on “prescriptivism” 15, 17 – 18, 154 – 5; on presumption of innocence 170 – 3; on private elements in philosophical life 185, 187; on “professionalization” within humanities 184 – 5, 187 – 8; on proto-practical deliberation ix; on realism vs. anti-realism in semantics vii – ix; on reasonable causation ix; on reasons for actions and desires ix, 4 – 5, 9, 13 – 17, 28 – 30, 33, 115 – 16, 129 – 30, 135 – 7, 150 – 2, 155 – 6, 163, 179; on responsibility x; on secondary qualities 6 – 7, 150; on social human rights 167; on “transcendentalism” 5 – 10; on truth-conditional semantics vii – ix, 2 – 4, 115 – 16, 136, 138, 186; on utilitarianism 4 Popper, K. 141, 184 “prescriptivism” 15, 17 – 18, 154 – 5 presumption of innocence: formulation of 83, 89 – 93, 169 – 72; supposed consequences for pre-trial practices 85 – 6, 172 – 3; supposed consequences for trial procedures 87 – 9 privacy 56, 86, 93, 185 – 6 private elements in philosophical life 185, 187 “professionalization” within humanities 174, 184 – 5,  187 proto-practical deliberation ix Proust, M. 37 Putnam, H. viii, 97, 156 Quine, W. V. O. 116, 118, 121 – 2, 133 – 5, 138, 149, 157, 178 – 9 Quinn, W. 151 radical interpretation 115 – 16, 121, 133 – 7, 139, 178 – 80 rationality see irrationality Rawls, J. 14, 62, 64, 71, 75 realism vs. anti-realism in semantics vii – ix, 1 – 4, 10 – 12 reasonable causation ix, 155 – 6 reasons for actions and desires 4 – 5, 8 – 12, 13 – 17, 26 – 32, 32 – 43, 129 – 30, 149 – 52, 155 – 6

196 Index responsibility x rights: criminal 83 – 6, 88, 90, 169, 173 Rousseau, J. J. 163 Ruiz Manero, J. 173 Russell, B. 99, 131 – 3, 179 saccades 124 – 5, 127 – 9,  141 Scanlon, T. M. 52 – 3, 164 Schiffer, S. 150 – 1 scientism 180 – 1 Searle, J. 26, 45 secondary qualities 6 – 8, 150 self-deception 16, 23, 31, 37 Sen, A. 64 sense-data 132 – 4, 178 – 9 Shakespeare, W. 188 Smith, M. 35 – 9, 47 Snowdon, P. F. vii, 162, 174 – 7 Sobel, D. & Copp, D. 37 Stampe, D. 42 Stocker, M. 48 Stoics 50 Strawson, P. F. vii, 150, 159, 162, 174, 181, 183, 185 – 6 Stroud, B. 149 – 52, 160 subjectivity 117, 150, 160 Sunstein, C. 82 system in ethics 51, 55 – 6, 162, 164 Tait, D. 88 Tarski, A. vii Tasioulas, J. 76 Taylor, B. vii Taylor, C. 64 Tenenbaum, S. 48 “transcendence” see “transcendentalism” “transcendentalism” 5 – 10, 148 – 50

truth vii – viii, 2, 5, 9, 23, 27, 30 – 2, 38 – 44, 136, 158 – 9, 185, 186 – 7 truth-conditional semantics viii – ix, 1 – 4, 10 – 12, 136 – 8, 178 – 80, 181, 186 Turing, A. 142, 184 Understanding 1 – 4, 11 – 12, 19, 21, 97, 101 – 6, 108 – 13, 115, 135 – 6, 149, 175 – 7,  186 utilitarianism 4, 50, 53 Vázquez, R. 164, 166 – 9 Velleman, D. 48 vulnerabilities 61, 64, 66, 77, 163 Walker, A. D. M. x Walker, R. 153 – 7, 163 Warnock, G. J. 154 Waugh, E. 183 Wiggins, D. vii – viii, 3, 78 Williams, B. 174 – 5, 184 – 5,  187 Wittgenstein, L. 97 – 113, 116, 121, 130 – 1, 133, 149, 162, 176 – 7, 178, 186, 187; on “circumstances” 110 – 12, 177; on meaning and use 97, 100 – 3, 116, 121, 130, 177; on motivation 116; on philosophy 98 – 101, 175 – 6, 188; on private languages 97 – 8; on rule-following 97 – 8, 100 – 4, 106 – 13, 174 – 6; on understanding 97, 101 – 6, 109 – 13, 176 – 7 Wollheim, R. 40 Woolf, V. 156 Zangwill, N. 39