The Act and Object of Judgment: Historical and Philosophical Perspectives [Hardcover ed.] 1138351385, 9781138351387

Citation preview

The Act and Object of Judgment

This book presents 12 original essays on historical and contemporary philosophical discussions of judgment. The central issues explored in this volume can be separated into two groups—namely, those concerning the act and object of judgment. What kind of act is judgment? How is it related to a range of other mental acts, states, and dispositions? Where and how does assertive force enter in? Is there a distinct category of negative judgments, or are these simply judgments whose objects are negative? Concerning the object of judgment: How many objects are there of a given judgment? One, as on the dual relation theory of Frege and Moore? Or many, as in Russell’s later multiple relation theory? If there is a single object, is it a proposition? And if so, is it a forceneutral, abstract entity that might equally figure as the object of a range of intentional attitudes? Or is it somehow constitutively tied to the act itself? These and related questions are approached from a variety of historical and contemporary perspectives. This book sheds new light on current controversies by drawing on the details of the distinct intellectual contexts in which previous philosophers’ positions about the nature of judgment were formulated. In turn, new directions in present-day research promise to raise novel interpretive prospects and challenges in the history of philosophy. Brian Ball is Senior Lecturer in Philosophy at New College of the Humanities, London and Associate Member of the Faculty of Philosophy at the University of Oxford. He works in the philosophies of mind and language, epistemology, and metaphysics, and has published papers in these areas in journals including Analysis, Erkenntnis, Mind and Language, Philosophical Psychology, and Philosophical Quarterly. Christoph Schuringa is Lecturer in Philosophy at New College of the Humanities, and has recently been a visiting scholar at the Universities of Leipzig and Pittsburgh. He works in the history of German philosophy and in practical philosophy, and has published in journals including History of Philosophy Quarterly and International Yearbook of Hermeneutics.

Routledge Studies in Contemporary Philosophy

Questions of Practice in Philosophy and Social Theory Edited by Anders Buch and Theodore R. Schatzki The Phenomenology of Real and Virtual Places Edited by Erik Malcolm Champion Context, Truth, and Objectivity Essays on Radical Contextualism Edited by Eduardo Marchesan and David Zapero Good Thinking A Knowledge First Virtue Epistemology Christoph Kelp Epiphenomenal Mind An Integrated Outlook on Sensations, Beliefs, and Pleasure William S. Robinson The Meanings of Violence From Critical Theory to Biopolitics Edited by Gavin Rae and Emma Ingala A Defense of Simulated Experience New Noble Lies Mark Silcox The Act and Object of Judgment Historical and Philosophical Perspectives Edited by Brian Ball and Christoph Schuringa For more information about this series, please visit: www.routledge.com/ Routledge-Studies-in-Contemporary-Philosophy/book-series/SE0720

The Act and Object of Judgment Historical and Philosophical Perspectives Edited by Brian Ball and Christoph Schuringa

First published 2019 by Routledge 52 Vanderbilt 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 © 2019 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 Names: Ball, Brian Andrew, 1978– editor. | Schuringa, Christoph, editor. Title: The act and object of judgment : historical and philosophical perspectives / edited by Brian Ball and Christoph Schuringa. Description: New York : Routledge, 2019. | Series: Routledge studies in contemporary philosophy ; Volume 118 | Includes bibliographical references and index. Identifiers: LCCN 2019005335 | ISBN 9781138351387 (hardback) | ISBN 9780429435317 (ebk.) Subjects: LCSH: Judgment (Logic) | Act (Philosophy) | Object (Philosophy) Classification: LCC BC181 .A33 2019 | DDC 128/.4—dc23 LC record available at https://lccn.loc.gov/2019005335 ISBN: 978-1-138-35138-7 (hbk) ISBN: 978-0-429-43531-7 (ebk) Typeset in Sabon by Apex CoVantage, LLC

Contents

Introduction

1

BRIAN BALL

1 Affirmation, Judgment, and Epistemic Theodicy in Descartes and Spinoza

26

M A RTI N L I N

2 Locke and Leibniz on Judgment: The First-Person Perspective and the Danger of Psychologism

45

M A R I A VA N DE R SCH A A R

3 Kant’s Logic of Judgment: Against the Relational Approach

66

A L E X A N D R A N E WTO N

4 Time and Modality in Hegel’s Account of Judgment

91

PA U L R E D D I N G

5 Bolzano’s Theory of Judgment

110

MARK SIEBEL

6 Correctness First: Brentano on Judgment and Truth

129

M A R K TE X TO R

7 Judgment, Reasons, and Feelings

151

S I M O N B L A CKB URN

8 Twardowski on Judgment

165

P E TE R S I M O NS

9 Attitudinal Objects: Their Ontology and Importance for Philosophy and Natural Language Semantics F R I E D E R I K E MO LTMAN N

180

vi

Contents

10 About vs Concerns

202

D A N I E L M O RGAN

11 Predication and Two Concepts of Judgment

217

I N D R E K R E I LA N D

12 How Is Logical Inference Possible?

234

C H R I S TO P H E R P E ACO CKE

Contributors Index

256 259

Introduction Brian Ball

After many years in the philosophical wilderness, judgment is enjoying something of a resurgence. Traditionally, it has been a central subject of philosophical inquiry: it was investigated by ancient theorists (Nuchelmans 1973; Perala 2013), and it played a prominent role in modern philosophy (Nuchelmans 1984; Moltmann and Textor 2017).1 As recently as the early and even mid-twentieth century, philosophers such as Bertrand Russell (1910, 1912, 1913/1984) and Peter Geach (1957) devoted themselves to developing theories of judgment. But in the late twentieth century, the topic fell out of favour, with belief becoming the primary target of investigation in this area of philosophical psychology: for instance, ‘belief’ and ‘propositional attitude’ both get lengthy entries in the index of W. V. O. Quine’s enormously influential (1960) book, Word and Object; ‘judgment’, by contrast, does not figure. It is an interesting question why this recent historical situation should have arisen. Was the decline of judgment due to the increasing influence of behaviourism (e.g. Skinner 1953) and a concomitant emphasis on overt actions (cf. Ryle 1949),2 or did the philosophical methodology of the ‘linguistic turn’ (cf. Rorty 1967) direct attention to attitude ascriptions rather than the attitudes themselves, somehow thereby privileging belief over judgment?3 Whatever the reasons—themselves worthy of investigation4—what is important here is that the recent neglect of judgment is over.5 The present volume aims to contribute to our understanding of both judgment itself and the modern history of philosophical engagement with it. On the one hand, the notion of judgment is directly involved in matters of central significance in logic, construed broadly, as it was historically, so as to include “topics that would fall today under the heading of theory of knowledge” (MacLachlan 2018: introduction)6 and the philosophy of mind; and it is arguably, though perhaps more indirectly, related to fundamental issues in metaphysics. An understanding of judgment therefore promises to be illuminating to contemporary philosophers in a variety of ways. On the other hand, our predecessors in the history of philosophy have many insightful, and sometimes neglected, things to teach us as we

2

Brian Ball

are developing and applying theories of judgment. Moreover, new light can be shed on that very history through a consideration of our contemporary discussions of the issues they tackled. Thus, both philosophers and historians of philosophy can benefit—in a two-way interaction—from the kind of engagement with the philosophy of judgment and its history offered here.7 Allow me to illustrate by briefly tracing some key elements of the history of the topic before providing a more conceptual orientation to the subject matter.

Initial Historical Background “The definition commonly given of judgment, by the more ancient writers in logic”, wrote Thomas Reid in his 1785 Essays on the Intellectual Powers of Man, “was, that it is an act of the mind, whereby one thing is affirmed or denied of another” (Reid 1983: 251; italics original). This view, which originates with Aristotle (cf. Marušić 2017: introduction), has been extremely influential throughout the history of philosophy.8 In the early modern period, John Locke (1690) held a version of it (Marušić 2017: section 1; Schaar, this volume), and it also figured in the Port-Royal Logic of 1662—“the most influential logic text from Aristotle to the end of the nineteenth century” (Buroker 2017: introduction). “After conceiving things by our ideas”, wrote its authors, Antoine Arnauld and Pierre Nicole, “we compare these ideas and, finding that some belong together and others do not, we unite or separate them. This is called affirming or denying, and in general judging” (1662: 82—cited in Marušić 2017: section 1). The Port-Royal Logic also adopted “the traditional view that there are four mental operations required for scientific knowledge: conceiving, judging, reasoning, and ordering” (Buroker 2017: section 2). It was organized accordingly, and the second of its four parts was devoted to an investigation of judgment. The first three of these acts especially continued to receive attention from philosophers thereafter.9 Thus, in his 1740 Treatise of Human Nature, David Hume spoke of “the vulgar [i.e. common] division of the acts of the understanding, into conception, judgment, and reasoning” (Treatise, 1.3.7.5);10 and when Francis Herbert Bradley wrote his Principles of Logic in 1883, it was still common for logic texts to be organized in three corresponding parts.11 Bradley’s views in this area are historically significant, for it is with his attack on Bradley’s theory of judgment—as well as the views of Immanuel Kant—that George Edward Moore (1899) launched the analytic tradition of philosophy in Britain, upturning the dominant idealist approach of the day and replacing it with a variety of metaphysical realism. It will therefore be worth dwelling, briefly, on Bradley’s view and Moore’s criticisms of it. “Judgment”, wrote Bradley, “is the act which refers an ideal content (recognized as such) to a reality beyond the act” (1883: 10—quoted in

Introduction

3

MacLachlan 2018: section 3). An ideal content, on Bradley’s view, is a kind of abstraction from concrete mental episodes; it is a logical rather than psychological idea that is universal rather than particular in character, and it accordingly has no independent existence. In the act of judgment, it is isolated from the content of more immediate and particular experience, and it is then “asserted” (1883: 13—quoted in MacLachlan 2018: section 3) of its subject, “the Absolute”. Moore criticized Bradley’s view on at least two points. First, he argued that it leads to regress. Roughly, the worry is that, if the ideal content of a judgment is to be isolated from the more particular content of experience, this must be done knowingly (the ideal content must be “recognized as such”); but this requires a prior judgment (of which that knowledge is the upshot), which in turn requires an infinite regress of judgments, each prior to the last. Since this is absurd, Moore concluded that what is judged has existence independently of any act of the mind; it is not a Bradleyan ideal content but a complex of concepts standing in necessary relations to one another, constituting a proposition. Second, he complained that truth can’t consist of a relation between a truth-bearer and reality that implies a difference between the former and the latter, as he took Bradley’s view to require. The reason, it seems, is that Moore thought it is “necessary . . . to regard the world as formed of concepts” (1899/1993: 5). But then there is no difference between the world and the propositions that are judged when one judges truly, and so the latter can’t stand in any relation to the former that implies such a difference. Instead, Moore held that truth is a primitive feature that some propositions have,12 while other propositions (e.g. their negations) are primitively false. Whether or not these objections to Bradley are successful, Russell accepted Moore’s view in his (1903) Principles of Mathematics. Later, he came to reject the idea that there are objective falsehoods, and he accordingly developed an approach to judgment that had features in common with the earlier views of Aristotle, Locke, and the Port-Royal logicians, as we shall see. Both of these views of Russell’s continue to be influential today.

Conceptual Preliminaries It is time, now, to turn to more conceptual matters, and it will be helpful, so as to focus our discussion, to begin by saying a few words by way of delineation of our topic. In typical (though perhaps not all) cases of judgment, a subject arrives at a kind of doxastic decision regarding the truth of some matter; he does not, for example, resolve a (standard) practical deliberation concerning what to do. It is with such theoretical rather than practical judgment (cf. Hanna 2016) that we are concerned here. Bearing this in mind, we can perhaps say that (such theoretical) judgment is a mental act (Geach 1957) that is directed towards an intentional object (Brentano 1874). Thus, like assertion, for example, it is an act—something

4

Brian Ball

done,13 whose performance is an occurrence, and it has an object—so that something is judged, when one judges, just as something is asserted, when one asserts. Unlike assertion, however, judgment is a mental phenomenon, and in this respect, it is like the mental states of belief and desire (which, arguably, are also object-directed) or, indeed, sensations like pleasure and pain (which, arguably, are not)—or so it has been suggested—and these are at least useful hypotheses from which to launch our discussion. In what follows, I begin with a discussion of the act of judgment. I then consider the credentials of the act-object analysis before turning to the nature of the object. Along the way, I will mention some more of the key moments in the history of the philosophy of judgment, and I will at least raise, and I hope illuminate, the following questions: (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)

In what sense is judgment an act? How does it relate to belief and knowledge? Does judgment constitutively involve answering a question? How is it related to assertion (and assent)? Are there (irreducibly) negative judgments (to be understood in terms of dissent)? Is judgment a unifying act of synthesis, or does it have some other character? How many objects does the act of judgment have? Are the objects of judgment (if any) propositional (and truth-apt)? Do judgments have both (narrow) contents and (wide) objects? Are the objects of judgment (if any) structured complexes? Are they ontologically independent of mind and language? Are they neutral with respect to (assertoric) force?

After engaging these issues in the three-part structure indicated above, I will conclude by introducing the papers that comprise this volume.

The Act of Judgment What does it mean to say that judgment is a mental act? Joelle Proust (2010: 209) distinguishes two senses of the phrase ‘mental act’. On the first of these, she says, the phrase is synonymous with ‘mental event’—it picks out any mental occurrence.14 Following Aristotle, Proust says a mental act in this sense involves the actualization of a potentiality. This last characterization seems apt for capturing a minimal notion of mental act, but I would add that the potentiality in question must be a capacity—roughly, a potentiality with a purpose (for which it was selected, and hence not a mere disposition, or power, such as a stone might have to fall); thus, the actualization in question, the mental act, will itself be the manifestation of such a capacity—the exercise of a mental faculty. By contrast, says Proust, in the second sense, “a mental act is the process of intentionally activating

Introduction

5

a mental disposition in order to acquire a desired mental property” (2010: 209). In other words, it is a species of (intentional) action, differentiated from other members of the genus by its (represented) goal, that of effectuating a mental change in the subject (cf. Proust 2001). It seems clear that judgment is a mental act in (at least) the first, more minimal sense: we have the (naturally selected) capacity to judge, and we exercise that capacity in judging. Is it (also) an act in Proust’s richer, second sense? Is judging an intentional action? René Descartes (1642) famously thought so. He held that error (i.e. false belief) is possible, despite God’s existing, creating us, and being no deceiver, because we exercise our unlimited will in judging matters that extend beyond the grasp of our limited understanding. Moreover, he maintained that judgment is an exercise of the will, even in those cases in which the understanding presents us with ‘clear and distinct’ ideas to judge. Thomas Hobbes objected to Descartes’s invocation of the will in this context on the grounds that there are many things one cannot, and other things one cannot but, believe (cf. Williams 1978: 161). And Bernard Williams notes that Descartes’s view raises the prospect of “someone who had no evidence whether p was true deciding at will to believe that p” (1978: 161). An alternative view, which avoids this kind of concern, is that judgment is an involuntary act of the understanding. Reid (1983), for instance, held that judgment is an “operation of the intellect” and is involved in other such acts, including perceiving15 and remembering. The question of whether judgment is an intentional action remains very much a live one. I turn now to a quite distinct cluster of issues surrounding the act of judgment—namely, how it relates to other intentional states and acts, such as belief, knowledge, wondering, and assertion. To address these issues, it will be useful to have to hand a concrete contemporary model of what judgment might be. Ernest Sosa, in his (2015) book, Judgment and Agency, “take[s] judgment that p to be a certain sort of alethic affirmation, in the endeavor to get it right on [the question] whether p” (Sosa 2015: 52, emphasis original). This proposal stands in need of commentary. First, note that on Sosa’s view, judgment is a kind of alethic affirmation. Sosa contrasts this with pragmatic affirmation, which aims (for instance) at instilling confidence to enhance performance (2015: 52)—for example, saying ‘you can do it’ to oneself when competing at sport. The idea here appears to be that (theoretical) judgment, as it is sometimes said, ‘aims at truth’ (cf. Williams 1970), whereas the promotion of confidence does not; and it is, accordingly,16 correctly undertaken only if the proposition judged is true.17 Second, Sosa says “a certain sort” of alethic affirmation because he thinks that when one guesses that p, one does not judge that p, yet one engages in alethic affirmation in the endeavour to get it right on the question whether p. Thus, judging that p requires more; in particular, it requires “[a]ffirmation in the endeavor to answer correctly and also competently, reliably enough, even aptly” (2015: 55). Now, according to

6

Brian Ball

Sosa, some acts, or performances, have objectives, or aims. An archer’s shot, for instance, aims at its target; moreover, it “might be accurate by hitting its target; it might be skillful or adroit [with or without being accurate]; and, finally, it might be apt: accurate because adroit” (2015: 1). Moreover, on Sosa’s virtue-theoretic approach to epistemology, for a belief to constitute knowledge, it must meet the “triple-A” (2015: 1) standard of “accuracy, adroitness, and aptness” (2015: 1). In other words, it must be true (i.e. accurate) because the result of the exercise of a capacity for such belief. Given this background, Sosa’s proposal seems to be that to judge that p is to affirm that p in the attempt to achieve knowledge on the question whether p. If the attempt is successful, the result will be knowledge;18 if not, the judgment will yield mere belief. A third point to note is the central role played by questions in Sosa’s account of judgment. “Once a question is given”, says Sosa, “there arises the familiar threefold issue: affirmation, denial, suspension” (2015: 44). Thus, on Sosa’s view, the theoretical question of whether p appears to the subject under the guise of the practical question of what to do when confronted with this question, and indeed, Sosa claims that “in cases of conscious judgment . . . [t]he epistemic agent faces a choice among three intentional actions” (2015: 44)—namely, those mentioned above. Thus (and this is now a fourth point), for Sosa, judgment is intentional, not only in the sense that it is object-directed but also in the further sense that it is “agential” (2015: 53) and “volitional” (2015: 54);19 in short, it is an intentional action—a mental act in Proust’s second sense above. As we have seen, Descartes would accept this fourth claim, but Hobbes and Reid would not. As for the third, clearly, in certain (arguably paradigmatic) cases of judgment, we wonder whether p, and then, having gathered enough evidence, we conclude either that p or else that not p; and yet it is not entirely obvious that possession of such an “interrogative attitude” (Friedman 2013) is constitutively necessary for an act of judgment to occur. For instance, it does not seem that the visual system ponders any questions20 before delivering its pronouncements; thus, if Reid’s view is correct and judgment is constitutively involved in (e.g. visual) perception, it would appear that judging does not require a prior question-directed act of wondering what’s the case.21 Finally, as already noted, on Sosa’s view, judgment is a kind of (alethic) affirmation. “Affirmation” of this alethic variety, says Sosa, “can be either public, through assertion, or private, to oneself” (2015: 66). This suggests that an affirmation might take the form of ‘assent’ (2015: 52; cf. Quine 1960), revealed in, for example, the use of a sentence to make an assertion or in certain sorts of responses to others’ uses of that sentence (such as saying, “I agree”) or related expressions (for instance, answering yes to the corresponding yes-no question). And of course, assent so understood can be internalized, taking place, as it were, in the privacy of one’s own thoughts. Now, assent in this sense is a relation to a sentence,

Introduction

7

not a proposition, but we might take judgment, as Sosa understands it, to be the relation one bears to a proposition when one assents to a sentence that expresses it; and if so, it will not be an act that is ever performed by non-linguistic animals (just as Sosa himself maintains).22 Such a view of judgment is certainly not unprecedented; for instance, Geach (1957: sections 17–23) endorses what he calls the “analogical” theory of the concept of judgment, on which it is grasped, or exercised, on analogy with saying; and he seems to think this requires us to deny that animals make judgments. But neither is the view forced upon us.23 Two further issues arise in connection with the act of judgment that are perhaps best approached through the lens of the history of philosophy. First, as we have seen, Locke and the Port-Royal logicians took there to be a distinction between judgments involving affirmation and those involving denial. On their view, we might say, some judgments (those involving denial) are irreducibly negative. The orthodox view, embraced by both Moore and Gottlob Frege, by contrast, is that there are no such judgments. Rejecting, or dissenting from, the proposition that p is just accepting or assenting to the proposition that not p; thus, negation is part of the object of judgment, not the act. But the idea that there are irreducibly negative judgments is not (obviously) entirely foreign to the analytic tradition; for instance, Ludwig Wittgenstein (1921/1974) held that negation is an operation that reverses the sense of a proposition, where this meant its agreement or disagreement with the various possibilities, and such reversal might conceivably be construed as a type of act.24 Second, there is the question of whether judgment is a synthetic act; for instance, both Locke and the Port-Royal logicians appear to have held that (at least in the affirmative case) judgment is a unifying act, bringing together disparate representations of subject and predicate.25 Moreover, Wayne Martin (2006: chapter 3) argues that whereas for Kant judgment remains synthetic, for Franz Brentano it is thetic; it takes a single object as input, which it leaves unchanged through its operations.26 For the sake of completeness, we can also consider the possibility that judgment is an analytic act, separating out elements, or components, of a single given object.27,28 I do not take myself to have answered any of the questions I have raised about the act of judgment. And, of course, much more could be said, even by way of introduction. Nevertheless, I turn now to considering the actobject analysis hypothesized above.

The Act-Object Analysis Is judgment best analysed as involving a relational, act-object structure? Some initial support for this proposal comes from the semantics of sentences used to report the act. Like beliefs and assertions, judgments can, on the face of it, be reported with sentences of the form ‘S Vs that p’; and a standard analysis of such sentences suggests that what replaces ‘S’ names a

8

Brian Ball

subject; the verb replacing ‘V’ picks out a two-place relation; and the thatclause designates the object of that relation. In particular, then, we might think that the verb ‘said’—at least when used in such constructions (rather than in direct discourse)—stands for the speech act of asserting; the verb ‘thinks’ picks out the psychological attitude of believing; and the verb ‘judges’ denotes the mental act of judging. On this semantic approach, it turns out that judgment, like other intentional acts and attitudes, consists in a relation between a subject and an object. Some, however, might question the act-object analysis. Uriah Kriegel (2008), for instance, has argued that intentional acts and states in general do not constitutively involve a relation between a subject and an object (though they may do so contingently);29 they consist simply of mentally acting in, or being, certain ways (Kriegel 2008: 84). Thus, according to Kriegel, these acts and states are monadic characteristics of individuals, not dyadic relations between them and certain further entities, their objects. An advocate of such a view will, of course, need some account of the semantics of reports of intentional acts and states. And he might, it seems, accept that such reports have a relational semantic structure (as above) while insisting that the (dyadic) relation expressed by the verb is not that of assertion, belief, or judgment itself—for, after all, he will maintain that no such relations exist! Metaphysically speaking, he might say, there are just clusters of (monadic) acts subjects may perform and states she may be in; thus, judging that p, judging that q, and so on, form a family of related mental acts; and believing that p, believing that q, etc., on the other, constitute a collection of relevantly similar mental states. In this respect, it might be argued, mental acts and states would be not unlike certain physical features (cf. Field 1980; Stalnaker 1984; Matthews 1994, 2007). For instance, there are many masses, or temperatures, that an object may have; the masses, like the temperatures, form a family of properties (Peacocke 2015). Of course, we might relate a subject to an object in reporting the act she performs or the state she is in; we might say that she judges that p or believes that q. Similarly, we might relate an object to a number in reporting its mass or temperature; thus, we might say that it has a mass of 21 grams or a temperature of 99.9 degrees Fahrenheit, and this might be construed as the claim that the object stands in the hasa-mass-in-grams relation to the number 21 or the measures-in-degreesFahrenheit relation to the number 99.9. It does not follow that the mass or temperature that we thereby ascribe really consists of a relation to a number; indeed, it seems implausible to suppose that these numbers form integral parts of the properties of having 21 grams of mass or measuring 99.9 degrees Fahrenheit, for what are intuitively the same features of the object might be measured using different scales, in which case our attributions would relate the object to different numbers—its mass in pounds, for example, and its temperature in Kelvin. Analogously, then, the objects of judgment and belief reports, for instance, are not really constituents of

Introduction

9

the acts and states in question either on the view under consideration,30 and if the verb ‘judges’, for example, expresses a relation, it is not that of judgment (recall, there is no such relation on this view); rather, it is something like the judges-in-a-way-that-can-be-measured-by relation.31 In any case, not everyone accepts the above account of the logical structure of the reports of intentionality. “The relation involved in judging or believing must . . . be taken to be a relation between several terms, not between two”, according to Russell (1912 [1959]: 125); thus, he took ‘judges’ to be variably polyadic, taking different numbers of arguments in different cases. Wittgenstein (1914–1916/1979, 1921/1974) worried, however, that the logical form of a belief or judgment should not vary with its subject matter as it would on this proposal (cf. Geach 1957: 49). In some recent work, however, Friederike Moltmann (2013a) has suggested that the verbs in act and attitude reporting constructions are not profitably regarded either as expressing two-place relations between a subject and a single object designated by the complement clause, nor as variably polyadic; rather, she argued, they express relations between the subject, on the one hand, and the plurality of entities designated by the various components of the complement, on the other. Accordingly, she advocated a version of the ‘multiple relation theory’, on which judgment is (at least roughly) an act with not one but many objects, just as Russell suggested; yet, on the view espoused, the Wittgensteinian objection is overcome.32 More recently still, however, Moltmann (2013b, 2017) has argued that in reports of attitudes and acts, that-clauses serve as predicates of what she calls “attitudinal objects”. The idea is that a sentence of the form ‘S Vs that p’ involves quantification over events (cf. Davidson 1967)— intuitively, the act or state mentioned by the verb—and the complement clause then predicates something of an entity standing in a certain relation to that act or state. In taking this line, Moltmann draws on the view of Kazimierz Twardowski (1912), according to which, in general, for any given action there is something distinct from it—namely, its product, to which it gives rise. On Twardowski’s view, some such products are enduring (for example, the letters that result from writings), while others are not, lasting only so long as the corresponding action is under way. Moreover, Twardowski maintains that the products of judgments, which are actions, are certain non-enduring mental items, also called ‘judgments’, in a distinct but related sense (cf. Betti 2017: section 3.2). On Moltmann’s more recent view, then, the entity of which the complement clause predicates something is the product of the event (i.e. act or state) quantified over in the report; and intuitively, what this predication does is classify that product according to its representational type, or kind.33 I will not attempt to decide which, if any, of these views is correct; in what follows I will, for expository purposes, simply assume the standard act-object analysis. Having done so, we can ask, what is the nature of the

10

Brian Ball

(one) object of judgment? It is this question that will exercise us in the next section.

The Object What is it that we judge when we judge? If “judgments just are beliefs”, as Trenton Merricks (2009: 211) asks us to suppose, then obviously the objects of the former will be the same as those of the latter. But this assumption seems reasonably safe, even on much looser conceptions of the relation between the act and the state—e.g. on which judgment is the act of belief formation or, alternatively, the expression of belief (Boyle 2009). Contemporary work in this area therefore suggests that the object of judgment, like that of belief, is a (truth-apt) proposition.34 This has not, however, always been the accepted view. According to Locke, for instance, “whatsoever is the Object of the Understanding when a Man thinks” (1690/1975: book 1, chapter 1) is an idea—that is, a certain sort of mind-dependent entity. Moreover, as Williams notes, ideas were introduced in Descartes’s (1642) Meditations “as ideas of (possible or actual) things, as the idea of a triangle, of God, of a chimera, etc.” (1978: 167). Yet Williams is concerned that it makes no sense to say that one can assent to a thing, or to the idea of a thing. I can assent only to something of the nature of a proposition: one believes, or refuses to believe, that such-and-such is the case. (1978: 167) In short, ideas, as understood by the early moderns, are not propositional and truth-apt,35 yet the objects of belief and assent or judgment must be; so, their objects cannot be ideas (so understood).36 Compelling as this argument may seem to contemporary ears, it would not have been accepted as sound by Brentano, for whom what we act on in judgment is always an intentional object. Brentano is at pains to stress that the intentional object that is judged need not be a thing: it might, for instance, be what Alexius Meinong later (1904/1960) called an objective or Wittgenstein (1921/1974) called a state of affairs—such as snow’s being white. Nevertheless, it can be a thing, and accordingly, Brentano would not accept that one cannot assent to or acknowledge “a thing, or the idea of a thing” (in Williams’s words), for one does so precisely by taking the thing in question to exist, on his view, and one can certainly accept the existence of a thing. Brentano also held that the intentional object of any mental phenomenon is contained “within” (1874/1995: 68) the act. But, as Wolfgang Huemer (2017) says, this led some of his students—who included not only Meinong but also Sigmund Freud and Edmund Husserl—to suggest that the ontological status of intentional objects was unclear. In particular,

Introduction

11

should these objects be regarded as objective and mind-independent, or should their immanence be taken as evidence that Brentano took them to be mind-dependent? Whatever the case may be,37 Twardowski (1894) drew a distinction between the object of a presentation and its content.38 The former was (typically) non-mental; the latter was an inseparable (i.e. abstract) part of the mental act of presentation. This distinction is often said (e.g. by Betti 2017) to be a psychologistic version of Frege’s (1892) distinction between sense and reference. Be that as it may, it seems that Twardowski was an early advocate of the idea that there are two aspects of the intentionality of a given state or act (or ‘components of content’— cf. Chalmers 2002), one narrow (the content) and one wide (the object), so that we need something like a two-dimensional semantics (cf. Chalmers 2006).39 In any case, the standard view nowadays is that propositions are the objects (or perhaps contents)40 of belief and judgment, and that these are abstract entities, existing independently of any mental or linguistic act, essentially and intrinsically possessing truth conditions and therefore susceptible of truth or falsity. This view, which is widely known to have been advanced by Frege (1918/1956) and which Moore also advocated, already makes a mature appearance in the work of Bernard Bolzano (1837). More specifically, Bolzano makes a key departure from theorists of the early modern period in hypothesizing abstract (hence nonmental) representations in themselves and, as a special case, propositions in themselves; in other words, like Frege, he rejects psychologism in favour of platonism about intentional objects in general and the objects of judgments—propositions—in particular. Moreover, according to Jan Sebestik, “for Bolzano, propositions are primary, undefined objects, and ideas in themselves . . . are defined as parts of propositions that are not themselves propositions” (2016: section 5). In short, propositions, for Bolzano, as for Frege and Moore, are sui generis abstract objects. There are, broadly speaking, three standard versions of this orthodox view. According to a first, Tractarian view, propositions are sets of possible worlds; more specifically, the proposition that p is the set of worlds in which p is true (Wittgenstein 1921/1974; Stalnaker 1984; Lewis 1986).41 A second, neo-Russellian view treats propositions as metaphysically complex entities having worldly objects and universals as constituents (Russell 1904/1980) and structured in a way that mirrors, at least roughly, the syntactic structures of sentences expressing them. And finally, a third, neo-Fregean view differs from the neo-Russellian one only in taking the constituents of propositions to be, not worldly items, such as objects and universals, but rather ‘modes of presentation’ thereof, or senses (Frege 1892/1948, 1918/1956). Whatever its merits, the orthodoxy (in all its versions) has recently come under attack42 (Hanks 2015; King 2007; Soames 2010); in particular, a number of theorists have felt that propositions cannot be sui generis,

12

Brian Ball

primitively truth-conditional entities, but that they must instead be naturalized (King 1994); that is, they must be reduced to, or at least explicated in terms of, entities that are recognized by the natural sciences (such as psychology or linguistics). An alternative view has therefore been gaining traction, on which the object of judgment is dependent upon our cognitive or linguistic activities. Perhaps, for example, what one judges when one judges is essentially some type of mental or linguistic act, maybe even a type of judgment, much as what one dances when one dances is essentially some type of dance (Ball 2011; Husserl 1903; cf. Moltmann and Textor 2017: x–xi). And while on the standard view, truth-apt propositions can be the objects not only of cognitive acts and attitudes, such as judgment, belief, and knowledge, but also of conative ones, such as desire and intention, so that one and the same thing can be, for example, believed and desired, this is up for grabs on the alternative approach.43 Allow me to briefly expand on this last point. “A central component of [the orthodox] Fregean picture is”, as Peter Hanks notes, “the distinction between content and force” (2015: 9; cf. Geach 1965). But Hanks rejects this distinction, on which (for example) “there is nothing distinctively assertive” (2015: 9) about the propositions we assert and believe.44 Contents (or objects),45 he thinks, are not force-neutral; rather, they “are individuated using concepts of force” (2015: 9). Different acts and attitudes can therefore have different kinds of (force-laden) objects, on his view, as appropriate. Much more could, of course, be said, but I hope that what I have said will do, for present purposes, by way of survey of the different accounts of the objects of judgments. In particular, though this was not always so, most nowadays will hold that the objects of judgment are propositions. Some maintain that propositions are structured, while others deny this (advocating the Tractarian view); amongst the former, some (neoRussellians) take propositional constituents to be worldly items, while others (neo-Fregeans) take them to be modes of presentation thereof. But perhaps most importantly, there are orthodox positions on which the objects of judgment are ontologically independent of both mental and linguistic activity, and there are alternatives on which they are not— including views on which the objects themselves carry the force of the act.

Contributions to This Volume Both Descartes and Benedict Spinoza aim to reconcile the possibility of error with God’s perfection, according to Martin Lin, but Spinoza does not distinguish the faculties of intellect and will, and even insists that all ideas are affirmations (roughly, subpersonal judgments). On Spinoza’s metaphysics, there is “a hierarchy of increasingly complex bodies” culminating in God’s, each of which has a striving for self-preservation, or conatus. Moreover, “every idea is identical to the body that it represents”;

Introduction

13

this includes the human mind, which is the (complex) idea of the human body. Ideas are true or false only relative to a mind, and “the conatus associated with an idea pushes [a mind to which it belongs] towards actions that would conduce to the self-preservation [of that mind and the body it represents] if the idea were true” relative to it. This makes room for false ideas (though not in God’s mind), and it allows an idea to be denied in a (human) mind that contains, in addition to a given affirmation, a further idea with a conflicting content yet stronger conatus. It does not make room for a notion of belief, according to Lin; but this is to be expected, since belief is a normative notion, while Spinoza is a naturalist. Maria van der Schaar compares the views of Locke and Gottfried Wilhelm Leibniz on judgment in her contribution to this volume. Each reacted to Descartes’s first-person epistemological method in his own way. Locke took knowledge to involve an infallible perception of the agreement or disagreement of our ideas, and he introduced judgment as a fallible act that deals in probabilities, not certainties. One can tell from the first-person perspective that one knows, when one does, according to Locke; but only God can discern, from a third-person point of view, whether one’s judgment is correct, and even one’s justification in such cases must be explicable to others. By contrast, Leibniz allows that we can have knowledge of probabilities, and our justification in such cases may be available only to ourselves; but this first-person approach is moderated by an appeal to a universal language, which reveals the natural order of explanation for truths and not merely the temporal order of their discovery. Accordingly, when the reason for our judgment is the reason for the truth of what is judged, that truth is first-personally available. The paper concludes with a fascinating discussion of particles and their role in logic. As Alexandra Newton explains, a number of scholars maintain that there are two conceptions of judgment at play in Kant: first, there is a forceless act—roughly, that of thinking a propositional content; and second, a forceful act—that of asserting it to be true. Both acts are relational in character; the representations involved are distinct from the acts of thinking them and from the individuals that think them. But Newton argues that on this relational approach, logic is infected with psychology, and the ground of truth of a judgment ends up being external to its content. Newton herself holds instead that Kant employs a single, forceful notion of judgment—that of “a synthetic unity . . . that is constitutively self-conscious”. On this synthetic approach, the ‘I’ that thinks is simply “the capacity for intellectual synthesis shared by any rational thinker”, and psychologism is avoided. Moreover, the predication that occurs in thinking is the same as that involved in its truth—or as Newton herself puts it, “Thinking and being are the same”; accordingly, “all truth is internal to the self-consciousness of judgment”—even when the predicate concept is not (antecedently) contained within the subject concept.

14

Brian Ball

Paul Redding examines Hegel’s accounts of tensed and modal judgments. His discussion offers fascinating insights into Hegel’s conception of human self-consciousness. If we consider a simple judgment like this rose is fragrant, that judgment may be correct now but incorrect a fortnight hence. Its content is not true simpliciter; it is only true of a time. Such tensed judgments are therefore sensitive to a feature of the context in which they occur. On Hegel’s view, however, in order to be a proper judge, one must be able to decontextualize; for instance, one must be able to assess an earlier judgment from the perspective of a later time, and this requires consciousness of oneself as the same at those two times. When we turn to, for example, the rose is a plant, it, too, can be thought of as true only relative to a feature of the context—roughly, a possible world; “less dramatically”, it is true only relative to a systematic set of inferentially connected beliefs about the world that might be held by another person, and grasping this requires a distinctively social kind of self-consciousness. In this way, Redding suggests, Hegel offers an alternative modal metaphysics to both Leibnizian possibilism and Aristotelian actualism. According to Mark Siebel, Bolzano opposed Kant and anticipated Frege’s anti-psychologism, yet was unrealistically optimistic in his treatment of actual thinkers. He distinguished acts of judging from their propositional contents: “a judgment in itself is a sentence in itself [i.e. proposition] that is the matter [i.e. content] of a judgmental act”. But such ‘objective sentences’ are not merely the possible contents of judgments since some cannot be judged, being evidently false; indeed, neither the concept of an objective sentence nor that of an act of judgment can be analysed for Bolzano. Believing, however, is (roughly) being disposed to judge. Mere thinking is distinguished from judging in terms of content, not just the relation the subject bears to it. And our unmediated judgments are infallible (though we may err about which these are), while our mediated judgments are inferred from premises that support them. Given Bolzano’s anti-psychologism (about support), this last claim is not true by definition but is rather an optimistic commitment regarding our mental lives—as is the contention that the degree of confidence of a judgment is (psychologically) determined by the degree to which it is supported by the judgments it is inferred from and their degrees of confidence. Mark Textor argues that Brentano advocated an appealing ‘correctness first’ view. The general notion of the correctness of an action is explanatorily prior to the specific notion of truth in connection with the mental act of judgment on this view: something is good if and only if it is correct to love it. Similarly, a true content is one that it is correct to judge. Brentano also aimed to give an account of our acquisition of the concept of truth from our awareness of correct judging. Textor thinks an account along these lines might be made to work: we can be aware of the correctness of our judgment in a certain special class of self-evident judgments (which is not to be understood in terms of self-evident truth), and we can then

Introduction

15

apply the notion more broadly to judgments in general. The paper concludes with some reasons to think that, acquisition aside, correctness is indeed first in the order of explanation. This accounts for the evaluative nature of truth, and given that objects (such as Fido) as well as states of affairs (such as Fido’s being a dog) can be acknowledged in judgment, this explains the logical (roughly, redundant, or non-determining) character of both truth and existence. Simon Blackburn defends a form of sentimentalism (as opposed to rationalism) in his piece: there can be reasons for feeling certain ways, and feelings can themselves be reasons for action or judgment; thus, reasons being normative, so, too, are feelings. Blackburn is careful to distinguish the aesthetic responses and moral sentiments he has in mind from sensations, such as pain: the latter do not “involve our awareness of things outside ourselves” but are merely caused, though even a headache can, he thinks, be a reason (e.g. to decline a party invitation). He finds support for his view in a number of writers, not least Brentano, for whom the normativity of feelings is phenomenologically given. If one experiences distress at hearing a child crying, for example, one equally apprehends the appropriateness of feeling this way; and, says Blackburn, one feels a certain aversion towards those who lack this response. As this already suggests, Blackburn finds the source of this normativity in “our nature as social animals”; in our sentimental responses to things, we incur the risk of failing to find “fellow feeling” with others and even with our future or possible selves who have “expanded [our] current sensibilities”. Peter Simons expounds and assesses Twardowski’s account of judgment. Twardowski maintained that judgments involve only a single idea and are (positive or negative) existential in form, and he distinguished the content of an idea from its object. Twardowski also held that there are non-existent objects. This allowed him to account for (true) negative existentials, but Simons thinks he would have done better to hold that these are (higher-order) judgments to the effect that the content of the idea in question is objectless. Simons admits, however, that Twardowski’s own position affords a theory of truthmakers: there are objects whose existence necessitates the truth of corresponding judgments. Twardowski is also commended for maintaining that the truth of a judgment is an absolute—not a relative—matter, though Simons thinks he mistakenly locates elements of the context in which a judgment takes place within its content. Finally, Simons is sceptical about the application of the actionproduct distinction to judgments: he finds it unclear what the relation could be between the product and the content of an act of judging, and he suggests that Twardowski misattributes the causal powers in this vicinity. Ultimately, Simons’s view is that “Twardowski’s account of judgment . . . fall[s] below the level of development of his account of ideas”. Attitudinal objects, for Friederike Moltmann, are “concrete, agentdependent entities that come with truth or satisfaction conditions”, such

16

Brian Ball

as judgments, beliefs, claims, and promises. Moltmann provides considerable linguistic evidence for their existence: they figure as referents of nominalizations like ‘John’s promise’; kinds of them (individuated by content similarity) are picked out by such expressions as ‘the judgment that God exists’, and they or their kinds are quantified over in, for example, ‘John believes something surprising’. She rejects act-object accounts of the phenomena: they can’t readily account for co-predication, as in ‘John heard Mary’s false remark’, since the act of remarking cannot be false (having no truth conditions), and its alleged propositional object cannot be heard (being abstract); and a part of John’s claim is neither a (temporal) part of (the action or event of) his claiming nor a part of any proposition. Attitudinal objects are not all Twardowskian (non-enduring) products (construed as artefacts), according to Moltmann, since intentions and judgments, for example, are not the results of any actions; nor are any of them states since a believing, for instance, has its time of occurrence essentially, while the belief had does not. These agent-dependent entities, with their causal powers and content-based properties, force us to reconsider our ontological assumptions about concreta in general. Daniel Morgan defends a “naïve” idea about intentionality: a judgment is about an object if, and only if, that object figures in its truth conditions. Some authors have suggested that a child’s judgment that it is 4 pm, for example, is about the time of day, while it merely concerns the time zone she inhabits. Yet Morgan argues that any such about/concerns distinction must be linked to a variety of phenomena in order to be explanatory of any: after all the hypothesis of a dormitive power is of no theoretical use; it does not explain the production of sleep but only labels its occurrence. One might think that some objects figuring in the truth conditions of a judgment are contributed by its content and others by the act itself, but this would not explain why only some of the commitments we incur in judging are immune to error—assuming the idea of such ‘innocent commitments’ to be intelligible; nor is it plausible to suggest that ‘parochial’ elements of truth conditions which the subject cannot vary in her judgments are contributed by the act—only a gerrymandered taxonomy of the mental could yield this result. Better to stick with the naïve view and recognize that judgment types are individuated by their truth conditions. In his contribution, Indrek Reiland aims to do two things. First, he defends an account of representation on which, when we predicate a property P of an object o, we judge that o is P. This yields a Twardowskian product that represents o as P. Propositions (if needed) can be regarded as product-types that exactly match in representational content. Reiland holds that predication is forceful; only in this way, he thinks, can we account for the fact that the judgment is correct if and only if o is P. And he maintains that when we judge, for example, that o is not P, we grasp the type of judgment product whose content is (the proposition) that o is P and predicate negation of it. Second, Reiland distinguishes the above

Introduction

17

minimal notion of judging, which he calls “semantic”, from a more robust “epistemic” judging and maintains that these are acts at different “levels”; that the latter, but not the former, is governed by a subjective norm requiring evidence for correctness, and that the former, but not the latter, can occur subpersonally. He closes by showing how certain debates surrounding the relationship between belief and judgment can be illuminated by this distinction. Christopher Peacocke, in his piece, discusses a puzzle surrounding the question of how it can be rational to draw an inference on the basis of one’s knowledge of a primitive logical principle, or indeed to judge that a logical axiom is true. He points out that the inference or judgment in question cannot be rationalized inferentially, or else the principles in question would not be primitive; and he argues that they are not self-justifying either. Instead, he suggests that “[a]n intellectual seeming that the axiom or principle is valid plays a crucial role in the thinker’s mental action of judging that the axiom or principle is valid”, that this intellectual seeming is based on the tacit knowledge that constitutes the subject’s understanding of the content in question and the concepts it involves, and that it serves to rationalize, or justify, the judgment—and so the inference made, if any. Peacocke notes that his puzzle arises in particularly “sharp form” for Frege, who held that the most basic laws of logic neither need nor admit of proof, but who nevertheless offered what appear to be arguments in support of them when introducing them; and he contends that his preferred response to the puzzle vindicates Frege’s practice.

Notes 1. Descartes, Locke, Spinoza, and Leibniz all have accounts of judgment that figure prominently in their work; Kant made it explicitly the subject of one of his Critiques, and Hegel also engaged the issue. Likewise, Bolzano, Brentano, and Frege discuss judgment at length, as does Twardowski. These theorists’ work are all discussed in this volume. 2. Ryle argued against the very idea of mental action—as some take judgment to be (see below)—on the grounds that it leads to regress. Suppose A is some overt, non-mental action, and that what makes it an action is its being caused by an intention to A. Then, if intending to A is also an action, it must be accompanied (and caused) by an intention to intend to A. And so on ad infinitum. 3. Judgment has often been investigated from a first-person perspective, using broadly phenomenological methods, but the semantics of attitude ascriptions is primarily investigated from a third-person point of view, using logical or analytical methods. (Note that I am here exploiting the first vs third-person distinction to roughly opposite effect as in van der Schaar’s contribution to this volume.) 4. Candlish (2007: 75–7) suggests some clues. Russell was enormously influential in philosophy in the twentieth century. Once he gave up on his multiple relation theory of judgment (see below) in 1919, he (1) took the objects of the attitudes to be symbols (i.e. interpreted sentences) and (2) was increasingly inclined towards behaviourism, understanding the attitudes in terms of their causal roles. This is consonant with the hypotheses floated above.

18

Brian Ball

5. Two recent debates in particular have seen the notion return to prominence. The first (see, for example, Boyle 2011; Moran 2001; A. Peacocke 2017; C. Peacocke 2008; Shah and Velleman 2005; Silins 2012) surrounds the nature of self-knowledge, and in particular the so-called ‘transparency’ of belief (cf. Evans 1982) and, to some extent, other mental states (Byrne 2011; Moore 1903). The issue here is that of explaining the ‘peculiar and privileged access’ (Byrne 2005) we have to our own mental states in general and to our beliefs in particular. One suggestion has been that we can make progress on this by considering how we know what we are doing when we act (cf. Anscombe 1957), and accordingly, the act of judgment and its relation to belief have come under investigation. The second debate is quite different in character. It derives from a perceived failure to have resolved an issue with which Russell (1903) grappled, namely that of the unity of the proposition (Linsky 1992)— and, more generally, to develop an adequate account of the metaphysics of these objects of our attitudes. Recently, a number of authors (e.g. Hanks 2015; King 2007; Moltmann 2013a; Soames 2010) have looked to illuminate this issue by appealing to the relation between these objects and certain associated acts, including judgment in particular (see also Moltmann and Textor 2017). Some of the issues and findings from these debates will be discussed below, along with those of two recent book-length treatments of the subject (Sosa 2015) and its history (Martin 2006). Finally, I should also mention a pair of recent collections of essays on related matters, namely those of Textor (2013) and Schaar (2013). 6. MacLachlan is here describing Bradley’s view of logic, which we will touch on below, but the point applies more generally, as we shall see presently. 7. One caveat: as I hinted at the beginning of this paragraph, the historical engagement of this volume is limited to the modern period. There are two things to say about this. First, much contemporary philosophy engages only with its very recent history, dating back to Russell and Frege—yet there is much to be learned from both early and late modern theorists that bears directly on current debates. In this respect, my co-editor and I are encouraging increased historical engagement. Second, a single volume can only do so much. The modern debates are, in many respects, easier for contemporary theorists to get to grips with and learn from than ancient and medieval ones, and we have accordingly restricted our attention to them. Nevertheless, I will say a few words presently by way of orientation. 8. Though, of course, it has not been universally accepted. Reid, himself, despite speaking favourably about this proposal, in fact takes the view that the notion of judgment is neither susceptible nor in need of definition—it is primitive. And Descartes and others in the early modern period also took an alternative view (cf. Marušić 2017). 9. Though, as van der Schaar argues (this volume), Leibniz objected to Locke for failing to order items of knowledge appropriately. 10. A key issue was, and is, which of these various mental operations, or acts, is explanatorily primary. 11. As Candlish and Basile note, “Bradley both inherits and transforms this tradition [in Principles], keeping the three-part format but devoting the first to Judgment and both second and third parts to Inference [or reasoning], thus dropping the separate treatment of Conception” (2017: section 5). 12. It is a short step from here to Wittgenstein’s claim that “the world is all that is the case . . . the totality of facts, not of things” (1921/1974: 5). 13. Not necessarily intentionally—see below for further discussion. 14. It is in roughly this sense that Geach (1957) uses the phrase. 15. Thus, very roughly, on his view seeing (for example) is (a kind of) believing.

Introduction

19

16. Shah and Velleman (2005) argue, by contrast, that the fact that belief is subject to a truth norm is not to be explained in terms of its aiming at the truth. 17. It is this that has been held to give rise to explain our inability, discussed above, to simply believe at will (cf. Williams 1970). 18. Locke, it seems, held that judgment can only yield belief, not knowledge (see Schaar [this volume]). Ordinary language considerations, however, seem to me to repudiate this suggestion. 19. It is a little unclear exactly what this amounts to. One hypothesis is that it means that judgment is both a personal, rather than subpersonal, level act (cf. Dennett 1969) and under voluntary control. This raises further avenues of exploration in connection with the question of whether judgment is a mental action, but due to limitations of space, they will be left to one side here. 20. Semanticists have found it useful to regard questions as sets of propositions— namely, those that answer the question (cf. Hanks 2007: 147–8); thus, if propositions are modelled as sets of possible worlds, for example (see below), questions are treated as sets of sets of possible worlds. These (former) sets are (typically taken to be) mutually exclusive and exhaustive (so that no world belongs to more than one set and every world belongs to at least one); accordingly, if the question admits of a yes or no answer (as the question whether p does), it partitions the set of all worlds into two subsets (those in which p and those in which not p). 21. Perhaps what this disagreement between, for example, Descartes and Reid shows is that the notion of judgment is not univocal. Reiland (this volume) suggests that there are two notions of judgment (one semantic, the other epistemic). 22. One might compare here Williams’s (1970) discussion of the relation of belief to assertion. Williams holds that the most straightforward expression of a belief that p is an assertion that p. But he allows that non-linguistic animals have beliefs in “a somewhat conventionalized sense” (1970: 140), and he denies that asserting that p is either necessary or—and he especially emphasizes this—sufficient for believing that p, stressing that one can assert insincerely. And it is, perhaps surprisingly, here that he finds a need for the will in connection with belief. 23. Williams (1970) thinks that the notion of belief can be illuminated through its relation to assertion (see previous note), but he explicitly acknowledges that non-linguistic animals can have beliefs. Whether someone sympathetic to this view should extend this line of thought to the case of judgment is, of course, another matter. 24. A Wittgensteinian proposition might be regarded as a yes-no question (in the sense described above of a set containing two sets of worlds), together with a sense (specifying which of these is the agreement set). 25. Indeed, the view can be traced to Aristotle’s term logic, as we have seen. 26. More specifically (and precisely), he argues that it is in order to resolve a crisis in the theory of judgment as a synthetic act engendered by Kant’s own account of existential judgments that, after a series of innovations due to less well-known figures, Brentano arrives at the view that judgment is thetic. 27. Given his view of propositions (see below), perhaps Bolzano can profitably be regarded as accepting the third possibility, that judgment is an analytic act—one in which ideas in themselves are discerned as constituent parts of propositions in themselves. Of course, there is nothing to preclude combining the Bolzanian theory of propositions with the view that it is in some act distinct from judgment in which one discerns their parts. Equally, it seems worth exploring whether Leibniz held such a view. After all, the distinction between analytic and synthetic truths, for Leibniz, concerns whether the predicate

20

28.

29.

30.

31.

32.

33.

34.

35. 36. 37.

Brian Ball concept can be revealed to be contained within the subject concept after finitely or infinitely many steps. But I do not mean to put forward positive interpretive claims here; I am only suggesting avenues for future research. This issue about the nature of judgment seems especially worthy of consideration in the current climate given that virtually all of the reaction to the (thetic) orthodoxy in recent years has been to suggest that judgment—or at least predication—is a synthetic act. The possibility that the metaphysics of propositions might be illuminated by an alternative way of looking at such acts therefore presents itself. It strikes me as too strong to hold, as Kriegel appears to do, that it is merely an accidental matter of contingent fact concerning those intentional acts and states that are in fact object-directed that they are so. A more moderate position that would nevertheless allow for some such acts and states to fail to be objectdirected would maintain that the constitutive relationship between such acts and states, on the one hand, and objects, on the other, is normative: intentional acts and states are those that (essentially) ought (in some sense) to have objects. To the best of my knowledge, this view has not been explored in the literature. The advocate of such a view would need an account of the intrinsic natures of the families of acts and states the objects in question (p, q, etc.) serve to measure, just as advocates of the analogous view have independent accounts of the intrinsic nature of mass and temperature properties. (Thanks to Chris Peacocke on this point.) This kind of view might be bolstered by a sententialist account of the logical form of attitude reports, according to which ‘complement clauses refer to themselves’ (Higginbotham 2006), rather than the more standard propositionalist view on which they refer to propositions. Of course, this will only be so if one does not regard the attitudes as relations to such (semantically interpreted) sentential clauses. Cresswell and von Stechow (1982) argue that that-clauses pick out sequences of entities, one a (zero or more place) universal, the others (if any) entities of appropriate type to instantiate that universal; but the word ‘that’ is, in effect, variably polyadic on their view—or as they say, ‘systematically ambiguous’. Their view accordingly occupies a kind of middle ground between act-object and multiple relation views. In these works, Moltmann took the product of a state to be identical with the state itself. In her paper in this volume, however, Moltmann modifies her view further, arguing that attitudinal objects are never states, and not all of them can be Twardowskian products either. She retains the basic proposal regarding logical form. Lewis (1979), by contrast, thought that the objects of the attitudes in general are properties (some but not all of which are propositional). Chierchia (1989) accordingly suggested that the logical forms of at least some reports (those of attitudes de se) involve the explicit representation of properties, rather than propositions, as the objects of the relevant attitude verbs. That said, even if one thinks that some attitudes are to be understood in terms of relations to properties, one need not abandon the idea that the objects of the attitudes themselves are propositional, as the approach of Cresswell and von Stechow (1982) makes clear. As Frege (1918/1956) effectively noted in his criticism of the correspondence theory of truth. This worry, raised by Williams against Descartes, might equally be pressed against Hume’s theory of belief. For what it is worth, the former interpretation strikes me as more plausible given the influence of Aristotle on Brentano and the fact that for Aristotle

Introduction

38. 39.

40.

41.

42.

43. 44. 45.

21

being is said in many ways. This opens up the possibility that immanent being is distinct from formal being, without the thing that is, in these two different ways, itself differing, i.e. being mind-dependent in one case and mindindependent in the other. We might compare, in this respect, the medieval distinction between formal and objective reality that is at play in Descartes. As Betti notes, however, the distinction “was not new in Twardowski’s time” (2017: section 2.1). Indeed, Siebel (this volume) suggests that Bolzano had already made it. Suppose I am presented with an apple in perception: its redness is part of the content of the presentation, but its being a fruit is not—though it is an essential part of the object. An advocate of two-dimensional semantics might therefore argue that there are two intensions associated with the presentation: first, one that is determined by the content and so yields a red thing (though not necessarily a fruit) at each world; and second, one that is determined by the object and which yields, at each world, a thing having all of the essential properties of that object, including that of being a fruit. Modal issues, however, were less discussed in the late nineteenth century than they are now. As for Twardowski’s view of the contents and objects of judgments, see Simons (this volume). I have been following Moore (1901–02) in speaking of what is believed or judged as the object of belief or judgment. Many, however, take these propositions to be the contents of such acts and attitudes—for instance, Grzankowski (2016), who reserves the term ‘object’ for the individuals (if any) a given mental state concerns. Below, for ease of exposition, I use the term ‘content’ where any author I discuss does so. What are possible worlds? According to David Lewis (1973, 1986), they are spatio-temporally maximal concrete objects; that is, the actual world is a thing, just like us, only bigger, including whatever stands in any space-time relation to us whatsoever, and other possible worlds are similar in kind. Such worlds are so specific, or complete, that they decide the truth value of any given proposition. If one of them contains, for example, (only) red snow, then the proposition that snow is red is true in that world, and it is false there otherwise; thus, on the Tractarian view, a proposition (such as the proposition that snow is red) can be simply identified with the set of those worlds in which it is true (i.e. the set of worlds containing red snow). On an alternative conception of possible worlds via Robert Stalnaker (1976, 1984), they are not concrete objects but abstract properties, or universals; in particular, they are ways the world as a whole might have been. Accordingly, non-actual possible worlds can exist, without, for example, any red snow existing (though not at any space-time distance from us). This view of worlds is preferred by many who are inclined to respond to Lewis’s view that other concrete universes exist with an ‘incredulous stare’ (Lewis 1986: 135), but it does not fundamentally alter the argument for identifying propositions with sets of worlds; and accordingly, those tempted by the Tractarian conception of propositions may wish to endorse it. Though see Merricks (2015) and Speaks (2014) for two interesting recent defences, each involving intriguing (but different) departures from the standard accounts discussed above. Unfortunately, limitations of space preclude discussion of these works. That said, Merricks (2015) defends the orthodox account of propositions; but, interestingly, his (2009) paper argues that while belief is a propositional attitude, desire (for example) is not, and its object is not truth-apt. This is what Hanks calls the ‘constitutive’ version of the distinction. I set aside the ‘taxonomic’ version. I am here using these terms interchangeably.

22

Brian Ball

References Anscombe, G. E. M. (1957), Intention. Oxford: Blackwell. Arnauld, A. and P. Nicole (1662/1996), Logic or the Art of Thinking, trans. and ed. J. V. Buroker. Cambridge: Cambridge University Press. (Commonly known as the Port-Royal Logic.) Ball, B. (2011), “Review: What Is Meaning?”, The Canadian Journal of Philosophy 41(4): 485–503. Betti, A. (2017), “Kazimierz Twardowski”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, URL = https://plato.stanford.edu/archives/fall2017/ entries/twardowski/. Bolzano, B. (1837/1973), Theory of Science, ed. J. Berg. Dordrecht: D. Reidel. Boyle, M. (2009), “Active Belief”, The Canadian Journal of Philosophy 39(sup. 1): 119–47. Boyle, M. (2011), “Transparent Self-Knowledge”, Proceedings of the Aristotelian Society 85(sup.): 223–41. Bradley, F. H. (1883), The Principles of Logic. London: Kegan Paul. Brentano, F. (1874/1995), Psychology From an Empirical Standpoint, trans. L. McAlister, A. Rancurello, and D. Terrell, ed. L. McAlister. London: Routledge. Buroker, J. (2017), “Port Royal Logic”, in E. N Zalta (ed.), The Stanford Encyclopedia of Philosophy, URL = https://plato.stanford.edu/archives/spr2017/ entries/port-royal-logic/. Byrne, A. (2005), “Introspection”, Philosophical Topics 33(1): 79–104. Byrne, A. (2011), “Transparency, Belief, Intention”, Proceedings of the Aristotelian Society 85(sup.): 201–21. Candlish, S. (2007), The Russell/Bradley Dispute and its Significance for Twentieth Century Philosophy. Basingstoke: Palgrave MacMillan. Candlish, S. and P. Basile (2017), “Francis Herbert Bradley”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, URL = https://plato.stanford.edu/ archives/spr2017/entries/bradley/. Chalmers, D. (2002), “The Components of Content”, in D. Chalmers (ed.), Philosophy of Mind: Classical and Contemporary Readings. Oxford: Oxford University Press. Chalmers, D. (2006), “The Foundations of Two-Dimensional Semantics”, in M. Garcia-Carpintero and J. Macia (eds.), Two-Dimensional Semantics: Foundations and Applications. Oxford: Oxford University Press. Chierchia, G. (1989), “Anaphora and Attitudes De Se”, in R. Bartsch, J. van Benthem, and P. van Emde Boas (eds.), Semantics and Contextual Expression. Dordrecht: Fortis. Cresswell, M. and A. von Stechow (1982), “De Re Belief Generalized”, Linguistics and Philosophy 5: 503–35. Davidson, D. (1967), “The Logical Form of Action Sentences”, in N. Rescher (ed.), The Logic of Decision and Action. Pittsburgh: University of Pittsburgh Press. Dennett, D. (1969), Content and Consciousness, London: Routledge and Kegan Paul. Descartes, R. (1642), “Meditations on First Philosophy”, in J. Cottingham, R. Stoothoff, and D. Murdoch (eds.), The Philosophical Writings of Descartes. Cambridge: Cambridge University Press, 1984–91.

Introduction

23

Evans, G. (1982), The Varieties of Reference, ed. J. McDowell. Oxford: Clarendon. Field, H. (1980), Postscript to his “Mental Representation”, in N. Block (ed.), Readings in the Philosophy of Psychology. Cambridge, MA: Harvard University Press. Frege, G. (1892/1948), “Sense and Reference”, Philosophical Review 57(3): 209–30. Frege, G. (1918/1956), “The Thought: A Logical Inquiry”, trans. A. M. and M. Quinton, Mind 65(259): 289–311. Friedman, J. (2013), “Question-Directed Attitudes”, Philosophical Perspectives 27(1): 145–74. Geach, P. (1957), Mental Acts: Their Content and Their Objects. London: Routledge and Kegan Paul. Geach, P. (1965), “Assertion”, The Philosophical Review 74(4): 449–65. Grzankowski, A. (2016), “Attitudes Towards Objects”, Noûs 50(2): 314–28. Hanks, P. (2007), “The Content-Force Distinction”, Philosophical Studies 134: 141–64. Hanks, P. (2015), Propositional Content. Oxford: Oxford University Press. Hanna, R. (2016), “Kant’s Theory of Judgment”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, URL = https://plato.stanford.edu/archives/ win2016/entries/kant-judgment/. Higginbotham, J. (2006), “Sententialism: The Thesis that Complement Clauses Refer to Themselves”, Philosophical Issues 16: 101–19. Huemer, W. (2017), “Franz Brentano”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, URL = https://plato.stanford.edu/archives/fall2017/ entries/brentano/. Hume, D., D. F. Norton and M. J. Norton (eds.) (1740), A Treatise of Human Nature. Oxford: Clarendon, 2007. Husserl, E. (1903), “Review of Der Streit der Psychologisten und Formalisten in der modernen Logik, by Melchior Palágyi”, in Aufsätze und Rezensionen (1890–1910). Husserliana 22. Dordrecht: Reidel 1979, 152–62. King, J. (1994), “Can Propositions Be Naturalistically Acceptable?”, Midwest Studies in Philosophy 19(1): 53–75. King, J. (2007), The Nature and Structure of Content. Oxford: Oxford University Press. Kriegel, U. (2008), “The Dispensability of (Merely) Intentional Objects”, Philosophical Studies 141: 79–95. Lewis, D. K. (1973), Counterfactuals. Oxford: Blackwell. Lewis, D. K. (1979), “Attitudes De Dicto and De Se”, Philosophical Review 88(4): 513–43. Lewis, D. K. (1986), On the Plurality of Worlds. Oxford: Blackwell. Linsky, L. (1992), “The Unity of the Proposition”, Journal of the History of Philosophy 30: 243–73. Locke, J. (1690/1975), An Essay Concerning Human Understanding, ed. P. Nidditch. Oxford: Oxford University Press. MacLachlan, D. L. C. (2018), “F. H. Bradley: Logic”, in The Internet Encyclopedia of Philosophy, ISSN 2161-0002, https://www.iep.utm.edu/brad-log/ Martin, W. (2006), Theories of Judgment: Psychology, Logic, Phenomenology. Cambridge: Cambridge University Press.

24

Brian Ball

Marušić, J. (2017), “Judgment and Belief in Early Modern Philosophy”, in Routledge Encyclopedia of Philosophy, Taylor and Francis, URL = www.rep.routledge. com/articles/thematic/judgment-and-belief-in-early-modern-philosophy/v-1/, doi:10.4324/0123456789-DA090-1. Matthews, R. (1994), “The Measure of Mind”, Mind 103(410): 131–46. Matthews, R. (2007), The Measure of Mind: Propositional Attitudes and Their Attribution. Oxford: Oxford University Press. Meinong, A. (1904/1960), “The Theory of Objects”, trans. R. Chisholm, I. Levi, and B. Terrell, in R. Chisholm (ed.), Realism and the Background of Phenomenology. Glencoe: Free Press. Merricks, T. (2009), “Propositional Attitudes?”, Proceedings of the Aristotelian Society 109(3): 207–32. Merricks, T. (2015), Propositions. Oxford: Oxford University Press. Moltmann, F. (2013a), Abstract Objects and the Semantics of Natural Language. Oxford: Oxford University Press. Moltmann, F. (2013b), “Theoretical Alternatives to Propositions: Propositions, Attitudinal Objects, and the Distinction Between Actions and Products”, Canadian Journal of Philosophy 43(5–6): 679–701. Moltmann, F. (2017), “Attitude Reports, Cognitive Products, and Attitudinal Objects: A Response to G. Felappi On Product-Based Accounts of Attitudes”, Thought 6: 3–12. Moltmann, F. and M. Textor, eds. (2017), Act Based Conceptions of Propositional Content. Oxford: Oxford University Press. Moore, G. E. (1899), “The Nature of Judgment”, in his Selected Writings, ed. T. Baldwin. London: Routledge, 1993. Moore, G. E. (1901–02), “Truth and Falsity”, in his Selected Writings , ed. T. Baldwin. London: Routledge, 1993. Moore, G. E. (1903), “The Refutation of Idealism”, Mind 7: 1–30. Moran, R. (2001), Authority and Estrangement: An Essay on Self-Knowledge. Princeton: Princeton University Press. Nuchelmans, G. (1973), Theories of the Proposition: Ancient and Medieval Conceptions of the Bearers of Truth and Falsity. Amsterdam: North Holland Publishing Company. Nuchelmans, G. (1984), Judgment and Proposition: From Descartes to Kant. Amsterdam: North Holland Publishing Company. Peacocke, A. (2017), “Embedded Mental Action in Self-Attribution of Belief”, Philosophical Studies 174: 353–77. Peacocke, C. (2008), Truly Understood. Oxford: Oxford University Press. Peacocke, C. (2015), “Magnitudes: Metaphysics, Explanation, and Perception”, in D. Moyal-Sharrock and V. Munz (eds.), Mind, Language, and Action: Proceedings of the 2013 Kirchberg Symposium. Berlin: de Gruyter. Perala, M. (2013), “Ancient Theories of Judgment”, in S. Knuuttila and J. Sihvola (eds.), Sourcebook for the History of the Philosophy of Mind: Philosophical Psychology from Plato to Kant. Dordrecht: Springer. Proust, J. (2001), “A Plea for Mental Acts”, Synthese 129: 105–28. Proust, J. (2010), “Mental Acts”, in T. O’Connor and C. Sandis (eds.), A Companion to the Philosophy of Action. Oxford: Wiley-Blackwell. Quine, W. V. O. (1960), Word and Object. Cambridge, MA: MIT Press. Reid, T. (1785/1983), Inquiry and Essays, ed. R. E. Beanblossom and K. Lehrer. Indianapolis: Hackett.

Introduction

25

Reiland, I. (this volume), “Predication and Two Concepts of Judgment”. Rorty, R. (1967), The Linguistic Turn. Chicago: Chicago University Press. Russell, B. (1903), Principles of Mathematics. Cambridge: Cambridge University Press. Russell, B. (1904/1980), “Russell to Frege: December 12, 1904”, in G. Gabriel, H. Hermes, F. Kambartel, C. Thiel, A. Veraart, B. McGuinness, and H. Kaal (eds.), Frege: Philosophical and Mathematical Correspondence. Oxford: Blackwell. Russell, B. (1910), “On the Nature of Truth and Falsehood”, in his Philosophical Essays. London: Longmans, Green. Russell, B. (1912), The Problems of Philosophy. London: Williams & Norgate. Russell, B. (1913/1984), “Theory of Knowledge: The 1913 Manuscript”, in The Collected Papers of Bertrand Russell , Vol. 7, ed. E. Ramsden Eames and K. Blackwell. London: Allen & Unwin. Ryle, G. (1949), The Concept of Mind. London: Hutchinson and Co. Schaar, M. van der, ed. (2013), Judgment and the Epistemic Foundation of Logic. Dordrecht: Springer. Schaar, M. van der (this volume), “Locke and Leibniz on Judgment: The FirstPerson Perspective and the Danger of Psychologism”. Sebestik, J. (2016), “Bolzano’s Logic”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy, URL = https://plato.stanford.edu/archives/fall2017/ entries/bolzano-logic/. Shah, N. and J. D. Velleman (2005), “Doxastic Deliberation”, Philosophical Review 114(4): 497–534. Siebel, M. (this volume), “Bolzano’s Theory of Judgment”. Silins, N. (2012), “Judgment as a Guide to Belief”, in D. Smithies and D. Stoljar (eds.), Introspection and Consciousness. Oxford: Oxford University Press. Simons, P. (this volume), “Twardowski on Judgment”. Skinner, B. F. (1953), Science and Human Behavior. New York: MacMillan. Soames, S. (2010), What Is Meaning? Princeton: Princeton University Press. Sosa, E. (2015), Judgment and Agency. Oxford: Oxford University Press. Speaks, J. (2014), “Propositions Are Properties of Everything or Nothing”, in J. King, S. Soames, and J. Speaks (eds.), New Thinking About Propositions. Oxford: Oxford University Press. Stalnaker, R. (1976), “Possible Worlds”, Noûs 10(1): 65–75. Stalnaker, R. (1984), Inquiry. Cambridge, MA: MIT Press. Textor, M. (2013), Judgement and Truth in Early Analytic Philosophy and Phenomenology. Basingstoke: Palgrave. Twardowski, K. (1894/1977), On the Content and Object of Presentations: A Psychological Study, trans. R. Grossmann. The Hague: Nijhoff. Twardowski, K. (1912/1999), “Actions and Products”, trans. A. Szylewski, in J. Brandl and J. Wolenski (eds.), On Actions, Products, and Other Topics in Philosophy. Amsterdam: Rodopi. Williams, B. (1970), “Deciding to Believe”, in his Problems of the Self. Cambridge: Cambridge University Press. Williams, B. (1978), Descartes: The Project of Pure Inquiry. London: Harvester Press. Wittgenstein, L. (1914–1916/1979), Notebooks, 1914–1916, trans. G. E. M. Anscombe, 2nd ed. Oxford: Blackwell. Wittgenstein, L. (1921/1974), Tractatus Logico-Philosophicus, trans. D. F. Pears and B. F. McGuinness, 2nd ed. London: Routledge.

1

Affirmation, Judgment, and Epistemic Theodicy in Descartes and Spinoza Martin Lin

Both Descartes and Spinoza develop their theories of judgment in the service of what may be called epistemic theodicy, in that they seek to reconcile human error with the existence of a perfect God. They have, however, very different conceptions of God and the relationship of the human mind to that God and, therefore, develop very different conceptions of judgment. Descartes’s God is a transcendent being in whose image the human mind has been made and whose commandments we seek to fulfil. Human error is compatible with the existence of a perfect God because we are free to use our cognitive capacities well or poorly. In particular, judgment (affirming, denying, or suspending judgment with respect to the content of an idea) is a matter of a free act of will. God has given us a faculty that allows us to discover a rule for applying our free will in judgment so that we will never affirm a false proposition and that allows us to unerringly affirm certain truths. And, thus, when we err, the responsibility for error falls to us and not God. For Spinoza, unlike Descartes, God is not a transcendent being; rather, he is the one substance in which all else, the human mind included, inheres. Indeed, for Spinoza, the human mind is an idea in God’s infinite intellect and constitutes God’s knowledge of the human body. This entails that every idea in the human mind is identical to an idea in God. Since all ideas, insofar as they are in God, are true and, indeed, known, all ideas, in and of themselves, are affirmed. This thesis has at least two troubling consequences. First, it seems difficult to reconcile with the apparent existence of cognitive attitudes such as denial and suspending judgment. Second, it appears to entail that all ideas are believed. In this paper, I begin by setting out Descartes’s account of judgment and how it accomplishes an epistemic theodicy. I then discuss Spinoza’s official argument for the claim that every idea is affirmed and note certain difficulties with it that prompt recent interpretations according to which Spinoza’s notion of affirmation is reducible to his notion of conatus or striving for self-preservation. I argue that such interpretations must be rejected on the grounds that they fail to preserve the connection between

Affirmation, Judgment, and Epistemic Theodicy

27

affirmation and truth that Spinoza endorses. I suggest instead that, for Spinoza, every idea is affirmed in the sense that every idea purports to represent the world as it really is. Striving for self-preservation does not ground this affirmation, but it looks to it for guidance so that each idea pushes us to act in ways that would be conducive to our self-preservation if it were true. When different ideas push us in different directions, how we act is a function of the respective degrees of power associated with each idea, considered as an individual in its own right. I conclude by arguing that Spinoza’s notion of affirmation should be regarded as a technical one that is unconnected with the ordinary conception of belief, which, I argue, plays no role in Spinoza’s psychology.

1.

Descartes’s Theory

In the Third Meditation, Descartes proposes the following rule: whatever I clearly and distinctly perceive is true.1 But before he accepts this rule, he must prove that God exists and is no deceiver because he needs to foreclose the possibility that God designed me in such a way that my faculties mislead me and what I clearly and distinctly perceive is false.2 Proving the existence of such a God is the principal task of the Third Meditation. At the start of the Fourth Meditation, Descartes confronts the worry that he has proved too much. If the existence of a non-deceiving God gives us a reason to trust our clear and distinct perception, wouldn’t it also give us a reason to trust all of our perceptions?3 After all, if God is no deceiver, wouldn’t he design us in such a way that our cognitive capacities never deceived us? This question is especially troublesome to Descartes because he thinks that our sensory perception of the world provides us with ample opportunities to err: it presents the world to us as having qualities that it does not, in fact, have. It presents, for example, an apple to me as red, fragrant, and sweet when in fact it is none of those things. (At least it is none of those things in the way they are presented to me in sense perception, that is, as intrinsic properties of the apple distinct from its geometrical ones.) Rather, the only properties that the apple truly has are modes of extension: size, shape, and motion. Why then did God, who is no deceiver, design me in such a way that my sensory capacities consistently give misleading testimony? This is where his theory of judgment comes in, which is, as we will see, designed to show that we, and not God, are responsible for our errors. For Descartes, the mind is composed of two faculties or powers: intellect and volition.4 The intellect is that aspect of my mind that accounts for its ability to represent the world and various possibilities concerning it. For example, sense perception, memory, perception of universals, and essences are all ideas that a mind possesses in virtue of the intellect. The will is the aspect of my mind that accounts for my ability to take certain attitudes to the contents represented by my ideas. For example, affirming,

28

Martin Lin

denying, wanting, and fearing are all volitions that a mind possesses in virtue of the will. Merely having an idea, all by itself, cannot constitute an error, even if that idea represents what is not the case. Only if we affirm an idea that represents what is not the case or deny an idea of what is the case do we err.5 When we affirm or deny an idea, we make a judgment. Judgment, thus, for Descartes, involves the cooperation of two distinct mental faculties: intellect and volition. It is useful to contrast Descartes on this score with some of his scholastic predecessors, such as Aquinas, who thought that (apart from the special case of religious faith) judgment was an act of the intellect alone. According to this scholastic tradition, the first operation of the intellect involves cognizing accidental and substantial form.6 By themselves, the representations of these forms are neither true nor false, but when we relate them to each other by a process that Aquinas calls combining and separating, we form judgments, which are either true or false.7 For example, the first operation of the intellect may abstract from sense perception of a red apple a representation of the accidental form of red and a representation of the substantial form of an apple. A judgment results when I combine my representation of redness with my representation of an apple into the form the apple is red. There are at least two ways in which Aquinas and Descartes differ on the issue of judgment. The first difference concerns the nature of the object that receives the action. In the case of Aquinas, the objects of the action are the mental representations of the forms. They are joined together in such a way that one is predicated of the other. For Descartes, on the other hand, the object of the action is an idea, the content of which already has propositional structure. For example, in the case of thinking about a red apple, judgment for Aquinas is the process by which the concept red is combined with the concept apple. We judge of the apple that it is red. An interesting result of this difference is that Descartes’s theory of judgment allows for attitude to vary independently of content, whereas Aquinas’s does not. For example, take the idea that the apple is red. We could affirm or deny that very idea, on Descartes’s view, merely by changing the quality of our will with respect to it. But, for Aquinas, the actions of combining and separating are not directed at the bearers of propositions but at sub-propositional elements, and thus separating cannot be like judging a proposition to be false. What is more, separating cannot merely be refusing to combine. Not predicating redness of an apple is not the same as judging the proposition that the apple is red is false. A natural solution to this difficulty for Aquinas would be to hold that separating is somehow like negation. For example, it would result in a judgment that it is not the case that the apple is red. The content of this judgment is, however, different from the content of the judgment that the apple is red. Thus, for Aquinas, unlike Descartes, attitude and content cannot vary independently.

Affirmation, Judgment, and Epistemic Theodicy

29

The second point of contrast is that judgment for Aquinas is an operation of the intellect alone, whereas, for Descartes, judgment requires the cooperation of the will, the same faculty that is responsible for purely conative attitudes such as wanting and fearing. This is a curious feature of Descartes’s account. Why think that the attitudes involved in wanting that p and affirming that p are the products of the same faculty? Descartes’s motivation is revealed by the context in which he introduces his theory of judgment. As mentioned previously, Descartes is concerned that the existence of a non-deceiving God is incompatible with the fact of human error. Descartes’s strategy for absolving God of guilt for our errors is to show that (1) he did not design us imperfectly in giving us the intellect that he did; (2) he did not design us imperfectly in giving us the will that he did; and (3) it lies within our control to use our will and intellect in a way that will never result in error. Our intellect, Descartes argues, is perfect in its kind. We are, of course, finite beings, and as such there are many things about which we have no ideas. But we must, from Descartes’s perspective, distinguish between ideas that we merely lack and ideas that we ought to have but lack. If a lack of ideas is to count as a defect or privation, it would have to be the case that we should have ideas of these things. There is no reason to think, Descartes claims, that we should have ideas of everything.8 After all, a craftsman, no matter how skilled, is under no obligation to include everything in every design. Thus, this limitation is not a defect or privation. We also have ideas, especially those deriving from sense experience, that are confused and obscure. But even these involve no error so long as we do not pass judgment on them. Therefore, that we have confused and obscure ideas does not, by itself, show that the design of our intellects is to blame for our errors. Next, Descartes argues that our will, too, is perfect in its kind. Indeed, we are perfectly free because our ability to affirm, deny, pursue, and avoid is not limited in any way. Of course, the manifestation of these attitudes in action is limited by our various cognitive and physical imperfections, but the attitudes themselves are not. It is in virtue of this unlimited power of will that we understand ourselves “to bear in some way the image and likeness of God”.9 The divine will may be more efficacious than ours in virtue of God’s greater knowledge or power, but with respect to the will considered in itself, ours is as perfect as his. What remains to be shown is that God designed us in such a way that the interaction of will and intellect that results in judgment can always take place in a way that is free of error. That this is so is suggested by the rule that Descartes takes himself to have established in the Third Meditation: whatever I clearly and distinctly perceive is true. Thus, if I only judge true what I clearly and distinctly perceive and in every other case suspend judgment, then I will never fall into error. Because judgment involves the application of the will, or freedom of choice, if I judge something true

30

Martin Lin

that I did not clearly and distinctly perceive, the fault lies with me and not God.10 Descartes’s theory of judgment has been criticized on many grounds, but perhaps the most serious is his claim that judgment involves the exercise of a free will. By affirming that p, I come to believe that p. Thus, if affirmation is a free volition, then what I believe is up to me. This has struck many commentators as implausible.11 There are two main reasons for this. First, if belief were a matter of decision, then belief would be sensitive to practical reasons. Belief, however, is insensitive to practical reasons. For example, there is some amount of money that would make it prudent for me to believe that the moon was made of blue cheese, and yet there is no amount of money that, in reality, could get me to believe that the moon is made of blue cheese. It is simply not in my power to believe on such a basis. Second, how I respond to considerations that do affect belief (i.e. epistemic reasons) does not appear to be under my control. For example, if, upon reflection, I conclude that my reasons to believe that the number of grains of sand in the Sahara Desert is even are no greater or less than my reasons to believe it is odd, then it isn’t under my control whether or not to affirm or deny either proposition. Of course, I can verbally affirm or deny anything, and such an affirmation or denial can be responsive to practical considerations or unresponsive to what I take to be my epistemic reasons for belief, but such verbal affirmations or denials are not what Descartes is talking about. Rather, he is talking about the mental acts that such verbal affirmations or denials express when they are sincere. Now, in fairness to Descartes, he does allow that some reasons to believe are so powerful that I cannot but judge in accordance with them.12 Such judgments are still free for Descartes because he distinguishes freedom of indifference from freedom of spontaneity. For example, I do not experience freedom of indifference with respect to occurrent clear and distinct perception. Rather, the clarity and distinctness of that perception command my assent, just as perception of something good commands my desire. I do, however, still enjoy freedom of spontaneity because it is the nature of the will to seek the true and the good. Thus, ultimately, for Descartes, I choose what to believe. My account of Descartes’s theory of judgment has assumed that, for him, the will’s control with respect to judgment is (1) direct rather than indirect and (2) includes affirmation and denial rather than merely passing and suspending judgment. Both assumptions are controversial. With respect to (1), some commentators have claimed that the will’s influence over judgment is only indirect and is mediated by direct control over some other mental act—for example, selective attention.13 With respect to (2), some commentators have argued that the will controls not affirmation and denial but rather whether to pass or suspend judgment.14 Once that decision is made, affirmation and denial are determined by reasons for belief and not the will itself.

Affirmation, Judgment, and Epistemic Theodicy

31

My reasons for making these assumptions is that they are, in addition to being plausible readings of Descartes’s text, presupposed by Spinoza in his critique of Descartes, as we will see in the next section. Yet, as we will see, Spinoza’s ultimate reasons for rejecting Descartes’s theory of judgment are independent of these assumptions. Even if Descartes’s theory is that the will merely indirectly controls whether we pass or suspend judgment, Spinoza’s reasons for rejecting direct control over affirmation and denial generalize to any theory on which the will has unconditional control over any mental act.

2.

Spinoza on Judgment

Unlike in Descartes, there is no systematic development of a theory of judgment in Spinoza. Rather, there are a few unsystematic remarks strewn throughout part 2 of the Ethics. Not only are these remarks unsystematic, but, as we will see, they are not sufficient to explain the role that affirmation and judgment play in Spinoza’s psychology. I will argue, however, that it is possible to reconstruct a Spinozistic account of judgment from materials drawn from elsewhere in the Ethics. Let us begin by considering those remarks where Spinoza is principally concerned with criticizing Descartes’s account of judgment. He writes in 2p48: The Mind is a certain and determinate mode of thinking (by 2p11), and so (by 1p17c2) cannot be a free cause of its own actions, or cannot have an absolute faculty of willing and not willing. This amounts to a rejection of Descartes’s doxastic voluntarism, the claim that what we decide to believe is up to us in the sense that it involves the exercise of a causally unconditioned faculty of free will. Spinoza’s reasons for rejecting it stem from what we might call his naturalism. He thinks that nature is uniform and that every natural phenomenon is governed by the same laws. This includes human beings. They are not, in his words, a “kingdom within a kingdom” that “disturbs, rather than follows, the order of nature”.15 This entails that when someone affirms a proposition, their action is governed by natural laws and conditioned by causal antecedents that necessitate it. His second objection is expressed in his claim that “the will and the intellect are one and the same”.16 In other words, judgment does not involve two independent factors: the intellect, which allows us to understand or consider some content, and the will, which ultimately judges the truth or falsity of that content. Rather, it is the same faculty by which we understand and affirm a proposition. But, as Spinoza’s discussion of this claim makes clear, his thesis is stronger than that a single faculty is responsible for judgment. Rather,

32

Martin Lin

he believes the much more radical proposition that every idea is both simultaneously a representation of a proposition and the affirmation of that proposition. Moreover, an affirmation is nothing over and above the idea of the affirmed content. In short, ideas and affirmations are identical. On the face of it, this claim is highly counterintuitive. Surely we cognize propositions that we do not affirm. For example, we can entertain propositions, the truth of which we deny or doubt. Why, then, does Spinoza believe this radical thesis, and how can he deal with the phenomenon of denial and doubt? Spinoza’s official argument for the claim that every idea is an affirmation (found in 2p49d) begins by saying that we can’t have an idea of a triangle without affirming that the triangle is such that its interior angles are equal to two right angles, and that we can’t affirm this proposition without having an idea of a triangle. Thus, the affirmation of the proposition that a triangle is such that its interior angles equal two right angles is nothing over and above the idea of the triangle. Because, Spinoza claims, this example was selected at random, we can infer a universal generalization: the affirmation of any proposition is nothing over and above the idea of its subject; that is, affirmations and ideas are identical. There is much to criticize in this argument, but perhaps the most serious problem is that not every affirmation is entailed by an idea of the subject of the proposition affirmed. The claim that every affirmation is identical to some idea should not be confused with the more plausible claim that every idea entails some affirmation. This more plausible claim can be explicated with the following schema: for all x, there is some F, such that x is inconceivable unless x is F is affirmed. This would be true if, for example, we conceived of things via their essences, and doing so entails that we judge that things satisfy some essential description. Spinoza’s claim is, rather, the much more implausible claim that every affirmation is such that its subject is inconceivable without that affirmation. This is clearly false. For example, I affirm the proposition that Paris is the capital of France, but it is not true that no one could have the idea of Paris without affirming that it is the capital of France.

3.

Belief, Affirmation, and Conatus

Spinoza’s claim that every idea is an affirmation appears to entail that every idea is believed because the affirmation of a content is often thought to constitute or give rise to a belief. After all, on the face of it, I cannot affirm that Paris is the capital of France without believing that Paris is the capital of France, at least so long as I affirm it. But if every idea is an affirmation, then, it is natural to think, every idea constitutes or gives rise to a belief.17 The claim that every idea is a belief would be problematic for Spinoza for at least two reasons. First, Spinoza thinks that a mind can contain

Affirmation, Judgment, and Epistemic Theodicy

33

ideas whose contents exclude one another. For example, Marlow might have an idea that represents Ms Wade as loving her husband and a different idea that represents her as not loving her husband. What is more, according to Spinoza, if one of these ideas is stronger (I will return to the question of what it means for one idea to be stronger than another presently) than the other, then it, and not the weaker, will determine behaviour. And yet, if every idea is a belief, then the weaker idea is still a belief despite not guiding behaviour. The weaker idea is, thus, likely to strike many philosophers as failing to satisfy the functional profile of a belief. Second, as Justin Steinberg has pointed out, Spinoza denies that all non-veridical ideas involve error.18 For example, he says that the idea that the sun is two hundred feet away does not, in and of itself, constitute an error but does so only in a context where the possessor of this idea lacks a stronger idea of the true distance. Suppose someone has such a stronger idea of the true distance. If every idea were a belief, then they would believe that the sun was only two hundred feet away and not commit an error. It would be hard to make sense of Spinoza if this were his position. Diane Steinberg has proposed an account of belief in Spinoza that denies that every idea is a belief. On her interpretation, all (and only) ideas that are stronger than ideas opposed to them are beliefs, where strength is characterized in terms of the power of the conatus of that idea. In other words, S believes that p just in case S has an idea i whose content is that p and i are stronger than any idea whose content excludes p.19 This account would explain why Spinoza thinks, for example, that a person who has an idea of the sun as two hundred feet away but also has a stronger idea of the true distance of the sun does not err. The person’s behaviour and subsequent thought are determined by the idea of the true distance and not by the idea of the sun as two hundred feet away, so it is natural to say that they believe the idea of the true distance, not the idea that represents it as closer than it really is. But Steinberg’s proposal isn’t satisfactory either. If we concede that the notion of belief plays any role in Spinoza’s psychology (a point about which I’m sceptical, as I will explain presently), then a counterexample to her account can be found in Spinoza’s discussion of thoughts about the future. According to him, our ability to think about the future depends upon the association of ideas. If someone has an experience of x and y together, subsequent experiences of x will cause thoughts about y and vice versa. Similarly, if someone has experiences of first x and then y, subsequent experiences of x will cause thoughts about y. He writes: Let us suppose, then, a child, who saw Peter for the first time yesterday, in the morning, but saw Paul at noon, and Simon in the evening, and today again saw Peter in the morning. It is clear from 2p18 that as soon as he sees the morning light, he will immediately imagine the sun taking the same course through the sky as he saw on the

34

Martin Lin preceding day, or he will imagine the whole day, and Peter together with the morning, Paul with noon, and Simon with the evening. That is, he will imagine the existence of Paul and of Simon with a relation to future time. On the other hand, if he sees Simon in the evening, he will relate Paul and Peter to the time past, by imagining them together with past time. And he will do this more uniformly, the more often he has seen them in this same order.20

If I first see Peter in the morning, Paul at noon, and Simon at night and then, the next day, I see Peter in the morning, I will, in virtue of psychological laws of association, automatically think of Paul at noon and Simon at night. But I will not think that Paul is currently present at noon because I have an idea of the sun being lower in the sky than it would be at noon. That idea is more powerful than the idea that represents the sun as having its noon position, and their contents are metaphysically incompatible. Therefore, the idea of Peter in the morning excludes the idea of Paul at noon. Assuming that Spinoza has a psychologically serious notion of belief, it seems natural to say that, although I will not believe that Paul is present at noon, I will, ceteris paribus, believe that Paul will be present at noon. This can be seen from the fact that I will act as if it were true that Paul will be present at noon. This, however, does not fit Steinberg’s model of belief in Spinoza. Recall that, on Steinberg’s interpretation, S believes that p just in case (1) S has an idea that p and (2) S’s idea that p is stronger than any idea that S has whose content excludes p. The problem is that, in the scenario described by Spinoza in 2p44s, I have no idea whose content is that Paul will be present at noon. I have, instead, an idea whose content is that Paul is present at noon and ideas that are associated with it by a temporal sequence. If there is a belief that Paul will be present at noon, then it is this complex that represents that content and no one idea. (Perhaps it will be objected that this complex is itself a complex idea. But the ideas that would constitute this complex idea are, by hypothesis, contrary to one another, in that their contents are metaphysically incompatible. Things that are contrary to one another cannot be in the same subject [3p5]. As 3p10 makes clear, being contrary in the sense that their contents are incompatible is a way of being contrary governed by 3p5.) I will return to the question of belief in Spinoza at the end of this section, but let us for now merely note the difficulty of locating a plausible notion of belief in the context of his claim that every idea is affirmed and turn now to the notion of affirmation itself. Several recent commentators have tried to explicate this notion in terms of Spinoza’s conatus doctrine. According to the conatus doctrine, “each thing, in and of itself, strives to persevere in its being”. This means that all of nature—from the stars in the sky to the smallest particle of matter—is animated by an urge to self-preservation. Human beings are

Affirmation, Judgment, and Epistemic Theodicy

35

no exception to this law, and Spinoza uses his conatus doctrine as the basis of his psychology. The human mind, for Spinoza, is a complex idea that represents the human body. It is a complex idea, in that it has parts that are themselves ideas that represent parts of the human body. Because each of these constituent ideas is, for Spinoza, a thing, they, too, strive for self-preservation. The conatus, or striving for self-preservation, of each thing is the ground of its causal powers. It determines both what something tends to do and how successful it will be if it is opposed by external causes. Can Spinoza’s conatus doctrine be used to explain why Spinoza thinks that every idea is an affirmation? Michael Della Rocca claims that every idea is an affirmation, for Spinoza, because every idea is bound up in the conatus of the agent who possesses it—that is, produces effects that are beneficial to the agent or that the agent regards as beneficial unless prevented from doing so by other ideas.21 He sees Spinoza as offering here both a theory of belief and affirmation according to which every idea is both affirmed and believed. In a similar vein, Diane Steinberg has argued that every idea is an affirmation because every idea strives to affirm (i.e. preserve) the existence of its object. Steinberg, however, denies that every idea is a belief and argues that only an idea that is more powerful than any idea that excludes it is a belief. Such proposals face a serious problem. As far as I can see, there is no way to analyse affirmation as the manifestation of the conatus that retains the connection between affirming an idea and the truth of that idea that Spinoza obviously intends it to have. He writes: [B]y will I understand a faculty of affirming and denying, and not desire. I say that I understand the faculty by which the Mind affirms or denies something true or something false, and not the desire by which the mind wants a thing or avoids it.22 It is clear from this passage that Spinoza thinks affirmation is related to truth in such a way as distinguishes it from “the desire by which the mind wants something”. Efforts to understand affirmation in terms of conatus, however, fail to preserve any connection to truth that distinguishes it in this way. Della Rocca’s interpretation fails because being such as to tend to produce effects that are beneficial or regarded as such doesn’t distinguish affirmation from desire. After all, desires often tend to produce effects that are beneficial or regarded by me as beneficial. For example, my desire that I eat leafy greens tends to produce effects that are beneficial to me or are regarded by me as such. Thus, on Della Rocca’s interpretation, such a desire would be an affirmation and, indeed, a belief. Steinberg’s proposal faces a similar objection. According to her, every idea strives to affirm (i.e. preserve) the existence of its object.23 But to

36

Martin Lin

see this as an account of judgment requires an equivocation on ‘affirms’ because I can affirm the existence of some object (that is, have a mindto-world directed attitude towards the proposition that it exists) without affirming its existence (that is, striving to preserve its existence) and vice versa. For example, that I affirm (in the sense that has a mind-to-world direction of fit) the existence of my coffee cup does not entail that I will try to protect it from harm. And I can, for example, perform actions that will protect an as of yet unborn child from harm without affirming that the child presently exists. The conatus doctrine only tells us that each idea affirms its object in the sense of protecting-from-harm and not in the sense of taking-to-exist. If the affirmation is not reducible to the conatus of each idea or of each mind, why then does Spinoza believe that every idea is affirmed? I propose that we can begin to find the answer to this question by looking at Spinoza’s claim that all sense perception is affirmed unless we have an imaginative idea (imagination, for Spinoza, as for most early moderns, pertains to imagistic thoughts, which include sense perceptions) that excludes it. Spinoza writes: If the human Body is affected with a mode that involves the nature of an external body, the human Mind will regard the same external body as actually existing, or as present to it, until the Body is affected by an affect that excludes the existence or presence of that body.24 Although it is not immediately obvious, this text concerns sense perception because an idea of a state of the body that has an external cause (“involves the nature of an external body”) constitutes, according to 2p16c1, sense perception of that cause. But Spinoza says that not only do we have sense perception of the external cause but we also judge it to exist (“the human mind will regard the same external body as actually existing”). That is, every perceptual idea entails an affirmation of the proposition that the object of the idea exists. His reasons for thinking this derive from his account of intentionality. According to him, there are two sources of intentionality. There is the primitive underived intentionality by which every idea represents the body to which it is identical. There is also derivative intentionality that has a causal/informational basis. If the body is in a state that has an external cause, then the idea of the body represents that state and also the external cause in virtue of the fact that the state carries information about the cause. This is the kind of causal/informational intentionality that allows us to infer fire from smoke and the age of a tree from the number of rings in its trunk. When we are in a state that implies the existence of an external cause (via causal/informational connections), we will represent that cause as existing unless we receive new information that tells us the cause no longer exists. This new information is encoded

Affirmation, Judgment, and Epistemic Theodicy

37

in a state of the body, which is incompatible with the present existence of the cause of the previous state. Spinoza’s argument for 2p17 is revealing. He writes: Dem: For whatever happens in the object of any idea, the knowledge of that thing is necessarily in God (by 2p9c), insofar as he is considered to be affected by the idea of the same object, i.e. (by 2p11), insofar as he constitutes the mind of something. Therefore, whatever happens in the object of the idea constituting the human Mind, the knowledge of it is necessarily in God insofar as he constitutes the nature of the human Mind, i.e. (by 2p11c), knowledge of this thing will necessarily be in the Mind, or the Mind will perceive it, q.e.d. In this demonstration, Spinoza says that the human mind is God’s idea of the human body—that is, the idea that constitutes knowledge of the body in the divine intellect. This being so, the idea of the body in God’s intellect—that is, the mind—must be affirmed because, presumably, knowing that p entails affirming that p. But how should we understand affirmation so that every idea, whether it is related to God’s mind or ours, is affirmed, yet not every false idea in the human mind constitutes an error? My proposal for making sense of Spinoza’s remarks on truth and affirmation has been prefigured by my criticisms of attempts to reduce affirmation to conatus. It is simply to read him as specifying the direction of fit that pertains to ideas. On this interpretation, the intended contrast between “affirming [. . .] something true” and “the desire by which the mind wants a thing” is that which obtains between that which purports to represent the world as it really is and that which doesn’t. Moreover, construing the connection between affirmation and truth as specifying a mind-to-world direction of fit allows Spinoza’s claim that every idea is affirmed to be consistent with his claim that not every false idea constitutes an error. There is no error in merely having a false mental representation with a mind-to-world direction of fit. This general point could be illustrated by the case of credences. Credences have a mind-to-world direction of fit, and yet I commit no error simply in virtue of having nonzero credences towards false propositions. For example, I commit no error if I evenly distribute my credences between the propositions that the fair coin will land heads and that the fair coin will land tails, although one of them is false. Although this answers the challenge of characterizing affirmation, serious difficulties remain. Just as Descartes faces the problem of showing how human error is compatible with the existence of a perfect God, so, too, does Spinoza face a similar problem. If our minds are fragments of God’s infinite intellect, every idea of which constitutes knowledge of the world, how can we account for attitudes such as doubt and denial in the

38

Martin Lin

human mind? A similar problem arises for error and misrepresentation, and consideration of how Spinoza addresses it will shed light on how Spinoza tries to account for doubt and denial. According to Spinoza, “[a]ll ideas, insofar as they are related to God, are true” and “there is nothing positive in ideas on account of which they are called false” because all ideas are modes of God. If there were ideas that were false in and of themselves, then God would have false ideas, which is incompatible with divine omniscience—or, as Spinoza would put it, it would be incompatible with the fact that “whatever follows formally from God’s infinite nature follows objectively in God from his idea in the same order and with the same connection”.25 And yet, this claim appears to be incompatible with the claim, which Spinoza endorses, that the human mind is an idea in God and that human minds can commit errors and misrepresent the world. Spinoza’s solution is that a given idea can be false relative to the human mind26 and true relative to God.27 He illustrates this with an example of an idea that represents the sun as being two hundred feet away.28 This idea is false only if it occurs in a mind that lacks knowledge of the true distance of the sun. A human mind can both have an idea of the sun being two hundred feet away and lack knowledge of the true distance of the sun, but God cannot. Therefore, an idea that represents the sun as two hundred feet away can be false in a human mind but cannot be false in God. Whether or not Spinoza’s account of falsity succeeds is a difficult question that I cannot take up here. Rather, I would simply like to note that Spinoza’s treatment of attitudes other than affirmation, such as doubt and denial, is structurally similar. First of all, just as there is nothing positive in an idea in virtue of which it is not true, so, too, there is nothing positive in an idea in virtue of which it is not affirmed. And just as falsity results from the interaction between ideas, there is evidence that Spinoza thinks that attitudes such as denial and doubt result from the interaction between ideas. For example, he gives the example of a child imagining a winged horse and not perceiving anything else. In this circumstance, he claims that the child will regard the horse as present. He writes: For if the Mind perceived nothing else except the winged horse, it would regard it as present to itself, and would not have any cause of doubting its existence, or any faculty of dissenting, unless either the imagination of the winged horse were joined to an idea which excluded the existence of the same horse, or the Mind perceived that its idea of a winged horse was inadequate. And then either it will necessarily deny the horse’s existence, or it will necessarily doubt it.29 In this text, Spinoza clearly states that we deny or doubt an idea x only if, in addition to idea x, we possess an idea y whose content excludes the content of idea x (in the case of denial) or we possess an idea z whose

Affirmation, Judgment, and Epistemic Theodicy

39

content entails that idea x is inadequate (in the case of doubt). This states a necessary condition on doubt and denial. (Diane Steinberg objects that all of our perceptual ideas are inadequate, and thus, upon recognizing this, Spinoza’s theory predicts that we suspend judgment, which strikes her as implausible.30 But Spinoza says that doubt results from this recognition and says nothing about suspension of judgment.31 Doubt, for Spinoza, is the state of being less than certain about something, which he also thinks is the correct attitude to take towards the deliverances of the senses.) Are there other conditions that must be in place for us to doubt or deny? Presumably yes because if we take exclusion to be a relation of logical or metaphysical incompatibility, then it is symmetric; that is, both ideas exclude each other.32 But in typical cases of denial, the idea that causes me to deny something is not itself denied. For example, if I have a perceptual idea that represents the distance of the sun as two hundred feet away and I have another idea that causes me to doubt it, say an idea that represents the true distance of the sun, then I deny the misrepresentation precisely because I don’t doubt the true one. At this point, I think that we can make use of the conatus to determine which cognitive attitude will obtain. Thus, like some of the accounts that we considered earlier and rejected, my account will appeal to Spinoza’s conatus doctrine, but unlike those accounts, my account does not reduce affirmation to the conatus or the causal powers of an idea. Rather, my proposal is that the conatus associated with an idea pushes towards actions that would conduce to self-preservation if the idea were true, including imaginative content that doesn’t represent the body parallel to the idea but rather the external causes responsible for its state. That is, the conatus takes an idea to be affirmed and determines us to act appropriately. When ideas conflict, their actions can generate attitudes of doubt and denial. Before explaining exactly how this works, it will be useful to review a few details of Spinoza’s theory of individuals and how it relates to the conatus doctrine. Let us begin by considering body under the attribute of extension and then apply the theory of complex individuality that Spinoza develops to modes under the attribute of thought. For Spinoza, simple bodies can join together to form complex bodies if they communicate their motions to one another according to a fixed pattern.33 What is more, a complex body can lose parts and still survive if those parts are replaced by functional equivalents.34 That is, both mereology and identity through change are determined by functional or causal properties. The resulting complex bodies can themselves be parts of even more complex bodies if their motions similarly realize a fixed pattern. There is, according to Spinoza, a hierarchy of increasingly complex bodies—each of which satisfies this condition—that terminates with a single superindividual that has every simpler body as a part.35 The human mind,

40

Martin Lin

according to Spinoza, is the idea of the human body, and every idea is identical to the body that it represents. Thus, the human mind and the human body are identical.36 The human body is a complex body that has many simpler bodies as parts that are bound together by their functional organization. Similarly, the human mind is a complex idea that has many simpler ideas as parts that are bound together by their functional organization.37 On this picture, the natural world is composed of a great many individuals at many different levels of complexity. Spinoza’s conatus doctrine says that each thing strives to persevere in its being.38 Because each of these individuals at each level of complexity is something in its own right, they exhibit this striving for self-preservation. But to the extent that they are parts of a more complex whole, their actions coincide with and are regulated by the striving of that more complex whole of which they are parts. For example, the human heart is both an individual with its own functional organization and part of the human body. As a genuine individual, the human heart strives to preserve its own functional organization and contributes to the human being’s striving to preserve its own functional organization. Now we are in a position to see how Spinoza could use his conatus doctrine to generate attitudes of denial and doubt from a collection of ideas that, insofar as they are in themselves and insofar as they are related to God’s mind, are all affirmed. Each idea, insofar as it is in itself, is affirmed because it has a mind-to-world direction of fit and determines, in virtue of its conatus, mental actions that would result in its self-preservation if its content were true. But insofar as they are parts of a mind that contains ideas that conflict with them, their actions will be transformed as a result of a struggle between them and their rivals. The outcome of this struggle will, in turn, be determined by the relative strength of those ideas. The relative strength of each idea is a matter of the conatus of each and the conatus of the mind of which they are parts. Insofar as an idea is part of a mind, what would preserve itself coincides with what would preserve the mind. But, insofar as an idea is a passion, its interests can diverge from those of the mind to which it belongs. It cannot, of course, destroy the mind because Spinoza thinks that nothing can destroy that in which it inheres,39 but it could seek to produce effects that result in the continued existence of the mind at a suboptimal degree of power.40 This divergence of interests can generate mental conflict that grounds attitudes such as denial and doubt. Denial results from an idea, which in itself is affirmed, being joined to ideas that exclude it (that is, their contents are logically or metaphysically incompatible with its content), that are stronger than it. Because of the weakness of the denied idea relative to the strength of the ideas that exclude it, the net result of their combined power is action that would conduce to the survival of the mind

Affirmation, Judgment, and Epistemic Theodicy

41

of which it is a part if its content were false. Doubt results from an idea being joined to stronger ideas whose contents entail that it is inadequate. Because of the relative weakness of the denied idea, the net result of its struggle with the ideas that indicate its inadequacy is action that would conduce to the survival of the mind of which it is a part if its content may or may not be true. In other words, doubt and denial are the result of a struggle between ideas, all of which purport to represent the world as it is. In themselves, they produce actions that would be beneficial to themselves and to the mind of which they are parts if they were true, but, in combination with each other, they produce actions that are determined by the net force that results from each of their individual strivings. In this way, affirmation begets denial and doubt. This account of affirmation, denial, and doubt entails that, whereas affirmation is fundamentally a subpersonal phenomenon, denial and doubt occur only at the level of human mind. (We can, of course, say of a mind that it affirms an idea, but this just means that it has an idea that is not denied or doubted. Nothing new need take place at the personal level for a mind to affirm an idea.) Ideas that are doubted or denied at the personal level are still affirmed at the subpersonal level. Those ideas still purport to represent the world as it really is, but their voices are drowned out by rival ideas. The mind containing them does not affirm them insofar as it doubts or denies them, but at the subpersonal level they are, as it were, still arguing their case. Earlier, we considered some accounts of belief in Spinoza and found them wanting. Having explained what affirmation is, for Spinoza, can we leverage this account of affirmation to give an account of belief? I am pessimistic about the prospects for doing so because I suspect that the ordinary conception of belief is irreducibly normative and such notions do not easily fit into Spinoza’s system. To see why belief might be normative, consider the difficulty of defining belief according to its functional role—that is, having certain inputs and outputs. For example, we might attempt to define belief as something that is formed on the basis of evidence and guides action. But this will not do because some of our beliefs are formed not on the basis of evidence but wishful thinking, and sometimes our actions are not guided by our beliefs but by representational states whose truth we do not endorse, as when we refuse to step out onto the glass floor of the CN tower despite believing the floor is strong and will not break. Rather, beliefs ought to be formed on the basis of evidence and ought to guide our action. I suspect similar problems will arise with any attempt to define belief in non-normative terms. The normativity of belief makes it difficult to square with Spinoza’s philosophy of mind. Spinoza’s philosophy is sometimes described as naturalistic because he thinks that all phenomena can be understood in terms

42

Martin Lin

of inviolable natural laws that are always and everywhere the same. Such naturalism is not a hospitable environment for the normative because natural law tells us what must happen given certain causal antecedents and not what ought to happen. Accordingly, Spinoza’s treatment of normative concepts such as perfection and imperfection, good and evil, and just and unjust reduces them to descriptive, non-normative elements. This strongly suggests that nature, for Spinoza, is ultimately non-normative. Thus, if belief is indeed an irreducibly normative notion, as I suspect it is, then Spinoza’s naturalistic philosophy of mind is not well-suited to accommodate it. Indeed, although Spinoza sometimes speaks of belief in passing, in the principal source of his psychological views, the Ethics, he never refers to it in rigorous formulations of his psychological principles. This being so, I propose that we abandon the attempt to define a Spinozistic notion of belief from his views of seemingly related topics such as affirmation, denial, and doubt and accept that Spinoza’s psychology is belief-free.

Conclusion I have argued that both Descartes’s and Spinoza’s theories of judgment result from their attempts to reconcile the fact of human error with the existence of a perfect God. The problem is particularly acute for Spinoza because the human mind is not only produced by God but is also, indeed, a part of the infinite intellect and constitutes God’s knowledge of the human body. For this reason, Spinoza thinks that every idea is affirmed, which I have argued means that every idea purports to represent the world as it really is. Thus, just as there is nothing positive in an idea in virtue of which it is false, so, too, there is nothing positive in them in virtue of which they are denied or doubted. Similarly, just as every idea is true relative to God’s mind, so, too, is every idea affirmed relative to God’s mind. Insofar as they are related to the human mind, however, ideas can be denied or doubted. These cognitive attitudes emerge from the interplay of ideas that are, in themselves, affirmed. This means that some ideas are simultaneously affirmed and denied or doubted. This is possible because affirmation and doubt and denial are phenomena occurring, in the first instance, at different levels. Doubt and denial attach to ideas at the level of the human mind, whereas affirmation is essentially at once subpersonal and divine. (The relationship between how things are quantum in se est and how things are in relation to God is a complex topic. Exploring this further would require an in-depth discussion of Spinoza’s notions of eternity, duration, and perception sub specie aeternitatis, which I cannot undertake here.) Because it is subpersonal, and so is consistent with denial and doubt, I have argued that affirmation, in Spinoza, should be understood as a technical notion that is unconnected to folk-psychological notions such as belief.

Affirmation, Judgment, and Epistemic Theodicy

43

Abbreviations and Conventions Passages of Spinoza’s Ethics are cited in the following way: app = appendix; a = axiom; c = corollary; d = demonstration or definition depending on context; p = proposition; s = scholium. AT CSM ST

Descartes, R., Oeuvres de Descartes, ed. Charles Adam and Paul Tannery. Descartes, R., The Philosophical Writings of Descartes, ed. J. Cottingham, R. Stoothoff and D. Murdoch. Aquinas, Thomas, Summa Theologiae.

Notes 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

AT VII 35/CSM II 24. AT VII 35–36/CSM II 25. AT VII 54/CSM II 38. AT VII 36–37/CSM II 25–26. AT VII 37/CSM II 26. ST I.75.6; 85.2. ST I.16.2. AT VII 56–57; CSM II 39–40. AT VII 57; CSM II 40. AT VII 59–60; CSM II 41. Williams (2005: 162–3), Curley (1975: 177–8), Wilson (1978: 127). AT VII 58–59; CSM II 41. Della Rocca (2006). Schüssler (2013). E3pref. 2p49c. Della Rocca (2003: 207). J. Steinberg (2018). D. Steinberg (2005: 151). 2p44s. Della Rocca (2003: 209). 2p48s. D. Steinberg (2005: 154). 2p17. 2p7c. 2p36d. 2p32, 2p32d, 2p33, 2p33d. 2p35s. 2p49s. D. Steinberg (2005: 408–9). 2p49s. See D. Steinberg (2005: 151). Definition following A´´ of the Short Physical Digression following 2p13s. L4 following A3 of the Short Physical Digression. L7s of the Short Physical Digression. 1p7s. 2p14.

44

Martin Lin

38. 3p6. 39. 3p4. 40. 3p11s.

References Aquinas, T. (1882–), Opera Omnia. Rome: Commissio Leonina. Curley, E. (1975), “Descartes, Spinoza, and the Ethics of Belief”, in E. Freeman and M. Mandelbaum (eds.), Spinoza: Essays in Interpretation. Chicago: Open Court. Della Rocca, M. (2003), “The Power of an Idea: Spinoza’s Critique of Pure Will”, Noûs 37(2): 200–31. Della Rocca, M. (2006), “Judgment and Will”, in S. Gaukroger (ed.), The Blackwell Guide to Descartes’ Meditations. Hoboken: Wiley-Blackwell. Descartes, R. (1965), Oeuvres de Descartes, ed. P. Tannery and C. Adam, 11 vols. Paris: Vrin. Descartes, R. (1984), The Philosophical Writings of Descartes, ed. J. Cottingham, R. Stoothoff and D. Murdoch, 3 vols. Cambridge: Cambridge University Press. Schüssler, R. (2013), “Descartes’ Doxastic Voluntarism”, Archiv für Geschichte der Philosophie 95(2): 148–77. Spinoza, B. (1925), Opera, ed. C. Gebhardt, 5 vols. Heidelberg: Winter. Spinoza, B. (1985), The Collected Works of Spinoza, trans. E. M. Curley, Vol. 1. Princeton: Princeton University Press. Steinberg, D. (2005), “Belief, Affirmation, and the Doctrine of Conatus in Spinoza”, Southern Journal of Philosophy 43(1): 147–58. Steinberg, J. (2018), “Two Puzzles Concerning Spinoza’s Conception of Belief”, European Journal of Philosophy 26(1): 261–82. Williams, B. (2005), Descartes: The Project of Pure Enquiry. London: Routledge. Wilson, M. D. (1978), Descartes. London: Routledge.

2

Locke and Leibniz on Judgment The First-Person Perspective and the Danger of Psychologism Maria van der Schaar

1.

Introduction1

In our time, philosophers prefer to speak about belief rather than judgment. Judgment is thought to be an old-fashioned term suited to a rationalist tradition and an out of date psychologistic logic. Belief is generally understood to be a primitive notion, and judgment may then be explained as occurrent belief (cf. Schwitzgebel 2015). The notion of belief is central to philosophy of mind and epistemology and is generally understood as the mental state in which one takes a proposition to be true. Beliefs may have grown in us in any way that may be accounted for by empirical means. Belief in this sense is an empirical notion. Recent literature on assertion can be understood, though, as a revival of the notion of judgment. Whereas belief is a mental state, assertion is primarily a speech act. Judgment is also primarily an act, and either assertion or judgment can be taken as primitive, while the one is explained in terms of the other. Although an older generation of philosophers is used to explaining the speech act of assertion in terms of belief (Searle 1979), such an approach is now seldom defended. One of the most convincing arguments against a belief account of assertion is given by Timothy Williamson. If someone makes an assertion, an interlocutor is entitled to ask: “How do you know?” (Williamson 2000: 252) This may mean that one is entitled to make an assertion only if one knows the asserted proposition to be true, as Williamson says. Or one may claim that being able to give a ground is enough for such an entitlement (Schaar 2011). If the asserter is not able to give a ground, the assertion needs to be withdrawn. There thus seems to be a tight relation between assertion and ground or reason, just as there always has been a tight relation between judgment and reason, as I will explain in section 5. Whereas belief is an empirical notion, assertion and judgment are normative notions, as they are to be understood in terms of knowledge or ground. John Locke, in his Essay Concerning Human Understanding (1690), was the first philosopher to use the term ‘belief’ synonymously with ‘judgment’ in a systematic way (see section 3). Does a neglect of the distinction

46

Maria van der Schaar

between judgment and belief bring in a form of psychologism in philosophy? In this paper, I will address this general question by comparing the different accounts of judgment given by Locke and Leibniz. Leibniz’s reading of Locke’s Essay, starting in 1695, is especially relevant to understanding the differences. After 1700, with the French translation of Locke’s Essay having appeared, Leibniz gave detailed comments on all sections of the Essay, intending to publish them in dialogue form as the Nouveaux Essais, which was finished at the end of 1704, though published only in 1765. Notwithstanding the fact that Leibniz takes his inspiration on epistemological questions from the fourth book of the Essay, Leibniz’s basic account of logical and epistemological notions differs on crucial points from Locke’s. Locke’s notion of judgment oscillates between a normative notion and an empirical notion of belief, and there is thus a danger of psychologism in Locke that Leibniz was well aware of. This difference between Locke and Leibniz may help us to understand a fundamental problem in modern philosophy, for it seems that we have lost something by substituting the notion of judgment for that of belief. The notion of judgment is needed in logic because logic aims at an answer to the question of how one should judge and reason. As one can see in Frege’s ideography, the presence of the normative notion of judgment in one’s logic does not imply a form of psychologism.2 Introducing the empirical notion of belief into logic, though, would reduce a normative notion to an empirical one and logical questions to psychological questions. In logic, we are not interested in what people actually believe but in the question of what one ought to judge. The red thread in this paper is that judgment and knowledge may be studied from two points of view. In modern analytic philosophy, we generally study these notions from a third-person point of view. We primarily aim to understand what it is for someone else to judge, believe, or know. We ask such questions as, under what conditions is one entitled to attribute belief or knowledge to someone else? Knowledge is understood in terms of knowledge attribution, and the same holds true for belief. When one understands knowledge from a third-person, empirical point of view, typical questions that arise are: Are our faculties and methods reliable? One thus deals with knowledge and belief as an empirical phenomenon in the world, and thereby hopes to be able to give a more objective account of these notions than when studying them from a first-person point of view. One studies judgment or knowledge from a first-person point of view when considering the question of what it is for oneself to make a judgment or to know. This first-person point of view is generally associated with the Cartesian thesis that we have infallible knowledge of our own mind and that we know our own mind better than other minds. As today philosophers no longer defend the Cartesian thesis, the first-person point of view is generally understood to result in a subjectivist philosophy. Both Locke and Leibniz reacted to the Cartesian method in their own way,

Locke and Leibniz on Judgment

47

not by neglecting the first-person point of view, but by arguing that we need to balance the first-person point of view with something else. An analysis of their views will enable us to address the main question of the paper: in what sense does the first-person point of view imply a form of subjectivism or psychologism, and in what sense may it be understood as a necessary element in our investigations of normative notions such as judgment and knowledge?

2.

Locke on Knowledge: The First-Person Point of View

The aim of Locke’s Essay is to determine “the Extent of humane Knowledge”, in opposition to opinion (E I.i.2: 43), in order to determine “the Bounds between Opinion and Knowledge” (E I.i.3: 44). Knowledge, which is infallibly certain, according to Locke, is possible in mathematics, but is in our practical life not attainable: we may there rely on probable opinion, which is enough for our preservation. Besides the faculty of knowledge, Locke is in need of a second faculty to explain error: the faculty of judgment, which is concerned with probabilities. In this and the next section, Locke’s distinction between knowledge and judgment is explained in terms of the distinction between first- and third-person perspective. For Locke, knowledge is “the perception of the connexion and agreement, or disagreement and repugnancy of any of our Ideas” (E IV.i.2: 525). The term ‘perception’ is ambiguous: it may mean either the act of perceiving or the perceived (dis)agreement. Which meaning is intended in the explanation of knowledge? A bit further on, Locke says, “When we know that White is not Black, what do we else but perceive, that these two Ideas do not agree?” (E IV.i.2: 525). He thus gives the act of perceiving as an example. And, when Locke explains the faculty of knowledge in terms of its actualizations, he writes: “Knowledge, whereby [the Mind] certainly perceives, and is undoubtedly satisfied of the Agreement or Disagreement of any Ideas” (E IV.xiv.4: 653). The notion in terms of which knowledge is explained is thus: the act of perceiving. There are three types of knowledge, according to Locke—intuitive, demonstrative, and sensitive—and thus three types of perceiving. In intuitive knowledge, “the Mind perceives the Agreement or Disagreement of two Ideas immediately by themselves, without the intervention of any other . . . the Mind . . . perceives the Truth, as the Eye doth light” (E IV.ii.1: 530–1). Intuitive knowledge is infallibly certain. Locke’s visual metaphor reminds one of the Cartesian criterion of truth and absolute certainty: whatever I clearly and distinctly perceive is true and absolutely certain. In this sense, intuitive knowledge is a first-person notion of knowledge: “I think, we may as rationally hope to see with other Mens Eyes, as to know by other Mens Understandings” (E I.iv.23: 101).3 Intuitive knowledge

48

Maria van der Schaar

is immediate in the sense that the two ideas in the known proposition are not mediated by a third idea. We directly perceive the agreement or disagreement. Locke’s question is not whether we, from a third-person point of view, are able to determine whether an agent is perceiving or knowing. Only the knowing subject can answer the question of whether he perceives the ideas of white and black to disagree. Intuitive knowledge also includes knowledge of our own existence and of our mental acts and appearances. There are thus, for Locke, first truths of reason as well as first truths of fact, although it takes Leibniz to make the distinction (NE 434). Locke says of the intuitive knowledge of our own existence that “it neither needs, nor is capable of any proof” (E IV.ix.3: 618), a point famously generalized to all primary truths by Leibniz (NE 434).4 Whether it concerns relations between ideas or our own existence, Locke explains intuitive knowledge from a first-person point of view. In demonstrative knowledge, the act of perceiving is mediated by other ideas, called proofs: “where the Agreement or Disagreement is by this means plainly and clearly perceived, it is called Demonstration” (E IV.ii.3: 532). Each step in the process of reasoning involves an act of mentally perceiving, and the conclusion is thus obtained by means of an act of demonstration. Whether there is demonstrative knowledge is something for the knowing agent to determine; demonstrative knowledge is thus a first-person notion. As mathematics and morality are demonstrative sciences, there is no doubt that a first-person point of view plays a role in Locke’s Essay. The truths of mathematics and morality are to be proved on the basis of primary truths, which are the result of an intuitive act of perceiving. And demonstrative knowledge is first-person, too. When an idea is “actually coming into our Minds by our Sense”, and we “inferr the existence of any thing without us, which corresponds to that Idea”, we have sensitive knowledge, or, at least, what “passes under the name of Knowledge” (E IV.ii.14: 537). By sensation, we perceive the existence of particular things actually present to our senses (E IV.iii.5: 539). As there is, according to Locke, “a very manifest difference between dreaming of being in the Fire, and being actually in it” (E IV.ii.14: 537), the knowing agent is in principle able to determine whether he has sensitive knowledge. Although Locke admits that sensitive knowledge is less certain than intuitive or demonstrative knowledge, whether one has sensitive knowledge can, in principle, be determined by the knowing agent, and sensitive knowledge is thus also a first-person notion. Although Locke’s primary notion of knowledge, the act of perceiving, may be called justified, the justification is not something added to the act of perceiving, for it is precisely the act of perceiving that gives the justification. In a similar way, one may say that knowing as act of perceiving is true, for an act of perceiving is infallible, on Locke’s account: truth is not an extra element to be added to the act of perceiving in order to have knowledge. For Locke, truth is implied in the act of perceiving.

Locke and Leibniz on Judgment

49

Although every piece of knowledge is thus true and justified, knowledge as the act of perceiving is not explained in these terms precisely because it is a first-person notion. It thus differs in an important sense from our modern understanding of knowledge. No external notion of truth is part of Locke’s concept of knowledge.5 Is Locke’s first-person point of view on knowledge able to prevent a subjectivist account of knowledge? I come back to this question in section 4. According to Locke, by comparing the ideas in one’s mind, one is able to determine the (dis)agreement between the two ideas. This means, though, that we have knowledge only of our own ideas. This makes the extent of knowledge very restrictive, as Locke himself points out, for it gives us no knowledge of the external world. As a result, Locke takes the realm of probable judgment to be of greater interest.

3.

Locke on Right Justified Judgment: The Third-Person Point of View

As the act of perceiving is infallible, according to Locke, the faculty of judgment is needed to account for error: Judgment, which is the putting Ideas together, or separating them from one another in the Mind, when their certain Agreement or Disagreement is not perceived, but presumed to be so; which is, as the Word imports, taken to be so before it certainly appears. And if it so unites, or separates them, as in Reality Things are, it is right Judgment. (E IV.xiv.4: 653) In the act of judgment, we might be mistaken. Locke’s fallibilism does not imply a form of scepticism: if it conforms to reality, our act of judgment is right. I come back to the notion of right judgment below. In the quote, Locke explains the act of judging as an act of presumption, a legal term, as Leibniz rightly notes, meaning that one accepts provisionally but not groundlessly while waiting for a proof to the contrary (NE 457). We may judge without grounds, in which case Locke calls it mere opinion, but we ought to judge on the basis of reasons. Whereas in knowledge the proofs can be found by intuition, which means that they are intrinsic to the content of our act of knowing, in judgment the arguments are “extraneous to the thing I believe” (E IV.xv.3: 655). Before it makes a judgment, The Mind . . . ought to examine all the grounds… for or against any probable Proposition . . . , and upon a due ballancing the whole, reject, or receive it, with a more or less firm assent, proportionably to the preponderancy of the greater grounds of Probability on one side or the other. (E IV.xvi.5: 656)

50

Maria van der Schaar

Our epistemic judgments are always provisional, waiting to be corrected when new evidence becomes available. Although Locke does not explain judgment in terms of reasons or grounds in the first quote above, there is a clear connection between the two notions: one is entitled to make a judgment only upon balancing the account of reasons. I come back to the relation between judgment and reason in section 5. Locke’s category of judgment is a complex one. Apart from the fact that Locke uses the term ‘judgment’ for the faculty of judgment, the term may also have different meanings when it concerns the actualizations of such a faculty. In the first place, Locke understands the act of judgment to be an all or nothing affair: there is assent, dissent, and suspension of judgment. Furthermore, Locke understands judgment as also involving degrees. Judgment in this sense is to be understood as conviction. The degree of conviction ought to be in proportion to one’s grounds. It may thus reach from belief, conjecture, and guess to doubt and disbelief (E IV.xvi.9: 663). Thus, Locke’s concept of judgment does not only include (1) the act of judging and (2) mere opinion, in contrast to knowledge, but also (3) conviction or belief, in varying degrees. Finally, it includes (4) the notion of religious faith, faith being our assent to revelation (E IV.xvi.14: 667). Depending on the context, Locke uses ‘faith’, ‘opinion’, ‘belief’, ‘assent’, and ‘judgment’ interchangeably (cf. Schaar 2008). As Locke is the first to give both the term ‘belief’ and the notion of belief a central place in philosophy, Locke’s account may have contributed considerably to our modern notion of belief, which involves all the ambiguities that Locke himself used with definite purposes in each case. Natural science cannot be called a ‘science’ in the strict sense, according to Locke: we can attain there no infallible knowledge, and we will never understand why an empirical proposition is true. Yet, the judgments we make there are crucial to our survival. For Locke, judgment is far more important than knowledge. Whereas knowledge is concerned primarily with our ideas alone, in our investigations of the world, we have to use the faculty of judgment. What we need in natural science and in practical matters of life is justified true judgment. Because the grounds are external to the judgmental content, the judging agent can be in error concerning the grounds and the truth of his judgment. Locke uses the fallibility of our judgments in one of his arguments for religious toleration in his Epistola de Tolerantia (Locke 1689: 123; cf. Schaar 2012: 56). As the magistrate is as fallible in his judgment concerning religious matters as any other human being, he might be mistaken in his faith and is therefore not entitled to force his faith upon others. For Locke, each individual has to judge for himself, for each will be individually held accountable for his faith when standing before God. In this sense, Locke’s account of judgment is individualistic (cf. Jolley 2016: ch. 4). Although our judgment may be justified and right, our judgment is fallible and gives us no entitlement to force our judgment or faith upon

Locke and Leibniz on Judgment

51

others. Only God can know whether our judgment is right, and this makes a right judgment very different from knowledge. God’s point of view on our judgments is preeminently a third-person point of view: his is a God’s eye point of view. For Locke, whether one’s judgment is right can thus, in the end, be determined only from an external, third-person point of view. Combined with the fact that he uses ‘judgment’ interchangeably with ‘belief’, and that he considers first-person knowledge to be less relevant, we may conclude that Locke has influenced our modern, third-person understanding of judgment, knowledge, and belief: the most we can obtain with respect to the world is justified true belief, not a firstperson insight or understanding.

4.

Leibniz’s Logical Point of View

Leibniz criticizes Locke’s strict definition of knowledge in terms of perception, and thereby also criticizes his absolute distinction between knowledge and judgment. For Leibniz, there is also knowledge of likelihoods (NE 373); that is, we can often measure the degrees of probability in empirical matters. Furthermore, truths of fact about sensible things outside us can be verified by means of truths of reason (NE 375). Optical appearances can, for example, be explained by geometrical principles insofar as we study the way these phenomena are linked to each other. Optics and magnetology are good examples of empirical sciences: “from a few assumptions grounded in experience we can demonstrate by rigorous inference a large number of phenomena” (NE 453). As in Locke, Leibniz’s concept of knowledge is primarily first-person. Leibniz takes mathematical knowledge to be the model of science, and mathematics is based on insights and demonstrations, resulting from a first-person act of insight or proving. There is an important difference with Locke, though, in the role the first-person perspective plays in Leibniz’s concept of knowledge. Like Descartes, Locke takes infallible certainty to be essential to knowledge, and he takes clear and distinct perception as a criterion for truth and certainty. As a visual metaphor is used to elucidate knowledge, there is a danger that knowledge becomes a purely psychological notion. Locke’s explanation of knowledge in terms of perception is not precise enough, according to Leibniz, for the perception may be confused (NE 452), especially because, for Locke, mathematical ideas have their foundation in the sensual. Neither is clarity enough as a criterion of truth, for we often make mistakes regarding the question of whether a perception is clear and distinct. The rules of common logic are not to be despised as criteria of the truth of propositions (Leibniz 1684: 425). We need formal features, and these are “accessible to ourselves and to others” (Leibniz 1675: 3). All mistakes in reasoning will become visible because they are shown in a wrong combination of characters. Error arises when we violate the rules governing the formation of expressions (“Scientia Generalis.

52

Maria van der Schaar

Characteristica” XV, GP VII: 205). In this sense, a first-person perspective needs to be balanced by a universal language, where reason is made public (cf. Losonsky 2001: 160–63). It is true that the main aim of the universal language is not communication but the improvement of our own thinking, making the deep structure of thought visible. It is also true that Leibniz’s ideal is to prove by meditation alone and not by means of blindly applying the rules of the calculus. At the same time, Leibniz’s ideal of such a language opens up the first-person perspective for others aiming at knowledge. As soon as one has acknowledged the axioms and definitions of a science, as well as the basic principles of Leibniz’s universal language, one is able to calculate blindly and thus come to new results by mechanical means. Considered purely from a first-person point of view, the universal logic also provides something more than certainty. Locke’s focus on the question of what things we know first in the temporal order leads to an arbitrary and huge amount of first truths, truths we all know by intuition, according to Locke—that is, by an act of perceiving the (dis)agreement of the ideas. Two plus two is four is considered by Locke to be a first truth, for we can determine its truth with infallible certainty without the use of other truths. The point of arithmetic, though, is not to establish the certainty of arithmetical truths but to understand how they can be proved from the appropriate axioms and definitions (cf. Frege 1884: §6, under reference to Leibniz, NE 413–14). Essential to scientific knowledge is the natural order of our knowledge (NE 411) and of truths: [I]n the natural order, we are not concerned . . . with the sequence of our discoveries, which differs from one man to another, but with the connection and natural order of truths, which is always the same. (NE 412, quoted with approval in Frege 1884: §17)6 Because Locke does not make a distinction between the temporal and the natural order of truths, his first-person view of (scientific) knowledge is psychologistic and subjective, for the temporal order may differ for different men (cf. Frege 1879: preface). We may obtain the natural order by demonstrating arithmetical truths such as two plus two is four—that is, by proving them from axioms and definitions. The axioms come first in the natural or explanatory order in the sense that they give the reason why the other truths are true. Such an order is common to all men and intelligences in general (NE 276), and it reflects the order and connection of things. Because such an order shows the reason why a proposition is true, Leibniz’s epistemic quest concerns understanding rather than Lockean certainty (cf. Wilson 1967: 356). This is not to say that Locke does not aim at capturing knowledge as understanding; it is, rather, that he fails at its aim because he relies too much on the metaphor of perception and takes self-evidence to be nothing but trifling obviousness. For

Locke and Leibniz on Judgment

53

Leibniz, knowledge as understanding means bringing different truths in an explanatory order rather than focussing on isolated certainties. This way, one may find the kernel from which other truths may be proved but which cannot itself be proved, for proving cannot go on to infinity. The kernel contains the general principles of “our thoughts, serving as their inner core and as their mortar” (NE 84; cf. Frege 1893: vi). The smaller this kernel of axioms the better, as there is thus a better survey. We thus can see from which primary principles everything is derived. Leibniz agrees with Locke that every particular identity proposition is self-evident (Leibniz prefers the simple “évident”, as “évident par ellemême” is not common in French [NE 408, IV.vii.2]). For Leibniz, though, their self-evidence is based on the self-evidence of the law of identity: “we shouldn’t here be contrasting the axiom with the example . . . but rather regarding the axiom as embodied in the example and as making the example true” (NE 413). The self-evidence of particular identity propositions is thus based upon the self-evidence of the axiom of identity, and in this sense these truths are not wholly independent, in contrast to what Locke claims (E IV.vii.10: 597). Self-evidence consists, for Leibniz, not simply in clearly perceiving that the ideas agree, for we may also understand that A is A is self-evident when we do not have a clear idea of A. Such a proposition is evident or known from the terms—ex terminis—as soon as they are understood, Leibniz writes under reference to Scholastic philosophers, but in the case of A is A, we do not have to understand what A is in order to determine its truth. A is A is true because of the law of identity, which is, according to Leibniz, somehow part of its particular instances, providing the reason why it is true. The law of identity is an axiom, a primary truth, in the sense that it is (1) immediate—known without mediation of a third idea; and (2) indemonstrable—not capable of being demonstrated by any truth prior in the epistemic, explanatory order (NE 406–8; cf. NE 434). Evidence can also be used for mediate truths. Evidence is “luminous certainty [une certitude lumineuse], where we have no doubt because of the way we can see the ideas to be linked together” (NE 445). The connection of the ideas can thus be made exact by definitions, axiomatic identities, and demonstrations (NE 452). The luminosity gives a certain perspicuity, “geistige Durchleuchtung”, as Frege puts it (NS, 171), providing insight and understanding in the way these truths are related to each other. From Leibniz’s point of view, Locke’s account of scientific knowledge is merely of psychological value, and his conception of geometry does not surpass that of the Egyptians (NE 452). Locke cannot explain the generality and necessity of mathematics. Leibniz’s first-person perspective on knowledge gives us a picture of knowledge as understanding without reducing understanding to mental perception. Understanding is relating ideas and truths to each other and thereby coming to know the reason why something is true. Even in the case of an axiom, though a borderline

54

Maria van der Schaar

case because it cannot be proved, one may distinguish a reason why it is true, Leibniz says. In case it is a primary truth of reason—that is, an identity—the reason is that it is known by itself: “thus for every truth there is a reason, either because the connection between the predicate and the subject is self-evident (evident per se), as in identities, or because it can be explained by an analysis of the terms” “Itaque cujuscunque veritatis reddi potest ratio, connexio enim preadicati cum subjecto aut per se patet, ut in identicis, aut explicanda est, quod fit resolutione terminorum” (Leibniz 1679: 295, 296). Elsewhere, Leibniz says that “of every truth (which is not immediate or identical) it is possible to give a reason” (Leibniz 1688: GP VII, 199), thereby implying that we cannot give a reason for immediate or identical truths. Below, in the next section, where the distinction is made between explicable and inexplicable reasons, it will become clear that Leibniz is not contradicting himself: although we cannot give a reason, this is not to deny that there is a reason. As far as scientific knowledge is concerned, Locke’s focus on the temporal order of truths makes it impossible for him to understand that scientific truths need to come in an order that is to provide first-person understanding. Locke’s central epistemic question of how one has arrived at certain truths is not free from psychologism. Leibniz’s first-person account of knowledge of necessary truths, aiming at understanding, sharply separates the psychological question of how we come to know a certain truth from the logical question of how these are grounded in the logical, demonstrative order of truths. Leibniz’s logical order is at the same time an order to be determined from a first-person point of view; necessary truths, though ultimately grounded in the existence of a necessary substance, can only be obtained by turning inwards, for “the pattern for the ideas and truths . . . [is] engraved in our souls” (NE 447). Leibniz thus goes beyond a consistent first-person point of view, for he is presupposing here the existence of a necessary substance, providing for a metaphysical order of truth being engraved in our soul. Locke’s perceptual metaphor makes the mind too passive, and, Leibniz adds, Locke is thus not able to explain how we can have knowledge of general, necessary truths (NE 49). We rather need a first person that is cognitively active: I am not a writing tablet or like wax (NE 110). This may be an unfair criticism, as for Locke, in the act of perceiving, the mind compares two ideas, and in judgment, the mind has to weigh the pros and cons, but it is true that Locke is not able to prevent a form of psychologism.

5.

Judgment and Reason

Although Locke and Leibniz give a different role to judgment in their epistemology, we see that both stress the relation between judgment and reason. We have seen that judgment is, for Locke, like balancing the

Locke and Leibniz on Judgment

55

account of reasons (E IV.xv.5: 656; cf. E II.xxi.67: 278). Leibniz develops Locke’s point by arguing for a logic that acknowledges degrees of probabilities: “balances . . . are needed to weigh likelihoods and to arrive at sound judgments regarding them” (NE 466). Besides the metaphor of a balance, Leibniz gives the metaphor of “games that combine chance with reason” (Leibniz 1688: 267; GP VII 201). Both indicate the possibility of an exact estimation of the degrees of probability. For Locke, our dim light of reason will never give us the certainty of knowledge regarding probabilities, but for Leibniz, this does not mean that we cannot come with exact measures and probabilities regarding predictions of future events. In general, and not merely with respect to probabilities, for Leibniz, “judgment consists in the scrutiny of propositions in accordance with reason” (NE 141). “A reason is a known truth whose connection with some less well-known truth leads us to give our assent to the latter” (NE 475). In the ideal case, the epistemic reason for our judgment is the reason or ground of the truth itself in accordance with the natural order of things and the explanatory order of truths. As the Aristotelian tradition, well known to Leibniz, puts it, the ratio cognoscendi for our judgment would also provide the ratio essendi for the truth of the judged proposition. For Leibniz, there is thus an essential connection between judgment and reason. Locke’s first definition of judgment, quoted at the beginning of section 3, does not mention reason. In the chapter on reason, though, Locke clearly relates judgment to reason, where a reason is, for Locke, an idea: “Judgment is the thinking or taking two Ideas to agree or disagree, by the intervention of one or more Ideas, whose certain Agreement or Disagreement with them it does not perceive, but hath observed to be frequent and usual” (E IV.xvii.17: 685; cf. E IV.xv.3: 655). In judgment, there is always a third idea involved, providing the reason for the judged proposition. Because there is no direct perception of the relation between the two ideas in judgment, there is always a third idea needed, which functions as the middle term in a syllogism. Acts of judgment are thus essentially acts of (probable) inference. Such a third idea, though, may be completely accidental to the content of the judgment, as when we make a rash judgment, resulting in mere opinion (E IV.xvi.3: 659). At first sight, it seems that Leibniz acknowledges acts of judgment for which there is no reason: judgments of taste, judgments about colours— for example, I am persuaded that a flower seems yellow to me.7 It is not true, though, according to Leibniz, that these judgments are completely without reason: But the reasons for our persuasion are of two kinds: some can be explained, the others are inexplicable. Those I call explicable can be proposed to others by a distinct reasoning; but inexplicable reasons consist only in our consciousness or perception, and in an experience

56

Maria van der Schaar of inner feeling which we would not be able to make others partake, if we do not find the means to make them feel the same things in the same way. For example, we are not always capable of telling others what it is that we find agreeable or disagreeable in a person, in a picture, in a sonnet, or in a stew. (Leibniz 1690: 311)

The reasons for these judgments are inexplicable. We cannot explain them in more basic terms (Leibniz 1690: 311), and we cannot explain them to others. In a similar way, the reason for a first truth of reason, an identity, is the experience of the fact that we see that the two terms of the relation of identity are identical. Each has to see for himself that this is an identity and there is nothing more to be explained. It is true that in the case of an identity, being a truth of reason, the reason is the same for everyone; the identity is evident per se, as we have seen above. Still, each has to make the truth evident to himself in order to grasp the identity as a known truth. For Leibniz, there are reasons that cannot be told to others; each has to see for himself, and in this sense, knowing a basic truth is essentially first-person. As far as judgment is concerned, Locke demands that reasons are explicable to others. Judgment as balancing an account is, for Locke, being accountable to others; the judging agent is an accountant, and the books should be open to others. Locke’s notions of right judgment and justified judgment are both third-person. Others should be able to evaluate whether one’s judgment is justified, as can most clearly be seen in the discussion of faith. As both Locke and Leibniz consider faith to be a form of belief or judgment, the above applies to matters of religious belief. For Locke, faith falls under the category of judgment, and faith is thereby not only fallible but also connected to reason. One’s faith, being firm assent, may not be very sensitive to arguments, but it should be. Locke’s central argument for mutual toleration is that it may improve one’s faith. By learning from other religions, one can hear the grounds these believers have for their faith, and one is thereby able to improve one’s own faith (Locke 1689: 59, 79; cf. Schaar 2012: 60–2). Faith and judgment are involuntary; therefore, faith cannot be commanded. At the same time, one is responsible for the faith one has; for each of us can influence the process towards judgment and faith, and it is precisely in this process that we collect the arguments pro and con. Leibniz summarizes and agrees: “a man is not responsible for having this or that opinion at the present time, but . . . he is responsible for taking steps to have it or not have it later on. So that opinions are only voluntary in an indirect way” (NE 456). In the fourth edition of the Essay from 1700, the edition that formed the basis of Coste’s translation into French and that was used by Leibniz in his extensive reworking of the Nouveaux Essais, Locke added a chapter on Enthusiasm (E IV.xix: 687–706). The enthusiast takes his persuasions

Locke and Leibniz on Judgment

57

as coming directly from God. How are the enthusiasts to be blamed if faith is involuntary? Their problem is that they neglect the rational aspect of faith. What is their reason that it is a revelation from God? “If they say they know it to be true, because it is a Revelation from GOD, the Reason is good: but then it will be demanded, how they know it to be a Revelation from GOD” (E IV.xix.11: 702). In the end, their reason is merely that they strongly believe it to be true. To them, it should be replied that either God will make a truth known through natural reason or he will give us “some Marks which Reason cannot be mistaken in. Reason must be our last Judge and Guide in every Thing” (E IV.xix.14: 704). Leibniz’s comments on the chapter show that he is, like Locke, critical of those people who use their powerful imagination aroused by passion to form sects: Their disputes show, at the least, that their inner witness needs outer verification if it is to be believed. (NE 507) At the same time, and in contrast to Locke, he seems to allow for utterances directly inspired by God: Still, such inspired utterances could bring their proofs with them; this would be the case if they truly enlightened the mind through the important revelation of some surprising truth which was beyond the powers of the person who had discovered it, unless he had help from outside. (NE 507) In the next chapter, Leibniz explains how this can happen: “inward grace will be making up for the absence of rational grounds for belief” (NE 510). “[T]hose who claim they find a divine inner light within themselves . . . base themselves on inexplicable reasons” (Leibniz 1690: 311). Michael Losonsky has shown that the contradiction in Leibniz is only apparent. In the second case, what is at stake is “the epistemology of religious belief from, roughly speaking, an internalist and first-person perspective” (Losonsky 2012: 718). One has in such a case “divine faith”, faith inspired by God, “a primary truth of fact that is immune from further justificatory requirements” (idem). It is based on reason but is not an explicable reason accessible to others. From a first-person point of view, one is fully rational when accepting such an inspiration from God. From a third-person point of view, though, “this alleged inner light is scarcely reliable” (Leibniz 1690: 311). “The way these people clash with one another should further convince them that their alleged ‘inner witness’ is not divine, and that other signs are required to confirm it” (NE 507). The terms ‘alleged’ and ‘reliable’ show that Leibniz is now speaking from a third-person point

58

Maria van der Schaar

of view. So, there is no contradiction involved, and it shows that Leibniz was aware of the distinction between the first- and third-person points of view. The first-person point of view is not open for Locke here because, for him, faith falls under judgment, and there are no direct, inexplicable reasons for judgment. One is essentially held accountable by others. As there is no human being who can provide the standard for right faith for us, there is nothing but human, fallible faith. Only on Judgment Day will our faith be evaluated. It is important to note that there is a crucial difference between the logical first-person point of view and the first-person point of view with respect to taste. Whereas the latter may differ for different people, and is thus subjective, the logical point of view is the same for all: each agent can see for himself that the law of identity is a truth, and we are thus able to know the same truth. Only the logical point of view is able to give a first-person point of view that is not subjective. The question whether one finds a divine inner light within oneself can only be answered by invoking the idea of God’s help, as Leibniz writes, and is therefore not consistently first-person.

6.

Judgments, Propositions, and the Particles

There is in the third book of Locke’s Essay a fascinating chapter on particles, and such particles also play a role in Leibniz’s universal language. Particles may be the signs of mental acts, and they may therefore be relevant for the question of whether there is a role for a sign of judgmental force in the universal language and thus for the question of whether the distinction between asserted and unasserted propositions is reflected in the philosophical grammar. But, there is also a risk of psychologism when particles signifying mental acts become part of logic. For both Locke and Leibniz, particles form a wider category than the Scholastic notion of syncategorematic term. The latter is a purely semantic notion explained as a term that can have meaning only together with other terms. The broader notion of particles may have been influenced by the rhetorical, humanist tradition, with its sensitivity for the different uses of language. Particles may be syncategorematic terms like ‘All’, ‘not’, and ‘or’, and they may be signs for relations between sentences. According to Locke, some signs, such as ‘Is’ and ‘Is not’, stand for the mental act of affirmation or negation, which connects the parts of propositions, the Ideas, into a proposition. Other particles, such as ‘but’, are signs of the actions of our mind in discourse, uniting several Affirmations and Negations “in one continued Reasoning or Narration” (E III.vii.2: 471). For Leibniz, the particles may also show the correct analysis of the ideas in a proposition: “particles connect not only the component propositions of a discourse, and the component ideas of a proposition, but also the parts of an idea made up of other ideas variously combined” (NE 330). In the

Locke and Leibniz on Judgment

59

latter case, particles are of importance for our capability to give a new demonstration through a new analysis of the ideas involved. Analysis of concepts and new definitions does not result in trifling propositions but is fruitful and gives us new insight because it makes new demonstrations possible.8 As Robert McRae has noted, for Locke, the importance of the particles lies not with logic but with philosophy of mind (McRae 1988: 157). For Leibniz, though, the particles are primarily a part of logic. Particles play an important role in the universal language because they reveal the various forms of the understanding (NE 330, 333). What are these forms? Not all linguistic words have a counterpart in the universal language: “genders are of no significance in philosophical grammar” (NE 330). What particles do we need in the universal language, and how do we determine their precise meaning? Take Leibniz’s example ‘Peter and John are learned’. We have to find a counterpart in the universal language for ‘and’. From the proposition that Peter and John are learned, it follows that Peter is learned and that John is learned. We may thus infer each of the members of the conjunction from the proposition that Peter and John are learned. The meaning of the particle ‘and’ in the universal language should reflect these possible logical inferences. A possible valid inference becomes “apparent from the characters themselves” (Leibniz 1686: 144). Only those words in natural language that are logically relevant will be represented by particles in the universal language, and these particles are to be understood in terms of possible inferences. For Leibniz, the forms of understanding expressed by the particles are those forms that are logically relevant. A similar point is made by Frege in his Begriffsschrift. The distinction between the passive and active forms of a sentence does not have any influence on the possible inferences we can make with it, which means that the distinction is not reflected in Frege’s ideography (Frege 1879: §3). There is a point where the topic of particles connects with the role of a first-person perspective in logic. For Frege, the sign for assertive force, the judgment stroke, is essential to logic, for there is no other way for us to express that a content is true (Frege 1915: 271, 272). Furthermore, the judgment stroke is a sign that the writer himself makes the judgment; a distinction between an asserted and an unasserted proposition is thus reflected in logic. The judgment stroke, though, does not include a name for the judging agent because no empirical fact is described. The sign is unique, showing, not describing, that a judgment is made. Because all axioms and theorems in Frege’s ideography are preceded by the judgment stroke, his logic is essentially first-person. This need not imply a form of psychologism because the judging agent is not an empirical notion. The asserter makes a truth claim and thereby claims that the content can correctly be asserted by anyone understanding the content or having made the demonstration. It is not relevant that Frege has made these assertions. To put it another way, the judging agent in Frege’s logic is not an empirical

60

Maria van der Schaar

subject but a logical subject (cf. Schaar 2018). As Leibniz is generally understood as making the distinction between asserted and unasserted propositions (cf. Barth forthcoming), one may ask whether the distinction is reflected in Leibniz’s universal language by a particle similar to Frege’s judgment stroke. Already at first sight, some of Leibniz’s particles show a congeniality with Frege’s sign for assertive force. Leibniz acknowledges ‘Adverbia Assertionis’, such as ‘Whether’ (An), ‘Yes’ (Ita), and ‘Certain’ (Certe). “An is a means of asking which sign of assertion or pronouncement has to be posited” (Leibniz 1686: 153). The particle ‘an’ is a sign for interrogative force—more specifically, a sign for asking a whether-question. And when we have found the answer to such a question, we are entitled to use the particle ‘Yes’ in front of the following proposition: “to every proposition which counts as an answer one of the signs of affirmation and denial must be prefixed” (idem). The act of judgment may thus be understood as an inner saying of yes or no (Nuchelmans 1983: 221). And there are several epistemic modifications of assertive force, such as certe and forte. In contrast to logical systems that are mere calculae, a sign of asserted force is needed in a logic conceived as universal language. As Ita is a sign for the act of judgment, the judging agent is taken account of within Leibniz’s universal language. Such a sign of assertive force is not to mean that the writer (here Leibniz) has actually made the judgment, for that would be a mere psychological fact. The ‘Yes’ sign is to be understood as a nonsemantic sign, a sign that whoever understands the system up to now and has accepted the relevant axioms, definitions, and demonstrations will be entitled to make the judgment at this particular moment. Leibniz’s understanding of the particles is primarily a logical one, but it exceeds the scholastic, semantic understanding of these terms insofar as it reflects an interest in the use of language in its full logical sense. Leibniz’s account of particles for assertions and questions exceeds, though, Frege’s use of the judgment stoke, for Leibniz also introduces a sign of interrogative force. Leibniz is able to relate questions and judgments to each other in his logic. They have a common core—the proposition—so that each judgment can be understood as an answer to a whether-question (NE 368).9 Questions come in two sorts: yes/no or whether-questions and wh-questions— ‘by whom and how’, including the important ‘why’ questions (idem). This interest in assertions and questions seems to be motivated by Leibniz’s interest in the art of controversies. The main aim of the universal language is to present the natural order of things, but it can also be used to solve controversies concerning contingent truths that cannot be decided by the calculus. Here, the language is in need of signs for the different acts put forward in a dialogue. As we would put it today, speech act theory can play a role in a logic of controversies. Leibniz also makes room for the idea of a dialogue in some of his explanations of fundamental logical notions like demonstration: “to ‘demonstrate’ is to reason from what

Locke and Leibniz on Judgment

61

ought to be granted. Those propositions ‘ought to be granted’ which those who speak with one another have agreed to grant” (Leibniz 1676: 55).10 He then elucidates these notions in terms of ‘speaker’ and ‘hearer’. Besides Michael Losonsky, Christian Barth has recently written on judgment in Leibniz (Barth forthcoming). According to Barth, “Affirmative judgments make propositions available as premises for future reasoning in response to a whether-question” (Barth forthcoming: section 2.2). Each answer to a question can function as a reason, a known truth (NE 475), for other judgments to be made. Because Leibniz has a sign of assertive force, he is able to make visible the distinction between asserted and unasserted propositions in his universal language. He acknowledges that different acts—judging and questioning—may have the same proposition as their content. The presence of a sign for assertive force in the universal language shows that Leibniz’s notion of proposition differs in an important sense from Locke’s concept of proposition. In “Locke and Arnauld on Judgment and Proposition” (Schaar 2008), two notions of proposition are distinguished, and it is argued that Locke’s notion of proposition is an example of the second, traditional notion: (1) the modern notion of proposition, to be represented by a that-clause (or, in Latin, by an accusativus cum infinitivo [ACI] construction) (2) the traditional notion of proposition, a declarative sentence together with its meaning or its mental counterpart, which is not to be represented by a that-clause but by a full declarative with the declarative mood Because the latter notion includes (the meaning of) mood, it is not apt to account for the fact that we may have different propositional attitudes with the same content. On Locke’s account, questions, orders, and judgments are made manifest by interrogative, imperative, and declarative sentences, respectively, and there seems to be no common core, besides the ideas involved. For Locke, there is no straightforward account for different propositional attitudes, as he does not have the modern notion of proposition. This is not to say that Locke is not able to make the distinction between asserted and unasserted proposition: a full declarative sentence may also occur unasserted.11 The mental act that unifies the ideas into a proposition need not be an act of judging (or perceiving). Although the standard view is that Locke needs the act of knowing or judging to unify two ideas into a proposition, I have argued that one may read Locke in a different way (cf. Schaar 2008). Locke seems to allow for propositions not yet judged or known. If the mind proceeds rationally, it “ought to examine all the grounds of Probability . . . for or against any probable Proposition, before it assents to or dissents from it” (E IV.xv.5: 656).

62

Maria van der Schaar

Leibniz, in contrast, acknowledges that the same proposition may be the content of a judgment and of a question as well. He is also willing to express the proposition A is B as A’s being B (that A is B, tò A esse B). In this sense, there is in Leibniz a forerunner of the modern notion of proposition, also in the sense that the proposition is independent of language and individual thoughts. For Leibniz, the distinction between asserted and unasserted proposition is of crucial importance. If one makes the proposition ultimately dependent upon a mental act of unifying, as in the case of Locke, whether this is an act of judgment or any other mental act, one makes the bearer of truth and falsity dependent upon the human mind. In his argument against Locke’s notion of proposition as being either verbal or mental, Leibniz claims that truths are independent of language and the human mind. We need a notion of truth-bearer that is independent of the signs we use, whether these signs are verbal or mental (NE 397). Does Leibniz also acknowledge that false propositions are independent of the human mind? Yes. In the earlier dialogue on the connection between words and things, one of the characters claims that truth and falsity belong to possible thought, not to actual judgments or thoughts (Leibniz 1677: 190). Although these possibilities are in us only in an implicit way, they are actualities in God’s mind, for God thinks everything that does not contain a contradiction. Leibniz’s notion of cogitatio possibilis was recognized by Bernard Bolzano as a forerunner of his notion of Satz an sich (Bolzano 1837: I, § 27), and it may thus also be considered a precursor of Frege’s Gedanke. From Leibniz’s point of view, the traditional notion of proposition, being dependent on language and the human mind, is not able to account for the objectivity of logic. Instead, he proposes a modern notion of proposition as independent of language, but this does not mean that the notion of judgment—and a corresponding sign of assertive force—is excluded from the universal language.

Conclusion The Cartesian first-person point of view regarding the certainty of our judgments seems to imply a form of subjectivism. Locke’s solution to this problem is to widen the scope to a third-person point of view on the notion of judgment, thereby shaping the modern notion of belief. As a result, the normative notion of judgment and the empirical notion of belief are no longer distinguished, and epistemology and philosophy in general are threatened by a form of psychologism. Leibniz’s solution to Cartesian subjectivism is to strengthen the idea of first-person knowing. For Leibniz, the first-person point of view is essentially a logical point of view. As the act of judgment is not confused with the empirical notion of belief, and asserted and unasserted propositions are clearly distinguished, there is no threat of psychologism in Leibniz’s logic.

Locke and Leibniz on Judgment

63

Notes 1. I thank Christian Barth for comments on an earlier version. 2. In Schaar 2018, I have argued that the presence of the judgment stroke, a sign that a judgment has been made, in Frege’s logic does not imply a form of psychologism. 3. As Nicholas Jolley puts it, “For Locke, as for Descartes, knowing is knowing for oneself; no one else can know for me” (Jolley 2016: 59). 4. As Tyler Burge notes, “The formula of basic truths and axioms neither needing nor admitting of proof can be found verbatim in Leibniz, from whom Frege surely got it” (Burge 2000: 362). 5. This section is a reformulation in terms of the first-person perspective of some of the results of my paper “Locke on Knowledge and the Cognitive Act” (Schaar 2009). 6. Although the influence of Leibniz on Frege’s logic may be disputed, Frege carefully read the Nouveaux Essais when presenting the philosophical account of his ideography in the Grundlagen der Arithmetik (1884), as Tyler Burge (2000) has pointed out. Frege’s critique of the traditional notion of judgment and truth in “Der Gedanke” (1918) seems to be influenced by Leibniz’s critique of Locke’s position (NE 396). 7. Leibniz uses the term ‘judgment’ in a stricter sense, as essentially related to reason, and in a wider sense, including assent to primitive truths. Cf. Barth (forthcoming, note 39). 8. There is here an agreement with Frege’s idea of fruitful definitions (1884: §88). There is an important difference, though, between Leibniz and Frege regarding the analysis of contents and the possibility of new definitions. For Frege, analysis applies to judgeable content, and, regarding the same judgeable content, it may yield different functions with a different number of arguments, thereby making a proper account of relations and multiple quantification possible. For Leibniz, analysis applies not to the judgeable content as a unity, but to the concept functioning as subject term; it thus applies to pregiven concepts or terms. 9. In NE 356, Leibniz writes that questions have a special content known as themes, which are midway between ideas and propositions. It seems, though, that the whether-questions have a proposition as their content, while the wh-questions have a non-propositional theme as their content—that is, a proposition where a part is left blank—also called problems by the mathematicians (NE 368). 10. Apparently, to convince is also a logical term: “to ‘convince’ is to reason from what is granted” (ex concessis rationari [Leibniz 1676: 54]). 11. Jennifer Marušić presents a somewhat different view of Locke’s position. She argues that Locke is able to distinguish between an asserted and an unasserted proposition insofar as Locke allows for a distinction between a proposition judged by oneself and a judgment made by someone else without oneself judging the relevant proposition, for “Locke and the Port-Royalists implicitly distinguish between performing an act of affirmation or denial and conceiving of someone else’s act of affirmation or denial” (Marušić 2014: 274)—an interpretation that could be put in terms of the difference between the firstand third-person points of view introduced above.

References Barth, C. (forthcoming), “Judgement in Leibniz’s Conception of the Mind: Predication, Affirmation, and Denial”, Topoi, published online: https://link.springer. com/article/10.1007/s11245-016-9422-z

64

Maria van der Schaar

Bolzano, B. (1837), Wissenschaftslehre. Leipzig: Felix Meiner,1929–31. Burge, T. (2000), “Frege on Apriority”, in his Truth, Thought, Reason; Essays on Frege. Oxford: Clarendon, 2005. Frege, G. (1879), “Begriffsschrift; eine der arithmetischen nachgebildete Formelsprache des reinen Denkens”, in I. Angelelli (ed.), Begriffsschrift und andere Aufsätze. Hildesheim: Olms, 19712. Frege, G. (1884), Die Grundlagen der Arithmetik. Breslau: Koeber. Frege, G. (1893), Grundgesetze der Arithmetik. Jena: Hermann Pohle; facsimile Hildesheim: Georg Olms, 1998. Frege, G. (1915), “Meine Grundlegenden Logischen Einsichten”, in NS, 271–2. Frege, G. (1918), “Der Gedanke”, Beiträge zur Philosophie des deutschen Idealismus 1: 58–77. Frege, G. (1983), NS = Nachgelassene Schriften, ed. H. Hermes et al. Hamburg: Meiner. Jolley, N. (2016), Toleration and Understanding in Locke. Oxford: Oxford University Press. Leibniz, G. W. (1675), “On Mind, the Universe, and God”, in De Summa Rerum; Metaphysical Papers, 1675–1676, trans. and ed. G. H. R. Parkinson. New Haven and London: Yale University Press, 1992. Leibniz, G. W. (1676), “On the Elements of Thinking”, in De Summa Rerum; Metaphysicial Papers, 1675–1676, trans. and ed. G. H. R. Parkinson. New Haven and London: Yale University Press, 1992. Leibniz, G. W. (1677), “Dialogue on the Connection between Things and Words”, GP VII. Leibniz, G. W. (1679[?]), “De synthesi et analysi universali seu arte inveniendi et judicandi”, GP VII. Leibniz G. W. (1684), “Meditationes de cognitione, veritate et ideis”, GP IV. Leibniz, G. W. (1686), “Analysis particularum”, ed. F. Schupp, in Die Intentionale Logik bei Leibniz und in der Gegenwart, ed. A. Heinekamp and F. Schupp, Studia Leibnitiana, Sonderheft 8 (1979): 133–53. Leibniz, G. W. (1688[?]), “An Ars Characteristica for the Rational Sciences”, ed. M. Dascal, in G. W. Leibniz, The Art of Controversies, trans. and ed. M. Dascal. Dordrecht: Springer, 2006/ “Scientia Generalis. Characteristica. XIV”, GP VII. Leibniz, G. W. (1690), “Confronting the Catholic Hardliners A”, in G. W. Leibniz, The Art of Controversies, trans. and ed. M. Dascal. Dordrecht: Springer, 2006. Leibniz, G. W. (1704), NE = New Essays on Human Understanding, trans. and ed. P. Remnant and J. Bennett. Cambridge: Cambridge University Press, 1996/Nouveaux essais sur l’endentement humain, in G. W. Leibniz, Sämtliche Schriften und Briefe, Berlin: Akademie-Verlag, 1962, Vol. VI.6. First publication 1765. Leibniz, G. W. (1875–90), GP = Die philosophische Schriften von Gottfried Wilhelm Leibniz, ed. C. I. Gerhardt, Vols. IV and VII. Berlin, 1880/90. Locke, J. (1689), Epistola de tolerantia/A Letter on Toleration, ed. R. Klibansky, trans. J. W. Cough. Oxford: Clarendon, 1968. Locke, J. (1690), An Essay Concerning Human Understanding (E), ed. P. H. Nidditch. Oxford: Clarendon, 1975. Losonsky, M. (2001), Enlightenment and Action From Descartes to Kant. Cambridge: Cambridge University Press.

Locke and Leibniz on Judgment

65

Losonsky, M. (2012), “Locke and Leibniz on Religious Faith”, British Journal for the History of Philosophy 20: 703–21. Marušić, J. S. (2014), “Propositions and Judgments in Locke and Arnauld: A Monstrous and Unholy Union?”, Journal of the History of Philosophy 52: 255–80. McRae, R. (1988), “Locke and Leibniz on Particles”, Synthese 75: 155–61. Nuchelmans, G. (1983), Judgment and Proposition: From Descartes to Kant. Amsterdam: North Holland Publishing Company. Schaar, M. van der (2008), “Locke and Arnauld on Judgment and Proposition”, History and Philosophy of Logic 29: 327–41. Schaar, M. van der (2009), “Locke on Knowledge and the Cognitive Act”, Grazer Philosophische Studien 78: 1–15. Schaar, M. van der (2011), “Assertion and Grounding; A Theory of Assertion for Constructive Type Theory”, Synthese 183: 187–210. Schaar, M. van der (2012), “Locke on Judgement and Religious Toleration”, British Journal for the History of Philosophy 20: 41–68. Schaar, M. van der (2018), “Frege on Judgement and the Judging Agent”, Mind 127: 225–50. Schwitzgebel, E. (2015), “Belief”, in Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/entries/belief/ Searle, J. (1979), Expression and Meaning. Cambridge: Cambridge University Press. Williamson, T. (2000), Knowledge and Its Limits. Oxford: Clarendon. Wilson, M. D. (1967), “Leibniz and Locke on ‘First Truths’”, Journal of the History of Ideas 28: 347–66.

3

Kant’s Logic of Judgment Against the Relational Approach Alexandra Newton

Judging is often assumed to be a mental act or attitude that relates an individual subject of judgment (=I) to that which she judges. ‘Judge’ is thus thought to be a psychological, relational predicate attached to a subject (I). Let us call this the ‘relational’ approach to judgment. The relational approach presupposes a distinction between the act of judging (or ‘force’) and content judged, or between thinking and being. Thus, it requires distinguishing the psychological laws of thinking from the laws of being; reflection on the laws of thinking will not, as such, afford insight into the laws of being. In this paper, I will argue that Kant rejects the relational approach in his reflections on judgment in general logic, a logic that he thinks provides us with an exhaustive and complete ‘table’ of functions of judging (A70/ B95, A76/B102). The logical ‘I think’ (or ‘I judge’) is a representation of the non-relational logical act of synthesis internal to what is judged so that reflection on what is judged is a reflection on the act of judging it. This will have two consequences, to be discussed in the second and third sections of this paper. First, that which is judged—namely, what is—itself must be understood through the act of judging it. Thinking and being—in a logical sense of ‘being’ (the copula ‘is’)—are understood as originally the same. Second, it will emerge that the logical ‘I’ that judges does not refer to a particular epistemic agent that performs the act of judging, but is nothing other than the logical unity internal to the judgment itself. I will argue on these grounds that there is no basis, within general logic, for a distinction between thinking ‘p’ and thinking about thinking ‘p’, or between logic and psychology. The sameness of thinking and logical being (the copula) cannot be appreciated on a certain reading of the act of judging that has become standard in the literature on Kant, according to which judging is an act of comparison or reflection, through which I subsume lower representations under higher ones. In the first part of the paper, I will show how interpreters have attempted to derive this relational conception of judgment from Kant’s general logic and have supplemented it with another relational conception of judgment as an act of synthesis from Kant’s transcendental

Kant’s Logic of Judgment

67

logic. In the second and third sections, I will argue against these relational conceptions of judging, and of the I that bears these relations, on both exegetical and philosophical grounds. Kant has a unitary conception of judgment as a self-conscious act of synthesis, both in general and in transcendental logic. In conclusion, I will indicate how these points about the logic of judgment will enable the discovery of the structures of material or objective being from within the self-consciousness of thinking or judging.

1.

The Relational Approach to Judging

Kant appears to endorse two distinct accounts of judgment in his critical writings. First, a passage from the Amphiboly chapter of the first Critique suggests that he adopts a traditional conception of the understanding as a faculty that compares and reflects on representations with regard to their identity and differences, agreement and opposition, etc., in forming judgments:1 The relation, however, in which the concepts in a state of mind can belong to each other are those of identity and difference, of agreement and opposition, of the inner and the outer, and finally of the determinable and the determination (matter and form). [. . .] Prior to all objective judgments we compare the concepts, with regard to identity (of many representations under one concept) for the sake of universal judgments, or their difference, for the generation of particular ones, with regard to agreement, for affirmative judgments, or opposition, for negative ones, etc. (A261–62/B317–18) Since Kant ends this passage with “etc.” [usw.], he suggests that the concepts of comparison [Vergleichung] (identity, difference; agreement, opposition; inner, outer; determinable, determination) can each be aligned with the first two moments under the four titles on his table of judgments (universal, particular; affirmative, negative; categorical, hypothetical; problematic, assertoric). Thus, the functions of judging appear to be ways of comparing and reflecting on representations. But in other passages, Kant defines judgment as a synthetic unity or connection [Verknüpfung] of various concepts, suggesting that the functions of judging are ways of unifying representations: The unification [Vereinigung] of representations in a consciousness is judgment. (P §22 4:304) [Through judgment] many possible cognitions are [. . .] drawn together [zusammengezogen] into one. (A69/B94)

68

Alexandra Newton all categories are grounded on logical functions in judgments, but in these combination [Verbindung], and therefore unity of given concepts, is already thought. (B131)2

On standard readings of Kant’s conception of judgment, these passages provide two distinct conceptions of the act of judging.3 On the one hand, judging is an act of bringing representations to consciousness through comparison [Vergleichung] and reflection. On the other hand, judging is an act of synthesis or combination [Verbindung],4 through which I make a truth claim or assertion about objects in the world.5 Since “general logic”, according to Kant, completely abstracts “from all content of cognition, i.e. from any relation of it to the object”, it does not, on this reading, consider judgment to be an act of synthesis (A55/B79). I will argue in the second section that this is wrong; the act of analysis or reflection internal to any judgment is the self-consciousness of an act of synthesis. So, contrary to appearances, we do not have here two distinct acts in judgment (of comparison and of synthesis), but one: judgment is a self-conscious act of synthesis and can be understood as such, even in general logic. But first, I will spell out the relational approach, according to which judgment involves two relations: comparison and synthesis. 1.1.

Judgment as a Relation of ‘Bringing Representations to Consciousness’

The restriction of judgment to an act of comparison or reflection in general logic is often taken to be implied by Kant’s “explanation of judgment in general [überhaupt]” from the Jäsche Logic: Judgment is the representation of the unity of the consciousness of various representations, or the representation of their relation [Verhältnis] insofar as they constitute a concept. (JL 9:101) Henry Allison has followed Béatrice Longuenesse in interpreting the “unity of the consciousness” in this passage as an analytic unity of consciousness, which both authors say is first “produced” or brought about through an act of analysis (Allison 2004: 84; Longuenesse 1998: 81, 85f.). Analytic unity is the unity of a concept, a reflected representation of one feature or mark of things as common to (or “identical” in) many (B133). Concepts are formed through the logical operations of comparison, reflection, and abstraction, which Kant sometimes refers to collectively as “analysis” (A76/B102, JL 9:94, Refl. 2876 16:555). In comparing different representations of firs, elms, and birches, for instance, I reflect on that which is identical or common in all of them—such as ‘a trunk’ and

Kant’s Logic of Judgment

69

‘branches’—and abstract from ways in which they differ (the sizes of the trunk, shapes of the branches) to form the general concept . The concept that emerges from these logical operations, unlike intuitions, has a logical quantity or a “sphere” [Umfang, extension], i.e. it is a general representation that ‘contains’ other representations ‘under’ it. Judgment is an analytic unity of the consciousness of various representations because it brings to consciousness what is contained in, and common to, various representations through analysis. This holds both for analytic and for synthetic judgment. In analytic judgment, the predicate concept brings to consciousness a mark common to everything thought under the subject-concept; and in synthetic judgment, both subject and predicate concepts reflect something already contained in, and common to, the prediscursive, intuitive representation(s) of the object(s) the judgment is about.6 This appears to capture Kant’s meaning in the metaphysical deduction when he says that “all judgments are [. . .] functions of unity among our representations” (A69/ B94). A few lines earlier, he had defined a function as “the unity of the action of ordering different representations under a communal [gemeinschaftliche] one” (A68/ B93).7 As Longuenesse understands this passage, judgment orders a representation under a higher, common one that comprises the former in its extension, and thus is an act of subordinating representations under concepts: “several concepts, and with them the representations contained ‘under’ them, are thought under one and the same concept of greater universality” (Longuenesse 1998: 85, 111; see also MacFarlane 2000: 122; Longuenesse 2006: 140). Since Allison and Longuenesse assume that analytic unity of consciousness— what is common to many—is first produced by analysis, they must also say that the mark that is common to all of my representations, which makes them all collectively ‘mine’, first emerges through analysis. Prior to analysis, I am in my representations, but I do not yet own them—they are not yet ‘mine’; their becoming mine (or sharing ‘mineness’ in common) is due to an active effort on the part of the understanding. The activity of thinking thus extends beyond the original, prediscursive representations of sensibility, adding something to them that was not there before—namely, their mineness. This is the act that gives them analytic unity of consciousness; analysis “produce[s]” analytic unity (Allison 2004: 84; Longuenesse 1998: 81, 85f.).8 So we must distinguish the objects (of sensibility) that are represented from the (common) act of representing them under concepts. The ‘I think’ is a consciousness of the latter, unifying mental activity (in thinking), and is thus a psychological representation.9 Although a concept represents something other than myself (‘red’, ‘table’, etc.), concepts all contain, in addition, a representation of myself within them (‘being mine’) since they are general representations only insofar as I am conscious of them as mine—i.e. as containing analytic unity of apperception (B133n.). This reading faces the difficulty that it appears to make (general) logic subjective and incapable of becoming objective (as a transcendental logic).

70

Alexandra Newton

It is only in their relation to me, as a particular reflecting subject, that things in the world have general or common features at all: generality is first produced by my acts of analysis. The logical feature of concepts— their generality—in fact emerges as a merely subjective, psychological feature of subjects who can think. Likewise, the logical form of a judgment is understood as a subjective relation between a subject and her representations—namely, as the relation of bringing representations to consciousness as mine. Logical forms (of concepts and judgments) are subjective because they reflect the ‘mineness’ of representations, understood as distinct from their objective content. 1.2.

Judgment as a Relation of ‘Making an Assertion’

Notice that on the above reading of judgment as a relation of ‘bringing representations to consciousness’, judging is understood as a psychological act of mere thinking, for even when I merely entertain the thought ‘the table is brown’, without judging or staking a claim about whether this thought is true, I bring to consciousness what is represented by the concept ‘table’ as subsumed under what is represented by ‘brown’. In this sense, it is often thought that Kant treats judging as a mere act of thinking in general logic. Only in transcendental logic does he turn to a distinct conception of judgment as an act of making a truth claim. If the act of judging were merely an act of concept subordination, or of grasping [be-greifen] various representations under one concept [Begriff ], the unity of the judgment would reduce to the unity of whichever concept occupies the role of the predicate.10 But as Kant himself observes in a letter to Beck from 1792, the “connection of representations in a concept” is distinct from the “connection of representations in a judgment”. Kant marks the distinction by saying that in the concept of a ‘black man’, “one thinks of a concept as determinate”; but in the judgment ‘the man is black’, “one thinks of the activity of my determining of this concept” (Letter to J. S. Beck, 3 July 1792, Corr. 11:346). Judgments are themselves acts of determining a subject concept through a predicate concept, and thus are acts of combining subject with predicate, while concepts represent what is already combined.11 Observations such as these led Klaus Reich to proclaim that judgment in the above ‘explanation’ from the Jäsche Logic is a synthetic (not analytic) “unity of consciousness of various representations” because it is the unity of an act of synthesis, not of mere analysis (JL 9:101; Reich 1992: 10).12 Other interpreters maintain that this conception of judgment as a synthesis of representations becomes Kant’s focus in transcendental logic but was not yet present in his general logic (Allison 2004: 84f.; Longuenesse 1998: 73ff.). Through analysis, the concept brings to consciousness what is contained in an already determined lower representation: it reflects on the ‘black man’ as ‘black’. But this analysis presupposes an act of synthesis

Kant’s Logic of Judgment

71

that descends from the higher concept to the lower one by determining the higher concept: the genus ‘black’ gets determined in its species, ‘black man’. Now, if judgment merely represents the lifeless products of these acts of synthesis, then synthesis would only play the role of generating the combinations of marks that are brought to consciousness through analysis. But since judgment is distinct from the conceptual representation of what is already determined, judgment must be the activity of synthesis itself (the ‘activity of determining’). Thus, synthesis is not merely presupposed by judgment, as the act that generates the logically lifeless content that is brought to consciousness through a concept, but rather judging is itself a ‘living’ activity of synthesis. How must this connection or living bond of concepts in a judgment be understood?13 Klaus Reich has argued that Kant understood it as an act of assertion through which concepts are related to an object (Reich 1992: 71ff.).14 Whereas a mere concept represents one complex thing (black man), a judgment makes a truth claim by asserting one thing of another (the man is black).15 Assertion says that the judgment “is considered actual (true)” (A75/B100). Even the analytic judgment ‘the bachelor is unmarried’ differs from the mere concept ‘bachelor’ since the former makes an assertion, while the latter does not. Kant emphasizes this aspect of judgment when he says that a judgment is a “rule” (P 4:306). A rule is defined as an “assertion under a universal condition” (JL 9:121; see Longuenesse 1998: 93). Kant adopts the phrase ‘assertion under a condition’ from Christian Wolff, who also uses it to express the relation a judgment bears to truth.16 Wolff says that the ‘condition’ [Bedingung] in a judgment articulates the “ground or principle [. . .] from which I may know, that in a given case the Proposition holds” (Wolff, German Logic III §8, 69). The condition lies either explicitly or “concealed in the subject”, while assertion (or assertoric force) is associated with its predicate insofar as the predicate contains the copula, that is, the act through which the subject is determined (ibid., 68). We have now distinguished two different roles for synthesis to play in judgment. First, synthesis produces the ‘unities of marks’ that are reflected, or brought to consciousness, through a concept. This role of synthesis was presupposed by the first conception of judgment since its lifeless products constitute the complex representations brought to the ‘I think’ through reflection and comparison. Second, synthesis is the (‘living’, articulated) activity of combining representations in making an ‘assertion’.17 Consciousness of the first role of synthesis guarantees consciousness of the second in analytic judgment, for in becoming aware of the marks contained in a subject concept S, I am already conscious that P holds of S, and thus already assert P of S. Analytic judgments are, we might say, semantically self-contained: they contain their truth conditions within themselves, since the condition of assertion lies in the content of the subject-concept.18 But in synthetic judgment, the two roles of synthesis do seem to come apart;

72

Alexandra Newton

for I can be aware of what the judgment says of the object—i.e. what combination of marks it supposes to be contained in the object subsumed under both subject and predicate concepts—without being aware that the predicate concept can be asserted of the subject concept. The grasp of the content and the assertoric force of synthetic judgment thus appear to be two separate acts. According to Longuenesse and other proponents of the relational approach, the possibility of synthetic judgment requires Kant to introduce a new kind of logical ‘condition’ into his conception of judgment: “Wolff’s characterization of the condition of a judgment is not acceptable to Kant” (Longuenesse 1998: 96). Although Wolff was right to say that concepts serve as the conditions of analytic judgments, the possibility of synthetic judgment requires that we think of a third thing outside of the subject and predicate concepts as the condition grounding the validity of the assertion. Thus, Longuenesse maintains that, for Kant, “sensible intuitions [. . .] are [. . .] ultimately the true conditions of judgments” (Longuenesse 1998: 97). Lanier Anderson, in his recent book on Kant, argues along similar lines that Kant’s impetus for introducing the distinction between analytic and synthetic judgments was that the truth of synthetic judgments rests on a non-logical ground: an object, not a concept (Anderson 2015: 198; see also Beck 2002: 5).19 It is now the object, not a concept, that provides the ‘glue’ or “cement” for the living bond or combination of representations in synthetic judgment (Longuenesse 1998: 107). We have seen that the relational approach to judgment distinguishes two different acts in synthetic judging. In general logic, judging is understood as a relation of ‘bringing representations to consciousness’ or of ‘subsumption’, while in transcendental logic, judging is understood as a relation of ‘purporting to be true’ or of making a claim about what is the case. Both conceptions of the act of judging distinguish it from what is judged. The first distinguishes the contents of representations from the act of bringing them to consciousness, while the second distinguishes the contents of judgment from the act of asserting them to be true. In Longuenesse’s words, “In [Kant’s] usage ‘judgment’ refers, on the one hand, to the act of judging, and on the other hand, to the content of the act (what we would call the proposition)” (Longuenesse 2006: 164n.). On the first conception of judgment, the content of judgment can be identified with its object, while the second conception distinguishes the object from the content. Thus, the two conceptions of judgment require two senses of ‘object’. On the one hand, there is the object subordinated under concepts (the intensional object), and on the other hand the object of reference (extensional object). Longuenesse draws this distinction between two senses of ‘object’ as follows: “objects [are] defined both logically as instances of concepts, and intentionally as what our representations are representations of (the intentional correlate of our representations)” (Longuenesse 2005: 204).20 Although Longuenesse thinks that the nature

Kant’s Logic of Judgment

73

of the intentional object is not a topic for general logic but only for transcendental logic, she thinks that reference to the intentional object must be present in the logical form of judgment in general to ensure that this is the form even of synthetic judgments: “every judgment of the form ‘A is B’ can be developed into the following form: ‘To everything x, to which the concept A belongs, belongs also the concept B’” (Longuenesse 1998: 86–7, 106).21 This reading not only places the I outside of the content judged, as the subject that bears a relation to that content (the relation of judging it), it also distinguishes the intentional object from the content as that to which the content refers. The logical copula ‘is’ in a judgment thus carries an ambiguity within it: it can either express the act of synthesis or combination performed by the subject (the act of assertion), or it can express what is the case concerning the object judged about. It either expresses thinking or it expresses being. In the following, I will challenge this reading and argue that the copula is unambiguous because thinking and being are the same, at least in Kant’s general logical discussion of judgment.

2.

Synthetic Approach to Logical Form

We can begin by first reflecting on the logical copula, the verb ‘to be’, that is employed in any judgment and most paradigmatically in categorical judgments of the form ‘S is P’. Kant himself isolates a general logical sense of ‘being’ in section 12 of the first Critique, in a passage leading up to the transcendental deduction. The “transcendental philosophy of the ancients”, he notes, distinguished three senses of being (ens): “quodlibet ens est unum, verum, bonum” (B113). Being can mean ‘unity’, ‘truth’, and ‘perfection’. The first two senses of being, which I will focus on here, can be lined up with predicative being and veritative being, respectively. To judge ‘S is P’ is to think of ‘P’ as belonging to ‘S’ in one predicative unity (predicative being) and to think that ‘S’ is truly ‘P’ (veritative being). The verb ‘to be’ thus expresses both unity and truth. Kant’s point in this passage is that the tradition was wrong to treat ‘being’ in these senses as ontological determinations. Instead, these three “supposedly transcendental predicates of things are nothing other than logical requisites and criteria of all cognition of things in general” (B113– 14). That is, the logical copula (‘is’), or logical being, expresses ‘unity’ and ‘truth’ not as features of the object cognized, but as logical requisites of the act of cognizing them. Indeed, as I will argue here, logical unity and logical (or formal) truth are inseparable from acts of cognition. Predicative or propositional unity is itself the unity of an act of synthesis in thinking, i.e. there are no unified, propositional contents outside of that self-conscious act. And truth is the self-agreement of an act of cognition— i.e. there are no truths outside of self-consciousness of acts of cognizing.

74

Alexandra Newton

Being (in the logical senses of ‘predicative being’ and ‘being true’) and thinking (or cognizing) are, I will argue, the same. 2.1.

Thinking and Predicative Being (Unity)

First, a judgment is not a mere list of representations but a “unity of the consciousness of various representations” (JL 9:101). We saw that Allison and Longuenesse interpret this to mean that it is an analytic unity of consciousness that results from analysis (comparison and reflection) of various representations. But in the opening passages of the B-edition transcendental deduction, where Kant is concerned only with logical acts of the understanding in abstraction from sensibility, Kant says that the “analytical unity of apperception” presupposes not analysis but a synthetic unity of apperception (B133). The analytic unity of apperception is said to be the “identity of consciousness” in various representations—not the reflection of that identity through analysis (ibid.). Representations are ‘mine’ even prior to the act of the understanding in analysis—i.e. they already contain something in common. Analytic unity presupposes synthetic unity because what all my representations share in common— what it is that is identical in all of them and that makes them all mine—is the ‘synthetic unity of apperception’: the possibility of their unity in one consciousness, one act of synthesis; for it is “only because I can comprehend their manifold in a consciousness [that] I call them all together my representations” (B134). Thus, Kant says in section 15 of the B-deduction that the “ground of the unity of diverse concepts in judgment”—what holds them together in one unity of consciousness, one thought or predicative unity—is to be found not in the object, but “higher” even than any concept of an object or category—namely, in the “original synthetic unity of apperception”. This unity is original because it is identified with a “faculty”—namely, the “understanding” (B134n.), which gives rise to, and thus originally grounds, acts of synthesis.22 Kant identifies this original unity with the “qualitative unity” of cognition from section 12, which he had said is “the unity of the comprehension of the manifold of cognition [. . .], as say, the unity of the theme in a play, a speech, or a fable” (B114). The unity of theme in a speech is a unity that precedes the parts of a speech by governing the unfolding of the speech into its parts and by giving them each a function to play in its articulation.23 Just as the same theme is contained in the different parts of the speech, it is the same original synthetic unity of apperception contained in the different parts of a judgment. Kant characterizes the representation of this unity, the representation ‘I’, as a “simple” rather than complex unity because the original ground of unity in a judgment is not like the unity of an empirical intuition (KrV B135, A345–46/B404). It doesn’t contain parts in the way in which my intuition of a book is a complex whole that comprises intuitions of its

Kant’s Logic of Judgment

75

cover and of its pages. It is instead the ground of an act of synthesis that ‘first divides’ (ur-teilen) into parts (Teile). Although this act thus contains different parts, such as the subject and predicate concepts in a categorical judgment, this does not jeopardize the simplicity and indivisibility of its original unity; for these parts are not to be understood as parts out of which the whole is constituted, but rather they are parts that emerge out of its unity, just as the parts of the speech emerge out of its theme or—to use another analogy that Kant sometimes appeals to—the functions of the organs in a body each articulate the function of the whole. Like the expansion and contraction of the lungs that each contribute to breathing, the use of the subject and use of the predicate concepts in a judgment contribute to the function of the judgment as a whole. They are each ways in which unity is given to a manifold or in which the simple original synthetic unity of apperception executes its function (namely, cognition).24 Now, thoughts do not have predicative unity (or predicative being) independently of consciousness of unity (like the mind-independent propositions of some Fregean philosophers); for if they did, their unity would have to be given to consciousness. But if, in addition to the representations unified (in a categorical thought), their unity is also given, the unity of these three items would also have to be given, thus launching an infinite regress. In section 15 of the transcendental deduction, Kant, for this reason, asserts that “combination” is “among all representations [. . .] the only one that is not given through objects but can be executed only by the subject itself, since it is an act of its self-activity” (B130). Although the combination or unity of a thought cannot be given, it also cannot be a (dead) product of an act of synthesis; for if the predicative unity of a thought resulted from the synthesis of its elements, this unity would be a third thing relating the two parts to one another, and one could then ask what unifies these three things together in one thought about an object, thus again launching an infinite regress. (This is similar to the Bradley regress that arises from thinking of the copula in a judgment as a real relation, for example, of exemplification.) Thus, one can say that the unity of thinking ‘the table is brown’ is already contained in, and common to, both the use of the subject concept ‘table’ and the use of the predicate concept ‘brown’; the unity of the whole judgment is, in this way, prior to its parts. We judge from a consciousness of the predicative or propositional unity of our judgments and not with the aim of producing their unity. Thus, we can say that a judgment is a synthetic unity—i.e. the unity of an activity of synthesis—that is constitutively self-conscious. Since comparison and reflection are the acts of bringing representations to the ‘I think’, even synthetic judgment is thus at once both an activity of synthesis and of comparison and reflection (analysis): there are not two acts here, but one. The act of analysis or reflection is not a second, higherorder awareness of a first-order, synthetic unity but is identical with the

76

Alexandra Newton

synthetic unity it is conscious of, i.e. it is a self-consciousness. Therefore, I cannot think of what is thus and so without thinking my act of thinking it—there is no consciousness of a predicative unity (something’s being thus and so) apart from self-consciousness in thinking it. So there are not two topics here—one of psychology or epistemology that investigates the act of thinking (or the correctness of such acts) and one of logic that investigates what is thought about (and the truth of such thoughts). Rather, the investigation of what is thought about is the same as the investigation of the act of thinking it. There is no distinction here between thinking and predicative being and thus no distinction between consciousness of thinking and consciousness of being. Contrary to the first relational account of judgment, predicative unity thus is not the unity of a dead complex content to which a subject bears a relation in judging. Rather, predicative unity is the unity of a ‘living bond’ or activity of synthesis. In judgment, this act of synthesis is the same as an activity of assertion through which I am conscious of the truth of what I judge. So there are not two separate unities in judgment—one of content and one of force or the act of asserting it—but a single unity of the act of asserting that is also the unity of what is asserted. Judging (asserting) and being (what is asserted) here are the same. It is significant that Kant describes judgment in the metaphysical deduction as a ‘function of unity’ that subordinates lower representations under “a communal [gemeinschaftliche] one”—not a common one—at A69/B94. This suggests that judgment is an act of synthesis that brings representations into community with one another—an act of ordering representations under the formal concept of the unity of this very same act. That is, the materials of the judgment are subordinated under its form in the act of synthesis. So to judge is to hold representations together (synthesis) from a consciousness of their ‘mineness’ or unity in one consciousness, as expressed by the copula.25 Thus, the logical form of judgment does not bring to consciousness the lifeless (and dry) relations of material concept subordination that are presented on Porphyrian trees of species and genera, which would explain why Kant discusses these relations between mere concepts in the chapter of the Jäsche Logic on the Concept, not in the chapter on Judgment (JL 9:96ff.). Instead, judgment subordinates the parts of a judgment under the formal activity of judging itself—a bit like the way in which an animal subordinates the materials of its environment to its life activity when it eats; it incorporates the materials by ordering them under the activity of its own living body. Hegel may appear to capture this aspect of the Kantian conception of judgment when he says that spontaneity, for Kant, demands that the “positive reality of the world must be as it were crushed and pounded, in other words, idealized” (he compares this crushing to the process of chewing in an animal): “the ‘I’ is as it were the crucible and the fire which

Kant’s Logic of Judgment

77

consumes the loose plurality of sense and reduces it to unity” (Hegel, Enz. §42, p. 69). Judgment reflects on representations not as containing lifeless (already determined) concepts or marks within them, but as containing the activity of combining or determining them, which, for Hegel, involves crushing and pounding them. But although this correctly emphasizes the living or ‘active’ character of Kantian predicative unities, Kant himself does not appear to think of this activity as involving the negation or destruction of an opposing manifold of sensibility. Sensibility is not a force that the act of judging must crush and pound; it is an enabling condition of the activity of judging. Thus, the activity is not one of destroying or negating but one of maintaining or sustaining its own unity. In judging, we hold what we think together, in one logical unity of consciousness, under conditions of sensibility. 2.2.

Thinking and Veritative Being (Truth)

Alongside predicative being or unity (unum), ‘logical being’ also expresses truth (verum). When I judge that the ‘table is brown’, the copula ‘is’ signifies that it is true to think this. To judge is to be conscious of the validity [Gültigkeit] or truth of so judging. But what exactly am I conscious of when I am conscious of being in the sense of being true? In abstraction from the relation my thought bears to the object, my consciousness of its validity consists in consciousness that the elements of the thought should be combined together in one unity of consciousness, or that they agree with one another. Thus, the copula ‘is’ expresses what Kant calls “formal truth”, the “agreement of a cognition with itself” (JL 9:51). To think the thought to be true is to think the self-agreement in the act of thinking it. The bearer of truth in this sense is not an act-independent content of judgment but the act of judging itself. Thinking and (formal) veritative being—being in the sense of being true—are not distinct; consciousness of formal truth is self-consciousness.26 Even a mere thought, which does not, like judgment, make an assertion or stake a claim to objective truth, must be minimally in agreement with itself. That is, the elements within the thought must agree in the sense that they can be brought together in one unity of consciousness, one act of thinking. Only then is there an identical act of combining common to both the use of the subject and the use of predicate terms (an analytic unity of consciousness) that can be reflected through the ‘I think’. If two concepts entirely exclude or contradict one another (as, for instance, ‘bachelor’ and ‘married’ do), they cannot be brought together in one affirmation and so cannot be accompanied by the ‘I think’: a contradiction “entirely annihilates and cancels” the thought (KrV A150/B189).27 So contradictions may be called ‘false thoughts’ in the same way that we speak of ‘false friends’. They are not thoughts at all because there is no living bond (no synthesis) between the elements of the thought.28 The principle of contradiction is

78

Alexandra Newton

thus a criterion of formal truth or self-agreement and a requirement of self-consciousness (JL 9:51). The ground of formal truth in this sense must lie in the unity of apperception—not in the object given in sensibility—to distinguish contradictory from merely contrary concepts. Two attributes are contrary, or in a relation of ‘real opposition’, when they cancel one another out, as warm cancels out cold, under specific conditions—namely, in accordance with the laws of the nature (and causal powers) of the object in which they inhere.29 But concepts are contradictory when they entirely or absolutely exclude one another (entirely annihilating their possible union), without restriction to conditions. So logical opposition (contradiction) doesn’t presuppose a given nature or substance; indeed, the ‘I’ in which concepts are logically opposed must lack any given nature or limitation. Therefore, the I cannot be a substance or an object but must be an empty, infinite capacity—the capacity represented by ‘I think’. Concepts lack contradiction—and thus are ‘formally true’ in the weak sense of ‘internally consistent’—only if they are acts of combining concepts in one pure (non-empirical) unity of consciousness. The principle of contradiction is a sufficient positive criterion of analytic truth, since the opposite of an analytic judgment is a contradiction (S is not S). But it is only a negative criterion of truth in synthetic judgment (S is P), since you can think the opposite of a synthetic judgment (S is not P) without contradiction.30 In making a synthetic judgment, we do not merely think the predicate in combination with the subject term; we assert it of the subject term. And this means that we think of the predicate as holding of the subject term, not just for me, but necessarily for any rational subject, so as to exclude the opposite judgment from consciousness as false (albeit not contradictory).31 We saw above that the ground of this act of assertion—and thus the ground of the judgment’s internal necessary unity—does not lie in the subject concept alone since the predicate concept is not already contained in the subject concept. The relational approach, as we have seen, concludes from this that the ‘ground’ or ‘condition’ of a judgment lies outside of the judgment, in the object or intuition. But this contradicts Kant’s Copernican turn, according to which cognition is true, or agrees with its object, not because it conforms to the object, but because the object conforms to it (Bxvi). So it is not the object that determines our cognitions but cognition that determines the object. “[Cognition is] the determination of the object” (B166n.). More specifically, in judgment, the concept of an object in the subject position is determined by the predicate concept: “determination is a predicate, which goes beyond the concept of the subject and enlarges it” (A598/ B626). Since the object is not the source of this determination, it must be the concept of the object that determines itself and thus expands or augments itself in the act of synthetic judging. Although it can do this only under sensible conditions of an

Kant’s Logic of Judgment

79

object given from elsewhere, the object as it is given to me is not the source of its expansion.32 This allows us to see that even in synthetic judgment, the condition of assertion lies not in the object of intuition, but in the understanding (JL 9:51f.).33 To gain a proper understanding of this internal grounding of assertion in synthetic judgment, it is crucial that judgment not be understood as an act of assertion of a forceless (lifeless) content, which would place the recognition of truth (assertion) outside of that which is true (content), as on Stoic and Cartesian conceptions of judgment.34 That is, truth cannot be separable from consciousness of truth; for what is grounded is what is true, not merely an act of the mind through which I relate to, or become conscious of, what is true (assertion). There are not two acts here: one of grounding ‘S is P’ and another of grounding my act of judging that ‘S is P’. Rather, to ground ‘S is P’ is to reveal the necessity of thinking these concepts together in one unity of consciousness and to exclude its opposite from consciousness as false.35 The grounding is thus logical-psychological, not merely subjective-psychological. Thus, Kant maintains that the “principle of sufficient reason”, or principle demanding that synthetic judgment become conscious of its sufficient reason, belongs to the “universal criteria of formal truth” since it is by conforming to this principle (i.e. by providing a sufficient reason for judging as I do) that my judgment is revealed to be in necessary agreement with itself (JL 9:51). When my judgment is thus grounded, I can become conscious that ‘S is P’ is not something I merely happen to think, but that I (and any other rational thinker) must think S and P together, in one unity of consciousness.36 To ground a judgment is thus to show that the judgment is in agreement “with the universal laws of the understanding and of reason” (JL 9:51); it is to show that the representations internal to it belong together necessarily in one unity of consciousness or that they are (non-accidentally) my representations, and thus that they have their ground in me (the understanding). The principle of sufficient reason is therefore a principle of self-consciousness and a criterion of formal truth (ibid.).37

3.

The Subject (‘I’) of Judgment

Our elucidation of judgment shows that, contrary to the relational approach, the logical ‘I’ of judgment is not a subject that bears a relation to an already unified predication but is nothing other than the unity of predication itself (or the unity of the proposition, in complex judgments). The elements in the predication belong to one ‘I’ insofar as they belong together in one act of thinking them together, in one thought.38 As Kant emphasizes in the Paralogisms chapter, the ‘I think’ merely expresses the “form of every judgment of understanding whatever”, i.e. it represents the copula unifying the judgment ‘S is P’; it does not add a substance or

80

Alexandra Newton

underlying subject to the judgment itself (A348/B406). To say that in self-consciousness I am conscious of the unity of my representations in one I is not to say that I am conscious of an item or thing that performs an act of combining them. The capacity of understanding is not to be understood here as an attribute or power of an individual being. Rather, the understanding is itself the subject of judgment, or the “original synthetic unity of apperception” that constitutes the logical unity of the act of judging itself, and of all other acts of judging insofar as they belong together in a single system of cognition (B134n.). The logical ‘I think’ thus cannot be understood as a referring expression used to pick out an individual thinker: the “wholly empty representation ‘I’ [. . .] is not even a representation distinguishing a particular object, but rather a form of representation in general” (A346/B404). We saw above that according to the relational approach, the thought ‘S is P’ and the thought ‘I think S is P’ are distinct thoughts: the former is about the object thought under the concept S, while the latter is about me.39 Thoughts about me are contained in thoughts about S, or indeed in any conceptual representation, but they are not equivalent thoughts. By contrast, the synthetic approach identifies these thoughts, because the ‘I think’ is a representation of the unity of intellectual synthesis that is the same as the predicative unity of ‘S is P’. The articulation of this capacity in thinking ‘S is P’ is the same as the articulation of S’s being P, as expressed by the copula.40 So there is no sense in which logic is subjective-psychologistic since there is no distinction, at this level of reflection, between psychology and logic, or between thinking and being. To be conscious of the act of thinking in ‘I think S is P’ is the same as being conscious of the logical connective, the copula, internal to ‘S is P’. No new thought—i.e. no new, first-person thought—is introduced by thinking ‘I think’ since the I just is the logical unity of thought itself (not a substance or subject that bears a relation to unified thoughts). This is of course only true if the ‘I’ is the logical I, the capacity for intellectual synthesis shared by any rational thinker, not the empirically determined ‘I’ of inner sense (an individual I).41 The ‘I think’ of relevance to logic is a logical, pure representation of synthetic unity, not an empirical-psychological representation of a temporally extended, individual act of the mind. This reading is corroborated by several passages in Kant’s critical writings. For instance, Kant says that the “I think” is the “form of every judgment of the understanding in general”, because the forms of judgment are each representations of the unity of consciousness of various representations (A348/B406).42 They are each acts of maintaining a unity through consciousness of this unity. Strictly speaking, this consciousness (‘I think’) is not a concept, because it lacks all content, and is instead “a mere consciousness that accompanies every concept” (A346/B404). But one can think of it as a formal concept (without content) that accompanies any concept, and hence any use of concepts, in judgment.43 For it reflects

Kant’s Logic of Judgment

81

what is common to representations—namely, an identity or analytical unity of consciousness—and thus represents the form (“universality” or commonality) of any concept whatsoever (JL 9:91). The ‘mineness’ of representations thus isn’t a subjective, psychological feature of them, as on the relational conception, but is just the logical universality of any concept at all. There is no distinction, here, between the universality of the concept ‘red’ and the ‘mineness’ (togetherness in one analytical unity of consciousness) of all my representations of red, for ‘redness’, as a general concept, has the form of ‘mineness’. Hence, one can think of the logical I not as a particular concept, but as the form of any concept, i.e. the Concept in general.44 Philosophers often think of first-person thoughts as confining or limiting our thought to a certain subjective perspective (namely, the firstperson perspective), but that’s not how Kant viewed the I-thoughts of logic. Rather than confining our thought to a peculiar, inner realm, the first-person representation of predicative and veritative being (through the concept ‘I think’) is what enables us to think objectively valid thoughts beyond the bounds of any limited perceptual or individual perspective. That is, it enables us to think objectively about what is the case. That is why general logic, as a ‘logic of truth’, can understand itself to be “a selfcognition of the understanding and of reason”, i.e. why it can be a study of self-consciousness (JL 9:14). It is nevertheless tempting to construe this first-person approach to synthetic judgment in general logic as limiting our thought to the individual’s subjective perspective. If only we could jump outside the self-consciousness of judgment in logic—e.g. by supplementing it with an external relation to the object in transcendental logic—we would see what is objectively true and not just what seems to be true from our own perspective. Kant distinguishes “material truth”, or the “agreement of cognition with an object”, which is studied in transcendental logic, from formal truth or self-agreement studied by general logic (JL 9:51). Thus, one might think that transcendental logic introduces an external perspective on judgment to supplement the internal, first-person standpoint of general logic. But this supposed external perspective on objective truth from outside of the act of judging (or externalism, as it’s called in epistemology) is a fantasy: “I can compare the object with my cognition [. . .] only by cognizing it” (JL 9:50). The self-consciousness of judgment at issue in general logic does not obscure our vision of truth, or hide it from us, but is what first opens up the possibility of grasping objective truths from within selfconsciousness; there is no perspective on objective or material truth from outside of self-conscious judging or cognizing or outside of formal truth as such. To echo Wittgenstein, the logical I of judgment is not a limited perspective on the world from within the world but is the limit of my world.45 The I of logic (reason) is aligned not with a capacity for mere (subjective) representation from a limited perspective, but with the understanding as a

82

Alexandra Newton

capacity for knowledge [Erkenntnis], which is conscious of itself as valid for any thinker. So transcendental logic must be a development internal to general logic; it adduces relations to an object from within the selfconsciousness of logical functions of thinking and judging. The dependence of material on formal truth shows that Kant’s insight into the role of self-consciousness as the logical form of judgment will have important consequences for understanding the transition from general logic to transcendental logic—or from the logical copula (logical being) to the objective copula (objective being)—and from formal truth (self-agreement) to material truth (agreement with the object). Rather than thinking of this transition as turning away from the internal relation of representations to a subject (=I) and towards the external relation of judgment to an object (=x), our reading suggests that the transition must be internal to judgment; for if all truth is internal to the self-consciousness of judgment, material truth cannot be construed as a relation of correspondence between a self-conscious judgment and an object external to it, a relation that could be described ‘from sideways on’ (outside of the ‘I think’). Material truth must instead be understood as an internal correspondence between the form of a judgment and its sensible matter (intuitions or objects): it is a correspondence internal to the ‘I think’. For instance, a judgment ‘S is P’ is true if the assertion of P corresponds with the object thought under the subject concept S, and thus is internal to the synthesis of S and P; the material veridical sense of ‘to be’ is thus parasitic on the copulative, logical (and formal veridical) sense of ‘to be’. What it is ‘to be’ an object—or, more generally, what the contents of cognition are—thus depends on, and presupposes, the formal concept of ‘to be’ expressed by the copula in the logical form of judgment.46 And since we said that the ‘being’ expressed by the copula is the act of assertion (an act of thinking in combining predicate and subject concepts), this means that the contents of judgment cannot be determined outside of the act of assertion. Rather, all contents of judgment, all specifications of its object, are themselves acts of assertion, considered under the sensible conditions that provide the materials for these acts. The contents of judgment therefore are first made possible in and through the logical form of judging, i.e. in the self-consciousness of an act of assertion.

Conclusion I have argued that the ‘I think’ that attaches to all thought and judgment is a representation of the logical, synthetic unity internal to thinking (the ‘unity of consciousness of various representations’), not a merely psychological representation of the relation a subject bears to synthetic unities in thinking them. The ‘thinking’ it represents is the same as the ‘being’ of what is thought about. We can read Kant’s general logic already as concerned with being qua being, in the senses of being articulated above:

Kant’s Logic of Judgment

83

unity and truth. But its concern with being is merely logical, not ontological, because the concept of being is still undetermined. General logic thus is not formal or empty because it concerns itself with mere structures of ‘thinking’, in abstraction from the ‘being’ of what its acts are about. Instead, logic is formal because the concept of logical being (the logical copula) that it is concerned with is itself empty and has not yet been filled with the “transcendental content” that constitutes the relation to an object studied by transcendental logic (A79/B105). The distinction between general and transcendental logic thus does not require a distinction between the act of judging and the content judged, or between force and content; instead, both general logic and transcendental logic work with a conception of judging that overcomes that (contemporary Fregean) distinction. I have also argued that this self-reflection in thinking represents a judgment’s ‘formal truth’ or agreement with itself, and that all material truth or agreement with an object depends on a judgment’s formal truth. Contrary to a relational approach to judgment, which takes the primary bearers of truth to be mind-independent contents of judgment, I have argued that the primary truth-bearer is the act of judging, and that contents of judgment are internal to this act. The implications of this reading for transcendental logic and for any future development of metaphysics are profound and warrant a more extensive study.47

Notes 1. Kant also emphasizes the importance of comparison in his precritical writings: “to judge [is] to compare something as a characteristic mark with a thing” (FS 2:47). Kant goes on to state that the characteristic mark is a predicate, and that the ‘thing’ it is compared with is the subject concept in a predication. 2. Also: “if one thinks two representations as they are combined [verbunden] together and together constitute one cognition, this is a judgment” (VL 24:928). “To think [. . .] is to unite representations in a consciousness. [. . .] The unification of representations in a consciousness is judgment. Therefore, thinking is the same as judging or as relating representations to judgments in general” (P 4:304). 3. The dual relation approach can be found in many interpretations of Kant’s logic. Sometimes the two relations are correlated with intensional and extensional contexts of judgment, e.g. in Schulthess (1981: e.g. 182–3). 4. Kant uses various terms for the synthesis of representations in judgment. The most general is “Verbindung” (combination), but he also refers to judgment as a “Verknüpfung” (connection or nexus) when his focus is on the discursive relations (Verhältnisse, Relationen) among representations in judgment (B201n.). 5. By assertion here, I do not mean the linguistic act of stating what I judge but the mental activity of judging. Assertion in this sense has three logical modalities: problematic, assertoric, and apodeictic judging. See the discussion below about judgment as an ‘assertion’ under a ‘condition’. 6. Kant suggests that the object subsumed under concepts in a synthetic judgment must contain the marks reflected by these concepts in the Schematism

84

7.

8.

9.

10.

11. 12.

13.

14.

Alexandra Newton chapter of the first Critique: subordinations of objects under concepts require that the “former must be homogeneous with the latter, i.e. the concept must contain that which is represented in the object that is to be subsumed under it” (KrV A137/B176). In the Guyer/Wood translation, and in Longuenesse’s translation of this passage, a function is said to be “the unity of the action of ordering different representations under a common one”. But the word in German is “gemeinschaftlich” (communal), not “gemeinsam” (common); this will be important for the interpretation I develop in the next section. According to Longuenesse, “When Kant says that ‘the thought “I think” must be able to accompany all my representations’, he is not claiming that it must actually accompany all of them, or even less that any time I think ‘p’ I necessarily also think ‘I think p’ or ‘I think that p’. [. . .] When Kant says: ‘The “I think” must be able to accompany all my representations’, what he means is that in order to be mine (that is, to be recognized, thought as mine, what I see, hear, imagine, and so on), a representation must be taken up in such a process of combination and comparison, which—the argument will continue—is also the process in virtue of which the object of the representation is recognizable under a concept, or thought” (Longuenesse 2008: 15). This reading of the ‘I think’ as a psychological representation of thinking comes out clearly in Longuenesse’s more recent work: “The term in the argument place, ‘I’, is a concept that refers to the current thinker of the thought ‘I think this is a tree’, namely the current agent of the activity of reviewing the reasons for asserting ‘this is a tree’” (Longuenesse 2017: 27; see also 31, 81, 87). Alois Riehl adopts this position in his interpretation of Kant without finding it problematic: “The relation of representations in a judgment is the relation of their conceptual unity, namely their unity in the concept of the predicate” (Riehl 1908: 417). Judgment brings concepts to life: it is “the act by which it [the concept] becomes real” (Falsche Spitzfindigkeit §6). Reich reads the passage at JL 9:101 as saying that the logical form of judgment represents the unity of an act of combining representations through which judgment relates to an object (objective synthetic unity of consciousness but not analytic unity of consciousness) in both analytic and synthetic judgments. The function of giving unity to representations in a judgment is an action not of mere subsumption or analysis, but of synthesis. The act of subsuming or ordering representations under a common one (“functions of unity under our representations”) from A68–69/B93–94, which is an act of analysis, should not be identified with the “function of unity in judgment” from A79/B105, which is an act of synthesis (Reich 1992: 17–18). By calling this a ‘living bond’, I mean to echo Plato’s characterization of the connection of representations in judgment as ‘alive’ in the Sophist 248e6–9a2. But I do not mean to suggest that the activity of thinking is itself an activity of living. Moreover, by calling products of the imagination ‘lifeless’, I mean to suggest that they are like artefacts, produced by us, not that they are disconnected from logical life (judgment) entirely. Longuenesse concedes that the conception of judgment as assertion (rather than as analysis or mere concept subordination) constitutes “the form of judgment proper, that is, the copula or connective, thus the synthesis in judgment” (Longuenesse 1998: 146). But the question is how to understand this new dimension of judgment. Longuenesse has argued that the conception of judgment as a rule is to be understood in terms of a judgment’s role in syllogisms: a judgment ‘S is P’ functions as a rule insofar as the subject

Kant’s Logic of Judgment

15.

16. 17.

18. 19.

85

concept ‘S’ serves as a condition for the subsumption of further concepts under ‘P’ (Longuenesse 1998: 91ff.). For instance, ‘man’ serves as a condition or ground for the subsumption of ‘Socrates’ under ‘mortal’ in the syllogism that begins with the major premise ‘All men are mortal’ because Socrates is a man, and ‘man’ is subordinated under ‘mortal’. However, notice that this understanding of the syllogism reduces the ‘living bond’ of representations in a judgment to the ‘dead unity’ of representations under a concept, and thus does not add a genuinely new dimension to Kant’s conception of judgment. For now, the reason why ‘Socrates is mortal’ lies simply in the fact that Socrates is contained under the concept ‘man’, and ‘man’ is contained under the concept ‘mortal’. This is not sufficient to distinguish the unity of the judgment ‘Socrates is mortal’ from the unity of representations under the concept ‘mortal’. Whereas the assertion in the judgment ‘Socrates is mortal’—the living bond between these representations—excludes the opposite judgment ‘Socrates is not mortal’ from within the act of determining ‘Socrates’ to be ‘mortal’, the mere concept of ‘being mortal’ does not exclude, but is simply different from, the concept of ‘not mortal’ (or ‘immortal’). A proper understanding of assertion must be able to explain how assertion is not simply different from but also excludes the opposite judgment from within the act of combining or synthesis itself. Frege seems to think that the conception of judgment as a mere act of subsumption would be sufficient to characterize its unity: “in the sentence ‘Two is a prime’ we find a relation designated, that of subsumption . . . This creates the impression that the relation of subsumption is a third element supervenient upon the object and the concept. This isn’t the case . . . the object engages immediately with the concept without need of special cement. Object and concept are fundamentally made for one another, and in subsumption we have their fundamental union” (Frege 1979: 178). This overlooks the distinction Kant draws between the unity of representations under a concept and the unity of synthesis in a judgment. It is also present in Meier: “all true judgments have a ground and a sufficient ground of their truth §16. This ground is called the condition of the judgments (hypothesis, conditio iudicii)” (Meier AA 16:642). Some interpreters have tried to downplay the importance of synthesis (in the second sense) in Kant’s theory of judgment. Wayne Martin, for instance, has argued that Kant moves towards Frege by construing existential judgments (an F exists) as lacking a synthesis of two concepts; instead, a single concept is asserted or posited. I cannot find support for this view in Kant since he insists that although ‘existence’ is not a real predicate, it is still a logical predicate (A598/B626f.; Martin 2006). I have adopted the phrase ‘semantically self-contained’ from Irad Kimhi (2018). In his reading of Kantian synthetic a priori judgments, Frege maintains that we must “invoke a pure intuition as the ultimate ground of our knowledge of such judgments” (Frege 1960: 18; see also Coffa 1991: 17). At first glance, it may appear that this interpretation is in agreement with Heidegger, who argues that the object of any judgment is a “standard” for the judgment; in analytic judgments, the “object is a standard . . . but solely within its concept”, whereas in synthetic judgment, “the object itself participates as the foundation and grounds” (Heidegger 1967: 164). Indeed, this new sense of condition or ‘ground’ is what distinguishes Kant’s logic from previous ones: “Kant only added something to the [traditional] definition of judgment which had been omitted up till then”—namely, its reference to an “object” of intuition (ibid., 158–59). However, it is not clear that Heidegger thinks of the object in Kant as external to logical self-consciousness.

86

Alexandra Newton

20. In Allison’s early work, the distinction appears to map onto that of Gegenstand and Objekt in his discussion of the transcendental deduction (Allison 1983). It can also be found in a different form in Tolley’s distinction between the content of an intuition (intensional object) and object of intuition (extensional object) (Tolley 2011: 211f.). 21. According to Longuenesse, this is something Kant “says” (1998: 86). However, she provides no citation of where Kant says that the logical form of judgment can be developed in this manner, and I have been unable to find any passage that she might be referring to. Kant does use the term “x” to articulate the distinction between analytic and synthetic judgments in the Jäsche Logic: “an example of an analytic proposition is, To everything x, to which the concept of body (a + b) belongs, belongs also extension (b). An example of a synthetic proposition is, To everything x, to which the concept of body (a + b) belongs, belongs also attraction (c)” (§36, p. 117, 118). However, it is questionable whether ‘x’ is being used here as a variable for the object, or set of objects, contained under a and b. As I will suggest later on, we must instead understand x to be a concept that is further determined by a, b, and c. 22. Kant does say that “we can represent nothing as combined in the object without having previously combined it ourselves” (B130). But notice that he does not say that we can represent nothing as combined in a judgment without having previously combined it ourselves, for there is no representation of combination in judgment that is not a representation of the action of combining. 23. It is of course important to make room for reason’s “goal” of bringing cognitions “to the highest degree of agreement with itself through systematic unity” (A702/B730). However, I would maintain that this goal does not belong to the logical form of judgment, but instead presupposes logical form. 24. Inhaling and exhaling are not species of breathing but functions of breathing. Similarly, the different functions of judgment (under the four titles) are not species or kinds of cognition but functions of cognizing. For a contrary view, see Stang (2014: 231). 25. Kant does, in the metaphysical deduction, go on to talk about the material subordination of one (species) concept under a higher (genus) one. However, material concept subordination in judgment rests on formal concept subordination. Thus, when the representation ‘body’ is subordinated under ‘divisible’ in judgment, the predicate must be understood as containing the copula (form of the judgment) within it: ‘body’ is subordinated under ‘being divisible’. 26. In his logic lectures, Kant suggests that these three dimensions of thinking, or of logical being, line up with the four titles on the table of functions of judging: quality and quantity correspond to predicative unity or being, both as presupposed by (quality) and identical in (quantity) the elements of cognition (synthetic and analytic unity of apperception); relation corresponds to veritative being or truth; and modality corresponds to being as actuality. He discusses these dimensions under the four headings of the “logical perfection” of cognition as to quantity, quality, relation, and modality (JL 9:40ff.). 27. “[W]hatever conflicts with [the principle of contradiction] is obviously nothing (not even a thought)” (On a discovery, 8:195). “The logical mark of the impossibility of a concept consists, then, in this: that under the presupposition of this concept, two contradictory propositions would be false simultaneously; and since between these two no third proposition can be thought, through this concept nothing at all is thought” (P 4:341). Of course, two contradictory concepts (like ‘bachelor’ and ‘married’) can be brought together in one negative judgment: the bachelor is not married.

Kant’s Logic of Judgment

87

28. Hegel notes that the correlative sense of truth, implicit in ordinary phrases like ‘true friend’, is the more philosophical one, presupposed by what Kant calls the nominal definition of truth (correspondence with an object). 29. For an interpretation that shows all real opposition of realities to be dependent on the causal powers of the substances in which they inhere, see D. Warren (2001: 25ff.). 30. Put differently, the principle of contradiction does not rule out false synthetic judgments: “I can think whatever I like, as long as I do not contradict myself” (Bxxvi). 31. Kant says that all judgments of experience have “objective validity”, which “signifies nothing other than its necessary universal validity . . . [it] should be valid at all times for us and for everyone else” (P 4:298). “Die Vorstellung der Art, wie verschiedene Begriffe (als solche) zu einem Bewustseyn (überhaupt (nicht blos meinem)) gehören, ist das Urteil. Sie gehören zu einem Bewustseyn theils nach Gesetzen der Einbildungskraft, also subjectiv, oder des Verstandes, d.i. objektiv gültig vor jedes Wesen, das Verstand hat” (R 3051). 32. Here, I follow Engstrom, who argues that Kant’s Copernican turn requires that we think that “the determination” in theoretical cognition must “be self-determination on the part of the concept of the object, the concept’s selfenlargement through the universally valid incorporation of another concept within itself” (Engstrom 2017: 41). 33. This becomes clear in section 12, where Kant aligns truth with the “qualitative plurality of the marks that belong to a concept as a common ground” (B114). The ‘ground’ here refers to the ground of assertion, and it lies not in the object but in the ‘concept’ or qualitative unity of the judgment. For instance, in a categorical judgment, the concept of the unity of the judgment as a whole (the concept of synthesis articulated above) serves as a ground of the assertion of the predicate. Truth is inseparable from this act of assertion or from the “consequences” of this concept as a ground (ibid.). 34. On some readings of Descartes, he thought of judgment as an impetus of the will towards dead contents (see discussion in Rosenberg 2005: 92), and the Stoics construed judgment as assent to lifeless lekta. Kant can to some extent agree with Spinoza, who famously made fun of Descartes’s theory of judgment as presupposing contents of thoughts like “dumb pictures on a tablet” (E2p49s). 35. It may strike contemporary thinkers as strange that Kant thinks of truth in logic not as a predicate of forceless propositions or contents of thoughts, but as internal to an act of assertion as an exercise of the understanding. But what Kant emphasizes, and what contemporary thinkers have trouble comprehending, is that truth excludes falsity. This exclusion is mysterious if the bearers of truth are mind-independent contents of thoughts: what should it mean to say that one forceless content excludes another? If instead truth is internal to the act of asserting, and if assertion is internal to the predicative unity of the judgment, the exclusion can be made intelligible. 36. In the B-deduction, Kant says that the “relation of the representations to the original apperception and its necessary unity” is contained even in a judgment that “itself is empirical, hence contingent, e.g., ‘Bodies are heavy’” (B142). 37. Truth itself is inseparable from self-consciousness, as Kant shows in section 12 of the first Critique, immediately preceding the transcendental deduction. Kant thought that the error of what he called the “transcendental philosophy of the ancients” was to suppose that “unity, truth, and goodness” are nonempirical, “transcendental predicates of things”, and thus that they belong to the general features of the transcendental contents of representations, when in fact they are empty “logical requisites and criteria of all cognition of things

88

38.

39.

40.

41.

42.

43.

44. 45.

46.

47.

Alexandra Newton in general” (B113–14). It is not a property of things that they are unities of various features or marks, or that they have being (or truth). Rather, unity and truth are unintelligible outside of reflection on our cognition of things. As Kant goes on to explain, they belong to that of which we are conscious in the self-consciousness of cognizing (or judging). This contradicts Allison’s reading, according to which representations may belong to one ‘I’ without belonging to one thought, e.g. qua mere representations of sensibility. But Kant says that without the spontaneity of the understanding in combining representations, my I would be as diverse and manifold as my representations. This suggests that bringing representations together in thought is required for consciousness of them as mine. This view comes out not only in Longuenesse’s analysis of judgment in Kant and the Capacity to Judge but also more recently in her book on self-consciousness: as Longuenesse understands pure apperception in ‘I think’, the ‘I’ makes reference to a particular, existing thinker (Longuenesse 2017: 163f.). As Kant says, “being [. . .] in the logical use [. . .] is merely the copula of a judgment” (A598/B626). It is important to note that I am ascribing to Kant the identity of thought and being only in this formal sense of being (qua copula). As Kant argues in the Paralogisms, the logical I is not an underlying substance that bears a relation to her thoughts in thinking them; it is nothing more than the unity of thought itself (A348/B406ff.). Frege, in criticizing the idealists, says that thoughts about the moon differ in their sense or content from thoughts about myself: “To assume that in the sentence ‘The Moon is smaller than the Earth’ the idea of the Moon is in question, would be flatly to misunderstand the sense. If this is what the speaker wanted, he would use the phrase ‘my idea of the Moon’” (Frege 1997: 156). This overlooks the possibility that ‘mineness’ does not add any sense or content to the original thought because it is the ‘mineness’ of pure logical apperception, not of empirical apperception. It is important to note that this self-awareness in the act of thinking is not the privilege of the logician. In the Dohna-Wundlacken Logic, Kant distinguishes between the “logic of the healthy understanding, sensus communis” and “the logic of the speculative understanding; this gives rules in abstracto” (24:696). In everyday thinking (common understanding), we are aware of the forms of thought in concreto. What the logician does is merely to bring those same rules to consciousness in abstracto. Since I can become conscious of my representations as mine only through my consciousness of the capacity to use them in combination with others in judgment, it is because of the possibility of its use in judgment that every concept contains the ‘I think’. “The analytical unity of consciousness pertains to all common concepts as such” (B133n.). Wittgenstein (1961: 5.6, 5.62). This does, however, raise the question of whether Kant would acknowledge Wittgensteinian ‘limits’ of the ‘world’. Unlike Wittgenstein, Kant is not concerned with the limits of language, but rather the capacity for cognition. Support for this view can be found in Kant’s definition of an object in transcendental logic as “that in the concept of which the manifold of a given intuition is united” (B137). This suggests that to be an object, for Kant, is to be the object of a concept. There is no object outside of the concept, and hence no object outside of logic. This paper would not have been possible without several lengthy discussions with Stephen Engstrom about thought, judgment, and Kant versus Hegel. I also thank Irad Kimhi for getting me to think more about thinking and being.

Kant’s Logic of Judgment

89

References Allison, H. E. (1983), Kant’s Transcendental Idealism, 1st ed. New Haven: Yale University Press. Allison, H. E. (2004), Kant’s Transcendental Idealism, 2nd ed. New Haven: Yale University Press. Anderson, R. L. (2015), The Poverty of Conceptual Truth: Kant’s Analytic/ Synthetic Distinction and the Limits of Metaphysics. New York: Oxford University Press. Beck, L. S. (2002), Selected Essays on Kant. Rochester, NY: Rochester University Press. Coffa, A. (1991), The Semantic Tradition From Kant to Carnap. Cambridge: Cambridge University Press. Engstrom, S. (2017), “Knowledge and Its Object”, in J. O’Shea (ed.), Kant’s Critique of Pure Reason: A Critical Guide. Cambridge: Cambridge University Press. Frege, G. (1960), The Foundations of Arithmetic: A Logico-Mathematical Enquiry into the Concept of Number. New York: Harper & Brothers. Frege, G. (1979), Posthumous Writings, ed. H. Hermes et al. Oxford: Blackwell. Frege, G. (1997), The Frege Reader, ed. M. Beaney. Oxford: Blackwell. Hegel, G. W. F. (1975), Hegel’s Logic: Part One of the Encyclopaedia of the Philosophical Sciences (1830). (Enz.), trans. W. Wallace. Oxford: Clarendon Press. Heidegger, M. (1967), What Is a Thing? Indiana: Gateway. Kant, I. (2007), Correspondence (Corr.), ed. P. Guyer and A. W. Wood. Cambridge: Cambridge University Press. Kant, I. (1998), Critique of Pure Reason (KrV), trans. and ed. P. Guyer and A. W. Wood. Cambridge: Cambridge University Press. Kant, I. (2000), Jäsche Logic (JL). In: Lectures on Logic, trans. and ed. J. M. Young. Cambridge: Cambridge University Press, 1992. Kimhi, I. (2018), Thinking and Being. Cambridge, MA: Harvard University Press. Longuenesse, B. (1998), Kant and the Capacity to Judge. Princeton: Princeton University Press. Longuenesse, B. (2005). Kant on the Human Standpoint. Cambridge: Cambridge University Press. Longuenesse, B. (2006), “Kant on a priori Concepts. The Metaphysical Deduction of the Categories”, in P. Guyer (ed.), The Cambridge Companion to Kant and Modern Philosophy. Cambridge: Cambridge University Press. Longuenesse, B. (2008), “Kant’s ‘I think’ versus Descartes’ ‘I Am a Thing that Thinks’”, in D. Garber and B. Longuenesse (eds.), Kant and the Early Moderns. Princeton: Princeton University Press. Longuenesse, B. (2017), I, Me, Mine: Back to Kant, and Back Again. Oxford: Oxford University of Press. MacFarlane, J. (2000), What Does It Mean to Say that Logic Is Formal? PhD dissertation, University of Pittsburgh. Martin, W. (2006), Theories of Judgment: Psychology, Logic, Phenomenology. Cambridge: Cambridge University Press. Reich, K. (1992), The Completeness of Kant’s Table of Judgments , trans. J. Kneller and M. Losonsky. Stanford: Stanford University Press.

90

Alexandra Newton

Riehl, A. (1908), Der Philosophische Kritizismus, Vol. I, 2nd ed. Leipzig: Engelmann. Rosenberg, J. (2005), Accessing Kant: A Relaxed Introduction to the Critique of Pure Reason. Oxford: Oxford University Press. Schulthess, P. (1981), Relation und Funktion. Eine systematische und entwicklungsgeschichtliche Untersuchung zur theoretischen Philosophie Kants. Berlin: De Gruyter. Stang, N. (2014), “Kant, Bolzano and the Formality of Logic”, in S. Lapointe and C. Tolley (eds.), The New Anti-Kant. London: Palgrave Macmillan. Tolley, C. (2011), “Kant on the Content of Cognition”, European Journal of Philosophy 22(2): 200–28. Warren, D. (2001), Reality and Impenetrability in Kant’s Philosophy of Nature. New York: Routledge. Wittgenstein, L. (1961), Tractatus Logico-Philosophicus, trans. D. F. Pears and B. F. McGuinness. London: Kegan Paul.

4

Time and Modality in Hegel’s Account of Judgment Paul Redding

1.

Introduction

At the outset of the twentieth century, Bertrand Russell championed breakthroughs in formal logic that had gathered pace during the later part of the nineteenth, especially with the work of Gottlob Frege. Russell argued that this new, properly scientific, approach to logic had finally broken the grip of the Aristotelian term logic that had dominated philosophy for over two millennia, and that, he claimed, had adversely shaped metaphysical thinking (e.g. Russell 1914: 48). Very crudely, the Aristotelian tradition had assumed a subject-predicate conception of judgment structure, thus restricting predication to that of one-place predicates and preventing the logical representation of relations between objects. Such a limited conception of predication was, in particular, behind the idealist metaphysics that had recently dominated Cambridge philosophy based on the idea of an “absolute” substance that was the ultimate subject of all predication. This logic and metaphysics, he went on, had been most clearly on view in the philosophy of Hegel, who had strongly influenced the British idealists. But following Frege, one could start with the idea of a proposition as primary and, removing substantive terms, come up with a multiplaced conception of predication. While in fact there had been anticipations of such an approach in the earlier history of logic (Redding 2007: 3–7), Russell tended to present this as an entirely new development. Historically, Russell’s elimination of Hegel from serious subsequent consideration among analytic philosophers turned out to be exceedingly successful; however, the weaknesses of the position from which Russell diagnosed the problems of Hegel’s logical thought would soon emerge. Russell had construed the new logic to be employed in the radical renovation of metaphysics extensionally, but soon after the publication of the first volume of Principia Mathematica, the adequacy of the principle of extensionality for the understanding of logic itself would be brought into question by C. I. Lewis (Lewis 1912, 1914). Lewis, who had studied logic with Josiah Royce, the prominent American “Absolute Idealist”, argued that the extensionality behind Russell’s conception of material inference

92

Paul Redding

could not capture the full meaning of the concept “infer” as used in science. In order to capture the necessity implicit in this concept, Lewis went on to introduce the first modern, systematic study of modal logic qua the logic of necessarily and possibly true judgments (Lewis 1918). Moreover, Lewis would also underline the historical connection between the idealist tradition and the “intensional” conception of inference required for logic to be a truly scientific discipline (Lewis 1930). Lewis’s propositional modal logic was to remain a specialist interest through the first half of the twentieth century, with modal issues being largely dismissed by the predominantly positivist views of the time. However, in the second half of the century, following the development of quantified predicate modal logic by Kripke (1959, 1963) and others, modal issues would again come to be widely discussed. In particular, a difficulty arose for the principle of extensionality when applied to the new modal logic. “Quantification theory”, as it had developed in the context of non-modal logic, had simply conceived of quantification as ranging over domains of actual objects and their sets, but in the new modal context, quantification needed to range over non-actual possible objects—in fact, over complete “worlds” of such objects. The most dramatic defence of extensionality in the new context was that of David Lewis, who was happy to affirm an ontology of possible worlds as concrete, real alternatives to the actual world, from which they were spatio-temporally and causally disconnected (Lewis 1973, 1986). However, as Lewis noted, this claim was mostly met by colleagues with “incredulous stares” (Lewis 1973: 86). In recent decades, many wanting to take modal talk seriously but unwilling to accept Lewisian modal realism have been drawn back to the type of logico-metaphysical conceptions found in Aristotle that Russell had confidently proclaimed to be dead.1 Russell’s elimination of Hegel from the range of alternatives, however, seems to have been so complete that Hegel’s views are rarely, if ever, mentioned within these debates,2 and yet I suggest Hegel offers a largely unexamined alternative to the opposing “possibilist” directions of Leibniz and Lewis and the Aristotle-inspired versions of “actualism” as found, for example, in the work of Alvin Plantinga.3 In the following sections, I reconstruct Hegel’s logical treatment of tense and modality in his account of judgment in The Subjective Logic, Volume Two of his Science of Logic (Hegel 2010).

2.

Tensed and Untensed Judgments in Hegel’s Logic of Judgment

In his treatment of judgment, Hegel works through a series of specifications of judgment and then inference types, progress within which is generated in a way such as that found in the series of objective categories of Volume One, The Objective Logic. There, each successor in the series

Time and Modality in Hegel

93

of categories, starting with the purportedly most general, “Being”, had been presented as a solution to a problem concerning the logical coherence of its predecessor. The category “Being” had seemed the place to start because of its immediacy and generality—i.e. everything that “is” surely falls within its scope. However, this apparent advantage was also its shortcoming. Meant as comprehensive in this way, nothing is left with which this concept is able to contrast, and this undermines its very determinacy as a concept (Hegel 2010: 59, 21.69).4 However, a solution to this problem is found in the further concept “Becoming”, because something’s becoming is understood as its passage from nothing to being. “Becoming”, then, can be understood as intensionally containing the concepts of “Being” and “Nothing” as its “moments”, and with this third category uniting the first two, a new cycle in this process commences. The series of cycles working through The Objective Logic concludes with the category Actuality, and at this point, The Objective Logic transitions into Volume Two, The Subjective Logic, in which we find Hegel’s accounts of judgment and inference. Here, it should be remembered that what is under examination is, as in the earlier Objective Logic, a series of concepts assessed in terms of their internal coherence—in this case, of concepts specifying what exactly judgments and inferences in fact are. And as with the conceptual triads of objective logic, progress through the conceptions of judgment will typically contrast two initial, opposed judgment forms, the opposition between which will be resolved in a new third form that commences as the first term of the following cycle. Thus, running through the series of judgments and inferences will be found two alternating judgment forms that Hegel distinguishes by the different conception of predication involved in each: predication as inherence of the predicate in the subject on the one hand and predication as subsumption of the subject under the predicate on the other (Hegel 2010: 555, 12.58; Redding 2014). This distinction will be crucial to his conceptions of the role of both tense and modality in judgment, but for the moment, let us concentrate on issues of tense. The simple expressive structure in which an individual subject is joined to a universal predicate will provide us with a convenient starting place— Hegel calls this the initial “positive” subtype of the judgment of existence [das Urteil des Daseins] (Hegel 2010: 557–68, 12.59–70),5 this judgment being the first instantiation of predication as inherence. With an individual concrete object for its subject, the judgment of existence instantiates a typical de re judgment in which a single predicate is said of that subject, and as indicated above, we should expect some logical shortcoming to affect this concept of a judgment—some shortcoming that will be rectified by the succeeding form. Here, Hegel will say that the positive judgment is “not true”, and that it will have its “truth” in the succeeding “negative judgment”. A clue as to why such positive judgments are not “true”—that is, not true judgments—suggests itself in the array of examples that Hegel provides.

94

Paul Redding

This judgment form is instantiated in such actual judgments expressed as “The rose is red”, “The rose is fragrant”, and “Cicero was a great orator in Rome”, but one example in particular provides the clue for Hegel’s reasons for finding all such judgments as not proper or “true” judgments. This is the example “It is daytime now” (Hegel 2010: 562, 12.65)—a version of an example familiar from his Phenomenology of Spirit (Hegel 1977: §95), “Now is night”. As Hegel had pointed out in the Phenomenology, truth should not be lost by being written down, and if written down, “Now is night” will become false if read tomorrow at midday. But the overt indexicality of “It is daytime now” also applies to judgments like “The rose is fragrant” or “The rose is red”, as it is clear that Hegel intends by such judgments ones that refer to specific, individual entities—a particular rose, for instance, that exists (sein) here (da), and, of course, some particular rose that is fragrant now need not still be fragrant in one or two weeks’ time. It is clear, I suggest, that the positive judgment of existence is meant as indexical and thereby not a “true” judgment. Everyday English usage is happy to ascribe truth or falsity to such contextual assertions, and such was apparently the case for Aristotle and ancient philosophical thought generally.6 However, in contemporary philosophy, it is conventional to say that such assertions have incomplete contents that need completing by the addition of a temporal “index”, specifying, for example, the time at which the judgment is made. With this, the index can be taken as supplying another argument place, with respect to which the “incomplete” original judgment now functions as a predicate7 and the content of the judgment thus now contains, as with Russell’s analysis, some multiplaced predicate. That Hegel sides with the modern approach can be seen from the fact that he denies to such indexical statements the value of truth, “Wahrheit”, affording them only the lesser normative status of correctness or accuracy [“Richtigkeit”] (Hegel 2010: 562, 12.65). Such an interpretation might help explain the “untruth” of the positive judgment of existence, but why should Hegel claim that a positive judgment, S is P, as in “The rose is red”, has its “truth” in the negative judgment, S is not P? Clearly, were we to think of the negative judgment as somehow non-indexical and as containing a “complete” proposition, then Hegel’s claim would coincide with the widespread assumption today that the proper content of a tensed sentence such as “The rose is fragrant (now)” can be expressed by an indexed one that preserves the truth value of assertions made over time, as in “The rose is fragrant at time t1”. But what is the connection here to negation? Again, I suggest the clue is provided by a feature of the popular modern philosophical approach to judgments, in that now negation is typically treated as applying “externally” to complete propositions which, qua abstracta, are not themselves thought of as subject to temporal change. However, in order to avoid the assumption that we are here reading modern logic back into Hegel’s

Time and Modality in Hegel

95

approach in anachronistic fashion, earlier precedents for this approach should be noted. The Stoics, for example, had regarded the contents of judgments, “axiomata”, in this propositional manner and had treated negation externally (Bobzien 2003); and, closer to Hegel’s time, a similar approach can be found as part of Leibniz’s project of a universal characteristic. There is clear evidence in the Science of Logic and elsewhere that Hegel was familiar with these logical innovations, and indeed, the logic taught at the Tübinger Stift when Hegel was a student there was heavily influenced by such “progressive” elements of Leibniz and not limited to traditional Aristotelian syllogisms (Pozzo 2010). This suggests that Hegel had in mind a connection between negation and a properly “propositional” conception of the content of judgment. Indeed, clear supporting evidence for this is found in Hegel’s actual treatment of negation in these sections. Negation is clearly central to Hegel’s method in generating his series of judgment forms, just as it was in the genesis of the series of the objective categories; and here as well as elsewhere, Hegel will talk of two different stages of negating processes—a “first” negation creating the second element of a triad, followed by a “second” negation, the so-called “negation of negation”, which results in the third stage of the cycle (that is at the same time the first stage of the next cycle). Here, Hegel’s two stages of negation coincide with the two broad judgment types based on different conceptions of predication involved and align with the two different approaches to negation found in the logical tradition: term negation, as found classically in Aristotelian logic, and sentence negation, as found in Stoic logic as well as in the post-Fregean logic dominant today. In this process, the first negation of an initial positive judgment of existence will produce a distinctly limited form of negation in which some other one-placed predicate contrary to the first is said of the same subject. We might then think of the judgment “The rose is yellow” or, more generally, “The rose is non-red” as the first negation of “The rose is red”. Hegel describes this negation of the original positive judgment as applying only to the “determinateness [Bestimmtheit]” of the general predicate, leaving its “determination [Bestimmung]” unnegated. What Hegel means here is that while the particular colour of the rose is negated, the assumption that it is nevertheless coloured remains: “if the rose is not red, it is nonetheless assumed that it has a color, though another color” (Hegel 2010: 565, 12.68). From a logical point of view, we might describe negation here as of narrow scope, leaving the positive judgment that the rose is (some) non-red (colour). In this sense, this judgment of limited-scope negation is still a positive judgment, saying something positive of the rose (its nonredness), and will in turn be subject to a further negation. This initial form of the negative judgment manifests the type of predicative structure that W. E. Johnson, early in the twentieth century, treated as involving the division of a universal qua “determinable” into

96

Paul Redding

its “determinates”. Here, the determinable colour divides into particular colours—red, green, blue, and so on (Johnson 1921: ch. 11). Johnson argued that this determinable-determinate relation differed from that produced when a genus is divided into its species. In the former, there is no specific differentiating property separating determinates of a determinable in the way that, say, the feature rational divides the genus “animal” into rational and arational species.8 Hegel here describes the logical structure of this first “positive expression of the negative judgment” as “the singular is a particular” (Hegel 2010: 563, 12.65), because the predicate is not simply a universal but a particular instance of a universal (a particular determinate of the determinable colour) that is thereby determined by its contrasts with other particular determinates of colour. This first positive expression of the negative judgment can in turn be negated. Hegel writes: “this negation of the negative judgment appears, when one starts from its positive form, to be again a first negation. But this is not what it is”. The negation of the first (positive) negative judgment, he goes on, “negates the determinateness [Bestimmtheit] of the predicate of the positive judgment, its abstract universality, or, considered as content, the singular quality that it possesses of the subject” (Hegel 2010: 655, 12.69). Using Johnson’s terms, we could say that the first negation is restricted to the predicate qua particular determinate of its determinable, while the second negates the more general determinable that the first predicate instantiates.9 That the output of this operation is effectively the type of external negation found in Stoic propositional logic becomes apparent from the type of judgment into which the negative judgment transitions: the “infinite judgment”. This is a type of anomalous judgment that seems to presuppose some type of category mistake. Hegel’s examples include “The rose is not an elephant” and “The understanding is not a table” (Hegel 2010: 567, 12.70). There is a sense in which such judgments can be considered “true”—again, in the limited sense of accurate or correct, i.e. “richtig” (Hegel 2010: 567, 12.70)—but it is clear that one cannot draw the parallel inferences from such “truths” as one can from the earlier “positive” negative judgments. That the rose is not an elephant does not imply that it is some other animal. Hegel is clearly treating the infinite judgment as a degenerate form of judgment; it offers a type of truth that is “nonsensical and fatuous” (Hegel 2010: 567, 12.70), but its very possibility has been brought about by the way the “negation of negation” has negated the determinateness of the predicate. Now, a predicate is simply treated as either true or false of the subject simpliciter, that is, irrespective of the kind of entity the subject is. It is this that allows for the possibility here of such category mistakes.10 The infinite judgment signals the transition from the judgment of existence into the judgment of reflection—the first full example of a judgment in which predication is understood as subsumption of the subject under

Time and Modality in Hegel

97

the predicate. While in the judgment of existence the subject term had been implicitly treated as a particular instance of some genus (some rose), it is now, Hegel says, determined as a “singularity as such”—that is, as what is usually discussed as a “bare particular”, it is no longer thought of as an instance of a “kind” term—a result already observed in the “infinite judgment”. And also in contrast to the earlier judgment of existence, the predicate is now understood as an “essential universal” that “constitutes the basis against which the subject is to be measured and determined accordingly” (Hegel 2010: 569, 12.72). Thus, in the judgment “The rose is red”, the predicate “red” is no longer understood as a Johnsonian determinate of the determinable “colour” but as “the basis” against which the subject is “measured and determined”—an abstract universal that sorts the things of the world into two groups, those of which “red” can and cannot be predicated. Moreover, the question now arises about the status of the genus concept “rose” in the subject term. Given that the subject itself is conceived as something singular, the answer seems to be that the concept “rose” could only relate to this singular subject in the way that the predicate concept “red” does, that is, it must also “subsume” the subject. This elimination of “kind” terms as well as their treatment as the same as other predicates is a feature of post-Fregean logic. Thus, what emerges here is a conception of the logical structure of the judgment that is close to the one used by Russell in his criticism of the Aristotelian conception of judgment—the criticism that he also applied to Hegel himself. Following Frege, Russell had treated the traditional categorical judgment as a conditional, a judgment such as “All Greeks are mortal” being understood as saying, of all things, if that thing is a Greek, then it is mortal. Hegel, in his later discussion of Leibniz’s “mathematical” approach to inference (Hegel 2010: 602–9, 12.105–111), points to this type of analysis as found in Leibniz’s universal characteristic. Leibniz had treated subject terms as predicates that could be truly said of some array of indeterminate things (that Hegel describes as “thirds”) of which the predicate of the judgment can also be said.11 In short, Hegel’s “negation of the negation” has produced a radically different conception of logical structure to that de re conception found in judgments of inherence. It has produced a type of de dicto judgment structure, the contents of which will be conceived as “complete” propositions rather than the traditional subject-predicate structure. In short, Hegel had thus already employed the form of analysis against Aristotle that Russell was later to use against him. Hegel’s polarity of tensed judgments of inherence and untensed judgments of subsumption in fact exemplifies a general distinction now used by modal logicians between sentences in modal and non-modal languages, respectively (Blackburn et al. 2001). This is because, in recent times, “modal” has come to cover a range of judgments beyond the traditionally conceived judgments of possibility and necessity, the underlying idea being that modal judgments are ones whose truth values are in some

98

Paul Redding

way contextual. Traditional judgments of possibility and necessity can be regarded as contextual in the sense that they are quantified over possible worlds from the actual world, but the idea of contextual sensitivity is most easily seen in the case of the tensed judgments we are discussing. And just as tensed sentences can be translated into untensed ones with the addition of an extra argument place, the task of modal semantics has come to be understood as that of translating a sentence from a modal language into a non-modal one in a way that eliminates context dependence. One feature of Hegel’s logic must be stressed, however, that contrasts with the way such translation has come to be understood in the context of modern modal logic. Standardly, the non-modal language is treated as the metalanguage into which the modal language is translated, although this has been resisted by some who, treating modality as primitive, have regarded the non-modal language as an artificial extension of the modal (e.g. Prior and Fine 1977: 9–10). Hegel, however, clearly resisted the tendency to treat either modal or non-modal judgment forms as more basic. Rather, a cyclical process is put in place in which a content is progressively transformed, moving from modal to non-modal forms, and then back again to some new modal form, with the result that the context within which the new form is to be understood is treated as of greater scope than that of the first. In Hegel’s series, the classically “alethic” modal judgments of necessity and possibility will thus come to be understood as contextually sensitive, like tensed sentences, but in such a way that the context is more inclusive than the spatio-temporal context of simple tensed positive judgments of existence.

3.

From Tensed to Modal Judgments

When Hegel comes to describe the first instantiation of a major judgment form showing the “subsumptive” type of predication—the judgment of reflection—the content of such a judgment will, by virtue of its properly propositional nature acquired by the end of the earlier cycle, be conceived of as free of that particular contextual dependency that marked the starting judgment form. But Hegel insists that this new reflective judgment will itself be subject to its own type of “untruth”. Hegel had earlier described judgments of inherence as related to the function by which a judgment acquires content and judgments of subsumption as linked to form (Hegel 2010: 560, 12.62), presumably the logical form allowing formal logical relations to other judgment contents. We might expect, then, that isolated from any relation to a judgment of inherence, a judgment of reflection will be in danger of losing its semantic content. In fact, something like this has been anticipated by the “untruth” of the “infinite judgment” from which the reflective judgment issues. The infinite judgment surely seems cognitively vacuous. Just what does one learn about the understanding by being told it is not a table?12

Time and Modality in Hegel

99

Hegel presents the logical structure of the judgment of reflection as underlying that of particularly and universally quantified judgments, but the limitations of determinacy of such judgments require the judgment of reflection to transition into a type of higher-order de re judgment of inherence, as in Aristotelian categorical judgments about “secondary substances” or “kinds”. This forms the first subtype of the following “judgment of necessity” (Hegel 2010: 575, 12.77), but with this, the earlier problem of the indeterminacy of isolated de re judgments will return. This progression through the series of judgments, which proceeds by a cyclical alternation of de re and de dicto forms, will only stop with the next major type of judgment, the “judgment of the concept” (Hegel 2010: 581–7, 12.84–9), which is a type of explicitly evaluative version of the earlier categorical judgment. Here, it will be shown that the “truth” of that judgment is not another judgment but, in fact, the syllogism. This thesis that judgments only find their proper semantics as parts of larger inferences is now familiar in the work of philosophers of language, especially that of Robert Brandom (1994), who uses his own “inferentialist semantics” to interpret Hegel’s approach to judgment (2017). While broadly in agreement with Brandom’s inferentialist reading of Hegel, I want to explore the significance of the alternation of de re and de dicto forms in Hegel’s progression, which I suggest is crucial for Hegel’s handling of modality. In that it generates a series of judgment forms in which object-involving de re judgments of inherence keep recurring, the logical process has come to involve actual objects of increasing logical complexity. First were particular potentially perceivable objects (“this rose” or “Cicero”),13 but the cycle driven by negation has produced judgments “about” higher-order objects, first in the form of the “secondary substances” of the categorical judgments of necessity (the rose as such, deemed to be a plant) and later in the case of the judgment of the concept, singular “objects” of evaluative judgments—particular human actions or particular products of human making (Hegel 2010: 583–4, 12.85–6). Moreover, in virtue of their “subject-locating” forms, this series of recurring de re judgments has encoded places for actual judging subjects within the contents themselves,14 and the place of these subjects becomes explicit in the transition to the syllogism, as there normatively accessible human activities (acting, building, and so on) become the “objects” that are judged in a way that acts of judging itself come to be grasped as judgeable. In short, with the syllogisms that are shown to be concrete, thought is now construed as necessarily instantiated in the activities of embodied, related cognizers— normatively assessable agents who are the bearers of “thought”. These are the activities that Brandom, following Sellars, describes as language games of the “giving and asking for of reasons” (Brandom 1994; Sellars 1997)— those normative social practices in which judgments are expressed that make the process of judging itself a self-correcting one. Here, however, I want to remain focussed on the relevance of the alternating predicational

100

Paul Redding

forms within this cyclical process, as this ties into Hegel’s distinctive approach to the nature of human consciousness and self-consciousness. For Hegel, proper consciousness of an object requires self-consciousness, and we might appreciate this point in relation to time. Were one’s judgments always simply “in the moment”, one would be oblivious to the kinds of problems concerning the lack of stable truth values for tensed judgments. Thus, in order to be a proper judge, one needs to be able to reflect on and assess the content of one’s judgments made then, from the new perspective of now. It is only with this that one might then grasp the “untruth” of the judgment form exhibited by “Now is night” when uttered the next day, for example. In short, a judge must be able to grasp her- or himself as the same judge existing now as had existed then, and this presupposes a certain ability to “step beyond” the immediate context of one’s present experience, because one will need to grasp that “now” will in turn become “then”. Hegel’s solution to this problem involves the necessarily social nature of self-conscious beings. For a subject to be adequately aware of an object, she or he has to be aware of other subjects for whom that object can be an object. That is, a judge must recognize other judges, and this means being aware of them as worldly beings who, like other objects, are located in time and space but who are not thereby merely natural objects. And each judge must come to reflexively understand that they must be recognized by other judges in the same dual way. This is a theme that Hegel develops under the broad heading of the socially “recognitive” conditions for any individual self-consciousness (Redding 1996), and it is an idea presupposed by his conception of judging as part of a social practice that unfolds in time. Against this background, the syllogisms of the Science of Logic can be read as attempts to map the broad logical forms of such practices in which individual self-conscious beings are linked to each other and to aspects of the world of which they are conscious.15 Against the background of this broader approach to the “pragmatic” conditions of self-consciousness, we might now make some sense of the cycling of de re “modal” and de dicto “non-modal” forms in Hegel’s succession of judgments. I suggest that while de re judgments provide the appropriate contents for immediate judgments that are made from some located “first-person” position, de dicto ones provide appropriate contents of the sort that one ascribes to differently located other subjects. Time and place provide the first contexts within which this oscillation of judgment forms is appropriate, and the familiar “tensed” and “untensed” analyses found there represent the need for a judge to be self-conscious about the temporal contextuality of her or his own judging in order to compensate for it.16 Thus, to properly cognize the objects of her world, Alice needs to understand indexical facts about them from other “points of view”. Thus, she must be able to know the contents of Jim’s judgments and grasp them as expressing how the world appears from times and

Time and Modality in Hegel

101

places that she does not actually occupy. Following a suggestion by his Hegelian teacher John Findlay, Arthur Prior developed a “tense logic” as a type of modal logic concerning how this type of coordination takes place within natural languages. But with Hegel’s judgment of necessity, more classically aletheic modal notions are now required. And while they are to be treated in a way that is broadly similar to that of the lower level, in Hegel’s account, the new judgment form will operate with a conception of “context” that is now of greater scope. This is apparent when Hegel moves to the classical categorical judgment as a judgment essentially about the genus that had been only implicit in the earlier and simpler form of de re judgment of inherence, the judgment of existence. The classical categorical judgment is the first judgment of “necessity”. Understood in this form, “the rose” as subject term will no longer be taken as referring to some particular instance of the genus rose but to the genus itself: the rose as such. With such a judgment, as in, “the rose is a plant” (Hegel 2010: 576, 12.78), the necessity of those concepts belonging to the subject’s genus will be now contrasted with the contingency of ones attributed to it in the predicate of the simpler judgment of existence about some particular rose. The rose is essentially a plant, but this rose (which, qua rose, is essentially a plant) is only contingently red. It is negation that drives Hegel’s series of judgment forms forward, and if we consider how negation works at the level of the categorical judgment, it will be seen that the negating form invokes an idea of a “context”, to which truth is sensitive, that is wider than that found in the case of tensed judgments. A denial that some particular rose is fragrant can be understood as coming from some other spatio-temporal location in the actual world—let’s say from the point of view of an observer of the rose two weeks later. But the denial that a rose is a plant can’t be understood in this way. Here, it is tempting to say that rather than coming from some different location in the world, it comes from a “different world”—one in which roses are not plants.17 A little less dramatically, it might be said that the claim comes from the perspective of an interlocutor for whom the world is radically different to the way it is for the person who hears that claim—the perspective of a person with different theoretical beliefs about the world. In short, entertaining the negation of the claim “The rose is a plant” will involve entertaining ideas about the world being systematically different to the way that it is, or is thought to be. In the case of tensed judgments, grasping other judges as “differently located” simply meant differently located within the same spatio-temporal framework. That was why some shared temporal reference framework could resolve those problems of indexicality. With the judgment of necessity, however, rather than as contextualized within some particular spatiotemporal location within the world, the judgment is conceived as if “the world” itself is to be considered “indexically”, qua the actual world, as belonging to some broader totality that includes other “possible worlds”

102

Paul Redding

(Lewis 1973: 86). Of course, this is not to say that Hegel is committed to some ontology of possible worlds, but only that he envisages the communicative process as able to bring together interlocutors who can bring to the dialogue systematically different sets of beliefs. It is this feature of the relativity of claims to background beliefs that would allow the sentence modifier “necessarily” to be treated in Leibnizian fashion by Kripke (1959) as true in all possible worlds and “possibly” as true in some possible worlds.18 Obviously, possible-world semantics was not available to Hegel, but the treatment of modality in terms of the framework of possible worlds was, as it had been a part of Leibniz’s treatment of modality.19 Moreover, it is relatively easy to see how something like possible-world semantics naturally follows the type of de dicto conception of judgment that Leibniz introduced with his idea of a universal characteristic. The distinctly “modern” philosophical sense of the primacy of de dicto over de re judgments embraced by the Stoics had reappeared in the nominalist tradition and was again taken up by Leibniz, and this opened up the possibility for some explicitly propositional approach to alethic modal logic. Among the axioms of such a logic will be one linking the three modalities in the following way: if a proposition is necessarily true, then it must be (actually) true, but, of course, a proposition’s being actually true does not imply its being necessarily true. Similarly, if some proposition is judged to be (actually) true, then it must be possibly true, but—and as Hegel notes (1991: §143 Zusatz)—a proposition’s being possibly true does not imply its being (actually) true. These modal axioms thus portray the scope of possible truths—truths about “possibilities”—as wider than the scope of actual truths, which in turn is wider than the scope of necessary truths. From this, it is a short step to explaining the greater scope of possibility over actuality with the idea of “possible worlds”, of which the actual is one—an idea that had been given various religious forms of expression in the medieval period. I have suggested that the reflective point of view, compensating for the contextual specificity of the positive judgment of existence, be thought of as representing a view on the subject matter of the judgment from a differently located other within a shared world. At this higher level of judgment, however, one must be able to entertain hypotheses about how the world might be that differ systematically from one’s beliefs about how the world is. We might think of the earlier form of reflexive decontextualization as bound up with the subject’s self-conscious capacity to move (in time) through space, but with the categorical judgment, a new type of “movement” of the subject is necessarily the self-conscious capacity to move (in time) within logical space or, using a related metaphor, the Sellarsian “space of reasons”. That is, it presupposes the capacity to reason inferentially. Again, this link is made explicitly by Hegel. With the categorical judgment, as with the earlier de re judgment, limitations of its form will lead to a more elaborate version of the de dicto form earlier seen in the judgment

Time and Modality in Hegel

103

of reflection. Within the overarching framework of the judgment of necessity, this is the “hypothetical judgment”—what effectively is now treated as a conditional—that is the successor to the categorical judgment (Hegel 2010: 576–8, 12.79–80). In the hypothetical judgment, the original relation between subject and predicate now gets expressed as a logical relation between antecedent and consequent, considered as separate, logically connected contents—contents “determined in terms of the relations of reflection as a relation of ground and consequence, condition and conditioned, causality etc.” (Hegel 2010: 577, 12.79). With this, the de dicto structure of the earlier reflective judgment returns, but at a higher level, the hypothetical judgment instantiating a logical relation between two such judgments.

4.

Time and Modality in the Limitations of Aristotle’s Metaphysics

It is clear from his discussion of modality in the section “Actuality”, with which The Objective Logic concludes, that Hegel is aware of that style of thought that, based on the primacy of the de dicto conception of judgment as seen in Leibniz’s universal characteristic, gives rise to “possibility” being conceived as of wider scope to “actuality”: “the notion of possibility appears initially to be the richer and more comprehensive determination, and actuality, in contrast, as the poorer and more restricted one” (Hegel 1991: §143 Zusatz). However, he continues, “actuality is what is more comprehensive, because, being the concrete thought, it contains possibility within itself as an abstract moment”. This, we might describe as Hegel’s critique of the “one-sided” reduction of judgment structure to the de dicto model. At the same time, however, as we have seen, he does not affirm as fundamental or ultimate any de re conception of judgment structure, with its one-placed predication, or the corresponding de re conception of modality. Rather, the cyclical alternation of de re and de dicto conceptions of judgment allows for a similar alternation of de re and de dicto conceptions of modality. With these features in place, it might now be useful to contrast Hegel and Aristotle in terms of their respective approaches to time and modality. It is sometimes said (e.g. Hartmann 2013: 7) that in his emphasis on “real” as opposed to logical or “formal” modality, Hegel returns to Aristotle’s ontological or objective approach to possibility or “potential” (dunamis) the material from which individual substances are actualized (in the process of energeia). However, what we have said about the role of logical conceptions of modality in Hegel’s cycles of decontextualization casts doubts upon the accuracy of this view. As found in one of his own examples, Aristotle conceives such potentiality as akin to a capacity, such as a person’s capacity to build, which a person may be said to possess when not actualizing it (Aristotle 1984: 1046b29–33).

104

Paul Redding

This dunamis-energeia pair is closely bound up with the form-matter distinction of Aristotle’s metaphysics: wood and stone are potentially a house, the form of the house becoming actualized in the process of its building. Were Hegel to closely align his position on possibility with this general picture, this would surely confirm Russell’s original diagnosis of his commitment to a basically Aristotelian metaphysics grounded in term logic. However, while there is certainly a sense in which Aristotle’s this-worldly metaphysics plays an important role as a model for Hegel, too much emphasis on this relation is in danger of ignoring Hegel’s conceptions of the limitations of this classical picture as well as his clear intention to incorporate the modern “subjective” standpoint into it—a standpoint that relies on the more formalizable considerations of logic seen in the conception of judgments with de dicto content. I have argued for the important role played by mediating de dicto logical relations in Hegel’s treatment of modality in The Subjective Logic, along with that of the relation of modality to issues of temporality, but we also see this anticipated in his treatment of the category “actuality”, with which The Objective Logic terminates. Thus, while broadly in line with Aristotle’s “actualist” approach, and aiming to integrate “possibility” into the fabric of the actual, Hegel’s way forward here does not look at all Aristotelian. Hegel starts with the modern logically based notion of “formal” possibility and its shortcomings. The formal approach conceives of the realm of the possible as an “indeterminate receptacle of everything in general”—a receptacle of possibilities that simply contains “everything . . . that does not contradict itself” (Hegel 2010: 479, 11.382)—an expression of the plurality of possible worlds approach. But internal problems with this conception lead to the notion of real possibility, and yet in contrast to Aristotle’s conception of an individual thing’s actualizable potential or capacity, Hegel’s account of real possibility starts from the idea of the multiple conditions of some entity or state of affairs’ existence, conditions that would be grasped via the complex hypothetical judgment, as discussed in relation to the categorical judgment in The Subjective Logic: “the real possibility of a matter [Die reale Möglichkeit einer Sache] is . . . the immediately concretely existing multiplicity of circumstances, which are related to it” (Hegel 2010: 482, 11.386 translation modified)—“the totality of conditions, a dispersed actuality which is not reflected into itself, but is determined to be the in-itself of an other and intended in this determination to return to itself” (Hegel 2010: 483, 11.387). All this depends on the logical capacity for hypothetical reasoning, a dependence made explicit in the transition to the syllogism. Thus, in conceiving of real possibility, we approach the possibility of something not yet actual as at first the “dispersed actuality which is not [yet] reflected into it”—the set of conditions that is necessary for the new thing or state of affairs to come into existence. The starting point here is more the modern conception of causality understood as the set

Time and Modality in Hegel

105

of necessary and sufficient conditions for something’s occurrence than anything found in Aristotle, but Hegel goes on to point out that in order to grasp these conditions modally as constituting the possibility of something actual, they must be understood in relation to the actuality that results. That is, the starting idea that the real possibility of some A just is the totality of its necessary conditions omits the thought that this totality itself is entirely indeterminate if conceived of independently of the A to which it gives rise.20 The conditions need to be conceptually gathered, as it were, into the subject place of a judgment that they can be conceived as “the possibility of A”, and such activity, for Hegel, as for Kant—but not for Aristotle—presupposes the activity of some unified subject doing such “gathering”.21 But unlike the picture found in Kant’s universal “I think”, such unified subjects in Hegel are plural and understood as not only already spatio-temporally located in the actual world but also selfconsciously cognitively “located” in relation to some possible conception of that world—the world as that subject conceives it. Within the resources of Aristotle’s term logic, it is difficult to see how the idea of the “external conditions” for the actualization of a potential might play any real role within his metaphysics, and yet such conditions were widely discussed in ancient philosophy, and examples used by Aristotle himself show him as clearly aware of the need to take appropriate conditions into account. A sperm, for example, might be considered to be a potential man, but it is only when it can “further undergo a change in a foreign medium”—presumably, undergoing changes that depend on its being implanted in a uterus—that it is genuinely a potential man (Metaphysics 1049a14–18). But from the point of view of both his metaphysics and his logic, it is unclear how to account for such conditions. Of the four aitia, it might be thought that “to kinoun”, the so-called “efficient cause”, might here be invoked, but Aristotle’s concept here is not the modern one. The being-in-a-womb is not an “efficient cause” of the sperm’s becoming a man because Aristotle regards the relevant kinoun here as the actual man from whom the sperm issued. In contrast, it would seem that the Stoics, who had a more propositional form of logic with which to work, were better equipped to conceive of the role of conditions here. For example, unlike the picture presented in Aristotle’s sperm example, the Stoics considered that although external conditions might prevent something from happening, that event could still be considered as a possibility. Wood at the bottom of the ocean still had the potential to burn even if, in fact, it were to remain there forever and never actually burn (Bobzien 1999: 110). As seen in the issue of the modal status of the sperm, Aristotle often seems to collapse modal issues into temporal ones. Thus, we might think of the problem of the sperm as an example of problems besetting a more general principle found in Aristotle—the so-called “Principle of Plenitude” (Hintikka 1973: ch 5)—which says that any potential A, given

106

Paul Redding

enough time, will become A. Another expression of this principle is that what is said of the past must thereby be necessarily true, possibility being restricted to the as yet unrealized future states of the world. Such notions, however, seem to go against everyday modal intuitions, which support counterfactual conditionals. Perhaps more importantly, they would rule out the conception of possibility that is required for the modern notion of subjective freedom, with its accompanying sense of a person’s having being able to act in ways contrary to the way in which they actually acted. Hegel was critical of the limitations of Aristotle’s metaphysics, especially in relation to modern notions of subjectivity and freedom, and aimed at integrating the modern principle of subjectivity into a metaphysical view of the world that nevertheless, like Aristotle’s, offered a “this-worldly” alternative to Platonism. This, I suggest, is on display in his treatment of temporal and modal considerations in relation to judgment. But Hegel’s way forward was not simply to replace Aristotle’s limited logic with the new, more powerful version—i.e. to translate broadly “modal” judgments into their non-modal equivalents. Thus, he perceived the need for the two types of logic—modal and non-modal—to be somehow integrated, and the pursuit of this took him into areas of logic and metaphysics that have once again become central to philosophy.22

Notes 1. See, for example, the contributors to Tahko (2012). Particularly influential in the return to a type of Aristotelian metaphysics in this connection have been Plantinga (1978) and Fine (1994). 2. For a rule-proving exception, see Borghini (2016: 46). 3. Moreover, I suggest the influence of Hegel can be discerned even in the actual revival of modal logic in the twenthieth century and not just as part of the background to C. I. Lewis’s early attempts to treat modal notions within logic. In particular, in the mid-1950s, the New Zealand philosopher and logician Arthur Prior explored connections between modal logic and the logic of tensed judgments—“tense logic” (Prior 1957)—in a way that was important for the subsequent development of quantified modal predicate logic (Copeland 2002). Prior had been explicit about the influence, in this regard, of a former teacher, John N. Findlay, whom, on the basis of a paper published on time (Findlay 1941), Prior had proclaimed the “founding father” of tense logic (Prior 1967: 1). Findlay had himself been a philosopher with diverse interests, but his first and most lasting influence in philosophy had been Hegel. I explore these connections in more detail in Redding (2017). 4. Here, page numbers to the English translation are followed by volume and page numbers to Hegel (1968–). 5. We should not think of this as an existentially quantified judgment in the modern sense. That the judgment of existence picks out some specific observable entity in the world that is present to the judge is suggested by the morphology of “Dasein”, the being (Sein) of something there (da). In line with this interpretation, Hegel sometimes uses demonstrative phrases in subject position for such judgment types (e.g. Hegel 1991: §172 add.). 6. According to Crivelli, this view “was widely shared in antiquity—in fact, it remained unchallenged” Crivelli (2004: 184).

Time and Modality in Hegel

107

7. Effectively, the “index” is taken as a proper name of a temporal point of which the incomplete sentence can be predicated. Said at time t1, the purported sentence “It is raining” is taken as predicating “It rains” as true of that point in time. 8. One cannot invoke some specific differentiating property possessed by “red”—say, but lacking in “blue”, for example—other than its being red. This is the feature we have observed in Hegel’s very first instantiation of negation between the categories Being and Nothing. 9. Plantinga (1978: 173–4) utilizes this de re/de dicto ambiguity of negated sentences in a similar way. 10. It is here that the reversal of the original subject-predicate form becomes significant. “The rose is non-red” can be interpreted as both having a singular subject and a particular predicate and also as having a particular subject and a singular predicate. This latter reading brings out both the individual shade of red instatiated by the rose (it is this red) and the fact that the subject of the judgment is a particular instance of the genus rose. 11. Treating subjects as predicates of some unnamed “third” thus forms the model for the way of adding argument places to the judgment that we have seen in the case of tensed judgments. This is basically the way Leibniz conceives of the process of “analysis”, whereby clear but otherwise “confused” judgments are, in a series of steps, to be converted into clear and distinct ones. 12. A similar idea can be seen in some contemporary approaches to semantics in which the cognitive content of an assertion is understood in terms of the possibilities that are thereby eliminated by that assertion. On this analysis, the assertion “The rose is red” allows the hearer to eliminate a range of possibilities concerning the rose—that it is yellow, blue, pink, and so on. But as we have seen, this analysis begs a Johnsonian “inherence” account of predication. 13. Frege had criticized Russell’s early conception of the proposition qua content of judgment, in that such propositions seemed to contain actual objects, such as Mont Blanc. In this sense, Hegel’s conception of judgment content is similar to Russell’s: content can be concrete and hence object-containing. 14. In the way that a judgment of Dasein locates the judge at the “da” of the object. 15. I explore this idea in relation to Stalnaker’s conception of shared intentional content in Redding (2018). 16. This point is made by Sebastian Rödl (2012). Rödl’s position, however, is closer to Kant than Hegel. 17. One might imagine roses as colonies of unicellular organisms, after the model of slime moulds, for example. 18. In modern modal logic, these worlds are limited to those “accessible” from the actual world, a feature not found in Leibniz. 19. Brandon Look points out that “while Leibniz does not explicitly define necessity and contingency in terms of possible worlds . . . his conception of possible worlds underlies his account” (Look 2016: 196). 20. Similarly, for Aristotle, the matter of a substance cannot be independently conceived as that which has the power to become the substance qua formed matter, as the matter is identified in terms of what it is that becomes substance and so cannot be identified independently of this. On some of the problems inherent in Aristotle’s dunamis-energeia relation, see Bechler (1995). 21. In this sense, possibilities are mind-dependent, making Hegel a “possibilityidealist”, as described and defended by Nicholas Rescher (1979). As he is committed to the existence of possibilities in the actual world, Hegel must thereby be committed to the existence of mind in the actual world, but this is an ontologically trivial thesis if one takes “actual” to function as an indexical— the actual world being the possibility that we, minded creatures, are in. 22. Work for this paper was supported by Australian Research Council Discovery grant DP130102346.

108

Paul Redding

References Aristotle (1984), Metaphysics, in J. Barnes (ed.), The Complete Works of Aristotle: The Revised Oxford Translation. Princeton: Princeton University Press, Vol. II. Bechler, Z. (1995), Aristotle’s Theory of Actuality. Albany: State University of New York Press. Blackburn, P., M. di Rijke, and Y. Venema (2001), Modal Logic. Cambridge: Cambridge University Press. Bobzien, S. (1999), Determinism and Freedom in Stoic Philosophy. Oxford: Oxford University Press. Bobzien, S. (2003), “Logic”, in B. Inwood (ed.), The Cambridge Companion to the Stoics. Cambridge: Cambridge University Press. Borghini, A. (2016), A Critical Introduction to the Metaphysics of Modality. London: Bloomsbury. Brandom, R. B. (1994), Making It Explicit: Reasoning, Representing, and Discursive Commitment. Cambridge, MA: Harvard University Press. Brandom, R. B. (forthcoming 2019), A Spirit of Trust: A Reading of Hegel’s Phenomenology. Cambridge, MA: Harvard University Press. Copeland, B. J. (2002), “The Genesis of Possible Worlds Semantics”, Journal of Philosophical Logic 3: 99–137. Crivelli, P. (2004), Aristotle on Truth. Cambridge: Cambridge University Press. Findlay, J. N. (1941), “Time: A Treatment of Some Puzzles”, Australasian Journal of Philosophy, repr. in A. Flew (ed.), Logic and Language. Oxford: Blackwell, 1951. Fine, K. (1994), “Essence and Modality”, in J. E. Tomberlin (ed.), Philosophical Perspectives 8: Logic and Language. Atascadero: Ridgeview. Hartmann, N. (2013), Possibility and Actuality, trans. A. Scott and S. Adair, with an introduction by R. Poli. Berlin: de Gruyter. Hegel, G. W. F. (1968–), Gesammelte Werke. Hamburg: Meiner. Hegel, G. W. F. (1977), Phenomenology of Spirit, trans. A. V. Miller with analysis of the text and foreword by J. N. Findlay. Oxford: Clarendon. Hegel, G. W. F. (1991), The Encyclopedia Logic: Part 1 of the Encyclopaedia of Philosophical Sciences with the Zusätze, trans. with introduction and notes by T. F. Gereats, W. A. Suchting, and H. S. Harris. Indianapolis: Hackett. Hegel, G. W. F. (2010), The Science of Logic, trans. and ed. G. di Giovanni. Cambridge: Cambridge University Press. Hintikka, J. (1973), Time and Necessity: Studies in Aristotle’s Theory of Modality. Oxford: Clarendon. Johnson, W. E. (1921), Logic: Part 1. Cambridge: Cambridge University Press. Kripke, S. (1959), “A Completeness Theory in Modal Logic”, Journal of Symbolic Logic 24: 1–14. Kripke, S. (1963), “Semantical Considerations on Modal Logic”, Acta Philosophica Fennica 16: 83–94. Lewis, C. I. (1912), “Implication and the Algebra of Logic”, Mind 21: 522–31. Lewis, C. I. (1914), “Review of A. N. Whitehead and Bertrand Russell, Principia Mathematica”, Journal of Philosophy 11: 497–502. Lewis, C. I. (1918), A Survey of Symbolic Logic. Berkeley: University of California Press.

Time and Modality in Hegel

109

Lewis, C. I. (1930), “Logic and Pragmatism”, in G. P. Adams and W. P. Montague (eds.), Contemporary American Philosophy, Vol. 2. New York: Macmillan. Lewis, D. K. (1973), Counterfactuals. Oxford: Blackwell. Lewis, D. K. (1986), On the Plurality of Worlds. Oxford: Blackwell. Look, B. C. (2016), “Leibniz’s Theory of Necessity”, in M. Cresswell, E. Mares and A. Rini (eds.), Logical Modalities From Aristotle to Carnap: The Story of Necessity. Cambridge: Cambridge University Press. Plantinga, A. (1978), The Nature of Necessity. Oxford: Clarendon Press. Pozzo, R. (2010), “Gottfried Ploucquet”, in H. F. Klemme and M. Kuehn (eds.), The Dictionary of Eighteenth-Century German Philosophy. New York: Continuum, Vol. 2. Prior, A. N. (1957), Time and Modality. Oxford: Clarendon Press. Prior, A. N. (1967), “Precursors of Tense-Logic”, in Past, Present and Future. Oxford: Clarendon. Prior, A. N. and K. Fine (1977), Worlds, Times and Selves. Amherst: University of Massachusetss Press. Redding, P. (1996), Hegel’s Hermeneutics. Ithaca: Cornell University Press. Redding, P. (2007), Analytic Philosophy and the Return of Hegelian Thought. Cambridge: Cambridge University Press. Redding, P. (2014), “The Role of Logic ‘Commonly So Called’ in Hegel’s Science of Logic”, British Journal for the History of Philosophy 22: 281–301. Redding, P. (2017), “Findlay’s Hegel”, Critical Horizons 18: 359–77. Redding, P. (2018), “Hegel, Modal Logic, and the Social Nature of Mind”, Inquiry. doi:10.1080/0020174X.2018.1484007. Rescher, N. (1979), “The Ontology of the Possible”, in M. J. Loux (ed.), The Possible and the Actual: Readings in the Metaphysics of Modality. Ithaca: Cornell University Press. Rödl, S. (2012), Categories of the Temporal: An Inquiry into the Forms of the Finite Intellect, trans. S. Salewski. Cambridge, MA: Harvard University Press. Russell, B. (1914), Our Knowledge of the External World. London: Allen and Unwin. Sellars, W. (1997), Empiricism and the Philosophy of Mind, with an introduction by R. Rorty and a study guide by R. Brandom. Cambridge, MA: Harvard University Press. Tahko, T., ed. (2012), Contemporary Aristotelian Metaphysics. Cambridge: Cambridge University Press.

5

Bolzano’s Theory of Judgment Mark Siebel

1.

Introduction

The Bohemian philosopher, mathematician, and theologian Bernard Bolzano was born in 1781, the year Kant’s Critique of Pure Reason was published; and he died in 1848, the year Frege was born. These biographical data blend in well with Bolzano’s philosophy. In opposing Kant, he developed ideas showing a striking resemblance to Frege’s. However, he worked not only in the fields Frege was active in but was also concerned with religious philosophy, ethics, and political philosophy. Due to the fact that his thoughts in the latter domains were quite liberal, Bolzano, after holding the chair of religious studies at Charles University in Prague for 15 years, was summarily dismissed in 1820. Since he was also barred from public teaching and many of his books were indexed, his influence on nineteenth-century philosophy was limited. Fortunately, scholars such as Brentano and Husserl made sure that Bolzano’s ideas started to attract the attention they deserve at the beginning of the twentieth century. Among these ideas is Bolzano’s theory of judgment. Section 2 presents one of the many places where Bolzano anticipates Frege’s anti-psychologistic notion of a third realm, which complements the inner realm of mental appearances and the outer realm of perceivable objects. In particular, Bolzano strictly distinguishes between judgments as mental acts and the contents of such acts. In section 3, it is shown how he tries to draw the line between judgments and acts of merely entertaining a thought. Section 4 focuses on the formation of judgments. Of prime importance is the distinction between mediated and unmediated judgments because it is intimately connected with epistemic issues. Section 5 deals with intrinsic qualities of judgments, such as vividness, degree of confidence, clarity vs obscurity, and distinctness vs confusedness.1 The notion of judgment occupies centre stage in Bolzano’s analyses of epistemic concepts. It is not only crucial to his explication of belief (Meinung) as a disposition to judge but also to his explications of cognition (Erkenntnis) as true judgment and conviction (Überzeugung) and knowledge (Wissen) as attitudes towards judgments. In the interest of

Bolzano’s Theory of Judgment

111

brevity, I will not go into this conceptual enterprise. Instead, it will be pointed out that Bolzano’s theory of judgment includes ingredients one would hardly expect when being told that he anticipated Frege’s antipsychologistic views.

2.

Judgments as Mental Acts and Contents

Bolzano applies the term ‘judgment’ (Urteil) in three senses. We find judgments as mental acts, judgments as the contents of such acts, and judgments as the products of them. A judgment in the third sense is “something that is brought about through judging”, i.e. the act’s “effect” (WL I, § 20, 82). However, this sense does not play an essential role in Bolzano’s theory. Much more important is his precise distinction between judgments in the first and the second sense.2 Bolzano’s official introduction of what he means by ‘judgment’, in § 34 of the Theory of Science (Wissenschaftslehre), is geared towards mental acts. There, he alternatively uses the nominalization “judging” (Urteilen), and he points out that judgments in this sense are “activities of our mind” (WL I, § 34, 155).3 Elsewhere, he describes them as “changes occurring in our souls”, which have “a beginning and an end” and are thus of “finite duration” (WL III, § 297, 118; cf. EG, 47). A judgment in this sense is an individual event; it is spatio-temporally located and part of the causal order.4 In Bolzano’s words, it has “actuality” (Wirklichkeit) insofar as it is causally efficacious, e.g. by eliciting further judgments.5 Furthermore, a judgment is an “adherence” (Adhärenz), i.e. a dependent entity, because it “does not have its existence by itself but only in the [. . .] mind of a particular being, which for this very reason is called the judging being” (WL I, § 34, 155).6 This entails that acts of judging have to be individuated not only by the time in which they take place but also by their bearers—namely, the subjects who perform them. Although Ann’s judgment that three is a prime number has something important in common with Ben’s judgment that three is a prime number—so that Bolzano is willing to speak about “equal judgments” (WL III, § 292, 111f.)—they are nonetheless numerically distinct acts. The term ‘judgment’ can also be used to refer to what Ann and Ben’s acts have in common—namely, their content. In this second sense, Ann and Ben’s judgments are thus strictly identical. In § 155 of the Theory of Science, Bolzano makes use of the term “judgments in themselves” (Urteile an sich) to denote the contents of acts of judging (WL II, § 155, 128). This expression is a descendant of his coinage “sentences in themselves” (Sätze an sich). A judgment in itself is a sentence in itself that is the content of an act of judging (cf. WL I, § 22, 86). Sentences in themselves, or “objective sentences” (objective Sätze; cf. EG, 47), are not linguistic signs, but rather the “sense” (Sinn) of certain word compounds.7 Strictly speaking, even this characterization is too

112

Mark Siebel

narrow because a sentence in itself is existentially independent of linguistic signs insofar as it might never be put into the words of some language.8 The German word ‘Satz’ is used in this sense within expressions such as ‘Satz des Pythagoras’ because it might have happened that Pythagoras’ theorem was never formulated by anyone (cf. WL I, § 49, 219f.). I will frequently use the shorter term ‘propositions’, and I will refer to them by putting sentences in square brackets—i.e. [Three is a prime number] is the proposition expressed by ‘Three is a prime number’. In contrast to a judgmental act, the sentence in itself that is its content lacks a position in space and time, does not stand in causal relationships, and is independent of the existence of thinking beings or languages.9 That is, while one may ask when and where an act of judging took place, what caused it, and which effects it had, as well as whose act it was, none of these questions makes sense when it comes to the content of the act. Just like Frege’s (1918) “thoughts”, sentences in themselves are conceived by Bolzano as the primary bearers of the unrelativized truth values true and false.10 This means, first, that other things, such as sentences or judgments, have their truth values in virtue of the truth values of the propositions that are their contents. Second, what is expressed by, say, ‘I am hungry’ in each case does not possess relativized truth values like true/false with respect to person S at time t. Rather, it includes elements specifying a particular time and person, which makes it unqualifiedly true or false. Third, there are neither truth value gaps nor further truth values: every sentence in itself is either true or false.11 A sentence in itself consists of sub-propositional parts called “ideas in themselves” (Vorstellungen an sich) or “objective ideas” (objective Vorstellungen)—i.e. [Three is a prime number] can be decomposed, among other things, into an idea of the number three and an idea of the property of being a prime number. These ideas in themselves are also neither linguistic symbols nor mental entities but abstract objects that can turn into the contents of such things without existentially depending on them. In contrast to propositions, however, they are not true or false but “objectless” (gegenstandlos) if nothing falls under them or “objectual” (gegenständlich) if there is something that is represented by them.12 Just as a sentence in itself is composed of objective ideas, so an act of judgment consists of partial acts called “subjective ideas” (subjektive Vorstellungen; WL III, § 291, 109; EG, 47). When Ann judges that three is a prime number, she has, inter alia, subjective ideas of the number three and the property of being prime. Like the whole judgment, these subjective ideas are mental events that are caused by something and may themselves become causally operative. Their contents are constituents of the sentence in itself that is accepted in the judgment, such as the objective idea [three] and [primality] (cf. WL I, § 48, 217). Via these ideas, Ann’s judgment refers to the number three and the property of being prime.

Bolzano’s Theory of Judgment

113

Although Bolzano sometimes describes judgmental acts as “containing” a proposition, his official expression is not “content” (Inhalt) but “material” (Stoff ).13 The reason for this choice is that he already introduced “content” as a technical term for the constituents of objective sentences and ideas.14 With regard to ideas, Bolzano notes that “material” must not be misunderstood as referring to the “object” (Gegenstand) of the idea, i.e. the entity represented by it. For example, while the object of Ann’s thinking of the moon is the moon and thus something in space and time, the content of this act is an abstract entity. And even if the object is an abstract entity, too, such as in the case of thinking of Pythagoras’ theorem, it is not identical with the objective idea that is the content of this act.15 Judgments can have objects in an analogous sense—namely, things they are about. The idea in a judgment’s content that represents these things is called the “subject idea” (Subjektvorstellung); the idea representing the property attributed to these things is the “predicate idea” (Prädikatvorstellung; WL II, § 127, 9). Objectual ideas, whether subjective or objective, may cover exactly one entity or at least two. In the first case, Bolzano dubs them “singular ideas” (Einzelvorstellungen); in the second case, he uses the term “common ideas” (Gemeinvorstellungen; cf. WL I, § 68). Furthermore, ideas can be distinguished not only by the number of entities falling under them, and thus by an external feature, but also by their internal structure. Some ideas are simple insofar as they do not contain any parts, while other ideas are complex, such as [lioness], which consists of [female] and [lion].16 By combining these distinctions, Bolzano splits the class of (objective and subjective) ideas into “intuitions” (Anschauungen), “concepts” (Begriffe), and “mixed ideas” (gemischte Vorstellungen). A concept is defined as an idea that neither is nor contains an intuition; a mixed idea is an idea containing at least one intuition. Intuitions, in turn, are ideas that are both simple and singular, meaning they do not contain any proper parts and nonetheless represent exactly one object. They are expressed by the word ‘this’ in its deictic use and (in the case of human beings) represent a current mental event of the thinker in question. Since a simple idea cannot pick out its object by identifying properties, Bolzano speculates that an intuition receives its object by the natural fact that it is directly caused by it.17 It will become clear that intuitions occupy a special position within Bolzano’s epistemology. We have already seen that judgments as contents can be defined with recourse to acts of judging. Drawing on Bolzano’s terminology, the definition states that a judgment in itself is a sentence in itself that is the matter of a judgmental act. In light of this definition, it is tempting to assume that not only propositions that are accepted as true by someone but all propositions are definable with recourse to judgmental acts, for it appears that even propositions that never become contents of judgments may nonetheless be characterized as possible contents of such acts. However, Bolzano

114

Mark Siebel

reveals that he does not accept this explanation. The reason is that there are sentences in themselves that are evidently false, such as [A triangle is not a triangle], entailing that it is not possible to make a corresponding judgment.18 Bolzano is not only sceptical towards the claim that the concept expressed by ‘sentence in itself’ is open to analysis but also adopts the same critical attitude towards an analysis of ‘judgment’ in the mentalact sense (cf. WL I, § 25, 91, § 35, 157). He appears to regard it a basic notion, entailing that its explication must aim at recognition “from use or context” (WL IV, § 668, 547). Such an explication is provided by entrusting to the reader theorems in which the term ‘judgment’ is used in the relevant sense, such as “Every judgment contains a proposition that is either true or false” (WL I, § 34, 154). Since, typically, judging is coming to believe, one might think that ‘S judges that p’ can be defined by ‘S acquires the belief that p’. Bolzano would dissent from this proposal because he assumes just the opposite direction of explanation. For Bolzano, holding the belief that p is being “constantly committed to” the proposition that p (WL III, § 306, 200), which means that the believer “must not only have formed the [. . .] judgment [that p], but he must still be prepared to form the same judgment if asked for it” (WL III, § 307, 208). This analysis comes close to what many recent philosophers understand by a belief—namely, a disposition to form the corresponding judgment (cf. Evans 1981: 133). Whatever is to be made of such explanations,19 Bolzano seems to have recognized that judgments need not only produce new beliefs but may also be based on already existing beliefs. If Ann is asked whether 193 is a prime number, she might answer in the affirmative because she became engrossed in a list of prime numbers five minutes ago. That is, apart from being the acquisition of a belief, a judgment can also be its actualization.20 The fact that Bolzano is not bewildered by the ambiguity of the term ‘judgment’, but strictly distinguishes acts of judging from their contents, is part of his quite remarkable approach to logic. Logic in the wide sense, also called “Theory of Science” by Bolzano, provides the “rules that we must follow when we divide the total domain of truths into individual sciences and present them in their respective treatises” (WL I, § 1, 7; cf. § 6). It includes “Theory of Knowledge” (Erkenntnislehre), which deals with “the conditions underlying the ability to know truths, especially for humans”, and “the Art of Discovery” (Erfindungskunst), which offers “rules to be observed in reflection when the goal is to discover truths” (WL I, § 15, 59). Therefore, logic in the wide sense has to “teach us how truth may be found and error be discovered”. Since this is impossible “without attending to the way in which especially the human mind obtains its ideas and knowledge” (WL I, § 13, 54), it is dependent on psychology (cf. George 1997: 235).

Bolzano’s Theory of Judgment

115

However, logic in the wide sense contains what Bolzano dubs “Theory of Elements” (Elementarlehre) or “pure logic” (reine Logik; WL I, § 15, 59, § 16, 67). Pure logic does not establish “the laws of thought”—i.e. the rules of “appearances in the mind of a thinking being”—but is “the theory of ideas, sentences, true sentences, and inferences in themselves” (WL I, § 16, 62, § 15, 59).21 There are ties between the sphere of thinking and the an sich sphere (cf. WL III, § 276, § 294). For example, “every idea in itself is composed of at least as many parts as we distinguish in the thought idea [i.e. subjective idea] whose matter it is”. But since “it may occur that we think a complex idea [. . .] without being conscious of the thinking of its individual parts” (WL I, § 56, 244), examining acts of thinking to learn something about their contents can lead us astray. Similarly, there is a crucial difference between the “ground for recognition” (Erkenntnisgrund) of a truth and its “genuine (or objective) ground” (WL III, § 313; cf. WL II, § 198, 341). Consider Ann, who recognizes that it is 20 degrees Celsius by looking at her thermometer. Her ground for recognizing that it is 20 degrees Celsius is to be found in her judgment that the thermometer displays this temperature. But since the thermometer reading does not explain why it is 20 degrees Celsius (in fact, quite the reverse is true), the objective ground for this truth is to be looked for elsewhere.22 The relation between an objective ground and its consequence, called “grounding” (Abfolge) by Bolzano, is not an epistemic relation. It is, rather, an “objective connection that is independent of our accidental subjective recognition” and thus “holds among truths in themselves regardless of our ideas” (BM, 39f.; WL II, § 198, 341).23 It is characteristic of Bolzano’s approach to logic that he turns away from an epistemological investigation of thinkers and their judgments to turn to what Frege (1918: 69) dubbed “a third realm”. Most noticeable is Bolzano’s treatment of the a priori/a posteriori distinction.24 He applauds Kant, who used these terms to distinguish non-empirical and empirical ways of recognizing truths. But he also takes the Kantian distinction between judgments to be less fundamental than his own distinction between “intuitional propositions” (Anschauungssätze) and “conceptual propositions” (Begriffssätze). While the former include at least one intuition, the latter consist of nothing but concepts. Hence, in the case of a conceptual proposition, “persuading yourself of its truth requires no more than paying careful attention to the concepts of which it is composed” and is thus an a priori enterprise in Kant’s sense (WL I, § 42, 180f.). Briefly, propositions are not a priori because the judgments containing them are, but judgments are a priori because the propositions contained in them are. This is but one of many places where the anti-psychologistic stance that made Husserl (1913: 225) praise Bolzano as “one of the greatest logicians of all times” comes into view. In sections 4 and 5, however, it will be shown that Bolzano’s theory of judgment includes some, in this regard, unexpected ingredients.

116

3.

Mark Siebel

Judgments and Mere Thoughts

Judgments have to be distinguished from acts of merely entertaining a thought. According to a widespread view tracing back to Frege (1892: 34, fn.; 1918, 62), the essential difference lies in the relation between subject and content. The same content is just grasped in the case of a mere thought but is also acknowledged in the case of a judgment.25 If Ann merely entertains the thought that 257 is a prime number because, say, she wonders whether it is true, she grasps this thought without admitting its truth. But if Ann judges that 257 is a prime number, she accepts that 257 is a prime number, entailing that she has the corresponding belief. Bolzano clearly separates judgment from mere thought and, in this context, makes use of a standard example—namely, conditional judgment. A thinker who judges that q if p need neither believe that p nor believe that q; she merely has to entertain the given thoughts.26 Furthermore, like Frege (1919: 149–55), and unlike Brentano (1874: book II, ch. 7.7), Bolzano seems to assume that the relation between judger and content is always positive. While Brentano claims that there are two independent species of judging—namely, accepting and rejecting something—Bolzano takes rejecting to be reducible to accepting. For example, he would say that rejecting that four is a prime number consists in accepting that four is no prime number. When explaining what a judgment is, Bolzano nowhere makes use of negative verbs, such as the German counterparts to ‘reject’, ‘deny’, or ‘negate’, but deploys positive verbs, such as the counterparts to ‘assert’, ‘accept as true’, ‘acknowledge to be true’, ‘concur’, ‘recognize the truth’, and ‘believe’.27 In addition, he explicitly says that “[t]o affirm something is nothing other than to claim that something is true; to deny something is nothing other than to claim that something is not true” (WL I, § 23, 94). Surprisingly, however, Bolzano maintains that judgment and mere thought do not differ only in the relation between subject and content but also in their content. While the content of a judgment is a sentence in itself, he takes the content of an act of mere thinking to be an idea of a sentence in itself. That is, entertaining the thought that p is thinking of the proposition that p. It is a non-propositional act we also perform when we doubt that p.28 Consequently, the proposition that p is not the content but the object of a mere thought that p.29 Note that this object must be given in a special way because it is possible to think of the proposition that p without entertaining the thought that p. One can think of Goldbach’s conjecture in an explicit way—namely, as the proposition that every even number is the sum of two prime numbers, or just as the proposition Goldbach assumed to be true (whatever that proposition may be). Since entertaining the thought that every even number is the sum of two prime numbers requires an explicit representation, it cannot consist in grasping the idea [the proposition Goldbach assumed to

Bolzano’s Theory of Judgment

117

be true] but has to consist in grasping the idea [the proposition that every even number is the sum of two prime numbers]. Thus, under the terms of Bolzano’s account, the proposition that is the object of a mere thought is also part of its content.30 At first glance, Bolzano’s conception can easily be dismissed by means of the following argument. Even if someone who entertains a thought need not acknowledge it to be true, it is always appropriate to ask whether the entertained thought is true. But Bolzano makes it impossible to ask this question, for he identifies the content of entertaining the thought that p with the idea [the proposition that p], and ideas do not have truth values.31 However, on closer inspection, this argument is fairly weak because it rests on the assumption that ‘the entertained thought’ refers to the content of the act of entertaining the thought. Bolzano can thus reply that asking whether the entertained thought is true is not inquiring after the truth value of the act’s content, i.e. [the proposition that p], but only after the truth value of one of its constituents—namely, [p]. Briefly, it is asking whether the proposition the thinker has an idea of is true. There is an objection that carries more conviction.32 Consider Ben, who judges that the proposition that every even number is the sum of two prime numbers is dubious. The content of this judgment includes the idea [the proposition that every even number is the sum of two prime numbers], and this idea in turn contains Goldbach’s conjecture [Every even number is the sum of two prime numbers] (cf. WL I, § 23, 99f., § 58, 252f.). However, Bolzano holds that it is impossible to grasp a compound of objective ideas without having subjective ideas of all of its constituents (cf. WL III, § 281, 39). Hence, Ben’s judgment contains an act whose content is the proposition [Every even number is the sum of two prime numbers]. Since Ben does not accept Goldbach’s conjecture as true, this shows that Bolzano is bound to admit of non-judgmental acts having propositions as contents after all. It is therefore odd that he does not simply say that merely entertaining a thought is such an act, instead of holding the needlessly complicated view that mere thoughts have a non-propositional content.

4.

Formation of Judgments

As to their formation, Bolzano distinguishes “mediated” (vermittelte) from “unmediated” (unvermittelte) judgments. Mediated judgments are caused by other judgments. The corresponding “action of mind” is called an “inference” and the causative judgments “premises” (WL III, § 300, 123). This is close to Armstrong’s (1968: 193–200) naturalistic theory of inferences as processes in which beliefs cause the acquisition of beliefs. As to the premises, however, Bolzano’s account is superior because it substitutes beliefs with judgments. To infer something from certain assumptions, the corresponding beliefs have to become active, and the act of

118

Mark Siebel

judging fits this role quite well. As to the conclusions, we must not forget that a judgment need not be the acquisition of a belief; it can also be made if the belief is present for quite some time. In contrast to Armstrong’s account, which requires the thinker to gain a belief, Bolzano’s thus allows for inferences to conclusions that the thinker already assumed to be true (cf. WL III, § 300, 125). Finally, since both judgment and belief include acceptance, Bolzano and Armstrong are equally committed to the Fregean claim that inferring always starts with assumptions taken to be true (cf. Frege 1923: 85f.; 1980: 104, 110). At first glance, Bolzano’s naturalistic account is not able to discern inference and association. Bolzano himself mentions that the judgment ‘Caius is a scholar’ may occasion the judgment ‘Scholars are often vain’ if “the concept of a scholar brought about the associated concept of vanity” (WL III, § 300, 124). If inference were nothing but causation of judgments by judgments, the thinker in question would have inferred that scholars are often vain from the assumption that Caius is a scholar. Fortunately, Bolzano’s theory of judgment provides resources to solve this difficulty. First, Bolzano holds that the conclusion of an inference “follows upon [the premises] in time, but in such a way that they have not altogether disappeared when it comes about” because, generally, “there must obtain a certain relation of simultaneity between an effect and its causes” (WL III, § 300, 125).33 This suggests that Bolzano is mainly concerned with instantaneous causation and thus direct inference. In a direct inference, judgments instantaneously induce a further judgment. An indirect inference can then be defined as a chain of direct inferences, i.e. a chain of instantaneous causations of judgments by judgments. This way, the causal chain leading from ‘Caius is a scholar’ to ‘Scholars are often vain’ in the association case is not even an indirect inference, for the associative transition from the idea ‘scholar’ to the idea ‘vanity’, which is part of this chain, can hardly be described as the causation of a judgment by a judgment. Second, Bolzano argues that “a judgment M can only be mediated by others A, B, C, D, . . . if one of the following three cases occurs: either (a) the propositions A, B, C, D, . . . [. . .] stand to M in the relation of objective ground to its consequence, or (b) proposition M must [. . .] be deducible from A, B, C, D, . . ., or finally, (c) the proposition M, if not completely certain with respect to A, B, C, D, . . ., nonetheless must have a certain degree of probability” (WL III, § 300, 126). The relation mentioned under (a), grounding, was outlined in section 2. For the sake of brevity, I will not introduce Bolzano’s conceptions of deducibility and probability but will summarize (a) to (c) by saying that, in all inferences, the premises support the conclusion (in one of the three ways).34 That is, given that the premises are true, the probability that the conclusion is also true is higher than 0.5 and assumes the highest value of 1 if the conclusion is deducible from, or even grounded in, the premises. This condition is

Bolzano’s Theory of Judgment

119

not fulfilled in the association case because ‘Scholars are often vain’ is not even probable in the light of ‘Caius is a scholar’. Bolzano’s constraint on inferences allows for “error” (Irrtum) insofar as it allows for false belief (cf. WL III, § 307, 208). Of course, a deductively valid derivation from true premises results in a true conclusion (cf. WL III, § 309, 212f.). But in inferring something from assumptions that make it only probable, the thinker exposes herself to the danger of acquiring a false belief.35 Nonetheless, the constraint appears to be unreasonably optimistic. If the premises support the conclusion, at least by making them probable, then, even though our inferences could still exhibit deductive invalidity, they would never be inductively invalid. Hence, the rules of logic would constrain thinking insofar as they tell us when an inference does not occur—namely, if the premises do not speak for the conclusion. Bolzano’s reply is that, in the case of a fallacy, the conclusion “M is not generated by the judgments A, B, C, D, . . . alone, but by them together with the (if only tacitly assumed) judgment that from propositions like A, B, C, D, . . . a proposition like M is deducible” (WL III, § 300, 129). Let us ignore the quite severe restriction that fallacies are always based on the extra judgment that the conclusion is deducible from the premises and let us assume that the thinker may also take the premises to support the conclusion in one of the other two ways. Then Bolzano’s account of fallacies does not falsify the claim that the mediating judgments always support the conclusion because the extra judgment that the further premises make the conclusion at least probable is among them (cf. WL III, § 317, 265). The distinguishing feature of fallacies would therefore not be that the premises do not speak for the conclusion but that an invalid mode of inference is accepted in one of them. To be sure, Bolzano denies that every inference includes a judgment to the effect that the other premises support the conclusion because this would lead to an infinite regress. But we sometimes depend on such an extra judgment, as, for instance, in the case of a Baroco syllogism or other inferences that are not self-evident (cf. WL III, § 300, 127f., 135). Since “every kind of inference that we have adopted not by reflection but, as it were, instinctively must be correct” (WL III, § 300, 135), fallacies belong to the reflective type. We make use of “a rule made by ourselves, i.e. a judgment” (WL III, § 309, 212), but, unlike Baroco, this rule is not valid (cf. WL III, § 310, 221f.). Since this account of inferences intimately connects descriptive facts with normative rules, it is a legitimate subject for debate. Why should the causal order, if human beings do not intervene, follow logical standards? As a devout Christian, Bolzano could reply that it is God who, in his benevolence, initially endowed us with valid rules. Since we are free but imperfect creatures, however, it happens that we fabricate further rules that are invalid. But is there a plausible answer that does not rely on God?

120

Mark Siebel

Does an evolutionary story, in which invalid rules threaten the survival of the species, offer a way out? Although Bolzano might thus be guilty of what may be called logical optimism, it needs to be noted that his anti-psychologism goes unquestioned thereby. As to the latter, the crucial question is what someone takes to be the defining feature of valid rules of inference. Bolzano could answer that this feature is not to be found in our actual inferential behaviour but in the fact that valid rules lead from true premises to conclusions that are at least probable. Of course, if (there are circumstances in which) we commit no fallacies, then the fact that we made an inference (in these circumstances) entails that the inference is valid. However, this merely means that such facts can be used as indicators of validity—just as one can use litmus paper to find out whether a substance is acidic. These facts would still not define validity—just as acidic substances are not defined by their reaction to litmus paper but by their chemical makeup. This way, logical optimism is still a bold claim, but a psychological one that leaves logic its own subject. Unmediated judgments also have their causes, but these are not judgments. Bolzano offers two classes of examples: (1) judgments by which the thinker attributes a current mental event to herself, such as ‘I am in pain’, and (2) judgments by which a property is attributed to a current mental event, such as ‘This (pain) is strong’.36 The subject idea in judgments of the second type is an intuition in Bolzano’s sense (see section 2), and since he maintains that the idea expressed by ‘I’ is [the substance that has this (experience)], judgments of the first type also contain an intuition.37 In light of these examples, it is tempting to think that all unmediated judgments relate to the inner life of the judger and thus arise from introspection. But Bolzano adds that “purely conceptual judgments”, i.e. judgments containing no intuition at all, can also be unmediated (WL III, § 300, 132). He gives no example but points out that the cause of these judgments is the attentive inspection of the corresponding concepts.38 Unmediated judgments occupy a significant position within Bolzano’s epistemology. He argues that “judgments which are formed without mediation must be true” because “[b]y the very way in which they are produced, these judgments are secure from the danger of error” (WL III, § 311, 225).39 Here, Bolzano’s epistemic foundationalism comes to light. In his view, all our knowledge originates—either directly or via inferences—from an infallible foundation consisting of unmediated judgments. To be sure, this does not mean that all beliefs arising from this infallible bedrock are themselves true. If a thinker starts from premises that are true but make it merely probable that the conclusion is also true, it is not guaranteed that she comes to a true conclusion. Bolzano emphasizes that not all judgments by which a current mental event is attributed to oneself, or a property to such an event, are immune from error (cf. WL III, § 300, 131). If Ann judges that she feels a pain in

Bolzano’s Theory of Judgment

121

her right leg, an error cannot be ruled out because she might suffer from a phantom pain; and if she judges that the pain results from an injury, she might be wrong as well. According to Bolzano, however, such examples do not contradict the claim that unmediated judgments are infallible because these judgments are mediated. There is a crucial difference between the purely phenomenal judgment that one feels a pain and the advanced judgment that one feels a pain in a certain limb. The former is unmediated and therefore infallible because it merely points to a sensation and does not localize it in parts of the body. In contrast, since the latter amounts to the judgment that “a particular sensation [. . .] is caused by the specified part”, it is based on an inference (AT, 30; cf. WL III, § 300, 133). Ann infers that she feels a pain in her right leg from the assumption that she feels a pain and the assumption that the pain is produced by some change in her right leg. The second assumption creates the possibility of error because, among other things, the pain could be a phantom pain. The person in question need not be aware of the inference because it can be a piece of subconscious information processing. Bolzano thinks that “most of [the judgments we form] remain obscure, and we are accordingly unable either to recall or become aware of them” (WL III, § 300, 126; cf. 132f.). Particularly, while the thinker is under the impression that her judgment directly springs from the corresponding sensation, it is in fact the result of a deduction.40 Since the impression of immediateness can be misleading, Bolzano’s epistemic foundationalism concedes infallible judgments without offering an infallible criterion for them (cf. Konzelmann Ziv 2011: 28, 33f.). As plausible as an (unconscious) inference might be in the case of ‘I have a pain in my right leg’, there remain cases casting doubt on Bolzano’s account of unmediated judgments. The reason is that, in common with his account of mediated judgments, it incorporates a strong tie between descriptive facts and normative matters by stating that a judgment that is not caused by other judgments is infallible. Imagine a hypnotist who sees to it that Ann, when waking up from hypnosis and hearing a finger snap, judges that she is Napoleon’s wife. Just because the judgment is false, Bolzano has to assume that it is the result of an inference. But what could be the premises? In focussing on (1) other judgments, (2) introspection, and (3) inspection of concepts as causes of judgments, Bolzano seems to have underestimated the multitude of further causes, including unusual ones like hypnotic triggers. Behind this narrow view could be the main aim of his Theory of Science, which is to provide rules for acquiring and presenting knowledge. Because of this epistemic aim, Bolzano is primarily interested in sources of knowledge and thus in causes of judgments that are reliable—namely, those that engender more true than false judgments. Put differently, he concentrates on those causes that are active in an epistemically ideal being and thus neglects unreliable sources. As a result,

122

Mark Siebel

he advocates connections too close between descriptive and normative features of judgments.

5.

Intrinsic Qualities of Judgments

Bolzano captures Leibniz’s distinction between distinct and confused ideas, as well as the one between clear and obscure ideas, and, after some modifications, applies them to judgments. In his understanding, a judgment of a thinker S at time t is distinct if and only if S knows at t of which ideas the judgment is composed and in which way these ideas are combined. Otherwise, it is confused.41 Bolzano demands of the thinker not only knowledge of the included ideas but also knowledge of the way in which they are put together. Knowing that a judgment consists of ‘Ann’, ‘loves’, and ‘Ben’ does not suffice for distinctness because these ideas could be parts of ‘Ann loves Ben’ as well as ‘Ben loves Ann’. Moreover, distinctness is defined in a dispositional way: someone who knows the constituents and composition of her judgment need not actually think of them. A judgment of S at t is clear just in case S has an intuition of the judgment at t, that is, refers to it by a simple and singular idea (see section 2). Otherwise, it is obscure.42 The singularity of the idea ensures that a judgment does not become clear just because the judger gives thought to a whole set of judgments including it. Similarly, picking out the judgment by a complex idea would not guarantee that the judger knows it in the sense of being directly acquainted with it (cf. WL III, § 280, 28f.), for it would be possible then to think of the judgment in a purely descriptive way, as, for instance, by representing it as the 43rd judgment one made today. On the other hand, referring to the judgment via an intuition guarantees that the thinker is immediately aware of it. Note that, while Bolzano defines distinctness dispositionally, he makes clarity require a current thought. It is easy to modify the latter definition, however, by merely stating that the thinker is able to have an intuition of her judgment, which does not imply that she actually thinks of it (cf. Centrone 2010: 266). More far-reaching is Bolzano’s approach to another characteristic of judgments: their degree of confidence. He stresses that “a judgment is not a mere sum of ideas but consists in a certain efficacious combination of them” (WL III, § 293, 112).43 Unfortunately, Bolzano declares himself unable to spell out this causal specification (cf. WL III, § 291, 110). But he introduces two magnitudes determining the efficacy of judgments— namely, “vividness” (Lebhaftigkeit) and “confidence” (Zuversicht). The vividness of a judgment is a function of the vividness of the ideas included in it. The vividness of the ideas depends, among other things, on the period of time in which the objects causing the ideas affect the thinker. By contrast, the degree of confidence results from the (unexplained) “mutual influence of these ideas upon each other” (WL III, § 291, 110).44 Consider

Bolzano’s Theory of Judgment

123

a voluptuary who judges that lust engenders pain. This judgment will remain almost ineffective because, even though the voluptuary’s idea of lust is extremely vivid, he does not put much confidence in the judgment (cf. WL III, § 293, 113). The notion of confidence at play here resembles today’s notion of degree of belief, i.e. the notion by which friends of so-called subjective probability interpret probability values (cf. Hájek 2012). Similar to what came to light in the previous section, Bolzano ties himself down to astonishingly strong links between descriptive facts and normative matters,45 for he contends that “nothing can immediately determine the degree of our confidence in a judgment M except (a) the degree of absolute probability [. . .] which accrues to the proposition M with respect to the totality of judgments A, B, C, D, . . . that are currently in our mind [. . .]; and (b) the degrees of confidence of the judgments A, B, C, D, . . . themselves” (WL III, § 318, 276). In particular, if a conclusion is drawn through a deductively valid inference from premises that possess the maximal degree of confidence, then the confidence put in the conclusion is also maximal (cf. WL III, § 319, 279f.). Here, Bolzano resorts to his conception of conditional probability introduced in § 161 of the Theory of Science. Without going into the details, probability in this sense belongs to the an sich realm because it is propositions that possess a certain degree of probability in the light of other propositions, independently of what thinking beings know about these propositions or the state of affairs described by them. Thus, a proposition’s probability is an objective affair that serves as a rule for our actual trust. Against this background, Bolzano’s constraint on confidence displays tremendous optimism about the match between the confidence we actually place in our judgments and the confidence we should place in them according to this standard. Just as he takes inference to be guided by rules that are at least inductively valid, so he takes confidence to be guided by the rules of probability (cf. WL II, § 161, 187). Once again, one can hardly avoid the impression that Bolzano, due to his focus on the acquisition of knowledge, involuntarily assumes an epistemically ideal thinker (cf. Konzelmann Ziv 2010: 251f., 287f., 312f.); for his remarks on confidence make sense when read as saying that such a thinker, besides taking no account of invalid inference rules, also does not run afoul of the rules of probability. This is no real excuse for Bolzano’s overstating the case by talking about all thinkers, whether ideal or not, but it at least offers an explanation.

Notes 1. Some of the following considerations may also be found in Siebel (1999) and Siebel (2004). 2. Early Bolzano’s anticipations of this distinction may be found in BM, 39f., and AM, 178. There are considerably more places where he distinguishes

124

3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

29. 30. 31.

Mark Siebel words from the concepts denoted by them (cf. BM, 52f.; AM 144, 147f., 160; EL, 141–6, 160f., 163). Cf. § 35, 158; WL II, § 155, 128. Cf. Beyer (1996: 82, fn.), for Bolzano’s inconclusive position on the spatial localisation of mental events. Cf. AT, 85; WL III, § 291, 108f., § 297–300, 118–23; EG, 47. Cf. WL I, § 48, 217; WL III, § 272, § 291, 109; AT, 26; Frege (1919: 152). For more on Bolzano’s distinction between substances and adherences, see AT, 21–7; Schnieder (2002). Cf. WL I, § 28, 121; WL II, § 148, 89; EG 47. For a complication, see Dähnhardt (1992: 70–4). Cf. WL I, § 19, 77, § 25, 112; EG, 47. Cf. WL I, § 19, 77f., § 25, 112; WL II, § 122, 4; EG, 47. For a comprehensive comparison of Bolzano and Frege’s primary truthbearers, see Künne (1997). Cf. WL I, § 24, 108, § 25, 113; WL II, § 125, 7, § 147, 77f. Cf. WL I, § 48, 216–8; § 49, 220; § 54, 237f.; § 66, 297; EG, 47f., 51; Fréchette (2010). Cf. WL III, §291, 108; WL I, § 22, 90, § 34, 154. In WL I, § 19, 78, Bolzano uses the locution ‘content of a judgment’. Cf. WL I, § 56, 244; WL II, § 123, 5f.; WL III, § 271, 9; EG, 48. Cf. WL I, § 49, 218–20; WL III, § 280, 31. Cf. WL I, § 56, 243f., § 61; WL III, § 277. Cf. WL I, §§ 72–4; WL III, § 278; § 286, 89; VZ, 141; Textor (1996: ch. 2), George (2004), Morscher (2004), and Kripke (1980) for the most famous causal theory of reference. Cf. BE, 105; WL I, § 23, 92, § 42, 175f.; Dähnhardt (1992: 96–106). van der Schaar (2006: 38) flirts with the idea that sentences in themselves can be captured by her notion of a “judgment-candidate”. However, if the latter is assumed to denote possible judgments, as van der Schaar (2006: 37, 45f., 2007: 75f., 2015: 314–21) suggests at many places, Bolzano himself has pointed out that this is an inadequate interpretation. One difficulty is that these explications do not allow us to distinguish beliefs from dispositions to acquire a belief (cf. Audi 1994; Siebel 1999). Cf. Künne (1996: 55f.), and contrast Martin-Löf (1991: 143f.), where the term ‘judgment’ is restricted not only to an act of acquisition but also, on top of that, to the acquisition of knowledge. Cf. WL I, § 12, 47, § 16, 61–4. Cf. WL II, § 198, 340; WL III, § 312, 229. Cf. § 162, §§ 198–222. The most comprehensive study of Bolzano’s notion of grounding is Roski (2014). Cf. Textor (1996: 191–214), Roski (2013: 104–6). See Textor (2010) on Frege’s conception of judging as acknowledging the truth. Cf. WL I, § 34, 157, § 22, 86; Frege (1892: 43, 1919: 145f). Cf. WL I, § 19, 78, § 23, 99, § 34, 154f., 157. Cf. WL I, § 23, 99, § 34, 157; WL III, § 290, 108, § 306, 199; and on doubting WL I, § 34, 155, § 40, 173. Contrast WL I, § 19, 78; WL II, § 122, 4, where Bolzano deflects from his own view by stating that a proposition is the content of a mere thought. Cf. Morscher (1973: 55, 86), Casari (1986: 65), Dähnhardt (1992: 59). Cf. Dähnhardt (1992: 61–4); Künne (1997: 165, 174). van der Schaar (2015: 313f.) fields an argument of this type against Twardoski’s (1912: 186f.) claim that sentences on stage do not express possible judgments but concepts whose objects are possible judgments.

Bolzano’s Theory of Judgment

125

32. Cf. Dähnhardt (1992: 64f). See also Künne (1997: 174f.) for a further argument against Bolzano’s account. 33. More on Bolzano’s conception of causation and the required simultaneity in AT, 73–80; RW, 200–2; Krause (2004: ch. III.1). 34. More on Bolzano’s conceptions of deducibility and probability in WL II, § 155, § 161, § 255; Siebel (1996, 2002); Konzelmann Ziv (2010). 35. Cf. WL III, § 309, § 317, 263; AT, 147. 36. Cf. WL III, § 300, 131, 123, 134, § 302, 139. Specific examples of the first type can be found in WL III, § 300, 134, § 319, 278; RW, 38; AT, 54. 37. Cf. VZ, 73, 90; Textor (1996: 96–102). 38. Cf. WL III, § 303, 140; WL I, § 42, 180f. 39. Cf. WL I, § 181; WL III, § 309, 212, § 313, 231, § 317, 263; RW, 38; AT, 55f. 40. Cf. WL III, § 300, 125f., 132f.; § 303, 162. 41. Cf. WL III, § 281, 41, § 296, 117, and Centrone (2010) for a highly instructive comparison of Bolzano and Leibniz’s definitions. 42. Cf. WL III, § 280, 29, § 275, 18, § 295, 116; AT, 150; and again Centrone (2010). 43. Cf. WL III, § 291, 109f.; AT, 39. The addendum ‘efficacious’ suggests that, in the back of his head, Bolzano favours a functionalist theory of judgments, i.e. a theory singling them out by their causal relations to other mental phenomena and behaviour (cf. Levin 2018). 44. Cf. § 293, 112, § 298, 120. 45. I do not go into Bolzano’s further assumption that unmediated judgments are formed with maximal confidence, regardless of whether the thinker recognizes that the judgment is unmediated (cf. WL III, § 319, 278).

References Bolzano, B. AM = Allgemeine Mathesis (1810–12), in Bernard Bolzano Gesamtausgabe 2, A, 5, ed. J. Berg. Stuttgart: Frommann-Holzboog, 1977, 13–64. ——— AT = Dr. Bolzano’s Athanasia oder Gründe für die Unsterblichkeit der Seele. Ein Buch für jeden Gebildeten, der hierüber zur Beruhigung gelangen will. Sulzbach: Minerva, 1838; repr. Frankfurt: Minerva, 1970. ——— BE = Briefwechsel mit Franz Exner 1833–1844, in: Bernard Bolzano Gesamtausgabe 3, 4, 1, ed. E. Morscher. Stuttgart: Frommann-Holzboog, 2016. ——— BM = Beyträge zu einer begründeteren Darstellung der Mathematik. Erste Lieferung. Prague: Caspar Widtmann, 1910; repr. Darmstadt: Wissenschaftliche Buchgesellschaft, 1974; Eng. trans.: “Contributions to a Better-Grounded Presentation of Mathematics”, in The Mathematical Works of Bernard Bolzano, ed. and trans. S. Russ. Oxford: Oxford University Press, 2004, 83–138. ——— EG = Einleitung zur Größenlehre (1830–35), in Bernard Bolzano Gesamtausgabe 2, A, 7, ed. J. Berg. Stuttgart: Frommann-Holzboog, 1975, 25–216. ——— EL = “Etwas aus der Logik” (ca. 1812), in Bernard Bolzano Gesamtausgabe 2, A, 5, ed. J. Berg. Stuttgart: Frommann-Holzboog, 1977, 139–68. ——— RW = Lehrbuch der Religionswissenschaft, ein Abdruck der Vorlesungshefte eines ehemaligen Religionslehrers an einer katholischen Universität, von einigen seiner Schüler gesammelt und herausgegeben. Sulzbach: Seidel, 1834; repr. in Bernard Bolzano Gesamtausgabe 1, 6, 1–1, 8, 4, ed. J. Louzil. Stuttgart: Frommann-Holzboog, 1994–2005. ——— SW = “Selbstkritik der Wissenschaftslehre” (1848), in Bernard Bolzano Gesamtausgabe 2, A, 12, 2, ed. J. Berg. Stuttgart: Frommann-Holzboog, 1981, 185–9.

126

Mark Siebel

——— WL = Wissenschaftslehre. Versuch einer ausführlichen und größtentheils neuen Darstellung der Logik mit steter Rücksicht auf deren bisherige Bearbeiter. Sulzbach: Seidel, 1837; repr. in Bernard Bolzano Gesamtausgabe 1, 11, 1–1, 14, 3, ed. J. Berg. Stuttgart: Frommann-Holzboog, 1985–2000; Eng. trans.: Theory of Science, ed. and trans. P. Rusnock and R. George. New York: Oxford University Press, 2014. ——— VZ = Verbesserungen und Zusätze zur Logik (1832–48), in Bernard Bolzano Gesamtausgabe 2, A, 12, 2, ed. J. Berg. Stuttgart: Frommann-Holzboog, 1977, 53–184. Armstrong, D. M. (1968), A Materialist Theory of the Mind. London: Routledge. Audi, R. (1994), “Dispositional Beliefs and Dispositions to Believe”, Noûs 28: 419–34. Beyer, C. (1996), Von Bolzano zu Husserl. Eine Untersuchung über den Ursprung der phänomenologischen Bedeutungslehre. Dordrecht: Kluwer. Brentano, F. (1874), Psychologie vom empirischen Standpunkte. Leipzig: Duncker & Humblot; 2nd ed. Leipzig: Meiner, 1924; repr. Frankfurt: ontos, 2008; Eng. trans.:x Psychology From an Empirical Standpoint. London: Routledge, 1973. Casari, E. (1986), “Bemerkungen über die Bolzanosche Wissenschaftslehre”, in Logik und Grundlagenforschung. Festkolloquium zum 100. Geburtstag von Heinrich Scholz. Münster: Aschendorff. Centrone, S. (2010), “Bolzano und Leibniz über Klarheit und Deutlichkeit”, Archiv für Geschichte der Philosophie 92: 256–89. Dähnhardt, S. (1992), Wahrheit und Satz an sich. Zum Verhältnis des Logischen zum Psychischen und Sprachlichen in Bernard Bolzanos Wissenschaftslehre. Pfaffenweiler: Centaurus. Evans, G. (1981), “Semantic Theory and Tacit Knowledge”, in S. H. Holtzman and C. M. Leich (eds.), Wittgenstein: To Follow A Rule. London: Routledge. Fréchette, G. (2010), Gegenstandslose Vorstellungen: Bolzano und seine Kritiker. Sankt Augustin: Academia. Frege, G. (1892), “Über Sinn und Bedeutung”, Zeitschrift für Philosophie und philosophische Kritik 100: 25–50; repr. in Funktion, Begriff, Bedeutung, ed. M. Textor. Göttingen: Vandenhoeck & Ruprecht, 2002, 2nd ed. 2007; Eng. trans.: “On Sense and Meaning”, in Collected Papers on Mathematics, Logic and Philosophy, ed. B. McGuinness. Oxford: Blackwell, 1984. Frege, G. (1918), “Der Gedanke. Eine logische Untersuchung”, Beiträge zur Philosophie des deutschen Idealismus 2: 58–77; repr. in Logische Untersuchungen, ed. G. Patzig. Göttingen: Vandenhoeck & Ruprecht, 1966, 3rd ed. 1986; Eng. trans.: “Thoughts”, in Collected Papers on Mathematics, Logic and Philosophy, ed. B. McGuinness. Oxford: Blackwell, 1984. Frege, G. (1919), “Die Verneinung. Eine logische Untersuchung”, Beiträge zur Philosophie des deutschen Idealismus 1: 143–57; repr. in Logische Untersuchungen, ed. G. Patzig. Göttingen: Vandenhoeck & Ruprecht, 1966, 3rd ed. 1986; Eng. trans.: “Negation”, in Collected Papers on Mathematics, Logic and Philosophy, ed. B. McGuinness. Oxford: Blackwell, 1984. Frege, G. (1923), “Gedankengefüge”, Beiträge zur Philosophie des deutschen Idealismus 3: 36–51; repr. in Logische Untersuchungen, ed. G. Patzig. Göttingen: Vandenhoeck & Ruprecht, 1966, 3rd ed. 1986; Eng. trans.: “Compound Thoughts”, in Collected Papers on Mathematics, Logic and Philosophy, ed. B. McGuinness. Oxford: Blackwell, 1984.

Bolzano’s Theory of Judgment

127

Frege, G. (1980), Gottlob Freges Briefwechsel, ed. G. Gabriel, F. Kambartel and C. Thiel. Hamburg: Meiner. George, R. (1997), “Psychologism in Logic: Bacon to Bolzano”, Philosophy and Rhetoric 30: 213–42. George, R. (2004), “Intuitions: The Theories of Kant and Bolzano”, in M. Siebel and M. Textor (eds.), Semantik und Ontologie. Beiträge zur philosophischen Forschung. Heusenstamm: ontos. Hájek, A. (2012), “Interpretations of Probability”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2012 Edition), URL = https://plato. stanford.edu/archives/ win2012/entries/probability-interpret/. Husserl, E. (1913), Logische Untersuchungen. Band 1: Prolegomena zur reinen Logik, 2nd rev. ed., Halle: Niemeyer; Eng. trans.: Logical Investigations, Vol. 1, trans. J. N. Findlay. London: Routledge and Kegan Paul, 1970. Konzelmann Ziv, A. (2010), Kräfte, Wahrscheinlichkeit und “Zuversicht”. Sankt Augustin: Academia. Konzelmann Ziv, A. (2011), “Bolzanian Knowing: Infallibility, Virtue and Foundational Truth”, Synthese 183: 27–45. Krause, A. (2004), Bolzanos Metaphysik. Freiburg and Munich: Karl Alber. Kripke, S. (1980), Naming and Necessity. Cambridge, MA: Harvard University Press. Künne, W. (1996), “Thought, Speech, and the ‘Language of Thought’”, in C. Stein and M. Textor (eds.), Intentional Phenomena in Context: Papers From the 14th Hamburg Colloquium on Cognitive Science. Hamburg: Graduiertenkolleg Kognitionswissenschaft. Künne, W. (1997), “Propositions in Bolzano and Frege”, in Versuche über Bolzano: Essays on Bolzano. Sankt Augustin: Academia, 2008, 157–95; orig. published in W. Künne, M. Siebel and M. Textor (eds.), Bolzano and Analytic Philosophy (Grazer Philosophische Studien 53), Amstermda: Rodopi. Levin, J. (2018), “Functionalism”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Fall 2018 Edition), URL = https://plato.stanford.edu/ archives/fall2018/entries/functionalism/. Martin-Löf, P. (1991), “A Path From Logic to Metaphysics”, in Società italiana di logica e filosofia delle scienze (eds.), Atti del Congresso “Nuovi problemi della logica e della filosofia della scienza” II. Bologna: CLUEB. Morscher, E. (1973), Das logische An-Sich bei Bernard Bolzano. Salzburg and Munich: Pustet. Morscher, E. (2004), “Kann es in der Welt 3 Indexikalität geben?”, in M. Siebel and M. Textor (eds.), Semantik und Ontologie. Beiträge zur philosophischen Forschung. Heusenstamm: ontos. Roski, S. (2013), “A Priori Knowledge in Bolzano, Conceptual Truths, and Judgments”, in M. van der Schaar (ed.), Judgment and the Epistemic Foundation of Logic. Dordrecht: Springer. Roski, S. (2014), Bolzano’s Notion of Grounding and the Classical Model of Science. PhD thesis, Free University Amsterdam. Schnieder, B. (2002), Substanz und Adhärenz. Bolzanos Ontologie des Wirklichen. Sankt Augustin: Academia. Siebel, M. (1996), Bolzanos Begriff der Ableitbarkeit. Sankt Augustin: Academia. Siebel, M. (1999), “Bolzanos Erkenntnistheorie”, in E. Morscher (ed.), Bernard Bolzanos geistiges Erbe für das 21. Jahrhundert. Sankt Augustin: Academia.

128

Mark Siebel

Siebel, M. (2002), “Bolzano’s Concept of Consequence”, The Monist 85: 580–99. Siebel, M. (2004), “Bolzanos Urteilslehre”, Archiv für Geschichte der Philosophie 86: 56–87. Textor, M. (1996), Bolzanos Propositionalismus. Berlin and New York: de Gruyter. Textor, M. (2010), “Frege on Judging as Acknowledging the Truth”, Mind 119: 615–55. Twardowski, K. (1912), “Funktionen und Gebilde”, manuscript, Varsovian Twardowski Archive, folio 70; publ. by J. Brandl in Conceptus 29: 157–89; repr. in Kasimir Twardowski. Gesammelte deutsche Werke, ed. A. Brozek, J. Jadacki and F. Stadler. Cham: Springer Nature Switzerland, 2017; Eng. trans. of the Polish version of the paper: “Actions and Products. Some Remarks from the Borderline of Psychology, Grammar and Logic”, in On Actions, Products and Other Topics in Philosophy, ed. by J. Brandl and J. Wolenski. Amsterdam: Rodopi. van der Schaar, M. S. (2006), “On the Ambiguities of the Term Judgment. An Evaluation of Twardowski’s Distinction Between Action and Product”, in A. Chrudzimski and D. Łukasiewicz (eds.), Actions, Products and Things: Brentano and Polish Philosophy. Frankfurt: ontos. van der Schaar, M. S. (2007), “The Assertion-Candidate and the Meaning of Mood”, Synthese 159: 61–82. van der Schaar, M. S. (2015), “The Things We Call True”, Grazer Philosophische Studien 91: 303–21.

6

Correctness First Brentano on Judgment and Truth Mark Textor

1.

Introduction

Truth is one of our most central concepts. Many philosophers tried to get clear about truth by giving definitions of this concept that decompose it into its marks. Franz Brentano took this approach to be of limited value. According to him, the primary question about any concept is how we acquire it, not how to define it. He argued that the concept of truth is derived from our awareness of correct judging, where correctness is a notion prior to truth. Truth stands to judgment as goodness to love: x is good if, and only if, x is correctly loved; x is true if, and only if, x is correctly judged. In current philosophy, Brentano’s correctness is often called ‘fittingness’.1 In this terminology, Brentano proposed that fittingness is the primitive notion that allows us to understand value in general and truth in particular. In this paper, I will use an objection made by Moore to develop and defend Brentano’s story of how we come to acquire the concept of correctness. In particular, I will argue that we need to revise our conception of self-evident judgment if we want to be a fittingness-first theorist like Brentano.

2.

The Manifold Senses of ‘True’ and Brentano’s Methodological Advice Now, the concept of truth undoubtedly arises from inner perception. But in order for this to become quite clear, attention must be drawn to the fact that the word ‘true’ is employed in several meanings which are to be distinguished from one another. (FCA, 137 [85])2

We say of things of different kinds that they are true/false. Here are some of Brentano’s examples:3 His aim was true. He is a true friend. His dream became true.

130

Mark Textor Some sentences in this book are true. This picture is a true representation of the situation. What he said was true.

The meanings of these different occurrences of ‘true’ are related. Brentano’s model for their relation is ‘healthy’. There are healthy people, healthy foods, healthy activities. Here, ‘healthy’ expresses different concepts, but one among them is central: It is the healthy body that is healthy in the strict or proper sense; other things are called healthy because they impart, enhance, or establish health. (OCT, 4 [6]) Similarly, among the many related concepts expressed by uses of ‘true’, Brentano takes one to be central. It figures in explanations of all other truth concepts, while none of the other concepts figure in its explanation. Which concept is that? Brentano picks out the core meaning of ‘true’ by the kind of thing that is true in this sense: Truth, in its proper sense, is found in judgments. (OCT, 138 [86)]) Brentano’s guiding hypothesis is that if something other than a judgment is called ‘true’, then the meaning of this use of ‘true’ has to be explained in terms of the application of ‘true’ to judgments. Take some examples: a true statement is a statement that expresses a true judgment, a true friend is a friend that induces us to make a true judgment to the effect that he is our friend, and so on.4 In turn, the application of ‘true’ to judgments cannot be explained any further. Why not? Because we can be aware of judgments, and by attending to our awareness, we learn what a true judgment is. His lecture on truth ends therefore in a methodological sermon directed to those philosophers who try to define truth: We have been concerned with a definition, i.e., with the elucidation of a concept connected with a name. Many believe such elucidation always requires some general determination, and they forget that the ultimate and most effective means of elucidation must always consist in an appeal to the individual’s intuition, from which all our general marks [Merkmale] are derived. What would be the use of trying to elucidate the concepts red or blue if I could not present one with something red or with something blue? All this has been disregarded by those who were concerned with the nominal definition of truth, whose history we have pursued. If, as I hope, we have succeeded in

Correctness First

131

clarifying this clouded [getrübt] concept, we have done so only by focussing primarily on examples of true judgements. [. . .] Even now, after the elimination of confusions and misunderstandings, our definition would convey nothing to one who lacked the necessary intuition. (OCT, 17 [29], my emphasis, and I have changed the translation) The ultimate and most effective means of conceptual elucidation consists in an appeal to our intuition. What does this mean? Brentano aims to bring about that he and his readers perceive or episodically remember perceptions of the property that falls under the concept to be clarified. If we can attend to this property when we remember perceivings of it, we will be able to answer philosophical questions pertaining to its concept. He is even more explicit in other writings: The task of determining a concept is very closely connected to the question as to the source from which we attain it. The explanation of a term is in the last analysis a reference to certain phenomena. (FCA, 84–5 [135], my emphasis) For example, if you episodically remember how it was for you when you saw one billiard ball hit another, you can episodically remember the immediate feeling of determination of one perception by another. You have then a starting point to assess views about causation. Brentano implements this methodology in his lecture “The Origin of Our Knowledge of Right and Wrong” (OKRW) which is a constructive counterpart to the mainly critical lecture on truth. He articulates the aim of his lecture as follows: In order to answer these questions [What do ‘the best’, what does ‘good’ mean? How do we recognize what is good?] in a satisfactory way we have first and foremost to find the origin of the concept of good which, like the origin of all our concepts, lies in certain concrete intuitive presentations. (OKRW, 8 [16], in part my translation) Brentano aims to get his readers to have an intuitive presentation of one or some instances of the property, the concept of which is fundamental for our understanding of value. This concept is not being good but correctness. He holds that our concepts of truth and goodness are based on, as Olson nicely put it, “an experiential feature of correctness with a distinctive phenomenology” (2017: 116). The basic normative property is not only conceptually primitive but is also supposed to be given to us when we are aware of our judgments. Brentano’s lecture is only successful if he manages to make us aware of correctness. Let’s see whether this methodology is successfully employed.

132

3.

Mark Textor

The Origin of Truth: Opposition and Correctness

In OKRW, Brentano outlined a story of how we acquire the concept of a right or correct attitude. I think it is a false start, but working through it will be helpful for the further discussion. Brentano starts by comparing and contrasting mental phenomena. The main division is between those that have opposites (judgment, love) and those that lack them (presentation): In the case of judgment there is the opposition between affirmation or acceptance, on the one hand, and denial or rejection, on the other. In the case of the emotions there is the opposition between love and hate or, as we may also put it, the opposition between inclination and disinclination, between being pleased and being displeased. But in the case of mere presentation—in the mere having of an idea—there is no such opposition. (OKRW, 10 [18]) Acknowledgement (love) and rejection (hate) are opposites: one cannot acknowledge (love) and hate (reject) the same thing at the same time. What is the modality expressed by ‘cannot’? Brentano does not explicitly answer this question in OKRW. But in later work he argued that “the unification of opposed judgment relations in one mental subject is not obviously impossible . . .” (V, 21). It at least seems possible that one can acknowledge and reject the same thing at the same time. For instance, in the waterfall illusion, one has a visual experience of movement and one of stillness at the same time. One acknowledges as well as rejects the movement in perception. Hence, the ‘cannot’ Brentano has in mind is plausibly taken to be deontic: if I love X as well hate X at the same time, I get something wrong. One way to articulate this further is to ascribe to love (judgment) correctness conditions: if I correctly acknowledge (love) X, I incorrectly reject (hate) X (and the other way round). In contrast, presentations are mental acts that lack opposites and are therefore neither correct nor incorrect. For example, there is no opposition between visual experiences independently of an incompatibility between the objects they are directed on. It is not possible that two distinct experiences with the same content are opposed to each other. The notion of correctness or rightness (‘fittingness’) will play a key role in the following sections. So let’s see whether we can get an initial grip on it. Correctness applies to actions in general and mental acts in particular. Things we do or that happen to us can be said to be correct or incorrect: Horst pronounced ‘thespian’ (in)correctly. John mixed the martini (in)correctly. John judged (in)correctly that it will rain.

Correctness First

133

In contrast, these adverbs and corresponding adjectives don’t ‘fit’ sentences and propositions.5 Consider: The proposition that 2 + 2 = 4 is right. (?) The proposition that 2 + 2 = 4 is correct. (?) The proposition that 2 + 2 = 5 is incorrect. (?) We can distinguish correctness from other properties pertaining to mental acts, such as being justified by saying what makes a judgment correct. The following statements ring true: What makes it correct to judge that it is raining is that it is raining. What makes it correct to judge that blood is red is that blood is red. If we replace ‘correctness’ with ‘justification’, we arrive at implausible statements: What justifies judging that it is raining is that it is raining. What justifies judging that blood is red is that blood is red. In sum, correctness seems to be the most general dimension of evaluations for activities. I can, for example, correctly admire, love, or hate as well as correctly judge, deny, or question. Brentano argues that this general notion of correctness is the source of our understanding of value, with truth and goodness being special cases. Hence, we should revise the guiding hypothesis: any object that is not a judgment is true/false in virtue of a relation to a correct/incorrect judgment. Now, if we have recognized that there are pairs of mental acts of which only one can be correct, we have found what we have been looking for. We have arrived at the source of our concepts of the good and the bad, along with that of our concepts of the true and the false. (OKRW, 11 [19]) The explanation of goodness and truth is now supposed to be straightforward: We call a thing true when the acknowledgement relating to it is correct. We call a thing good when the love relating to it is correct. In the broadest sense of the term, the good is that which is worthy of love, that which can be loved with a love that is correct. (OKRW, 11 [19], my emphasis) If we have the notion of a correct attitude, we can easily arrive at the notion of an object towards which it is correct to bear this attitude.6

134

Mark Textor

In PES, Brentano had already described why the step from correct attitude to truth is forced upon us: Our presentations of truth and falsity, too, presuppose and are acquired by reflection upon judgements, as no one would doubt. If we say that every acknowledging judgement [anerkennendes Urteil] is an act of taking something to be true, and every rejecting judgement [verwerfendes Urteil] an act of taking something to be false, this does not mean that the former consists in predicating truth of what is taken to be true and the latter in predicating falsity of what is taken to be false. Our previous discussions have shown, rather, that what the expressions denote is a particular kind of intentional reception of an object, a distinctive kind of mental reference to a content of consciousness. The only correct interpretation is that anyone who takes something to be true will not only acknowledge the object, but, when asked whether the object is to be acknowledged, will also acknowledge the object’s to-be-acknowledgedness [das Anzuerkennensein], i.e. its truth (which is all that is meant by this barbarous expression). (PES, 185–7 [II, 89], my emphasis and in part my translation) When I judge that p, I am committed to the correctness of my judging; that is, when the question arises of whether the attitude I have towards the content is one that one ought to have, I can only, on pain of irrationality, deny that this is so. If the content acknowledged is an object, we arrive at the notion of existence (more about this later). If the content acknowledged is a state of affairs, we arrive at the notion of truth: (TrueC) x is true if, and only if, it is correct to acknowledge x. (TrueC) is not a definition of the primary notion that anchors our uses of ‘is true’; it explains the meaning of an extended use of ‘true’ in which it applies to what is acknowledged in a judgment: It is hardly necessary to observe that when I spoke in the lecture [OKRW] of things being true and false, I was using the terms in their derivative sense and not in their strict and proper sense. (OKRW, 48 [59]) The derivative sense of ‘true’ can be defined, but only by appeal to a more fundamental sense of ‘true’, which is less misleadingly expressed by ‘correct’.

4.

Moore’s Objection

Has Brentano been able to guide us to the perception of correctness, thereby putting us in a position to acquire the core concept that

Correctness First

135

allows us to understand the many meanings of ‘true’? Prima facie, the answer is NO. We have been given a relation (opposition) between mental acts, and if we want to understand why the relation obtains, we will need to introduce a property like correctness. But while one may argue that the obtaining of the relation is given in consciousness, the property that grounds the obtaining of the relation is not given in consciousness. In his review of Brentano’s OKRW, Moore highlighted the problem. If the correctness/incorrectness of my judgments were introspectable, just paying attention to my judgments would allow me to revise the incorrect ones. An incorrect judging would immediately be retracted. Hence, it seems implausible that correctness/incorrectness is an introspectable property. Moore elaborated this point with respect to goodness, but his remarks carry over to truth: Obviously the conception of “good”, as Brentano defines it, cannot be derived merely from the experience of loving, but only from that of “right loving”—from the perception of the rightness of a love: its origin cannot be merely the perception of a love which is right, but in which this quality is not perceived, it can only be a perception in which it is itself contained. But whereas the experience of loving has all the marks which are suggested by calling it “a concrete impression of psychical content”, the “experience of right loving”—i.e., the perception of the rightness of a love—has not. The quality of rightness is not a psychical content and the perception of it is not an impression in the ordinary sense of these words. A single mark is sufficient to distinguish it: by a “psychical content” we always mean at least an existent, and by “impression” the cognition of an existent, and “rightness” is not an existent. (Moore 1903: 118) I can see John drunk in the yard without seeing his being drunk. Similarly, I can be aware of judging correctly without being aware of the correctness of my judging. Moore goes on to argue that rightness is not an ‘existent’, an object that can be encountered in awareness. On one understanding, this is right: the property cannot be encountered in awareness; no property can be so encountered. But particularized instances of the property can be encountered. The main challenge Moore poses for Brentano is therefore to show how awareness of correct judging can also provide us with awareness of correctness. Otherwise, Brentano has not guided us to a perception of correctness. (Moore surmises that we perceive rightness or correctness but not in the ordinary sense. We may come to see that a judgment is correct, but this is a kind of judgment.) In sum, Brentano has so far failed to put us into contact with the property of correctness. Can he do better?

136

5.

Mark Textor

Answering Moore’s Objection: Self-Evident Judgment and Correctness

To counter Moore’s objection, Brentano needs to argue that we are aware of not only correct judgings but also their correctness. Moore is right that, in general, we are not aware of correctness when we are aware of our judging or loving. But Brentano holds that self-evident judgments are an exception.7 On a widespread and prima facie plausible characterization of selfevidence, this seems to be the wrong starting point. When philosophers write about self-evidence, they have propositions in mind that have the following feature: to recognize their truth, it suffices to grasp them. For example, Frege holds that “the truth of a logical law is immediately evident from itself” (Frege 1923: 16 [50]). Now, if one recognizes the truth of such a thought, one makes a judgment that needs no further justification. One judges that p in this case because, as some authors put it, one ‘sees that (the proposition) that p is true’. For instance, Plantinga takes this phrase to describe the phenomenology of coming to, for example, the judgment that everything is self-identical. It has a peculiar sort of phenomenology with which we are all well acquainted, but which I can’t describe in any way other than the phenomenology that goes with seeing that such a proposition is true. (Plantinga 1993: 106) If one starts with self-evident propositions, one arrives at the following widely held and influential alethic conception of self-evident judgment: (EvidentTruth) A judgment that p is self-evident only if in judging that p one is directly aware of the truth of p and this awareness is the epistemic basis for the judgment. If self-evidence is a matter of obviousness of truth, self-evident judgments cannot be the source of the concept of correctness; for they don’t put us into contact with this property. Rather, truth seems to be the primary property. Hence, anyone who holds the alethic conception of self-evidence will find Brentano’s position unmotivated and even puzzling. For several months, Brentano unsuccessfully tried to convince his pupil Kraus of his view of truth. In a letter to Kraus, Brentano recognized that acceptance of the alethic conception of self-evident judgment blocks people from following his argument: I returned again to the question about correctness [Richtigkeit] and truth of a judgement in order to make clear what makes it so difficult for me to convince you. It seemed to me that the reason is that you

Correctness First

137

think that the immediately self-evident judgement is only characterised as true, but not also as logically correct, while the opposite is the case. It bears the character of a judgement, as it ought to be. Therein lies implicitly that it is true, what minimally means that it is logically correct, in that it suffices [for truth] to match the logically correct judgement in quality, object, time aspect and modality. (Letter to Kraus, 17 May 1916. In AVR, 306, my translation and emphasis) The intended opposite of ‘only characterized as true and not also as (logically) correct’ seems to be ‘only characterized as (logically) correct and not also as true’. Saying that a judgment is characterized as correct does not merely mean that it is correct; it must disclose its correctness.8 The clearest statement of this idea can be found in LRU: There are judgements that announce their correctness [sich als richtig kundgeben]; that have the character of insight. One calls them self-evident. (LRU, 141, my translation and emphasis) This is not meant as an explanation of self-evident judgment. One learns what a self-evident judgment is by making such judgments and comparing them to non–self-evident or blind ones. But it is a necessary condition, for being a self-evident judgment means the judgment announces its correctness. Further necessary conditions are infallibility and indubitability. I will set them aside for the further discussion. For our purposes, the central feature of self-evident judgments is that they are ‘characterized as correct’ or that they ‘announce’ their correctness. In what sense does a judgment ‘announce’ itself as correct? If a judgment ‘announces’ its correctness, the judger is aware of her judging as well as the correctness of it. It is like seeing a bright red apple: you see the apple, and its redness cannot escape you. You don’t need to infer or ascertain in some other way that the apple is red: seeing the apple is also seeing its redness. In the letter quoted above, Brentano also describes self-evident judgments as judgments that are as they ought to be. When you consciously judge that hearing is going on now, you are thereby aware that you get things right, and you judge as one ought to judge. That there is such a primitive sense of ‘ought’ seems independently plausible. Ginsborg argues that a child who is sorting geometrical shapes will have it: When she puts a cube together with the other cubes rather than with the spheres, her action, even if unhesitating, is not ‘blind’: she does it with a sense that it is the appropriate thing to do, that this is where

138

Mark Textor the cube belongs, that this is what she ought to be doing with the cube. (Ginsborg 2006: 361)

One does not need to possess the concepts of sorting or correctness to have the sense of correctness under consideration: the child can simply be aware that what she does right now should be done in this way.9 Ginsborg makes an additional point that is important for our purposes. She says the normativity involved in (perceptual) judgment is primitive: We are simply aware that we are perceiving as we ought, without that awareness depending on the appreciation of anything either about our way of perceiving or about the object, in virtue of which we ought to perceive the object that way. Relatedly, the normativity involved is not the normativity associated with veridicality or truth. (Ginsborg 2015: 185) If the ‘ought’ in ‘judges as one ought to judge’ were not primitive in this sense, Brentano would not have identified a property distinct from and prior to truth, for there would be a need to be a prior feature of the judgment that made the judgment correct. The most likely candidate for this feature is truth: An attitude is fitting when its object meets its standard. For instance, belief has truth as its standard of fittingness, and so a belief is fitting when what is believed is true. (McHugh and Way 2018: 165)10 On this view, our awareness of judging correctly is due to our awareness that the judged content is true. Hence, truth and seeing the truth of a content would be more fundamental than judgmental correctness, and Brentano’s project would not get off the ground. According to Brentano (see letter to Kraus quoted above), in self-evident judgment, we are aware of the correctness of our judgment without being aware that it meets an independent standard—namely, truth. The activity itself sets the standard for its correctness.11 We then work out that correctness goes hand in hand with further properties: truth of content judged, and everyone who judges like us also makes a correct judgment. But these properties neither figure in an explanation of correctness nor are standard for correctness. To sum up, self-evident judgments are ones where judging and awareness of correctness of what one does are inseparable: (EvidentCorrect) A judgment j is self-evident only if in making j one is directly aware of j’s correctness.

Correctness First

139

If one is aware of the judgment’s correctness—the property of its being as a judgment should be—one is aware of its being knowledge without an independent reason. In contrast, blind judgments only commit us to the judgment that they are correct (see previous section). We don’t make them in the awareness of their correctness. The awareness itself cannot be a propositional judgment; otherwise, we would already need to bring the concept of correctness to bear. The awareness is a direct uptake of a case of correctness, an instance of this property. We are supposed to be confronted with a clear and unmissable case of correctness in awareness from which we can learn which property is under consideration.12 Now we have found the source of the concept of correctness. Selfevident judgments and only self-evident judgments put us in contact with the property of correctness. This quasi-perceptual contact combined with selective attention enables us to form the concept of correctness. If we then go with Brentano’s line in OKRW, we can explain truth as a derived concept that applies to contents of judgments: a true content is a content that it is correct to judge. Because correctness is a fundamental concept, we need to make self-evident judgments to acquire the concept of correctness and the derived concept of truth: [The concept of truth] can only be extracted [entnehmen] from such instances of evident judgments. Anyone who wishes to understand what the word ‘true’ says has to be referred to them. But no rhetoric, no analytic acumen, can teach their sense [Sinn] to someone who has never had the experience of an evident judgment and who is therefore not in a position to compare it with other judgments which lack this characteristic. (FCE, 88 [141]) We make self-evident judgments all the time without effort and intellectual sophistication, for awareness of our mental life consists in evident acknowledgements of our present mental activities: Every mental act is [. . .] accompanied by a twofold inner consciousness, by a presentation which refers to it and a judgment about it, the so-called inner perception which is an immediately evident cognition [Erkenntnis] of the act. (PES, 203 [I, 111]) When I listen to C ♯, I acknowledge my listening to C ♯ with evidence. My attitude to my mental activity is correct and I am aware of ‘doing the right thing’ when I acknowledge it. So paying attention to our mental life will provide us with an experience of correctness. Since we make such judgments all the time from early on, we can easily acquire the concept of correctness by paying attention to them.

140

Mark Textor

Brentano’s answer to Moore’s problem is that while, in general, correctness is not given to us when we are aware of our judgment, there is a special class of judgments whose distinctive feature is that one cannot make them without being aware of their correctness. We acquire the concept of truth/goodness from our awareness of self-evident judgings and likings. Our awareness of evident judgings provides us, then, with ‘contact’ with correctness. It is worth emphasizing again that Brentano’s aim is not to provide a definition of truth. He wants to put us in direct contact with the property of correctness that will then put us in a position to derive the concepts expressed by ‘true’. He has achieved the task he has set himself when we are aware of correctness and can start to define the derivative concepts expressed by ‘true’ in its many meanings. Brentano is therefore unjustly taken to task for providing an implausible epistemic definition of truth.13 For instance, saying, as Schlick (1910: 43) did with reference to some Brentanians, that ‘truth consists in the feeling of inner evidence’ reveals a serious misunderstanding. Self-evidence plays a role in an account of the acquisition of a conceptually primitive property of correctness, but it does not play a role in the explanation of correctness. Making a self-evident judgment enables us to acquire the concept of correctness, but self-evidence is neither correctness nor truth; it is, at best, a special kind of correctness.

6.

Objections to Brentano’s Acquisition Account

I will now address two objections that arise with regard to Brentano’s account. The first one is well known, the second one not because not only Oskar Kraus et al. failed to take on board the normative character of Brentano’s notion of self-evidence. The Ehrenfels objection first:14 it is obvious that there are blind judgments that are correct. How do we manage to apply the primitive concept of correctness to them? Brentano’s attempts to answer this question are unsatisfactory. Sometimes he moves from an account of acquisition to an elucidation of the concept of correctness. He defines ‘truth’ directly as evident judgment: “True judgment” is ambiguous. In its original sense it means as much as evident judgment; in the derived sense one calls a blind judgment that corresponds to an evident judgment in all parts true. (VE, 149, my translation) Brentano claims here that there is not one, but two senses of ‘true’, and ‘true’ is supposed to apply in both senses to judgings. This is just one ambiguity too much, I fear. If ‘truly’ had a different sense with respect to different judgment contents, an inference such as

Correctness First

141

John truly/correctly judged that 1 = 1 John truly/correctly judged that 17 + 5 = 22 There are at least two things that John truly/correctly judged would be fallacious. But it isn’t. Are there any independent reasons to say that ‘judged truly’ is ambiguous? I am not aware of any. When commenting on the formula ‘veritas est adequatio ad rei et intellectus’, Brentano suggested a general way of saying when one judges ‘truly’ correctly:15 Truth pertains to the judgment of the person who judges correctly— to the judgment of the person who judges about a thing in the way in which anyone whose judgments were evident would judge about the thing; hence it pertains to the judgment of one who asserts what the person whose judgments are evident would also assert. (TE, 82, [139]) Ehrenfels is reported to have argued against Brentano’s proposal as follows:16 there may be facts, such as that there is a diamond weighing exactly 100 kilograms, that are unrecognizable. A fortiori, such facts cannot be acknowledged or rejected in self-evident judgments. Hence, by Brentano’s explanation, the judgment that there is a diamond weighing exactly 100 kilograms can’t be true. But I can make a blind judgment that there is a diamond weighing exactly 100 kilograms and get lucky: my judgment is correct, and it is one that one ought to make. How can Brentano respond to the Ehrenfels objection? Kriegel (2018) works through several options and finds them all unconvincing. He suggests that one should resist any explanation of correctness in terms of self-evidence: In my opinion, Brentano’s best move here is to go primitivist about belief-fittingness directly and construe self-evidence as just a particularly acute or manifest instance of fittingness. On this view, the only way to grasp the nature of belief fittingness is to contemplate side by side fitting and unfitting beliefs in things, and this is easiest to do with the most starkly fitting beliefs, namely, the self-evident ones. (Kriegel 2018: 141) I agree, and this is congenial to the partial account of self-evidence offered here. We can come into contact with the property of correctness by making self-evident judgments. Then we need to apply it in different cases. Here, we need reasons to apply the same property. But that does not make the property conceptually complex. As we develop different methods for checking that something is red, we need to develop different methods for checking that something is correct.

142

Mark Textor

Onto the second objection, the objection from the alethic character of self-evidence: Brentano’s partial description of self-evident judgment as judgment that announces its correctness is controversial. Brentano holds that the fundamental concept is correctness or fittingness, and we get in contact with the corresponding property by making evident judgments. His opponents hold that evident judgment puts us in contact with truth (see section 4). Hence, correctness would be derived. How can Brentano fend off the critics who hold that he has confused ‘immediately seeing the truth of a proposition’ with ‘immediate awareness of correct judgment’? As a first step, Brentano can argue that acknowledgement is a nonpropositional attitude. I will not rehearse Brentano’s reason for this view here, but outline how this notion is used in his account of self-evidence.17 The basic idea is that one can acknowledge objects without acknowledging they exist. For example, consciousness of hearing is acknowledgement of present hearing, not judging that one is hearing or that hearing is going on. When I am aware of hearing a note, I acknowledge hearing a note. Such acknowledgements are self-evident, yet they are not endorsements of propositions whose truths are obvious.18 But can’t one say that an acknowledgement of an activity is self-evident if one is directly aware of A? Yes, but it is no philosophical progress; for being directly aware of an activity is just acknowledging it with self-evidence. Now, why not say that the judgment itself, the mental act, is true and in a self-evident judgment one cannot fail to be aware of its truth? Judgings themselves cannot be true.19 I cannot judge true; I can only judge truly. But this adverbial notion comes close to ‘correct’: we use it to assess acts as worthy of being done.20 In sum, if we have reasons to think there are non-propositional acts like acknowledging an object, there is room for a notion of self-evidence that is not seeing the truth.

7.

Two Reasons for the Priority of Correctness

We know now how we acquire the concept of truth that applies to contents on the basis of a primitive sense of judgmental correctness. Brentano assumed that judgmental correctness comes first in the order of concept acquisition. Now, whether Brentano is right about this order is a matter of psychological research. Is he right about the order of conceptual explanation between correctness and truth? Is correctness prior to truth? We had better present reasons for Brentano’s positive answer because, as we have seen in section 5, it is tempting to explain judgmental correctness in terms of (propositional) truth. In general, one will ask how correctness of judgment can be independent of truth of content. If one wants to make Brentano’s primitivism about correctness and the connected view that correctness is prior to truth plausible, one needs to provide independent

Correctness First

143

reasons for Brentano’s correctness-first picture. In the next three subsections, I will give reasons that speak in favour of Brentano’s picture. 7.1.

The Pervasiveness of Correctness

Consider a contemporary account of the source of the concept of truth. Field describes a model for the first stage of acquiring the concept of truth: Let’s note one obvious fact about how the word ‘true’ is standardly learned: we learn it to apply to utterances of our own language first, and when we later learn to apply it to other languages it is by conceiving the utterances of another language more or less on the model of utterances of our own language. The obvious model of the first stage of this process is that we learn to accept instances of the schema (T) X is true if and only if p where ‘X’ is replaced by a quotation mark name of an English sentence S and ‘p’ is replaced by S. (Field 1972: 372) Now, what is it to accept an instance of the T-schema? If one does not want to go round in a very tight circle, one cannot say that accepting an instance of a T-schema is judging that the instance says something true or is holding it to be true. The notion of acceptance cannot involve or presuppose the notion of truth on pain of circularity. But it can involve and presuppose the notion of correctness: mastery of the concept of truth consists in being disposed to make self-evident judgments whose contents are the (non-paradoxical) instance of the T-schema. The proposition that snow is white is true if, and only if, it is right to judge that snow is white without further reason. All instances of the T-schema or the corresponding denominalization schema seem to us to be such that if we judge them, we are judging as one should. If we need the notion of correctness to complete an account of truth anyway, why not start with it and fashion the notion of truth out of it? 7.2.

The Evaluative Nature of Truth

The predicate ‘is true’ has a normative or evaluative dimension.21 Saying that something is true recommends believing and assenting to it. If the sentence ‘It is true that it is raining’ is uttered with assertoric force, and you grasp what I say and the force with which I say it, you have a prima facie reason to judge that it is raining. Brentano’s ‘correct judgment first’ account makes obvious why content truth is an evaluative concept. Saying that a content is true is just saying that it is correct to judge it, that it merits the attitude of judgment. So when I know that a content is true, I thereby know that the right thing to do with respect to this content is to judge it. If I am motivated to do what is right, I will judge it.

144

Mark Textor

In contrast, if we put propositional truth first, this feature of ‘is true’ is hard to accommodate. The fact that there is a true proposition does not directly bear on what I ought to believe or assent to.22 Consider, for illustration, the view that propositions are complex, mind-independent objects that are the primary bearers of truth and falsity. My knowledge that the proposition that London is a city is true does not give me a reason to judge that London is a city. It only seems so if one hears ‘is true’ as expressing the concept Brentano has in mind. The propositional truth–first theorist can respond by postulating a further norm: One ought to, if the proposition that p is true, believe that p. Why does this norm hold? Why does ascribing a property to a proposition that is independent of beliefs, etc., result in a commendation to believe the proposition? We want to know more about this property to be convinced that it has normative import. The propositional truth–first theorist can appeal to the instrumental value of believing true propositions. In general, if you believe true propositions, you are likely to act in a way that will satisfy your desires. Therefore, one ought to believe true propositions. But, as Lynch (2004: 502) points out, truth is not just an instrumental good. We care about believing the truth independently of the practical consequences of our beliefs and often even if they have no practical consequences: For example, with regard to at least some extremely abstract mathematical conjectures, knowing their truth would get us no closer to anything else we want. None the less, if we were forced to choose between believing truly or falsely about the matter, we would surely prefer the former. Even when guessing about such things, we prefer to guess correctly. (Ibid.) If I know that p is true, I have a prima facie reason to judge that p, even if I know in addition that my judgment has no practical value whatsoever. Brentano has a good explanation for this. A propositional content is true if, and only if, it is correct to judge it. Hence, if I recognize that the content that p is true, and I want to do what is right, I will judge that p. 7.3.

Truth and Existence as Logical Properties

In the literature on truth, philosophers often draw an analogy between the concept of truth and existence.23 Both are ‘logical properties’. Part of what is meant by this is that both ‘exists’ and ‘is true’ seem to be first-order

Correctness First

145

predicates, but neither of them determines the object to which they are applied any further. Asay illustrates this similarity as follows: If I ask you to imagine a kangaroo, and then to imagine an existing kangaroo, I have not asked you to do two separate things. Similarly, if I ask you to consider whether kangaroos live in Australia, and then to consider whether it’s true that kangaroos live in Australia, I have not asked you to do two separate things. Hence, just as adding existence to an idea adds nothing, so too does adding truth to a thought add nothing. (Asay 2013: 509) Brentano’s view of truth offers an explanation of the redundant/omnipresent character that both truth and existence exhibit, for the concept of existence is just the concept of truth that applies to the contents of a particular kind of judgment—namely, simple acknowledgements. For Brentano, one acknowledges an object. The concept of existence is like the concept of truth derived from correct judgment: Aristotle had already recognized that [the concept of existence] is acquired by reflection on the affirming judgment. (TE, 27 [45], my translation) Now, if it is correct to acknowledge the object, say the dog Fido, we hardly want to say that Fido is true. What we want to say is that Fido exists. So does Brentano: One may say that an affirmative judgment is true, or one may say that its object is existent; in both cases one would be saying precisely the same thing. Similarly for saying that a negative judgment is true, and saying that its object is non-existent. We may say that, for every (simple) affirmative judgment, either it or the corresponding negative judgment is true; and we may express precisely the same logical principle by saying that, for every such affirmative judgment, either its object is existent or its object is non-existent. (TE, 28 Fn [45 Fn], my translation) In short: An object x exists if, and only if, it is correct to acknowledge it. If we take some judgments to have propositional content, we can say what their truth consists in by deriving truth from correctness: A propositional content is true if, and only if, it is correct to judge it.

146

Mark Textor

Hence, the concepts of truth and existence are both explained in the same way, yet with recourse to different kinds of judgments. The strategy to put correct judgment first allows Brentano to give a unified explanation of features of truth and existence. Why does applying the concepts of truth and existence not determine the objects to which they are applied? Brentano’s answer for the concept of existence draws on a parallel between ‘exists’ and ‘good’: In calling an object good we are not giving it a material (sachliches) predicate, as we do when we call something red or round or warm or thinking. In this respect the expressions good and bad are like the expressions existent and nonexistent. In using the latter [‘existent’ and ‘non-existent’], we do not intend to add yet another to the determining characteristics of the thing in question; we wish rather to say that whoever acknowledges a certain thing and rejects another certain thing makes a true judgment. (FCE, 90 [144], my emphasis) Let us explain the idea by starting from the concept of goodness. When I utter with assertoric force ‘Champagne is good’, I don’t intend my audience to come to a judgment about champagne to the effect that champagne has a non-evaluative property shared by all things that are good: ‘good’ is not a real or determining predicate. I intend them to judge that one ought to like champagne and, ideally, to come to have such an attitude towards champagne. Similarly, when I utter with assertoric force ‘The thought that p is true’ and ‘Gravitational waves exist’, I intend them to judge that the thought that p merits the attitude of judgment and to judge that gravitational waves deserve to be acknowledged. In which sense is considering whether kangaroos live in Australia the same act as considering whether it is true that kangaroos live in Australia? When you consider whether kangaroos live in Australia, you want to make up your mind whether it is correct to judge that kangaroos live in Australia. Considering whether kangaroos live in Australia is therefore the same thing as considering whether the content that kangaroos live in Australia is to be acknowledged—that is, according to Brentano, whether it is true that kangaroos live in Australia.24 In which sense is imagining a kangaroo the same act as imagining an existing kangaroo? When you imagine a kangaroo, you imagine yourself perceiving such a thing. When you imagine perceiving such a thing, you imagine yourself being in a situation in which you ought to acknowledge a kangaroo. Hence, if Brentano is right about the concept of existence, imagining a kangaroo is imagining an existing kangaroo. According to Brentano, truth and existence are two versions of the concept of something that is correctly acknowledged. This does justice to

Correctness First

147

intuitive views of truth and existence and therefore speaks in favour of Brentano’s approach to these notions.

Conclusion I have argued that Brentano proposes a correctness-first theory of truth and explored how self-evident judgment can put us in the position of acquiring the basic primitive of this theory.25 The core of Brentano’s view of truth is an account of how we acquire the concept of correctness, which can then be used to define propositional truth and existence. Because these concepts are defined in terms of correct judgment, they have features of normative concepts. This is an advantage that makes Brentano’s basic idea attractive.26

Notes 1. See, for example, McHugh and Way (2018: 165): “‘fitting’ is here something of a term of art. We could also express the relevant judgements by saying that it is correct to admire Mandela, that one would be getting it right in admiring him, [. . .]”. 2. References to the pagination of the German texts are in square brackets. 3. For Brentano’s lists of uses of ‘true’, see OCT, 3 [5]; FCE, 137 [86]. 4. Russell changed tack and defined a derivative notion of truth for statements in terms of ‘uptake’: “a statement is true when a person who believes it believes truly, and false when a person who believes it believes falsely” (Russell 1910: 117). 5. See Wedgewood (2007: 157). 6. The focus on correctness brings Brentano close to Windelband (1882: 46), for whom a true judgment is one that “deserves to be acknowledged”. Brentano and Windelband agree that our consciousness of our own judgments reveals to us the norm of judgment: correctness. They disagree about the nature of acknowledgement: Windelband takes it to be an emotion; Brentano doesn’t (see Kastil’s comments in TE, 101). 7. The importance of the notion of self-evidence for Brentano’s account of truth is well known; see, for instance, Brandl (2017: 12), Chisholm (1986: 35), Kriegel (2018: 138f), and Montague (2017: 120). The account proposed in this section differs from these, in that I argue that self-evidence is fundamentally normative. 8. In OKRW, 14 [23], love is not only correct but also characterizes itself as correct. See also L&H, 146 [151–52], where feelings of love and hate are said to ‘disclose’ their correctness (als richtig verraten). (The translation ‘love that is experienced as correct’ sounds incorrect to me.) 9. Ibid.: 362. 10. See also Wedgewood (2002: 267). 11. This idea undoubtedly needs more unpacking. Ginsborg (2015: 80ff.) does a good deal of it. 12. Do we get an infinite regress is awareness is itself a case of evident acknowledgement, that is do we acknowledge the correctness of evident acknowledgement in a further evident acknowledgement whose correctness is acknowledged in third evident acknowledgement? No, we have an example where a judging is directed upon itself as well as a distinct object. Brentano (PES, 98 [I, 180]) has

148

13. 14. 15.

16. 17. 18. 19. 20. 21. 22.

23. 24. 25.

26.

Mark Textor given independent reasons to explain why every judging has several objects. Self-evident judgments are a special case, as they acknowledge, among other things, not only themselves but also their correctness. Künne (2003: 377) takes Brentano to endorse a biconditional that specifies the extension of true judgment. Our knowledge of what truth is cannot consist in knowledge of the biconditional for reasons given above. It is rehearsed by, for instance, Chisholm (1986: 39), Parsons (2004: 190), and Kriegel (2018: 114). This will remind contemporary readers of deflationary theories of truth. For instance, Parsons commented: “from our own perspective, we might summarize what Brentano says as that someone who judges that p judges truly if and only if p” (Parsons 2004: 187). See §13 of Kraus’s introduction to TE. See Textor (2017: 5–8). On awareness as self-evident judgment, see, for instance, FCE 88 [141]. See Bar-Hillel (1973: 304). See Aune (1967: 222). See Dummett (1959: 2ff). The topic is further developed in Lynch (2004) and Wright (1992: 16ff). According to Wright, ‘true’ and ‘warrants assertion’ coincide in normative force. See Aune (1967: 222). Aune comes very close to Brentano when he characterizes the basic function of truth “as assessing acts as worthy of being done”. In a footnote, he comments that, strictly speaking, ‘is true’ applies to nominal objects of such acts. The nominal objects are, of course, not worthy of being done; but they are worthy of the acts under consideration. See Field (1992: 377). See Asay (2013: 509). In this paper, I was primarily interested in developing Brentano’s account of the acquisition of the concept of correctness and truth. I leave, therefore, discussion of alternative proposals that see him as proposing a definition of truth for another occasion. I presented forerunners of this paper at the NCH workshop Judgment: Act and Object in London in 2016 and at a seminar at the University of Bielefeld in 2017. I am grateful to both audiences for feedback. Special thanks to Johannes Brandl for comments on a first draft of the paper.

References Asay, J. (2013), “Primitive Truth”, Dialectica 67: 503–19. Aune, B. (1967), “Statements and Propositions”, Noûs 1: 215–29. Bar-Hillel, Y. (1973), “Primary Truth Bearers”, Dialectica 27: 303–12. Brandl, J. (2017), “Was Brentano an Early Deflationist About Truth?” The Monist 100: 1–14. Brentano, F. FCE = The Foundation and Construction of Ethics, trans. E. H. Schneewind. London: Routledge, repr. 2009. References in brackets to Grundlegung und Aufbau der Ethik. Mayer-Hillebrand, F. (ed.). Repr. Hamburg: Meiner. Brentano, F. LRU = Die Lehre vom Richtigen Urteil, ed. F. Mayer-Hillebrand. Bern: Franke 1956. Brentano, F. OCT = “On the Concept of Truth”. Repr. in TE. Brentano, F. OKRW = The Origin of Our Knowledge of Right and Wrong, trans. R. Chisholm and E. H. Schneewind. London: Routledge Revivals 2009. Page

Correctness First

149

references in brackets to Vom Ursprung sittlicher Erkenntnis, ed. O. Kraus. Repr. Hamburg: Meiner, 1969. Brentano, F. PES = Psychology from an Empirical Standpoint, 2nd ed. London: Routledge, 1995. Page references in brackets to Psychologie vom Empirischen Standpunkt, 2nd ed. 1924, ed. A. Kraus. Repr.t Hamburg: Meiner, 1971. Brentano, F. TE = The True and Evident, ed. R. Chisholm, trans. R. Chisholm, I. Politzer and K. Fischer. London: Taylor & Francis e-Library 2009, 2–18. Page references in brackets to Wahrheit und Evidenz (WE), ed. O. Kraus. Repr. Hamburg: Meiner, 1974. Brentano, F. V = “Von der Natur der Vorstellungen (Diktat 1903)”, Conceptus 1987: 25–31. Brentano, F. VE = Versuch über die Erkenntnis. Leipzig: Meiner, 1903. Chisholm, R. (1986), Brentano on Intrinsic Value. Cambridge: Cambridge University Press. Dummett, M. (1959), “Truth”, Proceedings of the Aristotelian Society 59: 141–62. Reprinted in his Truth and Other Enigmas. London: Duckworth, 1978. Field, H. (1972), “Tarski’s Theory of Truth”, The Journal of Philosophy 69: 347–75. Field, H. (1992), “Critical Notice: Paul Horwich’s Truth”, Philosophy of Science 59: 321–30. Frege, G. (1923–26), “Gedankengefüge”, Beiträge zur Philosophie des Deutschen Idealismus 3: 36–51. Translated R. H. Stoothoff as “Compound Thoughts”, Mind 72(1963): 1–17. Ginsborg, H. (2006), “Empirical Concepts and the Contents of Experience”, European Journal of Philosophy 14: 349–72. Ginsborg, H. (2015), The Normativity of Nature. Oxford: Oxford University Press. Kriegel, U. (2018), Brentano’s Philosophical System. Oxford: Oxford University Press. Künne, W. (2003), Conceptions of Truth. Oxford: Oxford University Press. Lynch, M. (2004), “Minimalism and the Value of Truth”, The Philosophical Quarterly 54: 497–514. McHugh, C. and J. Way (2018), “What Is Good Reasoning?”, Philosophy and Phenomenological Research 96: 153–74. Montague, M. (2017), “Brentano on Emotion and the Will”, in U. Kriegel (ed.), Routledge Handbook of Brentano and the Brentano School. London: Routledge. Moore, G. E. (1903), “Review of F. Brentano: The Origin of the Knowledge of Right and Wrong”, International Journal of Ethics 14: 115–23. Olson, J. (2017), “Two Kinds of Ethical Intuitionism: Brentano’s and Reid’s”, The Monist 100: 106–19. Parsons, C. (2004), “Brentano on Judgement and Truth”, in D. Jacquette (ed.), The Cambridge Companion to Brentano. Cambridge: Cambridge University Press, 168–97. Plantinga, A. (1993), Warrant and Proper Function. New York: Oxford University Press. Russell, B. (1910), “On the Nature of Truth and Falsehood”, in his Logical and Philosophical Papers 1909–13, ed. Slater. London: Routledge, 1992.

150

Mark Textor

Schlick, M. (1910), “Das Wesen der Wahrheit nach der modernen Logik”, Vierteljahresschrift für wissenschaftliche Philosophie und Soziologie 34, 386–477. Reprint in his Philosophical Papers I (1909–22). Dordrecht: Textor, M. (2017), “Towards a Neo-Brentanian Theory of Existence”, Philosophers’ Imprint 17: 1–20. Wedgewood, R. (2002), “The Aim of Belief”, Philosophical Perspectives 16: Language and Mind: 267–97. Wedgewood, R. (2007), The Nature of Normativity. Oxford: Oxford University Press. Windelband, W. (1882), “Was ist Philosophie?”, in his Präludien: Aufsätze und Reden zur Einleitung in die Philosophie. Tübingen: Mohr, 1907. Wright, C. (1992), Truth and Objectivity. Cambridge, MA: Harvard University Press.

7

Judgment, Reasons, and Feelings Simon Blackburn

1.

The Field of Reason

It is possible to hold that reasons are the philosopher’s proprietary field, and that since there can be no reasons for feelings, philosophers interested in judgments, whether we are thinking about people, or in morals or aesthetics, need not be concerned with feelings at all. In a statement that at first seems to sound the thirteenth strike of the clock for his brand of rationalism, Derek Parfit suggests such a view, asserting roundly that “we never have reasons to enjoy, or be moved by, great music”.1 We might initially think that this is because in the bloodless atmosphere of rationalism, reasons beam only on our cognitive states. But that cannot be the explanation, because Parfit does suppose that we have reasons to want some things and to try to make some things happen.2 So when x is a reason for y, y need not be a belief or other purely cognitive change, such as a shift in a probability assignment. We can have reasons for wants, intentions, and decisions as well. One might also hazard that it is the generally involuntary nature of enjoyments and emotions that removes them from the sphere of reasons. But that, too, cannot be the problem since Parfit thinks that often enough our responses to reasons are themselves involuntary.3 So involuntary movements of reason can lead to some non-cognitive states, leaving it unclear why the embargo covers the states of mind it does and how far it extends. It seems to be what Parfit calls hedonic likings and dislikings that are the main culprits—the things for which there cannot be reasons. Thus, he compares aesthetic reactions to brute sensations: the squeaking of chalk or the taste of milk chocolate. And “whether we like, dislike, or are indifferent to these various sensations, we are not responding or failing to respond to any reasons”.4 We may have reasons to want to enjoy some things, but we cannot have reasons for being in one hedonic state or another, either enjoying or hating, as the case may be. Perhaps Parfit’s point hinges on the way in which sensations are outside the field of reasons. There is then some justice in his claim. There are things that cause sensations, of course, but these are not usually reasons for a sensation. Sensations such as twinges and aches just happen, and

152

Simon Blackburn

there is no room for pride or shame, discursive attention or criticism, about the way in which they happen. A person may be unfortunate in the way in which, say, damp weather sets off their rheumatism, but it is not a fault of character. It is more controversial to suggest, as Parfit does in the quotation above, that the sensations themselves do not give us reasons for anything—that, as well as being improper values for y in the schema x is a reason for y, they are improper values for x. One might have thought that the migraine gives its sufferer a reason to cry off going to the party or to lie still in the dark or to try remedies, for instance. But the more serious question is whether it is reasonable to assimilate aesthetic and moral feelings to sensations in the first place. Parfit evidently sees no difficulty here. At any rate, he frequently and cheerfully asserts that moral attitudes, as they are conceived in the anti-rationalist, sentimentalist, or expressivist tradition that assimilates them to ways in which we feel about things, would then be like headaches, which are paradigm sensations.5 If I can never have a reason for how I feel about things, then as well as having no reason to enjoy great music, for instance, I have no reason to feel gratitude for your good offices, to sympathize with your distress, or to feel elated at your joy—strange and eerie, if not downright sinister, absences in a treatise on ethics. Are there really no reasons for feelings? The mother, one might have thought, has every reason to enjoy her child’s health and exuberance and every reason to feel anxiety if they falter. If I ask why you are distressed and you tell me you have failed in something you had hoped to do, surely you have given me the reason for your distress—and a readily understandable one (similarly, if you feel elated because you have succeeded with flying colours). And if you are my friend, your distress or elation gives me, in turn, reason for sympathetic feelings of the same kind, so feelings are in both the domain and the range of the relation x is a reason for y. Even Kantians, surely, agree that if you freely help me in some enterprise, I have a reason to feel gratitude to you, and they tend, one gathers, to think that your rational agency is a reason for anyone to respect you. But if as a creature of flesh and blood I often enough do have reasons to feel these ways, it is hard to see why I cannot have reason to enjoy other things, including great music. The sublime expression of the transitions between despair and joy in the third movement of Beethoven’s Op. 132 string quartet invites thoughts of grief and dejection on the one hand and deliverance and joy on the other.6 These are my reasons for enjoying the work, although enjoyment is, of course, a shorthand for a whole complex of responses: admiration, gratitude, the feeling of having a resource for facing the world and its trials, a sense of peace or completeness, and many more. The work has an existential import. And, in turn, one’s enjoyment gives reason to feel grateful to Beethoven and glad that the world can produce such beauty—two more feelings, the first of which, as I have already suggested, is of more than a little ethical importance.

Judgment, Reasons, and Feelings

153

The trains of thought and feeling that a work of art, or a landscape, might engender may be pleasant or unpleasant. The elation at success is pleasant, and the dejection at failure is unpleasant, as sensations are. But, of course, such feelings are not just sensations. They involve our awareness of things outside ourselves, whereas sensations do not. In so doing, they invite a kind of thought that sensations do not. There is then scope for conversation and persuasion, and even shame and regret at either having missed the point or having responded crassly or inappropriately. The revaluation may involve a moral dimension: perhaps the thing you failed at was not worth attempting and you ought to get over it. Perhaps you are elated at your friend’s failure and ought not to be. In his extensive investigation into psychological taxonomy, Franz Brentano argues persuasively that there is a continuum between feeling and willing, and that both are to be identified in terms of one underlying state of affection, which he calls love (Liebe).7 He considers the following series, starting with a pure feeling and ending with a determination of the will: sadness, yearning for absent good, hope that it will be ours, the desire to bring it about, the courage to make the attempt, the decision to act. He says, plausibly, that there is no hiatus in this series; each element follows naturally from its predecessor: But is there not already a germ of the striving lying unnoticed in the yearning, which germinates when one hopes, and blooms when one thinks of the possibility of doing something oneself, when one wishes to act and then has the courage to do so, until finally the desire overcomes both the aversion to any sacrifice involved and the wish to reflect any longer, and it ripens into a decision?8 In this metaphor, feelings are always situated on the love/hate continuum and always contain the seeds of hope (for something to come about or continue/not come about or disappear) and desire.9 If they were not themselves subject to assessment, the decisions into which they naturally flower would lose most, if not all, of their susceptibility to the same criticisms. If the mother’s anxiety is reasonable, then so is her appropriate method of dealing with it, such as hurrying to the doctor. An unreasonable mother may feel anxiety when she need not and waste the doctor’s time. If her anxiety were immune to description as based on reasons (or not), then so would be the pestering of the doctor that is a natural and appropriate expression of it—and similarly for human decisions and actions in general, for the unity of feeling and will that Brentano champions means that they typically stand or fall together. And it is because we know of their power to incubate decisions and actions that we say that feelings can be, and often need to be, cherished or suppressed, neither of which applies to sensations. Brentano ended his series with a decision, but there are parallel series that end not with decisions but actions. In Robert Burns’s poem,

154

Simon Blackburn

Tam O’Shanter’s angry wife “nursing her wrath to keep it warm” does not need to decide to berate Tam when he gets home. She has already decided to do so, and the action is going to follow. Feelings about things and people and objects involve an orientation, and the orientation may be the result of care and attention, may be thought through, and may be the fruit of experience and comparison, or it may be careless and casual or inauthentic. None of these are terms that apply to sensations. All of them suggest the possibility of critical refinement and change. They can be compared with the affective responses and attunements of other people or those of one’s own earlier self. Like Brentano, Gilbert Ryle is particularly good on feelings, of which he distinguishes many families. In particular, he points out that no special sensation is had when I feel that there is a flaw in your argument. The questions “Where do you feel it? Would an aspirin allay it? Have you got it now? Does it come and go, or is it there all the time?” are not pertinent questions. My feeling that there is a flaw in your argument can be strong, but not intense or acute, continuous or intermittent. I am in no distress.10 The same is true when I feel that someone is creepy or that there is something sentimental about a Renoir painting. The feeling, Ryle suggests, is not yet a judgment. It has me leaning towards a judgment or tempted to make a judgment, but “the door is still ajar”. But just as you can ask me for my reasons for thinking that someone is creepy, you can also ask me for my reasons for feeling that he is creepy: why am I even leaning that way? Parfit does allow that I may have reasons to want to feel various ways, but this is surely a quite different condition to be in. Someone distressed by a child crying has reason for being distressed, but a cold-hearted and indifferent bystander is not likely to have reason to want to be distressed, although in some strange circumstances he might. I could only be said to have reasons to want to enjoy something if I do not currently enjoy it but think it would be somehow better if I could do so. Op. 132 might be my only disc on the desert island, and if I do not enjoy it, I shall have little else to enjoy, so I want to enjoy it, and if I find it difficult, I might persistently try to do so. But for me as I actually am, while I do think there are reasons for anyone to want to enjoy it, this of course derives from my own actual enjoyment, which I think it would be good for others to share, not the other way round. If I had not heard it and enjoyed it, there would be something off-colour or even dishonest in assuring other people that they should want to do so. Since the evidence seems so heavily stacked against Parfit’s embargo, here is a different way of looking at the matter. There are various ways of distinguishing reasons from brute causes. Reasons, it is widely supposed,

Judgment, Reasons, and Feelings

155

bring in an aspect of normativity: when we think of a feature to which we respond as a reason for our response, and not just a cause, we suppose that there is something right or justifiable about our response. We are prepared to acknowledge or stand by the way our minds were moved, and this acknowledgement can vary from simply being satisfied with the way we responded to being disappointed or even horrified by anyone who does not share the response. In a paper rightly insisting that expressivists must give some account of this, Allan Gibbard gives an example where this was not so.11 He confesses that in the turbulent 1960s, he found himself disgusted by young men with long hair, yet also worried that he had no reason for this reaction. His problem was that he could not single out any quality of which this long hair might be a signal and which he would be prepared to defend as a cause of his discomfort. And it is precisely this that marks his reaction as caused but not, in his view, having any reason (he might equally say having no good reason). Now, those of us who are used to the neutral import of long hair on young men will agree that he had no reason—or no good reason—for feeling as he did. All this applies in just the same way whether our minds are moved by apprehension of some aspect of things to a belief or a shift of confidence or a desire, intention, or feeling. Like most of the human race, I find the nearby crying of a child distressing. But I think that, as well as causing me to feel this way, the crying gives me a reason for this feeling, and this means I am prepared to defend the view that this is an appropriate way to feel, and I feel rapport with those who do feel the same and even look with some degree of aversion on anyone in whom such feelings are absent. As is often remarked, phenomenogically, the perception and the sense of the fitting and appropriate nature of the response may come in one packet. It is as if one heard not merely the baby crying but also the gerundively tinged fact of its being necessary to do something about it. It calls for alleviation, and if there is nobody else to help, one might well be miserable if one does not know what to do or cannot do anything about it. Brentano voiced the same idea: As we prefer insight to error, so also, generally speaking, we prefer joy (unless indeed it be joy in what is bad) to sadness. Were there beings among whom the reverse held good, we should regard such conduct as perverse, and rightly so. Here too it is because our love and our hatred are qualified as right.12 In this late essay, Brentano sometimes seems to couch his discussion in ‘realist’ terms, as if the rightness of a response were itself a further feature of the object to which we are responding, apprehended as its

156

Simon Blackburn

other features are. But earlier, he had been absolutely clear that this is not intended. Phenomena of feeling and will are not judgments: I do not believe that anyone will understand me to mean that phenomena belonging to this class are cognitive acts by which we perceive the goodness or badness, value or disvalue of certain objects. Still, in order to make such an interpretation absolutely impossible, I explicitly note that this would be a complete misunderstanding of my real meaning. In the first place, that would mean that I viewed these phenomena as judgments; but in fact I separate them off as a separate class. Secondly, it would mean that I would be assuming quite generally that this class of phenomena presupposes presentations of good and bad, value and disvalue. This is so far from being the case, that instead I shall show that such presentations can stem only from inner perceptions of these phenomena.13 He later describes the inner perceptions as modes of intentional reception of objects. It is the intentional reception of an object and a reception that is itself infused with love. There is no hint that the love is itself a response to any new and sui generis feature of the object of our feeling—a mysterious addition to the features that in our minds make it lovable. Should the inarticulate nature of many of our enjoyments—our inability to put our finger on why we feel as we do—impugn their classification as formed for reasons? Here is Gilbert Ryle again: It is no objection to this that sometimes we do not know the features of a work that have moved us. It is not even always requisite or desirable that a person should try to move on from feeling that something is the case to producing an articulate case for it being so. It is the business of the judge, but it is not the business of the members of the jury to give reasons for a verdict. Their business is to give the right verdict, not to satisfy a Court of Appeal that it was the right verdict to give. Nor, necessarily, should the general prepare himself to satisfy critics that he chose the right moment for the counter-attack. He might be a good general but a bad lecturer at the Staff College; there is no contradiction in saying that he had good reasons for ordering the counter-attack at that moment, though he could not then or subsequently tell himself or anyone else what they were. We are all pretty much in his position in the business of sizing up people’s characters. Our goodness or badness at making such estimates need not and does not vary concomitantly with our goodness or badness at producing justifications for them. We should contrast cases where our minds are simply caused to move in some direction with cases where they are guided by some consideration

Judgment, Reasons, and Feelings

157

towards some change. The paradigm cases of guidance would be ones where we are consciously in control of the movement: we appreciate the fact about the situation and, weighing it together with whatever other circumstances come to mind, make a decision. We should not identify the contrast between mere causes and reasons with that between mere causes and guides since we should allow a category of subliminal or subdoxastic forces that may or may not deserve to be called reasons. We should also avoid thinking that guidance involved an essentially different kind of causation, such as one working in the medium of consciousness. There are cases in which by exciting people’s imagination of someone watching them, an agent makes them behave better in a situation in which previously they behaved badly. If we ask whether this should be regarded as a case of brute causation or of being guided by a reason we may well find the answer indeterminate, although I think we should allow the more generous description. After all, an officer might excite his men’s imaginative awareness of the dangers they face in the hope that the reasonable ones amongst them will take more precautions. I suggest that ‘guidance’ has exactly the same normative slant as ‘reason’. In other words, we describe ourselves or others as guided in just the same way that we describe ourselves or others as responsive to a reason. Each vocabulary signals the same content with the kind of movement in question and potentially our distress or discomfort at those who remain unmoved. And this applies whether a belief, want, intention, or feeling is the endpoint of the movement. Ryle does not talk of enjoyment, but he does say something highly pertinent: On the other hand, there is an important connection between, for example, feeling tired and feeling like sitting down, or between feeling indignant and feeling like writing a protesting letter to the Times. That he should, from time to time, feel like doing certain sorts of concrete things is one of the things that we expect of a person who is in this or that mood or general condition. He is in the mood, inter alia, to have such ideas not only occur to him but occur seductively. We expect the angry person not only to talk gruffly, scowl, and slam the door, but also to entertain fancies of doing all sorts of hostile things—most of which, of course, he cannot do and would not permit himself to do, even if he could. Like being angry at something, enjoying something, we might suggest, is being in a “mood or general condition”. Its manifestations can include wishing it to continue, hoping to hear it again, feeling better for having encountered it, being disposed to draw it to your friend’s attention, and so on. It is a sustaining state, and one would wish it to be sustained. And, like finding someone creepy, we can ask for its reasons. Why on earth

158

Simon Blackburn

would anyone enjoy the work of Philip Glass or Damien Hirst, one might innocently ask, and there may be more or less convincing replies. I would hope nobody would ask the same question, with the same surprise, about Op. 132. So, however it may be in the desiccated world of rationalism, down on earth there are often reasons to enjoy or be otherwise moved by things. It is, incidentally, a complete mistake to think that anti-rationalists such as Hume deny this. Hume is rightly comfortable with reasons and reasonings that speak to our aesthetic feelings: “in many orders of beauty, particularly those of the finer arts, it is requisite to employ much reasoning, in order to feel the proper sentiment; and a false relish may frequently be corrected by argument and reflection”.14

2.

Fellow-Feeling and Risk

In the first section, I argued that there is no reason to exile feelings from the realm of the normative, signalled by allowing that there are reasons for feelings and ‘general conditions’ in this sphere as much as in any other. But I gave no positive explanation of how normativity got in on the act. Brentano was right about the phenomenology, whereby some reactions present themselves as required by what we are looking at or listening to. But the question is why we feel this way. Why subject our own reactions and those of others to the stern eyes of objectivity, improvement, truth, or reason? For someone thinking like Parfit, it seems that what Kant says about the agreeable could just as well be applied across the whole domain of feeling: As regards the agreeable everyone concedes that his judgment, which he bases on a private feeling, and in which he declares that an object pleases him, is restricted merely to himself personally . . . this applies not only to the taste of the tongue, the palate, and the throat, but to what may with anyone be agreeable to eye or ear. . . . With the agreeable, therefore, the axiom holds good: Everyone has his own taste (that of sense).15 This echoes Parfit’s view about straightforward sensations, but it is not plain that Kant is right about it, for even with the simple agreeable tastes of “the tongue, the palate, and the throat”, people do try to persuade and nudge others to share their enjoyment. People can relish the wrong things. You might think they are missing something—valuable enjoyments are escaping them. You might also worry that sticking with some crude taste is, as it were, lazy or slovenly, a wilful indifference to the pleasures that cultivation brings. It may even suggest a moral flaw, such as a cynical attitude towards ‘so-called experts’, born of ignorance and complacency. In other cases, insouciance about what might be found agreeable “to eye

Judgment, Reasons, and Feelings

159

or ear” becomes absurd: imagine that your job takes you to the scene of a car crash and within earshot of the cries of the injured, and your colleagues, hugging themselves with pleasure and lit up by enjoyment, remark with evident relish on the agreeable sensations and feelings the scene gives them. Unless there is some extraordinary excuse, their lack of any sense of the horror or disgust of the scene makes them suddenly alien, and one draws away baffled or even frightened. I think the first natural place to look for an explanation of our normative tendencies, even in areas where morality is not involved, is our nature as social animals. Finding that our feelings about things chime in with those of our peers is both pleasurable in itself and cementing of our social bonds. Finding that our feelings do not harmonize gives us a jolt, fractures a little bit of the relationships we have and that we need to have. This needs careful handling and some qualification, for sometimes, indeed, it need not matter. Kant was right that in some matters, there are what are usually described as no-fault disagreements, even if he was wrong about their extent. However, the terminology of ‘no-fault disagreements’ is not entirely apt. First of all, the question is not usually one of fault. It is a question of whether, by having or failing to have a certain feeling about things, a person might be manifesting a flaw or incapacity or disability rather than a fault. It need not be his fault that he has a poor palate, an uneducated eye, or the lack of experience that issues in wayward reactions to things. But these are flaws, nevertheless. One would not wish them on oneself or anyone one loves. In other cases, even talk of a flaw is inappropriate. I cannot appreciate Russian poetry since I do not speak Russian, and my appreciation of Indian ragas is thin and halting because I have very little experience of listening to ragas. Here, it is not so much a question of a flaw—still less of a fault—but of an incapacity: something that I am currently unable to do. So we might now ask whether there are any ‘no incapacity’ or ‘no disappointment’ disagreements. But then the point arises that once we are convinced that the diversity is one in which neither side thinks the other is manifesting a flaw or disability, then disagreement also flies out of the window. There is simply diversity. Hume’s own example serves as well as any: A young man, whose passions are warm, will be more sensibly touched with amorous and tender images, than a man more advanced in years, who takes pleasure in wise, philosophical reflections concerning the conduct of life and moderation of the passions. At twenty, Ovid may be the favourite author; Horace at forty; and perhaps Tacitus at fifty. Vainly would we, in such cases, endeavour to enter into the sentiments of others, and divest ourselves of those propensities, which are natural to us. We choose our favourite author as we do our friend, from a conformity of humour and disposition. Mirth or passion, sentiment

160

Simon Blackburn or reflection; whichever of these most predominates in our temper, it gives us a peculiar sympathy with the writer who resembles us.16

But Hume goes on to say that it is an error in a critic to make anything of this. The young man has no reason to criticize the older man, nor vice versa. It should not license disagreement, any more than there is disagreement when I describe someone as my rival and you say that he is not yours. The person who admits to liking an author need not go on to insist that it is right to like that author or compulsory to do so, nor be disappointed in someone who does not. But he may hope you agree that it is at least permissible to like that author. His taste, he hopes, does not deserve mockery or contempt. So in some cases, then, we may agree to live and let live. But in others, we cannot let live without, silently perhaps, thinking there is a flaw or incapacity or disability in the other person. Rather than simply thinking the boundary lies between judgments of what is agreeable and what is supposed to be beautiful, which is where Kant drew the distinction, I suggest that a more nuanced account is needed, and ideally such an account should cast some light on the whole phenomenon of feeling that a response is right or called for or even compulsory. We might approach this by asking why it is potentially such a source of pleasure to find that another’s response chimes with ours, such that we have noticed the same things and taken the same delight (or lack of delight) in them. The phenomenon is widespread enough. Hoping to enjoy a picnic, we hope as well that the others enjoy it as much and in the same way, and just one person not doing so can spoil it for all. Any lack of harmony in how we feel about things can begin to chill a friendship. It may be just a common fact about human nature, analogous to a liking for sugar or an aversion to vinegar. There might at best be an evolutionary just-so story, finding adaptive advantages in psychological identities shared across a tribe. Such an evolutionary story would have to stay in the background, for as Adam Smith noticed, no idea of the advantages of fellow feeling plays any role in our responses. It is not part of the phenomenology or overt psychology: Man, they say, conscious of his own weakness, and of the need which he has for the assistance of others, rejoices whenever he observes that they adopt his own passions, because he is then assured of that assistance; and grieves whenever he observes the contrary, because he is then assured of their opposition. But both the pleasure and the pain are always felt so instantaneously, and often upon such frivolous occasions that it seems evident that neither of them can be derived from any such self-interested consideration. A man is mortified when, after having endeavoured to divert the company, he looks round and sees that nobody laughs at his jests but himself. On the contrary,

Judgment, Reasons, and Feelings

161

the mirth of the company is highly agreeable to him, and he regards this correspondence of their sentiments with his own as the greatest applause.17 Smith also makes an extremely important point about the inception and decay of at least some kinds of aesthetic pleasure, pointing out that when we are no longer amused by a book, we can still take pleasure in reading it to the right companion: To him it has all the grace of novelty; we enter into the surprise and admiration which it naturally excites in him, but which it is no longer capable of exciting in us; we consider all the ideas which it presents rather in the light to which they appear to him, than in that which they appear to ourselves, and we are amused by sympathy with his amusement which thus enlivens our own. On the contrary, we should be vexed if he did not seem to be entertained with it, and we could no longer take any pleasure in reading it to him.18 Smith here recognizes that although, as he often stresses, we judge of other people’s sentiments by their concordance with our own, it is not always our own of the present moment that is the measure, but it can be our own as it had been when we were in a different, and more appropriate, situation to enjoy the potential pleasure. This is parallel to the way in which one might be certain that something or someone is beautiful, although one is not currently in a position to see them, or one might be sure a wine is excellent, although currently, having a cold, one cannot taste it properly. Of course, in Smith’s moral philosophy, this is a step towards recognizing the position of the “impartial spectator” whose sympathy with, or divergence from, our actual responses gives us a public standard. This desire for what Richard Rorty called solidarity may be part of the story of normativity in matters of feeling. But there is a gap between finding that our response harmonizes with that of others and being sure we are on the right track, responding with discrimination to those features of a work that deserve to engender enjoyment; for we can worry that we and our peers are all wrong in some way, taken in by something superficial or sentimental or ultimately undeserving of the praise we are offering it, or, on the contrary, we may fear that we are all missing something, and that the artist has mysteriously beaten us to it, if only we could see or listen aright. The only cure for this worry is, of course, to look or listen harder, to invoke comparisons with other works, to ask more people who have track records of discrimination and knowledge, and so on. It is here that risk comes in—the risk that one has simply not seen things the right way. Even Rorty allowed what he called a “cautionary” use of truth, one that is on show when we think or say that we may after all be wrong in thinking as we now do.19 And the same is true of feeling.

162

Simon Blackburn

Here, we may suppose, the difficulties of articulation that we described in section 1 play an important role. When the distress of a nearby child causes, and gives the reason for, my own distress, there is little more to say. We know both the input and the output, as it were. But with more delicate appreciation, neither is so readily identifiable. As we have seen, both the features to which we have responded and the precise nature of the response can be extremely hard to articulate. A fortiori, the extent to which those features are found in other situations or other works, to which we may be prone to respond quite differently, is also hard to articulate. All this implies a vulnerability and a fallibility. It is not only that I may not harmonize with other people but also, and perhaps even more worryingly, that I may not harmonize with my own future self or even my own possible self that has taken more in or otherwise expanded my current sensibilities in a way that would lead me to regard my present self somewhat askance. The priority of first-hand experience in making a judgment of one’s own is by no means confined to aesthetics. In Ryle’s case, I may be told that the general attacked at the right time, and I may believe my informant. But I cannot judge that the general did so except by acquaintance with the exact dispositions, movements, threats, opportunities, and probabilities that guided him, and as Ryle explained, even he may not be able to put his tongue to them all. I cannot share my companion’s feeling that some third party is creepy if I have no acquaintance with the demeanour and combinations of behaviour that might elicit the response. I cannot judge that a football match was exciting unless I have seen it, either in the flesh or via some medium, although I can believe that it must have been exciting because all the pundits say that it was. This is because our sensory awarenesses and, equally, our responses are endlessly complex. They have a fullness that outstrips the words we find to articulate them. If there is more to extract from a list of descriptions of things, it must be by way of inference from what the descriptions tell us. But with a full perceptual experience, we can do more than infer from what we have already noticed. We can come to notice new aspects of things. With new awarenesses come new responses and adjustments to old ones. Here, learning and profiting from the taste and experience of someone who knows what they are talking about are not confined to inferring more from settled descriptions. Nevertheless, with a companion, we can often be certain enough that we have experienced the same things in the same ways and have the same feelings about them. The general may be sure that his second-incommand saw the opportunities, dangers, and chances he did, if his attention seemed fixed on the same things and his responses are similar enough. Having been to many galleries with a friend, I can be confident that we see the same things the same ways, and nothing cements a friendship more. On the other hand, the dissonance of finding that we do not can

Judgment, Reasons, and Feelings

163

be uncomfortable, unsettling, and even, as we have seen, disappointing. We cannot get along with too much dissonance, and it is a short step from that to casting a critical eye on the dissident: it is the soil in which normativity grows. Alongside Brentano’s series, we might put another: diversity, discomfort, disappointment, dislike, disapproval. Or, if we believe ourselves to be in the presence of someone from whom we can learn: diversity, puzzlement, self-doubt, admiration, discipleship. We do not always traverse these series, but it is often remarkably easy to do so. Many of what I have called the gerundively tinged words of appreciation and appraisal invite us to do so, although temperaments, of course, vary in this and so much else. Some are inclined to live and let live, tiptoeing past opportunities to be censorious; others rather relish them. Some are more ready to learn than others. In any case, Brentano’s insight is fundamental. The role of feelings in incubating actions is enough to put us in the general arena of normativity. It matters to us how people feel because it matters to us how people behave.

Notes 1. Parfit (2011: vol. 1: 53). 2. Parfit (2011: vol. 1: 47). 3. Parfit (2011: vol. 1: 49), where it is insisted that our responses to epistemic reasons, and to reasons for desires and preferences, must alike be, in general, involuntary. 4. Parfit (2011: vol. 1: 53). 5. Parfit (2011: vol. 2: 393, 399, 477), etc. Feelings as a topic make no reappearance in On What Matters. They are not present at all in Volume I. One might be reminded of William James’s accusation that rationalism “seems too buttoned-up and white-chokered and clean-shaven a thing to speak for the vast slow-breathing unconscious Kosmos with its dread abysses and its unknown tides” (James 1912: 277–8). 6. Initially entitled the ‘Heilige Dankgesang eines Genesenen an die Gottheit’: the holy thanksgiving of a convalescent to God. 7. Brentano (1973 [1874]: ch. 8), “Feeling and Will United”. Unlike the English term ‘love’, the German word Liebe occurs in many compounds having to do with preference, like, and dislike in general. 8. Brentano (1973 [1874]: 237). 9. Instead of distinguishing love and hate, Brentano uses the one category. Hate or aversion to a thing is the love of its absence or disappearance or change. 10. Ryle (1950: 196). 11. Gibbard (2015). 12. Brentano (1902: §27). This late lecture is confusing, and it is difficult to clear it from the charge of being itself confused. 13. Brentano (1973 [1874]: 239). 14. Hume, An Enquiry Concerning the Principles of Morals, Section 1, §9, SelbyBigge, p. 173. 15. Immanuel Kant, The Critique of Judgment, ed. James Meredith, Oxford: Oxford University Press 1986. Hereafter CJ. pp. 51–2. I have slightly corrected the spelling. 16. Hume, “Of the Standard of Taste”.

164

Simon Blackburn

17. Smith (1976: 13–14 (I.i.2.2)). 18. Smith (1976: 14). 19. Rorty (1991: 128).

References Blackburn, S. (2010), “The Majesty of Reason”, in Practical Tortoise Raising. Oxford: Oxford University Press. Brentano, F. (1902), Of Our Knowledge of the Origin of Right and Wrong, trans. C. Hague. London: Archibald Constant and Co. Brentano, F. (1973 [1874]), Psychology From an Empirical Standpoint, ed. O. Kraus. London: Routledge. Gibbard, A. (2015), “Improving Sensibilities”, in R. N. Johnson and M. Smith (eds.), Passions and Projections. Oxford: Oxford University Press. James, W. (1912), Essays in Radical Empiricism. New York: Longmans Green and Co. Kant, I. (1986), The Critique of Judgment, ed. J. C. Meredith. Oxford: Oxford University Press. Parfit, D. (2011), On What Matters. Oxford: Oxford University Press. Rorty, R. (1991), “Pragmatism, Davidson, and Truth”, in Objectivity, Relativism, and Truth: Philosophical Papers, Vol. 1. Cambridge: Cambridge University Press. Ryle, G. (1950), “Feelings”, The Philosophical Quarterly 1: 193–205. Smith, A. (1976), The Theory of Moral Sentiments, ed. D. D. Raphael and A. L. Macfie. Oxford: Oxford University Press.

8

Twardowski on Judgment Peter Simons

Ale wasza mowa niech będzie: Tak, tak; nie, nie; a co nadto więcej, jest od złego. But let your communications be: Yea, yea; Nay, nay: for whatsoever is more than these cometh of evil. Matthew 5:371

1.

Context

As the father of modern Polish philosophy, Kazimierz Twardowski (1866– 1938) holds an important place in the recent history of philosophy in Europe. His ideas influenced his contemporaries, such as Meinong and Husserl, and two generations of students, such as Łukasiewicz, Leśniewski, Ajdukiewicz, Kotarbiński, and Ingarden, but also their students and students’ students, etc. Twardowski was born in Vienna and studied at the University there under Franz Brentano and Robert Zimmermann, and it was from Brentano that Twardowski took two crucial methodological principles, from neither of which he ever deviated. First, philosophy is no less scientific than any other discipline and should be pursued with the unerring aim of uncovering and clearly stating important truths. Second, psychology plays a crucial foundational role in philosophy, so that without sound psychological underpinnings, a philosophy will be at best incomplete and at worst hopelessly mistaken. In pursuing the former principle, Twardowski arguably surpassed his teacher Brentano, both in the clarity of his statements and in his ability to bring forth brilliant students who were able to continue the pursuit of these aims with an unusually low level of the antagonisms and acrimonious disputes that normally characterize revolutionary movements. It is not that there were not the usual share of personal animosities among his students and grandstudents, but these tended to be both less vituperative and less dysfunctional than those among Brentano’s frantically squabbling followers. Critical editions of Twardowski’s published and unpublished writings being under

166

Peter Simons

way, one may with some confidence expect future histories of twentieth century philosophy to assign Twardowski a far more important, central, and influential position than hitherto.

2.

Brentano on Judgment

Twardowski’s theory of judgment is closely modelled on that of Brentano, so it is useful briefly to summarize the latter’s theory before describing Twardowski’s own contribution. Brentano followed Descartes in dividing mental phenomena—which possess that ineluctable property of intentionality—into three major classes: ideas (Vorstellungen), judgments (Urteile), and a third class, covering emotion, desire, and volition, which Brentano usually referred to as “phenomena of love and hate”, but which, for brevity, I shall call attitudes. These three classes are intended as an exhaustive and mutually exclusive division. Whether they are or not (several Brentano students preferred to see attitudes as not one but two or more basic classes), judgments, the second class, clearly form a crucial class for both logic and epistemology. The standard theory of judgment current in philosophy before Brentano was that judgments, at least simple categorical judgments, belong to one of two kinds. Positive judgments consist in the mental bringing together of two ideas, a subject idea and a predicate idea, into an affirmation in which the predicate idea is predicated of the subject, that is, the object or objects presented by the subject idea—for example, in the true judgments (one singular, one general) Socrates is a philosopher Horses are quadrupeds Negative judgments likewise bring two ideas together, but in a way that makes clear that they exclude one another, as in the (also true) judgments Socrates is not a banker Horses are not vegetables Often, this second type was described in terms of ideas being separated, but this is clearly not to be taken seriously. Brentano disagreed with the prevalent view because there are two kinds of simple judgment that do not fit either of these schemes. The first are existential judgments, such as Socrates exists Horses exist

Twardowski on Judgment

167

Zeus does not exist Unicorns do not exist where Brentano, following Hume, denies that ‘exist’ predicates anything of Socrates or of horses. The second are so-called subjectless judgments, like It’s raining Baby, it’s cold outside In place of the two-idea analysis, Brentano proposes (taking note from these examples) a one-idea analysis,2 where the judgment is a positive or negative existential one, as, for example, Philosopher Socrates exists Non-quadrupedal horses do not exist Banker Socrates does not exist Vegetable horses do not exist Present local rain exists External local cold exists (, Baby) Because ‘exist’ is not logically a predicate, Brentano understands judgment as the acceptance or rejection of the object or objects intended by the subject idea, so, for example, accepting Philosopher Socrates but rejecting vegetable horses. This was then termed the idiogenic theory of judgment, by contrast with the two-idea allogenic theory.3 More complicated cases are covered, as indicated here, by exploiting nominal conjunction and negation, and Brentano used this theory to give a highly streamlined system of syllogistic.4

3.

Content and Object of Ideas

When Brentano first introduced intentionality in his 1874 Psychology, he made no distinction between the mental content of an act and its (usually non-mental) object—for example, the content of the idea of Socrates and Socrates himself. This defect was first pointed out in print by Höfler in 1890,5 but it was Twardowski’s 1894 Habilitationsschrift, On the Theory of the Content and Object of Presentations, which gave several unanswerable arguments for distinguishing the two. In the case of ideas, the primary object of Twardowski’s treatise, this is very straightforward. For judgments, as we shall see, it is less obvious. In this book, Twardowski does not focus principally on judgments, but since ideas must figure in judgments, he often mentions judgment in passing. When I think of Socrates, I have an idea of him, which occurs when I do the thinking, and the content of this idea presents the object, the real Greek

168

Peter Simons

philosopher who died long ago. Twardowski also at times treats contents as abstract, as the general type of Socrates content as distinct from the individual content in someone’s mind at a certain time, particularly when he wishes to link content with linguistic meaning. He is in this piece rather cavalier—or sloppy—about passing between the individual mental content and the abstract content. For reasons into which we do not go here, Twardowski considers that every mental act has an object as well as a content. This has the following consequence: if I am thinking about a mythical or fictitious object or one wrongly postulated by false science, such as Zeus, Othello, the planet Vulcan, or the substance caloric, the object of my thought does not exist, but there still is an object. Twardowski does not take this view very far: it was his older contemporary Alexius Meinong who developed the theory in much greater detail, and to greater notoriety. But it was Twardowski who put Meinong on this path.6

4.

Judgment in Twardowski’s Dissertation

Twardowski’s doctoral dissertation, Idee und Perzeption in Descartes (1891, 1892), contains a few rather uncontroversial remarks on judgment, obviously as related to Descartes’s own writings. In the dissertation, Twardowski is clearly presupposing Brentano’s views as the background. The main thrust of his remarks is to the effect that while Descartes’s ideas can be understood in the sense of Brentano’s Vorstellungen, his perceptions fit neatly neither into the category of idea nor into the category of judgment. Perception is not judgment for Descartes, first because judgment consists in the affirming or denying of one idea of another, and this dual structure is clearly absent from perception, and second because Descartes takes judgment to require an act of will, and this, too, is absent from perception. On the other hand, perceptions are not mere ideas because, unlike the latter, they embody a truth-like aspect. A veridical perception reveals the world as it (in part) really is, and according to Descartes, clear and distinct perceptions guarantee truth. The upshot is that Descartes’s treatment of perception does not fit his own tripartite division of thoughts (mental acts) into ideas, judgments, and attitudes, the division taken over and maintained by Brentano. Perceptions fall into a gap between ideas and judgments, sharing aspects of these two categories. For Brentano, on the other hand, there is no ambiguity: perceptions are positive judgments, a position made possible by his rejection of the need for a judgment to contain more than one idea. Nor is there need for an act of will for each perception as there would be for Descartes. Not only do Brentano’s judgments not presuppose acts of will (which fall into the attitude category) but the passive flow of ideas thrown up in sensation can become a flow of perceptions provided an overall “blanket” of trust or acceptance is held over them, which therefore does not require a separate individual judging, one for each new idea.

Twardowski on Judgment

5.

169

Judgment in On Content and Object

By the time of his Habilitationsschrift, three years later, Twardowski had matured and refined his position on the distinction between contents and objects, as mentioned above. The now firmly justified distinction means he can be very clear in defending Brentano’s position on the idiogenic nature of judgment. When I judge that A exists, I am accepting A, not the connection of A with a property called “existence”, and when I judge that A does not exist, I am not thereby rejecting a connection between A and existence, but rejecting A itself. It is therefore the object A that is the object of a judgment, whether the object is accepted or rejected.7 In the case of positive judgments, this is not especially difficult, but in the case of negative judgments, we seem to be caught in a dilemma. A negative judgment rejects some object, and for it to reject this object, there has to be an object for it to reject. On the other hand, by rejecting, we are saying it does not exist, so how can there be an it to be rejected? This problem, that of negative existentials, has frequently bedevilled philosophers and was epitomized by Quine as “Plato’s beard”. Twardowski’s own solution is to make a distinction between being an object and existing. Clearly then, ‘there is’ must be understood more widely than ‘there exists’, the point taken up and developed further by Meinong. So to reject A in a negative existential is not to deny that there is an A but to deny that this A exists. Care is needed in order not to be caught in a contradiction. It is a consequence of Twardowski’s position that he is unable to allow for a kind of judgment rejecting an object altogether, as distinct from denying that such an object exists. An alternative and more ontologically parsimonious theory might have been to regard existential judgments as being not flatly about the object accepted or rejected but as conveying something about the judgment’s content, so that a positive existential judgment would be true just in case its content did denote an object and false if it did not, while a negative existential judgment would be true and false under the opposite conditions. Especially where ‘content’ is understood to mean the general notion rather than the individual mental content, this would be the sort of theory proposed by Kant, Bolzano, and Frege—that existential judgments (appearances to the contrary) are second-order; they are judgments not about objects but about concepts. Bolzano—who is cited by Twardowski as often as is Brentano—understood an existential proposition of the form As exist as having the true logical form the idea-in-itself of an A has objectuality. For Bolzano, some ideas lack objectuality or are objectless, such as round square or golden mountain. Twardowski maintains that Bolzano thinks this because he confuses not having an object with not being presented, but this is both wrong about Bolzano and unconvincing if considered as a theory in its own right. The only good point Twardowski has against Bolzano is that Bolzano’s example nothing (of an objectless idea) is syncategorematic and so not in the business of being about objects at all.

170

Peter Simons

It is ironic then, having distinguished content from object, that Twardowski does not exploit the distinction to offer a subtler account of existentials. He rather prefers to adhere closely to Brentano’s theory of the nature of judgment and allow non-existent objects. While not driving this theory to the extremes that Meinong did, he must take some responsibility for setting Meinong on that track. By contrast, Twardowski’s older contemporary Edmund Husserl, who studied On Content and Object closely, firmly rejected non-existent objects and maintained that when I think of Zeus, while I am minded in a certain way (zeusly), so there is a content, there is no object, existent or not, that my thought targets or refers to. When I judge at a particular time that Socrates was a philosopher, I accept the object formerly living philosopher Socrates. The judgment’s object is this past person, but its content (whether particular or abstract) cannot have ceased to exist thousands of years ago. That is one of the clear reasons for distinguishing content from object. Twardowski is, however, not very explicit here about the content of judgments and only mentions it once in On Content and Object. Noting that judgments, like ideas, have contents as well as objects, and because judgments are “wrapped up in a peculiar way”8 with the existence and non-existence of objects, since to accept A is to judge that A exists and to reject A is to judge that A does not exist, the contents of the respective judgments are to be taken as the existence of A and the non-existence of A, respectively.9 How these two indicated items are to be understood, as mental or abstract—and how something suitable to be called the existence of something or the non-existence of something can be either—is not made clear. The remarks are simply too sketchy to build on at this stage: we return to judgment contents below.

6.

Indexical Elements of Judgment

One of Twardowski’s most important and influential papers is his 1900 essay “On So-Called Relative Truths” (Twardowski 1900).10 In it, he argued that truths, which are true judgments, are absolute, that is, invariant or unchanging in truth value with respect to the time, place, person, or other context from which they are considered. This appears to go against the fact that someone who judges “It is raining” may judge truly, while another person judging “the same thing” may judge falsely, because in the locale of the first person, it is then raining, while in that of the second, it is not. Twardowski’s answer to such cases is that the judgments are not “the same”, even in abstract content, because the expression of the judgment by the same (type) sentence is in each case incomplete. For a complete specification of the content of the judgment, the context-dependent or indexical elements of the sentence—tense, person, place, etc.—need to be brought to expression, so that a more adequate expression of the content

Twardowski on Judgment

171

would be something like “At noon Central European Time on 1 January 1900 by the Gregorian calendar, it is raining on Castle Hill in Lvov and its surroundings”. While Twardowski’s aim in securing the absoluteness of truth is (in my opinion) admirable and correct, it must be recognized that his way of trying to do so—by packing the judgment content with all the elements needed to eliminate any context-relativity—is not the way to avoid relative truth. For one thing, the phenomenology of such judgments discloses no such elements. For another, a small child can correctly judge “It’s raining” while having not the slightest inkling of dates, calendars, or time zones, for example. Most obviously, even an adult who is—for whatever reason—in the dark as to where she currently is and/or what time it is can quite correctly judge that it is raining—namely, where she is at the time of judging. Twardowski is, I think, confusing a judgment’s content with its truth conditions. But all that is needed to secure the truth or falsity of a judgment like this (one with indexical elements) is precisely given by the judgment’s being made in a context. Given when and where the judgment is made—whether in Lvov in 1900, London in 2000, or Lusaka in 2100—its truth or falsity depends solely on the state of the weather there and then. The relations of the judgment to its context are not parts of the judgment, overt or covert. Similar remarks apply to other contextual or indexical features that go together with a judgment and the way the world is to determine the judgment’s truth value.11

7.

Judgment Objects and Contents More Generally

We saw that in the 1894 monograph, the content of a judgment was briefly referred to as being “bound up with” the existence or nonexistence of its object. In other writings, though not in any way systematically, Twardowski generalizes the view from overtly existential judgments to others that are not of this form. His examples, from scattered places in his writings, include All radii of a circle are equal to one another Two times two is four Poseidon was the god of the sea The first two are relatively straightforward. The first is a negative existential, rejecting the object Inequality of the radii of a circle while the second is positive, accepting the object Equality of the product of two and two with four

172

Peter Simons

The final one is also apparently positive and seems to accept Poseidon being god of the sea But Twardowski thinks that such a judgment would implicitly accept the existence of Poseidon, so he claims the wording is misleading and the subject should be not ‘Poseidon’ but ‘thing called “Poseidon”’.12 Quite how this gets him out of the problem, I fail to see, unless the whole sentence is modified roughly as The thing called “Poseidon” is a thing called “the god of the sea” This seems to accept something called “Poseidon”, but presumably this is not an existing object. Twardowski does not, however, elaborate further. The upshot of this hint of an analysis is that when we have a categorical judgment, we obtain a designation for the object of the judgment, positive or negative, by nominalizing the main verb and retaining the terms in a suitable way. With non-symmetric relational judgments such as Four is greater than two the object would then be the being greater than two of four or perhaps four’s being greater than two while with verbs taking more than two objects, like Jan gave Maria a ring for her birthday the object would be something like The giving of a ring to Maria for her birthday by Jan or Jan’s giving Maria a ring for her birthday Assuming something like this can be made to work, we still have to raise the question as to what kind of things these items are whose existence is accepted or rejected. They are certainly not concrete things like Jan, Maria, or a ring. Twardowski does not go into the matter, and seems

Twardowski on Judgment

173

curiously incurious as to their status. Others of Brentano’s former students, such as Marty, Stumpf, Meinong, and Husserl, make a good deal of noise about such items, which Marty calls ‘judgment contents’, Meinong calls ‘objectives’, and Stumpf and Husserl call ‘states of affairs’. If there is anything like an acceptance of states of affairs in Twardowski’s work, it is extremely underdeveloped, but I hazard that Twardowski would in fact have rejected the ontological category of states of affairs. When mentioning the possible objects of judgment, he lists person, thing, phenomenon, and, “most importantly”, relation.13 But, again, he does not elaborate. The contents of judgments like these are always, for Twardowski, existence or non-existence of the object in question. Clearly, by ‘existence’, he cannot mean solely concrete, spatio-temporal existence; something like an equality cannot be a concrete entity. Sometimes he uses the term ‘actuality’ instead of ‘existence’, and in certain cases this is more idiomatic. We would be more likely to say an equality is actual than that it exists. Meinong used the word bestehen, subsist or obtain, for such cases and reserved ‘exists’ for concrete entities only, but Twardowski uses ‘exists’ in the broader sense, where it covers ideal as well as real being. In a short paper of 1907, “On Idio- and Allogenetic Theories of Judgment”,14 Twardowski again upholds Brentano’s theory, denying that all judgments need two ideas that are combined or synthesized and affirming again that the content of all judgments is existence (accepting or rejecting being the act-quality, not the content). More surprisingly, however, he declares that idiogenic theories “do not consider representations to be essential constituents of the judgment, but rather conditions for it”.15 Quite what this brief remark intends is not clear. Perhaps Twardowski has something like the following in mind. Suppose someone raises the question whether the ancient British King Arthur really existed, and the evidence for and against is rehearsed in a group discussion. Eventually, perhaps, a participant in the debate, who had up until then, as we say, withheld judgment, opts for one of the alternatives: yea or nay. This opting and putting an end to the indecision is the judging, and since presentations (ideas) of King Arthur had been in play throughout the discussion, there does not need to be a special idea-of-King-Arthur as a part of this act of judging itself. This is not to deny that many judgments do contain ideas of objects judged to exist or not, or which enter in indirectly, as when I judge that there is no present King of France (here, the judgment contains not just an idea of a present King of France but also an idea of France), only that the ideas of objects that figure in the judgment are not essential parts of the judgment as such.

8.

Judgment and Truth

One thing that is bound to strike a modern reader is that almost all of Twardowski’s discussion, as indeed that of Brentano, concerns very simple judgments, whether categorical or overtly existential. In modern logic,

174

Peter Simons

however, such simple judgments form only the basis for a recursion in which ever more complex forms of judgment can be built up, without theoretical limit. We need only think of the connectives of propositional logic, or the quantifiers of predicate logic, to see that the idiogenic theory could only be stretched to cover such cases by extremely artificial means. Even simple predicative and relational judgments, as we saw above, require a nominalization in order to show their existential form. How a highly complex formula, such as we find, for example, in Frege’s logical treatises, could be tortured into existential form defies credulity. There is indeed a precedent for seeing all propositions, even the most complex, as being reducible to a very simple form (namely, that of subject–copula– predicate) in a source of which Twardowski was well aware—namely, Bolzano’s Theory of Science. This is, however, one of Bolzano’s more unbelievable doctrines, and the artifices that his reductions are compelled to employ are far from convincing; indeed, in many cases, they change the sense of the proposition.16 Whether Twardowski had a similar sort of reduction in mind is moot. On the question what constitutes the truth of a judgment, Twardowski is brief and unexcitingly orthodox: a judgment is true if things are as it says they are. However, given the idiogenic theory, this means that the object of a true positive judgment is—in more recent terminology—a truthmaker for the judgment in question, that is, an object that, because it exists, makes the judgment true. It also, thereby, and with no further ado, makes the contradictory negative judgment false. The case of true negative judgments is more interesting. Since Twardowski accepts non-existent objects, there is an item in play. A rather piquant suggestion would be to add a dual to truthmaking, whereby, by not existing, this item makes it false that it exists, and so without further ado makes it true that it does not exist. Less expansive ontologies than Twardowski’s lack this option.

9.

Judging and Its Products

Twardowski is well aware that the term ‘judgment’ (in Polish, sąd) has a number of meanings. He principally uses it for the act of judging, but in his influential essay ‘Actions and Products’ of 1912, he distinguishes between the act of judging and the product of this, for which he now reserves the term ‘judgment’. For the act, he prefers (at least in this context) to speak of “passing [or making a] judgment” (wydawać sąd—cf. German ein Urteil fällen). In English, we can use the -ing form: judging. He also notes that the term can be used for the disposition to judge (in a certain way), as when we say a person has good judgment. And, finally, he mentions that some writers use the term for the public enunciation of a judgment. The judging as an act is now familiar: it is a concrete, dated mental event. Its product, the judgment, is less clear. Unlike, say, the product of an

Twardowski on Judgment

175

activity of building, it is not an enduring object that outlasts the judging event. Nor does it obviously concern the type or abstract content—when we say, for example, Jan judged that Warsaw is larger than Kraków Here, the that-clause may indeed be held to express the abstract sense or content, the proposition, that Warsaw is larger than Kraków. But such an abstract entity could hardly be considered the product of a mental act by Jan or, more problematically still, by anyone else who judges similarly, but at best the type of such a product, a token of which is produced in Jan’s so judging. But it is then unclear whether this product is or is not the same as the judgment’s content. A judging may certainly have results: not only does it (trivially) result in this judgment having been made; it typically results in a state or disposition in the judger of believing the judgment’s content and may also be a causal contributor to a judger’s acting in a certain way. For example, my judging that a nearby child is dangerously close to the edge of a station platform may result in my pulling them back before a train passes. But that which has effects must be an event, not the non-event product thereof. I conclude that, probably because of his enthusiasm for the action/product distinction, Twardowski fails to give a definitive sense to the idea of judgment as the product of an act of judging.

10.

Judging and Logic

In 1902/03, Twardowski lectured on logic in Lvov, and he prefaced his lectures with a discussion of the theory of judgments, which was published as a fragment in 1996, with an English translation in 2014.17 Despite the title of the fragment, ‘The Theory of Judgments’, it is, in truth, a disappointing piece. Twardowski mentions that judgments may have different kinds of parts: physical (parts that can exist separately from the whole), metaphysical (parts that can exist only within their whole), and logical or one-sidedly separable, e.g. that colour is part of the notion of blue but not vice versa. But he does not follow up on the promising start. Elsewhere, he notes—uncontroversially—that the distinctive mark of judgments is their susceptibility to being true or false. But for the most part, the fragment is about images as distinct from judgments and about philosophical methodology. Among Twardowski’s students who turned more definitively to logic, the notion of a judgment was replaced by that of a sentence (zdanie). This was part of the linguistic and anti-psychologistic turn in logic, which spread from Frege and Russell to Poland, starting with Jan Łukasiewicz and being carried forward by Leśniewski, Ajdukiewicz, Tarski, and others. Twardowski himself was persuaded by Husserl’s Logical Investigations of

176

Peter Simons

the falsity of psychologism in logic18 but persisted in treating judgments as the primary bearers of truth and falsity, and indeed cautioned, in the 1921 essay “Symbolomania and Pragmatophobia”, against treating the formalism of symbolic logic as something taking on a life of its own, independently of meaning (1921). While the main overt target of his essay is a piece by the French physicist Henri Bouasse, the essay may also have served as a covert warning to those of his compatriots who were becoming overly enamoured of symbols at the expense of their meaning. At any rate, the need to consider the meaning of the terms in symbolic logic was not a view lost on Łukasiewicz, Leśniewski, Ajdukiewicz, or, later, Tarski, no matter how virtuosic their treatments in symbolic logic. None of the Polish logicians was a formalist in the sense of Hilbert. What was left behind, though, was any consideration of the mental for its own sake.

11.

Continuing Loyalty

It is a mark of Twardowski’s loyalty to Brentano’s idiogenic theory of judgment that as late as 1925, when he gave a lecture course entitled ‘Theory of Knowledge’,19 he continued to propound and defend the theory with the same arguments as before, especially the inability of orthodox (allogenic) accounts to give a satisfactory treatment of impersonal (subjectless) and existential judgments. Indeed, he invoked Aristotle’s famous definition of truth and falsity,20 To say of what is that it is not, or of what is not that it is, is false, while to say of what is that it is, and of what is not, that it is not, is true as supporting the view that propositions are in general of existential form, rather than as supporting a correspondence theory of truth such as one may find in Aquinas or Russell, the latter of whose version of the correspondence theory and his multiple relation theory of believing are criticized at some length by Twardowski.21 Certainly, the Greek verb for ‘to be’ can be used existentially—to be is to exist. But it can also be used factively—to be is to be the case. Aristotle’s account is, in truth, schematic enough to cover a number of possible theories, but it will only completely sustain an idiogenic theory if all propositions can be put into existential form. While Brentano showed this can work for categorical propositions, it is, as we saw, highly problematic to extend it beyond these to propositions of any possible complexity. It is notable that just a few years later, Tarski was invoking the same passage as background to his semantic theory of truth, which is neither a correspondence theory nor an idiogenic theory, but employs the more general notion of satisfaction to allow a treatment to be given of quantificatory sentences of any complexity.

Twardowski on Judgment

12.

177

Verdict

As we have seen, Twardowski’s theory of judgment is based on that of Brentano and, in the main points, never deviates from it. The addition of the content of judgment to its object does not contribute to a general theory of judgment contents beyond taking them all to be of existence or non-existence. We already noted the difficulties in putting all kinds of judgment into positive or negative existential form as well as the incompletely sustained distinction between judgment contents as individual aspects of concrete acts of judgment vs contents as abstract, shareable and communicable, vs judgments as products of judgings and, finally, the unsatisfactory account of the role of context in indexical judgments despite Twardowski’s merits in highlighting the issue and endeavouring to defend absolute truth without platonism (such as we find in Frege or Husserl). Nor does Twardowski go into other types of mental acts that are truth-valuable, such as propositional memory, as well as such acts as doubting, wondering, surmising and conjecturing, or dispositions such as believing or inclining to believe. Twardowski’s older contemporaries Meinong and Husserl were much more thorough in their discussion of the phenomenology of mental acts, which, in the analytic tradition, are known as ‘propositional attitudes’. It seems that Twardowski was largely content to maintain Brentano’s account of judgment, having once distinguished content and object. Twardowski’s insistence that all mental acts have an object, carried forward and further by Meinong, is a point on which majority opinion has been against him, and, indeed, his reasons for taking the view—namely, that apparent cases of objectless acts are a result of confusing having an object with presenting an act—are unconvincing. All in all, Twardowski’s account of judgment, despite some interesting and developable aspects, appears to me—at this point anyway—to fall below the level of development of his account of ideas (presentations).

Notes 1. Twardowski alludes rather obliquely to this (most appropriate) biblical passage in his paper on Actions and Products: vide Twardowski (1999: 112 n). 2. Brentano (1874: Ch.7). 3. Hillebrand (1891: § 16, p. 26). Hillebrand’s terms were ‘idiogenetic’ and ‘allogenetic’, and Twardowski followed this. But, as was pointed out later by D. Tennerówna, the shorter terms ‘idiogenic’ and ‘allogenic’ are etymologically more appropriate. Cf. Twardowski (1999: 99 n). 4. See Simons (1987). 5. Höfler (1890: 6). 6. This was pointed out in J. N. Findlay’s dissertation: cf. Findlay (1963 : ch. 1). Not coincidentally, Findlay begins his treatment of Meinong’s theory of objects with a discussion of Twardowski’s distinction. 7. Twardowski (1894: 8). 8. “In einer eigenthümlichen Weise verwoben”: Twardowski (1894: 8). 9. Twardowski (1894: 9).

178 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

Peter Simons On the influence, see Woleński and Simons (1989). Cf. Simons (1992, 2003). Twardowski (1894: 28). Twardowski (1999: 202). Twardowski (1999: 99–102). Twardowski (1999: 99). Cf. Simons (1999). Twardowski (2014: 161–80). As stated in his short autobiography, Self-Portrait: cf. Twardowski (1999: 31). Twardowski (1999: 181–242). Twardowski (1999: 192). Twardowski (1999: 194 ff). This, together with Twardowski’s discussion of William James and the pragmatic theory of truth, shows that, like many of Brentano’s former students, he was familiar with contemporary English language writings, in addition to those in Polish, German, French, and, perhaps, further languages.

References Works by Twardowski Twardowski (1892), Idee und Perception. Eine erkenntnis-theoretische Untersuchung aus Descartes. Vienna: Hölder. Twardowski (1894), Zur Lehre vom Inhalt und Gegenstand der Vorstellungen— Eine psychologische Untersuchung. Vienna: Hölder, 1894 (repr. Munich and Vienna: Philosophia Verlag, 1982). Eng. trans. On the Content and Object of Presentations, trans. R. Grossmann. The Hague: Nijhoff, 1977. Twardowski (1900), “O tak zwanych prawdach względnych” (On So-Called Relative Truths). Księga Pamiątkowa Uniwersytetu lwowskiego ku uczczeniu pięćsetnej rocznicy fundacji Jagiellońskiej Uniwersytetu krakowskiego. Lwów, nakładem Senatu Akademickiego Uniwersytetu lwowskiego, 1900; repr. in Twardowski (1965), 315–36; Modified German transl. Über sogenannte relative Wahrheiten (M. Wartenberg), Archiv für systematische Philosophie 8, 1902: 415–47; repr. in Pearce, D. and Woleński, J. (eds.), 1988, Logischer rationalismus. Ausgewählte Schriften der Lemberg-Warschauer Schule. Frankfurt: Athenäum, 38–58. Eng. trans. from Polish by A. Szylewicz in (1999), 147–69. Twardowski (1912), “O czynnościach i wytworach—Kilka uwag z pogranicza psychologii, gramatyki i logiki” (On Actions and Products—Some Remarks from the Borderlines of Psychology, Grammar and Logic), in Księga Pamiątkowa ku uczczeniu 250-tej rocznicy założenia Uniwersytetu lwowskiego przez króla Jana Kazimierza. Tom II. Lwów, nakładem Uniwersytetu lwowskiego, 1912, 1–33. Reprint in (1965), 217–40. German version: “Funktionen und Gebilde” (J. L. Brandl, ed.), Conceptus XXIX (75), 1996, 157–89. Eng. trans. on the basis of German and Polish by A. Szylewicz in (1999), 103–32. Twardowski (1921), “Symbolomania i pragmatophobia” (Symbolomania and pragmatophobia). Ruch filozoficzny VI: 1–10; reprinted in (1965), 394–406. Eng. trans. by A. Szylewicz in (1999), 261–70. Twardowski (1965), Wybrane pisma filozoficzne. Warsaw: PWN, 1965.

Twardowski on Judgment

179

Twardowski (1999), On Actions, Products and Other Topics in Philosophy, ed. J. Brandl and J. Woleński. Amsterdam: Rodopi. Twardowski (2014), On Prejudices, Judgments and Other Topics in Philosophy, ed. A. Brożek and J. J. Jadacki. Amsterdam: Brill/Rodopi.

Works by Other Authors Brentano, F. (1874), Psychologie vom empirischen Standpunkte. Leipzig: Duncker & Humblot (2nd, enl. ed. O. Kraus, 1924, Leipzig: Meiner). Eng. trans., Psychology From an Empirical Standpoint, trans. A. C. Rancurello, D. B. Terrell and L. McAlister. London: Routledge, 1973 (2nd ed. 1995). Findlay, J. N. (1963), Meinong’s Theory of Objects and Values. Oxford: Clarendon Press. Hillebrand, F. (1891), Die neuen Theorien der kategorischen Schlüsse. Vienna: Hölder. Höfler, A. and A. Meinong (1890), Philosophische Propädeutik—I Theil: Logik. Prague, Vienna, and Leipzig: Tempsky/Freytag. Simons, P. (1987), “Brentano’s Reform of Logic”, Topoi 6: 25–38. Simons, P. (1992), “Verità atemporale senza portatori di verità atemporali”, Discipline filosofiche 2: 33–47. Simons, P. (1999), “Bolzano über Wahrheit”, in E. Morscher (ed.), Bolzanos geistiges Erbe für das 21. Jahrhundert. Sankt Augustin: Academia. Simons, P. (2003), “Absolute Truth in a Changing World”, in J. Hintikka, T. Czarnecki, K. Kijania-Placek, T. Placek and A. Rojszczak (eds.), Philosophy and Logic: In Search of the Polish Tradition. Essays in Honour of Jan Woleński on the Occasion of His 60th Birthday. Dordrecht: Kluwer. Woleński, J., and P. Simons (1989), “De veritate: Austro-Polish Contributions to the Theory of Truth From Brentano to Tarski”, in K. Szaniawski (ed.), The Vienna Circle and the Lvov-Warsaw School. Dordrecht: Kluwer.

9

Attitudinal Objects Their Ontology and Importance for Philosophy and Natural Language Semantics Friederike Moltmann

There are two categories of entities that are generally considered central for propositional attitudes and illocutionary acts. The first is events (including actions and states), the second propositions. Both are generally taken for granted and considered well motivated, both for the purposes of philosophy and for the purposes of natural language semantics. In this paper, I argue that a third category of entities should be given even greater importance—namely, what I call attitudinal objects (Moltmann 2003a, b, 2013a, b, 2014, 2017a, b, c, 2018a). In addition to presenting general philosophical and linguistic arguments for this category, the paper provides significant revisions and refinements of the notion of an attitudinal object as it was developed in my previous work. Attitudinal objects are entities like judgments, claims, beliefs, decisions, desires, fears, intentions, promises, and requests. Attitudinal objects are concrete, agent-dependent entities that come with truth or satisfaction conditions. It is common in both philosophy and linguistics to take nouns like judgment to stand either for an act or for a proposition, depending on the context. I reject that view: nouns like judgment always stand for attitudinal objects. Attitudinal objects bear a range of types of properties that jointly characterize them as an ontological category and distinguish them from events, actions, and states, as well as from propositions. Attitudinal objects are well reflected in natural language, not just in nominalizations of attitude verbs, but also in a range of generalizations regarding attitude reports, even though attitudinal objects have hardly been recognized as such in philosophy of language and linguistic semantics. However, in a way, attitudinal objects had been recognized by one philosopher in the past—namely, Twardowski (1912)—who took them to be the (non-enduring) products of acts, in the context of a general distinction between actions and products. In this paper, I will reject this as a general characterization of attitudinal objects. Attitudinal objects should not generally be regarded as non-enduring abstract artefacts even when they result from actions. Certain types of mental acts are, in fact, ontologically derivative with respect to the attitudinal objects that are their results.

Attitudinal Objects

181

Twardowski’s characterization would also not be applicable to statelike attitudinal objects such as beliefs, intentions, desires, and fears, and one might exempt them as states from the action-product distinction. In this paper, however, I will argue against an identification of state-like attitudinal objects with mental states. A belief has very different sorts of properties from a belief state, and the latter, I will argue, is ontologically derivative upon a belief and plays different semantic roles. Attitudinal objects thus are at least, to an extent, ontologically prior to acts or states. Moreover, attitudinal objects allow us to dispense with propositions. Attitudinal objects, it will be argued, play a central role as concrete bearers of truth or satisfaction conditions in our mental life and in communication. It is, moreover, attitudinal objects rather than abstract propositions that, on the present view, are involved in the semantics of attitude and illocutionary act reports and, of course, nominalizations of attitude verbs. The paper will first lay out the standard view about propositional attitudes and introduce the alternative view based on attitudinal objects. Second, it will present the various motivations for attitudinal objects: the semantics of nominalizations of attitude verbs, semantic generalizations about attitude reports, and intuitions about the roles and properties of attitudinal objects. Third, it will discuss the difficulties for Twardowski’s action-product distinction and the relation between attitudinal objects and states and address the question of the Davidsonian event argument of stative attitude verbs. The paper will conclude with some indications about how certain puzzles regarding the individuation of attitudinal objects may be addressed.

1.

The Standard View About the Objects Associated With Mental and Illocutionary Acts

On the view that is standard both in philosophy and in natural language semantics, there are two sorts of objects associated with propositional attitudes and illocutionary acts: (a) mental acts or states and speech acts (b) propositions as the objects or contents of propositional attitudes or illocutionary acts. Standardly, beliefs, desires, hopes, and intentions are considered mental states, on a par with mental acts or events. Actions, events, and states are generally taken for granted ontologically or at least felt not in further need of explanation. They are equally well accepted in linguistic semantics and, in particular, posited as implicit arguments of verbs, following the influential Davidsonian analysis of action sentences (Davidson 1967).

182

Friederike Moltmann

Propositions are considered much more controversial. Propositions are generally taken to be mind- and language-independent objects that serve the roles of shareable contents of attitudes, of truth-bearers, and of the meanings of sentences (relative to a context). The notion of an abstract proposition in this sense has been the subject of various sorts of critique, both philosophical and linguistic. The philosophical critique concerns questions such as how propositions can be grasped, how propositions can act as the contents of mental attitudes, and how propositions can be true or false and have the particular truth conditions they are meant to have (the problem of the unity of the proposition) (Jubien 2001; Soames 2010; Hanks 2011; Moltmann 2003a, 2013a). It also concerns the status of propositions as objects of attitudes, that is, the view that attitudes are relations between agents and propositions. To make use of a distinction of Brentano’s, propositions should be contents, rather than objects of attitudes, according to the critique. On the standard view, the logical form of attitude reports and speech act reports looks as follows, making use both of propositions and of events (including actions and states): (1) a. John thinks that Mary is happy. b. ∃e(think(e, John, [that Mary is happy])) In (1b), [that Mary was happy] is the meaning of Mary is happy, the proposition that Mary is happy. The logical form in (1b) makes clear the role of propositions in propositional attitudes, as entities propositional attitudes are directed towards, rather than as contents of attitudes. I will outline a view according to which a third category of entities plays a central role for propositional attitudes and illocutionary acts—namely, attitudinal objects. Attitudinal objects consist in act-related attitudinal objects, such as judgments, decisions, claims, requests, and promises, as well as state-related attitudinal objects, such as beliefs, intentions, desires, and fears. In my previous work, I called the former cognitive and illocutionary products and the latter mental states, terms I now consider problematic. Attitudinal objects do not just form a list of entities; rather, they share characteristic properties that jointly distinguish them from other, related types of entities. Attitudinal objects share properties of content-based causation, perception, evaluation, and memory. Moreover, they share content-related properties—namely, truth or satisfaction conditions—a part structure based strictly on partial content and the ability to stand in similarity relations based on shared content only. Attitudinal objects are mind-dependent particulars that generally have only a limited lifespan. In particular, act-related attitudinal objects generally do not last beyond the act that has produced or triggered them. Despite being mind-dependent particulars, attitudinal objects can account for the shareability of content due to their content-related properties

Attitudinal Objects

183

Attitudinal objects allow us to dispense with propositions as truthbearers. In fact, there are good reasons to consider attitudinal objects the primary truth-bearers and to take propositions only derivatively, if at all, to play that role (Moltmann 2018a). Unlike propositions, attitudinal objects do not act as the meanings of sentences. Attitude reports report attitudinal objects, but without attitudinal objects being the semantic values of that-clauses. How, then, do that-clauses relate to the attitudinal object that is reported? On the present view, clausal complements of attitude verbs act semantically as predicates of the reported attitudinal object, specifying its truth or satisfaction conditions (Moltmann 2014, 2017a, 2018a, b). If the clausal complement of an attitude verb just has the function of conveying a property of the reported attitudinal object, this has important consequences regarding notorious problems for propositions. Propositions as abstract objects that are both meanings of sentences and objects of attitudes raise the question of how they can have truth values and of how, if they are structured objects, their truth values are determined (the problem of the unity of the proposition). Attitudinal objects are minddependent particulars, and thus their ability to represent their truth- or satisfaction-directedness can be attributed to the intentionality of the mind itself. Only abstract meaning objects pose the problem of the truthdirectedness and, if they are structured, of the unity of the proposition. Also on the present view, truth-bearers are no longer treated as the objects of attitudes; rather, having a propositional attitude means engaging (as agent or experiencer) in an attitudinal object whose truth or satisfaction conditions are given by the complement clause. Having a propositional attitude thus does not mean standing in an attitudinal relation to a meaning object and a bearer of truth conditions. Could attitudinal objects dispense with events? I will address this question in detail later. For the moment, let us just note that attitudinal objects and Davidsonian events (that is, events in their roles as implicit arguments of verbs) have very different motivations. Davidsonian events are meant to be the objects to which adverbials apply, whereas attitudinal objects are mind-dependent entities that are bearers of truth or satisfaction conditions. As will be discussed in greater detail later, events are not bearers of truth or satisfaction conditions or other content-related properties. This is reflected in our intuitions about the attribution of truth or falsehood and, even more strikingly, of satisfaction or violation. An act of judging or claiming is intuitively not true or false, unlike a judgment or claim. Even more strikingly, an act of promising cannot be fulfilled or broken, unlike a promise; an act of demanding cannot be complied with, unlike the demand; and an act of recommending cannot be followed or taken up in the way a recommendation can. I will follow the common assumption in semantics that all verbs have an additional argument position for events (including actions and states), so

184

Friederike Moltmann

that temporal and other adverbials can act as predicates of those implicit event arguments and sentences themselves will involve existential quantification over events (Davidson 1967). Applied to attitude verbs, this means that the logical form of an attitude report as in (2a) will be as in (2b): (2) a. John claims that Mary is guilty. b. ∃e(claim(e, John) & [that Mary is guilty](att-obj(e)) In (2a), the that-clause acts semantically as a predicate of the attitudinal object related to the implicit event argument e of the verb, that is, att-obj(e). The questions I will address, then, are the following: what are the philosophical and semantic motivations for attitudinal objects, what are their characteristic properties, and how are attitudinal objects to be understood ontologically? One might be tempted to assimilate attitudinal objects to familiar ontological categories, in particular to actions, events, and states. I consider this the wrong move. First, it is unclear whether those categories are really better understood than attitudinal objects. Second, it is important to first focus on the types of properties that attitudinal objects have, and they are, in fact, rather different from the types of properties that are characteristic of actions, events, and states. Given the distinctness of attitudinal objects from events, actions, and states, it appears that some types of attitudinal objects are, in fact, ontologically prior to the actions or states they may be correlated with; that is, their correlated actions or states will have to be defined in terms of the attitudinal objects rather than vice versa.

2.

The Reflection of Attitudinal Objects in Natural Language

Attitudinal objects are extremely well reflected in natural language. In particular, attitudinal objects are the referents of nominalizations of attitude verbs, as in (3a, b): (3) a. John’s judgment that Mary is guilty b. John’s claim that Mary is guilty This means that the ontology of attitudinal objects will be reflected in the semantic behaviour of such nominalization. There is a standard view, though, according to which such nominalizations are ambiguous or rather polysemous, standing either for events (with their causal and temporal properties) or for propositions (with the truth-related properties), depending on the predicate with which they occur. The reason is that the entities that nominalizations of attitude verbs stand for seem to display both properties of concreteness characteristic

Attitudinal Objects

185

of events and content-related properties characteristic of propositions, illustrated in (4) and (5), respectively: (4) a. John’s claim caused astonishment b. John’s claim yesterday was astonishing (5) a. John’s claim is true b. John’s claim implies that Mary is guilty There are serious difficulties for that view, though. First, there are cases of co-predication involving the attribution of an event-related (causal or perceptual) predicate and a proposition-related (truth-related) predicate to the same entity: (6) a. John overheard Bill’s claim, which implies that Mary is guilty b. John’s obviously false claim yesterday caused astonishment Co-predication arguments are notoriously problematic, though, and various approaches dealing with them have been developed that would not involve positing a single entity as the target of the two predicates.1 A better argument against polysemy comes from the fact that there are predicates applicable to what such nominalizations stand for that could neither be predicated of events nor of propositions. First, these are predicates of satisfaction, such as satisfy, fulfil, comply with, follow, take up, violate, ignore. The agent-related predicates of satisfaction keep and break illustrate the contrast particularly well:2 (7) a. John kept/broke his promise. b. ??? John kept/broke the proposition that S. c. ??? John kept/broke his speech act. Second, part-related expressions, such as part of, show that nominalizations of attitude verbs could stand neither for events nor propositions: (8) a. part of John’s claim b. part of John’s (act of) claiming (9) a. part of John’s promise b. part of John’s (act of) promising (10) ? part of the proposition (8a) and (9a) have a very clear meaning: part of here stands for partial contents (Moltmann 2017a, b).3 Parts of events or actions are temporal parts. But such parts cannot, under any reading, be picked out when part of applies to nominalizations of attitude verbs as in (8a) and (9a). Part of, when applied to propositions as in (10), does not have a clear

186

Friederike Moltmann

understanding in the first place. Since proposition is a technical term, it depends on the conception of a proposition. Structured propositions would have as parts the components of a structured proposition. But, clearly, this is not what part of picks out when applied to claims or promises. Also, considerations regarding identity statements support the view that nominalizations of attitude verbs do not stand for propositions (on one of their two putative readings). Thus, the following identity statements appear false: (11) a. ??? John’s thought that it will rain is also his remark that it will rain. b. ??? John’s discovery that it will rain is his hope that it will rain. c. ??? John’s desire to leave is his decision to leave. This is because thoughts, remarks, discoveries, hopes, and desires are simply not propositional contents. Thus, satisfaction predicates, part-related expressions, and identity statements support the view that non-gerundive nominalizations of attitude verbs stand for attitudinal objects, rather than being polysemous between referring to propositions and referring to events or actions. Later, I will discuss in further detail the properties attitudinal objects have. Attitudinal objects come in kinds, kinds whose instances are maximal classes of exactly similar products. At least this is what natural language reflects with the availability of kind terms in Carlson’s (1977) sense, as in the examples below: (12) a. The belief that god exists is widespread. b. John often encounters the expectation that he should become famous. Kinds need not be conceived as single abstract objects, but may rather be viewed as pluralities of (possible and actual) instances (Moltmann 2013a). However they may be conceived, kinds should inherit truth properties from their instances, as below: (13) a. The belief that John won the race is true. b. The expectation that John would become famous was not fulfilled. Reference to kinds of attitudinal objects is important, in that it permits reporting the sharing of a propositional content: (14) John and Bill share the belief that Mary is guilty Here, the belief that Mary is guilty stands for a kind of attitudinal object.

Attitudinal Objects

187

Attitudinal objects and kinds of them have another important reflection in natural language besides the semantics of nominalizations of attitude verbs—namely, in the semantics of quantifiers and pronouns that can take the position of clausal complements, what I call ‘special’ or ‘nominalizing’ quantifiers and pronouns (Moltmann 2003a, b, 2013a, 2014, 2017a). In English, these are quantifiers like something or several things and the pronouns what and that, as below: (15) a. John claims/knows/fears something. b. John imagines/expects that. c. John claims what Mary claims. On the standard view, such quantifiers and pronouns are taken to stand for propositions. Only if they stand for propositions, according to a common assumption, can they validate inferences, as in (16a, b): (16) a. John thinks that Mary is happy. John thinks something. b. Mary believes everything Bill believes. Bill believes that it is raining. Mary believes that it is raining. However, the actual semantic behavior of special quantifiers and pronouns shows that such quantifiers and pronouns cannot stand for propositions, but rather stand for attitudinal objects or kinds of them. Thus, restrictions of special quantifiers cannot generally be understood as predicates of propositions; rather, what they are predicated of is attitudinal objects or kinds of them, as illustrated in the examples below: (17) a. John said something nice (namely, that S). b. John thought something daring (namely, that S). c. John said something that made Mary very upset. It is not a proposition that is said to be nice in (17a), but rather something like John’s remark or John’s claim. It is not a proposition that is said to be daring in (17b), but a thought. Moreover, it is not a proposition that could have made Mary upset according to (17c), but rather it is a claim or remark that did so. In general, restrictions of special quantifiers are to be understood as predicates of (kinds of) attitudinal objects, not (abstract) propositions. Reports of sharing of content among different attitudes with special quantifiers or pronouns make the same point. Unlike what the standard view would predict, such reports are not really available when the two propositional attitudes are significantly different, as below: (18) a. ?? John screamed what Mary believes—namely, that Bill was elected president.

188

Friederike Moltmann b. ?? John expects what Mary believes—namely, that Sue will study harder. c. ?? John said what Mary believes—namely, that it will rain.

The unacceptability of (18a–c) makes clear that what Mary believes cannot just stand for a proposition. The unacceptability, or rather falsehood, of the examples below, with the corresponding nominalizations, indicates that what Mary believes stands in fact for a belief, an attitudinal object, rather than an abstract proposition: (19) a. ?? John’s scream was Mary’s belief. b. ?? John’s expectation is Mary’s belief. c. ?? John’s claim was Mary’s belief. In (19a–c), what Mary believes will actually stand for a kind of attitudinal object, of the sort the belief that S, rather than a particular attitudinal object, of the sort Mary’s belief that S. Thus, special quantifiers range over attitudinal objects or kinds of them when they take the complement position of a clausal complement. They are nominalizing quantifiers, in the sense of quantifiers that range over the sorts of things the nominalization of the verb would stand for, rather than what could be the semantic values of a that-clause, and similarly for special pronouns. This means that the logical form of (17b), repeated as (20a), will be as below: (20) a. John thought something daring. b. ∃e ∃e’(think(e, John) & daring(e’) & e’ = att-obj(e)) (20b) involves existential quantification associated with the Davidsonian event argument of think as well as existential quantification associated with the nominalizing quantifier something. Reports of sharing of the content of attitudes of different agents as in (21a) involve existential quantification associated with the Davidsonian event argument positions of the two attitude verbs as well as existential quantification associated with the special pronoun what: (21) a. John thought what Mary thought. b. ∃e e’e’’(think(e, John) & e’ = att-obj-kind(e) & think(e’’, Mary) & e’ = att-obj-kind(e’’)) There is further support for the semantic involvement of attitudinal objects in attitude reports, and that comes from the availability of complex attitudinal predicates instead of simple attitude verbs. Complex attitudinal predicates involve a light verb and a noun or noun phrase standing for an attitudinal object—for example, have a belief, make a

Attitudinal Objects

189

judgment, or give advice. Sometimes simple attitude verbs alternate with complex predicates (think—have the thought that S, believe—have the belief, claim—make a claim); sometimes the complex form is the only option (have the impression, German Angst haben ‘have fear’). The compositional semantics of the complex predicate as in (22a) obviously involves attitudinal objects, as, roughly, in (22b), and thus comes close to the semantics of simple attitude verbs that was given earlier, as in (23b) for (23a): (22) a. John has the thought that S. b. ∃!d(have(John, d) & thought(d) & that S(d)) (23) a. John thought that S. b. ∃e(think(e, John) & [that S](att-obj(e))) Attitudinal objects thus are explicitly involved in the semantics of both complex attitudinal predicates and special quantifiers. The semantics of special quantifiers raises the question of what enables them to quantify over attitudinal objects or kinds of them. There are two options to consider. One of them is that the morpheme -thing in (20a) incorporates into the verb think, leading to V-thing expressing a relation between agents and attitudinal objects (Moltmann 2003a, b, 2013a). Another option to consider is that the verb think in (20a) is somehow interpreted on the basis of the complex predicate have the thought. Special quantifiers and pronouns would then match the contribution of the nominal part the thought, rather than the complement clause. Both options are associated with a range of linguistic issues and challenges that will have to be discussed in greater detail elsewhere.

3.

How Can Clauses Act as Predicates of Attitudinal Objects?

So far, one important question has been left open; namely, what properties should that-clauses express so that they can act as predicates of attitudinal objects? As was mentioned, when predicated of an attitudinal object, a that-clause will specify the satisfaction conditions of the attitudinal object; it will thus express a property of attitudinal objects. For formulating the meaning of a that-clause for that purpose, possibleworlds semantics would not be a viable approach. Possible-worlds semantics is not able to provide a single property applicable to both attitudinal objects that come with the modal force of necessity and attitudinal objects that come with the modal force of possibility, say a demand and a permission (Moltmann 2018b). The most plausible property of attitudinal objects that a that-clause that S should express given possible worlds semantics is λd[∀w(w ∈ f(d)  S is true in w)], for f(d) being the set of

190

Friederike Moltmann

worlds compatible with d. But this property would be suited only for attitudinal objects with the force of necessity, not possibility. (John gave permission for Mary to leave does not require that all worlds compatible with the permission are worlds in which Mary leaves; just some worlds need to be that way.) By contrast, truthmaker theory, along the lines of Fine’s (2017) recent truthmaker semantics, allows formulating a single property applicable to attitudinal objects of both sorts (Moltmann 2018a, b). Truthmaker theory in its application to attitudinal objects is based not on entire worlds, but on situations or actions standing in the relation ╟ of exact truthmaking or exact satisfaction to a sentence or attitudinal object. More precisely, an attitudinal object d will be associated with exact truthmakers or satisfiers (entities s that stand in the relation ╟ to d) as well as exact falsemakers or violatos (entities that stand in the relation ╢ to d). Unlike possible-worlds semantics, truthmaker theory allows for a straightforward notion of partial content, and the idea then is that thatclauses specify a partial content of the attitudinal object to which they apply. Formally, this is given below: (24) Sentence Meanings as Properties of Attitudinal Objects [S] = λd[∀s(s ╟ d → ∃s’(s’╟ S & s < s’) & ∀s’(s’╟ S  ∃s(s ╟ d & s < s’)) & ∀s(s ╢ S→ s ╢ d, in case neg(d) ≠∅] That is, a sentence S expresses the property of attitudinal objects d such that every satisfier of d is part of an exact satisfier of S, and every satisfier of S contains a satisfier of d as part, and moreover all violators of S are also violators of d, if d has violators. Crucially, attitudinal objects of possibility (permissions, offers, invitations) have only satisfiers, but no violators, which is what accounts for the difference between attitudinal objects of the two different forces.

4.

The Roles and Properties of Attitudinal Objects

Attitudinal objects are not only well reflected in the semantics of attitude reports and nominalizations of attitude verbs but, arguably, also play a central role in our mental life as well as in communication. In their status as concrete content bearers, they naturally serve as objects of various forms of content-based causation. This is well reflected in the way we use causal predicates with attitudinal objects as opposed to with the corresponding actions (Moltmann 2013a 2014, 2017a). If Mary’s claim caused a commotion, this implies that the content (conveyed by Mary) was causally responsible; by contrast, if Mary’s speech act caused a commotion, this implication does not hold. If an answer caused surprise, this implies that the content was the subject of surprise; but not so if an act of answering caused surprise. Also,

Attitudinal Objects

191

mental attitudinal objects engage in content-based causation. A decision may cause an action on the part of the agent, and that can only be in virtue of its content. This is not so for a mental act of deciding (whose exhausting nature may be the trigger of an act of taking a break from further decision-making). Propositions as abstract objects cannot play causal roles and thus leave content-based causation a puzzling phenomenon. Mental attitudinal objects also act as the targets of content-related memory. We remember thoughts, beliefs, decisions, and intentions, rather than propositions. We may remember acts of thinking or acts of deliberating without recalling their content, and thus this would not be content-related remembering. In addition to the roles of attitudinal objects for content-based causation and remembering, attitudinal objects have properties relating only to their contents. There are three important types of content-related properties of attitudinal objects: (a) Attitudinal objects have truth conditions or, more generally, satisfaction conditions. John’s claim and John’s judgment may be true or false, as may be John’s belief. By contrast, this does not intuitively hold for acts: a speech act or an (act of) claiming cannot intuitively be true or false, and neither can an act of judging. It also fails to hold for mental states described as such: a belief state is not something one would naturally say is true or false, but a belief is. Other attitudinal objects do not have truth conditions, but rather satisfaction conditions. Thus, a request can be fulfilled or ignored, a decision implemented, a command executed. Even more so than truth predicates, entities described as acts or states resist predicates of satisfaction. An act of requesting (or a speech act) cannot be fulfilled, an act of deciding (or a mental act) cannot be implemented, and an act of command cannot be executed. Even more striking are contrasts with agent-related predicates. Advice can be followed, but an act of advising cannot be followed in that same sense. A recommendation can be taken up or ignored, but an act of recommending cannot, at least not in the same sense. Again, mental states described as such are not bearers of satisfaction conditions. A state of desiring or hoping cannot be fulfilled, but a desire or a hope can. Attitudinal objects generally come with inherent truth or satisfaction conditions of some sort or other, but acts and entities described as states do not. (b) Attitudinal objects that are of the same sort (involving the same kind of physical realization and force) enter similarity relations strictly on the basis of being the same in content. The relation of exact or close similarity in natural language is conveyed by is the same as. Thus, (25a) says that John’s and Mary’s beliefs are the same in content: (25) John’s belief is the same as Mary’s

192

Friederike Moltmann

The sentences below, by contrast, sound false, and that is because a thought and a remark do not involve the same physical realization, and a hope and a claim do not have the same force: (26) a. ?? John’s thought is the same as his remark. b. ??? John’s hope is the same as Mary’s claim. Is the same as does not apply in that way to actions or states. For two actions or states to be the same, they need to share features of their performance or constitution (if it makes even sense to apply is the same as to them); sameness of content will not be enough, as the contrast between (27a) and (27b) makes clear: (27) a. John’s thought is the same as Mary’s. b. John’s thinking is the same as Mary’s. Thus, for exact similarity to obtain, two attitudinal objects need to be of the same type and share their content. (c) Attitudinal objects have a part structure based strictly on partial content, not the temporal part structure of events or states. This is most obvious from the way part of is understood when applying to an attitudinal object. ‘Part of John’s decision’ cannot be ‘part of the action of deciding’. ‘Part of John’s claim’ cannot be ‘part of the speech act of claiming’. ‘Part of John’s answer’ cannot be ‘part of John’s act of answering’. Similarly, ‘part of John’s belief’ and ‘part of John’s hope’ can be only partial contents. It is remarkable that even physically realized attitudinal objects fail to have a physical part structure. They differ in that respect from physically realized artefacts like books. The book as a materially realized artefact has two part structures at once. The parts of a book as an information object are partial contents, the parts of the physical copy material parts. ‘Recalling a part of the book’ can mean recalling either a part of the information object or a part of the physical object; recalling part of the claim can mean recalling only a partial content. To sum up, attitudinal objects are characterized by two types of properties: properties of content-based causation, remembering, and evaluation and pure properties of content. The next question then is, how should one make sense of attitudinal objects ontologically? One approach is to assimilate them to an already familiar ontological category, such as that of an abstract artefact. Twardowski’s distinction between actions and product can be considered an attempt in that direction.

5.

The Action-Product Distinction

Twardowski (1912) was an early analytic philosopher who, in opposition to Frege, argued for a mind-dependent notion of a truth-bearer, one that

Attitudinal Objects

193

would not be subject to the objections to psychologism that were around at the time. Twardowski basically argued for the notion of an attitudinal object, but in the context of a distinction he drew between actions and products. Twardowski’s action-product distinction follows closely the linguistic distinction between two sorts of nominalizations in natural language. While Twardowski focussed on Polish, German, and French, the same kind of distinction is present in English. In English, terms for actions are generally gerunds, whereas terms for products are other, simple or derived, nominalizations. Thus, pairs of terms for actions and products are thinking–thought, judging–judgment, believing–belief, claiming– claim, deciding–decision, demanding–demand, screaming–scream. For Twardowski, the action-product distinction includes mental actions and their products as well as illocutionary actions and their products.4 Twardowski took products of actions like thinking, claiming, judging, deciding, and demanding to be on a par with material products like a piece of writing as a product of an act of writing or a drawing as a product of an act of drawing. The latter differ from the former only in having a material realization, which enables them to endure beyond the act that produced them, whereas the former, for Twardowski, last only as long as the action producing them. Products are as agent-dependent as actions, but, crucially, they enter similarity relations on the basis of shared content only. Importantly, distinct products, dependent on different agents or pertaining to different times, can share their content, in which case they are similar. Twardowski distinguished actions and their products in terms of the predicates that can be true of them, without, though, being very systematic about the range of types of predicates. Most importantly, predicates of truth and satisfaction can be true or false of products, but not of actions, and thus products but not actions are bearers of truth or satisfaction conditions. Products last only as long as the actions that produce them. However, products allow for similar products to be produced at a later point in time, which gives rise to the emergence of an apparent enduring content. For two agents to share the same propositional content, they must be engaging in actions that produce similar products. Thus, products, on Twardowski’s view, are particulars that are concrete and depend on a specific agent, yet have important content-related properties—in particular truth or satisfaction conditions—and the ability to stand in similarity relations based on shared content. Twardowski was not explicit about how products are to be viewed ontologically. However, a very plausible interpretation of the action-product distinction is that products are the non-enduring artefacts produced by the actions, that is, ‘abstract’ artefacts in Thomasson’s (1999) sense. Artefacts, in general, are considered mind-dependent objects that may lack a physical or material realization (e.g. poems or musical compositions that

194

Friederike Moltmann

have not been written down) and thus may fail to endure. Artefacts may have physical properties as well as content-related properties. Books, for example, are artefacts that come with two distinct facets, as physical objects and as information objects. Artefacts thus appear to share characteristic properties with attitudinal objects. However, there are types of attitudinal objects that do not fare well with the action–produced artefact distinction. First of all, attitudinal objects such as beliefs, hopes, intentions, and desires cannot generally be viewed as products of actions. Attitudinal objects such as beliefs may be produced by an action, but need not be. Intentions are states that are presupposed by the intentional action set out to realize them and could not be produced by an intentional action themselves on pain of regress (Searle 1983). While attitudinal objects such as beliefs, desires, hopes, and intentions are generally considered mental states, the notion of a state is not actually suited for them, as will be discussed in the next section. Let me therefore call such attitudinal objects instead state-related attitudinal objects. Another difficulty for Twardowski’s notion of a product is the existence of a category of objects closely related to attitudinal objects—namely, modal objects. Modal objects are, for example, needs, obligations, permissions, invitations, offers, and abilities. Modal objects may exhibit features of concreteness, in particular having a limited lifespan and perhaps being causally efficacious.5 Most importantly, modal objects exhibit the same content-related properties as attitudinal objects (having satisfaction conditions, standing in similarity relations based on shared content only, having a part structure strictly based on partial content) (Moltmann 2017a, 2018a). While ‘heavy’ (or explicit) obligations and permissions, to use von Wright’s (1963) term, are generally products of acts (of demanding or permitting), this is not the case for ‘light’ (or implicit) permissions and obligations, various sorts of needs, and abilities (Moltmann 2017c, 2018a). State-related attitudinal objects as well as modal objects exhibit the very same characteristic features as attitudinal objects that can be considered products of acts. Since they cannot generally be regarded products of acts, those characteristic features, and the notion of an attitudinal object as such, cannot be traced to the nature of a product as an artefact. The notion of an artefact thus is not illuminating as regards the nature of attitudinal objects. In addition to state-related attitudinal objects, there are also act-related attitudinal objects that do not fare well with the action-product distinction understood as the distinction between an action and the produced artefact. These are attitudinal objects that are not entities agents generally intend to produce with the action in question. Attitudinal objects associated with eventive epistemic verbs, for example, can hardly be considered artefacts produced by a mental action. A recognition that S

Attitudinal Objects

195

and a realization that S are not the intended products of epistemic acts; rather, they are occurrences that, if anything, may have answered a state or act of inquiry. A particular conclusion is not the intended product of an act of reasoning; the act of reasoning may have as its intended product only some conclusion or other, but not a specific one. Moreover, any act describable as an act of concluding that S is individuated by the conclusion reached, not by the mental activity pursued as such. That is, an act of concluding depends for identity on the conclusion reached, rather than the conclusion depending for its identity on an intentional action being performed. In the area of speech acts, the same is the case for perlocutionary acts, such as an act of persuading or an act of achieving an emotional effect on an audience by performing an illocutionary act. An act of persuading is individuated by the effect it happens to have, the persuasion, not by realizing a type of action. Thus, the act of persuading someone that S depends for its identity on the attitudinal object that is the persuasion that S, rather than the attitudinal object depending on the act. Even an act of judging is of that sort. A judgment that S is not the realization of an intentional action but what an agent arrives at when evaluating a thought (or propositional content). For actions of recognizing, realizing, concluding, persuading, and judging, it is the attitudinal object that individuates the action that culminates in its, rather than the attitudinal object’s, being individuated as the intended product of the action.

6.

The Distinction Between Attitudinal Objects and States

While beliefs, intentions, hopes, and desires are not (or not generally) the products of actions, the notion of a state, as standardly understood, is not suitable for them either. This became already clear in the discussion of the characteristic properties of attitudinal objects. States, at least as entities we refer to as states, do not have the properties that attitudinal objects such as beliefs, intentions, hopes, and desires have. A mental state (of believing) is not intuitively something that could be true or false, but a belief certainly is. A state (of intending) cannot be realized, but an intention can. A state of hoping or desiring cannot be fulfilled, but a hope or desire can. The parts of a mental state are not intuitively partial contents, but the parts of beliefs, intentions, hopes, and desires clearly are. A part of a mental state is a temporal part, or perhaps better, a condition partly constitutive of the state (a condition that, together with others, obtains while the state endures). Also, similarity relations do not apply to states in the way they apply to attitudinal objects. Two mental states (of the same type) are not just the same if they are the same in content. Rather, our intuitions about John and Mary’s mental states being the same are that their constitutive features or conditions (including strength of the attitude) are the same.

196

Friederike Moltmann

A state is constituted by a condition enduring in time, which appears to be why states have such fundamentally different sorts of properties from attitudinal objects like beliefs, desires, and fears. That holds even though generally, for a state-related attitudinal object, there appears to be a corresponding state that obtains under the same circumstances (John’s belief that S and the state of John’s believing that S appear to exist under the same circumstances, and so for John’s desire that S and John’s desiring that S and for John’s fear that S and John’s fearing that S).6 There is one further intuitive difference between actions and states on the one hand and attitudinal objects on the other. Events (actions) and states are often identified with space-time regions or property instantiations in space-time regions, which means that events and states have their time of occurrence essentially. By contrast, the time of occurrences seems not to be essential for attitudinal objects. A thought or a scream could easily have occurred earlier than it did, and a promise could have been made later than it was.7

7.

State-Related Attitudinal Objects and Abstract States as Davidsonian Arguments

If state-related attitudinal objects are, in fact, distinct from the correlated states, this raises a question for the semantics of attitude reports; namely, what should the Davidsonian event argument of stative attitude verbs be, a state-related attitudinal object or the correlated state? This will have to depend on how adverbials behave since adverbials are considered predicates of the Davidsonian argument. Are adverbials of state-related attitude verbs to be understood as predicates of state-related attitudinal objects or of the correlated states? The data indicate that they are better taken to be predicates of the correlated states. First, at least certain temporal adverbials, such as for two weeks, in attitude reports as below need to be understood as predicates of states rather than attitudinal objects: (28) For two weeks, John believed that he would make the deadline The standard view about the semantics of for two weeks predicts such adverbials to apply to states but not attitudinal objects. For-adverbials are generally taken to require a homogeneous event predicate (or a predicate satisfying a closely related condition, cf. Krifka 1989; Moltmann 1991). A (one-place) event predicate P is homogeneous just in case the sum of any two events falling under P again falls under P and any temporal part of an event falling under P also falls under P. Homogeneity requires that the Davidsonian arguments of the predicate have temporal parts, but attitudinal objects have only content-related parts. Homogeneity also requires Davidsonian arguments to form fusions in time, but attitudinal objects

Attitudinal Objects

197

form fusions that are based on a merging of content, not fusions of two temporal entities (Moltmann 2018a). Thus, for two weeks in (28), given the standard view, requires the Davidsonian event argument of believe to be a state, rather than a belief. The restriction of for-adverbials to states is also reflected in their applicability to nominalizations (her week-long mental state is better than her week-long belief). Another piece of evidence that attitudinal objects are not Davidsonian event predicates comes from truth predicates across languages. Across languages, it appears, the adverbial versions of truth predicates do not generally attribute truth to the Davidsonian argument. Whereas in English truly (and correctly) does, in fact, convey the truth of the described attitudinal object, in other languages—for example, German, French, and Italian—the adverbial version of true in general does not convey the truth of the relevant attitudinal object, but rather the reality of the propositional attitude obtaining (Moltmann 2017a). Thus, the German, French, and Italian examples in (29a, b, c) are translations of (30a), not of (30b): (29) a. Hans glaubt wahrlich, dass S. b. Jean crois vraiment que S. c. Gianni crede veramente que S. (30) a. John really believes that S. b. John truly believes that S. The data indicate that the English adverbial truly does not share its meaning with ‘true’, but has a more derivative meaning, stating the truth of the attitudinal object associated with the Davidsonian argument, rather than of the Davidsonian argument itself. Given such linguistic indications, it appears then that the Davidsonian argument of stative attitude verbs is, in fact, a state distinct from the attitudinal object. In the case of John believes that S, this would be the state of John’s having the belief that S, the state constituted by the bearerhood relation between John and the belief that S. This would mean that believe that S and have the belief that S have the same meaning, and also that the Davidsonian argument of have itself is a state, the state of standing in the relation of possession or bearerhood to the object argument.8 The entity that is John’s believing that S will then correctly come out as distinct from John’s belief that S. John’s believing that S would be John’s having the belief that S, rather than John’s belief that S.9

8.

Conclusion and Outlook: The Ontology of Attitudinal Objects

Even though attitudinal objects are not generally recognized as an ontological category in contemporary philosophy, they are extremely well reflected in natural language as well as in our general intuitions about

198

Friederike Moltmann

mental attitudes and speech acts. They are thus part of the ontology we implicitly accept, the ontology that is the subject matter of descriptive metaphysics in Strawson’s (1959) sense and more specifically natural language ontology (Moltmann 2017d). Attitudinal objects are characterized by a range of types of properties that jointly distinguish them from other ontological categories, such as actions, events, and states. Most importantly, attitudinal objects play the roles of bearers of truth conditions (or satisfaction conditions) and of inferential relations. By entering similarity relations strictly on the basis of a shared content and forming corresponding kinds, they allow for an account of the sharing of content. Attitudinal objects also play important roles in the ontology of the mind as the objects involved in content-based causation and remembering. Attitudinal objects are most obviously the semantic values of (nongerundive) nominalizations of attitude verbs, but they also serve as semantic values of special quantifiers and pronouns and arguably play a central role in the semantics of attitude reports not involving explicit reference to them. Nothing has been said in this paper about the structure or composition of attitudinal objects and, in fact, whether attitudinal objects even have a structure. The overall view of attitudinal objects this paper has presented certainly allows an attitudinal object to have satisfaction conditions without being associated with a structure. There is no reason to assume that state-related attitudinal objects come with a structure. Even more obviously, state-related modal objects do not come with a structure—objects that include light obligations and permissions, needs, and abilities. The part structure of attitudinal and modal objects is, in fact, based on partial content only, rather than, say, the structure of acts that may be involved in their creation or a temporal part structure.10 In that respect, attitudinal and modal objects differ from artefacts such as books, which have a physical part structure as well as a content-related part structure. The fact that a strictly content-related and thus abstract part structure is compatible with attitudinal objects having properties of concreteness is itself in need of explanation. Attitudinal objects may be objects of perception and enter causal relations, and they generally have a limited temporal lifespan. But their part structure is not related to those aspects of concreteness. The fact that the part structure of attitudinal objects is entirely independent of their features of concreteness does not cohere with standard assumptions about objects and their parts. A temporally extended concrete object should be able to have parts that are concrete as well, but that is not the case with attitudinal objects. This puzzle may be related to the status of attitudinal objects as ontologically dependent objects, dependent on the mind of an agent. It appears that ontologically dependent concrete objects may, in general, display gaps in property spaces that other concrete objects do not

Attitudinal Objects

199

display. Some ontologically dependent objects may lack an independent spatial location—for example, holes. A hole may be in the bag and the bag may be on the table, but the hole is certainly not on the table. Holes have a location relative to the object on which they depend, but not absolutely. Another example is tropes. The painting may have great beauty and be on the wall, but the beauty of the painting won’t be on the wall. Tropes, moreover, may have a part structure that is entirely independent of the spatial part structure of their bearer. The beauty of the painting may have as a part a particular colour combination, for example, but no parts that relate to its spatial location relative to its bearer. Similarly, the failure of attitudinal objects to have a physical part structure may be considered a gap in a property space that ontologically dependent objects in general may exhibit. Attitudinal objects as well as certain other types of ontologically dependent objects thus would not require the sort of completion under property specification that the standard view takes ontologically independent concrete objects to be subject to. Of course, this requires a revision of common assumptions about object individuation that will have to be pursued in greater detail elsewhere.

Acknowledgements Previous versions of this paper were presented at the Workshop on Judgment at the New College of the Humanities in 2016, organized by B. Ball and C. Schuringa; at the workshop Propositions and Linguistic/Cognitive Action in Duesseldorf in 2017, organized by G. Vosgerau; and the conference Objects and Properties: Generating Dialogue, Cambridge University, in 2017, organized by C. Rossi. I would like to thank the audiences for very stimulating discussions. I would also like to thank Brian Ball and Chris Peacocke for comments on an earlier version of this paper and Kit Fine and Bob Matthews for many conversations on the topic.

Notes 1. See, for example, Pustejovsky (1995). 2. The observation about satisfaction predicates when applied to nominalizations of illocutionary verbs was made by Ulrich (1976), who argued that claims, demands, and promises are objects sui generis. Twardowski (1912) already gave various examples with different sorts of attitudinal objects. 3. For the notion of partial content, see Yablo (2015) and Fine (2017), as well as section 3. 4. In addition, and more problematically, Twardowski assumed that even physical actions can come with a product (thus, an action of walking has as its product a walk and an action of jumping a jump). See Moltmann (2017a) for a critique of the physical action-product distinction. 5. Though modal objects have the ability to endure beyond the act that may have created them (Moltmann 2017a).

200

Friederike Moltmann

6. For further arguments against a conception of beliefs and other state-related attitudinal objects as states, see Machery (2016, 2017). Machery instead considers attitudinal objects like beliefs to be ‘traits’ on a par with courage, that is, as dispositions of a sort. 7. It appears at least less natural to say that about a process of thinking and particular acts of screaming or promising. 8. This would be parallel to sick vs be sick: sick takes a trope as an argument, but be sick a derivative state of being a bearer of the trope. See Moltmann (2015) and Maienborn (forthcoming). 9. There is one issue, and that is predicates of intensity, which do apply to believe but are not predicates of abstract states (John firmly believes that S) (Maienborn forthcoming). I will leave this as a puzzle for future research. 10. In Moltmann (2017b), I argued that attitudinal objects may come with a structure when they are products of locutionary acts, acts below the level of locutionary acts. Examples are thoughts, remarks, and screams.

References Carlson, G. (1977), “A Unified Analysis of the English Bare Plural”, Linguistics and Philosophy 1: 413–57. Davidson, D. (1967), “The Logical Form of Action Sentences”, in N. Rescher (ed.), The Logic of Decision and Action. Pittsburgh: Pittsburgh University Press. Fine, K. (2017), “Truthmaker Semantics”, in B. Hale, C. Wright and A. Miller (eds.), A Companion to the Philosophy of Language. Chichester: Wiley. Hanks, P. W. (2011), “Structured Propositions as Types”, Mind 120: 11–52. Jubien, M. (2001), “Propositions and the Objects of Thought”, Philosophical Studies 104: 47–62. Krifka, M. (1989), “Nominal Reference, Temporal Constitution, and Quantification in Event Semantics”, in R. Bartsch et al. (eds.), Semantics and Contextual Expressions. Dordrecht: Foris. Machery, E. (2016), “De-Freuding Implicit Attitudes”, in M. Braunstein and J. Saul (eds.), Implicit Bias and Philosophy. Oxford: Oxford University Press. Machery, E. (2017), “Do Indirect Measures of Biases Measure Traits or Situations?”, Psychological Inquiry 28(4): 288–91. Maienborn, C. (forthcoming), “Events and States”, in R. Truswell (ed.), Handbook of Event Structure. Oxford: Oxford University Press. Moltmann, F. (1991), “Measure Adverbials”, Linguistics and Philosophy 14(6): 629–60. Moltmann, F. (2003a), “Propositional Attitudes Without Propositions”, Synthese 135: 70–118. Moltmann, F. (2003b), “Nominalizing Quantifiers”, Journal of Philosophical Logic 35(5): 445–8. Moltmann, F. (2013a), Abstract Objects and the Semantics of Natural Language. Oxford: Oxford University Press. Moltmann, F. (2013b), “On the Distinction Between Abstract States, Concrete States, and Tropes”, in A. Mari, C. Beyssade and F. Del Prete (eds.), Genericity. Oxford: Oxford University Press. Moltmann, F. (2014), “Propositions, Attitudinal Objects, and the Distinction Between Actions and Products”, Canadian Journal of Philosophy 43(5–6): 679–701.

Attitudinal Objects

201

Moltmann, F. (2015), “States vs Tropes. Commentary on Marcyn Morzicki: ‘Degrees as Kinds of States’”, Natural Language and Linguistic Theory 33(3): 829–41. Moltmann, F. (2017a), “Cognitive Products and the Semantics of Attitude Reports and Deontic Modals”, in F. Moltmann and M. Textor (eds.), Act-Based Conceptions of Propositions: Contemporary and Historical Contributions. Oxford: Oxford University Press. Moltmann, F. (2017b), “Levels of Linguistic Acts and the Semantics of Saying and Quoting”, in S. L. Tsohatzidis (ed.), Interpreting Austin: Critical Essays. Cambridge: Cambridge University Press. Moltmann, F. (2017c), “Partial Content and Expressions of Part and Whole. Discussion of Stephen Yablo: Aboutness ”, Philosophical Studies 174(3): 797–808. Moltmann, F. (2017d), “Natural Language Ontology”, in Oxford Encyclopedia of Linguistics . http://oxfordre.com/linguistics/view/10.1093/acrefore/ 9780199384655.001.0001/acrefore-9780199384655-e-330 Moltmann, F. (2018a), “An Object-Based Truthmaker Theory for Modals”, Philosophical Issues 28(1): 255–88. Moltmann, F. (2018b), “Clauses as Semantic Predicates: Difficulties for PossibleWorlds Semantics”, in R. Bhatt, I. Frana and P. Menendez-Benito (eds.), Making Worlds Accessible. Festschrift for Angelika Kratzer. Amherst: University of Massachusetts (online). Pustejovsky, J. (1995), The Generative Lexicon. Cambridge, MA: MIT Press. Searle, J. (1983), Intentionality. Cambridge: Cambridge University Press. Soames, S. (2010), What Is Meaning? Princeton: Princeton University Press. Strawson, P. (1959), Individuals: An Essay in Descriptive Metaphysics. London: Methuen. Thomasson, A. (1999), Fiction and Metaphysics. Cambridge: Cambridge University Press. Twardowski, K. (1912), “Actions and Products. Some Remarks on the Borderline of Psychology, Grammar, and Logic”, in J. Brandl and J. Woleński (eds.), Kazimierz Twardowski: On Actions, Products, and Other Topics in the Philosophy. Amsterdam and Atlanta: Rodopi. Ulrich, W. (1976), “An Alleged Ambiguity in the Nominalizations of Illocutionary Verbs”, Philosophica 18(2): 113–27. von Wright, G. H. (1963), Norm and Action: A Logical Inquiry. London: Routledge and Kegan Paul. Yablo, S. (2015), Aboutness. Cambridge, MA: MIT Press.

10 About vs Concerns Daniel Morgan

I see the familiar form of a tiger emerge out of the long grass. I judge about it ‘That is a handsome animal’. Suppose the tiger is Tony Tiger. Tony is something my judgment is about. He is an object of my judgment. Tony also figures in the truth conditions of the judgment. My judgment is true if and only if Tony really is a handsome animal.1 How, in general, do claims about what a judgment is about, or claims about what its object is, relate to claims about what its truth conditions are? Focussing on the example just mentioned, a naïve answer is that they are equivalent. ‘Tony is something the judgment is about’ is equivalent to ‘Tony is something the truth of the judgment depends on’. We can put pressure on the naïve answer by focussing on three things other than Tony that the truth of my ‘That is a handsome tiger’ judgment appears to depend on. In each case, it doesn’t seem quite as clear that the judgment is really about that thing as it is that the truth of the judgment depends on that thing. First, in the actual world, Tony is handsome. In a different possible world, in which his hair has been badly singed, he isn’t. Pointing this out does nothing to undermine the idea that the ‘That is a handsome animal’ judgment I made is true (rather than false or neither true nor false). So the truth conditions of that judgment seem to involve the actual world rather than the different possible world in which Tony has been singed. My judgment is true if and only if Tony Tiger is handsome at the actual world. If the naïve answer just given is right, it follows that my judgment is about the actual world. The actual world is an object of the judgment. Second, at the moment of judgment, Tony is handsome. At some earlier point of time, before his hair had grown out after his most recent singing, he was not. My ‘That is a handsome animal’ judgment still seems true. So its truth conditions must involve the moment of judgment rather than the earlier moment when his hair hadn’t come in yet. If the naïve answer is right, it follows that the ‘That is a handsome tiger’ judgment is about the moment of judgment. The moment of judgment is an object of the judgment.

About vs Concerns

203

Third, suppose that ‘handsome’ is context-sensitive in the following way: the extension of ‘is handsome’, as used by me, is those things that are disposed to produce a particular kind of excited shiver in me and my peers. In that case, whether my ‘That is a handsome animal’ judgment is true depends on how me and my peers are (do we experience that kind of shiver when exposed to Tony?). If the naïve answer is right, it follows that the ‘That is handsome tiger’ judgment is about me and my peers. I and my peers are an object of the judgment. In this paper, I canvass the idea that there is no theoretically interesting difference between two ways of being involved in the truth conditions of a judgment that the example we’ve just looked at illustrates. So, if we say, with the naïve answer, that the judgment is indeed about not just Tony but also the actual world, the current moment, and me and my peers and leave it at that, there isn’t any theoretically interesting distinction we’ve failed to draw. There are, I think, all sorts of different differences that the example might be taken to illustrate. I will outline some of them. But, on the face of it, there does not appear to be any single distinction that unifies them in the way that has been widely and, I’ll argue, harmfully assumed. Suppose we label the kind of involvement Tony has in the truth conditions of the ‘That is a handsome animal’ judgment ‘about’-involvement and the type of involvement the actual world, the present moment, and I and my peers have in the judgment ‘concerns’-involvement. I am happy to accept that, even just with the distinction introduced by reference to these examples, there might be some level of agreement about how to apply the distinction in at least some new cases. For example, consider the property of being an animal. I predict that, with the ‘about’/‘concerns’ distinction explained by reference to the examples already given, most people will be fairly confident that the property of being an animal should fall on the ‘about’ side (along with Tony) rather than the ‘concerns’ side (along with the actual world, the current moment, and me and my peers). That alone isn’t a significant concession to the idea of an about/concerns distinction. We can compare Quine’s attack on the analytic/synthetic distinction and Grice and Strawson’s (1956) response to it. Grice and Strawson say the following: Those who use the terms ‘analytic’ and ‘synthetic’ do to a very considerable extent agree in the applications they make of them. They apply the term ‘analytic’ to more or less the same cases, withhold it in more or less the same cases, and hesitate over more or less the same cases. . . . In general, if a pair of contrasting expressions are habitually used in application to the same cases, where these cases do not form a closed list, this is a sufficient condition for saying that there are kinds of cases to which the expressions apply; and nothing more is needed for them to mark a distinction. (1956: 142)

204

Daniel Morgan

Suppose we accept that Grice and Strawson have pointed to some real data—e.g. most people will agree, without being drilled in these two examples, that ‘All vixens are foxes’ is a better candidate for being analytic than ‘All vixens are wily’. Suppose we also accept that it follows that there is some distinction that people’s use of ‘analytic’ and ‘synthetic’ is tracking. It doesn’t follow that there is a theoretically interesting distinction that people are tracking. It doesn’t even follow that there is a single distinction that different people who use ‘analytic’ and ‘synthetic’ are tracking. One person’s use of ‘analytic’ or ‘synthetic’ might be keyed to whether the truth in question is one they find utterly boring. Another person’s use might be keyed to whether they regard the truth as being close to the core rather than the periphery of our web of beliefs. These two people’s uses of ‘analytic’/‘synthetic’ are keyed to different distinctions (and, arguably, neither is keyed to a distinction that has the significance the analytic/synthetic distinction was supposed to have). But one can see how they might habitually use the terms in application to roughly the same cases of the sort Grice and Strawson think significant. Similarly, even if people habitually agree in their application of ‘concerns’/‘about’ to some degree—and we’ll see that there isn’t much agreement in the application of the distinction to cases; to that extent, the about/concerns distinction seems worse off than the analytic/synthetic distinction—it doesn’t follow that there is some single interesting distinction they are tracking. Why does it matter if there is a single distinction that is being tracked by theorists’ use of ‘about’/‘concerns’? That, I think, follows from a plausible account of what it takes for an about/concerns distinction to be explanatorily useful. Mark Sainsbury (2002) canvasses a parallel account in relation to the Fregean notion of ‘sense’. Sainsbury thinks the following about the term ‘sense’: if ‘sense’ is stipulatively tied to a single phenomenon (e.g. “If a rational subject simultaneously believes, of a given thing, that it is F and it is not F, then she must have two different senses via which she apprehends that thing”) and nothing further is said, then we can’t explain that very phenomenon by reference to the notion of sense (“She could believe that a is F and a is not F without being irrational because she deployed two different senses”). That would be like saying that a pill causes sleep because it has a dormitive virtue. This isn’t an explanation because our only handle on what a dormitive virtue is is that it is something that causes sleep. The notion of ‘sense’ will earn its explanatory spurs, assuming it has some, by unifying different phenomena—e.g. rational co-tenability of belief and disbelief in the same coarse-grained proposition, on the one hand, and intersubjective communication, on the other. As Sainsbury puts it: The explanatory value of sense lies in its putative subsumption of a number of apparently distinct phenomena. But if we take any one

About vs Concerns

205

phenomenon, . . ., to invoke difference of sense is simply to redescribe the phenomenon in the theorist’s preferred vocabulary. (2002: 127) We have more than one ‘hook’ onto the notion of sense, and the interesting thing is that they are both hooks onto the same thing. The idea in the current paper is to see whether there is any distinction between two different ways an element can be involved in the truth conditions of a judgment that fares well on this kind of test. If we introduce the about/ concerns distinction using one ‘hook’, can we infer anything interesting about where an element stands in relation to some other ‘hook’, with reference to which we might equally have introduced it? If we can, then the distinction has a clear explanatory use—the same kind of use in subsuming apparently different phenomena that, if Frege and Sainsbury are right, the notion of ‘sense’ has. If we can’t, then to invoke the about/ concerns distinction is simply to redescribe a certain phenomenon in the theorist’s preferred vocabulary (at best) or (at worst) to conflate different differences. I’ll begin by mentioning two target authors for this paper, which should also give a more general flavour of the kind of examples those drawing an about/concerns distinction have been trying to do justice to. Perry distinguishes things that a mental state is about and things it merely concerns; the label, as I am using it, originates with him (see 1986: 213). A famous example of his involves Z-landers, people who live in a small isolated community in a place called Z-land and never leave that place. They speak and think about rain, but since they never leave Z-land, they never have use for the judgment ‘It is raining’ outside of Z-land. The truth conditions for that judgment (and for utterances of that sentence) involve Z-land. When a Z-lander judges ‘It is raining’, what she judges is true if and only if it is raining in Z-land. But, Perry wants to say, the judgment is not about Z land, it merely concerns it. As he also puts it, Z-land is an unarticulated constituent of the judgment. Other examples of Perry’s include the thought about time of a four-year-old who hasn’t heard of time zones or the thought about seasons of Northern Hemisphere inhabitants who don’t understand that what season it is at a given time depends on which hemisphere you’re in. The four-year-old in Sacramento judges ‘It is 4 pm’, and the judgment is true if and only if it is 4 pm in the time zone that Sacramento is in even though the four-year-old is not thinking about a time zone and has no articulated constituent in his judgment that refers to a time zone. The Northern Hemisphere inhabitant judges ‘July is a summer month’, and the judgment is true if and only if July is a summer month in the Northern Hemisphere. But they are not thinking about the Northern Hemisphere. Kriegel (2015) distinguishes between things that are part of the content of a mental state vs those that are supplied by the attitude (or mode)

206

Daniel Morgan

of the state. For example, citing Brentano, he distinguishes the state of believing in ghosts from the state of believing that ghosts exist. The truth conditions of the two states are the same—both are true if and only if ghosts exist. But there is still a difference. In the first case, the content is just ghosts, while the commitment to their existence is supplied by the attitude type of the state (the attitude here is something like ‘believing in’). In the second case, by contrast, both the ghosts and the commitment to their existence are part of the content (the attitude here is something like ‘believing that’).2 The about/concerns distinction applies to judgments, but it doesn’t just apply to them. For example, as the Kriegel case brought out, the distinction applies to beliefs. And it also applies to states that are more naturally thought of in terms of satisfaction conditions than truth conditions. For example, some have wanted to distinguish John’s intention to dance from John’s intention that he himself dance using that distinction. Both intentions have the same satisfaction conditions: each is satisfied if and only if John dances. Neither involves John thinking about himself ‘as someone else’ or ‘as another’ in such a way that John would not be able to answer ‘me’ if asked who needs to dance for the intention to be satisfied. But the first intention is about the act of dancing and concerns John, whereas the second intention is about the act of dancing and about John. In some sense, the first intention is more thoroughly subject-involving: its involving its subject is built into its being the kind of state that it is and John being its subject, so that John does not even need to be something the intention is about to be relevant to its evaluation. By contrast, involving the subject is not built into the second’s being the kind of state that it is. John could have a state of the same kind about someone else. For example, John’s intention that James dance would be a state of the same kind about someone else.3 We can distinguish those kinds of mental states that have the kind of directedness that makes it possible to evaluate them as true, or as satisfied, from those that do not.4 Pains and pleasures are arguably in the second category. Judgments, beliefs, desires, and intentions are in the first category. The ‘about’/‘concerns’ distinction applies to things in the first category. As applied to things in the first category that have truth conditions, like judgment, it’s a distinction between two ways in which something can be involved in the state’s truth conditions—by being about it and the judgment concerning it. As applied to things in the first category that have satisfaction conditions, it’s a distinction between two ways in which something can be involved in the state’s satisfaction conditions. Even so, the case of judgment is, in this broader area, especially illuminating. Judgments ought to be a particularly clear case for the distinction since it’s so natural to think that we individuate judgments by their truth conditions. Conversely, much of the historical and contemporary controversy about judgment in particular has focussed on the right way of thinking of issues concerning the ‘object’ and/or ‘content’ of judgment.5

About vs Concerns

207

So, in getting the right account of judgment, it’s especially important to have a firm grip on terms like ‘object’, ‘content’, and ‘aboutness’. This paper is intended to clear up some of the rules of the game concerning the use of those terms. The paper is structured by consideration of two different ways in which the about/concerns distinction might be drawn and seen to be illuminating— ‘innocent commitments’ (section 1) and ‘parochialism’ (section 2).

1.

About vs Concerns: Regular Commitments vs Innocent Commitments

I’m going to discuss, as a locus for what I will call the ‘innocent commitments’ way of thinking about the about/concerns distinction, some work by Uriah Kriegel (2015) on the nature of temporal experience. Kriegel notes that, intuitively, perception and episodic memory differ in their felt temporal orientation. In episodically remembering rain outside one’s window, the rain is experienced as past. In seeing rain outside one’s window, the rain is experienced as present. This phenomenological datum appears to conflict with our best metaphysics of time, the B-theory. For, at least on one understanding of what B-theory says, if B-theory is true, there are no such temporal properties as being past, being present, and being future. So, if B-theory is true, on this understanding, the rain shower can’t have the temporal properties the experience presents it as having. There seem to be just three options. One might abandon B-theory. Or one might regard one’s experiences as illusory in what temporal properties they present things as having—one might accept an error-theoretic option. Or one might adopt an understanding of what B-theory says according to which it doesn’t say that things don’t have properties of a sort that experience presents them as having—e.g. perhaps all B-theory is committed to is the claim that things do not have monadic properties of pastness, presentness, etc., and all our experiences do is present things as having properties like being past or present without weighing in on whether these temporal properties are monadic. Kriegel wants to do none of these three things. Instead, he says: I want to offer a fourth approach that ‘squares the circle’: it does justice to the A-theoretic phenomenology while respecting B-theoretic metaphysics, yet without involving error theory. The very coherence of this alternative approach will expose a suppressed assumption that makes the three main approaches seem exhaustive. (2015: 408) Clearly, if there is a distinction that helps us see the possibility of a whole new approach to a familiar debate, that is a significant contribution. Kriegel sets out the background for his view as follows:

208

Daniel Morgan To appreciate the alternative approach, let us take a detour through existential belief. Consider the following two belief reports: (1) Aaron believes that ghosts exist. (2) Baron believes in ghosts. If we take these reports at face value, they appear to report two similar but structurally slightly different mental states. The similarity is this: both states commit to the existence of ghosts. The structural difference is this: while the commitment to ghosts’ existence is built into the content of Aaron’s mental state, it is built into the attitude of Baron’s. What Baron believes in is not ghosts’ existence, but simply ghosts; the commitment to their existence is built into the very attitude of believing-in. We may put this by saying that while Aaron’s belief represents ghosts-as-existent, Baron’s represents-as-existent ghosts. In the former the existence-committing element is a component of the represented, in the latter it is a modification of the representing. Thus while in Aaron’s belief, existence is part of what is represented, in Baron’s it is rather an aspect of how it is represented. (2015: 408)

In taking this view about existential belief, Kriegel is following Szabó (2003) and, ultimately, Brentano (1995 [1874]). His original move is to take an analogous view about temporal experience as a means to the end of not having to give an error theory about temporal experience. Focussing on perception, but referring back to Baron’s belief in ghosts, Kriegel writes: Consider a perceptual report such as: (3) Caron perceives rain. The idea is that we should read (3) on the model of (2): Caron’s perception encodes commitment to the rain’s presentness, just as Baron’s belief encodes commitment to ghosts’ existence, but this commitment is built into the attitude rather than content. Caron’s perception does not represent rain-as-present, but rather represents-as-present rain. The rain’s presentness is thus not part of what is experienced in Caron’s perception, but part of how the experiencing is done in it. This means that no property of presentness is attributed to the rain in Caron’s rain perception, just as no property of existence is attributed to ghosts in Baron’s belief. (2015: 410) Kriegel does not use the about/concerns terminology. He contrasts “content” with “attitude” and the “what of representation” with the “how

About vs Concerns

209

of representation”. Suppose we initially introduce the about/concerns distinction by reference to these notions: A state is aboutcontent a particular thing if and only if: that thing is relevant to the evaluation of the state in virtue of the content of the state—it is part of what is represented. (CONTRIBUTED BY CONTENT)

A state concernsattitude a particular thing if and only if: that thing is relevant to the evaluation of the state in virtue of the attitude the state involves—it is part of how the state represents. (CONTRIBUTED BY ATTITUDE) There is no obvious connection between the about/concerns contrast, introduced in this way, and the project of avoiding error theory. Still, Kriegel clearly thinks that making that distinction—and arguing that a perception’s commitment to presentness derives from the attitude and not the content—is enough to fend off error theory. That suggests a different way of thinking about the distinction. I’ll put it as follows: A state is aboutregular commitments a particular thing if and only if: that thing is relevant to the evaluation of the state. And: if the thing isn’t how the state presents it is as being, the state involves error. (REGULAR COMMITMENTS) A state concernsinnocent commitments a particular thing if and only if: that thing is relevant to the evaluation of the state. But: even if the element isn’t how the state presents it as being, the state does not involve error. (INNOCENT COMMITMENTS) As in the discussion of the Fregean term ‘sense’, one possibility is that introducing two hooks onto the notion of ‘concerns’ simply induces an ambiguity in the word ‘concerns’—the two relations are distinct (and similarly for the dual notion ‘about’). The other is that they are two different hooks onto the same relation. What Kriegel needs is that the second of these is the case. Only if that is true could he be right that buying his “attitudinal” account of felt temporal orientation is a way of escaping temporal error theory. There are two reasons, though, to think it isn’t true. First, on the INNOCENT COMMITMENTS understanding of ‘concerns’, the notion of ‘concerning’ seems paradoxical in a way that suggests nothing ever falls on that side of the distinction. Suppose the perception of the rain shower concerns the presentness of the rain shower in the INNOCENT COMMITMENTS sense. The rain shower is still presentedas-present. In that sense, the perception commits to the presentness of the

210

Daniel Morgan

rain shower. But also, even if B-theory is true and there is no presentness (i.e. nothing ever instantiates presentness, in particular no rain shower ever does), the experience is supposed not to involve error. In that sense, the perception is innocent of the commitment to the rain shower’s being present. A flat-footed question is, how is this combination possible? It seems to clash with the tautology that if a state represents-as-F x and x is not F, then the state involves error with respect to whether x is F. There is a second, independent problem. Assume that there could, in principle, be some things that fall on the ‘concerns’ side of the distinction, on the INNOCENT COMMITMENTS understanding. We still need a reason to think that we can infer from something’s falling on the ‘concerns’ side of the distinction, on the CONTRIBUTED BY ATTITUDE understanding, to its falling on the “concerns” side of the distinction, on the INNOCENT COMMITMENTS understanding. The very example Kriegel uses as a comparison to introduce his account of felt temporal orientation seems to illustrate that that inference isn’t safe. That is, the following seems like bad reasoning: The existential commitment to ghosts involved in Baron’s believing in ghosts is built into the attitude of believing in. So, even though ghosts don’t exist, Baron’s belief in ghosts does not involve error. But, in that case, the following reasoning, which instantiates the same pattern, must also be bad: The commitment to presentness involved in a perception of rain is built into the attitude of perceiving. So even though nothing is present, that perception does not involve error. On the face of it, the question of whether Baron represents-as-existent ghosts or represents ghosts-as-existent has no impact on what needs to be the case for Baron to avoid error. The fact that the commitment to existence falls on the ‘concerns’ (CONTRIBTUED BY ATTITUDE) side of the distinction does nothing to suggest that it falls on the ‘concerns’ (INNOCENT COMMITMENTS) side of the distinction. They are just two unrelated properties. We can also make the point in relation to an example like my ‘Tony is a handsome animal’ judgment. On a standard treatment, this judgment is supposed to concern (CONTRIBUTED BY ATTITUDE) the actual world. But how things are in the actual world is entirely irrelevant to determining whether the judgment is in error. If Tony is not handsome at the actual world, then my judgment is in error. In summary, arguably, the notion of concernsinnocent commitment is incoherent, and nothing instantiates it. Even if it is coherent, it doesn’t relate to the notion of concernsattitude in any obvious way.6 Kriegel’s discussion of time

About vs Concerns

211

is an example of the harm done by the assumption that there is a single unified distinction between two ways in which an element can be involved in the truth conditions of a state. We might define the about/concerns contrast in a way that has something to do with being contributed by the attitude rather than the content of a state. Or we might define it in a way that has something to do with the innocent commitments idea. Neither of these is clearly problematic. The problem comes when we assume they are tracking the same distinction. Kriegel’s approach to temporal experience assumes this. As a result, a false impression of a new, fourth solution to a problem that seemed to have only three possible solutions is generated. The effort to decide which of the three existing solutions to the problem is best is abandoned prematurely.

2.

About vs Concerns: Catholicism vs Parochialism

Perry would say that a benighted Northern Hemisphere inhabitant who judges ‘July is a summer month’ without awareness of the hemisphererelativity of being a summer month makes a judgment that concerns but is not about the Northern Hemisphere. The judgment concerns the Northern Hemisphere, in that the judgment is true if and only if July is a summer month in the Northern Hemisphere. But, Perry wants to say, the judgment is not about the Northern Hemisphere. Other examples—e.g. the child in Sacramento who judges ‘It is four o’clock’—work similarly. This suggests something like the following understanding of the about/ concerns distinction: A state is aboutcatholicism a particular element if and only if: that element is relevant to the evaluation of the state. And: the subject has the conceptual capacity to be in states to whose evaluation other values of the same parameter are relevant. (CATHOLICISM) A state concernsparochialism a particular element if and only if: that element is relevant to the evaluation of the state. But: the subject lacks the conceptual capacity to be in states to whose evaluation other values of the same parameter are relevant. (PAROCHIALISM) There is some vagueness to the about/concerns distinction, on this understanding. Children who clearly lack the concept of a time zone turn into adults who clearly have that concept. At some point in the transition, there will be indeterminacy as to whether the child really has the conceptual capacity to represent other values of the same time zone parameter. So it will then be indeterminate whether the state concerns or is actually about a particular time zone. But that doesn’t do anything to undermine

212

Daniel Morgan

these notions. The notions, as explained, are perfectly meaningful. They just don’t have any obvious explanatory use yet. One can’t say that the reason the Northern Hemisphere is relevant to the evaluation of the ‘July is a summer month’ judgment is because; although that judgment isn’t about the Northern Hemisphere, it does at least concern it. That would just be a dormitive virtue style explanation, at least so far. We may ask, though, what happens if we hook up this way of thinking of the about/concerns with a different way of thinking of it—in particular the (CONTRIBUTED BY ATTITUDE) way we have already looked at: A state is aboutcontent a particular thing if and only if: that thing is relevant to the evaluation of the state in virtue of the content of the state—it is part of what is represented. (CONTRIBUTED BY CONTENT) A state concernsattitude a particular thing if and only if: that thing is relevant to the evaluation of the state in virtue of the attitude the state involves—it is part of how the state represents. (CONTRIBUTED BY ATTITUDE) That makes for a potentially explanatory connection. The two ways of introducing the distinction might permit what Sainsbury calls the “subsumption of a number of apparently distinct phenomena” (2002: 127). The problem is that, on investigation, the phenomena do seem to be actually distinct. For example, when the benighted Northern Hemisphere inhabitant judges ‘July is a summer month’, the attitude type is the attitude type judgment. That attitude type no more intimately involves the Northern Hemisphere than the Southern. So, it isn’t the case that the Northern Hemisphere is something the judgment ‘concerns’ (CONTRIBUTED BY ATTITUDE). But it is the case that the Northern Hemisphere is something that the judgment ‘concerns’ (PAROCHIALISM). Of course, it isn’t completely obvious how much should be built into the ‘attitude’ part of a state. Attitude verbs are a natural guide to attitude types. One reason it seems obvious the attitude of, for example, judging that p differs from the attitude of supposing that p is that ‘supposing’ and ‘judging’ are different words, and they clearly are not synonyms. We might also think that supposing and judging are associated with different functional roles, and that is a further reason for regarding them as different attitudes. But, really, there is nothing to stop us from being much more fine-grained in how we individuate attitude types. We could say that when a Northern Hemisphere inhabitant judges ‘July is a summer month’, the fact that this judgment is the judgment of a Northern Hemisphere inhabitant—and even the fact that it’s the judgment of someone who never represents values of the hemisphere parameter other than the Northern Hemisphere—is built into the kind of attitude their state

About vs Concerns

213

involves (parochially-judging-month-of-year-northern-hemisphere-wise). On that way of saying what kind of attitude the state involves, it won’t be true the attitude type is too coarse-grained for it to be plausible that the attitude type contributes the Northern Hemisphere to the truth conditions of the judgment. But that is only because we’ve gerrymandered a notion of attitude type to fit our purposes. An analogy would be someone who introduced the notion of ‘sense’ by reference to rational co-tenability of belief and then claimed that the notion of sense, thus introduced, had some role to play in understanding intersubjective communication—e.g. perhaps understanding someone always involves thinking about the same thing they are thinking about using the same sense. In response to apparent counterexamples—e.g. a case where I intuitively understand you perfectly well but we don’t deploy the same sense—they reply that, on the notion of ‘understanding’ they mean, this isn’t a case of understanding. This is unexplanatory if the notion of understanding they mean is stipulated to involve two people thinking the same thing using the same sense. There are cases where it is genuinely unclear whether we should regard there being two different attitudes. For example, in the example Kriegel discusses, it isn’t clear whether we should (as Brentano, Kriegel, and Szabó do) really think that the state of believing in ducks involves a different kind of attitude from the state of believing that ducks exist. An alternative hypothesis is that English just possesses two different ways of ascribing the very same mental state. Similarly, it isn’t clear whether we should think that John’s state of intending to dance involves a different kind of attitude from the state of intending that John dance. Are there two kinds of attitude here—a more primitive kind of intention that perhaps animals can share with us (‘intentions to’) and a more sophisticated kind that only humans can have (‘intentions that’)—or is there just a shorter and a longer way of ascribing an identical mental state? It isn’t clear. The answer might depend on whether we can find a distinctive functional role for the two putative kinds of intention. But, however that kind of dispute turns out, it really doesn’t seem as though there is any non-gerrymandered notion of attitude type that is fine-grained enough to connect up with a notion of ‘concerns’ explained by reference to PAROCHIALISM. It’s worth pointing out that the PAROCHIALISM idea I’ve just looked at is tied to fairly stable features of an individual—whether they lack the concept of a time zone, for example. We could also introduce a notion of ‘concerns’ by reference to a kind of MOMENTARY PAROCHIALISM. I am an adult and I understand the time zone-relativity of time of day. On the other hand, most of the time, time zone isn’t very salient to me. I judge ‘It’s 4 pm’. My judgment is true if and only if it is 4 pm in GMT, but GMT is probably not something I would think of if asked what the truth of my judgment depends on. If I voice my thought in inner speech, there won’t be any term in inner speech corresponding to GMT. My judgment doesn’t concernparochialism GMT (because I can think of other time

214

Daniel Morgan

zones), but we might say that it concernsmomentary parochialism GMT (because on this occasion my judgment involved GMT as a kind of default). Again, there’s nothing wrong with a notion introduced in this way. But it only becomes explanatory when hooked up to something else. And it’s clear that concernsmomentary parochialism is no better a candidate for linking up with concernscontributed by attitude than is concernsparochialism. In summary, the kinds of examples discussed in relation to PAROCHIALISM are striking and interesting. It’s interesting that sometimes I make a judgment whose truth depends on what time zone I am in even though the time zone–relativity isn’t at that moment salient to me. It’s interesting that some people make judgments whose truth depends on what time zone they are in even though the time zone–relativity is never salient to them because they lack the capacity to represent different values of the time zone parameter. We shouldn’t think, though, that pointing out that there are such cases, or even introducing a label to refer to them, does anything to explain their possibility. Analogously, it’s striking and interesting that someone can rationally believe (of the same thing) that it is F and that it is not F. But merely pointing out that there are such cases—and introducing a label for them (e.g. ‘sense’)—does nothing to explain their possibility. Neither does introducing a label enable us to draw any conclusions we weren’t in a position to draw before introducing it, or disable us from drawing any conclusions we were in a position to draw before. For example, as mentioned at the outset, a naïve idea is that a judgment is about something if and only if that thing is one of the things that the truth of the judgment depends on; so if you know that some element is one of the things that the truth of a judgment depends on, you can infer that that element is one of the things the judgment is about. None of the examples discussed in relation to PAROCHIALISM do anything to undermine this naïve idea about what ‘aboutness’ is.

Conclusion I’ve argued that there is no explanatorily interesting distinction between two ways that an element can figure in the truth conditions of a judgment. This is based partly on a general account of what it would take for the distinction to be explanatorily interesting (inspired by Sainsbury’s account of what it takes for the notion of ‘sense’ to be explanatorily interesting). It’s also based partly on an examination of individual examples. Those examples are striking, but they don’t work out nearly as neatly as is generally supposed. On the face of it, they illustrate a number of different differences.

Notes 1. For comments and help with this paper, thanks to Brian Ball, Max Kölbel, Uriah Kriegel, and Léa Salje and the participants at an NCH workshop on judgment.

About vs Concerns

215

2. Recanati (2007)—whose work I don’t discuss in detail in this paper—distinguishes between complete content and explicit content. If I judge ‘It is raining’, my judgment is true if and only if it is raining at the time of judgment. But the time of judgment does not figure in the explicit content of the judgment, only in its complete content. If I judge ‘I am in pain’, on the basis of introspection, my judgment is true if and only if I am in pain. But the explicit content of the judgment just involves the property of being in pain. I figure in only the complete content of the judgment. This is partly inspired by Hume’s idea that when I introspect I don’t seem to encounter any object that is myself. It is also intended to do justice to the idea that when I learn by introspection that someone is in pain, there’s no real question about who that person is—I don’t need to check that the person whose pain I’m apprehending is my own. As the point is sometimes put, introspective judgments are “immune to error through misidentification relative to the first person” (see 2007: 23–5 for discussion). 3. See Rumfitt (1994) for discussion of the idea that intentions can have acts as their objects. In relation to this kind of example, Recanati uses the evocative phrases “the implicit de se” and “the explicit de se”. He thinks the intention to dance is implicitly de se where the intention that I dance is explicitly so (2007: 25–6). 4. I will use ‘state’ in a catch-all sense to cover act- and event-like mental phenomena as well. 5. See Brian Ball’s illuminating introduction to this volume [pp. 8–10] for further discussion of this. 6. Perhaps there are some domains that provide a more fertile application for the innocent commitments idea than the one to which Kriegel applies it. Consider ‘tastiness’ judgments. I judge ‘Rhubarb is tasty’, and you judge ‘Rhubarb is not tasty’. Rhubarb is tasty relative to my standard of taste, and rhubarb is not tasty relative to your standard of taste. Suppose we have the intuition that this really is a case of disagreement (so my judgment shouldn’t be regarded as the judgment that rhubarb is tasty-to-me, and your judgment shouldn’t be regarded as the judgment that rhubarb is tasty-to-you, in which case the idea of disagreement would be lost). But there is also no good sense in which either judgment could be at fault. (See Kölbel 2003 for discussion.) There is something many will find paradoxical in that suggestion. If I judge rhubarb is tasty and you judge it isn’t, why isn’t it the case that I am at fault if and only if rhubarb isn’t tasty and you are at fault if and only if rhubarb is tasty? But suppose we set this kind of objection aside—perhaps this domain of tastiness judgment just is a bit paradoxical. So we might assume that, at least in the case of judgments like ‘Rhubarb is tasty’, the notion of faultless disagreement makes sense. It might be possible to use that to introduce a notion of ‘concerns’ that works roughly the way the INNOCENT COMMITMENTS notion is supposed to (it does seem to be that, on Kriegel’s picture, B-theory and experience conflict but somehow there is no fault). But, of course, doing that wouldn’t do anything to suggest that paradigmatic faultless disagreement cases are interestingly like cases in which a commitment is built into the attitude a state involves—e.g. believing-in cases. If you believe in ghosts and I disbelieve in ghosts, then one of us is clearly at fault. The faultless disagreement idea has no purchase, and it would be a bad thing if the concerns/about distinction made it seem as though it must do.

References Brentano, F. C. (1995) [1874], Psychology From an Empirical Standpoint, trans. A. C. Rancurello, D. B. Terrell, and L. L. McAlister. London: Routledge.

216

Daniel Morgan

Grice, H. P. and P. F. Strawson (1956), “In Defense of a Dogma”, The Philosophical Review 65(2): 141–58. Kölbel, M. (2003), “Faultless Disagreement”, Proceedings of the Aristotelian Society 104: 53–73. Kriegel, U. (2015), “Experiencing the Present”, Analysis 75(3): 407–13. Perry, J. (1986), “Thought Without Representation”, Proceedings of the Aristotelian Society 60(sup.): 137–51. Recanati, F. (2007), Perspectival Thought. Oxford: Oxford University Press. Rumfitt, I. (1994), “Frege’s Theory of Predication: An Elaboration and Defense, With Some New Applications”, Philosophical Review 103(4): 599–637. Sainsbury, M. (2002), Departing From Frege: Essays in the Philosophy of Language. London: Routledge. Szabó, Z. G. (2003), “Believing in Things”, Philosophy and Phenomenological Research 66: 584–611.

11 Predication and Two Concepts of Judgment Indrek Reiland

Introduction What is it for us to cognitively represent the world as being a certain way in the sense of being in a state with propositional content and truth conditions?1 On the traditional Platonist or Fregean picture, we represent by standing in certain relations to intrinsically representational propositions qua abstract objects. For example, for you to represent this apple as being red is for you to stand in some relation to the proposition that this apple is red, which is what in the first place represents it as being red. On this picture, it is propositions that represent fundamentally and we who represent derivatively. The Platonist picture has recently come under attack. First, it leaves it completely unexplained how we could come to stand in the relevant relations to propositions qua abstract objects. Second, it leaves unexplained what propositions qua abstract objects are such that they could intrinsically represent. In other words, it can’t solve the problem of the unity of the proposition or the problem of explaining how propositions have truth conditions (Davidson 2005; Jubien 2001; King 2007, 2009; Soames 2010, 2014, 2015; Hanks 2015). On the critics’ alternative naturalist picture of cognitive representation, it is we who represent fundamentally and propositions that represent, if at all, derivatively. This requires an answer to the question what it is for us to represent. Suppose we can independently explain how we get to be in touch with objects, how we get to think of them or, in a thin sense, refer to them. For example, it’s a relatively common view in contemporary cognitive science and philosophy of perception that perception involves causally driven, non-conceptual, context-bound reference to objects (Burge 2010; Clark 2004; Fodor and Pylyshyn 2015; Pylyshyn 2007, 2009). Furthermore, many people think that perceptual reference somehow grounds conceptual, context-bound/demonstrative reference to objects, and it’s not a stretch to think that this, in turn, somehow grounds naming, which enables contextfree/non-demonstrative reference (Campbell 2002; Smithies 2011).

218

Indrek Reiland

One way to approach our question about representation is to ask, what do we need to do beyond thinking of or referring to an object to come to cognitively represent it as being a certain way? The age-old answer is that we need to further predicate, attribute, or ascribe a property to it. Peter Hanks and Scott Soames have both appealed to this answer in developing their act-based theories of cognitive representation and propositions. Their view has a common structure (Hanks 2015; Soames 2010, 2014, 2015). On the first step, they claim that to predicate the property of being F of O is to represent O as being F. They also equate this with performing the most basic act with propositional content. On the second step, they then claim that the relevant act-types can be identified with propositions. The central difference between Hanks and Soames’s views lies in how they think about predication and the most basic propositional act. Hanks thinks that predication is forceful, in that when you predicate being F of O, you take a stand on whether O is F—namely, that it is. To predicate being F of O is to judge that O is F. In contrast, Soames thinks that predication is neutral and to predicate being F of O is instead to entertain the proposition that O is F. I’m attracted to act-based theories of cognitive representation and have elsewhere started developing a third view that tries to improve on both Hanks and Soames (Reiland 2019). My first and main aim in this paper is to present the common part of our program in its strongest form while also outlining my own version of it. However, I also have a secondary aim. Like Hanks, I think that predication is forceful and to predicate is to perform an act that may be called judging. But this is potentially confusing because it’s common to also use ‘judge’ for a much more involved act of settling a question about how things are in the light of one’s evidence. Accordingly, my second aim in this paper is to show how we can clearly separate the thin predication-resultant notion of judging, S(emantic)-judging, from the much richer notion used in epistemology, E(pistemic)-judging. I will proceed as follows. I will first present the act-based program of explaining cognitive representation and my own version of it (section 1). Second, I will introduce the difference between S-judging and E-judging and give examples of philosophers talking about each (section 2). Next, I will elucidate the difference between them further in three respects: belonging to different levels of propositional acts and attitudes, normgovernance, and the possibility of subpersonal occurrence (section 3). Finally, I’ll clarify the distinction some more by showing how both types of judging have been related to belief (section 4). 1.

The Act-Based Program, Predication, and Judgment

The act-based program of explaining cognitive representation common to Hanks, Soames, and me starts with the following steps:

Predication and Two Concepts of Judgment Atomism:

Glue:

219

We presuppose an independent account of subpropositional acts of reference to objects and indicating properties. In the most basic cases, if one refers to an object, indicates a property, and predicates the property of the object, then one performs the most basic act with propositional content.

To illustrate, on Hanks’s view, if one refers to Arvo Pärt, indicates the property of being a composer, and predicates the latter of the former, then one judges that Arvo is a composer. Let me start fleshing these parts of the program out by explaining how to think of the notion of reference, why we need to distinguish between property-indication and predication, and how to think of the relation of the acts of reference and property-indication to the ensuing predication. (a) Thinking of vs reference: thinking of an object is mentally picking it out, bringing it to mind, and making it available to do something further with it. We can call this reference, but it must be made clear that this is reference in a thin sense. The problem here is that many people’s intuitions with the verb ‘refer’ seem such that we would only call an act of thinking of an object reference if it occurs while making it a subject of predication. For example, ask yourself in which of the following typical uses you refer to Arvo: (1) Using ‘Arvo is a composer’ to say that Arvo is a composer (2) Using ‘What is Arvo’s best composition?’ to ask what Arvo’s best composition is (3) Using ‘Arvo, write some more music!’ to beg Arvo to write some more music (4) Using ‘Arvo!’ to get Arvo’s attention I suspect many people will find it intuitive to say that in cases (1) and (2), you are referring to Arvo, but in (3) and (4), you’re not. This seems to be because, in the first two cases, you are thinking of Arvo to make him a subject of predication—or, in the case of (2), perhaps some questionrelated analogue. This doesn’t happen in the last two cases. On the other hand, there is something in common in all the cases, and that is using the name ‘Arvo’ to think of Arvo or to refer to Arvo in a cheap sense. Since the act-based program is an Atomist program, trying to build up from the constituents to the whole, it needs its sub-propositional act of thinking of or reference to be one that can occur independently of predication, and so it operates with this latter notion.2 I think that the distinction between mere thinking of and making a subject of predication helps us to better assess Frege’s context principle,

220

Indrek Reiland

according to which, roughly, we can only make sense of the parts (e.g. reference, predication) in terms of the whole (e.g. judging). Even though this is plausible for reference, understood in the richer sense, it is implausible for thinking of. As the above examples seem to show, the primary semantic function of names is to encode acts of thinking of objects, acts that also occur in (3) or (4), and not acts of reference in the richer sense, acts that only occur in the context of the whole. (b) Property-indication vs predication: indicating a property is similarly getting it to mind, making it available to do something further with it. Why do we need to separate it from predication? The reasons here are similar to the above but more widely appreciated. First, consider predicating complex properties like being not red or being red or green. Such complex properties must be formed through negating or disjoining simple properties. This requires that the simple properties be made subject of other acts or operations. And this means that there must be a way in which they’re gotten in mind that doesn’t involve predicating them (Soames 2010: Ch. 7). The same can be said of quantification. On standard treatments, this involves predicating second-order properties of properties themselves.3 For example, on Hanks’s view, to judge that everything is coloured is to predicate being universally instantiated of the property of being coloured. Again, this requires that properties themselves be made subject of predication, and this means that there must be a way in which they’re gotten in mind that doesn’t involve predicating them (Hanks 2015: 87). It’s a further and very interesting question what it is to indicate a property like the property of being a composer, something that is true of x iff x is a composer. It clearly can’t be thought of in terms of mentally picking out an intrinsically satisfaction-conditioned Platonic property because that would be no progress in explaining representation. Hanks therefore suggests that it is to be understood in terms of giving oneself a rule for sorting or predication (Hanks 2015: 206–27). However, this part of the program is currently undeveloped and awaits further work (see Hanks 2017: Reiland MS). (c) The relation between thinking of, indication, and predication: how are the acts of thinking of an object and property-indication related to predication? One way to think about it is that a judging (or entertaining) consists of a sequence of the acts of thinking of and property-indication + predication. Soames and Hanks sometimes present their view like this, and I’ve presented it myself in this way in the past (Reiland 2013). However, this version of the view has the obvious problem of raising the question of how the sequence of acts is unified into a whole with truth conditions (Speaks 2018). The alternative—and the only viable way to think about the view—is that the acts of thinking of and property-indication serve as preconditions for being able to perform the predication. They are not parts of the predication or judging. The predication and judging don’t

Predication and Two Concepts of Judgment

221

have parts at all, but are unities in and of themselves. It follows that if we take something suitably related to predication to be the proposition, then the proposition doesn’t really have parts either and is also a unity in and of itself. Now, like Hanks, I think that to predicate is to judge. If so, why do we need to separately talk of predication and judging? (A similar question arises on Soames’s view in relation to predication and entertaining.) Let’s look at this next. (d) Predication and judging: if to predicate is to judge, then the two acts are token-identical. However, Hanks and Soames think they’re not type-identical. The reason for insisting on this can be seen by looking at Wittgenstein’s objection to Russell’s multiple relation theory of judgment (Russell 1912). Russell treated judging as a multiple relation between a person and disjoint items—say, you, Arvo, and the property of being a composer. Wittgenstein objected that the judgment that Arvo is a composer must contain something that is capable of being true or false, but that this is not so on Russell’s theory, on which it is a relation to disjoint items (Wittgenstein 1995; for this interpretation, see Hanks 2015: 161–3). Hanks and Soames’s project starts from seeing that Russell had a real insight when he thought that a multiple relation between a person and disjoint items is what generates representation and propositional content. However, Wittgenstein’s objection seems damning. The insight can be captured while avoiding the objection if we distinguish between predication and judging qua types. The claim is that to predicate is to stand in a multiple relation to disjoint items. One predicates one thing, a property, of a different thing, an object. To predicate is to judge in the sense that judging is analysed in terms of predication. But to judge is not to stand in a multiple relation to disjoint items, but to perform an act with propositional content, an act that thus “contains” something capable of being true or false. Hanks and Soames fill in the details a bit more by saying that in predicating and judging (or entertaining), one represents the world as being a certain way, and one’s act-token similarly does as well. Then they go a step further and identify the act-types with propositions. They therefore sometimes say that judging and entertaining are tokening relations (Hanks 2015: 161). However, the claim that judging and entertaining are tokening relations is more confusing than helpful. This is because the actbased program’s grounding idea is that predication is more fundamental than judging and judging as a token act is more fundamental than the types. Thus, a judging can be a “tokening” relation where one tokens a type only in a derivative sense. I think the more correct way to put their view is that to judge is not to stand in a relation at all, but to instantiate a monadic property of performing an act with propositional content. After all, the further claim that the relevant act-types can be identified with propositions is an entirely voluntary add-on.

222

Indrek Reiland

In fact, I’ve never liked the third step of the Hanks-Soames view: Types:

The relevant act-types (e.g. judging-types, entertaining-types) are propositions.

Here are the most damning objections to this view. First, intuitively, identifying propositions with acts is a category mistake (King 2013: 90). Propositions are sorts of objects, not acts. This intuition can be made sharper by focussing on the claim that predication grounds representation and truth conditions. Intuitively, when I predicate being a composer of Arvo, I represent Arvo as being a composer. I represent, but my act doesn’t. But if my act doesn’t, then the token acts lack representational properties, and so do the types of acts.4 I think we can do much better if we follow Kazimierz Twardowski and Friederike Moltmann in further distinguishing between acts and their products (Moltmann 2013, 2017; Twardowski 1911). On this view, the idea is that to predicate is to judge and to do this is to represent. However, to predicate and judge is to produce a product, a judgment. I represent, my acts don’t, but the product, the judgment, does. In fact, I represent by producing the product that represents. My act has a propositional content only in the sense that it produces a product that has such a content. Let’s look at this more closely. (e) Acts, products, and propositions: to get a grip on the act-product distinction, consider first cases where an act results in a physical product that outlives the time of its production. For example, take the act of drawing, which produces a drawing. The act of drawing is a datable and locatable event, whereas the physical product is an enduring object that outlives the time of its production. They also differ in their properties: only the act can be careless or hasty, whereas only the product can be accurate. Next, consider cases where the physical product doesn’t outlive the time of its production. For example, take the act of making a face (producing a grimace, a smile, etc.). The act is a datable and locatable event, whereas the physical product, a face or facial expression, is something that doesn’t outlive the time of its production. Yet, they still differ in their properties: only the act of making a face can be impolite or disrespectful or difficult to perform, whereas only the product, the face or facial expression made, can be beautiful. Now, consider cases where an act doesn’t result in a physical product, but a psychological, social, or normative one that nevertheless outlives the time of its production. For example, the act of deciding and a decision, an act of promising and a promise, and an act of making a law and the law. The acts are again datable and locatable events, whereas the products are enduring statuses that outlive the time of their production. A decision stands as long as it isn’t changed, a promise exists as long as it

Predication and Two Concepts of Judgment

223

isn’t fulfilled—or one isn’t released from it—and a law is in force until it ceases to be so. Again, they also differ in their properties. Acts of deciding can’t be changed, but decisions can. Acts of promising can’t be broken, but promises can. Acts of law-making can’t be violated, whereas laws can. Finally, consider cases where the psychological, social, or normative product doesn’t outlive the time of its production. In this category, Twardowski and Moltmann would place the acts of judging and judgments. The acts are again datable and locatable events, whereas the products are objects that don’t outlive the time of their production.5 Nevertheless, the distinction is needed since they differ in their properties. First, as we saw above, intuitively, acts of judging can’t have truth conditions and be true or false; only the products, the judgments, can. Second, acts of judging enter similarity relations on a different basis than judgments. My judging and your judging are exactly similar when they’re performed in the exact same way. But for the products, the judgments, to be exactly similar, they must just have the same content (Moltmann 2017: 261). Thus, on my view, the acts of predicating and judging are acts of producing a product, a judgment. A judgment has a propositional content in the sense that it is itself something that has truth conditions and is capable of truth and falsity. Given all this, on this view, the products, things like judgments, are of primary importance. They are the primary bearers of truth and falsity and of modal properties. However, they can’t be identified with propositions because they are fleeting, mind-dependent creations, artefacts. Insofar as we need propositions at all—and whether we do so is an open question—we can perhaps identify them with types of products. (f) Predication and force: I agree with Hanks that predication is forceful, in that when you predicate being F of O, you take a stand on whether O is F—namely, that it is. To predicate being F of O is to judge that O is F. I think that taking predication to be forceful is the only coherent option. Hanks gives the following argument (Hanks 2015: 36–7). Performing the act of predication is supposed to result in representing-as and be something that has or results in something having truth conditions. If I predicate F of O, then I represent O as being F and do something incorrect, something that’s done falsely, if O is not F. Now assume that predication is neutral and doesn’t amount to taking a stand on whether O is F. Then my action wouldn’t be incorrect, wouldn’t be done falsely, if O is not F. It follows that if an act of predication is to have truth conditions, it has to be forceful in the sense of involving taking a stand. I find this argument convincing and won’t discuss it further here.6 Taking predication to be forceful raises the spectre of the Frege-Geach problem: the problem of accounting for embeddings in negations and disjunctions. I’ve developed the view in a way in which this problem doesn’t

224

Indrek Reiland

arise at length elsewhere. Here’s a quick summary of the main idea, now also given in terms of products. (g) Grasping and why there is no Frege-Geach problem: consider Hanks’s view of what it is to perform judgments of complex propositions like those involving propositional conjunction, negation, and disjunction. First, you perform the constituent judgments that make available the propositions as judgment-types to serve as targets of further predication. Then you predicate conjunction or negation or disjunction of those judgment-types (Hanks 2015: 99). For example, take the act of judging that O is F and O is G. On Hanks’s view, to perform this judgment, you must first perform the constituent judgments that make available the propositions as targets, after which you can predicate being jointly true or standing in the conjunction relation Conj of them. We can represent this as follows, where REF(O), represents the act of referring to O, IND(F) represents the act of indicating the property of being F, ├ represents forceful predication, and ├ ↑ represents target-shifted predication: ├ ↑ The idea is that one performs the first judgment, then performs the second judgment, and then shifts to the judgment-types as targets of predication for the final predication of Conj of them. But now take the acts of judging that it is not the case that O is F or that O is F or O is G. On Hanks’s view, to perform these judgments, you must again first perform the constituent judgments, only after which can you predicate untruth and being disjointly true or standing in the disjunction relation Disj of the relevant propositions: ├ ↑ ├ ↑ But it’s simply false that judging that it is not the case that O is F and judging that O is F or O is G require judging that O is F. And this is the essence of the Frege-Geach problem.7 I think there are serious problems with the idea that performing the constituent judgments is sufficient to make available the associated propositions qua judgment-types as targets of predication (see Reiland 2019). However, to prevent the Frege-Geach problem from arising, we have to also see that it’s not necessary. My alternative view defended in that paper is that once we have the capacity for performing certain judgments, we can grasp the propositions qua judgment-types independently of performing them. Grasping is an objectual attitude like thinking of, under a practical mode of presentation.8 And grasping is what is required to make available the propositions qua judgment-types as targets of predication. Thus, even though predication is forceful and grounds everything, grasping a

Predication and Two Concepts of Judgment

225

proposition qua a judgment-type is required for embedding and is how force gets neutralized in those contexts. After the shift to products, I would put the same idea in terms of them. Predication and judging are forceful, in that they result in producing a product that has truth conditions. But the product or product-type can be grasped, and it’s grasping that makes the product or product-type available as a target of further predication, which is required for embedding and is how force gets neutralized in those contexts. (h) Summary: let me sum up the essentials of my view. When one thinks of Arvo, indicates the property of being a composer, and predicates the latter of the former, then one judges that Arvo is a composer. Thinking of is reference in a cheap sense, to be distinguished from making something a subject of predication. Similarly, indicating a property is to be distinguished from predicating it. The two acts of thinking of and property-indication are preconditions to be able to perform the predication but are not parts of it. Predication is a multiple relation; judging is a two-place relation of production. To judge is to produce a judgment that is a product, an object with propositional content. Products are mind-dependent artefacts. Predication is forceful and truth-committal, which is why the product has truth conditions. Products or product-types can be grasped where grasping is an objectual attitude like thinking of, under a practical mode of presentation. Grasping is required for embedding and is how force gets neutralized in embeddings in negations and disjunctions. 2.

S-Judging vs E-Judging

Until now, we’ve been using ‘judge’ for the predication-resultant act in the performance of which one represents something as being some way. Many philosophers who are primarily focussed on philosophy of mind and language use the word ‘judge’ in this thin sense. Let’s call this S(emantic)-judging. The problem is that many other philosophers use the word ‘judge’ for a much more involved act of settling a question about how things are in the light of one’s evidence. Let’s call this E(pistemic)-judging. In this section, I will introduce examples of philosophers talking about each to try to get an initial grip on their difference. Let us start with some examples of philosophers using ‘judge’ in the sense of S-judging. First, consider philosophers who use the term ‘perceptual judgment’ (e.g. Lyons 2015; Johnston 2006; Toribio 2018). A perceptual judgment is supposed to be an act with propositional content that either forms a part of perception or immediately follows it. It’s clear that ‘judge’ is here used in the sense of S-judging because perceptual judgments are supposed to be formed immediately and are not made by the agent for reasons or based on evidence.

226

Indrek Reiland

Second, consider the standard translation of Kant’s ‘Urteil’ as ‘judgment’. According to the Stanford Encyclopedia of Philosophy entry ‘Kant’s Theory of Judgment’, here’s Kant’s view of judging: According to Kant, a “judgment” (Urteil) . . . is a higher-order complex conscious cognition that refers to objects either directly (via the essentially indexical content of intuitions/non-conceptual cognitions) or indirectly (via the essentially attributive or descriptive content of concepts); in which concepts are predicated either of those objects or of other constituent concepts. (Hanna 2017) Now, perhaps Kant has more packed in his ‘Urteil’ than just being the most basic act in the performance of which one represents. However, it should be clear that he’s not using ‘Urteil’ exclusively for acts performed for reasons or based on evidence. Let’s now look at examples of philosophers using ‘judge’ in the richer sense of E-judging: Judgement is a conscious rational activity, done for reasons, where these reasons are answerable to a fundamental goal of judgement: that it aims for truth. (Peacocke 1998: 88) Judging is a mental action. To judge that P is to do something and to do it for a reason. (Cassam 2010: 81) A judgment is a cognitive mental act of affirming a proposition (although, as we shall explain, not all affirmations are judgments). It is an act because it involves occurrently presenting a proposition, or putting it forward in the mind; and it is cognitive because it involves presenting the proposition as true—or, as we have said, affirming it. . . . A judgment, like a belief, is correct if and only if its content is true. Reasoning aims to issue or not issue in a belief that p in accordance with the relevant norm by first issuing or not issuing in a judgment that p in accordance with the corresponding norm. Strictly speaking, then, the question whether to believe that p is transparent, in the first instance, to the question whether to judge that p, which in turn is transparent to the question whether it would be correct to judge that p, and thence to whether p is true and, finally, to whether p. (Velleman and Shah 2005: 503) All of the above use ‘judge’ so that to judge is not merely to represent something as being some way, but to do so with the aim of getting things right and based on reasons and evidence one has.

Predication and Two Concepts of Judgment

227

Conor McHugh has worked out a view of E-judging that adds some detail. On his account, E-judging is an act, performed for reasons. It’s a way of making up one’s mind, settling a question about how things are—for example, whether p—while having the aim of doing so correctly.9 Having this aim entails not only that in judging that p one has normative reasons, considerations that count in favour of judging that p, but also that in judging, one is motivated by such reasons. In other words, judging is goal-directed in aiming at truth. In judging whether p, the thinker aims that the following state of affairs holds: she judges that p iff p. Having this aim means being motivated by consciously registered reasons conducive to the goal, that is, by evidence (McHugh 2011). Hopefully, we now have an initial grip on the difference between S-judging as the predication-resultant act in the performance of which one represents and E-judging as the act of concluding theoretical inquiry by settling the question how things are in the light of one’s evidence. In the next section, I’ll elucidate the difference between them further in three respects. 3.

Three Differences

(a) Levels of acts: consider Austin’s distinction between locutionary and illocutionary acts in speech act theory (Austin 1962). On my reading of Austin, on his view, to utter a sentence with (one of) its meanings— “with a more or less definite sense and reference” (Austin 1962: 95)—in a language, is to perform a rhetic or full locutionary act. Depending on the mood of the sentence, this act will be either one of saying something or asking something or telling someone to do something. For example, consider the humorous sentence ‘Koristame ruumi(d/t)’, which in Estonian means ‘Let’s clean the rooms’ and in Finnish means ‘Let’s decorate the corpses’. If you utter it with its meaning in Estonian, you tell a group including yourself to clean the rooms. If you utter it with its meaning in Finnish, you instead tell a group including yourself to decorate the corpses. Locutionary acts are usually performed to perform further illocutionary acts, acts that one performs in saying, asking, or telling (Austin 1962: 99). For example, in saying, one might be guessing or conjecturing or asserting; in asking, one might be inquiring or examining someone; and in telling someone to do something, one might be requesting or ordering or advising. Thus, there is a level difference between a locutionary act like saying and an illocutionary act like asserting. Saying is a part of any other illocutionary act of the declarative kind, whereas asserting is a distinctive act of the declarative kind at the same level as other such acts, like conjecturing. According to the act-based program, predication and S-judging are the most basic acts in the performance of which one represents propositionally.

228

Indrek Reiland

This means that they’re in one way or another implicated in any other act or attitude with propositional content. For example, consider mentally guessing. If you guess, you’re representing something as being a certain way, and thus you must be predicating and S-judging. The same is true of other acts, like hypothesizing and E-judging. Hence, there is a level difference analogous to that between locutionary acts and illocutionary acts between S-judging and E-judging: S-judging is a part of any other act with propositional content, whereas E-judging is a distinctive act at the same level as other such acts, like guessing and hypothesizing. (b) Norm-governance: it’s customary to distinguish between objective and subjective norms governing belief and discuss their relation. A belief is supposed to be objectively correct iff it is true. Thus, it is sometimes suggested, there is an objective norm of truth governing belief. On the other hand, a belief is subjectively correct or rational if it is based on adequate evidence. Thus, there is a subjective norm of evidence governing belief. S-judgings and E-judgings differ insofar as norm-governance. On the face of it, S-judgings are only governed by the objective norm of truth, whereas E-judgings, analogously to belief, are governed by both the objective norm of truth and the subjective norm of sufficient evidence. However, given my view, some nuance is needed here. Consider the question of whether predication or S-judging qua acts are governed by any norm. Well, one could say that they’re governed by an objective norm in the sense that they’re performed correctly iff they’re true. However, on my view, acts of predication and S-judging are not true. Their products are. Nevertheless, we can say the following: an act of predication or S-judging is performed correctly iff it results in a product that is true. In the case of E-judging, things are more straightforward. They’re governed by an objective norm insofar as they’re performed correctly iff they result in a product that is true. However, more importantly, they’re governed by a subjective norm on which the act is performed correctly iff one has sufficient evidence. This is why E-judgings belong in the sphere of epistemic agency and rationality. To violate the subjective norms governing the act is to fail rationally. (c) The possibility of subpersonal occurrence: it’s highly likely that many of our S-judgings are performed subpersonally. Consider what are known as seemings: occurrent events in which it seems to the subject that p. Such seemings can be divided into perceptual seemings and intuitions or intellectual seemings. They’re frequently taken to have propositional and conceptual content that represents the world as being a certain way (Reiland 2015). However, the agent is not active with regard to their formation, as she is with regard to her E-judgings. Rather, to the agent, it is as if they passively receive the seemings. Nevertheless, they have to be formed somehow, and one plausible story about them tells us that a subpersonal mechanism forms them by categorizing and predicating. This means that subpersonal mechanisms can

Predication and Two Concepts of Judgment

229

perform S-judgings that, on the personal level, are felt as received (Reiland 2015). In fact, some philosophers simply call them perceptual judgings (Lyons 2009). In contrast, E-judging is a personal level, epistemic act for which the agent is responsible. Thus, it’s clear that it couldn’t occur subpersonally. 4.

Judgment and Belief

Judging has been always thought to be closely related to belief. In this final section, I’ll clarify the distinction between S-judging and E-judging some more by showing the diverse ways in which they’ve been thought to be related. Some philosophers have suggested that we can use judging to define belief. Very roughly, the idea is that to believe that p is to be disposed to judge that p upon consideration. Here’s H. H. Price: We can say of someone, quite correctly, ‘he believes that Oxford will win the boat race this year’ and that he continues to believe it throughout the months of January and February. But on the traditional view, we mean by this that if at any time during that period he were to consider the proposition ‘Oxford is going to win the boat race’, an actual belief-occurrence would take place in him—a specific sort of experience which he could notice introspectively if he wished—and this proposition about the boat-race would be its object. (Price 1969: 21) Assuming that an “actual belief-occurrence” is to be thought of something like judgment, Price gets close to the above rough analysis. Similarly, here’s Wilfrid Sellars: Jones believes that-p = Jones has a settled disposition to think that-p, if the question occurs to him whether-p. (Sellars 1969: 523) Again, if the act of thinking that-p is something like judgment, then Sellars seems to subscribe to something like the rough analysis given above. Finally, here’s Uriah Kriegel’s analysis: S believes that p iff S is disposed to immediately affirm that p when p-entertaining-triggers obtain. (Kriegel 2013) Given that Kriegel is using ‘affirming’ and ‘judging’ for the same act and assuming that p-entertaining-triggers are best thought of in terms of considering p, he seems to again subscribe to this analysis.

230

Indrek Reiland

Should we understand ‘judge’ in this account in terms of S-judging or E-judging? It’s quite clear that it has mostly been intended in terms of the former. After all, the idea behind the analysis is, roughly, that if you believe that p, you would immediately, without any further thought, affirm that p upon considering it. It’s not like you need to go through the process of weighing your evidence and E-judging it again upon considering.10 Consider now a different putative relation between judging and belief. Some people have claimed that judging is the most fundamental way of forming a belief (Peacocke 1998: 88). Yet, others seem to deny this. For example, Quassim Cassam distinguishes between three fundamental ways of forming a belief: perception, testimony, and reasoning. He claims that when we take our perceptions—or perceptual seemings (see Lyons 2009; Reiland 2015)—at face value, we form a belief without judging. The same goes for testimony. It’s only in the case of reasoning that we form a belief by first performing a judgment. Since Cassam focuses on reasoning, it should be obvious that we should understand him as talking about E-judging. Indeed, E-judging is plausibly merely one way of forming a belief, whereas S-judging is involved in the formation of any propositional attitude and thus in all beliefs. A final way in which judging has been thought to be related by belief is illustrated by Nico Silins’s claim that conscious judging is a guide to belief. His final proposal is that when you answer the question whether p by judging that p, then your judging gives you immediate fallible justification to believe that you believe that p (Silins 2012). Again, since Silins is talking about conscious judging—and specifically in the context of answering a question—it’s clear that he’s talking about something like E-judging. To sum up, one view relating judging and belief tells us that to believe that p is to be disposed to S-judge that p upon considering. Furthermore, given that S-judging is involved in forming any propositional act or attitude, one might claim that it’s always implicated in the formation of belief. Yet, E-judging is merely one way (even if a very important one) of forming a belief. Finally, E-judging is perhaps also a guide to belief in the sense of giving us evidence about what we believe.

Conclusion Let me end with some terminological reflections. Given that S-judging and E-judging are different acts, one might wonder whether there’s a different term we could use for either. A plausible candidate for the predicationresultant act would be simply ‘thinking’. Thus, we could say that to predicate being F of O is to think that O is F. This would be fine were it not for the fact that ‘thinking’ is also frequently used to talk about belief. On this usage, to think that p is not to perform any particular act, but simply to believe that p (Cassam 2010: 85).

Predication and Two Concepts of Judgment

231

On the other hand, it’s hard to see what other term to use for E-judgings. In some languages, like Estonian, ‘judge’ translates as ‘decide’. This makes sense since E-judging is indeed a species of decision, a theoretical decision about what to believe rather than a practical one about what to do. However, given that in English ‘decision’ has strong practical connotations, this might not be feasible. It might therefore be that we need to continue to use ‘judge’ for both S-judgings and E-judgings. This makes it extra important that we clearly specify which one we’re talking about so as not to end in any unnecessary conundrums or talk past each other.

Notes 1. I’m here setting aside iconic, perceptual, etc., representation, which I take to mostly consist in being in a state with some sort of non-propositional content and accuracy conditions. For discussion, see Burge (2010) and Reiland (2015). 2. Contrast this with Stephen Neale’s sentence-level Holist way of understanding the Gricean program on which the notion of speaker meaning with its propositional content is the first unit of analysis and speaker reference is analysed in terms of it (Neale 2015: 259–60). It’s unclear what such a Gricean could say about the communicative intentions involved in the uses of names in (3) and (4). 3. Even though it should be clear that this doesn’t really amount to an analysis or account of quantification since our understanding of what it is for a property to be universally instantiated is plausibly derived from our understanding of everything having it, rather than vice versa. What this means is that, philosophically, we can’t really claim to understand quantification. 4. These sorts of objections have been leveled against Hanks and Soames’s view from the beginning. Neither of them is convinced, of course (see Hanks 2015: 66–74; Soames 2015: Ch. 10). 5. As an aside, and to jump a bit ahead, this is only plausible for acts of S-judging and S-judgments that are basically forceful thoughts. Acts of epistemically judging are like acts of deciding, the products of which stand as long as they aren’t changed. In fact, as we will see, epistemic judgings are ways of making up our minds about what to believe, ways of concluding theoretical deliberation analogously. This is analogous to how decisions are ways of making up our minds about what to do, and ways of concluding practical deliberation. 6. For discussion and elaboration, see Hanks (2015: 36–9) and Reiland (2017). The argument is also endorsed by François Recanati in Recanati 2016. For Soames’s reply, see Soames (2015: 219–23). 7. Hanks has, of course, proposed a solution in terms of his notion of cancellation. For criticism, see Reiland (2019). 8. For more on objectual attitudes, see Grzankowski (2016). 9. This entails that antecedently to judging, one has to have considered the question whether p (something that is not so in the case of S-judging). 10. For problems for such analyses, see Schwitzgebel (2010).

References Austin, J. L. (1962), How To Do Things With Words. Cambridge, MA: Harvard University Press. Burge, T. (2010), Origins of Objectivity. Oxford: Oxford University Press.

232

Indrek Reiland

Campbell, J. (2002), Reference and Consciousness. Oxford: Oxford University Press. Cassam, Q. (2010), “Judging, Believing and Thinking”, Philosophical Issues 20: 80–95. Clark, A. (2004), “Feature-Placing and Proto-Objects”, Philosophical Psychology 17: 443–69. Davidson, D. (2005), Truth and Predication. Cambridge, MA: Harvard University Press. Fodor, J. and Z. Pylyshyn (2015), Minds Without Meanings: An Essay on the Content of Concepts. Cambridge, MA: MIT Press. Grzankowski, A. (2016), “Attitudes Towards Objects”, Noûs 50: 314–28. Hanks, P. (2011), “Structured Propositions as Types”, Mind 120: 11–52. Hanks, P. (2015), Propositional Content. Oxford: Oxford University Press. Hanks, P. (2017), “Predication and Rule-Following”, in P. Stalmaszczyk (ed.), Philosophy and Logic of Predication. Bern: Peter Lang. Hanna, R. (2017), “Kant’s Theory of Judgment”, in Edward N. Zalta (ed.), Stanford Encyclopedia of Philosophy (Winter 2018 Edition). https://plato.stanford. edu/archives/win2018/entries/kant-judgment/ Johnston, M. (2006), “Better than Mere Knowledge? The Function of Sensory Awareness”, in T. S. Gendler and J. Hawthorne (eds.), Perceptual Experience. Oxford: Oxford University Press. Jubien, M. (2001), “Propositions and the Objects of Thought”, Philosophical Studies 104: 47–62. King, J. (2007), The Nature and Structure of Content. Oxford: Oxford University Press. King, J. (2009), “Questions of Unity”, Proceedings of the Aristotelian Society 109: 257–77. King, J. (2013), “Propositional Unity: What’s the Problem, Who Has It and Who Solves It?”, Philosophical Studies 165: 71–93. Kriegel, U. (2013), “Entertaining as a Propositional Attitude: A Non-Reductive Characterization”, American Philosophical Quarterly 50: 1–22. Lyons, J. (2009), Perception and Basic Beliefs: Zombies, Modules and the Problem of the External World. Oxford: Oxford University Press. Lyons, J. (2015), “Unencapsulated Modules and Perceptual Judgment”, in J. Zeimbekis and A. Raftopoulus (eds.), The Cognitive Penetrability of Perception: New Philosophical Perspectives. Oxford: Oxford University Press. McHugh, C. (2011), “Judging as a Non-Voluntary Action”, Philosophical Studies 152: 245–69. Moltmann, F. (2013), “Propositions, Attitudinal Objects, and the Distinction Between Actions and Products”, Canadian Journal of Philosophy 43: 679–701. Moltmann, F. (2017), “Cognitive Products, and the Semantics of Attitude Verbs and Deontic Modals”, in F. Moltmann and M. Textor (eds.), Act-Based Conceptions of Propositional Content. Oxford: Oxford University Press. Neale, S. (2015), “Silent Reference”, in G. Ostertag (ed.), Meanings and Other Things. Oxford: Oxford University Press. Peacocke, C. (1998), “Conscious Attitudes, Attention, and Self-Knowledge”, in C. Wright, B. Smith and C. Macdonald (eds.), Knowing Our Own Minds. Oxford: Clarendon Press.

Predication and Two Concepts of Judgment

233

Price, H. H. (1969), Belief. London: Routledge. Pylyshyn, Z. (2007), Things and Places: How the Mind Connects With the World. Cambridge, MA: MIT Press. Pylyshyn, Z. (2009), “Perception, Representation, and the World: The FINST That Binds”, in D. Dedrick and L. Trick (eds.), Computaton, Cognition, and Pylyshyn. Cambridge, MA: MIT Press. Recanati, F. (2016), “Force Cancellation”, forthcoming in Synthese. Reiland, I. (2013), “Propositional Attitudes and Mental Acts”, Thought 1: 239–45. Reiland, I. (2015), “Experience, Seemings, and Evidence”, Pacific Philosophical Quarterly 96: 510–34. Reiland, I. (2017), “Review of Peter Hanks’s Propositional Content”, Philosophical Review 126: 132–6. Reiland, I. (2019), “Predication and the Frege-Geach Problem”, Philosophical Studies, 176: 141–59. Reiland, I. (n.d.), “Naturalism About Propositions, Naturalism About Properties”. Russell, B. (1912), Problems of Philosophy. Oxford: Oxford University Press. Schwitzgebel, E. (2010), “Acting Contrary to Our Professed Beliefs or the Gulf Between Occurrent Judgment and Dispositional Belief”, Pacific Philosophical Quarterly 91: 531–53. Sellars, W. (1969), “Language as Thought and Communication”, Philosophy and Phenomenological Research 29: 506–27. Silins, N. (2012), “Judgment as a Guide to Belief”, in D. Smithies and D. Stoljar (eds.), Introspection and Consciousness. Oxford: Oxford University Press. Smithies, D. (2011), “What Is the Role of Consciousness in Demonstrative Thought?”, Journal of Philosophy 108: 5–34. Soames, S. (2010), What Is Meaning? Princeton: Princeton University Press. Soames, S. (2014), “Cognitive Propositions”, in J. King, S. Soames and J. Speaks (eds.), New Thinking About Propositions. Oxford: Oxford University Press. Soames, S. (2015), Rethinking Language, Meaning, and Mind. Princeton: Princeton University Press. Speaks, J. (2018), “Cognitive Acts and the Unity of the Proposition”. Unpublished MS. Toribio, J. (2018), “Visual Experience: Rich But Impenetrable”, Synthese 195: 3389–406. Twardowski, K. (1911/2017), “Actions and Products. Some Remarks on the Borderline of Psychology, Grammar, and Logic”, in F. Moltmann and M. Textor (eds.), Act-Based Conceptions of Propositional Content. Oxford: Oxford University Press. Velleman, D. and N. Shah (2005), “Doxastic Deliberation”, Philosophical Review 114: 497–534. Wittgenstein, L. (1995), Ludwig Wittgenstein: Cambridge Letters. Oxford: Blackwell.

12 How Is Logical Inference Possible? Christopher Peacocke

1.

The Puzzle

I will be stating some puzzles about logical inference and suggesting some interconnected solutions to them. The puzzles seem to me hard to solve. Whether or not the solutions I offer are correct, I hope this discussion will help by elaborating the many constraints on solutions to them. The puzzles lie at the intersection of several areas of philosophy, and any solutions to them have to draw on substantive theories in those areas. Those areas include the philosophy of mind, the theory of rationality, epistemology, the ontology of abstract objects, and the theory of rule-following—among many others. Any proposed solution to the problem is unavoidably committed to taking a stance on large-scale issues in philosophy. I start by focussing attention on a particular kind of case of logical inference made by a thinker. The kind meets three conditions. First, the logical principle involved in the inference is primitive. It is not a principle derived from more primitive principles. Second, the thinker makes the inference on the basis of her understanding of the logical connectives involved—rather than, for instance, on the basis of testimony that it is a valid principle. Third, our thinker makes the inference because she knows the principle to be valid. For brevity, I will refer to cases meeting these three conditions as ‘the target cases’ or as ‘the problem cases’ (which is not to imply that other kinds of cases do not equally raise other problems—they do). There are, every day, plenty of instances of these target cases. You and I, and many nonspecialists in philosophy or logic, frequently infer in accordance with the logical principles ‘A v B, ~B/A’; we do so on the basis of our understanding of alternation and of negation; the principle is a primitive one, it is not plausible that in ordinary thought it abbreviates some longer derivation (even if one can be given in some deduction systems); and we infer in accordance with the principle because we know it to be valid. This is just one of multiple examples of the target cases. It seems likely that these target cases are included amongst those of which the psychologist C. Lloyd Morgan was thinking when he wrote, “To this process of practical

How Is Logical Inference Possible?

235

inference through association in sense-experience, the term reasoning is applied by many writers. . . . There is, however, a narrower use of the word reasoning. This may be described as the process of passing on from proposition to proposition, the connections (transitions) between which involve, and are seen to involve, the logical relation symbolized by the word ‘therefore’ or ‘because’. Reason, in this narrower sense, implies the power of perceiving, and conceiving the logical relation as such” (Lloyd Morgan 1894: 282, Lloyd Morgan’s own italics). There are three puzzles about these target cases. Puzzle (1) concerns rationality. In the target examples, the logical principle is primitive, not derived from other principles. So how is it rationally held when the model of reaching it rationally from other principles or propositions is not available? How can acceptance of primitive principles be rational? We need a model of rationality in these cases that in the nature of the case cannot be a wholly inferential model. Puzzle (2) concerns the source of the thinker’s knowledge of validity of the principle. If in the target cases our thinker is relying on the validity of the principle, isn’t that then a further premise, and how can it be known? If the validity of the principle is said to be reached by inference from other premises, then since those inferences will involve logical principles, aren’t we on a regress? If knowledge of the validity of the principle is not attained by inference, how then is it attained? And how is acceptance of the validity rational? Puzzle (3) concerns the relation of the thinker’s knowledge of validity of the logical principle to the rationality of making the inferential transition. Whatever the nature of the source of the knowledge, that knowledge of the validity of the principle certainly seems to contribute to the rationality of the transition—certainly from the viewpoint of the thinker—but how does it do so? The puzzle can be formulated equally for logical axioms as well as for logical principles. We can consider the case in which a thinker accepts a primitive logical axiom, and does so because she knows the axiom to be valid, and knows this because of her understanding of the logical constants involved in the axiom. The axiom is primitive; that is why it is an axiom. It is not inferred from anything else; yet a thinker can know it, know it because it is valid, and this knowledge can be based on the thinker’s understanding of the logical expressions it contains. All three puzzles arise for this case of axioms too. The issues arise in a particularly sharp form in Frege’s writings. Frege had a classical rationalist conception of axioms, of the ‘general laws’ to whose truth the truth of a priori propositions could be traced back. He said that these general laws “neither need nor admit of proof”— “allgemeinen Gesetzen . . . die Selber eines Beweises weder fähig noch bedürftig sind” (1950: §3). Yet Frege’s practice in expounding his logical system in his Grundgesetze is seemingly in direct conflict with that view

236

Christopher Peacocke

of axioms. Axiom I of that system, in modern notation, is a version of the law that A→(B→A), where ‘→’ is the material conditional (2013: §18, p. 24). Frege precedes the statement of Axiom I with a ‘therefore’ (‘darum’). What precedes the ‘therefore’ is a simple argument from the truth function assigned earlier, in §12, to his sign for the conditional, to the conclusion which is the Thought expressed by Axiom I. This certainly looks at first blush like a commitment to what Axiom I states as something that admits of proof after all. The argument, as described here, is not about expressions, but about truth-functions and the establishment of a Thought, the Thought expressed by Axiom I. When someone is so concerned with rigour as was Frege, the use of ‘darum’ is not to be taken lightly. Frege also gives a parallel justification of his version of modus ponens as a rule of inference at §14, though as Tyler Burge notes, the formulations there mention expressions in a way that the apparent justification of Axiom 1 does not (2005: 332). It is of course an important question for Frege scholarship what is going on in these passages (and I will return at the end of this paper to comment on that). But the three puzzles above show that the apparent trouble in which Frege finds himself here is not just a problem generated by his own particular system and philosophy thereof. It is a puzzle for us too. It is important to circumscribe precisely the target cases in which the puzzles arise. In an important paper, Michael Rescorla has noted that what may superficially seem like a case of logical inference may in fact be explained by a Bayesian process of successively assigning subjective probabilities to various states of affairs (2009). An animal may superficially appear to be making an inference of the form A→B, A/B (or of the form AvB, ~A/B). To give an abbreviated form of the classical case of Chrysippus’s dog, an animal chasing a target may sniff at one path and, as a result of the sniffing, chase down the other path. Rescorla observes that that behaviour that might appear to be explained by an inference of the form AvB, ~A/B can equally be explained by the animal’s successive assignment probabilities to propositions of the form “The target went down this path”, “The target went down that path”. When the explanation is the successively updated probabilities assigned to these propositions, there is no genuine logical inference in the case, and hence no puzzles of the sort with which we are concerned. There are equally no puzzles (at least of the sort described above) in cases in which a transition is made automatically purely because it is of a certain form. Suppose there is a Fodorian language of thought, and some subject is disposed to infer B from the two premises A→B and A, and this disposition is explained by syntactic operations at the level of the Fodorian language of thought. Again, there is no puzzle here. The immediate explanation of any resulting inference is not the knowledge that the transition is valid, nor is the willingness to make the transition explained by the subject’s understanding of the connective ‘→’. We do, in fact, make

How Is Logical Inference Possible?

237

transitions in thought automatically much of the time; we are entitled to do so; and such transitions can transmit knowledge. Nonetheless, not all logical transitions we make in thought fall under this automatic case. It would not be a happy attempt to solve the puzzle by saying that we never really make logical inferences because we know that the principle of the inference is valid, and on the basis of our understanding of the expressions involved. So the puzzles remain. One proposal for a way out is to argue that logical laws are selfjustifying. Self-justification may indeed seem structurally the only way out of the problem. A proposal in this spirit was considered in lectures given over many years in Oxford by Michael Dummett, and it is developed in his book The Logical Basis of Metaphysics (1991). Dummett’s position is not at all that any arbitrary set of logical laws is self-justifying. On the contrary, he explicitly rejects that position, describing it as a kind of ‘logical formalism’ (1991: 208). The position that his writing suggests is rather that we can specify the meaning of a given logical operator as one that conforms to certain laws, provided that such a specification is in accordance with whatever is the correct general account of meaning, as applied to logical operators. Dummett himself was tempted by what in his later writing he called ‘justificationist’ accounts of meaning. Under that conception of meaning, a logical principle that functions as an introduction rule for a logical operator can be regarded simply as a stipulation of what is to count as a meaning-based justification for propositions that are the conclusion of the introduction rule. On the justificationist approach to meaning, a set of logical laws for a connective can be self-justifying, in that they do not collectively permit the establishment from a set of premises of anything that could not be established without that set of logical laws—a form of conservative extension requirement. That is much stricter than ‘logical formalism’. A corresponding position could be formulated for those who believe that meaning or content is individuated in terms of canonical consequences, rather than justification. This general position then allows the formulation of a proposal to answer our puzzle, a proposal that makes use of the notion of self-justification. The proposal has three parts. (a) When the relevant constraints on meaning are met, we may stipulate the meanings of logical operators as ones that conform to certain principles. (b) The meanings of our actual logical constants, in our ordinary natural language, are as if they were so stipulated. (c) The requirement in the statement of the puzzle that the principles are accepted because they are valid, and on the basis of the thinker’s understanding of the constants, is met because thinkers accept the law on the basis of their appreciation of their validity under the justificationist specifications of meanings for the constants involved. There are two salient problems with this Dummett-inspired proposal. The first can be called the ‘as-if problem’. What is the relation between the

238

Christopher Peacocke

hypothesized stipulation of meanings for some invented expression on the one hand and our actual words and concepts as expressed in our actual language, in which we as speakers and understanders have not stipulated anything at all? Take the rule of or-elimination, which states that if C follows from a set of premises {A, X1 . . . Xn}, and if C also follows from a set of premises {B, X1 . . . Xn}, then C follows from the alternation A or B together with that background set {X1 . . . Xn}. This is a primitive rule, and one that concerns our ordinary concept of alternation, as expressed by the existing English word ‘or’. Students have to work out that instances of this rule are valid. It is not something primitively written into their understanding of ‘or’, an expression they will have understood—and not merely partially—for years before considering instances of, or the general law of, alternation-elimination. From what do our students work out this rule? Plausibly, from some realization to the effect that if AvB is true, then either A is true, in which case there is a valid deduction of C, or B is, in which case there is also a valid deduction of C, so in either case, C can be deduced, all relative to the background set {X1 . . . Xn}. This mode of working out the validity of principle raises many question, some of which I will be addressing soon, but what matters for the present is that this working out does not at all mention possible justifications for the proposition A and possible justifications for the proposition B. This mode of working out mentions truth, not justification. Similar points could be made for existential quantification and the rule of existential-introduction. Dummett writes at one point, “No one can be said to understand either [the existential or the universal—CP] quantifier unless he at least knows both the introduction and elimination rules for it” (1991: 217). That seems much too strong. At most, we can say that someone who understands the quantifier must be in a position to work out both rules for both quantifiers. But the resources of understanding on which someone needs to draw in working out the rules, in making essential use of the notion of truth rather than of justification, are incompatible with the Dummettinspired suggestion for solving the puzzle. The second problem with the proposal is that, as formulated, it works only under an untenable account of meaning and content. Under a justificationist approach to meaning and content, a transition from premises to conclusion is valid just in case any means of justifying the premises can be transformed into a means of justifying the conclusion. But this would incorrectly validate the transition from the premise A to the conclusion It’s verifiable that A. That is not a truth-preserving transition, and it is truth-preservation that is to be required of valid transitions. Only if truth coincided with verifiability would this be avoided, which means that the proposal would work only with a commitment to verificationism. A dual problem arises for consequence-based accounts of meaning. If we are not identifying truth with the possibility of justification, there is also no good reason to accept an overarching constraint of conservative

How Is Logical Inference Possible?

239

extension, understood proof-theoretically, on legitimate new axioms and principles. It is fine—indeed it is positively desirable—that new axioms and principles for new vocabulary should give us means of establishing propositions in the old vocabulary we could not previously establish, provided those propositions were already true given the meaning of expressions in them. That is arguably precisely what Gödel’s theorem does for a sentence of arithmetic. It establishes that it is true, though not provable in the system to which the theorem is applied. It is good rather than bad to have that new means of establishing it.

2.

Five Desiderata

It will help us to formulate five desiderata for a good solution to the puzzle. (1) We need a philosophical account of what is involved in a thinker’s making a transition because the conclusion follows from the premises. This account has to be partially psychologistic. Psychologism about a given subject matter is the doctrine that the nature of the subject matter is constituted in part by mental properties and relations. Psychologism about the relation of logical consequence itself is verboten, for reasons so vividly explained by Frege himself. Psychologism about inference made because of appreciation of the validity of the principle involved, and based on understanding of the expressions involved, is not merely not verboten, it is required. Inference is a psychological notion; so is appreciation of validity; and so too are understanding and something’s being founded in understanding. Satisfactory theories of this psychological subject matter can hardly fail to involve the thinker’s psychology. (2) We need a philosophical account of what is involved in the thinker’s making the transition because of her understanding of the constants in question, an account that properly relates this understanding to appreciation of the validity of the transition. (3) We need an explanation of how acceptance of a primitive axiom or rule of inference can be a rational mental action. (4) We need an account of why, when an axiom (or principle) is known to be valid, this knowledge is of something that is guaranteed to be true (or truth-preserving) in the actual world, whichever world is the actual world. It holds fixedly actually, in the sense of Martin Davies and Lloyd Humberstone (1980). (5) This may already be implicit in (4), but if not, we need an account of why the knowledge of the primitive axiom or rule of inference, when known to be valid, is a priori.

240

Christopher Peacocke

The relation of broadly logical axioms or principles to metaphysical necessity, as opposed to the a priori, is more complex. The axiom schema “If Actually A, then A” should be in any good logic of “Actually”, but it is well known that many instances of “If Actually A, then A” are not metaphysically necessary. There is no general requirement, when we have ‘Actually’ in our language and conceptual repertoire, that what is logically valid should also be metaphysically necessary.

3.

A Way in

Perhaps unexpectedly, desideratum (3), the need to explain how acceptance of a primitive axiom or rule can be a rational mental action, provides a way in to addressing these issues. The mental action of accepting a principle of inference—and for the reason that it is valid—is, like many other actions, something done for a reason. What can this reason be? The reason is not a perceptual state; nor is it some sensation; nor is it some apparent action awareness; nor is it testimony, for we are concerned here with the thinker’s own appreciation of the validity of the axiom or rule and not with the case in which she is told of the validity by someone else. I suggest that in the cases with which we are concerned, the notion of an intellectual seeming, a state or event in which it seems to the subject that something is the case, has to play an indispensable role in the account of the thinker’s reasons for performing the mental action of accepting the axiom or principle. An intellectual seeming that the axiom or principle is valid plays a crucial role in the thinker’s mental action of judging that the axiom or principle is valid. Intellectual seemings are events and states that just occur to subjects. They are not themselves mental actions. Though these seemings are not themselves perceptual states, this feature of just occurring to the subject, rather than being the subject’s actions, is a property that intellectual seemings do share with perceptual experiences. Intellectual seemings are independent of judgment in multiple respects. An intellectual seeming may persist in a thinker even when its content is overruled by rational judgment, just as a perceptual illusion may persist after its content is overruled by judgment. Consider the seeming that to every property there corresponds a set of just those things having the property. We know from Russell’s paradox that this seeming is incorrect, and its content cannot be defended in its full generality. That does not make the intellectual seeming disappear. It seemed to Frege that his Axiom V was correct, though, as he himself said, it never had the same strength of intuitive support as his other axioms and rules in his Grundgesetze. This mention of intellectual seemings in an account of our understandingbased knowledge of primitive logical principles and axioms is not an ad hoc move. There is an intellectual seeming involved in rational judgment even in entirely non-logical cases—and in all cases of concept application of a kind

How Is Logical Inference Possible?

241

mentioned in the possession condition or understanding condition for the concept in question. The relevant intellectual seeming is that the concept is correctly applied in the given circumstances. In good cases of application of a concept in accordance with the concept’s understanding-condition, the relevant intellectual seeming is produced in the right way by the thinker’s meeting the understanding-condition. This applies even in the case in which the subject has an experience of an object as having a shade of what is in fact blue and the subject thereby judges the conceptual content that’s blue. It seems intellectually to the perceiver that that conceptual content is correct. It is a further mental action to make that judgment, the judgment endorsing the content of the intellectual seeming that in the given circumstances, the content is correct. The presence of an understanding-based seeming in certain cases is a feature that needs to be mentioned in an account of any concept, whether logical or non-logical. Intellectual seemings fulfil a crucial function in the theory of concepts, rationality, and judgment. They are not a mere add-on in the theory of possession conditions for concepts. Intellectual seemings allow us to steer between a Scylla and a Charybdis. Scylla in this perilous passage is any position according to which it is some kind of psychological law that a judgment with a certain content will be made in the conditions in which the thinker finds herself. This is an unacceptable Scylla because judgment is a mental action, made freely, and never follows by psychological law alone from the relevant conditions in which the judgment is made. It may be wholly reasonable to make a judgment, and unreasonable not to make it; but reflection on whether it should be made is always possible for a free, rational thinker. The Charybdis of this passages is any position according to which thinkers have inclinations to judge, inclinations that may be endorsed, but on which thinkers do not enjoy conscious states representing something as a reason to judge the content in question. According to these Charybdis positions, there are subpersonal states producing inclinations to judge, but there is no conscious mental state that provides any reason for accepting the content in question. By contrast, the safe middle way that intellectual seemings make possible is one that endorses this position: intellectual seemings are not judgments, but they can make free judgments rational, and the very content of the seemings specifies reasons that support the content in question. In good cases, the seeming is appropriately caused by the possession condition for the concept. Intellectual seemings make concept-based reasons conscious, and we can recognize this without treating judgments endorsing the contents as somehow a matter of unfree compulsion. This can be taken as a transcendental argument for the existence of intellectual seemings in the possession conditions for concepts. The conditions for the possibility of free and rational judgment, together with the reasoninvolving features of concepts, require these intellectual seemings.

242

Christopher Peacocke

So on this position, intellectual seemings are a deep feature of concept possession, something essential for judgment that is both reason-based and free. This position respects the point that concepts (as opposed to non-conceptual contents) are essentially denizens of the realm of rationality. There is, of course, a great deal more to be said about intellectual seemings, about their relation to such matters as Wittgenstein’s discussion of rule-following, and to classical conceptions of rational intuition. The position I have outlined is in fundamental agreement with Ernest Sosa’s point that “free judgments cannot gain epistemic status arbitrarily, with no rational basis for the gain” (2015: 201). All of these issues have to be matters for some other occasion. I hope just to have said enough to explain the motivation for a position under which the invocation of intellectual seemings in the logical case is a special case of a wholly general, and theoretically motivated, position about concepts and concept possession. So now let us return to the case of primitive principles and axioms and the role of intellectual seemings when a thinker accepts them rationally. Consider a thinker who rationally accepts the principle {AvB, ~B}/A and does so for the reason that it is valid. One way our thinker might work out its validity is by engaging in this thought process: AvB is true if A is true; and AvB is true if B is true; and these are the only ways for AvB to be true. But if all the premises in the principle {AvB, ~B}/A are true, it can’t be that B is true. So any way in which both premises of the principle are true is also a way in which A is also true. When a thinker comes to appreciate the validity of a principle by this route, the form of the transition is part of the explanation both of the validity and of the thinker’s knowledge of validity. It is because of the form, and the structural contribution of that form to truth conditions, that form to truth conditions that the principle is valid. The way the thinker comes to appreciate the validity depends on those features of the form. It does not, however, follow that the thinker needs to conceptualize the form as the form it is. It is enough that the content has a certain form, together with that form’s having a certain explanatory power. I suggest that the first sentence of the displayed description of a psychological process of a thinker in coming to accept rationally the validity of the relevant transition is a sentence that describes three intellectual seemings enjoyed by the thinker: that AvB is true if A is true, that AvB is true if B is true, and that these are the only ways for AvB to be true. There are corresponding judgments based on these seemings, but these judgments are made rationally only because these seemings are taken at face value. The judgments that AvB is true in these cases, and only in these cases, are not mere stabs in the dark, something no better than guesses, nor are they

How Is Logical Inference Possible?

243

made arbitrarily. The first two seemings, that AvB is true if A is true and that AvB is true if B is true, meet the condition Sosa mentions when he writes, “Some seemings need not be based on any other person-level mental states” (2015: 231), if “person-level mental states” is here intended to be restricted to conscious states. (Satisfying the understanding-condition for a concept is something done by the whole person, not something subpersonal, but it is not itself a state or event of consciousness.) The intellectual seemings are part of the required structure of personal level justification in these examples. Intellectual seemings can be mistaken. But in good cases, intellectual seemings that a content is correct in given circumstances are properly produced by an appropriate relation to a possession condition or understanding condition. A seeming that AvB is true when A is true can be appropriately produced by the thinker’s tacit knowledge that AvB is true when A is true, and this is part of the possession condition or understanding condition for alternation. This tacit knowledge contributes to the explanation of the thinker’s assessment of the truth values of complex contents on the basis of information about the truth values of the components of the complex content. Because they are founded in the tacit knowledge of the thinker, veridical and knowledgesupporting seemings are not epistemological or metaphysical mysteries. The later components of the intellectual seeming whose content is displayed above involve transitions from the initial understanding-based seemings. Such later seemings are what Sosa calls ‘resultant seemings’ (2015: 231). The transitions by which these resultant seemings are attained involve the use of logic within the scope of a seeming to generate a new resultant seeming. There is no such thing as an entirely logic-free account of the rationality of acceptance of primitive logical principles and axioms. All the same, the process of acceptance, as described, seems to be rational in a way in which a straightforwardly question-begging argument is not. The process is rational in part because it draws on an explanation of why the primitive axiom or principle is correct, is always truth-preserving. The explanation uses logic, as any explanation must. I turn now to a challenge to the position I have so briefly outlined, a challenge formulated at length in Elijah Chudnoff’s book Intuition (2013). I have been giving an account of the puzzle case that combines an understanding-based treatment of knowledge of logical validity with a role for intellectual seemings. Chudnoff argues that understanding-based seemings—what he would call intuitions—cannot provide the reliability that we need for knowledge (2013: ch. 4). He considers two cases at length: the Jordan curve theorem in topology and the axiom of choice. I will consider just his discussion of the Jordan curve theorem, which raises all the essential issues. The Jordan curve theorem (“the JCT”) states that any non–selfintersecting closed curve in a plane separates the plane into two disconnected regions, an inside and an outside. Chudnoff claims that ordinary

244

Christopher Peacocke

intuitions that the proposition expressed by the JCT is true are compelling. He writes, “We all find JCT intuitively compelling, and, on the face of it, when we do so we all gain justification for believing JCT” (2013: 129). Chudnoff’s view is that intellectual seemings (intuitions), even compelling ones, can justify without being reliable. As he writes, “In general, it is unsafe to make generalizations about all curves on the basis of what holds true for curves that can be visualized” (126). By contrast, I would say that we need to distinguish mere seemingjustifications from real justifications. Someone who finds the JCT immediately intuitively compelling, even though it takes pages of substantive argument to prove in topology texts, is almost certainly using perceptual imagination in finding it compelling. In perceptual imagination, it seems that any closed curve on a plane that we could perceive must divide the plane into an inside and an outside. But, of course, the theorem concerns arbitrary nondenumerable sets of points that are closed curves. Neither our perception nor our perceptual imagination extends to the nondenumerable. So imagination, and any thought that is merely about possible perceptions, cannot provide an adequate justification for accepting the JCT. With these points, Chudnoff would agree. But I propose that the right conclusion to draw from the example is rather that the exercise of perceptual imagination provides only a seemingjustification for the JCT. It is a mere seeming-justification because it overlooks the cases that perception and perceptual imagination cannot tell us about. A real justification does not overlook those cases, and that is what the long proof in the topology textbooks provides. Someone who has the appropriately understanding-based intellectual seemings involved in the justification we considered above for the principle {AvB, ~B}/A has a real justification, and not a mere seeming-justification, for accepting the principle. This thinker has not overlooked some cases. Mere seeming-justifications seem frequently to result from not considering the full range of cases covered by the proposition whose justification is in question. That applies both to attempts in imagination to justify the JCT and to attempts to justify Frege’s Axiom V—concepts that involve, in certain ways, their own course-of-values are left out of consideration. My position, in summary, is that when someone rationally accepts a primitive axiom or rule of inference for the reason that it is valid, and does so on the basis of her understanding or grasp of the concepts involved, the following conditions are met. First, the thinker has a real, and not merely a seeming, justification for the mental action of acceptance. Second, the intellectual seemings involved in the thinker’s reasons are properly founded in the thinker’s grasp of the concepts involved. Third, as a result of the acceptance having this rational etiology, acceptances reached in this way will be reliable. They will also be safe—will yield correct acceptances in nearby possible states of affairs—provided the

How Is Logical Inference Possible?

245

thinker employs the same methods, and the seemings are founded in the right tacit knowledge, in nearby possible states of affairs. The rationale I offered in the puzzle cases for a thinker’s acceptance of a primitive axiom or rule of inference uses the notion of a way for a proposition of a certain form to be true. The development of logic over time, and in particular its semantics, can be regarded as a progressive elaboration and precisification of this intuitive notion of a way of being true. The substitutional notion of validity often attributed to Bolzano—that a schema is valid if there are no substitution instances of it that are false—can be regarded as an intermediate stage on the road to a fuller articulation of the notion of a way of being true. It is well known that, with its tacit relativity to a background vocabulary from which the substituted expressions are drawn, the substitutional account of validity is too wide. But it is the case that each substitution instance of a schema that yields a true sentence corresponds to a way in which a proposition of that schematic form can be true. That is why the account attributed to Bolzano was certainly progress, and allowed some rigorous theorizing about validity. A Tarskian model-theoretic account of ways in which a proposition that is an instance of a schema can be true is certainly better in two respects than the substitutional account. First, it does not have the relativity to a background vocabulary, and correspondingly determines a narrower class of schemata as valid. Second, in appealing to the apparatus of reference, objects, and predication, the Tarskian account draws on the very notions that are fundamental to determining truth for sentences of first-order languages. This is a more powerful explanatory resource when we want to consider truth outright as the special case of truth in a particular designated model. We also correspondingly want a notion of validity that involves truth in all Tarskian models, not just truth of all substitution instances. For first-order standard languages, only the Tarskian notion of logical validity and logical consequence is guaranteed to be stable over an expansion of our non-logical vocabulary. This is not to say that the Tarskian notion of a model is the endpoint of the development of the notion of a way of being true. On the contrary, the very general notion of a way of being true has a certain open-ended character. We need it for any range of concepts and corresponding sentences and contents for which there is some notion of logical consequence, or something close to it. We need a notion of a way of being true for a modal proposition, a notion of a way of being true for a deontic proposition; and so forth. Even if there is a strict notion of logical constant on which these concepts go beyond the strictly logical, there is certainly a notion of consequence for modal, deontic, and other propositions. There are correspondingly modal, deontic, and other transitions that a thinker may make because she knows the propositions stand in the consequence relation, and knows that on the basis of her understanding or grasp of the concepts in question. We will need a notion of a way of being true for

246

Christopher Peacocke

all such propositions if we are to explain how here, too, a thinker can rationally accept a principle for the reason that she knows it to be valid and on the basis of her understanding. These points should not be taken to suggest that the notion of truth in a model is philosophically more fundamental than absolute truth. On the contrary, the notion of truth in a model is of interest because it formalizes the intuitive notion of a way of being true—that is, a way of being absolutely, non–model-theoretically true. Being true outright is indeed a special case of being true in a model, as it has to be if model-theoretic validity is to be of interest to us as a constraint on transitions that are truth-preserving. But in the order of philosophical explanation, absolute truth is more fundamental than truth in a model. The range of models we admit is wholly motivated by the notion of a way of being absolutely true. Constraints on model theory flow from considerations about absolute, non–model-theoretic truth and not conversely. I also suggest that the intuitive notion of a way in which something is true is an instance of Ian Rumfitt’s notion of a possibility at which something may be true (2015: ch. 2). Rumfitt proposes that logical laws are laws that hold for any implication relation, where an implication relation is a relation that satisfies the properties of Reflexivity, Monotonicity, and Cut (2015: 42).1 Rumfitt considers the following relation between A1, . . ., An and B: for any possibility in a background set of possibilities, when A1, . . ., An are all true at that possibility, then B is also true at that possibility. Rumfitt notes that this relation is an implication relation (2015: 46). If each way in which something may be true is a possibility in Rumfitt’s sense, then we equally have an implication relation in the relation in which A1, . . ., An and B stand when any way that makes all of A1, . . ., An true also makes B true. The laws of logic, taken as laws of any arbitrary implication relation, will then hold for that relation formulated in terms of ways of making something true.

4.

Issues of A Priori Status

We are inclined to say that when a thinker knows a primitive logical principle or axiom because it is valid and on the basis of her understanding of the logical expressions involved, then that knowledge is a priori. But it is a challenge to say how that can possibly be so. The standard intuitive characterization of the a priori that we offer to undergraduates is that a priori knowledge is knowledge that is justificationally independent of perception, sensation, memory, and action awareness—independent of conscious experience, broadly construed. But any sharp undergraduate given this characterization may say: If that is the test for the a priori, then the thinker who knows logical principles because they are valid does not have a priori knowledge

How Is Logical Inference Possible?

247

of them since conscious intellectual seemings are equally conscious experience, broadly construed. If you stipulatively exclude intellectual seemings from the list of mental states and events mentioned in the characterization of the a priori, then the resulting criterion will be intolerably ad hoc and of no use in philosophical explanation. But if you allow seemings into the characterization, knowledge of logical principles on the basis of their validity will, by the account you offered above, count as a posteriori after all. Is this sharp undergraduate right? Is there a rationale for omitting the intellectual seemings involved in knowledge of validity from the mental states and events mentioned in the characterization of the a priori? I suggest that there is a fundamental rationale. Perceptual experience, memory, sensation, and action awareness are world-involving in the sense that what makes an event one of these kinds involves its relations to the physical or mental world beyond the thinker’s own cognitive capacities. Bodily sensations already are mental events beyond the subject’s cognitive capacities. Genuine perceptions of the world involve events and objects beyond the subject’s cognitive capacities. Perceptual experiences that are non-veridical still involve relations to the world beyond the thinker’s cognitive capacities, but in a more complex way. What gives a class of perceptual experiences their content concerning the world is in part the fact that, when all is working properly and the subject is properly embedded in the world, they are of a kind whose instances are caused by what they are as of. There is much more to be said about these conditions, but their complexity and delicacy do not detract from their world-involving character. The same applies pari passu to events of action-awareness. By contrast, the intellectual seemings involved in the transition to rational acceptance of a primitive logical principle or axiom are worldindependent. These seemings result from the thinker’s meeting certain conditions for grasping the logical concepts in question. Possession of the tacit knowledge that a content (a Fregean Thought) of the form AvB is true just in case either A is true or B is true does not involve the empirical world of mental and physical events beyond the thinker’s standing cognitive capacities and knowledge. Nor does the seeming it produces when everything is working properly involve the world beyond the thinker’s cognitive capacities being a certain way. The content of the seeming, it should be emphasized, concerns concepts, not expressions in a public language, knowledge of which would have a certain empirical element. This distinction between the world-involving mental states and events and the world-independent mental states and events gives the materials for a principled distinction between the perceptions, memories, sensations, and action awarenesses on the one hand and the intellectual seemings on the other. We can answer the sharp undergraduate’s complaint by saying that in the characterization of the a priori, the mental states and

248

Christopher Peacocke

events of which a priori knowledge is justificationally independent are world-involving states and events. Intellectual seemings are, by contrast, world-independent. So understanding-based knowledge of the validity of a primitive logical principle or axiom is, after all, a priori. This account accords with the classical rationalist position that knowledge of a priori principles is founded in the understanding. That was a central tenet of Leibniz’s New Essays on Human Understanding (1982). A good account of how we attain a priori knowledge in any particular case has to be properly integrated with an account of what makes a content itself a priori. That is something I attempted in earlier work. For someone who thinks concepts are individuated by possession conditions of the form I discussed in A Study of Concepts (1992), there is a natural corresponding account of what makes something a priori. The corresponding account states that a priori contents are ones whose truth in the actual world is a consequence of three things, taken jointly: the possession conditions for the concepts in the content, the way those concepts are combined in the content, and the account of how the reference of the concepts is determined together with the world (the Determination Theory for the concepts). This was the account in Peacocke (1993). Later, I came to think that a concept should be individuated by its fundamental reference rule, and that grasp of any concept should rather be explained in terms of tacit knowledge of its fundamental reference rule (2008). Under this later account of concepts, there is a natural, very simple corresponding account of the a priori. A priori contents are those whose truth in the actual world is a consequence of two things, taken jointly: the fundamental reference rules for the concepts, together with the way they are composed in the content. What we have just been saying about a thinker’s rational a priori knowledge of a logical axiom or primitive principle integrates with this later account of what makes something a priori. In drawing on information about the conditions under which, for instance, a content of the form AvB is true, the thinker is employing the fundamental reference rule for alternation and drawing consequences from it and other fundamental reference rules. Our thinker’s rational steps in thought in coming to accept a primitive axiom or principle of inference trace out steps that would be involved in showing that the content is a priori under this later account of the a priori.2 We are also now in a position to show how the account meets desideratum (4) of our list of five desiderata above, viz. that the account explain why when primitive axioms and rules are known in the way outlined, their content holds Fixedly Actually—holds in the actual world, whichever world is actual. The information drawn upon in the intellectual seemings that explain that knowledge is information about the contribution made to truth conditions by the various components of a content, a Fregean Thought. Those principles will hold in any world and

How Is Logical Inference Possible?

249

therefore in any world which is the actual world. So what follows from those principles holds Fixedly Actually. We also have answers to our initial three puzzles. The proposed answer to the first puzzle is that intellectual seemings properly based on understanding-conditions can provide a source of rational judgment that gives us a model of rationality that is not purely inferential. The answer to the second puzzle, concerning the source of a rational thinker’s knowledge of the validity of a primitive principle, is that intellectual seemings properly founded in the understanding of the concepts in question can provide, via what Sosa called resultant seemings, ways of coming to know the validity of the principle in question. On the third puzzle, the treatment offered above gives the knowledge of the validity of the principle, reached via the right kind of intellectual seemings, a rational role in acceptance of the principle at the level of rational, free judgment. Our sharp undergraduate may still not be satisfied. The sharp undergraduate may agree that no particular relations to the empirical physical and mental worlds are required by the states and seemings involved in rational acceptance of a logical axiom or primitive logical principle. But what about relations to truth functions themselves, or to such higherlevel properties of contents as being true if at least one of its constituent contents is true? Are we not being equally arbitrary if we exclude consideration of such relations to the abstract world in characterizing the a priori? And if we do not exclude them, do we really have a contrast between the logical cases and the clearly a posteriori cases? These questions clearly need answers, and I turn now to a general position that can address them.

5.

Individuation Precedes Representation

We might try replying to our sharp undergraduate that there is no need to take logical constants as referring to anything at all. They refer, we might reply, neither to (unsaturated) truth functions nor to properties specifying the nature of the dependence of the truth of a complex content on the truth or falsity of its constituents. This reply is fine as far as it goes—which is not very far at all. The reply is inadequate for one smaller reason and one much larger reason. The smaller reason is that if we consider certain logics in which there is quantification over truth-functional operations, we will be required to take the logical vocabulary as referential. But it is plausible that knowledge of axioms and primitive principles in this logic is still a priori, so the challenge remains. The much larger reason the reply will not meet the need is that there are commitments to elements of vast domains of abstract entities of various kinds in the abstract sciences of number theory, set theory, category theory, and so forth. It is equally plausible that the axioms and any primitive principles of inference for these sciences are

250

Christopher Peacocke

also a priori. In these areas, the reply that there is no reference to abstract objects is completely implausible. I propose instead a general position on abstract objects, abstract properties, and abstract entities of any category that can be summarized in the slogan ‘Individuation Precedes Representation’. Unpacked, the slogan is intended to summarize a twofold thesis. Part (1) of the thesis states that for any abstract entity, there are conditions that individuate that entity; and correspondingly, for any given type of abstract entity, there are conditions of a general kind, instances of which individuate abstract entities of that type. Part (2) of the thesis states that to think about a particular abstract object (in a canonical way) is to have tacit knowledge of its individuating condition, knowledge manifested in judgments and other representational states. Similarly, to think about a given type of abstract object (in a canonical way) is to have tacit knowledge of the type of condition that individuates objects of that type, knowledge manifested in judgments and other representational states. In the case of the natural numbers, each natural number is individuated by the condition for it to be the number of Fs for an arbitrary concept F. That is what makes the natural number the number it is. Tacit knowledge of this individuating condition is shown in a thinker’s use of the numerical quantification “there are n Fs”. In the case of the real numbers, each real number is the ratio of a possible pair of magnitudes. Being a certain ratio of a pair of possible magnitudes is what makes the real number the entity it is. Tacit knowledge of that individuating condition can be shown in a thinker’s use of real numbers in measuring ratios of magnitudes in the world. Further development of the cases of the natural and the real numbers is given in The Primacy of Metaphysics (Peacocke, 2019: ch. 5). Here, our concern is with such abstract objects as truth functions or, better, such complex abstract properties as are specified by phrases of the form “the property of a complex condition of being true just under such and such conditions”. A subject can be thinking about a particular truth function or corresponding abstract property because her fundamental practice in assessing the truth value of complex contents on the basis of the truth or falsity of the constituents is given by precisely that function or abstract property—just as a thinker’s use of the numerical quantifier there are 3 Fs corresponds to the first-order condition with identity that individuates the number three. In both the number case and the logical case, this can be true without the thinker herself conceptualizing any connection between individuation and representation, or even conceptualizing facts about individuation at all. The thesis that Individuation Precedes Representation in these abstract domains is a philosophical theoretical thesis about the relations between ontology, individuation, and representation. It is not something the ordinary thinker is supposed to be formulating—or even possess the conceptual apparatus for formulating.

How Is Logical Inference Possible?

251

The thesis that Individuation Precedes Representation is a version of a metaphysics-first view of the domain of abstract objects in relation to intentional contents and meanings concerning those abstract objects. In general, a metaphysics-first view of the relation between a domain of entities and intentional contents or meanings about them states this: that the metaphysics of entities in the domain is prior, in the order of philosophical explanation, to both the theory of intentional contents concerning elements of that domain and to the theory of meaning for sentences about that domain. The thesis that Individuation Precedes Representation is a version of the metaphysics-first view because the very metaphysics of each abstract entity, the correct statement of what individuates that entity, is drawn upon in giving an account of what it is to represent that entity (in a canonical way) in thought or language. The point of interest about the thesis that Individuation Precedes Representation is that it shows that the range of cases in which a metaphysicsfirst thesis is correct goes beyond those in which there is causal interaction between elements of the domain in question and thinkers who employ the intentional contents or meanings. One familiar kind of domain in which the metaphysics-first view is plausibly correct is in the relation between spatial and temporal properties, relations, and magnitudes and the intentional contents of perceptual states and events concerning these spatial and temporal properties, relations, and magnitudes. Causal relations between instances of these spatial and temporal properties, relations, and magnitudes and perceivers of them have to be mentioned in an account of what it is to be capable of perceptual states concerning them. Furthermore, there is no plausibility in the spatial and temporal cases that those properties, relations, and magnitudes are in turn equally dependent upon their relations to perceivers or thinkers. The reality the intentional contents concern is, in those cases, wholly mind-independent. The abstract cases in which Individuation Precedes Representation show that we can have an explanatory priority of the metaphysics even in the absence of such causal interaction with the subject matter. We should be open to the possibility that there may be other instances of this phenomenon too. The position I am sketching and advocating here diverges in some fundamental respects from that of Carnap in his famous paper “Empiricism, Semantics, and Ontology” (1956). According to Carnap, for each fragment of a language, there is a set of “rules for forming statements and for testing, accepting, or rejecting them” (208). These rules, which presumably include axioms and principles of inference, settle certain questions that can be asked within that fragment of the language. These are said to be “internal” questions (1956: 206). Questions not so settled are said to be “external”. Carnap said of philosophers who attempt to engage in substantive discussion about those external questions that they “have so far not given a formulation of their questions in terms of the common scientific language. Therefore our judgment must be that they have not

252

Christopher Peacocke

succeeded in giving to the external question and to the possible answers any cognitive content. Unless and until they supply a clear cognitive interpretation, we are justified in our suspicion that their question is a pseudoquestion” (209). If we endorse the principle that Individuation Precedes Representation, the landscape looks very different from one conforming to Carnap’s description. First, the conditions that individuate abstract objects actually provide a rationale for what Carnap calls ways of speaking, subject to “rules” (206). The principles governing abstract objects of a given kind are not just rules of language, but are rather truths justified by the nature of their subject matter. These so-called rules of language are not purely conventional. Second, we can make sense of there being principles that at a certain stage of intellectual development are not then acknowledged, and so in no plausible sense are rules of the current language, but which are correct for the abstract entities in question when individuated in a certain way. Consider the omega-rule, that if in a system we can prove each proposition of the form F(t), where t is a numeral for a natural number, then the proposition For all natural numbers n, F(n) is also provable in the system. If that is not currently accepted, it ought to be added as a rule given the way in which each natural number is individuated, in terms of its application conditions in numerical quantifications of the form there are n things that are so and so. The same point applies to the Gödel sentence for any arithmetical system. Third, the rationales for the so-called rules of language, and the rationales for adding additional principles, are not purely internal in Carnap’s sense. The claims that natural numbers are individuated by their role in numerical quantifications, that real numbers are individuated as ratios of possible magnitudes of a given type, are claims of substantive philosophy, of the metaphysics of abstract objects. They are answers to external questions, they are cognitively significant, and they do not seem to be some kind of nonsense. I would, of course, make these points far beyond the case of abstract objects (as Carnap equally would—and did—for his position too). Though our present focus has been on abstract objects and the a priori, the same issues about rationales for what Carnap calls “rules of language” and their relations to a substantive metaphysics arise across the board. Carnap was also famous for his Principle of Tolerance, also formulated in that same paper “Empiricism, Semantics, and Ontology”. Carnap stated the Principle in italics: “Let us be cautious in making assertions and critical in examining them, but tolerant in permitting linguistic forms”. Despite the fundamental differences between Carnap’s conception and what I have been advocating, there is equally a Principle of Tolerance that can be endorsed if we accept that Individuation Precedes Representation. Any domain of abstract objects, if properly and consistently individuated

How Is Logical Inference Possible?

253

by a set of conditions, possibly including application conditions (as we have for the natural and real numbers), is intelligible. Pragmatic considerations, just as Carnap said, should determine whether we actually employ that domain of abstract objects in everyday and in scientific thought.

6.

Concluding Sympathetic Observation on Frege

The tension I noted in Frege lay in the combination of his view that, for axioms, justification is neither necessary nor possible, together with his actual practice of justifying axioms and primitive inference rules in his Grundgesetze. The position outlined above suggests a sympathetic attitude towards the problematic passages in the Grundgesetze. What Frege offers as axioms are indeed primitive in the system. We can, however, ask what would have to be true of someone who understands the new formal language Frege introduces in that book. Someone who understands the language, and who accepts the axioms and rules because they are valid, will have certain reasons. What in Frege’s text precedes the “therefore”, the “darum”, just before those axioms and rules are stated are the formulations of the reasons such a person has for rational acceptance. Those reasons start with the understanding-based knowledge (based on intellectual seemings) of lines of the truth table for the relevant connective and then proceed by consideration of relevant cases to a conclusion of the outright truth of the axiom or to the truth-preserving property of the rule of inference. Someone could use Frege’s language with understanding, and accept the axioms for the reason that they are valid, and make transitions in the language for the reason that they are truthpreserving, only if she were capable of such a process of acceptance.3 Frege is carrying out a task that anyone with a broadly rationalist view of the laws of logic, and more generally of the laws of the abstract sciences, is obliged to undertake: that of showing how acceptance of these laws can be rationally justified on the basis of principles in one way or another available to those who grasp the Thoughts in question.4

Notes 1. Where X and Y are sets of sentences and A and B are individual sentences, the relation ⇒ satisfies Reflexivity if A ⇒ A; it satisfies Monotonicity provided that if X ⇒ B, then X, A ⇒ B; and it satisfies Cut provided that if X ⇒ B for all B in Y and Y ⇒ A, then X ⇒ A. 2. Here, there are parallels with the account of tacit knowledge and its relation to semantic theory in Martin Davies’s account (1981) and with my own treatment of computational characterizations at level 1.5 in Peacocke (1986). 3. It will be clear to Frege scholars and to other close readers that here I am in partial disagreement with Tyler Burge’s discussion in his Truth, Thought, Reason (2005: 328–38) of the relevant passages in the Grundgesetze. This is not the place for extended interpretational discussion, but I do briefly note some points at issue. First, by way simply of clarification, the position I have advocated in

254

Christopher Peacocke

the text above does not essentially involve linguistic expressions in the thinker’s rational justification for the axioms and primitive rules. Everything can be carried out at the level of sense, Thoughts, and the truth functions corresponding to the expressions that are the traditional propositional logical constants. Second, Burge writes, “It seems to me plausible that if one understands Axiom I, one realizes that it is true” (2005: 336); and “Understanding the content of an axiom suffices to warrant one in believing it” (2005: 338). Axiom I, a Fregean version of the law A→(B→A), is far from immediately obvious to everyone who understands the material conditional. It is not a condition on understanding the material conditional that the Axiom is found immediately obvious without any working out at all. It seems to me rather that the following, weaker claim is true: if one understands the material conditional, one is in a position to work out from one’s understanding that Axiom I is true. The working out, on my view, draws on the understander’s tacit knowledge of the truth function referred to by (or corresponding to) the material conditional. I do not, of course, claim that Frege used the notion of tacit knowledge at this point. (Frege equally, in the relevant passages, did not use the notion of what is involved in understanding the logical operators.) A third issue is whether we should really regard Frege as offering fundamentally the same rationale for the Axiom in both the Begriffsschrift and the Grundgesetze. Once we have the sense/reference distinction, and a sense is individuated by the condition for something to be the reference of that sense, it is far from clear that the account of understanding, as grasp of sense, can still be the same as whatever was being relied upon in the Begriffsschrift. The interpretational issues are complex. It is also highly likely that there is some unresolved indeterminacy in Frege’s own thought on the more philosophical issues. 4. Earlier versions of this material were presented at a 2015 conference in Oslo organized by Anders Nes; in 2015 at Saul Kripke’s seminar at the CUNY Graduate Center in New York; at a 2016 workshop on logic in psychology and philosophy organized by Jacob Beck at York University, Toronto; at the 2016 New College of the Humanities London Conference on Judgement; at the 2016 Joint Columbia/PSL Philosophy Seminars in Paris; and at a 2017 meeting in Brooklyn of the Rutgers/Barnard/Columbia Philosophy of Mind Workshop. I learned from the discussions at all these events. Special thanks to David Chalmers and Ram Neta for pointing out a mistake in the version presented to the Brooklyn workshop and to Paul Boghossian for very helpful extensive recent discussions. I also received valuable comments that influenced the present text from Jacob Beck, Simon Blackburn, Jerome Dokic, Pascal Engel, E. J. Green, Saul Kripke, John MacFarlane, Jake Quilty-Dunn, Michael Rescorla, Susanna Schellenberg, Nico Silins, and Sebastian Watzl.

References Burge, T. (2005), Truth, Thought, Reason: Essays on Frege. Oxford: Oxford University Press. Carnap, R. (1956), “Empiricism, Semantics, and Ontology”, repr. in his Meaning and Necessity. Chicago: University of Chicago Press. Davies, M. (1981), “Meaning, Structure, and Understanding”, Synthese 48: 135–61. Davies, M. and L. Humberstone (1980), “Two Notions of Necessity”, Philosophical Studies 38: 1–31. Dummett, M. (1991), The Logical Basis of Metaphysics. Cambridge, MA: Harvard University Press.

How Is Logical Inference Possible?

255

Frege, G. (1950), The Foundations of Arithmetic, trans. J. L. Austin. Oxford: Blackwell. Frege, G. (2013), Basic Laws of Arithmetic, trans. P. A. Ebert and M. Rossberg with C. Wright. Oxford: Oxford University Press. Leibniz, G. (1982), New Essays on Human Understanding, trans. P. Remnant and J. Bennett. Cambridge: Cambridge University Press. Lloyd Morgan, C. (1894), An Introduction to Comparative Psychology. London: Walter Scott. Peacocke, C. (1986), “Explanation in Computational Psychology: Language, Perception and Level 1.5”, Mind and Language 1: 101–23. Peacocke, C. (1992), A Study of Concepts. Cambridge, MA: MIT Press. Peacocke, C. (1993), “How Are a Priori Truths Possible?”, European Journal of Philosophy 1: 175–99. Peacocke, C. (2008), Truly Understood. Oxford: Oxford University Press. Peacocke, C. (2019), The Primacy of Metaphysics. Oxford: Oxford University Press. Rescorla, M. (2009), “Chrysippus’s Dog as a Case Study in Non-Linguistic Cognition”, in R. Lurz (ed.), Philosophy of Animal Minds. Cambridge: Cambridge University Press. Rumfitt, I. (2015), The Boundary Stones of Thought. Oxford: Oxford University Press Sosa, E. (2015), Judgment and Agency. Oxford: Oxford University Press.

Contributors

Brian Ball is Senior Lecturer in Philosophy at New College of the Humanities, London, and Associate Member of the Faculty of Philosophy at the University of Oxford. He works in the philosophies of mind and language, epistemology, and metaphysics and has published papers in these areas in journals including Analysis, Erkenntnis, Mind and Language, Philosophical Psychology, and Philosophical Quarterly. Simon Blackburn is Professor Emeritus of Philosophy at the University of Cambridge, halftime Research Professor at the University of North Carolina, Chapel Hill, and Professor of Philosophy at the New College of the Humanities. He was Editor of Mind from 1984 to 1990. He is the author of Spreading the Word (1984), Essays in Quasi-Realism (1993), Ruling Passions (1998), and other works. Martin Lin is Associate Professor at Rutgers University, New Brunswick. He is the author of Being and Reason: an Essay on Spinoza’s Metaphysics (Oxford University Press, forthcoming) and articles on metaphysics and philosophy of mind in the seventeenth century. Friederike Moltmann is research director at the French Centre Nationale de la Recherche Scientifique (CNRS) and in recent years visiting researcher at New York University and visiting professor at the University of Padua. Her research focuses on the interface between natural language semantics and philosophy (metaphysics, but also philosophy of mind, philosophy of language, and philosophy of mathematics). She is author of Parts and Wholes in Semantics (1997) and Abstract Objects and the Semantics of Natural Language (2013), as well as numerous articles in philosophical and linguistic journals and edited volumes. Daniel Morgan is an associate lecturer at the University of York and has previously held postdoctoral positions at Oxford, University College London, and Barcelona. His main current research interests are indexicality, moral and epistemic praiseworthiness, and addiction.

Contributors

257

Alexandra Newton received her PhD in philosophy at the University of Pittsburgh, where she wrote a dissertation on Kant on logical form. She has worked as a ‘wissenschaftliche Assistentin’ at the University of Leipzig in Germany and is currently employed as an assistant professor of philosophy at the University of Illinois at Urbana-Champaign. Her interests include all topics discussed by Kant’s works, with an emphasis on the role of self-consciousness in logic, metaphysics, practical philosophy, and aesthetics. Christopher Peacocke is Johnsonian Professor of Philosophy at Columbia University, Collaborative Fellow at the Institute of Philosophy, School of Advanced Study, University of London, and Visiting Professor of Philosophy at New College of the Humanities. He was Waynflete Professor of Metaphysical Philosophy at the University of Oxford from 1989 to 2000. He is the author of nine books, including Sense and Content (1983), A Study of Concepts (1992), Truly Understood (2008), and The Primacy of Metaphysics (2019). Paul Redding is Emeritus Professor of Philosophy at the University of Sydney. His interests focus mainly on the history of German idealism and, in particular, the conceptions of logic at the heart of that movement. His books include Continental Idealism: Leibniz to Nietzsche (2009), Analytic Philosophy and the Return of Hegelian Thought (2007), and Hegel’s Hermeneutics (1996). Indrek Reiland finished his PhD at University of Southern California in 2014 and is currently a Teaching Fellow at the University of Edinburgh. He works in philosophy of language and philosophy of mind, with a focus on linguistic meaning, propositional content, and perceptual experience. His work on these topics has appeared in Philosophy and Phenomenological Research, Philosophical Studies, Synthese, American Philosophical Quarterly, and Pacific Philosophical Quarterly, among other places. Maria van der Schaar is Associate Professor in philosophy of logic and its history at the Institute of Philosophy, Leiden University. Her historical research focuses on early phenomenology, the origins of analytic philosophy, and early modern philosophy. Her philosophical work stands in the tradition of intuitionism and Swedish proof-theory, especially the work of Per Martin-Löf, and she has proposed accounts of judgment, assertion, knowledge, and meaning from this perspective. Unique in her research is the idea that the first person cannot be neglected in logic and in philosophy in general. She aims to bring problems in analytic philosophy to a deeper level of understanding by introducing ideas coming from phenomenology, the Kantian tradition, and early modern philosophy.

258

Contributors

Mark Siebel is full professor for theoretical philosophy at the University of Oldenburg in Germany. He did his PhD on Bolzano’s concept of consequence at the University of Hamburg. His main areas of research are (formal) epistemology, philosophy of language, Bolzano, Kant, and need-based justice. Publications include two monographs (on Bolzano and memory), two anthologies (on Bolzano and communication), a Festschrift for his dissertation adviser Wolfgang Künne, and articles in journals such as Analysis, Erkenntnis, Kant-Studien, Linguistics and Philosophy, Philosophical Studies, and Synthese. Mark Siebel was chairman of the German Society for Analytic Philosophy for many years. Peter Simons is Emeritus Professor of Philosophy at Trinity College Dublin. He specializes in metaphysics and ontology, pure and applied; the philosophy of logic and mathematics; and the history of philosophy in Central Europe in the nineteenth and twentieth centuries. He is the author or co-author of five books and some 300 articles. He is a member of the British, Irish, European, and Polish Academies. Mark Textor is Professor of Philosophy at King’s College London. He works on philosophy of language and mind as well as history of analytic philosophy and Austrian philosophy. His latest book is Brentano’s Mind (OUP 2017).

Index

about/concerns distinction 16, 202–14 acknowledgement 52, 116–17, 132–4, 139, 141–2, 145–6, 155 action-product distinction 15–16, 174–5, 192–5 affirmation 28, 30–2, 34–9, 41–2, 58, 116, 132, 168 Ajdukiewicz, Kazimierz 175–6 Allison, Henry 68–9 Anderson, R. L. 72 anti-psychologism 14, 120, 175 Aquinas, Thomas 28–9, 176 Aristotelianism 55, 91 Aristotle 2, 94, 103–6, 176 Armstrong, David 117–18 Arnauld, Antoine 2 Asay, Jasmin 145 assent 4, 6–7, 10, 30, 49–50, 55–6, 61 attitudinal objects 9, 15–16, 180–99 Austin, J. L. 227 axioms 53, 235–6, 239–40, 242–9, 251, 253 Barth, Christian 61 being: predicative 74–7; veritative 73, 77–9 Blackburn, Simon 15 Bolzano, Bernard 11, 14, 19n27, 62, 110–23, 169, 174, 245 Bradley, Francis Herbert 2–3 Brandom, Robert 99 Brentano, Franz 7, 10–11, 14, 15, 116, 129–47, 153–6, 158, 163, 165–9, 173, 176–7, 206, 208 British idealism 2–3, 91 Burge, Tyler 253–4n3 Carnap, Rudolf 251–2 Cassam, Quassim 230

Chudnoff, Elijah 243–4 clear and distinct perception; Descartes and 11, 27, 30, 47, 168; Locke and 51 combination and separation 28, 68, 70, 173 conative attitudes 12, 29 conatus (Spinoza) 26, 32–7, 39–40 confidence 122–3 context sensitivity 98 contradiction 77 conviction 50, 110 copula 66–83, 73, 174 correctness 14–15, 129–47 credences 37 Davidson, Donald 181, 183–4, 188, 196–7 de dicto and de re 97, 99–100, 102–4 Della Rocca, Michael 35 demonstrative knowledge (in Locke) 47–8 denial 6–7, 26, 30–1, 37–42, 60, 101, 116, 132, 168 Descartes, René 5–6, 10, 12–13, 18n8, 26–31, 37, 42, 79, 87n34, 166, 168 desire 4, 12, 30, 35, 144, 153, 155, 166, 181–2, 191–6 dissent 4, 7, 38, 61, 50 doubt 38–42, 50, 116, 177 doxastic voluntarism (in Descartes) 5, 30–1 Dummett, Michael 237–8 Ehrenfels, Christian von 140–1 enthusiasm 56–7 epistemic theodicy 26–42 error: Bolzano and 114, 119–21; Descartes and 5, 26–9, 42; Locke

260

Index

and 47, 49–51; Spinoza and 33, 37–8, 42 evidence 6, 17, 41, 50, 53, 218, 225–8, 230 faith 50, 56–8 fallibilism 49–50 feelings 151–63 Field, Hartry 143 Findlay, John N. 106n3 Fine, Kit 190 first-person point of view 13, 46–7, 59; in Descartes 62; in Hegel 100; in Leibniz 51–2, 54, 57–8; in Locke 47–51, 54; see also third-person point of view force (assertoric) 1, 12–13, 58–62, 66, 71–2, 143–4, 146, 218, 223–5; and content 66, 76, 83 foundationalism (epistemic) 120–1 Frege, Gottlob 11, 17, 52–3, 59, 62, 110–12, 116, 118, 136, 169, 204–5, 235–6, 239–40, 253, 253–4n3; ideography 46; third realm 115 Frege-Geach problem 223–4 Geach, Peter 1, 7 Gibbard, Allan 155 Ginsborg, Hannah 137–8 Grice, H. P. 203–4 grounding (in Bolzano) 115 Hanks, Peter 12, 218–21, 223–4 health (in Brentano) 130 Hegel, Georg Wilhelm Friedrich 14, 76–7, 91–107 Heidegger, Martin 85n19 Hobbes, Thomas 5–6 Huemer, Wolfgang 10 Hume, David 2, 158–60, 167 Husserl, Edmund 115, 170, 175–6 I (logical) 74, 79–82, 105 identity (and identity statements) 53–4, 56, 58, 67, 74, 186 imagination 33–4, 36, 38–9, 57, 146, 157, 244 indexicality 94, 100–1, 170–1, 177, 226 inference 14, 17, 55, 59, 92–3, 102, 117–21, 123, 162, 234–53 inferentialism 99 intellect 5, 12, 27–9, 31

intellectual seeming 17, 228, 240–4, 247–9 intentionality 9, 11, 16, 36, 166–7, 183 James, William 178n21 Johnson, W. E. 95–7 judgment stroke (Frege) 59–60 Kant, Immanuel 7, 13, 66–88, 105, 115, 158–60, 169, 226 Kriegel, Uriah 8, 141, 205–11, 213, 229 Kripke, Saul 102 Leibniz, Gottfried Wilhelm 13, 19n27, 46, 48, 51–62, 95, 97, 102, 122, 248 Leśniewski, Stanisław 175–6 Lewis, C. I. 91–2 Lewis, David 21n41, 92 Lin, Martin 12–13 Lloyd Morgan, C. 234–5 Locke, John 2, 7, 10, 13, 45–59, 61–2 Longuenesse, Béatrice 68–9, 72–3 Losonsky, Michael 57 Łukasiewicz, Jan 175–6 McHugh, Conor 227 Martin, Wayne 7 Marušić, Jennifer 63n11 material inference (Russell) 91–2 mathematics 51–3 Meinong, Alexius 10, 168–9, 173, 177 mental act 1, 3–5, 8, 11, 30–1, 40, 48, 58, 61–2, 66, 111–13, 132–3, 135, 139, 142, 168, 175, 177, 180–1, 191, 226, 239–41, 244 Merricks, Trenton 10 modal logic 92, 97–8, 101–2 Moltmann, Friederike 9, 15–16, 222–3 mood 61, 227 Moore, George Edward 2–3, 11, 134–6 Morgan, Daniel 16 naturalism 217; in Bolzano 118; in Spinoza 31, 41–2 Neale, Stephen 231n2 negation 7, 28, 58, 94–7, 223–5 negative existentials 15, 169 Newton, Alexandra 13 Nicole, Pierre 2 non-existent objects 15, 145–6, 170, 174 normativity 41–2, 45–7, 99, 121–3, 138, 155–63

Index objective sentence (sentence in itself) 14, 111–14 opinion (in Locke) 47, 55 opposition 67, 78, 132 Parfit, Derek 151–2, 154, 158 particles 58–62 Peacocke, Christopher 17 Perry, John 205, 211 Plantinga, Alvin 92 Port-Royal 2, 7 Price, H. H. 229 principle of extensionality 91–2 principle of plenitude 105–6 principle of sufficient reason 79 Prior, Arthur 101, 106n3 probability 13, 47, 49, 51, 55, 61, 118, 123, 151 psychologism 11, 13, 46–7, 54, 62, 239; see also anti-psychologism Quine, Willard Van Orman 1, 203 reasons 151–2, 154–8, 160, 162 Recanati, François 215n2 Redding, Paul 14 Reich, Klaus 70–1 Reid, Thomas 2, 5–6, 18n8 Reiland, Indrek 16–17 Rescorla, Michael 236 Rorty, Richard 161 Rumfitt, Ian 246 Russell, Bertrand 1, 3, 9, 11, 91, 97, 176, 221 Ryle, Gilbert 154, 156–7, 162 Sainsbury, Mark 204–5 Schaar, Maria van der 13 Schlick, Moritz 140 Sellars, Wilfrid 229 sensations 151–3, 158–9 self-consciousness 13–14, 67–8, 73, 76–82, 100 self-evidence 52–3, 136–42 sensitive knowledge (in Locke) 47–8 Siebel, Mark 14

261

sign of assertive force (in Leibniz) 60–2 Silins, Nico 230 Simons, Peter 15 Smith, Adam 160–1 Soames, Scott 218, 221 sociality 99–100, 159 Sosa, Ernest 5–7, 242–3 Spinoza, Baruch 12–13, 26–7, 31–42, 87n34 Stalnaker, Robert 21n41 Steinberg, Diane 33, 35 Steinberg, Justin 33 Stoicism 79, 87n34, 95–6, 102, 105 Strawson, P. F. 203–4 suspense of judgment 6, 26, 29–31, 39, 50 syncategorematic term 58, 169 Szabó, Zoltán Gendler 208 Tarski, Alfred 175–6, 245 Textor, Mark 14–15 third-person point of view 13, 46–7; Leibniz and 57–8; Locke and 47–8, 51, 56, 62; see also first-person point of view thought (Gedanke) (Frege) 62, 112, 248 thoughts, mere (Bolzano) 116–17 truth 129–47, 173–4; formal 82; material 82; relative vs. absolute 170–1 truthmakers 15, 174, 190 Twardowski, Kazimierz 9, 11, 15, 165–77, 180–1, 192–4, 222–3 unity of apperception (analytic and synthetic) 69, 74–5, 78, 80 universal language (universal characteristic) (Leibniz) 52, 58–60, 95, 97, 103 will 27–31, 35, 153, 168 Williams, Bernard 5, 10, 19n22, 19n23 Williamson, Timothy 45 Wittgenstein, Ludwig 7, 9, 10, 11, 81, 221 Wolff, Christian 71–2