Kant's Critique of Pure Reason: A Critical Guide 1107074819, 9781107074811

Kant's monumental book the Critique of Pure Reason was arguably the most conceptually revolutionary work in the his

587 118 5MB

English Pages 310 [313] Year 2017

Report DMCA / Copyright


Polecaj historie

Kant's Critique of Pure Reason: A Critical Guide
 1107074819, 9781107074811

Table of contents :
Half title
List of Contributors
List of Translations and Abbreviations
1 Kant on the Distinction between Sensibility and Understanding
2 Knowledge and Its Object
3 Transcendental Idealism and the Transcendental Aesthetic: Reading the Critique of Pure Reason Forward
4 Kant on the Ideality of Space and the Argument from Spinozism
5 How Precise Is Kant's Table of Judgments?
6 Kant's ''Transcendental Deduction''
7 Kant's Critique of the Layer-Cake Conception of Human Mindedness in the B Deduction
8 The Critical and ''Empty'' Representation ''I Think''
9 Kant's Mathematical Principles of Pure Understanding
10 Kant's Dynamical Principles: The Analogies of Experience
11 The Refutation of Idealism
12 The Antinomies: An Entirely Natural Antithetic of Human Reason
13 The Ideal of Reason
14 Knowledge, Discipline, System, Hope: The Fate of Metaphysics in the Doctrine of Method
List of References

Citation preview


Kant’s monumental book the Critique of Pure Reason was arguably the most conceptually revolutionary work in the history of philosophy, and its impact continues to be felt throughout philosophical debates today. But it is a notoriously difficult work whose basic meaning and lasting philosophical significance are both subject to ongoing controversy. In this Critical Guide, an international team of leading Kant scholars addresses the challenges, clarifying Kant’s basic terms and arguments and engaging with the debates that surround this central text. Providing compact explanations along with cuttingedge interpretations of nearly all of the main themes and arguments in Kant’s Critique, this volume provides well-balanced arguments on such controversial topics as the interpretation of Kant’s transcendental idealism, conceptualism and nonconceptual content in perception, and the soundness of his transcendental arguments. This volume will engage readers of Kant at all levels. james r. o’shea is Professor of Philosophy at University College Dublin. He is the author of Wilfrid Sellars: Naturalism with a Normative Turn (2007) and Kant’s “Critique of Pure Reason”: An Introduction and Interpretation (2012), and the editor of Sellars and His Legacy (2016).

cambrid ge c ritical g uides Recent titles in this series: Kant’s Groundwork of the Metaphysics of Morals edited by je ns t i m m e r man n Kant’s Critique of Practical Reason edited by and rews re at h and je n s tim m er m an n Wittgenstein’s Philosophical Investigations edited by arif ahme d Kierkegaard’s Concluding Unscientific Postscript edited by rick an tho n y f u rtak Plato’s Republic edited by mark l. mcpherran Plato’s Laws edited by c h r i s to ph e r b o b o n i c h Spinoza’s Theological-Political Treatise edited by y i t z hak y. m e l am e d and m i ch ael a. ro sen t h al Aristotle’s Nicomachean Ethics edited by j o n m i l l e r Kant’s Metaphysics of Morals edited by l ara de n is Nietzsche’s On the Genealogy of Morality edited by s imo n may Kant’s Observations and Remarks edited by r i c hard ve l k l ey and susan sh ell Augustine’s City of God edited by jame s we tze l Descartes’ Meditations edited by k are n de tle f se n Kant’s Religion within the Boundaries of Mere Reason edited by g o rdo n mich also n Kant’s Lectures on Anthropology edited by alix co he n Kierkegaard’s Fear and Trembling edited by dan ie l co n way Kant’s Lectures on Ethics edited by l ar a d e n i s and o l i ve r s e n s e n Aristotle’s Physics edited by m ar i s k a l e u n issen Aristotle’s Politics edited by t ho rnto n lo c k wo o d and th an assis sam ar as Aquinas’s Disputed Questions on Evil edited by m. v. do u g h e rt y Plato’s Symposium edited by pie r re d e s t r é e and z i n a g i a n n o p o u lo u Spinoza’s Ethics edited by yitzhak y. mel amed Kant’s Critique of Pure Reason edited by jame s r. o ’sh e a


edited by JAMES R. O ’SHEA University College Dublin

University Printing House, Cambridge cb2 8bs, United Kingdom One Liberty Plaza, 20th Floor, New York, ny 10006, USA 477 Williamstown Road, Port Melbourne, vic 3207, Australia 4843/24, 2nd Floor, Ansari Road, Daryaganj, Delhi - 110002, India 79 Anson Road, #06-04/06, Singapore 079906 Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781107074811 doi: 10.1017/9781139871389  C Cambridge University Press 2017

This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2017 Printed in the United Kingdom by Clays, St Ives plc A catalogue record for this publication is available from the British Library Library of Congress Cataloging-in-Publication data Names: O’Shea, James R., author. Title: Kant’s Critique of pure reason : a critical guide / James R. O’Shea, University College Dublin. Description: New York City : Cambridge University Press, 2017. | Includes bibliographical references and index. Identifiers: lccn 2017005279 | isbn 9781107074811 (alk. paper) Subjects: LCSH: Kant, Immanuel, 1724–1804. Kritik der reinen Vernunft. | Knowledge, Theory of. | Causation. | Reason. Classification: lcc b2779 .o84 2017 | ddc 121 – dc23 LC record available at https://lccn.loc.gov/2017005279 isbn 978-1-107-07481-1 Hardback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.


List of Contributors List of Translations and Abbreviations

page vii x



James R. O’Shea

1 Kant on the Distinction between Sensibility and Understanding


Eric Watkins

2 Knowledge and Its Object


Stephen Engstrom

3 Transcendental Idealism and the Transcendental Aesthetic: Reading the Critique of Pure Reason Forward


Lucy Allais

4 Kant on the Ideality of Space and the Argument from Spinozism


Michela Massimi

5 How Precise Is Kant’s Table of Judgments?


Michael Wolff

6 Kant’s “Transcendental Deduction”


Barry Stroud

7 Kant’s Critique of the Layer-Cake Conception of Human Mindedness in the B Deduction


James Conant

8 The Critical and “Empty” Representation “I Think” Patricia Kitcher v




9 Kant’s Mathematical Principles of Pure Understanding


Lisa Shabel

10 Kant’s Dynamical Principles: The Analogies of Experience


Kenneth R. Westphal

11 The Refutation of Idealism


Ralf M. Bader

12 The Antinomies: An Entirely Natural Antithetic of Human Reason


Graham Bird

13 The Ideal of Reason


John J. Callanan

14 Knowledge, Discipline, System, Hope: The Fate of Metaphysics in the Doctrine of Method


Andrew Chignell

List of References Index

280 293


lucy allais is jointly appointed as Professor of Philosophy at the University of the Witwatersrand, Johannesburg, and Henry Allison Chair of the History of Philosophy at the University of California, San Diego. She is a leading contributor to current debates on Kant’s transcendental idealism, on the idea of nonconceptual content in relation to Kant, and on the supposed contrast between transcendental and naturalistic philosophy, most recently in her book Manifest Reality: Kant’s Idealism and His Realism (2015). She has also published on forgiveness and other topics in ethics and is interested in moral psychology and free will. ralf m. bader is a Fellow of Merton College and an Associate Professor in the Philosophy Department at the University of Oxford. His research focuses on Kant, metaphysics, ethics, and political philosophy. graham bird is Emeritus Professor of Philosophy at the University of Manchester and a cofounder of both the UK Kant Society and the journal Kantian Review. His book Kant’s Theory of Knowledge (1962) was a groundbreaking work on Kant in the analytic tradition, and recently he has published a comprehensive commentary on Kant titled The Revolutionary Kant: A Commentary on the Critique of Pure Reason (2006). john j. callanan is Senior Lecturer in the Department of Philosophy at King’s College London. He has written articles on Kant on topics such as transcendental arguments, the holy will, and the mathematical method. He has also written on Strawson and Mandeville. He is the author of Kant’s Groundwork of the Metaphysics of Morals: A Reader’s Guide (2013). andrew chignell is Professor of Philosophy at the University of Pennsylvania. His primary research is on Kant and modern philosophy, with additional interests in contemporary epistemology, aesthetics, and philosophy of religion. vii


List of Contributors

james conant is Chester D. Tripp Professor of Humanities, Professor of Philosophy, and Professor in the College at the University of Chicago. He works broadly in philosophy and has published articles in philosophy of language, philosophy of mind, aesthetics, German idealism, and history of analytic philosophy, among other areas, and on a wide range of philosophers, including Kant, Emerson, Nietzsche, Kierkegaard, Josiah Royce, William James, Frege, Carnap, Wittgenstein, Putnam, Cavell, Rorty, and McDowell, among others. stephen engstrom is Professor of Philosophy at the University of Pittsburgh. He has published articles on various topics in Kant’s theoretical and practical philosophy. He is also the author of The Form of Practical Knowledge (2009) and coeditor (with Jennifer Whiting) of Aristotle, Kant, and the Stoics (Cambridge, 1996). patricia kitcher is Roberta and William Campbell Professor of the Humanities and Professor of Philosophy at Columbia University. She is the author of two books on Kant’s conceptions of cognition and the self, Kant’s Transcendental Psychology (1990) and Kant’s Thinker (2011). michela massimi is Professor of Philosophy of Science at the University of Edinburgh. She has published widely on philosophy of science and on Kant’s philosophy of natural science in particular. She was the principal investigator on the Leverhulme Trust international network Kant and the Laws of Nature (2012–15), and she is currently the principal investigator on an ERC grant on Perspectival Realism (2016–20), inspired in part by Kant. She is the editor of Kant and Philosophy of Science Today (Cambridge, 2008) and coeditor (with Angela Breitenbach) of Kant and the Laws of Nature (Cambridge, 2017). james r. o’shea is Professor of Philosophy at University College Dublin. He is the author of Wilfrid Sellars: Naturalism with a Normative Turn (2007) and Kant’s “Critique of Pure Reason”: An Introduction and Interpretation (2012), and the editor of Sellars and his Legacy (2016). lisa shabel is Associate Professor of Philosophy at the Ohio State University. Her book Mathematics in Kant’s Critical Philosophy (2003) and subsequent articles explore the role of mathematics in Kant’s Critique of Pure Reason. barry stroud is the Willis S. and Marion Slusser Professor of Philosophy at the University of California, Berkeley. He works mainly in epistemology, metaphysics, philosophy of mind and language, and the history of

List of Contributors


modern philosophy, and he has been a leading contributor to debates concerning Kant’s transcendental deduction. He is the author of Hume (1977), The Significance of Philosophical Scepticism (1984), The Quest for Reality: Subjectivism and the Metaphysics of Colour (1999), Engagement and Metaphysical Dissatisfaction: Modality and Value (2011), and three volumes of collected essays. eric watkins is Professor of Philosophy at the University of California, San Diego. Author of Kant and the Metaphysics of Causality (Cambridge, 2005) and well-known editor and translator of Kant’s works (most recently of the Cambridge University Press volume of Kant’s writings on science, Immanuel Kant: Natural Science, 2013), Watkins specializes in Kant’s metaphysics and epistemology across both the Critical and pre-Critical periods. kenneth r. westphal is Professor of Philosophy at Bo˘gaziçi Üniversitesi, Istanbul. His publications include several books on Kant, Hegel, and epistemology. His newest is How Hume and Kant Reconstruct Natural Law (2016); of special significance to his present contribution is his book Kant’s Transcendental Proof of Realism (Cambridge, 2004). michael wolff is Professor of Philosophy at Universität Bielefeld, Germany. Among his many other publications, his book on Kant’s metaphysical deduction and Table of Judgments, Die Vollständigkeit der kantischen Urteilstafel (1995), transformed debates concerning that central topic in Kant’s Critical philosophy.

Translations and Abbreviations

Unless authors indicate otherwise in their chapters, all translations are from the Cambridge Edition of the Writings of Immanuel Kant (CE), series editors Paul Guyer and Allen W. Wood (Cambridge, 1992–). Page citations are to Kant’s gesammelte Schriften, Ausgabe der Königlich-Preußischen Akademie der Wissenshaften (Berlin: Walter de Gruyter, 1902–), using the volume:page number format (e.g., 10:129) found in the margins of the Cambridge Edition volumes, and in many cases, the pages of the Cambridge Editions are provided, too. References to the Critique of Pure Reason, however, are by means of the standard “A/B” references to the first (1781) and second (1787) editions, respectively. The following are the abbreviations of Kant’s works used throughout this volume. Ak. Anthr

CE CJ Corr CPR CPrR Discov

Akademie Ausgabe Anthropology from a Pragmatic Point of View (Anthropologie in pragmatischer Hinsicht [1798], Ak. 7), in Anthropology, History, and Education, CE 2007, G. Zöller, R. B. Louden, eds., R. B. Louden, trans. Cambridge Edition of the Writings of Immanuel Kant. Critique of the Power of Judgment (Kritik der Urteilskraft [1790], Ak. 5), CE 2000, P. Guyer ed., trans., E. Matthews, trans. Correspondence, CE 1999 (Ak. 10–13), A. Zweig, trans., ed. Critique of Pure Reason (Kritik der reinen Vernunft [1781, 1787], Ak. 3–4), CE 1997, P. Guyer, A. Wood, trans., eds., cited by A (first edition)/B (2nd ed., 1787). Critique of Practical Reason (Kritik der praktischen Vernunft [1788], Ak. 5), in Practical Philosophy, CE 1996, M. J. Gregor, trans., ed. On a Discovery, according to which all Modern Critique of Pure Reason is alleged to be made Superfluous by an Earlier x

Translations and Abbreviations


FS G ID Inquiry



Critique (Über eine Entdeckung, nach der alle neue Kritik der reinen Vernunft durch eine ältere entbehrlich gemacht werden soll [1790], Ak. 8), in TP2 . Concerning the Ultimate Ground of the Differentiation of Directions in Space (Von dem ersten Grunde des Unterschiedes der Gegenden im Raume [1768], Ak. 2), in TP1 . Erdmann, Benno: Kant’s notes in his own copy of the first edition of CPR, referring to the original pages in Erdmann 1881, reprinted in Ak. 23, and as footnotes in the Cambridge Edition of CPR. The False Subtlety of the Four Syllogistic Figures (Die falsche Spitzfindigkeit der vier syllogistischen Figuren [1762], Ak. 2), in TP1 . Groundwork of the Metaphysics of Morals (Grundlegung zur Metaphysik der Sitten [1785], Ak. 4), in Practical Philosophy, CE 1996, M. J. Gregor, trans., ed. Inaugural Dissertation (Concerning the Form and Principles of the Sensible and Intelligible World; De mundi sensibilis atque intelligibilis forma et principiis, [1770], Ak. 2), in TP1 . Inquiry Concerning the Distinctness of the Principles of Natural Theology and Morality (Untersuchung über die Deutlichkeit der Grundsätze der natürlichen Theologie und der Moral [1764], Ak. 2), in TP1 . Jäsche Logic, in LL. Lectures on Logic, CE 1992, J. M. Young, trans., ed. Lectures on Metaphysics, CE 1997, K. Ameriks, S. Naragon, eds., trans. Lectures on the Philosophical Doctrine of Religion, in RRT Metaphysical Foundations of Natural Science (Metaphysische Anfangsgründe der Naturwissenschaft [1786], Ak. 4), in TP2 . A New Elucidation of the First Principles of Metaphysical Cognition (Principiorum primorum cognitionis metaphysicae nova dilucidatio [1755], Ak. 1), in TP1 . Notes and Fragments, CE 2005 (Ak. 14–23), P. Guyer, ed., C. Bowman, P. Guyer, F. Rauscher, trans. Attempt to Introduce the Concept of Negative Magnitudes into Philosophy (Versuch den Begriff der negativen Grössen in die Weltweisheit einzuführen [1763], Ak. 2), in TP1 .

xii OPA PM


Refl (R) RRT TP1 TP2

Translations and Abbreviations The Only Possible Argument in Support of a Demonstration of the Existence of God (Der einzig mögliche Beweisgrund zu einer Demonstration des Daseins Gottes [1763], Ak. 2), in TP1 . Physical Monadology (The Employment in Natural Philosophy of Metaphysics combined with Geometry, of which Sample I contains the Physical Monadology; Metaphysicae cum geometria iunctae usus in philosophia naturali, cuius specimen I. continet monadologiam physicam [1756], Ak. 1), in TP1 . Prolegomena to Any Future Metaphysics that will be Able to Come Forward as Science (Prolegomena zu einer jeden künftigen Metaphysik, die als Wissenschaft wird auftreten können [1783], Ak. 4), in TP2 . Reflexionen, from Kant’s handschriftliche Nachlaß (“handwritten remains”), by Reflexionen number (R), Ak. volume:page reference, and page in NF where applicable. Religion and Rational Theology, CE 1996, A. W. Wood and G. Di Giovanni trans., ed. Theoretical Philosophy, 1755–1770, CE 1992, D. Walford, trans., ed., with Ralf Meerbote. Theoretical Philosophy after 1781, CE 2002, H. Allison, P. Heath, eds., H. Allison, G. Hatfield, P. Heath, M. Friedman, trans.

Introduction James R. O’Shea

The publication of Kant’s Critique of Pure Reason in 1781 reconfigured the intellectual landscape in ways that still, to this day, shape our most fundamental debates not only about knowledge, perception, freedom, and God, but about the very nature of philosophy and the possibility of any future ‘rational metaphysics’ itself. From that date onward Kant’s book was widely known not only for its ‘all-crushing’ criticisms of the traditional alleged proofs of the existence of God and the immortality of the soul, but also for attempting to reorient entirely our understanding of how our theoretical concepts and our morally practical will are related to the reality that we attempt to know and to transform for the better. Each successive generation of readers of Kant’s Critique has been struck in equal parts with the novel transformative power of its complex ideas and arguments and with the unusually difficult task of interpreting and understanding those ideas. The aim of this Critical Guide to Kant’s Critique of Pure Reason is to present cutting-edge research by leading scholars representing a variety of interpretive perspectives, and to do so in such a way that the volume also serves to explain the most fundamental themes and arguments that are to be found in one of the most fertile and revolutionary texts in the history of philosophy. The relative newcomer to Kant’s first Critique – that is, the first of his three famous Critiques: of pure reason, of practical reason (1788), and of judgment (1790), respectively – should find that the chapters of this book cover nearly all of the main topics of Kant’s enormous edifice of argument in the Critique. The aim of each chapter in this Critical Guide, however, is not so much to introduce all of the main concepts and arguments that are involved in the given topic of which each chapter treats, but rather to illuminate and probe some particularly important or currently much discussed aspects of that topic. Serving the twin tasks of introduction and exploration has been one of the most successful features of Cambridge’s Critical Guides series, and it is hoped that this volume continues that helpful practice. 1


j ames r. o’shea

Those readers who are, on the other hand, more deeply acquainted with the sharp interpretive controversies that have continually accompanied Kant’s works over the last two centuries should find that the chapters to follow represent an appropriate selection of interpreters who take very different stands on some of those most hotly debated topics in Kant scholarship today: for example, in relation to ongoing disputes concerning the nature of Kant’s transcendental idealist distinction between “appearances” and “things in themselves,” or concerning the conceptual and nonconceptual dimensions of our cognition, and so on. But in line with the emerging tradition of Critical Guide volumes, I will keep this introduction short and close by letting the chapters speak for themselves, in a brief summary of their contents. The volume begins with certain key features of Kant’s thinking during his “pre-Critical” period (i.e., prior to the publication of the first Critique in 1781), but in ways that lead intelligibly to some of the most fundamental distinctions and themes of his Critical philosophy. Eric Watkins in Chapter 1, “Kant on the Distinction between Sensibility and Understanding,” seeks to clarify the foundations of Kant’s transcendental idealism by offering a novel account of the key distinction between sensibility and understanding as it develops throughout Kant’s pre-Critical career. The idea is that Kant thinks that because existence is not a real predicate, it cannot be cognized fully solely by the understanding, but also requires a distinct faculty of sensibility through which objects are “given.” This idea finds historical support in the fact that Kant focuses on the distinction between existence and real predicates early on and then throughout his pre-Critical career, and it finds philosophical support in the fact that it is plausible to think that existence is different from other kinds of properties and thus requires a different kind of analysis. Watkins’s account enables a compelling explanation of Kant’s relation to Leibniz on this issue, not only because it has Kant reacting to Leibnizian ideas, but also because it is reasonable to view Leibniz’s position as vulnerable in its explanation of the status of existence. Stephen Engstrom then follows with an analysis of similarly fundamental themes in Chapter 2, “Knowledge and Its Object.” His focus is on how Kant, in undertaking to explain how a priori knowledge of objects is possible, proceeds according to the idea that “the objects must conform to our knowledge.” Engstrom suggests that Kant’s reliance on this “Copernican” way of thinking is puzzling, since on the one hand most readers find it paradoxical, yet on the other the nature of Kant’s project in the Critique of Pure Reason precludes all reliance on questionable assumptions. Engstrom’s



chapter addresses this puzzle by articulating the basic self-understanding involved in theoretical knowledge. It argues that in this self-understanding, such knowledge is understood to have two essential features, which entail that it bears a relation to its object that implicates the Copernican way of thinking. The chapter also describes the main factors that contribute to the customary misconception that “our knowledge must conform to the objects.” With Lucy Allais’s Chapter 3 on “Transcendental Idealism and the Transcendental Aesthetic: Reading the Critique of Pure Reason Forward,” we move into the complex and dense short section in which Kant discusses the ontology of space and time, conditions of mathematical cognition, and the role of space in representing objective particulars. Here Kant introduces his novel and puzzling idea of a priori intuition and presents and argues for his complex form of idealism – transcendental idealism. This chapter presents an account of Kant’s transcendental idealism as it is presented in the Aesthetic, as well as an account of his argument for this position, starting with a focus on his notion of intuition and the role intuition plays in cognition. Allais argues that this role is that of giving us acquaintance with the objects of cognition, and that this explains how a priori intuition can provide Kant’s general answer to his question of how synthetic a priori cognition is possible. She argues that this general question, including in its application to geometry, does not concern how we justify synthetic a priori claims but how it is possible for them to concern given objects and therefore qualify as cognition. Kant argues that this requires a priori intuition and he takes this to lead to idealism. The historical and conceptual roots of Kant’s transcendental idealist conception of space are then pursued in Chapter 4, “Kant on the Ideality of Space and the Argument from Spinozism,” by Michela Massimi. She explains that Kant’s engagement with Newton’s account of absolute space was complex and problematic. The received view has it that after endorsing relationism about space in his Physical Monadology in 1756, Kant came to defend Newton’s absolute space in the 1768 text, Directions of Space. But Kant’s flirting with Newton’s absolute space was short-lived, soon to be ended with the Inaugural Dissertation in 1770, where the ideality of space was first introduced, and fully defended in the Critique of Pure Reason. In this chapter Massimi focuses on one particular aspect of Kant’s departure from Newton’s absolute space: what she calls the argument from Spinozism. First she clarifies the argument’s premises and structure, highlighting what she takes to be the Newtonian spirit of its premises. Second, she argues that Kant’s reasons for associating absolute space with Spinozism are to be


j ames r. o’shea

found not in the debates (including charges of Spinozism) that surrounded Newton’s own view and Newtonianism about space. Instead, they have to be looked for in an influential metaphysical tradition that – from Leibniz, to Baumgarten – addressed what she calls the problem of the world as a totality of substances in interaction. Thus Massimi ultimately argues that we should read and understand Kant’s defense of idealism about space in the argument from Spinozism against the latter intellectual backdrop. Moving now from the forms of space and time, as the sensible conditions on our knowledge, to the forms of judgment and the categories as its intellectual conditions, in Chapter 5, “How Precise Is Kant’s Table of Judgments?,” Michael Wolff explains how Kant’s Table of Judgments provides the ground plan for his Critical philosophy and for the systematic form of all its parts. This is why Kant emphasizes that this Table is complete and precise: complete insofar as it can be proved that there is no logical function and no logical form which does not contribute to what Kant calls the ‘quantity,’ ‘quality,’ ‘relation,’ or ‘modality’ of a judgment, and precise insofar as it can be shown that in each of the four cases there are no less than three ‘moments’ of the logical form of a judgment. Michael Wolff has long been a leading commentator on these aspects of Kant’s thought, so Ken Westphal’s translation of this new essay by Wolff into English will certainly be welcomed by many English-speaking readers of Kant’s first Critique. From the logical forms of judgment and Kant’s ‘metaphysical deduction’ of the categories, as it is called, we then come to Chapter 6 and Barry Stroud’s analysis of the leading ideas behind “Kant’s ‘Transcendental Deduction.’” Overall Stroud provides a schematic description of the goal of Kant’s “Transcendental Deduction,” the general structure of the argument, the conditions of its success, and its implications for the defense of “transcendental idealism.” Both those new to Kant’s argument and seasoned readers of the secondary literature on the Deduction will find much of value in Stroud’s careful delineation of its aims and implications, including the difficult questions that Stroud takes the argument to leave us with. James Conant argues in Chapter 7 that the argument of Kant’s Transcendental Deduction has generally been misinterpreted by the main lines of twentieth century English-speaking commentary on Kant’s Critique of Pure Reason. In “Kant’s Critique of the Layer-Cake Conception of Human Mindedness in the B Deduction,” Conant suggests that according to many commentators, the point of the Transcendental Deduction is to show that the categories of the understanding represent conditions on the thinkability for us of that which is heterogeneously given to us in a self-standing form of sensible consciousness. He argues that Kant should rather be read as



taking aim at the central assumption that underlies such a reading – namely, the assumption that our nature as sensibly receptive beings, in so far as it makes a contribution to cognition, represents a self-standingly intelligible aspect of our nature. According to Conant a proper understanding of the B Deduction requires appreciating how it involves a rewriting of the A-Deduction with an eye to highlighting why the standard way of summarizing the teaching of the Critique of Pure Reason involves a fundamental misunderstanding of its teaching. Patricia Kitcher then takes up a different fundamental theme in the Deduction, one that is also central to Kant’s “Paralogisms of Pure Reason” later in the Transcendental Dialectic section of the Critique. In Chapter 8 on “The Critical and ‘Empty’ Representation ‘I Think,’” Kitcher explains how in the Transcendental Deduction Kant describes the principle of the transcendental unity of apperception as the “highest principle” of cognition (e.g., A117n, B136), the principle from which much (other) a priori cognition can be gleaned. According to this principle, any mental state or representation must belong to a common cognitive subject, an ‘I think’ in Kant’s terminology. Yet in the Paralogisms chapter, Kant characterizes the representation ‘I think’ as ‘empty’ (A345–46/B403–4). The tension between these central doctrines (the ‘I think’ principle is the most important principle of cognition; the representation ‘I think’ has no content) has led scholars to reject Kant’s claims for the importance of the principle of apperception. In this chapter, Kitcher attempts to establish that Kant has a solid argument for the transcendental unity of apperception and that, when we understand how that argument works, we can also understand his puzzling claim about the emptiness of the representation ‘I think.’ In Chapter 9 Lisa Shabel then investigates “Kant’s Mathematical Principles of Pure Understanding,” a chapter which clarifies key concepts and principles that are intimately related to other key sections of the Critique as well, such as the Transcendental Aesthetic, the Schematism, and the Discipline of Pure Reason. As Shabel explains, the mathematical principles of pure understanding (the Axioms of Intuition and the Anticipations of Perception) are those judgments that are meant to establish the application of the categories of quantity and quality to the objects of possible experience. The principles of pure mathematics are the axioms or fundamental truths of the mathematical sciences. Both sets of principles comprise judgments that are, according to Kant, intuitively certain, synthetic, and a priori knowable, but the former (mathematical principles of pure understanding) ground the “possibility and objective a priori validity” of the latter (principles of pure mathematics). In this chapter Shabel explores the relation


j ames r. o’shea

that the mathematical principles of pure understanding bear to the principles of mathematics proper, while also exploring Kant’s very notion of a “principle,” whether of pure understanding, mathematics, or sensibility. From the mathematical principles we then move to Chapter 10 and Kenneth R. Westphal’s analysis of “Kant’s Dynamical Principles: The Analogies of Experience.” Westphal explains that Kant’s justification of a transeunt account of causal interaction – contra Hume – is not in the Second Analogy of Experience alone, but in all three Analogies conjointly. Officially the Critique of Pure Reason aims to justify our use of the general causal principle: every event has a cause. The relevant causal principle is more specific: every spatiotemporal event has a distinct spatiotemporal cause. The Critically justified use of this specific principle is still more specific, according to Westphal, because this regulative principle of causal inquiry obtains constitutive significance only by making true and justified causal judgments about particular causal relations among spatiotemporal phenomena. On Westphal’s analysis of Kant’s arguments, identifying actual causal relations requires conjoint use of all three principles of causal judgment because causal judgments are discriminatory: we can identify any one causal relation only by distinguishing it from causally possible alternative scenarios. According to Westphal, Kant’s analysis of legitimate causal judgments bears upon such issues as ‘relevant alternatives’ in epistemology, justificatory fallibilism, the role of imagination in cognitive judgment and the semantics of singular cognitive reference (predication as a cognitive achievement, not merely as a grammatical or logical form). Kant’s analysis of causal judgment and its a priori transcendental conditions hold independently of transcendental idealism, Westphal argues, because Kant’s ‘Analytic of Principles’ (to which the ‘Analogies’ belong) is a transcendental ‘Doctrine of the Power of Judgment’ (B171ff.). In Chapter 11 Ralf M. Bader then analyses Kant’s “Refutation of Idealism” in the B-Edition of the Critique of Pure Reason by examining the conditions that must be satisfied for inner states to be objectively determined in time, focusing in particular on the question to what extent their temporal ordering is parasitic on an objective ordering of outer states. Such a dependence of the ordering of inner states on that of outer states would show, contrary to the problematic idealist, that one’s existence (understood in terms of one’s mental states) cannot be objectively determined in time unless there is an external world. Bader carefully sorts through complex questions concerning the starting points and assumptions of Kant’s argument, as well as its implications for varieties of both skepticism and idealism.



Continuing to follow the broad structure of Kant’s first Critique itself, we then move from the Transcendental Analytic to the Dialectic, beginning in Chapter 12 with Graham Bird on “The Antinomies: An Entirely Natural Antithetic of Human Reason” (Kitcher’s chapter having addressed key themes from the Paralogisms). Kant refers, in his own terminology, to the traditional conflicts outlined in the Antinomies as an “entirely natural antithetic of human reason” (B433). Bird highlights how this terminology reflects Kant’s central aim in the Antinomies to resolve issues where reason inevitably “comes into conflict with itself” (Axii–xiii) and “precipitates itself into darkness and contradictions” (Aviii). That project, Bird explains, is an essential part of Kant’s wish to reform philosophy by laying bare the underlying errors which have encouraged the futile pursuit of these apparently insoluble conflicts. Bird argues that the upshot of Kant’s account, however, is not negatively to reject reason but only to restrict it by recognizing more positively its legitimate function. That conclusion is captured in his claim that reason has only a “regulative,” but not a “constitutive” role. Kant’s position has been much criticized, but recently among commentators such as Allison (1983), A Defense of Transcendental Idealism, and Grier (2001), Kant’s Doctrine of Transcendental Illusion, these criticisms have been modified. Bird’s chapter looks in detail at Kant’s account of the first and third Antinomies which exemplify respectively what Kant distinguishes as “mathematical” and “dynamical” antinomies. The analysis of the Antinomies leads smoothly into John J. Callanan’s Chapter 13 investigation of Kant’s conception of “The Ideal of Reason.” As Callanan explains, Kant’s critical analyses regarding the origin and basis of religious belief have often been interpreted negatively and as undermining ordinary attitudes, despite Kant’s intentions to the contrary. This suspicious reception, Callanan claims, stems not just from Kant’s famous attack on possible proofs for the existence of God but also on his account of the positive grounds for the origin of our concept of God. Kant’s account engages with a tradition that had raised the possibility of a radical difference between human and divine rationality and that also questioned the motives for perceiving any commonality between them. Callanan argues that Kant’s account can be seen to be premised on an acceptance of many such claims, yet nevertheless demands that such attitudes remain rationally warranted. Focusing on the notion of an archetype, Callanan contends that Kant’s Critical account demands the peculiar position that it is one of the “interests” of human rationality that it projects its characteristics onto the idea of a divine being, yet only for the purpose of subsequently viewing human reason as a copy of that original divine reason.


j ames r. o’shea

Finally we turn to the concluding sections of the first Critique, and in particular to Andrew Chignell’s Chapter 14 analysis of Kant’s views on “Knowledge, Discipline, System, Hope: The Fate of Metaphysics in the Doctrine of Method.” In this chapter Chignell highlights the apparent tensions between Kant’s very stringent critique of metaphysical speculation in the “Discipline of Pure Reason” chapter and his endorsement of Belief (Glaube) and hope (Hoffnung) regarding metaphysical theses in the subsequent “Canon of Pure Reason.” In the process Chignell examines Kant’s distinctions between the theoretical and the practical bases for holding a “theoretical” conclusion (i.e., a conclusion about “what exists” rather than “what ought to be”) and argues that the position is subtle but coherent. In the second part of the essay Chignell then focuses on Kant’s account of rational hope in the Doctrine of Method: its nature, scope, conditions, and role in the philosophy of religion generally. For the more detailed explanations and arguments, of course, we now turn to the chapters themselves. I would like to end this introduction, however, by thanking all of the above contributors to this volume for their exceptionally kind patience and persistent hard work in producing what I hope has turned out to be a helpful and thought-provoking critical guide to Kant’s extraordinary book. Special thanks also to Ken Westphal for his translation of Michael Wolff’s essay, to Fabio Gironi for his work on the Index, and to Hilary Gaskin of Cambridge University Press for her perceptive guidance from start to finish.

c h a p ter 1

Kant on the Distinction between Sensibility and Understanding Eric Watkins∗

Fundamental to Kant’s mature theoretical philosophy as it is expressed in the Critique of Pure Reason is his distinction between appearances, or things as they appear to us, and things in themselves, or things as they are in themselves. The distinction is necessarily presupposed, for example, by transcendental idealism, the view that we can have theoretical cognition only of appearances, which essentially depend on space and time as merely subjective forms of sensible intuition, and not of things in themselves, which cannot be given to our senses and thus must lie forever beyond the limits of our cognition, even if they can be objects of thought. But Kant’s distinction between appearances and things in themselves depends, in turn, on his distinction between sensibility and the understanding. For appearances, he maintains, can be given to us only through sensibility, and things in themselves can be thought by us only through the understanding.1 It is thus crucial that we understand the exact nature of the distinction Kant wants to draw between these two faculties and what argument he can offer in favor of drawing the distinction in the way that he does. His entire theoretical philosophy depends on it.2 However, it is possible to come to a full appreciation of the distinction Kant draws between sensibility and the understanding in the first Critique only if we take a broader historical view, one that includes several of his pre-Critical works as well as his relation to his predecessors. For on this ∗



Thanks to Lucy Allais, Tobias Rosefeldt, Clinton Tolley, and Marcus Willaschek for two rounds of comments on earlier versions of this essay. This should lay to rest any doubts about the possibility of supererogatory acts. The distinction between sensibility and understanding, which are faculties, differs from that between intuition and concept, which are representations. However, the fact that human intuition is, for Kant, sensible and not intellectual can, at times, make it difficult to discern whether it is the sensible or the intuitive character of a sensible intuition that is supposed to bear philosophical weight. The standard view, represented by Hans Vaihinger, Michael Wolff, and Lorne Falkenstein, is that Kant has no argument for the distinction and that it is thus a fundamental, un-argued-for assumption of the first Critique. See also Gloy (1990).



er ic wat k in s

latter point, Kant explicitly criticizes both his rationalist and empiricist predecessors for mistakenly conceiving of the distinction as involving a difference of degree rather than a difference in kind. Specifically, Kant charges, “Leibniz intellectualized the appearances, just as Locke sensitivized the concepts of the understanding” (A271/B327).3 That is, Locke “sensitivized” our understanding’s representations by rejecting innate ideas and trying to trace all of our ideas back to experience (whether that takes the form of sensation or reflection), while Leibniz “intellectualized” appearances by holding that our sensory representations are not in fact generated by external objects affecting the senses, but rather are simply confused modes of intellectual representations. As a result, Kant thinks, “instead of seeking two entirely different sources of representations in the understanding and sensibility . . . , each of these great men holds on only to one of the two, which relates immediately, in his opinion, to things in themselves, while the other does nothing but confuse or order the representations of the first” (A271/B327). Thus, rather than first positing a single faculty that privileges one kind of representation and then explaining the other kind away as a deficient exemplar of the former by appealing to some difference in degree between the two (e.g., in their degrees of clarity and distinctness), Kant asserts the need for “two entirely different” faculties of sensibility and the understanding that can then be used to account for the differences between these different kinds of representations. But this gives rise to two central questions. What exactly is the nature of each of these faculties? And why should we think that they are distinct in kind, as Kant supposes? Though we shall be able to appreciate the complexity of Kant’s answer to the first question only after we have considered some central pre-Critical texts (especially the Inaugural Dissertation (ID)), it is useful to have a brief preview of his main answer to the second question. Kant takes sensibility and understanding to differ in kind, I suggest, both because they are responsible for representations that have different and in fact irreducible representational characters, which allow sensible and discursive clarity, respectively, and because they fulfill different functions within cognition insofar as, unlike the understanding, sensibility allows for objects to be given to us in intuition in such a way that we are immediately aware of the existence of objects and can provide evidence that our understanding’s judgments actually refer to what exists. 3

All translations are my own, though I have frequently consulted the relevant volume of The Cambridge Edition of the Works of Immanuel Kant.

Kant on the Distinction between Sensibility and Understanding


In this essay, I first sketch the account of existence that Kant develops in his pre-Critical period, highlighting the notion of absolute positing that underlies it, and its difference from relative positing, which involves real predicates and the understanding. Second, I describe how Kant characterizes our different cognitive faculties in the Inaugural Dissertation and show how the first Critique picks up on this position in drawing the distinction between sensibility and the understanding in the way that he does there. Third, on the basis of these considerations, I present two arguments that support Kant’s view of the distinction between sensibility and the understanding as a difference in kind.


Kant’s Pre-Critical Views on Existence

The immediate context for situating Kant’s views on existence in his preCritical period lies in Crusius’s distinction between essence and existence.4 Early on in his most important philosophical work, Sketch of the Necessary Truths of Reason in 1745, Crusius holds that the first distinction to be made upon encountering a sensible object is between its essence and its existence, where, in agreement with Leibniz and Wolff, he maintains that the essence of a thing, which can be found in thought in the understanding, allows one to distinguish between that thing and any other. Unlike Leibniz and Wolff, however, he emphatically maintains that existence is a predicate that cannot be found in the thought of the essence of the thing.5 In another departure from Leibniz and Wolff, he then connects space and time to existence such that, necessarily, everything that exists does so in space and time. Crusius then goes on to advance several substantive and, in their historical context, controversial theses about existence. For example, he argues, against Leibniz and Wolff, that the existence of a thing has priority over its possibility, or essence, and that sensations are a special way of representing existence. Specifically, he defines sensation in such a way that it “is precisely that state of our understanding in which we are forced to think something immediately as existing, without first needing to cognize it through inferences.”6 Sensations are thus noninferential, or immediate representations of objects that exist, though, it should be noted, he attributes such representations to the understanding rather than to a distinct faculty of sensibility. 4 5 6

For fuller discussion of Kant’s pre-Critical account of existence and his break with Leibniz and Wolff, see Watkins (forthcoming). For Leibniz, the concept of existence is contained at least in the case of the concept of God. Christian August Crusius, Entwurf der notwendigen Vernunft-Wahrheiten, Leipzig, 1745, §16. This passage and others are translated in Watkins (2009).


er ic wat k in s

In the New Elucidation (NE) in 1755, Kant’s most basic goal is to chart a middle course between the Leibnizian-Wolffian and Crusian positions. He sides with Wolff against Crusius in accepting the principle of determining ground for all truths and all contingently existing things, despite his concession to Crusius that the principle should invoke “determining” rather than “sufficient” grounds (NE 1:393). At the same time, his restriction of the principle of determining ground to contingently existing things effects a fundamental break with Wolff (NE 1:396). In particular, he argues, against Descartes, Wolff, and Baumgarten, that whatever serves as the ground of all contingently existing things cannot also serve as the ground of itself, since nothing can contain the ground of itself (NE 1:394). For a ground must be prior to what it grounds, and nothing can be prior to itself. As a result, Kant asserts that whatever grounds the possibility of all contingent things must itself be ungrounded and exist prior to its own possibility. Such a necessary being is then identified with God (NE 1:395). Eight years later, in The Only Possible Argument (OPA), Kant provides a more detailed analysis of existence as part of the setup for his novel theistic proof, which turns on the basic idea already expressed in the New Elucidation that God must exist to serve as the ground of all possibility. He begins by noting, negatively, that existence is not a predicate or a determination of a thing. To put the point in Crusian terms, existence is not part of the essence of a thing, for you can “draw up a list of all the predicates which might be thought to belong to Julius Caesar . . . You will quickly see that he can either exist with all these determinations or not exist at all” (OPA 2:72). One of the arguments that Kant employs to establish this negative claim proceeds on the basis of broadly Leibnizian considerations in philosophical theology and metaphysics. When, according to Leibniz, God represents an individual as possible, God’s representation of that individual must be complete, since God, in virtue of his omniscience, cognizes every individual essence in its entirety as distinct from every other individual essence. However, if God’s representation of that possible individual is complete, then, if God decides to make it actual, it cannot be the case that the predicate of existence has been added to the representation of that individual, because the object was already represented as complete. Therefore, existence cannot be a predicate like those that God thinks of as constituting the essence of any possible thing. By thinking of existence (not just of God but of any object) as distinct from essence, Kant seems to be drawing on one aspect of Crusius’s position. Kant rightly notes, however, that one can still use ‘existence’ as a predicate, such as in the proposition ‘God exists.’ Yet in that case, it is, Kant

Kant on the Distinction between Sensibility and Understanding


suggests, not a predicate “of the thing itself” but rather “of the thought which one has of the thing” (OPA 2:72). This idea leads him to assert that “existence is the absolute positing of a thing” (OPA 2:73). Though Kant notes that the notion of positing “is so simple that it is not possible to say anything further by way of elaboration” (ibid.), he does draw an instructive contrast between absolute and relative positing. Relative positing occurs when one attributes a predicate concept to a subject concept in a judgment. It is, Kant seems to think, a purely logical relation between the predicate concept and the subject concept whose truth conditions are internal to the judgment and that does not require that the subject exist.7 Thus even an atheist can agree with the claim that God is omnipotent, since, on this understanding of the claim, it asserts only that omnipotence is posited with respect to God but not that such an omnipotent being exists. Absolute positing, by contrast, “posits” that an actual object corresponds to the subject concept. It is thus extralogical and involves what we would call the existential use of “is.” That is, unlike relative positing, which is (in the simplest case) a relation between two concepts in a judgment, absolute positing requires a different kind of relation, one between the judgment (its subject concept in particular) and something external to the judgment, namely, the thing to which the judgment refers; its truth conditions thus lie outside of the concepts that are involved in the judgment. Thus, “God exists” is not a case of relative positing (with “God” and “existence” being predicates posited relative to each other) but rather one that involves absolute positing, because the concept “God” is posited absolutely; its truth conditions require something in addition to any identity or containment relation that might obtain between concepts, for it requires that God actually exist in such a way that the concept of God refers to it. Kant’s use of the notion of absolute positing helps to explain how existence differs from other predicates that the understanding employs within its judgments. Though Kant is still heavily indebted to the rationalist tradition early in his pre-Critical period, he is also thinking through several fundamental features of the metaphysics of his predecessors in an independent way. Existence is no exception insofar as he draws on aspects of Crusius even while forging his own position on how existence is to be understood. Later, we shall see the relevance of these views to the mature Kant’s arguments for the distinction between sensibility and the understanding. 7

It is not clear that Kant consistently maintains this view throughout his pre-Critical period.



er ic wat k in s

The Distinction between Sensibility and the Understanding

Kant’s Inaugural Dissertation, published in 1770, bears the title On the Form and Principles of the Sensible and the Intelligible World. The title makes it clear that his basic intent is to distinguish between the sensible and intelligible worlds so as to be able to state the form and fundamental principles of each. In line with this, the Dissertation begins, in Section 1, with a discussion of the concept of the world in general, before distinguishing, in Section 2, between the sensible and the intelligible. Sections 3 and 4 then state and argue for the principles and the form of the sensible and intelligible worlds, respectively. Kant begins Section 2, in §3, with definitions of faculties that are relevant to representing what is sensible and what is intelligible. He thus offers definitions of what he calls “sensualitas” and “intelligentia”: Sensualitas is the receptivity of a subject in virtue of which it is possible for the subject’s own representative state to be affected in a definite way by the presence of some object. Intelligentia (rationality) is the faculty of a subject in virtue of which it has the power to represent things that cannot by their own quality come before the senses of that subject. (ID 2:392)

Despite some apparently trivial terminological issues, these definitions seem to express features that will become hallmarks of Kant’s mature conception of sensibility and understanding. For both appear to be characterized as faculties by means of which subjects are capable of forming representations, and what is distinctive about sensualitas is that it is receptive insofar as an existing object is able to affect the subject so as to influence its representative state. The faculty he calls intelligentia is defined as the power to represent things that cannot be represented by the senses. He then immediately states further definitions, e.g., of the sensible and the intelligible, that seem to follow from these initial definitions trivially. In the ensuing paragraphs in Section 2 (§§4–12), Kant then fills out his account. In §4 he explicitly connects “things that are thought sensitively” to “representations of things as they appear” and “things that are [thought] intellectual[ly]” to “representations of things as they are” (ID 2:392). The first move that is clearly distinctive of Kant’s position occurs when he distinguishes “in a representation of sense” between its matter, which he identifies with sensation, and its form, by means of which what affects the senses is “coordinated by a certain natural law of the mind” (ID 2:393). For neither the rationalists nor the empiricists who preceded him drew this kind of form-matter distinction for representations of sense. Furthermore,

Kant on the Distinction between Sensibility and Understanding


just as the matter of a sensory representation is evidence of the existence of the object that affects the subject (while the quality of that representation depends on the subject), so too the form “is undoubtedly evidence of a certain reference or relation in what is sensed . . . For objects do not strike the senses in virtue of their form” (ibid.). In this way, Kant infers that insofar as sensory representations are unified “into some representational whole,” the matter is due to the object affecting the subject, while the form is brought about entirely by the subject “in accordance with stable and innate laws” (ibid.). In line with the form-matter distinction, Kant then notes, at the beginning of §5, that cognition is called sensory (sensuales) because of its matter (i.e., sensations) and sensitive (sensitivae) “in virtue of the form, even if it were to be found free from all sensation” (ID 2:393). This passage makes it clear that Kant is not treating ‘sensory’ (sensualis) and ‘sensitive’ (sensitivus) as equivalent. Indeed, careful attention to the Latin terms reveals that Kant does not use these terms either haphazardly or as synonyms for ‘sensible’ (‘sensibile’), but rather with quite different meanings. Specifically, ‘sensory’ is whatever involves sensations and might be translated as ‘empirical’ if Kant did not also occasionally use the term ‘empiricus.’ By contrast, and as the quote above suggested, Kant uses ‘sensitivus’ and its various cognates (such as ‘sensitive’) to refer to a feature had by any representation that involves the form associated with the law of the mind that, among other things, coordinates the sensory elements in our representations. Kant makes this point quite explicit by saying that cognitions would be called sensitive even if they did not involve sensation at all (and therefore are not sensory). Thus, a representation could be sensitive even if it does not include a sensation in it. And a representation could also be both sensory and sensitive (as in an empirical intuition whose sensitive form takes up sensations and coordinates them).8 Moreover, since Kant is investigating the form of the sensible world, it clear that he is more interested in the sensitive than in the sensory, a fact that is also indicated by the frequency of his use of ‘sensitivus’ and its cognates (84) as compared to that of ‘sensualis’ and its close relatives (12). Therefore, though it might have initially seemed as if Kant were starting out with a soon-to-be standard contrast between sensibility and understanding, we can now see that this would fundamentally misrepresent his position. If, as we have just seen, the sensory and the sensitive are different, 8

A representation could presumably also be sensory without being sensitive, as would be the case with a sensation that is not taken up into an empirical intuition.


er ic wat k in s

then, when Kant defines sensualitas in terms of receptivity and passivity in §3, his definition does not for that reason apply to what is sensitive. For if what is sensitive need not involve a sensation, then sensitivity cannot be characterized in terms of receptivity or passivity per se. Instead, Kant maintains that sensitivity involves a stable and innate law of the mind that is to some extent active (e.g., in coordinating sensory representations). Furthermore, and more importantly, we can see that when Kant provides a definition of sensualitas, he is not thereby offering a definition of sensibility (sensibilitas). For, as we have seen, sensualitasis responsible for sensations. As a result, it is plausible to understand sensibile (sensible) as the genus that covers both sensualis and sensitivus and sensibility as the faculty that is responsible for both sensory and sensitive representations, though it is true that Kant never explicitly defines sensibility in the Dissertation.9 Instead, as noted above, his primary focus is on one aspect of sensibility, namely, the (sensitive) form of the sensible world. For similar (though less fully developed) reasons, intellegentia, which Kant immediately associates with rationality, is not obviously to be equated with intellectus, which Kant never explicitly defines in the Dissertation and which is best translated as ‘understanding.’ Kant’s view of the understanding must therefore be determined by the uses to which he puts it. When Kant turns to the understanding in §5, he immediately distinguishes its logical and real uses. The real use of the understanding concerns how its concepts are given, while its logical use compares different concepts (according to the principle of contradiction) so as to subordinate lower concepts to higher ones (e.g., in a taxonomic hierarchy). He then shows how the logical use applies to cognition and is required in the sciences. For in comparing cognitions, regardless of how they are given, a cognition either contains a given mark under it or is opposed to it (and thus does not contain it at all), and if it does contain that mark, then it must do so either immediately (in which case the cognition is distinct and can be expressed in a judgment) or mediately (in which case the containment relation can be expressed in a syllogism). The logical use of the understanding is thus valuable for establishing relations not only between concepts but also among cognitions so as to establish laws in science. 9

One might take the absence of a definition of sensibilitas as grounds for denying that Kant is committed to sensibility as a faculty in the Dissertation. However, in light of the prominent use of the term ‘sensibile’ and the need for a faculty to represent sensible objects (e.g., the sensible world), it is difficult to resist viewing Kant’s considered position as being committed to a faculty of sensibility and the absence of an explicit definition as an omission deriving from the haste with which the Dissertation was composed.

Kant on the Distinction between Sensibility and Understanding


Though this description leaves much obscure about the positive features of the understanding, Kant goes on to emphasize that the various uses of the understanding cannot change sensory or sensitive representations in such a way that they are no longer sensory or sensitive: It is of the greatest importance here to have noticed that cognitions must always be treated as sensitive cognitions, no matter how extensive the logical use of the understanding may have been in relation to them. For they are called sensitive on account of their genesis and not on account of their comparison in respect of identity or opposition. Hence, even the most general empirical laws are nonetheless sensory; and the principles of sensitive form which are found in geometry (determinate relations in space), no matter how much the understanding may operate on them by reasoning according to the rules of logic from what is sensitively given (by pure intuition), nonetheless do not cease to belong to the class of what is sensitive. (ID 2:393–94)

Two points in this passage are especially noteworthy. First, Kant clarifies that regardless of the extent of the logical use of the understanding, empirical laws are still sensory, and sensitive cognitions will always be sensitive. For, as he notes, a representation is sensitive because of its genesis and not because of the logical relations (of identity and diversity) that can be discerned through the understanding’s acts of comparison.10 That is, what makes a representation sensory is that it either is or includes a sensation, while what makes a representation sensitive is that it is formed in accordance with an innate law of the mind that coordinates sensory representations (if they are present). Nothing that the understanding does can change that. Second, the passage gives us a more concrete sense of what we should understand by the sensitive, as contrasted with the sensual. Those principles of sensitive form that concern space are contained in geometry, and what is sensitively given derives from pure intuition, which he defines later, in §12, as “an intuition devoid of sensation but not for that reason deriving from the understanding” (ID 2:397). Thus, in geometry, the mind forms a pure intuition that contains no sensations that have been caused by an external object, and is thus not sensory, but that is nonetheless still sensitive, since its status as sensitive is unaffected by the logical use of the understanding that would occur in, say, geometrical proofs. So an example of what Kant has in mind here as sensitive is the pure (nonempirical) representation of space that is used in geometry and that is based in an 10

To say that a representation is sensitive because of its genesis means, on this rendering, that it is sensitive because of the laws according to which sensibility operates when it forms a sensitive representation.


er ic wat k in s

innate (and nonintellectual) capacity of the mind to form a certain kind of representation (pure intuition). In light of all of this, the crucial question that emerges is the following: what, in all of this, is sensibility? More specifically, what is it that sensory and sensitive representations have in common such that they are both sensible (and derive from the faculty of sensibility)? Unfortunately, Kant does not address this question explicitly. However, my hypothesis is that they both share a certain kind of representational character that is distinct from that had by intellectual representations such as concepts.11 That is, both sensations and our pure intuitions of space and time have what one might call a sensible character that sensibility is responsible for by objects being given through it in a particular way, even if the one involves causal affection from without, while the other does not.12 This interpretive hypothesis can be illustrated and, to a certain extent, confirmed on the basis of a passage from §7. There, Kant explains, From this one can see that the sensitive is poorly defined as what is more confusedly cognized, and what belongs to the understanding as that of which there is a distinct cognition. For these are only logical distinctions that do not touch at all the things given, which underlie every logical comparison. Thus, sensitive representations can be very distinct, and representations which belong to the understanding can be extremely confused. We notice the first case in that paradigm of sensitive cognition, geometry, and the second case in the organon of everything that belongs to the understanding, metaphysics. (ID 2:394–95)

Kant’s main critical point here is that sensitivity is not to be defined in terms of confused representations, as some of his rationalist predecessors had maintained. For sensitive representations in, say, geometry can be very distinct, while the understanding’s representations can be extremely confused, as is made apparent by the “clouds of confusion” that are, on his view, prevalent in metaphysics. For example, the brightly illuminated pen right in front of my eyes is as clear as could be, though whether the concept of crimson or scarlet applies to it is not (due to my lack of clarity about the exact differences between these two concepts).13 11 12


Recall that the sensible character of sensible representations is distinct from their intuitive character, for not all intuitions are sensible, as is clear from the (possible) divine case. This view of sensibility is consistent with the faculty of desires being sensible. For what is essential to the faculty of desire is not so much that it involves causal affection from without but rather that it involves a certain kind of felt representation, i.e., a representation with a certain kind of phenomenal character. Kant clarifies “clarity” and “distinctness” in, e.g., the Jäsche Logik 9:62–63, LL 567–68.

Kant on the Distinction between Sensibility and Understanding


But this critical point has a positive implication, namely, that what distinguishes sensitive representations is that they can have a kind of clarity that is distinct from discursive clarity, since, as we have seen, the former can occur in the absence of the latter.14 Furthermore, since sensations will similarly lack discursive clarity, it is tempting to think that sensible representations in general (which would include both sensory and sensitive representations) have a special kind of representational character in virtue of which they are sensible, one that can be made apparent through the kind of clarity they admit.15 It is distinct from the intuitive character of an intuition through which it relates immediately to a singular object because intuitions can, in principle, be either sensible or intellectual (even if human beings cannot have intellectual intuition). So the idea is that in a sensible intuition, the intuitive character of the representation is responsible for the object being given, while its sensible character is responsible for its being given in the particular way that it is such that it can, under the right conditions, display sensible clarity. This suggestion coincides with the kinds of claims that Kant wants to make about both sensory and sensitive representations. Empirical objects can be given by way of sensations in empirical intuition, which accounts for how sensory representations can be sensible in this sense, while mathematical objects can be given in pure intuition, which accounts for how sensitive representations could be sensible in the same sense.16 On this account, then, sensibility is the faculty through which objects are given to us with the kind of representational character that admits of a distinctive kind of (sensible) clarity about objects, even in the absence of the kind of discursive clarity that can attach to concepts and render objects understandable. This account of sensibility can accommodate both (1) the sensory, insofar as sensations resulting passively via external affection allow empirical objects to be given to us in the requisite way in empirical intuition, and (2) the sensitive, insofar as pure intuition allows nonempirical objects (which could be constructed in pure intuition) to be given to us in 14



This idea finds some confirmation in Kant’s distinction in the Blomberg Logik between logical clarity and sensible clarity, where logical clarity involves subordination according to logical laws, whereas sensible clarity involves coordination according to “aesthetic laws” (24:129–30, LL 101). This reading gains additional support from Kant’s claim, in §3 of the Dissertation, that incongruent counterparts are objects that we can distinguish, though not on the basis of features that could be grasped purely intellectually, which indicates the need for a different kind of representation, one that would be sensible rather than intellectual. Eckart Förster (2012: 35) emphasizes this line of argument. Kant provides hints that he is already committed to the construction of mathematical objects at ID 2:389 and 2:402.


er ic wat k in s

a similar way. It thus fits nicely with the quite nuanced distinctions that Kant implicitly utilizes in the Dissertation between the sensory, the sensitive, and the sensible (where all three are distinct from the understanding and its real and logical uses of concepts). If this account captures Kant’s understanding of sensibility in the Dissertation, what is his account of the understanding? Unfortunately, Kant was in a hurry to publicly defend his Dissertation so that he could take up the professorship of logic and metaphysics that he had long coveted. As a result, his treatment of the understanding is rather underdeveloped. However, the basic view he expresses (in §8) is that the understanding forms concepts, which are distinct from sensible representations and are “abstracted from the laws inherent in the mind (by attending to its actions on the occasion of an experience)” (ID 2:395). He goes on to give an incomplete list of examples of such concepts – substance, cause, possibility, and necessity – which overlap with some of the first Critique’s categories, but he does not yet see a need for a deduction of them (either metaphysical or transcendental). Instead, he describes (in §9) the negative and positive ends of these concepts as elenctic and dogmatic, where the latter has theoretical and practical paradigms of noumenal (divine) and moral perfection. He also provides a brief account (in §§10–12) of the kinds of cognition that we can have (including symbolic cognition of things in themselves), but the distance between the Dissertation and the first Critique on the vast majority of these points is still immense. It is primarily in Section 3, where Kant focuses on the principles of the form of the sensible world – space and time – that he anticipates some of the first Critique’s most significant results. With this account of Kant’s understanding of sensibility and the understanding in the Inaugural Dissertation firmly in view, we can now turn, briefly, to his treatment of these faculties in the Critique of Pure Reason. Of course, Kant’s project in the first Critique is quite different from that of the Inaugural Dissertation. He is no longer focused on the form of the sensible and intelligible worlds. Instead, he is interested in the fate of metaphysics itself – cognition of God, freedom, and the immortality of the soul – whose scandalous past he hopes to leave behind by undertaking a full-scale analysis of the faculty of reason itself so as to be able to determine the possibility of synthetic a priori cognition. What emerges at the end of this analysis is a complex picture according to which theoretical cognition is possible, but only of appearances and not of things in themselves, such as the objects of traditional metaphysics, though metaphysics’s fate can be saved, he suggests, by moral considerations that are then discussed in detail later, in the Critique of Practical Reason.

Kant on the Distinction between Sensibility and Understanding


Given that Kant is undertaking an analysis of the faculty of reason in general in the first Critique, it is no surprise that sensibility and the understanding play a prominent structural role in its overall argument. Specifically, the Transcendental Aesthetic (the first of the two main sections of the Transcendental Doctrine of Elements) analyzes the faculty of sensibility, while the Transcendental Logic (the second main section of the Doctrine of Elements) dissects the understanding and reason. Accordingly, in the beginning of each of these sections we find definitions of sensibility and understanding. In the very first paragraph of the Aesthetic, sensibility, through which objects are said to be given to us in intuitions, is defined as “the capacity (receptivity) to acquire representations through the way in which we are affected by objects” (A19/B33), while the understanding is introduced as the faculty that thinks objects through concepts. In the Introduction to the Transcendental Logic, Kant similarly defines sensibility as “the receptivity of our mind to receive representations insofar as it is affected in some way” while the understanding is described as “the faculty for bringing forth representations itself, or the spontaneity of cognition” (A51/B75; cf. also A15/B29). Two points in these passages are relevant for current purposes. First, Kant offers a more detailed (and positive) characterization of the understanding as a spontaneous faculty of concepts that depend on a certain set of unifying functions (A68/B93). The new features that are ascribed to the understanding are due to the immense intellectual labor that was required to analyze the understanding and reason so as to develop the arguments of the Metaphysical and Transcendental Deductions and the Transcendental Dialectic. Second, these “definitions” are clearly not precise technical statements that would specify exactly what is required for a faculty to be either sensibility or the understanding. Instead, they read more like rough approximations, especially since on a precise reading of this definition of sensibility, our pure intuitions of space and time would not qualify as sensible. In fact, the contrast with the Dissertation is striking. Whereas the Dissertation is precise and distinguishes consistently between the sensory and the sensitive, the first Critique talks only about sensibility and what is sensible, even though Kant has not changed his position fundamentally insofar as he still distinguishes between the matter and form of a sensible intuition and continues to think of the former as sensations (in empirical intuitions) and the latter in terms of laws in accordance with which pure intuitions are formed.17 Part 17

Kant can still capture the difference between the sensory and the sensitive, but he would now do so in terms of a contrast between empirical and pure (sensible) intuition.


er ic wat k in s

of the reason for Kant’s terminology in the first Critique may be the relative paucity of the German language in this particular case (which has as an option only “sinnlich” as compared to Latin’s “sensualis,” “sensitivus,” and “sensibilis”). However, he may also be assuming that his readers are already familiar with his basic account of sensibility and the understanding (from the Dissertation) and now are simply curious about whether these faculties can in fact establish the kind of synthetic a priori cognition associated with claims of traditional metaphysics. Toward the end of the Transcendental Aesthetic, in §8, Kant situates his account of sensibility and understanding with respect to that of the Leibnizian-Wolffians. After reiterating that space and time are, on his view, the pure forms of sensibility and sensations its matter, he remarks: That our entire sensibility is nothing but the confused representation of things . . . is therefore a falsification of the concept of sensibility and of appearance. . . . The difference between an indistinct and a distinct representation is merely logical, and does not concern the content. . . . The Leibnizian-Wolffian philosophy has therefore directed all investigations of the nature and origin of our cognitions to an entirely unjust point of view in considering the distinction between sensibility and the intellectual as merely logical, since it is obviously transcendental, and does not concern merely the form of distinctness or indistinctness, but its origin and content, so that through sensibility we do not cognize the constitution of things in themselves merely indistinctly, but rather not at all. (A43–44/B60–62)

Again, we have Kant objecting to the rationalists’ view that sensibility is a confused mode of representation that thereby assimilates it to the understanding and to logical clarity and distinctness. Instead, Kant wants to insist, the distinction between sensibility and understanding is “obviously” transcendental and concerns “its origin and content,” just as was suggested above. Overall, therefore, there is no reason to think that Kant’s conception of sensibility has changed fundamentally. It is still the faculty by means of which objects are given in such a way that they have a particular (sensible) character, regardless of whether the subject has been affected by an external object (in the empirical case) or not (in the mathematical case). And the understanding, despite now being clearly demarcated from reason and having undergone much development, still contrasts with sensibility insofar as it produces a different kind of (intellectual) representation on the basis of a different kind of law (spontaneous).

Kant on the Distinction between Sensibility and Understanding


1.3 Kant’s Arguments for a Distinction in Kind With this account of Kant’s distinction between sensibility and understanding in hand, we are now in a position to consider why Kant maintains that the distinction is one in kind, and not degree. A number of arguments could seem attractive here. For example, given the prominent role synthetic a priori cognition plays in the first Critique, it is tempting to think that the analytic-synthetic distinction would form the basis for an argument for the distinction in kind between sensibility and understanding. For an analytic truth is one that is determined solely by the understanding, whereas synthetic truths must appeal to sensibility as a distinct faculty. However, in addition to doubts one might have about the adequacy of such an argument, we have seen that Kant already has the distinction between sensibility and understanding in place in the Dissertation, where he had not yet drawn the analytic-synthetic distinction.18 This makes it unlikely that Kant’s primary argument for the distinction being one in kind rests on the analytic-synthetic distinction. The same kinds of considerations make an argument based on the Antinomy of Pure Reason ill-suited to the task as well.19 Without making any claim to exhaustiveness, I suggest that Kant is moved by two kinds of considerations.20 One is based on the different characters that sensible and intellectual representations have, along with the impossibility of a gradual transition of the one into the other, while the other depends on the special kind of function that sensible representations play in cognition. Let me expand on each of these in turn. As was suggested above, Kant maintains that sensible representations have a distinctive representational character, one that is shared by sensations (in the guise of empirical intuitions) and by pure intuitions, that admits of a certain kind of sensible clarity for given objects, and that is 18



My own view is that Kant needs to have the distinction between sensibility and understanding in place before he can explain the distinction between analytic and synthetic truths (rather than the converse). This is not to say that Kant does not develop new or further arguments for the distinction in the first Critique. In fact, one of the arguments that I propose receives further development in the first Critique. But the basic idea, I contend, is already present by the time of the Dissertation. Nor is it to deny the significance of the “great light” Kant saw in 1769. For a helpful description of several argumentative possibilities, see Heidemann (2002: esp. 69). As should be clear from the arguments presented above, I am more optimistic about the ability of the “Rezeptionsthese” to deliver the kind of arguments that are needed than is Heidemann, in part, because I distinguish between arguments for the distinction between intuition and concept and those for the distinction between sensibility and understanding.


er ic wat k in s

distinct from the kind of intellectual representation by means of which we think objects and can attain a certain kind of discursive clarity about them. Furthermore, these representations are not merely distinct, but distinct in such a way that no amount of activity by the understanding will render a sensible representation nonsensible. For, as we saw above, Kant held it to be “of the greatest importance” to recognize that a cognition is sensible “no matter how much the understanding is operative” (ID 2:293–94). That is, there is no continuous transition from a sensible to an intellectual representation, no way of getting from the one to the other by some kind of gradual increase in degree. Now given the difference in the kind of representational content that sensible and intellectual representations have, and given that there is no way to transition from the former to the latter (e.g., by simply getting clearer and clearer about its content), it is most natural to think that these features are best explained in terms of a distinction in kind in the faculties from which they originate.21 Importantly, the difference at issue here does not pertain to the different degrees of activity and passivity involved in sensibility and the understanding. Though Kant often characterizes sensibility in terms of its passivity, this is not, as we saw above, one of its essential features since it does not pertain to pure sensible intuitions at issue in mathematics.22 Indeed, the Dissertation’s more careful distinction between the sensory and the sensitive made this clear. And even if sensibility were essentially characterized in causal terms (of activity and passivity), sensibility is still active insofar as it does not literally receive representations from without, but rather actively forms representations in response to affection from without. As a result, if sensibility were essentially characterized solely in causal terms, then both sensibility and understanding would be active and would differ only by means of degree, contrary to Kant’s position. Instead, what these reflections on the nature of sensibility suggest is that sensibility and the understanding must be fundamentally different in virtue of the kind of representational content that they are responsible for and due to the fact that they operate according to fundamentally different laws.23 The kind of clarity that sensible representations can have is different from the kind of clarity that intellectual representations can have and these differences are due to the different principles (or laws) that govern (the activity 21

22 23

It makes sense to speak not just of representations, but also of faculties as distinct in kind rather than degree if they operate according to fundamentally different principles (rather than, say, different degrees of activity). Kant would run the risk of begging the question against Leibniz if he were to define sensibility in this way (since Leibniz denies that sensibility requires causal interaction). For a similar view, see Tolley (2013).

Kant on the Distinction between Sensibility and Understanding


of ) each faculty. As a result, it is plausible to view the differences between these principles (or laws) and thus between the faculties themselves as a difference in kind. The second argument for taking the distinction between sensibility and understanding to be a difference in kind concerns the distinctive role that sensibility plays in representing the existence of objects in cognition.24 As we saw above, sensibility is the faculty through which objects are given, while the understanding is the faculty through which objects are thought. Insofar as any object that is given must also exist, sensibility is capable of representing the existence of objects.25 Now matters are complicated by the fact that in both the Dissertation and the first Critique, existence can also be represented by the understanding (through its real use or by way of the category of existence). But because the understanding is responsible only for thinking, its representation of the existence of an object does not guarantee that that object is given, for it can represent an object as existing even when it does not. That is, if we merely think of an object as existing, it can fail to exist and our thought then fails to refer and we fail to have cognition of it, since the givenness condition that Kant places on cognition is not satisfied in that case. If, by contrast, an object is given in sensibility, then its representation of the existence of the object does guarantee (or at least provide evidence) that its object exists.26 This is clearest in the case of empirical objects, since our empirical intuitions contain sensations, which are caused by objects that must exist in order to bring about these sensations. Indeed, this point was made quite clearly in the Dissertation’s characterization of the matter of an intuition in terms of a sensation that is “evidence for the presence of something sensible” (ID 2:393).27 Thus what is distinctive about sensibility is not simply that it represents the existence of objects, which the understanding can do as well, but rather the specific way it does so since its representations, unlike the understanding’s, can guarantee (or at least be evidence of ) the existence of the objects that are immediately given to it (by way of sensations in cases of empirical intuition). 24


26 27

Textual evidence to support aspects of this line of argument (e.g., the distinction between possibility and actual existence and an understanding of the latter in terms of absolute positing) can be found in §76 of the third Critique (CJ). Kant may not have been entirely clear about how to understand the existence of mathematical objects. He is aware that they do not exist in the same way that empirical objects do, but he does not specify what kind of existence they might have. This claim is true for intuition of sense and not of the imagination. Also, I remain neutral here on whether cognition is to be understood primarily in semantic or epistemological terms. Here I am treating ‘presence’ and ‘existence’ as synonymous (despite subtle, but important differences).


er ic wat k in s

This distinctive role that sensibility plays in cognition indicates that there must be a difference in kind between sensibility and understanding. For sensibility plays a role in cognition that the understanding is simply incapable of taking on. The understanding can represent objects (including their existence), but it cannot do so in the way that is required for its representations to guarantee reference to (or provide evidence for the existence of ) an object. Only through sensibility can objects be given to us in such a way that the reference of our representations can be established. It is thus plausible to think that a distinction in kind between these faculties is the best way of accounting for the satisfaction of these very different conditions on cognition. But note how this second argument in particular implicitly draws on several distinctive features of Kant’s pre-Critical position. For as we saw, early on, with Crusius’s help, Kant noted that the rationalists’ account of existence is inadequate. Existence cannot be adequately represented by a real predicate like those that constitute the essences of things and thus cannot be represented by the understanding. Instead, existence requires absolute positing, something extrajudgmental that can ensure that an object corresponding to the subject concept in a judgment actually exists. Concepts alone cannot do the job.28 Furthermore, in virtue of its causal origin, sensation provides immediate (noninferential and thus nondiscursive) cognitive access to (and thus evidence for) the existence of objects, a kind of access that the understanding cannot provide. In these ways, the second argument that we have considered for Kant’s distinction between sensibility and understanding picks up on Kant’s pre-Critical account of existence and the unique role that sensations can play in representing existence, a role that the understanding simply cannot play in virtue of the cognitive limits to its conceptual thought. This second argument is thus a natural development of the line of thought that Kant had initiated in his pre-Critical period in response to Leibniz, Wolff, and Crusius.

1.4 Conclusion It is tempting to think that Kant simply accepts the distinction between sensibility and understanding that is central to his theoretical philosophy as a fundamental assumption, one that could be discharged, if at all, only through the explanatory power of the account that he builds on it. At the 28

In light of his commitment to symbolic cognition of (the existence of ) things in themselves, it is unclear what exactly Kant’s position is on this point in the Dissertation.

Kant on the Distinction between Sensibility and Understanding


same time, given how distinctive his understanding of the distinction is and given that both his rationalist and empiricist predecessors would have found it highly controversial, it would clearly be preferable if we were able to read him as having an argument that would not blatantly beg the question against his predecessors. Fortunately, we have been able to identify two arguments that have a firm basis in the text and some philosophical force against his predecessors. For as we have seen, over the course of his career, Kant came to appreciate that existence is not like the standard properties of objects and for that reason requires a fundamentally different kind of treatment.29 By noting this point and Leibniz’s apparent inability to account for different kinds of representational content and clarity, Kant can be viewed as picking up on genuine tensions in Leibniz’s position rather than either begging the question or talking past him.30 The interpretation developed above thus satisfies crucial requirements on an adequate interpretation. Of course, much more would need to be done before we could claim to be in possession of a full-blown account of sensibility and the understanding in Kant’s theoretical philosophy. For example, we would need to explain both why and how these faculties can and must work together to make cognition possible. However, one should not underestimate the importance of a proper first foundation, which, in this case, includes an account of the nature of the distinction between two of our most basic cognitive faculties. 29


One need not think, with Frege and Russell, that existence is a second-order predicate to recognize that existence should be treated differently from how more standard properties of objects are considered. For discussion of whether Kant anticipates Frege on this point, see Rosenkoetter (2010). For nuanced discussion of Leibniz’s position, see Robert Merrihew Adams (1994: esp. Chapter 6).

ch a p ter 2

Knowledge and Its Object Stephen Engstrom∗

To explain how a priori knowledge of objects is possible, Kant develops a distinctive hylomorphic account of human knowledge. According to this account, our knowledge is composed of two specifically different kinds of representation, intuition and concept, which arise, respectively, from the receptive and the spontaneous stems of our cognitive capacity, sensibility and understanding. Since these components, being related as matter and form, are both requisite for cognition, our knowledge is limited to objects that can be sensibly intuited in experience, yet we can know something about those objects a priori. Although much can still be learned from this account, the impediments to a proper appreciation of it are considerable. Indeed, the very aim it serves – to comprehend how a priori knowledge is possible – is liable to arouse doubt and puzzlement, particularly in the climate of naturalism that has conditioned much recent discussion in epistemology.1 Nothing in the account, however, is likely to seem more paradoxical than the conception of cognition’s relation to its object that Kant advances to secure conceptual space for comprehending the possibility of a priori knowledge – the conception he expresses in his well-known “Copernican” proposition that “the objects must conform [sich . . . richten] to our knowledge” (Bxvi). According to contemporary conventional wisdom, this conception is exactly backward, contrary to plain common sense. As P. F. Strawson (1966: 22) remarked, Kant’s reversal, and his associated doctrine of transcendental idealism, “are undoubtedly the chief obstacles to a sympathetic ∗


This essay was presented, in earlier versions, at the 2011 meeting of the New England Colloquium in Early Modern Philosophy, held at Dartmouth College, at the 2013 Biennial Meeting of the North American Kant Society, held at Cornell University, and at a conference on Kant and Sellars, held in 2014 at the University of Chicago. I thank the participants for helpful comments and suggestions. There is an irony here, in that naturalism depends on the concept of nature, yet for Kant this concept would not be possible were there no possibility of a priori knowledge: “through no experience can we become acquainted with nature in general” (Prol 4:318). (Translations of quotations from Kant are my own.)


Knowledge and Its Object


understanding of the Critique.” The sense of paradox is reinforced by Kant’s depiction of himself as effecting a revolutionary break with tradition. Linking the long-prevalent assumption that “all our knowledge must conform to the objects” to the lengthy record of failure in metaphysics, Kant suggests that our prospects for success in this science will be improved if we reverse our way of thinking and suppose that the conformity runs in the opposite direction. Whatever we make of it in the end, this proposal is bound to strike almost everyone, at least upon first encounter, as a strange idea, contrary to our ordinary way of thinking.


Two Views of Cognition’s Relation to Its Object 2.1.1

The Apparent Paradox

An initial perspective on Kant’s reversal can be gained by noting that the traditional assumption figures as a component in a long-standing Scholastic conception of knowledge and truth, which informed the tradition of German school-philosophy to which Kant belonged. According to this view, our knowledge takes two forms, theoretical (or speculative) and practical. In each case, a relation of truth obtains between knowledge and object, but the cases differ as to which of the related terms is the standard for the other. The difference can be expressed in terms of a relation of “being true to”: theoretical knowledge is true to its object, while in the practical case the object is true to the knowledge.2 Thus, the science of astronomy is true to the stars, whereas an artisan’s handiwork is true to the art that guides its production. Of course, the relation Kant reverses is conformity, not truth. When he characterizes truth, he speaks, not of conformity, but of agreement (Übereinstimmung). Truth, he says, according to its nominal definition, is “the agreement of knowledge with its object” (A58/B82).3 But it is reasonable 2


Cf. Aquinas, Summa theologica Ia, q. 16, a. 1. See also Ia, q. 21, a. 2: “Truth consists in the equation of mind and thing. . . . Now the mind that is the cause of the thing is related to it as its rule and measure: whereas the converse is the case with the mind that receives its knowledge from things. When therefore things are the measure and rule of the mind, truth consists in the equation of the mind to the thing, as happens in ourselves. For according as a thing is, or is not, our thoughts or our words about it are true or false. But when the mind is the rule or measure of things, truth consists in the equation of the thing to the mind; just as the work of an artist is said to be true, when it is in accordance with his art.” (Translations from Aquinas follow Aquinas [1948, 1952].) Kant says the nominal definition is “granted,” and even if in saying this he is speaking only of parties to an old controversy about truth – as Gerold Prauss (1969) has argued – his own acceptance of it can be seen from his employment of it throughout the Critique (e.g., A191/B236, A237/B296, A642/B670). Although this definition will call to mind the familiar idea of truth as correspondence,


s t e ph e n en g s t rom

to surmise that he regards conformity as the ground of such agreement, a ground that would be lacking in the case of fortuitously correct opinion.4 So the Scholastic distinction could also be expressed in terms of conformity: theoretical knowledge conforms to the object, while in the practical case the object conforms to the knowledge. In broad outline, the Scholastic doctrine is still with us, though couched in different terms. Many contemporary philosophers accept some version of the view that what distinguishes cognition from desire is a difference in “direction of fit.” In desire, it is commonly said, the objects are to fit the representations, whereas in cognition, the representations are to fit the objects. This view enjoys such widespread currency that it may strike us as a platitude, and a long-recognized one at that. Against the background of this familiar contrast, Kant’s Copernican proposal may seem seriously confused, verging on a conflation of cognition with desire. Rather than saddle Kant with a mistake so gross, sympathetic interpreters may suggest that he means something much less radical. They may remind us that he introduces his proposal for a specific purpose, to enable us to understand the possibility of synthetic a priori knowledge, the sort of knowledge that holds special interest for reason and science. Kant’s idealism is not unqualified, but merely transcendental. It is complemented by his empirical realism. It might accordingly be suggested that Kant’s proposal concerns only a priori knowledge. As for the rest, he stands with the empiricists, acknowledging that the great body of truth we gain through experience must conform to the objects. This suggestion, however, though correct in linking the Copernican turn to the problem of synthetic a priori knowledge, is too quick and too easy. For Kant argues that the only way of explaining how such knowledge is


it expresses that idea in a traditional form and should not be understood as referring to an agreement with a fact or a state of affairs; “object” signifies the subject matter (e.g., the stars). In his letter of February 21, 1772, to Marcus Herz, Kant poses his problem as one of accounting for the necessary agreement between intellectual representations and their objects (Corr 134, 10:131); necessary agreement is also mentioned at Bxvii–xviii. Since Kant’s nominal definition speaks of cognition’s agreement with the object rather than of the object’s agreement with the cognition, it might be thought to carry at least an intimation about the direction of conformity, particularly when we consider that in the nominal definitions offered by some philosophers, the direction of conformity is explicitly indicated. Aquinas, for instance, states that “the judgment is said to be true when it equates to [adaequatur] the external reality” (De veritate q. 1 a. 3), and according to Descartes, “the word ‘truth,’ in the strict sense, denotes the conformity [la conformité] of thought with its object” (Letter to Mersenne, October 16, 1639; translation from Descartes [1991]). But Kant would reject as illegitimate a definition that would rule out a priori the possibility of a priori knowledge (cf. CPrR 5:9n, 5:12; B127–28). We may therefore presume that no intimation about the direction of conformity is intended. So understood, the definition is similar to what Kant elsewhere calls a “transcendental definition” (see CJ 5:177n).

Knowledge and Its Object


possible is by tracing it to the form of synthetic knowledge in general, a formal condition on which a posteriori knowledge, including experience, depends. The very concepts employed in experience are, on his account, nothing but a priori concepts in concreto (A567/B595). A posteriori knowledge is accordingly always a specification of a priori knowledge. The knowledge that the sun is of a certain size is a specification of the knowledge that a body in space has extensive magnitude, and the knowledge that the sun is warming this stone on which it is shining is a specification of the knowledge that substances coexisting in nature effect changes in one another. Kant thus holds that the combination of concepts in experience – indeed even the combination of the manifold of empirical intuition in perception – “can never come into us through the senses” but is, rather, an act of the understanding, “the capacity to combine a priori” (B129–30, 134–35; cf. A120n). Experience itself is thus “the product of the understanding out of materials of sensibility” (Prol 4:316, cf. A1). So experience cannot be determined by its object. It is the other way around. Experience “determines an object through perceptions” (B218). So insofar as the objects must conform to a priori knowledge, they must conform to a posteriori knowledge as well.5 2.1.2

The Copernican Proposal and the Project of a Critique of Pure Reason

Those less sympathetic to Kant’s program might be inclined to shrug off the apparent strangeness of his proposal, seeing it as another example of how philosophers may stray from the truth when their reflections are not tethered to common sense and experience, or another instance of the vain extravagance and heterodoxy to which they are prone. Kant does not hesitate to underscore the great differences separating his system from conventional views. By his own reckoning, the Critique of Pure Reason is a revolutionary work, “in conflict with all customary notions” (Prol 4:261).6 5


As for Kant’s empirical realism, it will suffice to note that his distinction between ideality and reality does not line up with the difference between the Copernican and the traditional ways of conceiving of cognition’s relation to its object. As can be seen from his initial explication of the notions of ideality and reality, his thought is not that a priori representations are ideal and empirical representations real, but that the same representation – e.g., space – is ideal with respect to things in themselves, but real, or objectively valid, with respect to appearances (A27–28/B43–44). And though reality is said to amount to objective validity, which entails truth, truth does not entail conformity to the object, but merely agreement with it (see preceding note). Although he does not directly describe his proposed reversal as strange or paradoxical, he acknowledges elsewhere, in speaking of a variant formulation of the same thought (viz. “that nature should


s t e ph e n en g s t rom

But this too would be an over-hasty and facile response, insufficiently attentive to Kant’s avowed purpose in undertaking a critique of pure reason. As Kant makes clear, the aim of his proposed revolution is to release metaphysics from the unhappy plight that has embarrassed it ever since the emergence of philosophy as an academic discipline. In established sciences such as mathematics and physics, canons and procedures of investigation are observed in a spirit of cooperation and common purpose, enabling a steady advancement of learning. Metaphysics, on the other hand, despite its aspiration to establish itself as a science, indeed as the fundamental science, has throughout its history remained in a condition of thoroughgoing conflict, reminiscent of a Hobbesean state of nature, a war of all against all. Yet as its title indicates, the Critique serves a deeper purpose than the preceding description might suggest. Traditional metaphysics conceives of itself as “the science of the first principles of human knowledge” (A843/B871). But since principles are known through reason, and since first principles, being absolutely a priori, are known through pure reason, or reason so far as it is capable of knowledge not derived from any experience (A11/B24; cf. A299–300/B356–57), the perennial conflicts within metaphysics call into question the coherence and integrity of human reason as a cognitive capacity in its own right. So what is needed is a critique, not of “books and systems,” but of human reason itself (Axii). As Kant later reports, he undertook this critical project “to remove the scandal of the seeming contradiction of reason with itself.”7 The ultimate purpose, then, is to remove the apparent conflicts among the judgments of reason, with a view to securing the principles on which human knowledge and science depend. In the service of this aim, the critique is to constitute a tribunal, before which the claims of reason may be adjudicated. According to its very idea, this tribunal of reason depends for its authority in matters theoretical on its independence from all contingent claims and suppositions, just as a court of law depends for its authority in matters practical on its independence from every faction and special interest. It must take as its starting point nothing but reason itself 8 and must be carried out on a plan that, as Kant expressly states, allows no opinions nor anything that even resembles



conform to our subjective ground of apperception”), that it “sounds very contrary to the senses and strange” (A114; cf. B163), and he describes in similar terms the heliocentric hypothesis of Copernicus, to which he likens his own proposal (Bxxiin). Moreover, he calls the parallel reversal that he advances in his practical philosophy a “paradox of method” (CPrR 5:62–64). Letter to Christian Garve, September 21, 1798 (Corr 552, 12:257–58). Kant is of course referring in particular to the antinomy of pure reason, which he expounds and resolves in the Transcendental Dialectic of the Critique. Prol 4:274; cf. letter to Garve, August 7, 1783 (Corr 198, 10:340).

Knowledge and Its Object


a mere hypothesis.9 Reliance on a questionable assumption is thus completely out of the question, strictly prohibited.10 Considered in the light of this aim and plan, the very fact that Kant’s Copernican proposal strikes us as strange should itself strike us as strange. Indeed, Kant makes clear that to achieve its aim, the criticism of traditional metaphysics must be immanent. The critique is self-critique: of pure reason, by pure reason, for pure reason. Its business is “self-knowledge” (Axi). If he is serious in that aim, then he had better not be thinking that the sort of knowledge to which traditional metaphysics aspires – knowledge that, echoing Aristotle and the Scholastics, he calls theoretical, knowledge of “what is” – must, according to our original understanding of it, conform to its object. Unless he is gravely confused about his own project, he must rather be thinking that the view that theoretical knowledge must conform to its object is what he in fact says it is, an assumption, and moreover one that is at odds with our basic understanding of such knowledge, an understanding that he and traditional metaphysicians share. 2.1.3

The Traditional View Reconsidered

To some, this assumption may seem so obvious as to fuel the suspicion that Kant is already off the mark to the extent that he implicitly denies that it figures in our original conception of theoretical knowledge. It might be noted, for instance, that on the Scholastic view I described earlier, the assumption serves as the differentia by which theoretical knowledge is marked off as a species of knowledge, distinguished from its practical counterpart.11 And while it is true that Kant’s modern predecessors standardly reject, as he does, the Scholastic doctrine of sensible species, and true as well that a number of them also depart from the Scholastic teaching that everything in the intellect must first be in the senses, they do not, by and large, break with the Scholastics on whether our knowledge of the objects we encounter in experience must conform to those objects. When we look back over the tradition, however, we find that those metaphysicians who investigate theoretical knowledge most deeply are careful to 9



Axv. Kant does say that he is proposing the alteration in our way of thinking “only as a hypothesis” (Bxxiin), and in introducing his proposal he represents himself as venturing an opinion (Bxv: “Ich sollte meinen . . . ”). But as Kemp Smith (1923: 25) notes, he makes clear that this language reflects the provisional character of the discussion in the Preface and is not meant to apply to the systematic exposition in the body of the Critique itself (Bxxiin). This is not to say that the propositions of the Critique are subject to a requirement that they be readily accessible to the educated public at large. Metaphysical investigations involve modes of attention and abstraction that call for a certain skill, acquired through long practice (cf. B1–2). See Summa theologica Ia IIae, q. 3, a. 5.


s t e ph e n en g s t rom

deny that conformity to the object is essential to it. In particular, both Aristotle and Aquinas make clear not only that such conformity is not essential, but further that in the primary, unlimited case – that of divine the¯oria – the conformity does not run in that direction. Such knowledge, as Aquinas puts it, is not “derived from things known.”12 Here it might be countered that although many traditional metaphysicians do recognize a sort of knowledge that does not conform to its object, they do so by shifting their consideration from human, finite knowledge to knowledge of an altogether different form. The difference is recognized by Kant himself, it being one of his principal doctrines that human knowledge is merely discursive, or conceptual, whereas infinite, divine knowledge is conceived as wholly intuitive (B71, 139). So the fact that some of his predecessors agree that infinite knowledge does not conform to its object is not to the point. The question is whether such conformity is entailed by the basic understanding of human or finite knowledge. This response, however, is not satisfactory. It overlooks the fact that, when Aristotle and Aquinas shift their attention from human to divine knowledge, they do not find it necessary to introduce some other term than ‘theoretical’ or ‘speculative’ to characterize the infinite form. If they thought that conformity to the object were basic to theoretical knowledge, they would deny that divine knowledge is theoretical, just as they deny that divine knowledge is discursive.13 Basic to these philosophers’ understanding of theoretical knowledge is the thought of activity. They recognize that, even though in humans it first arises in the experience of independently existing particulars, such knowledge lies in activity and so has an intrinsic completeness that distinguishes it from all process and coming-to-be. To the extent that it gives rise to scientific knowledge, through which particular things can be known a priori, from principles, theoretical knowledge gains greater completeness and a measure of independence from the particular objects encountered in experience. Thus science – and especially metaphysics, the science of the first principles of science – is seen by these thinkers as approximating the highest form of the¯oria, which is absolutely complete and independent. Since the objects of absolutely complete knowledge can exist only through the 12


In defending the doctrine that God has a speculative knowledge of things from the objection that “speculative knowledge comes by abstraction from things; which does not belong to the divine knowledge,” Aquinas replies, “The fact that knowledge is derived from things known does not essentially belong to speculative knowledge, but only accidentally in so far as it is human” (Summa theologica Ia, q. 14, a. 16). See also Aristotle, Metaphysics XII.9. Summa theologica Ia, q. 14, a. 7; Aristotle, Metaphysics XII.7, 9.

Knowledge and Its Object


knowledge itself, such knowledge cannot be conceived as conforming to what it knows. Human knowledge, in contrast, is limited, depending on the externally existing object, to which it must conform. For Aristotle and Aquinas, then, the¯oria lies essentially in activity, and only accidentally, in the finite case, does it also involve dependence. It appears, therefore, that the traditional assumption is not written into the very idea of theoretical knowledge. Rather, it serves, at least in part, to register such cognition’s dependence on its object in the human, finite case. Kant too calls attention to this dependence, stating that the object “must be given from elsewhere” (Bix–x, cf. B145). All sides agree, then, that our knowledge depends on the object. At issue is what implications, if any, this dependence has for the question of conformity. Space thus opens up for a reconsideration of how finite theoretical knowledge is related to its object. One approach is favored by long tradition; the other is supported by preliminary reflections on the a priori knowledge we understand ourselves to possess.


Recovering Cognition’s Original Understanding of Its Relation to Its Object 2.2.1

Resolving the Problem of the Paradox

A priori knowledge is distinguished by its self-conscious thought of necessity, which marks it out as originating in the cognitive capacity rather than experience (B1–3). The Copernican way of thinking is thereby implicated, as requisite for the comprehension of such cognition’s possibility. But if, as Kant argues (§2.1.1), a priori knowledge is possible only as a formal condition of empirical knowledge, experience included, then its thought of necessity must lie within the self-consciousness of our theoretical knowledge in general. So if a priori knowledge is possible, the Copernican way of thinking must be implicated in theoretical cognition’s basic self-understanding. By the same token, the traditional assumption must rest in a habitual misconception, indeed a self-misunderstanding, first arising at the level of reflection and philosophy. When Kant says, “Hitherto it has been assumed that all our knowledge must conform to the objects,” he is speaking of an assumption on the part of metaphysicians, not an assumption latent in common human understanding (Bxiv–xvi). To understand the Copernican way of thinking, therefore, it will be necessary to see how it is implicated in theoretical cognition’s basic selfunderstanding. And since part of the business of the critique should


s t e ph e n en g s t rom

be to help remedy the misconception responsible for the traditional assumption,14 some account will also be needed of the factors that contribute to that misunderstanding. 2.2.2

Theoretical Knowledge

Our starting point, plainly, must be theoretical cognition’s basic constitutive self-consciousness, noted in Kant’s famous proposition that “The I think must be able to accompany all my representations” (B131). Remaining within the scope of this self-consciousness will ensure that our account articulates what belongs to the basic self-understanding of theoretical knowledge. Two features of such knowledge merit consideration, one pertaining to it generically, as knowledge, the other specifically, insofar as it is finite. First, the self-consciousness integral to all knowledge and even to all thought is fundamentally consciousness of unity.15 Because selfconsciousness is not of anything distinct from itself, but rather identical with that of which it is a consciousness, it can never be a consciousness of a mere plurality. Consciousness of a plurality requires consciousness of distinct items. And the consciousness of each such item, being identical with it, would likewise be distinct from the consciousness of every other. But a mere plurality of items each conscious of itself does not yield consciousness of a plurality. A single consciousness must therefore be present in all the diverse elements figuring in whatever is self-conscious, constituting them as a totality of interdependent components, a whole having the intrinsic completeness of an activity. Such unity, then, belongs to all thought. In the case of knowledge, this unity is of a higher grade, constituting not only interdependence but agreement among the elements. Not only do the components of knowledge depend on one another (and hence never conflict); they also reinforce or confirm one another. Knowledge is “a whole of compared and connected representations” in which every element agrees with every other (A97, 104–5). Knowledge is therefore a self-sustaining, self-productive activity. It has staying power, which bare thought lacks, and this self-consciously self-productive character constitutes its inner validity as knowledge. Second, according to its basic self-understanding, our knowledge, though not itself any process or coming-to-be, nevertheless comes to be, in 14


Just as it is part of that project not only to identify the errors besetting traditional metaphysics but also to explain how they are occasioned by a certain illusion to which human reason is naturally subject. The points discussed in this paragraph receive fuller treatment in Engstrom (2013: §2.2).

Knowledge and Its Object


an advance from ignorance to knowledge. As self-productive activity, however, knowledge depends, insofar as it comes to be, on two conditions: an internal condition, namely, a capacity to know, of which it is the exercise, or actualization; and an external condition, which enables that capacity to be actualized. Yet the external condition cannot lie outside consciousness altogether, for cognition’s dependence upon it must be self-conscious. Finite cognition’s implicit understanding of this dependence thus rests on its distinguishing, within its consciousness, between its own self-conscious activity and modifications or affections of it, between what is self-produced, the exercise of its own capacity, and what is empirical, arising from a source outside itself. But the items thus distinguished must also be understood as standing in a relation suited to the possibility of knowledge. The affections of consciousness must stand in an enabling relation to knowledge and so must be affections of a receptivity that belongs to the cognitive capacity itself. (In Kant’s terms, this distinction is of course between the spontaneous exercise of our cognitive capacity’s determining stem, the understanding, and the receptive operation of its determinable stem, sensibility.) The sort of affection requisite for theoretical knowledge, where the object must be “given from elsewhere,” is sensation. For sensation, as affection by an object,16 enables the distinctive self-conscious relation that theoretical knowledge (whether empirical or a priori) bears to its object, a relation in which the object is understood, in the knowing of it, to be something on whose actuality the actuality of the knowledge depends. Obviously this object too, like the sensations it effects, is understood as standing in an enabling relation to knowledge. Theoretical knowledge thus depends not only on an inward capacity, the capacity to know, but also on a capacity on the outward side, the capacity to be known. Just as the cognizing subject is originally conceived as bearer of the capacity to know, the cognized object is originally conceived as bearer of the capacity to be known. Theoretical knowledge, then, constitutes an actual relation between a cognizing subject and a cognized object, in which these capacities to know and to be known are jointly actualized. 2.2.3

Knowledge and Its Object

So far, two features of theoretical knowledge have been elicited from its implicit self-understanding. We are now in a position to consider afresh the question about the ground of such cognition’s agreement with its object. As 16

In contrast to the feeling of pleasure or displeasure, which, as affection by a representation, is requisite for practical knowledge.


s t e ph e n en g s t rom

I will argue, the two features have implications regarding the constitution of the act of knowledge, implications that bear on how knowledge must be related to its object. It should be clear from the foregoing that theoretical knowledge presupposes a representation of its object as knowable, a single representation underlying all acts of cognition, making possible their agreement in one body of knowledge. The object, moreover, is originally represented merely as knowable, in that our knowledge relies on such a representation in understanding itself as having come to be. Only infinite knowledge could recognize its object to be always already known. But since knowledge is a self-productive unity, it cannot come to be through mere accretion of representations. The advancement from ignorance to knowledge must constitute, not merely an addition of further representations to a representation its cognizing subject already has, but an enlargement of the latter. It must lie, that is, in the growth of the very representation of its object that it presupposes: not a growth of the merely external sort that might be ascribed even to a heap of sand, but a spontaneous growth enabled by the presence of empirical consciousness available (once made general) for that representation to take up into itself.17 Hence there must be identity across the representation the knowledge presupposes and the representation it constitutes, even though the latter contains more than the former. It follows that the presupposed representation must be a general representation, or concept, and that the knowledge must lie in an act in which that concept somehow specifies or determines itself, in what Kant calls a determining judgment. In the basic case, this act is a synthesis of a predicate with the concept of the object. Yet the synthesis is no mere aggregation of concepts, but the self-determination of a concept through an incorporation of another into it. In short, then, finite knowledge is discursive, or knowledge through concepts, in that it presupposes, as its starting point, an original and hence universal concept of its object, the knowable in general, and enlarges it through self-determination, even though, as knowledge that comes to be, it depends for its actuality and growth on a condition outside its own activity. The implications just traced have a clear bearing on how theoretical knowledge and its object are related. Since discursive knowledge is selfdetermining, it cannot be determined by anything outside itself, including its object. Its agreement with its object must therefore rest on its being so related to it that the object is determined by the knowledge rather than 17

Cf. A833/B861.

Knowledge and Its Object


the knowledge by the object, so that the inner validity of knowledge also constitutes its objective validity, its validity in respect of the object. Moreover, determining judgment’s consciousness of its self-sustaining character constitutes a certain consciousness of necessity, grounding an implicit consciousness of the judgment’s validity as universal both in respect of the objects that can be known through the concept and in respect of the judging subjects that can know through that concept. This universality is independent of a judgment’s logical quantity. It figures implicitly even in singular judgments of experience. In cognizing that this stone is growing warm, for instance, I implicitly suppose both that anything constituted as this stone is and in such conditions as it is in would likewise grow warm, and that any subject constituted as I am and in such conditions as I am in would likewise cognize that this stone is growing warm. Judgments of experience are thus already by their form intimations of conclusions derivable from principles. The necessity and universality of determining judgments’ validity is accordingly the mark of rational knowledge, or knowledge from principles, and it gives discursive knowledge the character of law, or what we might call legislative form. On the one side, in respect of subjects, it constitutes such cognition’s legitimacy. On the other, in respect of objects, it constitutes its legislation, whereby its determination of those objects amounts to a kind of rulership or governance over them. This idea finds expression in Kant’s striking portrayal of the understanding as “the legislation for nature” (A126; cf. B163–65). As this characterization makes explicit, the legislative relation knowledge bears to its object holds fundamentally at the level of capacity, between the understanding, the capacity to know, and nature, the knowability of things. The foregoing considerations, if sound, reveal that the Copernican way of thinking is implicated in theoretical cognition’s basic self-understanding. But they also reveal that there is more to such cognition’s relation to its object than is made explicit in the Copernican proposition. In particular, they show that this relation has two aspects, corresponding to theoretical cognition’s two previously distinguished features. As finite, such knowledge depends on affection by objects in order to come to be. But on account of its self-productive unity, it is nevertheless self-determining and so stands in a universally determining relation to its object, a relation entailing that the object must conform to the knowledge. Distinguishing these two aspects makes it possible to understand how the Copernican way of thinking is implicated in the basic self-understanding of theoretical knowledge, even though such knowledge is of objects that must be given from elsewhere.


s t e ph e n en g s t rom

2.3 Elaborating the Account It would be desirable to secure a better understanding of how the two aspects are related, particularly in view of what may seem to be a discrepancy between the preceding account and what is stated in the Critique. I just now characterized cognition’s determining relation to its object in terms of its form as legislative. But Kant identifies cognition’s relation to its object with what he calls its content (Inhalt).18 To address this issue, we need to take into account Kant’s ideas of the form and the content of knowledge. Doing so will help provide a context for understanding the Copernican proposal and also help us see how the traditional assumption arises. 2.3.1

Form and Content

Traditionally, and in Kant’s thinking as well, form is set over against matter.19 Content, in the sense of interest here, can be characterized as matter, but certain qualifications are in order. The distinction between form and matter is most familiar from its use in the context of human productive activity, where materials are determined – united, or ordered – through the introduction and maintenance of objectconstituting form. Thus, an artifact presupposes certain materials, which as such are suited to be determined through the introduction of form originally residing in the artisan’s knowledge of the object to be produced. Matter is thus what is determinable by form, and so far as it is determined it is informed matter, or the object produced by such knowledge. Form, in contrast, is the determination of the matter. As determination, form is act, the production of the object, or the production and maintenance of order and unity in the matter; it is the efficacy of the productive knowledge in respect of the matter, or, what comes to the same, the dependence of the object on that knowledge for its possibility and existence. The notion of informed matter yields the idea of content. Carried over to knowledge and conscious representation, this idea signifies the united material these contain within themselves. The content of representation or cognition may thus be regarded as its matter, with the proviso that such matter is always informed matter, in that content is found only within 18 19

See esp. A55/B79, A58/B83, A62–63/B87. For Kant’s statement of the distinction, see A266–68/B322–24. An insightful comparison of Kant’s use of this distinction with Aristotle’s can be found in Matthew Boyle, “Kant’s Hylomorphism” (unpublished manuscript).

Knowledge and Its Object


representations.20 And here form is the determining act that constitutes the cognition or representation as such, giving unity to the manifold material belonging to its content. In theoretical knowledge, as we noted, this determining act is a judgment, a combination of concepts that, in the basic case, determines the concept of its object through predication.21 But we also noted that since knowledge is self-productive, this act must lie in the growth of that concept, and that the determination must therefore be self-determination on the part of the concept of the object, the concept’s self-enlargement through the universally valid incorporation of another concept into itself. At this point, it might seem that we face a problem. If matter is the determinable and form the determination, how can any distinction between them be sustained in this case of self-determination, where the concept determined is itself the concept that determines? The difficulty, however, is merely apparent. In self-determination, the contrast between form and matter has a different application from the one it has in determination from without. In the latter case, form is introduced into the matter it determines. When for instance logicians of old applied the contrast to their definitions, they “called the universal the matter, but the specific difference the form.”22 But in cognition’s self-determination, form is not introduced into matter; rather matter is introduced into form – or, more precisely, content (informed matter) is introduced into the concept of the object through the concept’s taking matter up into itself while retaining its identity, its form.23 The content of the knowledge the determining judgment constitutes is thus the difference between that knowledge and the concept it presupposes, the difference registered in the specificity of the former as compared to the latter. And the form must reside in the concept of the 20


22 23

Thus Kant identifies content with matter (A6/B9, A59/B83), but also states that it is through an act of synthesis alone that “the elements for knowledge” “are united to a certain content” and that it is this act that first gives rise to knowledge (A77–78/B103); he thus holds that content arises only through the cognition-constituting form lying in the unifying act. Nor do his occasional references to the “stuff” that constitutes the material condition of the possibility of representation and content (see e.g., A76–77/B102, B145) entail that such stuff is ever encountered outside contentful representation. Our primary concern here is with discursive knowledge and hence with conceptual content, the content of conceptual representation, but intuitive representation also has content, intuitive content. A266/B322; cf. the “First Introduction” to the Critique of the Power of Judgment, CJ 20:215. Here “the concept’s taking matter up into itself” does not refer to a process of aggregation, but signifies that, although the cognition is an act of self-determination on the part of the presupposed concept it contains within itself, it nevertheless depends on conditions outside itself insofar as it contains more than that concept.


s t e ph e n en g s t rom

judgment’s object, as that concept’s capacity for such self-determination as constitutes the judgment. Since content is introduced into form in the knowledge established in determining judgment, and since such knowledge can in turn serve as the concept that determines itself in a further judgment, what figures as content in one relation can belong to form in another. Thus we can say of ‘form’ what Kant says of ‘principle’: that the term can be used in an absolute and a relative sense.24 In a relative sense we can speak of form where one element in knowledge has priority over another. The concept substance (as permanent), for instance, stands as form in relation to the concept matter (as movable), while including content beyond what is contained in the concept subject (of predication). But form in the absolute sense must lie originally, prior to all content, in rational cognition’s universal concept of an object.25 2.3.2

Form and Matter of Cognition’s Relation to Its Object

Theoretical knowledge constitutes a relation between a subject, or bearer of the capacity to know, and an object, or bearer of the capacity to be known, and its hylomorphic constitution is reflected in the presence of formal and material aspects in both its relation to the subject and its relation to the object. The subject’s capacity to know must comprise two capacities, namely, spontaneity, the internal condition of the possibility of cognition’s determining or ordering its material, and receptivity, the internal condition of the possibility of that material.26 Correspondingly, two aspects are to be found in cognition’s relation to its object. The object is determinable 24 25


A300–302/B356–58. As should be clear from what has just been stated, to say that form is prior to content is not to say that form is any more possible without content than content is possible without form. Form is what is common to all knowledge, not something actual independently of its relation to knowledge. This relation to content and matter is built into Kant’s conception of form as determination, in that there can be no determination except where something determinable is determined. I describe the hylomorphic relation between these capacities in Engstrom (2006). Here I will only note that comprehending the possibility of a priori knowledge requires that spontaneity and receptivity be conceived as so related that spontaneity generically determines receptivity itself, yielding an original intuition, which constitutes receptivity as a capacity of intuition and which on account of its unity stands as form determining the manner in which this capacity is specifically determined to represent by spontaneous synthetic activity under empirical conditions (cf. B150–52, 159–61). Through this generic determination, the form of knowledge establishes itself as the form under which alone objects can be sensibly intuited and hence be even so much as perceived in experience. Yet to the extent that the latter form incorporates an aspect that, not being entailed by the concept of receptivity in general, may reflect the specific character of our receptivity (e.g., the threedimensionality of our representation of space), it includes a subjective limitation that does not trace to the cognitive form originating in spontaneity.

Knowledge and Its Object


in respect of its form or nature by cognition’s universal legislation, even though knowledge depends on its object’s existence for its actuality, in that the object’s capacity to affect the subject in respect of its cognitive capacity constitutes the external condition of the possibility of cognition’s content.27 If Kant does not explicitly state that cognition’s relation to its object has these two aspects, that is not because he overlooks them.28 His recognition of them is reflected in the fact that he describes the relation in contrasting ways, which will seem to conflict with one another if not appreciated in the context of his hylomorphic conception of human knowledge. Thus, on the one hand, Kant frequently emphasizes that only through intuition do concepts and knowledge gain relation to objects (e.g., A19/B33, A62/B87, A155–56/B194–95), and he holds that what enables knowledge to relate to an actual object is sensation, the effect of an object on consciousness, which corresponds to the matter, or existence, of the intuited object (A19–20/B34, B146–47). On the other hand, he says, “Thinking is the act that relates given intuition to an object” (A247/B304), and he adds that in the absence of thinking, the presence in the mind of intuition, as a mere affection of sensibility, “constitutes no relation of such a representation to any object at all” (A253/B309). Each of these aspects of cognition’s relation to an object depends on the other; they are inseparably bound together as its material and formal moments. Just as the determinate thought (judgment) that constitutes knowledge of an object is not possible in the absence of sensible material, so the intuition through which an object is given depends for its unity on the understanding (B144n). This interdependence is clearly implicated in Kant’s account of the origin of cognition’s content. In the metaphysical deduction he identifies an act of spontaneity – the imagination’s synthesis of a manifold of intuition – as what first gives rise to content (A77–78/B103), and in the transcendental deduction he traces this act to its source in the understanding’s unity of consciousness (B137). Hence Kant’s identification of cognition’s relation to an object with its content, far from excluding a formal, spontaneous aspect 27


Here we can see that the Copernican way of thinking issues in a generic form of transcendental – or formal (B519n) – idealism. Because the object depends on the knowledge in respect of form only, not also existence, it must affect the subject’s receptivity in order to be known. It is therefore not a thing in itself, but an appearance, an object first given in sensible intuition and hence under a form that may include, in addition to the cognitive form of unity, a subjective limitation (see preceding note). In fact, they are clearly set out in the first paragraph of §14 (A92–93/B124–25), though there they are contrasted rather than described in their complementary relation to one another. They are also registered in Kant’s repeated statement that knowledge requires both that objects be given and that they be thought (A50/B74, B146), which fits closely with the contrast drawn between the determinable and its determination in his explanation of matter and form (A266/B322).


s t e ph e n en g s t rom

from that relation, rather requires that such an aspect be included, in that it is through cognition’s determination of its object (in the understanding’s act of uniting a manifold of sensible intuition) that its content, its relation to that object, is first constituted. 2.3.3

Diagnosing the Traditional Misconception

I mentioned earlier that if the Copernican way of thinking is implicated in the basic self-understanding of theoretical knowledge, then the traditional assumption must reflect some misconception, a misconception that needs to be accounted for. Now that we have seen that the Copernican view figures as a component of an account that distinguishes formal and material aspects of cognition’s relation to its object, we have the resources at hand to venture a diagnosis. Kant and traditional metaphysicians agree, as I noted, that the objects of theoretical knowledge must be “given from elsewhere.” But the traditional view does not separate the thought that knowledge depends in this way on the objects from the thought that knowledge must conform to them. It regards affections of the senses as impressions of form, and it conceives of the basic concepts and principles of the intellect along broadly analogous lines (cf. A120n, Prol 4:282, B167). That the traditional view fails to separate dependence in respect of matter, or actuality, from dependence in respect of form and moreover mistakenly assimilates the direction of the latter to that of the former suggests that it is not informed by theoretical cognition’s basic consciousness of its self-productive, self-determining character, even though this consciousness is registered in the traditional insight that theoretical knowledge is essentially activity. That implicit awareness has evidently been somehow blocked or occluded from reflective awareness, leaving nothing in place to prevent the assimilation of the direction of conformity to that of material dependence. What might account for the occlusion? One factor, presumably, lies in the salience of the consideration that what is distinctive of theoretical knowledge is its material dependence on its object. This dependence distinguishes such knowledge at once from infinite knowledge and from practical knowledge, both of which are productive of their objects (A92/B125). It is thus understandable that in traditional metaphysical reflection on theoretical cognition’s relation to its object, the material aspect of that relation should have a special prominence. A more insidious and more potent factor, however, is that when subjects of theoretical knowledge turn their attention on such knowledge and

Knowledge and Its Object


its relation to its object, they are liable to retain the attitude of a theoretical knower, in an attempt to gain theoretical cognition of theoretical cognition’s relation to its object. In any such attempt, the relation will be misconceived as one whose terms are both theoretically knowable objects (for instance a picture and what it depicts), and cognition’s essential selfdetermining character will, as a result, inevitably be lost from view. For the objects of theoretical knowledge are concrete things, representable in intuition, and as such determinate in all respects (cf. A571–72/B599–600). In contrast to the concepts employed in cognition, which come to be determined through self-determination, things already determinate in all respects come to be determined through alteration. And where one thing is altered by another the direction of formal dependence must follow that of material dependence, as when one thing impresses a form on another. The risk of such misrepresentation increases when, as often happens, reflection takes place in response to the discovery of error in judgment. For error directs attention away from the self-determining capacity to know (which all possible judging presupposes) and onto the liability to err (which conflict among actual judgments first reveals), even though the liability depends on the capacity. If, with attention focused on my liability to err, I consider what is needed to ensure that my fallible judgments are in agreement with the objects, the thought may easily come to mind that I must look outside myself to find assurance, and in particular that conformity to the objects is required. Of course, the thought that my fallible judgments must conform to the objects if they are to be true is a familiar one, and it can be in keeping with ordinary prephilosophical thinking. It is innocent so long as the conception of those objects is not severed from the original conception of them as knowable. But where the originally and universally presupposed self-determining capacity to know is blocked from reflective awareness, the thought that one’s fallible judgments must conform to the objects will be readily confused with the thought that knowledge itself must conform to the objects, resulting in the traditional misconception.

c h a p ter 3

Transcendental Idealism and the Transcendental Aesthetic Reading the Critique of Pure Reason Forward Lucy Allais

3.1 Introduction The first section of the Critique of Pure Reason, the Transcendental Aesthetic, is complex and dense. In this short section Kant discusses the ontology of space and time, conditions of mathematical cognition, and the role of space in representing objective particulars, introduces his novel and puzzling idea of a priori intuition and presents and argues for his complex form of idealism – transcendental idealism. Given this range of topics, the arguments in the section seem frustratingly brief. Not only is there no agreement in the literature on how to interpret either the argument for transcendental idealism or the position itself, there is not even agreement on exactly which part of the text is supposed to be Kant’s argument for the position. Kant’s project in the Aesthetic must be understood in relation to the rest of the Critique, so one way commentators have chosen to deal with the interpretative difficulties is to read later sections of the Critique into the Aesthetic, seeing the Aesthetic as presenting an abbreviated, suggestive first-take on arguments which cannot properly be made without later material, and as making use of concepts which cannot be fully understood at the point at which they are presented, and indeed need to be thoroughly revised in the light of what comes later. In contrast, in this essay I attempt to approach transcendental idealism and the argument for it in the Aesthetic by focusing primarily on the position and arguments as they are presented in this section. My view is that this should be our starting point, that it should constrain our understanding of later sections, and that we should revise central claims and arguments in the Aesthetic only if there is no other way of making sense of his position. A central notion in the Aesthetic is intuition; in my view, the role of intuition in Kant’s idealism is central but often underemphasized. I will argue that transcendental idealism is completely established in the Aesthetic and that both the nature 46

Transcendental Idealism and the Transcendental Aesthetic


of the idealism and the argument for it essentially turn on Kant’s notion of intuition. A common reading of Kant’s technical notion of intuition provides a clear example of the interpretative tendency to revise the Aesthetic in the light of later sections of the Critique. Kant starts the Aesthetic by saying that intuitions give us objects. He argues that our representation of space is an a priori intuition on the basis of the claim that we represent space as one, and he says that space is a condition of our being presented with objective particulars (things outside and other than ourselves) in empirical intuition.1 However, many commentators think later sections of the Critique imply that we could not represent anything singular or one, and in particular that we could not represent perceptual particulars or singular, unified objects outside us, without applying concepts and therefore that we cannot take Kant’s claims in the Aesthetic at face value.2 More specifically, it is frequently thought that Kant’s notion of synthesis, which features centrally in the Deduction, refers to something that is needed to constitute intuition and that Kant’s argument in the Deduction that the categories are needed for thought to have relation to an object implies that he thinks we cannot represent objects in any way without concepts. However, Kant tells us in the Aesthetic that he is abstracting from the contribution of concepts to experience (A22/B36) yet still talks about presentations of individual objects. Reading the Critique forward suggests that intuitions do present us with singular objects independently of the role of concepts. As I will argue, this is in fact central to Kant’s account of what intuitions are and what their role is. A central argument in the Aesthetic turns on an appeal to conditions of the possibility of geometrical cognition; however, the Aesthetic is concerned only with intuition and, for Kant, we do not have cognition proper without both concepts and intuitions. This means that the Aesthetic does not contain the materials that, for Kant, are necessary to fully explain geometrical cognition. The section does not explain, for example, the nature of geometrical inference, and therefore does not explain how geometrical claims are justified or established.3 Since later sections of the Critique add 1

2 3

This is asserted throughout the Critique and other critical writings (A19; A50/B74A320/B377; The Vienna Logic 24:905, LL 349; A713/B741; Metaphysik Mrongoius 29:800 LM 154, 29:888 LM 256; Metaphysik Vigilantius (K3 ) 29:970–73, LM 441–44). Intuitions are opposed to concepts which are general, mediate representations that do not give us objects (A68/B93; A69/B94 A141/B180; A239/B298). For example, Falkenstein (2006: 141), Ginsborg (2006), Griffith (2010), Longuenesse (1998/2001), McDowell (1994, 1998), and Rosenberg (2005: 62–63). This is discussed in the Doctrine of Method (A712/B741–A738/B766), which presents Kant’s account of the nature of geometrical concepts and geometrical inference. The transcendental deduction discusses the conditions of cognizing space as the unified object studied by geometry (especially B158– 62) by bringing our representation of space under concepts.


luc y al l ais

considerably to Kant’s account of mathematical cognition, the arguments in the Aesthetic could be read simply as abbreviated or incomplete versions of the later concerns. Furthermore, in later sections of the Critique Kant argues that his transcendental idealism enables us to see how cognition of synthetic a priori metaphysical claims is possible. Since Kant opens the Critique with the question of how synthetic a priori judgments are possible (B19) and discusses the conditions of the possibility of synthetic a priori geometrical claims in the Aesthetic, his explanation of the possibility of metaphysics could be read back into the Aesthetic. The idea would be that transcendental idealism enables us to explain how geometric claims can be justified: roughly, synthetic a priori geometrical cognition is possible because it is cognition of objects that depend on our minds. In contrast, I will argue that Kant’s appeal to a priori intuition as a condition of the possibility of geometry in the Aesthetic is a specific concern distinct both from his complete account of geometrical cognition and from his explanation of the possibility of synthetic a priori metaphysical claims. His concern is with the specific question of how synthetic a priori geometrical claims can concern given objects and therefore qualify as cognition. Kant argues that this requires a priori intuition and he takes this to lead to idealism. Having established idealism, Kant then takes this to be essential to explaining the possibility of cognizing metaphysical claims. Kant’s transcendental idealism involves a distinction between things as they are in themselves and things as they appear to us (appearances). In the Aesthetic he argues that “the things that we intuit are not in themselves what we intuit them to be” (A42/B59). He takes this to follow from arguments for the claim that our representations of space and time are the a priori forms of our intuition. There is no agreement in the literature about how to understand Kant’s distinction between things in themselves and appearances, or about the claims Kant makes about each of appearances and things in themselves. There is disagreement as to whether his distinction is metaphysical or epistemological, about whether or not Kant is a metaphysical idealist about appearances, and if so, what kind of idealist (Berkeleyan phenomenalist? antirealist?), and about whether Kant’s notion of things as they are in themselves involves a metaphysical commitment to an actual aspect of reality which we cannot cognize.4 Looking at both 4

Allison (1983 [2004]), Bird (2006), and Prauss (1971, 1974) are influential and important representatives of the view that transcendental idealism is primarily an epistemological position. Guyer (2007) and Van Cleve (1999) see it as a metaphysical position involving an extreme, phenomenalist form of idealism. Ameriks (2003, 2006) and Allais (2004, 2007) are examples of commentators who see transcendental idealism as a metaphysical position while not seeing Kant as an extreme, phenomenalist

Transcendental Idealism and the Transcendental Aesthetic


the way Kant describes the position and his argument concerning a priori intuition, I argue that the Aesthetic clearly supports seeing transcendental idealism as some kind of metaphysical idealism but that it does not mandate Berkeleyan idealism or phenomenalism with respect to appearances, and is more straightforwardly read as presenting a form of antirealism that rejects experience-transcendence. With respect to things in themselves, I will argue that the most straightforward reading suggests that Kant thinks that there is an aspect of reality that we cannot cognize, although it is not impossible to read the text in a way that denies this.

3.2 Intuition Kant opens the Aesthetic talking about empirical intuition, rapidly introduces the idea of a priori intuition, presents arguments for the claim that our representations of space and time are a priori intuitions, and says that it follows from this that space and time are not features of things as they are in themselves, and that objects in space and time are mere representations. Understanding what intuition is, and what a priori intuition is, is clearly crucial. The first few sentences of the section emphasize the role of intuition in cognition, which is contrasted with the role of concepts. The claims concern cognition (Erkenntnis) which, for Kant, is not the same as knowledge (Wissen), a point to which commentators are paying increasing attention, and which is important for understanding transcendental idealism.5 In Kant’s account, cognition, unlike knowledge, can be false (B83), and what is relevant to whether or not something qualifies as cognition is not whether it has some specified kind of justification or warrant, but rather the kind of representation of objects with which it is able to provide us. Both at the empirical and the a priori level Kant’s primary concern is with what it takes for us to achieve a certain kind of objective representation of the world (cognition), rather than with the kind of warrant required for knowledge. In particular, Kant’s central argument in the Aesthetic does not present considerations about the justification of geometrical knowledge,


idealist. Bird (2006), Hanna (2006), and Senderowicz (2005) deny that Kant’s notion of things in themselves involves a commitment to an existing aspect of reality that we cannot cognize. Langton (1998) sees Kant’s distinction between things as they are in themselves and things as they appear to us as a metaphysical distinction between intrinsic and extrinsic properties, which need not involve any idealism. The literature on the topic is enormous, and these are simply a few representatives of a few central interpretative stances. See Schaffer (forthcoming).


luc y al l ais

but rather about a necessary condition that must be met for geometrical claims to qualify as cognition. Kant says that the specific contribution intuitions make to cognition is that of giving us objects and that being given an object is an immediate relation to it (A19/B33). He says that “all thought . . . must . . . ultimately be related to intuitions . . . since there is no other way in which objects can be given to us” (A19/B33). This is contrasted with concepts, which enable us to think objects. Kant does not explain what he means by objects being given to us. One possible reading is objects causally affecting us: the idea would be that an object is given to us when we have some sensory input (some sensations) caused by this object. However, this reading is made implausible by the fact that Kant immediately goes on to say that what is given in intuition has a form and a matter, that the latter corresponds to sensation, but the former is a priori. Being given an object in intuition therefore is something which is more than or other than simply having sensations through being causally affected by objects. Kant clearly thinks intuitions are not sensations, and that intuitions, and not sensations, give us objects. Another way the giving us objects by intuitions could be understood is that intuitions phenomenologically represent their objects as really seeming to be there. This is also not a straightforward reading of the opening text, where Kant says that intuitions give us objects not phenomenologically immediate–seeming images of objects. In my view, what Kant means by the claims that intuitions give us objects and that intuitions are immediately related to objects is that when we have an intuition we have acquaintance with its object: intuitions put us in direct mental contact with objects or present their objects to our consciousness. (So a mere mental image or a hallucination is not an intuition.) Understanding intuitions as giving us acquaintance with objects not only explains the immediacy of intuition, it also fits the singularity of intuition: intuitions individuate their objects uniquely because they actually present them (which need not be true of an image of an object, which could look like more than one thing). The contribution to cognition that intuition makes is to present us with the objects about which we can then make judgments, using concepts. Without acquaintance with objects thought does not constitute cognition; it fails to connect with a world. If we read the opening of the Aesthetic through Kant’s claim in the Deduction that the application of a priori concepts is needed for experience of objects, at the same time as failing to notice that ‘experience’ here is technical, and means empirical

Transcendental Idealism and the Transcendental Aesthetic


cognition, we may think that Kant cannot really mean that intuitions give us objects (Gegenstände). This is, however, what he says. Kant then says that “The underdetermined object of an empirical intuition is called appearance.” Two things are worth noting about this. One, Kant’s idealism is frequently expressed in claims he makes about appearances (which, he later claims, are mere representations), but at least this first use of ‘appearance’ need not be read as expressing idealism; it could simply refer to what it is that appears to us. In this sense, appearances are objects that are presented to us in perceptual experiences (as a result of affecting our senses), as opposed to, for example, nonsensible objects like numbers, Leibnizian monads or Cartesian souls. Second, as we have just noted, many commentators think that intuition cannot present us with objects without the application of concepts and the corresponding role of synthesis; however, Kant here speaks of intuition as presenting us with appearances and with objects. In fact he tells us that an object which is given to us and has not yet been brought under concepts and thereby determined is an appearance,6 and that such an appearance is what intuition gives us. What empirical intuitions give us are objects, not sensations; intuitions give us these immediately, and we can be given objects in empirical intuitions only if these objects affect our senses. Kant introduces his central and crucial notion of the a priori form of intuition very rapidly, and without much argument. In the introduction, he has told us what he means by a priori: “independent of all experience and even of all impressions of the senses” (B2).7 In the Aesthetic he says that appearance has a matter, which corresponds to sensation, and a form, and that the form is a priori: “[s]ince that within which the sensations can alone be ordered and placed in a certain form cannot itself be in turn sensation” (A20/B34).8 In addition to showing that on Kant’s account intuitions are not sensations, a further important point to note here is that although Kant thinks that the sensory input needs to be structured and ordered in order for us to be given objects, he does not speak of this as involving synthesis or concepts, but rather as involving a priori intuition. The conceptually 6 7 8

As Bird (2006: 105) says, “the terms ‘bestimmt’ and ‘Bestimmung’ are typically used for the descriptive properties we ascribe to experienced items.” He says that necessity and universality are “secure indications of an a priori cognition” (B4). Paton (1936: 94) suggests that rather than reading this as a (too) brief argument, we can read it as a statement of what is to be proved. In support of this, Kant does often present arguments in this way. For example, his central arguments in the Aesthetic, those presented under the headings of the Metaphysical Exposition and the Transcendental Exposition, all start by stating the claim to be proved.


luc y al l ais

governed syntheses Kant invokes in the Deduction are sometimes understood as what is needed to introduce order out of the mass of sensory input in such a way that we can represent objects (and therefore as corresponding to Kant’s solution to what is now called the binding problem).9 However, in the Aesthetic, the a priori ordering of sensation that Kant says is necessary for us to be given objects is not conceptually governed synthesis but rather an a priori form of intuition. This suggests that conceptually governed synthesis is doing something other than producing perceptual particulars out of the mass of sensory input.10 A final point to note is that Kant says that the form of intuition is also an intuition – pure or a priori intuition. Since intuitions give us their objects (the things they represent) immediately, and the form of intuition is an intuition, whatever it is that a priori intuition represents will be something that is present to us.

3.3 Space Is the A Priori Form of Our Intuition Kant says that we represent objects as outside us and in space, and asks what space is. He gives three possible alternatives: either it is an actual entity or it is (mind-independent) relations between things, or it consists of relations that “attach to the form of intuition alone, and thus to the subjective constitution of our mind” (A23/B37–38). He is going to argue for the third option. In the section titled the “Metaphysical exposition of this concept” he then gives four (very brief ) arguments, based on the way we represent space, for the claim that our primary representation of space is the a priori form of our intuition. (This is in the second edition; in the first edition there are five arguments, one of which is replaced in the second edition with the subsequent Transcendental Exposition.) The first two arguments are for the claim that our representation of space is a priori. First, he argues that our representation of space cannot be empirical, because we represent things as distinct from us and other things by representing them as spatial. Some commentators think that the argument concerns the kind of warrant we have for thinking objects have the geometrical spatial properties we represent them to have.11 Another possibility 9 10


See Treisman (2003) for a presentation of the binding problem in psychology. A possible view is that both conceptually governed synthesis and the a priori forms of intuition are needed for us to be given objects. However, Kant never says this, and in fact denies that concepts give us objects. Concepts do, of course, play a crucial role in experience, or empirical cognition of objects, by enabling us to think objects, but this is something more than simply being presented with objects. See, for example, Paton (1936: 83).

Transcendental Idealism and the Transcendental Aesthetic


is that since Kant later argues for a priori conditions of empirical knowledge, we could read this later concern into this argument, and see him as giving a very abbreviated argument for the claim that space is an a priori condition of empirical knowledge. However, he doesn’t argue for, but rather simply states that the way we represent things as distinct from ourselves is by representing them as in space, and he doesn’t say anything about needing assurance or certainty for certain kinds of knowledge claims.12 He seems to assume that whatever framework is the condition of our being presented with particulars cannot be abstracted from our representation of particulars. The second argument in the Metaphysical Exposition is that our representation of space is a priori because we can represent the structure of space independently of the particular arrangement of objects in space, but we cannot represent particular objects without representing them as spatial. Kant does not separate showing that our representation of space is an a priori intuition and that it is the form of our intuition, but we can see in these first two arguments why this is so, because of the structuring role this a priori intuition plays in our representing objects. These two arguments, in particular the claim that we cannot represent objects without representing their spatiality, are crucial to seeing why Kant will take the ideality of objects in space to follow, without further argument, from the ideality of space. Since, in Kant’s view, the representation of space is so pervasive to our representation of objective particulars that we cannot abstract spatiality from our representation of them, we will have no way of representing them that will not inherit from our representation of space its ideality. Kant then presents two arguments for the claim that our representation of space is an intuition. He says that our representation of space is essentially singular: we represent space as one and then represent parts of space (spaces) as limitations of this one space, rather than representing a whole of space by first representing parts (spaces) and then ordering or organizing these spaces together to make up a whole. And he says that the infinity of space is something given to us. Our representation of space immediately presents us with space as something infinite and singular or one, and does this a priori. The claim that space is given to us as infinite and as one might seem to imply that we somehow represent the whole of an infinite thing. But, for Kant, represented space is not a thing: it is a form.13 The form of the 12 13

Unlike, for example, the way we could construct a representation of color space or brightness space from our representation of particulars. (See Warren [1998: 202].) And he will argue later, in the first Antinomy, that we cannot represent space as a whole.


luc y al l ais

structuring representation in which all objective particulars are situated is that it always goes on (is infinite) and that every part of space is spatially related to every other part (space is one).14 Three important points can be noted from these arguments. First, since one of Kant’s argument for the claim that space is an intuition and not a concept turns on the claim that we represent space as singular, the conceptualist denial that intuitions could present something singular or one without concepts would make nonsense of the argument. Second, the conceptualist view that we could not represent any kind of unity without conceptualization is also undermined, since Kant clearly thinks that there is a specific way in which intuition presents oneness or unity: as a whole that is presented prior to its parts. The role of conceptually governed synthesis in the Deduction is to enable us to cognize objects as unified complexes of parts or properties, which is different from, and in fact depends on, having singular intuitional units. Third, Kant thinks that the infinite form of intuition is immediately given and not constructed. This, as we will see, is significant for how we understand the transcendental ideality of space, as many phenomenalist readings require seeing the empirically real world as a construction. In the next section, added in the second edition, the “Transcendental exposition of the concept of space,” Kant presents a further argument for the claim that our representation of space is a priori and an intuition. He says that geometry determines the properties of space synthetically yet a priori, and asks what the representation of space must be to make this possible. The answer is that it must be an intuition (in order to explain how geometry can be synthetic) and that it must be a priori (in order to explain the fact that geometrical claims are necessary, not contingent). It might be thought that Kant’s appeal to geometry here concerns the justification of geometrical claims – the idea that we need a certain representation of space to explain geometrical inference, or to explain how we can guarantee that geometrical claims apply to all objects. However, he does not say anything here about how geometrical inference works or how geometrical claims can be justified, and he does have an account of this at the end of the Critique (in the Doctrine of Method), involving constructions in pure intuition. Rather, his concern is with how it is possible that geometrical claims can qualify as synthetic a priori cognition. Geometrical claims will not be synthetic cognition unless they have an object given in intuition, but 14

As Rosenberg (2005: 65) puts it: “One can’t think of a point, line, place or solid, except as a point, line, plane or solid in space. But to be located in space at all is to be in one place rather than another, and so the whole of space is implicit in every spatial experience.”

Transcendental Idealism and the Transcendental Aesthetic


since they are necessary, this object must be given to us a priori (A48/B65). The only way this is possible is if something is given to us in a priori intuition. Since he started with the assumption that space is the object studied by geometry, he concludes that our representation of space is an a priori intuition. Two things should be noted about this argument. First, if Kant’s primary worry about the possibility of synthetic a priori claims is about how it is possible for them to concern objects with which we have acquaintance, and therefore qualify as cognition, he could have different accounts of how different kinds of synthetic a priori claims are justified or established.15 Second, his explanation of the possibility of geometrical cognition does not appeal to idealism: he does not say, for example, that it is because our minds make objects in certain ways that we can know a priori that certain claims about them are true. He simply says that synthetic a priori geometrical cognition is possible if our representation of space is both an intuition and a priori. Having argued for the claim that our representation of space is an a priori intuition, Kant then asks how it is possible for a representation to both be an intuition (a representation that immediately gives us its object) and be a priori, and answers that this is possible only if what is presented is something subjective: merely the form of subjects’ intuition (B41). Thus, he takes idealism to follow, and to follow immediately, from combining the notion of intuition with a priority. He states two conclusions: (a) “Space represents no property of any things in themselves nor any relation of them to each other, i.e., no determination of them that attaches to objects themselves and that would remain even if one were to abstract from all subjective conditions of intuition” (A26/B42) and (b) “Space is nothing other than merely the form all of appearances of outer sense, i.e., the subjective condition of sensibility, under which alone outer intuition is possible for us” (A26/B42). These conclusions (a) and (b) express the transcendental ideality of space.


The Argument for the Transcendental Ideality of Space

As the argument is presented, Kant takes the transcendental ideality of space to follow without further argument from having established that our representation of space is a priori and is an intuition. It might not seem 15

Which fits the fact that he justifies metaphysical synthetic a priori claims with transcendental arguments showing them to be conditions of the possibility of experience, and thinks mathematical synthetic a priori claims are justified by construction in pure intuition.


luc y al l ais

obvious why he thinks this. After all, the arguments concern our representation of space while the conclusion concerns space. Seeing space as the structure in which we organize the sensory input in our representation of the world seems compatible with realism about space.16 We might think that having given arguments for the claim that space is the a priori form of our intuition, Kant needs “some further argument to prove that space and time are nothing but these representations of them” (Guyer 2006: 66).17 A strategy for finding such further arguments could be to appeal to what is needed for our a priori knowledge of the conditions of the possibility of experience, or for assurance that geometrical claims will be true of objects.18 Idealism could be invoked to provide an explanation of how it is possible to justify synthetic a priori claims, or how it is possible to guarantee that they apply to objects. However, as we have seen, in the Aesthetic Kant talks about cognition rather than knowledge, and in fact about one specific ingredient of cognition: a priori intuition. And as he presents his argument, Kant does not think further argument is required to move from the claim that space is the a priori form of our intuition to the claim that space is merely subjective and does not present us with things as they are in themselves. In the first edition Kant takes his conclusions (a) and (b) to follow immediately from showing that our representation of space is an a priori intuition. The only further reasons he gives for this are the following. In support of claim (a) he says that no “determinations can be intuited prior to the existence of the things to which they pertain, thus be intuited a priori” (A26/B42). The specific claim is that we cannot intuit things a priori. In support of claim (b) he says that we can understand how the form of all appearances can be given prior to all actual perceptions if it is nothing but the a priori form of intuition. What he appeals to in both these explanations is the claim that we can make sense of something being intuited prior to or independent of experience only if it is something that doesn’t exist independent of our representing it. And this is precisely the explanation Kant adds in the second edition, where he asks how an outer intuition can “inhabit the mind that precedes the objects themselves?” He answers “Obviously not otherwise than insofar as it has its seat merely in the subject, as its formal constitution for being affected by objects and thereby 16

17 18

Even more strongly, Guyer (2006: 63) says, “If we somehow know a priori that we can only perceive objects distinct from ourselves in space, indeed in three-dimensional Euclidean space, why isn’t the explanation of our success in perceiving some particular outer object precisely that it really is spatial, indeed three-dimensional, quite apart from our representing it as such?” (see also Guyer 1987: 349). See also Hatfield (2006: 76, 82) and Setiya (2004: 67). See Gardner (1999: 82), Guyer (1987: 362), Parsons (1992: 83), and Wood (2005: 27).

Transcendental Idealism and the Transcendental Aesthetic


acquiring immediate representation i.e., intuition of them, thus only as the form of outer sense in general” (B41, my italics). Kant does not say that our insight into geometrical claims can be explained by our minds’ being responsible for the geometrical nature of objects. He simply says that something can be an outer intuition and a priori (precede the objects) only if it is something merely subjective. In order to understand Kant’s argument then, what we need is a reason for thinking that an a priori intuition couldn’t represent a mindindependent feature of reality. If we focus on this question, then rather than seeing the argument in the Aesthetic as a problematically abbreviated argument concerning the justification of certain synthetic a priori claims, we can find what we are looking for in the way Kant introduces intuition: the fact that intuition is immediately related to its object, or gives us its object. (‘Object’ here is understood in the very general sense of what it is that is represented; clearly space is not an object in the sense in which tables and chairs are objects.) On my reading, an intuition gives us acquaintance with its object: this means that the object that an intuition presents is present to consciousness. Kant thinks that the way things that are independent of us get to be present to our consciousness is by affecting our senses (A19/B33). What an a priori intuition presents is present to consciousness (it is an intuition) but does not involve anything affecting our senses (it is a priori). As I read him, Kant thinks this means that what an a priori intuition presents us with cannot be something that exists independently of our minds. If intuitions were merely a phenomenologically immediate–seeming image of an object, it would be hard to see why such an image could not represent something that exists independent of us. Why could we not have an image-like representation of something which seems to be really there without the thing affecting our senses? Why could we not have an imagelike representation that matches the form things have in themselves? If we read intuition in this way we might need to appeal to arguments about the kind of assurance that is needed for knowledge claims, for example, that we cannot be sure that our image matches mind-independent reality. But this is not what Kant appeals to. He does not say that we need a guarantee of a fit between our representation and reality, or appeal to other such thoughts about justification, but rather says that it is simply not possible to have an intuition a priori. He simply says that determinations of independent things cannot be intuited a priori (A26/B42), that an intuition can be a priori “[o]bviously not otherwise than insofar as it has its seat merely in the subject” (B41), and that “there is therefore only one way possible for my


luc y al l ais

intuition to precede the actuality of the object and occur in an a priori cognition, namely, if it contains nothing else except the form of sensibility, which in me as subject precedes all actual impressions through which I am affected by objects” (Prol. 4:282). As I read him, Kant says that we can understand how something can be immediately present to consciousness without anything affecting us only by seeing that it does not exist independently of our minds. As I understand Kant’s notion, intuitions are not representations which point beyond themselves to some object they picture or describe; they do not represent indirectly. (Something that represented its object without its object’s being present would not be an intuition, but rather a mere image.) It is because Kant thinks that intuitions do not represent their objects in this way that his conclusion follows. Space is not merely pictured in our a priori intuition but it is actually present in it. Since this presentation is independent of anything affecting us, Kant thinks that what is presented cannot be something that exists independent of us and our representing it. This is why he takes showing our representation of space to be an a priori intuition to show that space represents no property at all of things in themselves and is nothing other than merely the form of our intuition, without further argument being required.

3.5 Interpreting the Transcendental Ideality of Space We now have two guides for interpreting transcendental idealism as it is presented in the Aesthetic: the things Kant says to describe the position and the argument that he presents for it. I start with the former. We have seen that Kant’s conclusion about the status of space is that it is nothing more than the form of all appearances of outer sense and it is not a feature of things as they are in themselves independent of the subjective conditions of intuition. These claims have frequently been read as expressing an extreme, phenomenalist idealism about space. On the other extreme, some interpreters see transcendental idealism as merely empirical realism: Kant’s claim that we know only appearances is understood as saying that we can know only things that present themselves to our senses, rather than as saying that the things that present themselves to our senses are mind-dependent.19 In my view, a moderate, antirealist form of idealism is better supported by the text than either of these extremes, and is also better supported by the argument Kant gives for his idealism. 19

A clear example is Bird (2006).

Transcendental Idealism and the Transcendental Aesthetic


In support of the merely empirical realist interpretation, it can be pointed out that Kant’s initial statement of his conclusion about space is that space is nothing but the form of all appearances of outer sense. Since ‘appearance’ can be read in a neutral (nonidealist) sense as simply referring to that which appears to us, saying that the conditions of sensibility are restricted to things that can appear to us could be taken to mean that it is restricted to things that we can sense. Thus, Kant can be read as expressing the idea that space is something that belongs only to the kinds of objects of which we can have sense experience, as opposed to, for example, objects or entities like God, Leibnizian monads, or Cartesian souls, which do not affect our senses. Mere empirical realist interpreters tend also to deny that Kant is committed to there existing an aspect of reality that is in principle uncognizable by us. This can be made consistent with the text if we start by assuming that ‘things in themselves’ refers to pure reason’s experiencetranscending ‘ideas’ of such intelligible objects as Leibnizian monads or Cartesian souls: Kant’s saying that space represents no property of things in themselves could be taken to be reinforcing the point that the space that we represent belongs only to the objects that are presented to our senses, and not to these other kinds of objects. In my view, it is not impossible to reconcile this interpretation of Kant’s empirical realism with the text, but it is far from straightforward. If Kant merely meant to say that space is the form of the objects that affect our senses rather than of Leibnizian monads, it is hard to see that he needs much argument for this, much less the specific argument he presents. In fact he seems to already state this claim as one of his premises, in argument 1 of the Metaphysical Exposition. And if he meant his notion of things in themselves to refer specifically to nonsensible objects like Leibnizian monads, we would expect him to have noted this. A far more straightforward way of reading the notion is as referring to the way things are independently of their relations to us, or independently of their relations to other things. On this reading, the claim that our representation of space does not represent things as they are in themselves amounts to saying that it does not represent the way things are independently of their relation to our consciousness. And while, on its own, the claim that space is nothing but the form of appearances of outer sense need not be read in an idealist way, Kant goes on to clarify his position by saying that “[w]e can accordingly speak of space, extended beings, and so on, only from the human standpoint” (A26/B42). He says that “the proposition: “All things are next to one another in space,” is valid under the limitation that these things be taken as objects of our sensible intuition” (A27/B43). These claims express


luc y al l ais

at least some form of antirealism. Antirealism includes a family of positions which deny the reality of what transcends human experience or knowledge, although they do not see physical objects as existing literally in our minds.20 Different versions of antirealism have different accounts of what they hold reality to be limited to, with variations including knowledge, warranted assertability, and possible knowledge. The specific version Kant clearly expresses in these claims is that the world in space and time is limited to what can possibly be presented to our intuition. The feature of the text that most supports the extreme, Berkeleyan or phenomenalist idealism is Kant’s saying that appearances are mere representations (A28/B44; A30/B45). This can be taken to mean that outer objects exist only in subjects’ minds. However, this is not a compulsory reading: appearances could be representations in the sense that they are things that are or can be presented to human consciousness: things that are represented.21 Understood in this way, the claim that appearances are mere representations expresses, at least, antirealism, since it says that appearances do not go beyond what can be presented in human consciousness, they are nothing but what can be presented to human consciousness. Seeing Kant as an antirealist makes better sense of his claim that his position combines empirical reality and transcendental ideality than does seeing him as a phenomenalist (A28/B44). In the Aesthetic, Kant does not explain empirical reality, as a phenomenalist should, in terms of relations of coherence and consistency within representations in our minds, but simply in terms of things existing outside our minds. However, he claims that these things do not transcend our possible experience, and in this sense are transcendentally ideal. Antirealist readings can allow that objects exist outside us in space but do not exist independently of the possibility of our experience of them. So far we have looked at the passages in which Kant describes his transcendental idealism. The other feature of the Aesthetic relevant to interpreting the position is Kant’s argument for the position. I will give some reasons for thinking that this supports the antirealist interpretation better than it does the phenomenalist one. On my reading, Kant’s argument for the ideality of space is based on thinking that a representation cannot be an a priori intuition and yet present something that exists independent of our minds and our representing it. The reason for this, I suggested, is that intuitions involve the 20 21

See, for example, Dummett (1993), Walker (1995), and Wright (1992, 1996). Kant’s term Vorstellung, translated as ‘representation,’ could also be translated as ‘presentation’: something that is put before the mind

Transcendental Idealism and the Transcendental Aesthetic


presence to consciousness of their objects (objects being immediately given) and that Kant assumes that objects that are independent of us get to be present to our consciousness (given to us) by affecting our senses. As I read him, Kant takes it to follow from this that an a priori intuition (an intuition we have independent of anything affecting us) does not present us with something that exists independently of our actual or possible representation of it. It might be thought that this leads to some kind of phenomenalist interpretation, because it might be thought that something that does not exist independently of our representing it must exist merely in our minds. However, antirealism also gives a sense in which what we represent does not exist independently of our representing it. The version of antirealism we are considering here is one that limits space (and objects in space) to what can possibly be presented to us in intuition: what is presentable to us in intuition. To say that space is limited in this way is not to say that it exists literally in our minds. A straightforward realist can agree that space, and spatial features of objects, are presented to human consciousness; the antirealist agrees with this, but adds that space (and objects in space) does not transcend what can be presented to our consciousness. Thus, it does not exist independently of the possibility of our representing it. Kant’s argument does not turn on the idea of things existing literally in our minds, but of things existing in being presentable to consciousness. The argument is therefore compatible with an antirealism which says that space is a form or structure in which all objects outside of us are located, which does not exist apart from the possibility of its being present to our consciousness. The antirealist reading fits better than does the phenomenalist with Kant’s claim that space is immediately given in intuition. A straightforward phenomenalist sees space (and empirical reality) as a construction based on what is actually present to us in experience: a construction based on features of empirical ideas. But this clearly cannot be Kant’s account: he precisely denies that space could be constructed on the basis of features of empirical objects that are presented to us in experience. This is central to his argument for the claim that space is a priori. Berkeley thinks space is constructed by abstraction from ideas in us. Kant argues that our representation of space could not arise in this way: it could not be abstracted from our experience of particulars in space, because it is a requirement of perceiving distinct particulars. For Kant, truths about space are not just truths about what is actually in subjects’ minds or about what we have actually perceived, but neither are they truths about what can be extrapolated from or constructed on the basis of what is in subjects’ minds. Rather, they are truths about what could possibly be presented to subjects like us.


luc y al l ais

It is crucial to Kant’s arguments in the Metaphysical Exposition that our representation of space is not constructed but rather is immediately given: his account of our representation of the infinity of space is precisely not that we construct such a representation: the infinity of space is immediately given, not constructed. He says that we represent space as an infinite, immediately given, single structure that is present both in our experience of objects and when we do pure mathematics. He thinks that constructing a representation of an infinite space would require first representing parts of space and then running through and synthesizing these parts. Kant thinks, in contrast, that we represent parts of space by representing limits to the immediately given singular infinite structure that is presented in intuition, and that representing the parts of space in this way is subsequent to an immediate representation of space as one. Kant thinks that what an a priori intuition presents is limited to what can possibly be presented to our consciousness; he does not argue that it is limited to what we can construct. On the contrary, he thinks that what can be presented to us places limits on what we can construct. If space exists merely in our minds it must either exist as a property of what is actually in our minds or as something that can be constructed out of what is actually in our minds. But neither of these is Kant’s account of space. The antirealist reading, in contrast, sees space as something that does not exist apart from the possibility of its being presented to minds like ours.22 The space that we cognize is a structure or form that is present to us immediately, without boundaries (rather than being constructed out of bounded parts put together), that provides limits on what it is possible for us to perceive and what it is possible for us to construct.

3.6 Conclusions It is hard to make any sense of the argument in the Aesthetic unless we allow that intuitions present us with singular things. The central argument in the Aesthetic works at the point at which Kant presents it, without the need for further premises or reconstruction, if we allow (as fits straightforwardly with the opening claims of the Aesthetic) that intuitions actually present their objects to consciousness. The kind of idealism that Kant takes to follow from his argument is, at least, a form of antirealism which limits space and time and objects in space and time to what can possibly 22

The idea of space as something essentially presentable to us fits with points emphasized by O’Shea and Rosenberg, about how space, on Kant’s account, is essentially orientable and perspectivally presentable (Rosenberg 2005: 65; O’Shea 2014: 84–103).

Transcendental Idealism and the Transcendental Aesthetic


be presented to human intuition. This is a form of antirealism that rejects the (empirical) reality of anything which cannot be manifested in an experience, anything with which it is not, in principle, possible for us to have acquaintance. The explanation of mathematical synthetic a priori cognition in the Aesthetic does not involve saying that we have a guarantee of certain knowledge claims because we make objects in such a way that they come out as true. The explanation of mathematics is not the same as the explanation of synthetic a priori cognition in metaphysics, since Kant says at the end of the Aesthetic that we now have “one of the required pieces for the solution of the general problem of transcendental philosophy – how are synthetic a priori propositions possible?” In my view, our starting point in reading other central arguments (such as Kant’s account of synthesis and the notion of ‘relation to an object’ in the Deduction) should be to look for readings which fit with this account, and give up on this reading of the Aesthetic only if this proves to be impossible.

ch a p ter 4

Kant on the Ideality of Space and the Argument from Spinozism Michela Massimi∗

4.1 Introduction Kant’s engagement with Newton’s absolute space is complex and problematic. The received view goes that after endorsing relationism about space in Physical Monadology, Kant came to defend Newton’s absolute space in the 1768 text Directions of Space. But Kant’s flirting with Newton’s absolute space was short-lived, soon to be ended with the Inaugural Dissertation in 1770, where the ideality of space was first introduced, and fully defended in the Critique of Pure Reason. Yet, absolute space continues to appear in the Metaphysical Foundations of Natural Science, in the “Phenomenology” chapter, this time as an idea of reason (for an influential interpretation, please see Friedman [1992: Chapter 4; 2013: 474ff.]). In this essay, I focus on one particular aspect of Kant’s departure from Newton’s absolute space: namely, the role played by seemingly Newtonian assumptions behind one of Kant’s mature arguments for the ideality of space, what I call the argument from Spinozism. The argument from Spinozism is not Kant’s main argument for the ideality of space. It does not even feature in the Metaphysical or Transcendental Exposition of the Transcendental Aesthetic of the first Critique, and can be found instead in the Critique of Practical Reason, among other places (primarily, Kant’s lectures on metaphysics). I have two main reasons for focusing on the argument from Spinozism. First, this argument betrays, in my view, the real and profound reasons why Kant could not endorse Newton’s absolute space. I see the argument from Spinozism as expanding on and clarifying Kant’s criticism of the Newtonians to be found in the Transcendental Aesthetic. My first goal then is to ∗

I am grateful to Karl Ameriks, Peter McLaughlin, Jim O’Shea, and Eric Watkins for comments on an earlier draft of this essay. This research originates from a Leverhulme Trust international network grant IN-081 on Kant and the Laws of Nature, whose support is gratefully acknowledged.


Kant on the Ideality of Space and the Argument from Spinozism


clarify the argument from Spinozism, elucidate its premises and structure, and highlight what I take to be its main – Newtonian in spirit – premises. I hope to show that Kant’s official line in the first Critique against the Newtonians – portrayed as the “mathematical investigators of nature” positing “two infinite, eternal and self-subsisting non-entities” – and, hence, his rationale for endorsing idealism about space, hides in fact more worrisome considerations. I focus on the argument from Spinozism also for another reason. Despite the Newtonian-sounding premises of the argument, Kant introduced a further assumption about God’s omnipresence being the determining ground for motions (including human actions), and it is this assumption that carries the full weight of the Spinozistic charge. I argue that for the argument to go through, Kant had to introduce this further assumption, which is not to be found in Newton, but must instead be read into Newton’s view. Moreover, I show how Kant’s profound reasons for associating absolute space with Spinozism have to be found elsewhere; namely, not in the debate surrounding Newton’s own view and Newtonianism about space (although there certainly was such a lingering charge of Spinozism at the time). Instead, they have to be looked for in an influential metaphysical tradition that – from Malebranche, to Leibniz, and Baumgarten – addressed what I call the problem of the world as a totality of substances in interaction. In this essay, I argue that we should read and understand Kant’s defense of idealism about space in the argument from Spinozism against this intellectual backdrop. In Section 4.2, I briefly review Kant’s famous criticism of Newtonians in the Transcendental Aesthetic, and Newton’s own view about space – famously presented in the General Scholium to the Principia and in the unpublished De Gravitatione. In Section 4.3, I unpack the troublesome relation between God and space in Kant’s argument from Spinozism (as expounded in the Critique of Practical Reason) by elucidating the premises and structure of the argument. Finally, in Section 4.4, I illustrate the intellectual backdrop against which, I urge, we should read the argument from Spinozism for the ideality of space. I show how Kant during the Critical turn engaged with the problem of the world, and, as a solution to it, came to reinterpret Newton’s absolute space as a phenomenon of the divine omnipresence. In so doing, my hope is to offer a new slant on understanding Kant’s ideality of space, as a response to both Newton’s metaphysics of space and to outstanding metaphysical problems about substances and their interaction (left open by Kant’s German and French predecessors).


michela massimi


Kant against the “Mathematical Investigators of Nature”

In the Transcendental Aesthetic, Metaphysical Exposition and Transcendental Exposition (A23/B38), Kant famously defends the apriority, necessity, and ideality of space. Space is said to be a “necessary representation, a priori, that is the ground of all outer intuitions.” More precisely, space is “the condition of the possibility of appearances, not as a determination dependent on them, as is an a priori representation that necessarily grounds outer appearances.” Hence, the “ideality of space in regard to things when they are considered in themselves through reason, i.e. without taking account of the constitution of our sensibility” (B44/A28). Space cannot be a property or a determination of appearances, needless to say of things in themselves, otherwise it could not be intuited prior to the objects of which it is a determination, and its apriority would be jeopardized. But its apriority cannot be jeopardized since “the receptivity of the subject to be affected by objects necessarily precedes all intuitions of these objects”; hence, the form of all appearances must be given to the mind prior to all actual perception, and it must ground all outer intuitions. Authoritative readings of the Transcendental Aesthetic have focused on both the receptivity thesis and the related thesis of the unknowability of things in themselves behind Kant’s defense of idealism about space.1 In what follows, I concentrate on a different – and strangely overlooked – argument for the ideality of space, what I call the argument from Spinozism (to be found in the second Critique). In so doing, I hope to show another route to Kant’s idealism about space, a route that goes through Kant’s departure from and mature reinterpretation of Newton’s absolute space. In the Transcendental Aesthetic, the Newtonians (described as “the mathematical investigators of nature”) are praised for succeeding in making mathematical knowledge of nature possible by “opening the field of appearances for mathematical assertions” (A40/B57). Yet, they make the mistake of assuming “two eternal and infinite self-subsisting non-entities (space and time), which exist (yet without there being anything real) only in order to comprehend everything real within themselves” (A39/B56). The 1

Strawson (1966) argued that subjects like us can, pace Kant, produce spatiotemporal representations by being affected by things in themselves (where affection incoherently presupposes that objects are located already in space and time). Allison (1983) has reinterpreted appearances as spatiotemporal entities (phenomena), i.e., things insofar as they are viewed as subject to the conditions of human sensibility, whereas things in themselves are nonspatiotemporal entities not subject to the conditions of human sensibility. And Langton (1998) has interpreted the receptivity thesis as implying that our knowledge is confined to spatiotemporal phenomena, understood as relational properties of substances whose intrinsic properties (qua things in themselves) remain unknown.

Kant on the Ideality of Space and the Argument from Spinozism


metaphysicians of nature [i.e., the Leibnizians–Wolffians], on the other hand, are accused of identifying space and time with relations of appearances abstracted from experience, at the cost of disputing “the apodeictic certainty of a priori mathematical doctrines in regard to real things (e.g., in space)” (ibid.). The metaphysicians of nature seem to fare, overall, worse than the mathematical investigators of nature; for the latter are at least in a position to secure the apodeictic certainty of mathematical knowledge, while the former cannot “bring the propositions of experience into necessary accord with those assertions.” While acknowledging the ability of the Newtonian conception of space to secure mathematical and geometrical knowledge, Kant distances himself from its metaphysical underpinning. In particular, Kant’s main qualm against the mathematical investigators is about positing two “eternal and infinite self-subsisting non-entities.” The locus classicus is, of course, Newton’s General Scholium to the Principia, where the Lord God Pantokrator is introduced as an “eternal, infinite, and absolutely perfect being” (Newton 1999: 940), who with His omnipresence and eternity constitutes absolute space and absolute time: He is eternal and infinite, omnipotent and omniscient, that is, he endures from eternity to eternity, and he is present from infinity to infinity; he rules all things, and he knows all things that happen or can happen. He is not eternity and infinity, but eternal and infinite; he is not duration and space, but he endures and is present. He endures always and is present everywhere, and by existing always and everywhere he constitutes duration and space. ( . . . ) He is omnipresent not only virtually but also substantially; for action requires substance ( . . . ). In him all things are contained and move, but he does not act on them nor they on him. (Newton 1999: 940, emphasis added)

Even more clearly in the unpublished De Gravitatione (probably written a couple of years before the Principia, in 1685), Newton defined space as “an affection of a being just as a being. No being exists or can exist which is not related to space in some way. God is everywhere, created minds are somewhere, and body is in the space that it occupies; and whatever is neither everywhere nor anywhere does not exist. And hence it follows that space is an emanative effect of the first existing being, for if any being whatsoever is posited, space is posited. ( . . . ) So the quantity of the existence of God is eternal in relation to duration and infinite in relation to the space in which he is present” (Newton 2004: 25, emphasis added). By declaring space to be an “emanative effect of the first existing being,” and by making God constitute space and time “by existing always and everywhere,” Newton equated the omnipresence of God with absolute space in whom “we live, move, and have our being,” to echo St Paul (Acts 17:28).


michela massimi

To be clear, by defining space as “an emanative effect” of God, Newton was at pains to clarify that space is neither a substance nor an accident. It is not a substance, since “it is not among the proper affections that denote substance, namely actions, such as thoughts in the mind and motions in body. For although philosophers do not define substance as an entity that can act upon things, yet everyone tacitly understands this of substances, as follows from the fact that they would readily allow extension to be substance in the manner of body if only it were capable of motion and of sharing in the actions of body” (Newton 2004: 21, emphasis added). Thus, absolute space is not a substance as it is not capable of acting upon things, i.e., it is not capable of motion in and of itself, by contrast with bodily substances. Nor is it an accident inhering in the subject either “since we can clearly conceive extension existing without any subject, as when we imagine spaces outside the world or places empty of any body, whatsoever, and we believe [extension] to exist wherever we imagine there are no bodies” (Newton 2004: 22). We can then understand Kant’s qualm regarding the mathematical investigators of nature assuming “two eternal and infinite self-subsisting nonentities”: space (and time) are non-entities, because they are not substances capable of actions in and of themselves; yet they are not accidents either, because they are self-subsisting, infinite, and eternal to contain everything else that exists and moves. The relation between God and Newton’s absolute space has been at the center of an important literature among Newton’s scholars, which I cannot enter into here (see Janiak and Schliesser [2012] for the latest additions to this debate). It suffices to say that to the eyes of many of his generation, Newton’s view appeared dangerously close to a form of Spinozism, as Eric Schliesser (2013) has recently documented by focusing on the correspondence between Newton and Bentley and the changes to the second edition of Newton’s Principia. Absolute space as a receptacle of God led to the perilous association with Spinoza’s God, encompassing with His substance everything that exists. What matters for my purposes here is that the charge of Spinozism against Newton was (rightly or wrongly) present in the cultural milieu in which Kant worked, and Kant seems to have been familiar with it, as we find him raising this charge almost verbatim against Newton in a famous argument for the ideality of space, to which I now turn.

4.3 The Ideality of Space and Kant’s Argument from Spinozism Kant seems to have had more worrisome reasons against the mathematical investigators of nature than those expounded in the Transcendental

Kant on the Ideality of Space and the Argument from Spinozism


Aesthetic. These reasons can be found in a passage from the Critique of Practical Reason in which Kant discusses the relation between freedom and natural necessity, where the latter concerns the existence of things only insofar as they are determinable in time qua appearances, while the former concerns their causality as things in themselves.2 The distinction between freedom (key for moral laws) and natural necessity (which underlies what Kant calls the mechanism of nature in accordance with the natural law of causality) leads Kant to a general reflection about human actions as being both free with respect to moral laws, and mechanically conditioned with respect to the law of causality. It is in this context that we find what I call the argument from Spinozism: as soon as one admits that God as universal original being is the cause also of the existence of substance (a proposition that can never be given up without also giving up the concept of God as the being of all beings and with it his all-sufficiency, on which everything in theology depends), one must admit that a human being’s actions have their determining ground in something altogether beyond his control, namely in the causality of a supreme being which is distinct from him and upon which his own existence and the entire determination of his causality absolutely depend. In fact, if a human being’s actions insofar as they belong to his determinations in time were not merely determinations of him as appearance but as a thing in itself, freedom could not be saved. A human being would be a marionette or an automaton, like Vaucanson’s, built and wound up by the supreme artist ( . . . ). Therefore I do not see how those who insist on regarding time and space as determinations belonging to the existence of things in themselves would avoid fatalism of actions ( . . . ). On the other hand, it is quite easy for us to distinguish between the determination of the divine existence as independent of all temporal conditions and that of a being of the sensible world, the distinction being that between the existence of a being in itself and that of a thing in appearance. Hence, if this ideality of time and space is not adopted, nothing remains but Spinozism, in which space and time are essential determinations of the original being itself, while the things dependent upon it (ourselves, therefore, included) are not substances but merely accidents inhering in it. ( . . . ) Of such great importance is the separation of time (as well as space) from the existence of things in themselves that was accomplished in the Critique of pure speculative reason. (CPrR 5:100–103)

Kant’s argument from Spinozism surprisingly chimes with some of Newton’s aforementioned remarks. If God is the cause of the existence of substance, and if we regard ourselves as substances (or things in themselves), 2

This passage has been analyzed by Brewer and Watkins (2012) with a particular focus on the threat of theological determinism, and its relation to both Leibniz and Spinoza. In what follows, I concentrate on this passage from the perspective of Kant’s defense of ideality about space.


michela massimi

then it turns out that God is also the cause, or the determining ground of our human actions. But if so, freedom would be jeopardized and fatalism of action would follow. This undesirable conclusion can be avoided via a two-step maneuver: (i) By drawing a distinction between the divine existence as the existence of a being in itself and our existence as things in appearance; and (ii) By reallocating space and time from “essential determinations of the original being itself,” to us and our outer sense. Note that step (i) per se is not sufficient to rule out fatalism of action. Step (ii) is also crucially needed. Indeed, it is possible to conceive that even under the assumption of a distinction already in place between God as a thing in itself and us as things in appearance, if (ii) was not in place (i.e., if space remained an “essential determination of the original being itself,” as with Newton’s absolute space for example), any alteration of place that we may perform with our actions would still have its determining ground in something altogether beyond our control, i.e., it would take place in absolute space (as a determination of God), and hence would be dependent upon God’s own substance.3 Hence, Kant’s conclusion that the ideality of space and time is the best antidote against the Spinozistic danger lurking in the view that takes space (and time) as “essential determinations of God.” Interestingly, no explicit mention is here made of Newton’s absolute space (and in the longer passage which I have omitted in the quote above, reference is also made to Mendelssohn’s view, which I will not discuss here). But the overall discussion leaves little doubt that it is to a broadly Newtonian view that Kant is referring in relation to space (and time) as “essential determinations of the original being itself.” Only by distinguishing between appearances and things in themselves and by reallocating space (and time) to determinations of things as appearances, rather than essential determinations of the original being itself (God), can the charge of Spinozism be averted. Three comments are in order. First, this passage shows how the ideality of space (and time, which I will not discuss here) is not just a consequence of the unknowability of things in themselves. The argument from 3

It is worth noting here again that what Kant reports as fatalism of action following from God qua the ultimate ground of alteration of space (if understood as Newton’s absolute space) seems to be a report on what he took to be probably general concerns in the cultural milieu of the time about Newton’s view (with its perceived Spinozistic flavor). For Kant’s own considerate view on fatalism would require more than the assumption that our actions are grounded on God qua absolute space: instead, our own free will and choice (prior to outer action) would also have to be constrained or determined. Obviously, the problem of free will and moral choice does not feature in Newton’s view of space. I thank Karl Ameriks for drawing my attention to this point.

Kant on the Ideality of Space and the Argument from Spinozism


the unknowability of things in themselves needs be supplemented. We must reallocate space from an essential determination of God to us and our outer sense, because the mere distinction between appearances and things in themselves by itself cannot eschew the danger of fatalism of action. If space remained an essential determination of God (as with Newton’s absolute space), freedom would be jeopardized. Second, some of the premises in Kant’s argument from Spinozism have a Newtonian-sounding origin. Kant seems to be reacting against the Newtonian view well-captured by the expression that “space is an affection of a being just as a being.” If space is a determination of substances in general, then it follows that space is also, first and foremost, an emanative effect of the first existing being; or, as Kant puts it, an “essential determination of the original being itself” (God). Third, Newton’s view that space is an affection of a being just as a being; and that space is, first and foremost, an emanative effect of the first existing being – jointly – do not license any Spinozistic-flavored conclusion that everything that moves in space (including ourselves) is an accident inhering in God’s substance. For Kant’s argument from Spinozism to follow from Newton’s view of space, a further premise is required. Namely, that God’s omnipresence qua absolute space grounds, in the sense of being the determining ground for motions of bodies (including our own bodily actions).4 Only if we take God’s omnipresence as the determining ground for any alteration of place and motion, does the argument from Spinozism follow. But it is far from clear that this further premise can be found in Newton; or is in fact compatible at all with a Newtonian view about space. On the contrary, the fact that we can conceive of absolute space as bereft of matter and independently of bodies, clearly suggests that for Newton, absolute space is not the determining ground for motions, in the strong sense required for Kant’s argument from Spinozism to go through. Moreover, Newton’s aforementioned comment in De Gravitatione about space not being a substance (because not being capable of acting upon things) reveals once more Newton’s considered view on the matter. Thus, for Kant’s argument from Spinozism to go through, Kant needed to surreptitiously assume that God’s omnipresence (qua absolute space) grounded alterations of place. Kant’s overall argument from Spinozism can then be summarized as follows: 4

For the ambiguities surrounding the term “determining ground” in Kant’s moral philosophy, see Ameriks (2012).


michela massimi

(1) Suppose that space is an essential determination or property of substances. (2) A fortiori, space must be an essential determination of God, qua substance. (3) But God is not just a substance among substances. God is the cause of the existence of substance. (4) Thus, the spatial determination of God as omnipresence (e.g., Newton’s absolute space) is the ultimate determining ground for any alteration of place. (5) But then the spatial determination of God as omnipresence (e.g., Newton’s absolute space) is also the determining ground for human actions (after all, in God, ‘we live, move and are’). (6) The spatial determination of God as omnipresence blurs the distinction between causality according to natural laws and causality according to moral laws, i.e., between natural necessity and freedom. (7) Fatalism of action, or Spinozism, follows. (8) To avoid Spinozism, we must deny premise (2), i.e., that space is an essential determination of God, and embrace idealism about space instead. QED. Prima facie, Kant’s argument is puzzling in more than one way. Denying premise (3) in the argument above is not possible because, as Kant himself concedes, denying that God is the cause of the existence of substance would be tantamount to denying “the concept of God ( . . . ), on which everything in theology depends.” But accepting premise (3) seems, on the other hand, to land us on a slippery slope, where the causality of the supreme being becomes the determining ground for the causality we encounter in nature (including in our own human actions). Natural necessity according to laws of nature would reduce to a mere accident inhering in God’s own causality; and freedom, according to the moral law, would be jeopardized. These undesirable conclusions can only be averted – Kant claims – if we assume that our human actions (and any sequence of events in nature) unfold into space and time, not qua essential determinations of the original being (i.e., Newton’s absolute space and time) but qua forms of our sensibility. Thus, Kant’s defense of idealism about space in the argument from Spinozism is centrally tied to the causal role that God plays in the world, and space understood as an “essential determination of the original being.” While the Newtonian tradition had influentially portrayed God’s role in the natural world as mediated by absolute space – as the receptacle (organon) of God’s omnipresence in nature – Kant seems to warn us that going down this path would lead us to dangerous fatalism of action.

Kant on the Ideality of Space and the Argument from Spinozism


In the next section, I argue that it is unsurprising that some of the key premises in Kant’s argument cannot in fact be found in Newton. For the argument from Spinozism, despite the Newtonian-sounding premises, should be read and understood against the backdrop of a different metaphysical tradition that engaged with the problem of the world as a totality of substances in interaction (). If my analysis is correct, Kant’s argument from Spinozism hides more profound metaphysical reasons for endorsing idealism about space, reasons that have less to do with Newton’s own view, and more to do with the influential views of Malebranche, Leibniz, and Baumgarten.

4.4 Kant on God, Space, and the World as a Real Connection of Substances If my interpretive take in the previous section is correct, we have identified in some of the premises of Kant’s argument the culprit for his charge of Spinozism leveled against a broadly Newtonian view of space. Premise (4) in particular (i.e., that the spatial determination of God as omnipresence is the determining ground for any alteration of place) seems to bear the burden of the proof. Such premise, I have already noted, cannot be found in Newton. How can absolute space be a determining ground for any alteration of place, including our own bodily actions? And why should absolute space qua determining ground for alterations of place follow from premise (3), i.e., that God is the cause of the existence of substance? That God is the cause of the existence of substance is a noncontroversial claim with a long history in natural theology. So Kant’s argument from Spinozism cries out for an explanation of how premise (4) can legitimately follow from premise (3). In this section, I take some steps toward answering this question. First, I argue that the rationale for the troublesome step from premise (3) to premise (4) in the argument from Spinozism should be looked for in the metaphysical tradition that goes from Malebranche to Leibniz, from Crusius to Baumgarten. This metaphysical tradition provided the backdrop for Kant’s analysis of God as the cause of the world intended as a connection () of substances in interaction ().5 The puzzling step in the argument from God as the cause of the existence of substance to God as the determining ground of human actions can, in my view, be illuminated if we consider the way Kant understood the world as a totality of substances, and the way he provided his own answer to open problems 5

For a similar analysis of how Baumgarten’s treatment of Spinozistic fate and chance influenced Kant (both in the lectures on metaphysics and in the first Critique), see Watkins (2000).


michela massimi

left by his predecessors about the relation between God and the world, so understood. Second, as a result of Kant’s own solution to these problems (in terms of real connection among substances, I argue), Kant came to reinterpret absolute space in idealistic terms (whereby absolute space becomes as early as the 1770s the “phenomenon of the divine omnipresence,” Metaphysik L1 28:214, LM 36). Hence, I conclude that Kant’s real motivation for the ideality of space – in the argument from Spinozism – should not be sought in Kant’s engagement with a broadly Newtonian view, after all. His real motivation is instead downstream to a wider metaphysical view about God, space, and the world that Kant came to elaborate in the 1770s, in response to both occasionalism and Leibnizian pre-established harmony. If my analysis proves correct, the threat of Spinozism is a red-herring. Or better, it is a much later spin that Kant gave to his own idealistic solution to the metaphysical problem about God and the world, and not the real start-up problem for idealism about space. Although arguments against Spinozism can be found in Baumgarten’s Metaphysics, and again in Kant’s lectures on metaphysics, a closer reading reveals that idealism about space as the best antidote against Spinozism is a gloss (even a gimmick, one may say) that Kant gave to fully worked-out and independently motivated arguments for the ideality of space. Or so, I shall argue. Central to Kant’s defense of idealism is instead a reinterpretation of space as a phenomenon of God’s omnipresence in the world, as opposed to being a determining ground for the interaction () among substances.

4.5 Kant at the Critical Turn (Mid 1770s to 1783) on God and the World: Interaction among Substances, Real Influence, and Space as a Phenomenon of God’s Omnipresence Alexander Baumgarten’s Metaphysics was one of the most influential texts of the time, a text that Kant repeatedly used for his lectures on Metaphysics. In Part II, Cosmology, Section II “The negative concept of the world,” Baumgarten distinguished between fate (more precisely, Spinozistic fate) and chance. The former is the absolute necessity of events in the world; the latter is the unknowability of sufficient grounds for events that occur in the world (Baumgarten 2013 [1739/1757]: 171, §382–83). This distinction proves functional to Baumgarten’s discussion about the relation between world and God, whereby the world is said to be “neither an

Kant on the Ideality of Space and the Argument from Spinozism


infinite substance nor an internal determination of an infinite substance, and hence the world is not the essence, attribute, mode or modification of an infinite being. Hence every world is to be posited apart from the infinite substance, so this world also exists apart from the infinite being, which for this reason is called an EXTRAMUNDANE BEING, a being that is actual apart from this world” (Baumgarten 2013 [1739/1757]: 172, §388). And while “THEOLOGICAL SPINOZISM is the doctrine denying that God is an extramundane being and it is an error” (ibid., 291, §855), Baumgarten hinted also at the association between theological Spinozism and Newtonian absolute space by declaring that “in God there are no simultaneous things posited mutually outside one another, no parts, hence no space. Therefore, God is neither extended, nor does he fill up space in the sense that extended things are said to fill it up” (ibid., 288, § 841). Baumgarten’s influential characterization of God as an extramundane being posed, however, a pressing metaphysical problem, which the young Kant grappled with. What is the relation between God qua extramundane being, and the world? How could God as a first cause that does not however inhabit the world (extramundane being) be present in the world itself? The obvious advantage of the Newtonian system was to avoid this metaphysical question altogether, by offering absolute space and time as the receptacle (organon) of God, in which everything that exists, moves and is (unappealing as it was to think of God as ‘filling up’ space). In the German tradition that goes from Leibniz, to Crusius and Baumgarten, the problem remained wide open (for more details, please see Watkins 2006) and was the subject of lively debates that influenced the pre-Critical Kant. In New Elucidation (1755), Kant tackled this problem in a novel way. For he offered an alternative to the Newtonian system, whereby it was the action and reaction of substances on one other, their ability to act upon other substances that was said to constitute space. Kant went as far as identifying the connection of substances, by virtue of which they were said to determine space, with Newton’s gravitational attraction.6 A year later in Physical Monadology (1756) Kant clarified how space “is entirely free from substantiality and ( . . . ) is the appearance of the external relations of unitary 6

“If the external appearance of this universal action and reaction throughout the whole realm of space in which bodies stand in relation to one another consists in their reciprocally drawing closer together, it is called attraction. Since it is brought about by co-presence alone, it reaches to all distances whatever, and is Newtonian attraction or universal gravity. It is accordingly probable that this attraction is brought about by the same connection of substances, by virtue of which they determine space. It is also probable that it is the most fundamental law of nature governing matter, remaining constantly in force only in virtue of God’s immediately sustaining it, according to the opinion itself of those who declare themselves to be followers of Newton” (NE 1:415).


michela massimi

monads” (PM 1:479; for an analysis, see Massimi, 2017). Kant modeled the action and reaction among substances – a vexed issue in the German metaphysical tradition – on the natural sciences: i.e., by thinking of substances as physical monads acting and reacting on one another in virtue of fundamental forces of attraction and repulsion. This model allowed the young Kant to think of space as the appearance of external relations among physical monads (pace Newton) and to think about the world as a totality of substances in interaction via fundamental forces acting as efficient causes or determining grounds for a plurality of effects (e.g., attraction causes changes of motion; repulsion causes clouds, the maintenance of fire, among many others). As early as 1755–56, Kant framed the problem of the world (i.e., the problem of explaining the world as interaction among substances) on the model of the natural sciences, whereby substances were understood as physical monads and dynamical forces (as determining grounds) were said to necessarily cause effects in nature. What about God? In The Only Possible Argument (1763) ‘Second Reflection: Differentiation of the dependency of all things upon God into moral and non-moral dependency,’ Kant gave a splendid distinction between moral and nonmoral dependency of things upon God, whereby moral dependency is dependency through the will of God, while nonmoral dependency is dependency without the will of God. Thus, Kant claimed, when we say that God is the ground of the existence of things, this dependency is always moral: ‘in other words, things exist because God willed that they should exist.’ But when we say that God is the ultimate ground of the internal possibility of things, we do not mean this as a moral dependency: the internal possibility of things (with all fruitful and serendipitous consequences we observe in nature) does not depend on the will of God. These remarks already point in the direction of a helpful distinction between God as the cause of the existence of substances (through His will) vs. God as determining ground for the mechanism of nature (which Kant seems to reject as early as 1755–63). Yet in 1763 Kant had not yet worked out the exact relation between God (as the ultimate ground or cause of the world), and the world itself (as a totality of substances in interaction and following an order according to natural laws, and determining ground-effects causal relations). If God is an ens extramundanus (as Baumgarten claimed against theological Spinozism) and not part of any causal chain in nature (despite being the first cause of everything that exists), how can He also be present in the world? If Newton’s absolute space was not a live option for Kant to think about God and the world, what else could be?

Kant on the Ideality of Space and the Argument from Spinozism


In what follows, I show how in the mid 1770s Kant came to rethink Newton’s absolute space along idealistic lines (i.e., as a phenomenon of the divine omnipresence) precisely in the attempt to answer the problem of the world (i.e., the problem of understanding the relation between God and the world as a totality of substances). Some of the most interesting passages on this issue can be found in Metaphysik L1 , containing students’ notes from Kant’s lectures on Metaphysics in the mid 1770s (which featured Baumgarten’s Metaphysics as textbook). Latching onto Baumgarten’s discussion about the world as a totality of substances and God as ens extramundanus, Kant (Metaphysik L1 28:212–13, LM 34) marked an important distinction between the mutual interaction among substances and God as an absolutely necessary being, who cannot stand in interaction with substances because He subsists in and of Himself, without any need for interacting with other substances. Thus, necessary beings are isolated and cannot be in space, Kant now claims, because “to exist in space already means: to be in community; for space is a phenomenon of the general connection of the world and we want to have precisely the ground of this connection through space” (Metaphysik L1 28:213, LM 34, emphasis added). Once again, pace Newton, God cannot exist in space because existing in space means already being in community with other substances, and God is not a substance among substances. Nor is the interaction among substances themselves (qua necessary beings) possible through space either (again pace Newton) because, as we just noted, Kant claimed that the very notion of necessary being implies that they are isolated and not in interaction. Thus, how can substances (qua necessary beings) be, after all, in interaction for the world (as a totality of substances) to be possible? And how can God act as the ultimate ground for the world, while also not entering into any interaction with it? Here is Kant’s solution to this conundrum: Interaction () is thus possible not through space, but rather only through this, that they all are through One and depend on One; for otherwise those that depend on another would not stand in interaction with each other. Every world thus presupposes a primordial being, for no interaction is possible except insofar they are all there through One. As phenomenon, space is the infinite connection of substances with each other. Through the understanding, we comprehend only their connection, to the extent they all lie in the divine. ( . . . ) If we imagine this connection sensibly, then it happens through space. Thus, space is the highest condition of the possibility of the connection. Now if we sensibly represent the connection of substances, which consists in this, that God


michela massimi is present to all things, then we can say: space is the phenomenon of the divine [omni]L presence. (Metaphysik L1 28:214, LM 35–36)

Kant’s departure from Newton is complete; and it acquires new light if read against the backdrop of Kant’s engagement with the metaphysical problem of the world he inherited from Baumgarten. Substances qua necessary beings cannot be connected through space, for space presupposes community and mutual interaction and it is against the very concept of a necessary being to be in community. Thus, substances can only be connected through God as the One ultimate ground. While we can think of the connection among substances through God in our understanding, we can also represent such a connection sensibly via space. Thus, while not doing any genuine metaphysical work of connecting substances (for the reasons just explained), space is nonetheless the way in which we come to represent to our senses the presence of God in all things. Far from being the determining ground for the connection of substances, space (or better, what used to be Newton’s absolute space) is now the phenomenon of the divine omnipresence. Space is not God’s receptacle, through which we live, move, and are. Instead, space is the “highest condition of the possibility of the connection.” Space makes possible for us to represent things as being connected (and hence makes possible for us to know the world as a totality). But space is not itself the determining ground for such a connection. What is most interesting in this story are some corollary discussions about how to think of substances as being connected through God (but not through space). Along lines already to be found in New Elucidation, Kant argues that sheer co-presence or co-existence of substances () does not constitute per se interaction, pace Crusius’s physical influx. Instead, Kant continued, the commercium among substances that constitutes the world requires a ‘third ground,’ and is then called derivative interaction (). Derivative interaction comes in two varieties: via physical influence (but not in the “crude sense” of physical influx, Kant hastened to add), or via hyperphysical influence. Both are derivative influences according to laws of nature, but the difference is that hyperphysical influence refers to laws of nature that are “posited by another being” (28:213, LM 35). Or, more precisely, it is influence “according to the universal determinations of the extramundane being” (28:214, LM 36). And to illustrate hyperphysical influence, Kant gives an example of fatalism of action, as when “a third being moves my foot when I want to move it.” This can happen in two different ways.

Kant on the Ideality of Space and the Argument from Spinozism


Hyperphysical influence can be automatic () when “for every single case the highest cause has to arrange an agreement; thus where the agreement rests not on universal laws, but rather on a primordial arrangement which God put in the machine of the world” (28:215, LM 36). Kant identifies automatic harmony with Leibniz’s pre-established harmony as a harmony that is not generated through natural laws according to determining grounds–effects; but is instead instituted by God Himself via an original intervention in the machine of the world. Calling pre-established harmony “automatic harmony” is revealing, I think. The risk of reducing human actions to automata (as Kant will say a decade later in the argument from Spinozism) seems to be associated not so much with Spinoza in this passage of Metaphysik L1 but rather with the kind of derivative hyperphysical influence advocated by Leibniz. The second kind of hyperphysical influence is occasionalistic whenever “the ground is not arranged at the beginning such that at every occasion God accomplished the effect continuously with the continuation of the world” as in Malebranche and Descartes. Kant complains that both kinds of hyperphysical influence provide only an ideal connection () or interaction (), but not a genuine or real interaction among substances, because they posit the ground of the interaction among substances in a primordial being that acts to make possible such interaction (via either pre-established harmony or occasionalism). What is required for the concept of the world as a totality (), Kant continues, is that substances be in real connection (). This real connection has to be derivative and physical (but not in Crusius’s crude sense), while also being grounded ultimately “on the unity of the primordial being” (28:215, LM 37). These metaphysical reflections end here in Metaphysik L1 without going much further in terms of explicating how to reconcile the primordial being qua ultimate ground of the unity of nature with real connection () understood as a physical and derivative influence among substances. This dichotomy finds its final resolution in the Critical period, as becomes evident in corresponding passages of Metaphysik Mrongovius, dating to 1782–83. A year after the first Critique, Kant went back to the notions of vs. . Once again, he stressed how ideal connection is not real connection among substances or better what he now calls “things in themselves, but rather merely in the idea of the observer who considers them” (Metaphysik Mrongovius 29:866, LM 236), with the consequence that there is “no world but rather only an ideal whole in thoughts.” And again in this context, Kant mentions Descartes’s occasionalism and


michela massimi

Leibniz’s pre-established harmony as two examples of , the danger of which is this: “For God could also have allowed representations of body to come about in the soul, either as occasioned or as pre-established , without actual bodies being necessary. These representations could indeed always harmonise with the soul. ( . . . ) For since the souls effect nothing, without them God could still effect all alterations in bodies” (29:867, LM 237). The risk of reducing freedom of action to automata reappears in the same context of a criticism of Leibniz and Descartes’s system. As an alternative, and in defense of , Kant now gives his own mature solution: real connection is only possible in the noumenal world “if one assumes a common cause, i.e. God, which has already put that [real influence] in the nature of substance”; while in the phenomenal world, it is possible from the mere existence of substance in space: The concept of space accomplishes in the sensible world what the divine omnipresence does in the noumenal world , and one can therefore call it as it were a phenomenon of the divine omnipresence. Perhaps God wanted thereby to make his omnipresence sensibly cognizable to us. Newton called it the seat of the senses of the divine omnipresence. Perhaps space is also the only sensibility that belongs to all rational beings other than God. (Metaphysik Mrongovius 29:866, LM 236)

Kant’s final solution to the metaphysical problem of the world as a totality of substances can be found in Kant’s mature distinction between noumenal world and phenomenal world. Real connection among substances belongs to the former, by positing God as the common cause of the world. In the phenomenal world, Kant concludes, we no longer need to prove any real connection among substances “for it is nothing in itself. Here [in the phenomenal world] everything is interaction in virtue of space” (29: 868). The last vestige of Newton’s absolute space can be found in the phenomenal world where the “concept of space accomplishes what divine omnipresence does in the noumenal world.” Space does not ground any real connection among substances because it is only a phenomenon, it is a form of our sensibility though which we can sensibly represent the world as a totality. The unity and totality of the world can only be grounded in God qua the first cause in the noumenal world. Space, as a mere phenomenon, allows us to intuit the world as interaction . Kant’s idealistic turn is complete. God is not extended in space, He does not fill up space; nor does space – as a determination of God’s

Kant on the Ideality of Space and the Argument from Spinozism


omnipresence – ground any real connection among substances or our bodily actions. By distinguishing between noumenal world and phenomenal world and by relegating space to the latter and real connection among substances to the former, Kant could explain what escaped his predecessors. Namely, why the world forms a composite real of substances and how God, qua ens extramundanus, can ground the world (and be omnipresent in it) without yet entering into any real interaction with it. Kant’s idealism about space could at once solve the metaphysical problem that beset Baumgarten, Leibniz, Malebranche and Descartes, and do justice to God’s omnipresence in the world without the pitfalls of Newton’s absolute space (i.e., thinking of God as spatially extended). Fatalism of action results only if we conflate these two realms and take God as a ground for hyperphysical influence (as occasionalism and pre-established harmony did). Freedom is jeopardized only when we take appearances for things in themselves, i.e., if we take God’s omnipresence in space not as a phenomenon but as a thing in itself (as with Newton’s absolute space). What about Spinoza? After all, the argument that has concerned us in this essay is named after him. Interestingly, Spinozism does not enter into Kant’s metaphysical reflections about space until much later. After the reference to Spinozism that we found in the second Critique, Kant returns to the threat of Spinozism almost verbatim in Metaphysik L2 (1790–91), where space and time are said to be “not things themselves, not properties, not a constitution of things, but rather the form of sensibility. ( . . . ) If I assume space to be a being in itself, then Spinozism is irrefutable, i.e. the parts of the world are the parts of the divinity. Space is the divinity; it is united, all-present; nothing can be thought outside of it; everything is in it. ( . . . ) Space occurs only with things, as appearances. Appearances teach us nothing as to how the things are, but rather how they affect our senses” (Metaphysik L2 28:567, LM 331). Even more interestingly, in Metaphysik Vigilantius (K3 , 1794–95), Kant makes explicit the link between the threat of Spinozism and his earlier reflections about space as a phenomenon of the divine omnipresence. Once again, he comes back to Newton’s incorrect idea of space as the instrument of the divine omnipresence and the whole discussion is cast once more in terms of substances in the world having a reciprocal influence on each other and standing in real connection via God as a communal cause (Metaphysik Vigilantius (K3 29:1007, LM 476). Kant rehearses the by now familiar arguments against both Crusius’s physical influence as of substances (which cannot be in interaction in virtue of


michela massimi

their simple existence) and also against Leibniz’s pre-established harmony as a derivative hyperphysical influence which can only deliver the world as an ideal whole . The distinction between phenomenal and noumenal world features once again, with the phenomenal world providing space as that “which connects the substances, through which they are in interaction ,” while the interaction of substances in the noumenal world requires God as a common cause: If I assume all substances as absolutely necessary, then they cannot stand in the slightest community. But if I assume the substances as existing in community, then I assume they all exist through a causality, for only through that can their community be explained. – Space itself is the form of the divine omnipresence, i.e. the omnipresence of God is expressed in the form of a phenomenon, and through this omnipresence of God all substances are in harmony. But here our reason can comprehend nothing more. – Those who assume space as a matter in itself or as a constitution of things in themselves, are required to be Spinozists, i.e. they assume the world to be a summation of the determinations of a united necessary substance, thus only one substance. Space as something necessary would then also be a property of God, and all things exist in space, thus in God. (Metaphysik Vigilantius K3 29:1009, LM 478)

The charge of Spinozism could not be clearer. But, as I hope to have shown in this chapter, the motivation and profound reasons for such a charge should not be looked for in Newton’s own view, or the lingering charge of Spinozism against Newton’s view at the time. Rather, they should be looked for in the way Kant engaged and offered a solution to a long-standing metaphysical problem about the world, as a whole of substances in real connection. Kant’s mature solution relied on the distinction between phenomenal world and noumenal world, whereby space was relegated to the former qua phenomenon, and God as a first cause to the latter qua noumenon. The charge of Spinozism – I claimed – is a later spin Kant gave to his argument for the ideality of space. Those, who blur the distinction between phenomenal world and noumenal world, and take space not as a phenomenon but as a noumenon, are bound to face Spinozism.

c h a p ter 5

How Precise Is Kant’s Table of Judgments? Michael Wolff



Kant’s Table of Judgments is as fundamental to his Critical philosophy as it is obscure and questionable: how does Kant show that his Table with four groups or “Titles,” each comprising three forms of judgment, has neither too few nor too many forms? How does Kant show that his Table is complete? How does Kant’s Table fare when confronted with post-Fregean developments? This brief essay addresses the first question.1 There are good indications that the First Section of Kant’s “Transcendental Clue to the Discovery of all Pure Concepts of the Understanding” (A67–69/B92–94) reworks a sketch related to Kant’s treatment of that topic in his 1770 Dissertation (Part II, §5), which characterizes the “logical use of the intellect” as consisting in comparing and subordinating concepts (of whatever origin). Kant sketched this argument soon after mentioning to Herz his “Critik der reinen Vernunft,” which Kant expected to complete because he had reduced all the concepts belonging to completely pure reason . . . to a certain number of categories, . . . arranged . . . according to the way they classify themselves by their own nature, following a few fundamental laws of the understanding. (Letter to Herz, February 21, 1772; Corr. 10:132)

Here Kant first indicates his aim to derive the completeness of the table of categories from an a priori division according to principles of the understanding. In the Critique, Kant’s “Transcendental Clue” is enthymematic and yet replete with terms marking inferences and conclusions. Here I examine Kant’s basic assumptions underlying the completeness of his Table of Judgments (§5.2), the steps made by the First Section of Kant’s 1

The further questions are addressed in Wolff (1995, 2009a, 2010a, 2010b, forthcoming). This translation is by K. R. Westphal, in consultation with the author. This essay presents a new analysis, published here for the first time. (Translations from Kant follow the Cambridge Edition.)



m i c h a e l wo l f f

“Transcendental Clue” to demonstrating the precision of this table (§5.3) and the further steps in that demonstration taken in the Critique (§5.4). These considerations show that Kant’s Table has precisely the correct number of judgment forms: twelve.


The Logical Use of the Understanding in General

The First Section of Kant’s “Transcendental Clue” takes the first steps toward discovering all pure concepts of the understanding (A67–69/B92– 94). Because these concepts, like all concepts, rest upon acts of the understanding, discovering all such pure concepts requires enumerating completely all acts of the understanding. No complete enumeration can be empirically grounded; it requires instead a “concept” or “idea” of “the understanding,” from which these concepts “spring pure and unmixed” (A67/B92). Only if we know what kind of capacity our understanding is, and why we require its acts, can we identify which functions it performs, and what are their corresponding acts. Hence Kant’s single most important contention in the First Section is that all acts of the understanding are rooted in judging, so that the “understanding in general” – i.e., the “faculty for thinking” – can be represented as “a faculty for judging” (A69/B94). This same idea of the understanding as the “faculty for judging” Kant describes in the Third Section (in B: §10) of his transcendental “Clue” as the “common principle” which “systematically . . . generates” the 4-part “division” of the table of categories (A80–81/B106). This division presupposes the same kind of division as the four Titles of the Table of Judgments; both 4-part divisions rest in the same way on Kant’s idea that the “understanding in general” is the capacity to judge. Kant’s introduction to this chapter alludes to this idea as the “principle” according to which one must “seek” all the pure concepts of the understanding (A67/B92; cf. Prol. §39, 4:323.22–31).2 As pure concepts of the understanding, they must somehow relate to each other according to this idea, which provides a “rule by means of which the place of each pure concept of the understanding and the completeness of them together can be determined a priori” (A67/B92). Apparently Kant regards this “rule” as a complete disjunction by which the concept of judgment, qua act or function of the understanding, can be divided so that, to each member of the division, some functions of the understanding can be subordinated, thus constituting a complete system of functions corresponding to pure concepts of the understanding. 2

References such as “4:323.22–31” are to the line numbers 22–31 of page 323 (here, in volume four of the Akademie edition of Kant’s works, found also in the margins of the Cambridge Edition).

How Precise Is Kant’s Table of Judgments?


Kant’s claim in the First Section, that the understanding as a capacity to think is the same as the capacity to judge, provides the basis for the rule of Kant’s divisions; I shall call it the ‘Principle of Division’ or PD. This Principle (PD) first concerns Kant’s Table of Judgments, as the table of “logical functions in all possible judgments” (A79/B105), although Kant’s ultimate aim is the division of the table of categories. Kant’s PD is no mere presumption; the First Section is to justify it. Justifying PD likewise justifies Kant’s claim that all acts of the understanding are rooted in judging. Kant’s title to the First Section indicates that justifying PD requires examining “the logical use of understanding in general.” What this means, Kant does not explain in the Critique; apparently the reader should recall from Kant’s Dissertation that, by the specifically “logical” use of the understanding (usus intellectus), concepts, no matter whence they are given, are merely subordinated to each other, the lower, namely, to the higher (common characteristic marks [notis communibus]), and compared with one another in accordance with the principle of [non-]contradiction. (ID II, §5, 2:393, TP1 385)

Kant here claims that understanding is a capacity to use concepts, and that the logical use of concepts consists in their mutual subordination or comparison. Kant regards Aristotelian syllogistic as “general logic.” The syllogistic use of concepts in categorical, hypothetical, and disjunctive inferences can be explicated in terms of subordination and comparison of concepts. If β and γ are mutually exclusive concepts, both of which can be subordinated to a more general concept α, and if δ and ε are mutually exclusive concepts which can be subordinated to γ, then these five concepts form this hierarchy:

Figure 5.1

These relations among concepts can be conceived completely in terms of judgments. These judgments can form inferences (syllogisms), as diagramed in Figure 5.1, representing the universally valid rules of categorical, hypothetical, and disjunctive syllogisms,3 corresponding to categorical judgments in the traditional forms (A, E, I, and O) and to hypothetical and disjunctive forms of judgment; e.g.: 3

Cf. Plato, Sophist 254ff.; Kneale and Kneale (1975: 9–10).

86 (1) (2) (3) (4) (5) (6) (7) (8)

m i c h a e l wo l f f (A) each β is an α, (E) each γ is not a β, (I) some α is a γ, (O) some α is not a δ, if some α is an ε, then this α is a γ, if some α is a β, then this α is not a γ, either some α is a β, or this α is a γ, if some α is a γ, then this α is either a δ, or an ε, etc.

Similarly, Kant notes, “To judge is: to represent one concept as contained in another or as excluded from it” (Refl. 3045, 16:631, NF 59). According to Kant, an “adequate” or “complete” concept or cognition contains neither too many nor too few characteristics, and “fully completes” the relevant “series of characteristics” (Logik Dohna, 24:731, LL 467). According to Kant’s Dissertation, the logical use of understanding is “common to all sciences,” because regardless of its source, any cognition must be regarded either as contained under or as opposed to a characteristic mark common to several cognitions, and that either immediately and directly, as in the case of judgements, which lead to a distinct cognition, or mediately, as in the case of ratiocinations, which lead to a complete cognition. (ID II §5, 2:393, TP1 385)

Here Kant first identifies the logical use of understanding with the use of concepts (as an activity of the understanding qua “faculty of concepts”; A160/B199) and then links it to the higher activity, which Kant ascribes to the understanding in general, as the capacity to think. These activities, Kant indicates, consist first in judging, i.e., the immediate consideration of a given concept as either subordinated or opposed to another; and second in reasoning, i.e., the mediate consideration of such a concept in one such regard. Kant also distinguishes these two activities as serving distinct purposes or functions (common to all disciplines), regarding the distinctness or the adequacy of knowledge (cf. Refl. 3041, 3045; 16:629, 631; NF 59). The title to Kant’s First Section announces an inquiry into the use of the understanding described in the above passage; it begins by noting that understanding contributes to cognition by using concepts, and that concepts are based upon representations subordinate to some common representation (A68/B93). (Later Kant defines concepts as an “objective” representation or a “cognition” [cognitio] which relates to objects “by means

How Precise Is Kant’s Table of Judgments?


of a mark, which can be common to several things”; A320/B377.) Unlike the Dissertation, Kant’s Critique disregards disciplinary aims served by such hierarchies or classifications. However, the intrinsic, primary aim defining the logical use of understanding in all of its functions is to generate such hierarchies, classifications, or differentia. Kant expressly defines these functions as “the unity of the action [of the understanding] of ordering different representations under a common one” (A68/B93; 85.17–19).4 All acts of the understanding qua capacity to think trace back to this definition, according to which judgment can be explicated as the exercise of a function, a unitary act of ordering representations by subordinating them to some common representation. By this definition, one judgment may exercise several functions (just as the one activity of riding a bicycle unifies several subservient activities). Kant’s analysis of the logical use of understanding in the First Section aims to distinguish all the functions comprised within the function of judgment by specific aspects, to divide completely and to specify the functions of the understanding qua capacity to think. Accordingly, Kant presents the Table of Judgments as a “table of logical functions in all possible judgments” (Second Section; A79/B95) divided according to four aspects, where these functions are all designated “moments” contained within the “function of thinking” (A70/B95), insofar as these functions are exercised in a judgment; hence Kant designates these functions as “moments of thinking” (A71/B97) and as “moments of thinking in judgments” (A73/B98). The First Section does not consider these specific moments; instead it aims to describe the four aspects by which the functions of the understanding contained in judgments can be distinguished from each other and from judgment as such. This description affords the completeness of Kant’s quadruple division of these functions. According to this division, all acts of the understanding can be traced back to judgment; hence PD is the principle of this division. Kant reports its discovery in Prolegomena, §39.

5.3 Kant’s Quadruple Division of Logical Functions in Judgment To demonstrate that judgment is not merely one of three acts of the understanding, but that all acts of the understanding are rooted in judgment, Kant must show that: 4

All “page.line” references to CPR are to Ak. volume 3. Where the location in the A/B pagination has already been noted, the A/B reference will not be repeated in each case.


michael wolff

(1) Judgment consists in a unitary act which unifies the exercise of more than one function (in the sense defined); (2) All acts of the understanding consist in the exercise of such functions and so are rooted in judgment, because each function is exercised either in or through a judgment. Kant begins this proof (First Section, A68/B93) by claiming that the understanding “can make no other use” of concepts than “that of judging by means of them.” Kant argues that understanding is a capacity to know “through concepts.” Since knowledge is always knowledge of objects, knowledge through concepts is only possible by referring concepts to objects. Since concepts (unlike intuitions) cannot be referred immediately to objects, they can be referred to objects only mediately, by subordinating representations (whether intuitions or other concepts). Such a mediated knowledge of an object is a judgment about it, so that this knowledge is “the representation of a representation.” Consequently the understanding can make no other cognitive use of concepts than by “judging by means of them.” In “The False Subtlety” (1762, 2:58, TP1 102) Kant argued that a “distinct concept” is “only possible through a judgment.” The First Section speaks, not of “distinct concepts,” but of concepts as means for knowing objects. With this justification Kant implicitly indicates two distinct acts of the understanding: one is the use of a concept in judgment, the other is the judgment itself. Kant so describes these acts that they are clearly distinguishable. The use of a concept he describes consists in subordinating one representation (whether intuitive or conceptual) under another representation, which of itself (in contrast to judgment about an object) provides no knowledge of any object; this use serves only as a means to knowing objects. Because judgment is an act, which consists in judging, by which the subordination of a representation under a representation is used to achieve knowledge (which that judgment should express), judgment consists in exercising a function. In contrast to that, the function exercised by using a concept is exercised within a judgment. Consequently the concept of logical function within judgment can be divided as follows: Logical function in judging Judging

Use of a concept in judging Figure 5.2

How Precise Is Kant’s Table of Judgments?


In direct connection with justifying the thesis that the understanding can use concepts only by judging, Kant addresses the two theses mentioned above: (1) and (2). Kant first clarifies how, according to (1), judgment consists in exercising functions (A68/B93, 83.28–86.4), by using this sample judgment: (a) All bodies are divisible. (85.31) This judgment uses the concept “divisible” as a predicate, which stands for a general, common representation that subordinates (“comprehends”) many other representations, some of which are themselves concepts, though some are intuitions.5 Among the representations comprehended by “divisible” is the concept “body,” used in (a) as subject. To this subject concept other representations can be subordinated, including intuitions, which in (a) are immediately referred to objects specified as bodies by use of its subject concept. Accordingly, judgment (a) also refers the concept “divisible” to these objects, though only mediately, namely, via the concept “body” and the intuitions it can comprehend, i.e., those which can be subordinated to it. Kant directly generalizes this point (86.4–8); he claims that all judgments are “accordingly” (demnach; A69/B94) – just as judgment (a) – “functions of unity among our representations.” To elucidate this claim Kant identifies two tasks which can be accomplished by using a concept in a judgment. First (i), in any judgment, “instead of an immediate representation” (i.e., instead of an intuition), “a higher [representation], which comprehends this and other representations under itself, is used for the cognition of the object.” Second (ii), in any judgment “many possible cognitions are thereby” (i.e., by using a representation which is not an intuition, but instead a concept) “are drawn together into one” cognition (A69/B94). Kant’s first point (i) indicates that, in any cognitive judgment about an object, a concept is used which is nonempty, because it comprehends intuitions which refer immediately to that object, so that this concept, too, can be referred to that object, without mediation by any other concepts comprehended within that concept. Such a concept is the subject concept of the judgment. Its use specifies what is judged and to what its predicate does (or does not) pertain, if the judgment is true. Due to the reference to the object of the judgment established by this use of the subject concept, 5

Only sensory intuitions are relevant here; subsequently this qualifier is omitted.


m i c h a e l wo l f f

the subject concept so used provides the condition by which the predicate concept, too, is nonempty. Kant’s second point (ii) indicates that, in any judgment many possible cognitions are “drawn together” (zusammengezogen, integrated) in one cognition, insofar as the predicate concept (like the subject concept) comprehends many representations within it (or subordinates them under itself ), though in such a manner that the predicate concept is not referred to each of the representations it comprehends, but only to that specific subject concept. The unity of the cognition thus achieved consists in the unity of a judgment which, because it integrates many possible cognitions within itself (e.g., ‘this body is divisible,’ ‘that body is divisible,’ etc.), can pertain to more than one object and so can have a scope (Umfang; A71/B96, 87.22). The use of the predicate concept makes possible this unification because its reference to one nonpredicatively used concept mediately refers to more than one possible intuition comprehended under that concept, so that this concept may be referred to more than one object. By this explication, Kant’s claim that all judgments are “functions of unity among our representations” is tantamount to this: all judging consists in the exercise of functions (in the sense defined: A68/B93; 85.17–19), which themselves consist in various moments contained within an act of the understanding in judgment, which serve to unify various representations into the unity of a judgment, to afford knowledge by understanding. This claim prepares for Kant’s justification of the Principle of Division (PD), according to which judging is that “act of the understanding that contains all the rest and that differentiates itself only through various modifications or moments in order to bring the multiplicity of representation under the unity of thinking in general” (Prol §39, 4:323). Kant’s elucidation of this claim also provides justification for assumption (1), according to which judging is a unitary act integrating more than one function. According to Kant’s elucidation, the act of judging contains two distinct kinds of use of concepts; both consist in exercising a function. One of these: the subordination of intuitions under the subject concept, establishes the reference to the object judged, which is necessary for knowledge by understanding, without any further intermediation by other concepts used in that judgment. I designate this use of a concept: immediately referred to objects and nonpredicative. The other use of concepts consists in connecting the predicate concept of the judgment to its subject concept. This connection consists in subordinating intuitions, though mediately via the subject concept. Hence this connection can mediately comprehend as

How Precise Is Kant’s Table of Judgments?


many referential connections to objects as properly fall under (or are comprehended by) the subject concept. This second use of a concept in judgment I designate: predicative. These distinct uses of concepts in judgment afford this division: Use of a concept in judging Predicative

Non-predicative Figure 5.3

Excepting the last two sentences (A69/B94; 86.19–22), Kant devotes the remainder of the First Section to justifying assumption (2), according to which all acts of the understanding consist in exercising functions either in or through a judgment. Justifying this assumption also justifies PD; Kant introduces its formulation by suggesting that, if indeed we can “trace all actions of the understanding back to judgments,” then it follows directly, “that the understanding in general can be represented as a faculty for judging” (86.8–11). Kant’s speaking of all acts of the understanding suggests that to the understanding in general, as the capacity to think, must be ascribed, not only the acts of using concepts and of judging, but also of reasoning, not yet mentioned in the First Section. Later Kant states expressly that general logic, as “constructed on a plan that corresponds quite precisely with the division of the higher faculties of cognition” (i.e., “understanding, power of judgment, and reason”), “deals with concepts, judgments, and inferences, corresponding exactly to the functions and the order of those powers of the mind, which are comprehended under the broad designation of understanding in general” (A130–31/B169). The First Section considers “the logical use of the understanding in general”; hence one may expect the justification of assumption (2) to consider reasoning as a function of the understanding in general. Kant’s justification of assumption (2) again uses the sample judgment (a): all bodies are divisible (85.31). However, Kant now considers the subject concept ‘body’ in a new regard: namely, regarding everything to which the predicate ‘body’ pertains. Accordingly, in (a) a concept is used which also occupies the predicate position of a possible judgment (b) with the form ‘x is a body,’ but which (in contrast to (a)) does not leave undetermined what a divisible body is, but instead judges to what the predicate ‘body’ pertains. Kant’s sample judgment which determines this “still undetermined object” (86.13–14) x, is:


m i c h a e l wo l f f

(b) Each metal is a body. (A69/B94; 86.18–19) What does this extension of Kant’s example show? Kant claims that a concept “is a concept” only because “other representations are contained under it, by means of which it can be related to objects” (86.15–17); i.e., concepts are concepts only as “predicates of possible judgments” in which they are “related to some representation of a still undetermined object” (86.12–13). This holds, because for any concept another can be given as subordinate to it, just as according to the “principle of specification” (A659/B687) necessarily for any species subspecies are conceivable. Hence any possible subordination of one concept under another must be conceivable in the form of a possible judgment. Syllogistic inference is based on such ordering, as Kant indicates with the further extension of his example (construed according to Modus Barbara): (a) All bodies are divisible. (b) Each metal is a body. (c) Each metal is divisible. As the major premise of a categorical syllogism in the first figure,6 a judgment can perform its proper function only because the concept it uses nonpredicatively can also be the predicate of another possible judgment, in which it would be referred to an object which it does not determine. Thus from the premises (a) and (b), the conclusion (c) only follows because (a) already contains the thought: ‘x, for which holds: x is a body, is divisible.’7 Only because the subject concept in (a), by “mediation” (A69/B94; 86.17) of the representations contained “under it,” is referred to the “still undetermined object” x of the possible judgment (b), can (a) be the major premise of a valid syllogism. Apparently Kant’s elaboration of his example intends to show that there is also a third kind of use of a concept within judgment. This use is nonpredicative, but does not consist in referring the subject concept of the judgment to its object; hence it does not consist in referring the subject concept to an object without mediation by any concepts it comprehends. Instead, this use consists in referring the nonpredicatively used concept to 6


Regarding Kant’s justification of his view that categorical syllogisms only of the first figure are “pure,” since only they can be recognized to be valid regardless of whether “immediate” (nonsyllogistic) inferences (consequentiae immediatae; B141, §19n) are valid, see Wolff (2009b, 2010a, 2010b). Cf. Kant’s remarks on the Dictum de omni and the Dictum de nullo in “Die falsche Spitzfindigkeit” (“The False Subtlety of the Four Syllogistic Figures” (FS) §2, 2:49, TP1 91).

How Precise Is Kant’s Table of Judgments?


the still undetermined object of subsumable (possible) judgments which contain this concept as a predicative concept. This reference is thus mediated by other concepts, namely, by further concepts used in those subsumable judgments. This third use of a concept also serves to combine various representations within the unity of a judgment. Distinctive of this third use, however, is that it not only affords knowledge by understanding in the narrow sense (i.e., knowledge by use of concepts and judgments), but also affords knowledge by understanding in the sense of knowledge by reason (i.e., knowledge by inferences). (Only later can I consider which role is served by this third kind of use of concepts in judgments, such that it can provide major premises of noncategorical syllogistic inferences.) According to this explication, the nonpredicative use of a concept in judgment affords the following division: Non-predicative use of a concept in judging Immediately referred to an object

Not immediately referred to an object Figure 5.4

Kant himself never states explicitly the dichotomous divisions presented in Figures 5.1–5.4, though the First Section strongly suggests that these divisions are exactly those by which he sought to demonstrate that all acts of the understanding can be traced back “to judgments” and that the “understanding in general” is “a capacity to judge.” These divisions clearly provide a complete classification of acts of the understanding. If indeed the logical use of understanding in general consists in subordinating concepts under other concepts, then: (1) such subordination consists either in judgments themselves or in acts of the understanding occurring within judgments; (2) all acts occurring within any judgment are kinds of use of concepts; and (3) in the various kinds of use of concepts, and in judging itself, various functions of the understanding are exercised. Kant’s First Section concludes by describing these functions as “functions of unity within judgments” (A69/B94; 86.20). Apparently this means that, by their respective kinds of subordination of representations under a common representation, each of the four acts of the understanding serves to provide unity in judging. The four distinct ways to serve this end thus correspond to exactly four different “functions of unity” in judging (cf. Figure 5.5).


m i c h a e l wo l f f

This does not mean that these four functions already individuate all the “functions of the understanding,” though it does mean that any function of the understanding (in its logical use) is one of these four functions. Because the logical use of the understanding can be traced back to subordinating concepts to concepts (per Figure 5.1), one may expect to find all the functions of the understanding (insgesammt; 86.19) by coordinating each of the four “functions of unity in judgments” with corresponding “moments” (86.31) of logical form (Verstandesform; 86.29) in judgments. Hence Kant concludes the First Section thus: The functions of the understanding can therefore all be found together if one can exhaustively exhibit the functions of unity in judgments. The following section will make evident that this can readily be accomplished. (A69/B94; 86.29–22)

Kant here describes his heuristic procedure (in the Second Section; B: §9) for finding all the functions of the understanding as a “complete exhibition” of the “functions of unity in judgments.” This exhibition involves assigning, in the Table of Judgments, the four identified functions to corresponding “Titles” (A70/B95; 86.30) by which moments of logical forms of judgment can be distinguished. This assignment is to show that the complete quadruple division of acts of the understanding in the First Section also affords a complete quadruple division of all moments of logical forms, as proposed in the Second Section by the Table of Judgments.8 Kant’s contention that all the functions of the understanding can be found in this way rests upon two assumptions which he does not make explicit, but which can be explicated as follows. One assumption is that the moments of the logical form of a judgment (e.g., universality, particularity, affirmation, and negation in the logical form of an A, E, I, or O judgment) correspond to moments contained in the “function of thought” (86.30) which are expressed in a judgment. Regarding its logical form, each judgment corresponds to the unity of an act of the understanding which contains several functions as moments (e.g., generalization, specification, affirmation, negation, etc.). Accordingly, each of these functions, as moments of an act of thinking, is manifest in one of the moments of logical form of judgment. 8

By the “logical form” (Verstandesform; A70/B95) of a judgment Kant means that which remains if one abstracts from the conceptual content of a judgment (cf. 86.28): “In every judgment one can call the given concepts logical matter (for judgment), their relation (by means of the copula) the form of the judgment” (A266/B322). Hence the logical form of a judgment can be presented by replacing its conceptual content with concept variables α, β, γ etc., as in the expressions (1)–(8) on p. 86.

How Precise Is Kant’s Table of Judgments?


The second assumption is that each function of the understanding contained as a moment within the unity of an act of the understanding is rooted in a subordination of representations to other representations and thus is either a moment within the unity of a specific use of a concept in judgment or is a moment within the unity of the act of judging itself. Kant’s third, explicit assumption is that the three kinds of use of concepts distinguished in the First Section correspond to three “Titles” under which those moments can be arrayed, which specify the logical form of a judgment regarding its quantitative, qualitative, or relational characteristic. These correspondences are: (1) The Title ‘Quantity’ corresponds to the predicative use of a concept in judging, because this consists in referring its predicate concept to one subject concept, thus “drawing together many possible cognitions into one cognition” regarding that to which the subject concept can be referred. (2) The Title ‘Quality’ corresponds to the nonpredicative, immediately object-referred use of a concept in judging, because this consists in referring the subject concept (and mediately the predicate concept) to objects possibly given in intuition, without which the judgment could not be true. (3) The Title ‘Relation’ corresponds to the mediately object-referred, nonpredicative use of a concept in judging, because within the judgment this consists in using a predicate concept of possible subsumable judgments as a subject concept of the judgment, thus making this into a potential major premise of a syllogistic inference. The Title ‘Modality’ corresponds to judging itself, insofar as it is thinking, and thinking is either “a function of the understanding” (in the narrow sense), or “of the power of judgment,” or “of reason” (A74–75/B100n).

5.4 The Precision of Kant’s Table of Judgments The completeness of Kant’s Table of Judgments consists, first, in there being exactly four titles under which the moments of logical form in judgments, and hence also the functions of the understanding, can be arrayed, because the distinction among these titles corresponds to a complete, principled division of the acts of the understanding (85.31, 93.26), namely, the idea of the understanding as a “capacity to judge.” This division assigns each function of the understanding, and hence each moment of a logical form of judgment, a “place” (85.2) in Figure 5.5.


michael wolff Logical function of the understanding in a judgment Judging / 4. Modality

Use of a concept in a judgment

predicative non-predicative / 1. Quantity Immediately Mediately referred to referred to an object an object / \ 2. Quality 3. Relation Figure 5.5

This “exhibition” assigns each of the four kinds of functions of the understanding to one of the four corresponding titles, beneath which logical forms of judgment can be arrayed; it does not specify how many or which moments these are. Kant’s Table of Judgments first exhibits (stellt vor Augen; A69/B94) that such an assignment is possible. Three of Kant’s four titles in the Table are familiar from traditional logic texts: ‘Quantity,’ ‘Quality,’ and ‘Modality.’ Categorical syllogistic typically used ‘Quantity’ and ‘Quality’ to distinguish two moments of the forms of categorical judgment designated as A, E, I, or O, and to consider these moments as subspecies of two kinds of logical determination. The title ‘Modality’ is used in the nonassertoric part of syllogistic to designate a third kind of determination which distinguishes relevant forms of nonassertoric judgment from the assertoric form, as a moment of the logical form of judgments. The fourth title in Kant’s Table, ‘Relation,’ is Kant’s innovation, introduced to designate the kind of logical determination characteristic of both categorical and noncategorical judgments in the hypothetical and disjunctive parts of syllogistic, as a moment of the logical form of judgments. The correspondence of each Title within Kant’s Table to one of the indicated four kinds of logical determination, of which the moments of the logical form of a judgment are species, indicates that all the logical functions of the understanding can “be found” (A69/B94), if it is possible to identify completely the moments belonging to each of the four Titles; that would identify all the moments of logical forms of judgment, and all the logical functions of the understanding expressed in these moments. Kant claims that each Title in his Table heads exactly three moments (A70/B95). Justifying this claim would complete Kant’s heuristic procedure for discovering all the logical functions of the understanding, thus

How Precise Is Kant’s Table of Judgments?


providing a complete “ground plan” (Grundriß; A130/B169) for a systematic construction of “general logic,” as Kant regarded syllogistic. Within Kant’s Critique the result of this heuristic lacks the form of a normal completeness proof. In the Second Section (B: §9) Kant obtains this result in two steps. First, within the Table itself, each of the four Titles is assigned three of the adjectives (then largely9 ) familiar from (syllogistic) logic: ‘Universal,’












These adjectives are useful designations of logical characteristics of judgments, and thus are suitable designations of moments of logical forms of judgment. Kant’s second step occupies the remainder of the Second Section (A70–76/B96–101); it defends this assignment against possible objections due to “several, although not essential” regards in which the Table of Judgments “seems . . . to depart from the customary technique of the logicians” (A70–71/B96). Showing that these apparent divergences stem from “misunderstanding” constitutes Kant’s second step in justifying his completeness claim. The first misunderstanding Kant allays (§9, No. 1, 2; A71–73/B96–98) concerns the Quantity and Quality of judgments, specifically: that by focusing only “upon the use of judgments in inferences of reason [syllogisms]” (87.20–21), syllogistic theory of inference is “limited only to the use of judgments with respect to each other” (88.1–2) and only treats judgments of the forms A, E, I, and O, thus neglecting singular and infinite judgments. Kant grants that the theory of inference “rightly” (A71, 72/B96, 97) limits its scope in this regard, insofar as syllogisms remain valid if any singular or infinite judgments they contain are reformed as universal or affirmative judgments, so that the proof of their validity is provided by proofs of the validity of syllogisms which only contain A, E, I, or O forms of judgment. However, this does not show that the singular and infinite forms of judgment are not logically distinguishable from universal or affirmative forms. Kant’s point is logical, not transcendental or epistemological (cf. e.g., Logik Pölitz 24:577). The quantity of singular judgments is clearly distinct from that of universal judgments, as Aristotle expressly indicated (An. Pr. I.1:24b26–28). 9

The designation ‘assertoric’ appears to be Kant’s innovation; see Reich 1986 [2001]: 74.


m i c h a e l wo l f f

Their distinctness is also clearly evident in the fact that the logical constants expressing the singularity of a judgment: ‘this . . . is . . . ,’ are not contained within any of the constants occurring in A, E, I, or O judgment forms.10 In contrast to universal judgments, singular judgments (like particular judgments) do not refer to a totality of objects to which their (respective) subject concepts can be referred. In contrast to particular judgments, singular judgments do not leave unspecified to how many they refer, for (like universal judgments) they refer to something which their (respective) subject concepts can be referred as a “whole” (87.25).11 Likewise, the infinite judgment is, according to its Quality, logically distinct to the affirmative judgment. This distinction is clearly evident in the fact that the logical constants expressing the infinity of a judgment (‘is a non-’) are not contained in any of the constants used in A, E, I, and O judgment forms. (By subordinating one concept to another according to Figure 5.1, infinite judgments (e.g., of the form ‘some α is a non-β,’ or ‘each δ is a non-β’) result by substituting in that figure ‘γ’ by ‘non-β.’) Although syllogistic as “general logic” disregards “all content of the predicate” (A72/B97), so that within syllogistic theory one can replace the negative predicate (‘non-β’) of an infinite judgment by a positive predicate (e.g., ‘γ’) to obtain an affirmative judgment, this does not show that by their logical form infinite judgments would be affirmative judgments. Instead, due to the negation involved in its predicate (as also in negative judgments), by their logical form infinite judgments are opposed to affirmative judgments. Infinite judgments are also distinct to negative judgments insofar as they (like affirmative judgments) are only true if their subject concept is nonempty, so that their being true does not follow from the falsehood of their contrary affirmative judgments. (Kant does not treat negation truthfunctionally; only thus can he distinguish negative from infinite judgments according to their logical form.12 ) Although syllogistic neglects singular and infinite judgments on technical grounds, as “general logic” it nevertheless presupposes as familiar their distinguishability from the forms of A, E, I, and O judgments. Hence one may say regarding ‘Quality’ and ‘Quantity’ that Kant’s Table merely “appears” to depart from “the customary technique of the logicians” (87.17), 10

11 12

Kant says nothing about logical constants expressing the singularity of singular judgments. According to Christian Wolff (1740: §241), the subject of a singular judgment is indicated by a “terminus singularis,” which is either a proper name, a demonstrative phrase or a specific singular designation. Accordingly, the logical form of the consequent (dependent) clause in judgments of the forms (5)–(8) on p. 86 is the form of a singular judgment. In Wolff (forthcoming) I discuss further Kant’s account of the form of singular judgments. Kant’s treatment of infinite judgments is similar to Aristotle’s; see An. Pr. I.46:51b36–52a15.

How Precise Is Kant’s Table of Judgments?


since it merely appears to enumerate more moments than must be taken into account in proofs of the validity of syllogisms from singular or infinite premises. Consequently, Kant’s arguments in §9, No. 1, 2, only reply to objections which purport that Kant’s Titles: ‘Quantity’ and ‘Quality’ contain, not too few, but too many moments: Kant here argues, not for the completeness, but for the “precision” of his Table of Judgments; both are claimed in Prolegomena, §39. By “complete” Kant indicates that none are lacking, whereas “precise” indicates that none are redundant (cf. Jäsche Logik 9:142, LL 633; Logik Philippi, 24:416). Two further misunderstandings Kant addresses (§9, No. 3, 4) concern moments of the two remaining Titles in Kant’s Table: “Relation” and “Modality.” Kant’s innovative Title, ‘Relation,’ is a possible source of misunderstanding (No. 3; A 73–74/B98–99); Kant highlights this in the Bedition (§19). There he refers to “the explanation that logicians give of a judgment in general,” according to whom a judgment is nothing but “the representation of a relation between two concepts” (B 140).13 Kant rightly objects that this “explanation” “fits only categorical, but not hypothetical and disjunctive judgments,” since “the latter two do not contain a relation of concepts, but of judgments” (B141). If all judgments were representations of relations between concepts, then it would be impermissible to use ‘Relation’ as a Title for a logical determination, by which moments within the logical form of judgments can be distinguished. In a footnote to §19 (B141n) Kant refers expressly to §9, to underscore that, according to No. 3, it is “false” to devote attention “only to categorical judgments as those to which all others have to be related.” Kant thus objects to a widespread tendency in the logical tradition (especially in Christian Wolff’s school) to disregard the logical distinction between categorical and noncategorical (as ‘simple’ and ‘compound’ or ‘unconditional’ and ‘conditional’) judgments, and so to reduce syllogistic inferences from hypothetical or disjunctive premises to categorical syllogisms.14 Consequently, Kant aims to show, also regarding the Title ‘Relation,’ that the Table doesn’t enumerate too many moments, but rather that it merely appears to diverge from standard logics, insofar as, considered logically, the Table ascribes to them different status.15 This is to say, also in §9, 13 14 15

E.g., Wolff (1740: 157), Lambert (1764: §119), and Meier (1752: §292). E.g., Christian Wolff (1740: Chapter 4, §§18, 19: 170–71). Hypothetical and disjunctive judgments, treated as compound (categorical) judgments, are assigned to the same class as ‘copulative’ judgments (based on the conjunction ‘and’) (cf. Lambert 1764: §133). On the “analysis” of copulative judgments as “brief compounds of simple statements” into “just as many simple” statements, and on the corresponding treatment of inferences from copulative judgments, see Lambert (1764, §§133–35, resp. §§278–80).


m ichael wolff

No. 3, Kant regards it as necessary to argue for the precision of his Table of Judgments. Accordingly, Kant divides “all relations of thinking in judgments” into “those a) of the predicate to the subject, b) of the ground to the consequence, and c) between the cognition that is to be divided and all of the members of the division” (A73/B98, 88.33–35). Regarding these relations, the logical distinction which holds between judgments only concerns the forms of (a) categorical, (b) hypothetical and (c) disjunctive judgments. These judgments thus forge a relation, respectively, between “two concepts” (in categorical), “two judgments” (in hypothetical), and “several judgments” (in disjunctive judgments) (88.36–89.1). So described, in hypothetical and disjunctive judgments there is a predicative use of a concept in connection with a logical subject only insofar as this use occurs within those judgments which are their components. Strictly, in hypothetical and disjunctive judgments a concept is used predicatively only within (component) judgments as their predicate. Hence it is not a predicate of the hypothetical or disjunctive judgment itself; there is no such predicate.16 Instead, within hypothetical and disjunctive judgments the use of concepts is not in their regard predicative. Because hypothetical and disjunctive judgments contain a nonpredicative use of concepts, which in a different respect is predicative, they are comparable to categorical judgments. As in categorical judgments, this consists in using a nonpredicatively used concept as a predicate of another possible judgment (see §5.2, pp. 92–5); unlike categorical judgments, however, hypothetical and disjunctive judgments contain this possible judgment as their own component. This is tantamount to a judgment which can be subsumed under them in valid hypothetical or disjunctive syllogisms and from which, as (an assertoric) minor premise of such syllogisms, it differs merely in having the value of a problematic judgment. The basic inference rules of categorical and noncategorical syllogistic are based upon the logical relations holding between the concepts in categorical judgments and between judgments in noncategorical judgments. Regarding categorical syllogistic, Modus Barbara and Modus Celarent are its basic rules; their validity does not derive from that of other syllogisms (cf. Aristotle, An. Pr. I.1). As Kant suggests in the First Section (see §5.3, p. 92), the validity of Modus Barbara is based upon the meaning of the relational expression, ‘each . . . is a . . . ’ (i.e., the logical constant: A), according 16

In Kant’s view, the relation between subject and predicate in hypothetical and in disjunctive judgments is entirely irrelevant to their logical form; cf. Refl. 3063 (16:636, NF 60), Logik Pölitz (24:578– 79), Logik Busolt (24:662), Wiener (Vienna) Logik (24:932–33, LL 372–73).

How Precise Is Kant’s Table of Judgments?


to which the predicate concept β of any judgment of the form, ‘each α is a β,’ pertains to an “still undetermined object,” to which α – as the predicate of a possible judgment of the form ‘some γ (or another) is an α’ – also pertains. Exactly parallel to this, the validity of Modus Celarent is based upon the meaning of the relational expression, ‘no . . . is a . . . ’ (i.e., the logical constant: E), according to which the predicate concept β of any judgment of the form ‘no α is a β’ pertains to no “still undetermined object,” to which α – as the predicate of a possible judgment of the form ‘some γ is an α’ – pertains. In other words, a valid syllogistic inference according to Modus Barbara results from a categorical premise of the form, ‘Each α is a β,’ because (by the Dictum de omni) this premise implies a judgment of the form: ‘If any (as yet undetermined) γ is an α, then this γ is a β.’ Likewise, a valid syllogistic inference according to Modus Celarent results from a categorical premise of the form, ‘No α is a β,’ because this premise (by the Dictum de nullo) implies a judgment of the form: ‘If some (still undetermined) γ is an α, then this γ is not a β.’ Accordingly, the validity of syllogisms in the first figure rests solely upon the meaning of the logical constants A or E which occur in their major premises.17 In just the same way the basic inference rules for hypothetical and disjunctive syllogisms are based upon the meaning of expressions for relations.18 Thus ‘if . . . , then . . . ’ in hypothetical judgments expresses that the antecedent of such judgments stands to the consequent in the relation “of ground to consequence” (A73/B98, 88.34). Hence if one presupposes a hypothetical judgment as a premise, then from its antecedent one can infer its consequent (by Modus ponendo ponens), simply because it follows from that antecedent (due to the presupposition). Similarly, by Modus tollendo tollens one can infer the denial of its antecedent from the denial of its consequent. These relations hold because the ground-consequent relation presupposed by a hypothetical judgment simply means that it is impossible that the antecedent be true yet the consequent false. Consequently the rules of Modus ponendo ponens and of Modus tollendo tollens are valid due solely to the meaning of the logical constant which occurs in their major premise: ‘if . . . then . . . ’ In exactly the corresponding way, the validity of the basic inference rules of disjunctive syllogism is based upon the meaning of ‘either . . . , or . . . (, or . . . etc.),’ which is expressed in the logical form of disjunctive judgments 17 18

Likewise, Aristotle explains Barbara and Celarent as “valid” (and “perfect”) by appeal to his elucidation of the meaning of A and of E in An. Pr. I.1:24b28–30; see Wolff (2009a: 225–30; 2009b). For details, see Wolff (2009a: 189–98).


m ichael wolff

and which specifies the relation holding between its subjudgments. Due to this meaning, the subjudgments contained within a disjunctive judgment are related as mutually exclusive “members of a division” of a totality of possible cognition (A73/B98, 88.35). Exactly one of these members must be true, if the disjunctive judgment is true. Hence if one presupposes a disjunctive judgment as premise, one can infer on the basis of one of these subjudgments (by Modus ponendo tollens) the denial of the others, and also from their denials (by Modus tollendo ponens) one can infer that first subjudgment. Hence these modes of disjunctive syllogisms are valid solely due to the meaning of the logical constant which occurs in their major premise, ‘either . . . , or . . . (, or . . . etc.).’ Kant holds that the logical forms of categorical, hypothetical, and disjunctive judgments belong under the common “Title” of Relation due to the logically distinct relations of their component judgments. This is based upon what was just demonstrated, that the three distinct forms of syllogistic inference, and hence also three distinct logical uses of reason, are based upon these three distinct logical relations. This is because the basic rules of these forms of inference are all based upon the occurrence in their major premises of a nonpredicative use of concepts, which are predicates of possible (i.e., subsumable) judgments, not only (as Kant indicates in the First Section) in categorical, but also in hypothetical and disjunctive major premises. This circumstance also requires distinguishing judgments differing in modality. The possible judgments relevant here are to be distinguished from the hypothetical or disjunctive judgments in which they occur as components because these component subjudgments as such claim no actual, but only a possible truth; i.e., it remains unspecified (unausgemacht; 89.5) whether they are true or false. Hence unto themselves they are, by Kant’s Table, not assertoric, but only problematic. (According to Kant, the “possibility of a judgment” consists in its being neither assertoric nor apodictic; Refl. 2167, 16:257.) At the end of No. 3 (A74/B99, 89.24–25) Kant accordingly notes that his remarks in No. 4 concern the very close connection between the aspects of modality (89.26–90.27) and those of relation. (The special status of modality is based upon the dichotomy presented in Figure 5.2.) According to No. 4, the modality of judgments is a quite special function of them, which is distinctive in that it contributes nothing to the content of the judgment . . . , but rather concerns only the value of the copula in relation to thinking in general. (B99–100, 89.26–30)

How Precise Is Kant’s Table of Judgments?


Kant here indicates that the distinction between the Title ‘Modality’ and the first three Titles corresponds to the division of logical functions of judgment in Figure 5.5. By that division the functions of the first three Titles are based upon the various uses of a concept in a judgment. Accordingly, they can be assigned to the “content” of a judgment, insofar as they are exercised within that judgment and concern its “logical matter,” as Kant calls the relata of categorical and noncategorical judgments.19 In contrast, the functions assigned to ‘Modality’ concern the judgment itself; i.e., they concern various moments of the logical form of a judgment insofar as, by judging within some context, its “copula” is assigned a different “value” for “thinking in general.” Kant elucidates this point in No. 4 by using the same examples used in No. 3 to distinguish the relations involved in hypothetical and disjunctive judgments. According to this exposition, the subjudgments of major premises in syllogisms according to Modus ponens or tollens are problematic; because in this context they cannot be evaluated as true, they only express “logical possibility” (B101; 90.14). To count as logically possible, they must merely accord with the principle of noncontradiction (principium contradictionis). If these same judgments are used in these syllogisms as a minor premise, in order to use them to infer something other than the logical possibility of a judgment, which the major premise already evaluates as possible, then they must be evaluated as true. That is to say, they must be assigned “logical actuality” (90.16). By this evaluation problematic judgments become assertoric, without altering their content. A third logical value results from this consideration: the conclusion of syllogisms according to Modus ponens or tollens is a judgment, which by its content corresponds to a subjudgment of the major premise, but which as a conclusion has a distinctive value, different both from the value of this subjudgment and from the value of the assertoric minor premise. If the premises of a valid inference are asserted to be true, then the judgment which is its conclusion is not merely assertoric, but also (according to the principle of noncontradiction) “asserts” a proposition “a priori,” because its falsehood is inconsistent with the presupposed premises; hence it “expresses logical necessity” (90.21–22). This explanation of modal distinctions in judgments consists in tracing them back to the logical distinctions between “three functions” which 19

See note 7. In Refl. 3046 (16:631, NF 59) and in the Wiener (Vienna) Logik (24:932–33, LL 372–73) Kant calls the subjudgments in noncategorical, i.e., in hypothetical and in disjunctive judgments, their “matter” (Materie). This does not preclude concepts being the “logical matter” constitutive of all judgments, because the matter of noncategorical judgments, too, are concepts.


m ichael wolff

are equally just “as many moments of thinking in general” (90.26–27). These are “moments” of thinking “in general” (generatim) insofar as (1) the problematic judgment consists in producing distinct (logically consistent) concepts by the understanding, (2) the assertoric judgment consists in producing cognitive judgments by the capacity to judge, and (3) the apodictic judgment consists in forging logical connections between cognitive judgments through inferences of reason. Kant’s explanation of these modal distinctions is not traditional. He maintains that modality “is not a particular predicate” (Prol. 4:325); he holds that modality does not depend upon the use of ‘possible,’ ‘necessary’ or other predicates as modal operators – which according to Wolff, Meier, Lambert, and other logicians (at least originally) do not belong to logic, but to ontology.20 By contrast, in Kant’s view, modal distinctions are logical distinctions, originally depending upon judgments – whether as such or as components of judgments – being used within an inferential context, so that their role within this context gives their logical quality a specific modal-logical “value.”21 Kant’s explication of these modal distinctions does not exclude the explicit de dicto use of logical constants (such as ‘it is possible, that . . . ,’ ‘it is necessary, that . . . ’) in modal syllogisms. However, on Kant’s explication these modal constants only belong to general logic so far as they can serve explicitly to ascribe to a self-sufficient judgment the modal-logical value it has due to its role within a presupposed inferential context.22 According to Kant, the logical use of understanding itself contains the use of judgments of distinct modalities. It would be, he thinks, an unfortunate “misunderstanding” to object that his modal distinctions diverge “from the technique of the logicians,” because in fact judgments have different modality, even when (as in the judgments of the forms (1)–(8) according to Figure 5.1) no modal-logical expressions occur. Considering all four of Kant’s numbered paragraphs in his Second Section shows that they do not purport to prove the completeness of the four trichotomous divisions comprised in the Table of Judgments. Instead Kant’s aim regarding all four Titles is to show that they are precise, insofar as the Table of Judgments does not contain too many logical functions 20 21


Cf. Lambert (1764: §137). In their logical treatises, Wolff and Meier disregard modality. Kant also describes different modal values of a judgment as different in the degree of their acceptance, i.e., in their being “gradually incorporated into the understanding,” i.e., into the capacity to think (B101, 90.22–25). On the derivation of the de re use of modal-logical constants from their de dicto use, see Wolff (2009a: 242–311, 420–21).

How Precise Is Kant’s Table of Judgments?


or forms. Kant’s arguments also show his intent to demonstrate that all four trichotomies contain more functions than the “customary technique of the logicians” would seem to suggest (A70–71/B96; 87.17). Kant notes that logicians are accustomed for “technical” reasons to neglect some of the logical moments Kant’s Table identifies, or to discount them as not belonging to the “mere” logical form of the moments of judgment (bloßen Verstandesform: A70/B95; 86.29). Kant’s arguments make plain why, without claiming to prove the completeness of his Table of Judgments, he explicated these twelve functions of judgment so as to make their completeness “certain” (A81/B107).

c h a p ter 6

Kant’s “Transcendental Deduction” Barry Stroud

Kant’s unprecedented “transcendental” investigation of the human mind and its powers arises from a distinctive question. It is not the question how the “materials” we employ in thinking get into the mind in the first place. John Locke had tried to show in detail that sense experience alone can explain how the mind “comes to be furnished” with that “vast store” of ideas “which the busy and boundless Fancy of Man” is capable of (Locke 1975, II,1: 104). A successful explanation along those lines would leave no reason to regard some ideas as “native” to the human mind, or “stamped upon” it, or given “innately” to it from some other source. Whether anything like that is so or not is in a very broad sense a question of fact. Locke meant to answer that question of fact, and Kant concedes that “we are indebted to the celebrated Locke for opening out this new line of enquiry” (A86/B119) – “a certain physiology of the human understanding” (Aix). But the question Kant starts from is different, as he explains. Jurists, when speaking of rights and claims, distinguish in a legal action the question of right (quid juris) from the question of fact (quid facti), and they demand that both be proved. Proof of the former, which has to state the right or the legal claim, they entitle the deduction. (A84/B116)1

The “deduction” Kant thinks is needed for understanding the human mind would establish and explain our “right” or “entitlement” to something we seem in fact to possess: experience, concepts, and principles that we employ in “the highly complicated web of human knowledge” (A85/B117). For many concepts – so-called “empirical” concepts – no such question seems to arise, since “experience is always available for the proof of their objective reality” (A84/B116). That could be loosely called an “empirical deduction” of those concepts. But it would show only “the manner in which a concept is acquired through experience . . . not its legitimacy, but only its de facto 1

In this chapter, translations from the first Critique are from the Norman Kemp Smith translation, unless otherwise noted.


Kant’s “Transcendental Deduction”


mode of origination” (A85/B117). The question of “right” or “legitimacy” demands a different kind of “deduction.” There is a special problem about this because Kant thinks some of our concepts “are marked out for pure a priori employment, in complete independence of all experience” (A85/B118). Since “empirical proofs do not suffice to justify this kind of employment, we are faced by the problem how these concepts can relate to objects which they yet do not obtain from any experience” (A85/B117). Explaining how this can be so is the task of what Kant calls the “transcendental deduction” of those concepts (A85/B117). We will return to what he means by calling it “transcendental.” It is to be a “deduction” in the sense Kant explains: a justifying or legitimating explanation of how our use of those concepts can yield knowledge of what is so, despite their independence from all experience. Whether any of our concepts are in fact “marked out for pure a priori employment,” and so present us with this special problem of how they can relate to objects independently of experience, is also, in a broad sense, a question of fact. Is that actually true of some of our concepts, or not? Kant has no doubt that we do have many a priori concepts, and that we do know many things a priori, because he thinks we all know many things to be necessarily true. And for Kant necessity is a “sure criterion” of the a priori: “if we have a proposition which in being thought is thought as necessary, it is an a priori judgment” (B3). So the task of the legitimating “deduction” of our a priori concepts is to explain how we know the kinds of things Kant thinks we all know a priori. Kant does not address this problem in its full generality. The question of how a priori concepts “can relate to objects which they yet do not obtain from any experience” arises only for what Kant calls “synthetic” a priori knowledge. No such challenge arises for so-called “analytic” truths which can be known a priori simply by reflecting on a given concept and “extracting” its “constituent concepts” from it “in accordance with the principle of contradiction” (B12). But “synthetic” a priori judgments are “ampliative” and “add to the concept of the subject a predicate which has not been in any wise thought in it” (A7/B11). The problem then is to explain how those judgments can “relate to objects” they are about completely independently of all experience and not simply in accord with the principle of noncontradiction. The question of how synthetic a priori knowledge of that kind is possible is for Kant the same as the question of how metaphysics as reliable philosophical knowledge of the world is possible. A successful “transcendental deduction” that answered that question would therefore put metaphysics


bar ry st roud

“on the secure path of a science.” The quid juris of the claims of metaphysics would have been answered positively and their “legitimacy” would have been explained. Kant is optimistic about the prospects of this kind of metaphysics because “its subject-matter is not the nature of things, which is inexhaustible, but the understanding which passes judgment upon the nature of things” (A12–13/B26). That subject matter is “nothing other than reason itself and its pure thinking; and to obtain complete knowledge of these, there is no need to go far afield, since I come upon them in my own self” (Axiv). Kant believes that fruitful metaphysical conclusions about the nature of reality are to be expected from an investigation of our a priori ways of thinking about it because “these a priori possessions of the understanding, since they have not to be sought for without, cannot remain hidden from us, and in all probability are sufficiently small in extent to allow of our apprehending them in their completeness, of judging as to their value or lack of value, and so of rightly appraising them” (A13/B26–27). So Kant’s “deduction” is to vindicate metaphysical knowledge of objective reality through the proper understanding of “these a priori possessions of the understanding.” I will concentrate almost exclusively on the single central line this “deduction” follows to reach its desired end. It starts with the conditions of thought. Kant first establishes, of each of an identified set of very general concepts, that its employment is essential to the possibility of any thought. For Kant thought is possible only for those with a capacity for judgment – for thought of something or other’s being so or of a certain thing’s being so-and-so. By reflection on what he regards as the fundamental forms that any possible judgment must take he arrives at an exhaustive set of extremely general concepts that he thinks must be applied to some things for any judgments at all to be possible. If all human thinking must conform to those forms of judgment, those concepts can be known to be the fundamental “categories” that must be present in “the understanding” of anyone who can think about anything at all. The most intricate sections of Kant’s overall “deduction” concentrate on the complex relations among a rich set of further conditions that must be fulfilled for any thoughts within even these most general and abstract “categories” of thought to be possible. There is thought only if a capacity for thought is actually put to work on something. Any thought of objects, or any application of concepts to an object to make a judgment, requires at least that objects be present to us in some way. Objects that are present to us in experience obviously must conform to the “formal conditions” of

Kant’s “Transcendental Deduction”


“sensibility” – the conditions necessary for our being sensorily aware of something. Otherwise, whatever affected us would be as nothing to us as far as having thoughts about it is concerned. But although whatever we are aware of in experience must conform to the conditions of our being aware of it, that alone does not imply that the objects we are aware of in experience must conform to the conditions necessary for the possibility of thought of an object, for a judgment that such-and-such is so-and-so. It seems possible at first that “appearances” might present themselves in experience so randomly and in such confusion that “the understanding” could make nothing of them, and would find no objects to form thoughts about, and so could not think of anything at all (A90/B123). The heart of Kant’s “deduction” is the attempt to explain how this apparent possibility presents no real difficulty. The possibility of human thought and experience is to be accounted for by showing how and why objects can be given to us in experience only because the forms in accordance with which thoughts must be unified into judgments about something are also the forms within which whatever affects our “sensibility” must be unified to yield experiences of objects. That would bring the conditions of thought and the conditions of experience together as necessary for each other. If there could be no experiences that do not conform to the conditions necessary for the possibility of judgment, and all judgment involves application of the fundamental and therefore a priori concepts of the human understanding, that would explain how those a priori concepts can “relate to objects which they yet do not obtain from any experience.” The application of those concepts to some objects of experience is held to be a necessary condition of there being any thought or experience of any objects at all, so the “objective reality” of those concepts in general is secured simply by the fulfillment of that necessary condition, not through having to have experience of those objects themselves. For Kant, thinkers who apply such a priori concepts to something, and not solely “in accordance with the principle of contradiction,” would thereby know a priori that certain things are so in the world they experience. Showing how and why that must be so would explain how synthetic or “ampliative” metaphysical knowledge of the world is possible completely a priori. The quid juris of the claims of any metaphysical investigation pursued according to that plan would be answered positively and the “legitimacy” of its results accounted for. The success of this ambitious philosophical enterprise obviously depends on the success of the arguments for the necessity of the fundamental “categories” and of the many arguments throughout the Analytic of Principles


bar ry st roud

by which Kant reaches the conclusion that this or that more specific way of thinking is a necessary condition of the possibility of thought and experience in general. Interpretation and criticism of those difficult arguments have kept philosophers busy ever since the Critique of Pure Reason was first published, and continues to this day. Our concern here is not with the validity of those arguments but only with the special character of the Kantian “transcendental” project itself and the conditions of its success. That remains a formidable problem even if it is conceded that Kant’s arguments for the necessity of the conditions he identifies are all correct. The arguments are meant to establish a number of interrelated conclusions that could be summed up loosely and in general terms as follows. They start from the idea that for experiences and thoughts to belong to a conscious subject that subject must be able to think of those experiences and thoughts as belonging to him. That requires a capacity on the thinker’s part to think of those experiences and thoughts as belonging to him but distinct from him, and also to think of whatever he takes those experiences to be experiences of. That requires that thinkers be capable of thoughts about something other than themselves and their experiences, and so to that extent are capable of thoughts about something objective. Some of the objects of any thinker’s thoughts or experiences must be thought of as existing or as being as they are independently of their being thought about or perceived by anyone. This requires in turn that some of those objects be thought of as enduring independently in space and time and that states of affairs involving those objects be thought of as occurring in an order not necessarily the same as the order of thinkers’ perceptions and thoughts about them. That independent order must be thought of as containing enduring objects capable of entering into events which are connected with other events in accordance with causal laws. I think this loose, informal list of a few of the very general “results” Kant takes himself to have reached about the necessary conditions of thought and experience will suffice for our purposes. To understand the general overall structure of the “deduction” we must understand not only those “results” themselves but also the precise role they are meant to play in the “deduction” even if they are correct. It is important to observe that the conclusions of Kant’s arguments identify certain propositions or ways of thinking that all thinkers must understand and accept as a condition of their having any thoughts or experiences of anything at all. The idea is that whoever can think at all must think that each of the very general propositions Kant identifies describes some aspect of the way things are in the world, at least in general. That is, for Kant, a condition of thought.

Kant’s “Transcendental Deduction”


Even if Kant’s arguments for these conclusions are sound and the “results” arrived at in the Analytic of Principles are taken as established, we can ask about the distinctive epistemological and metaphysical status those conclusions would thereby have been shown to have, and how their having that status serves Kant’s announced purposes. The arguments identify a number of propositions that all thinkers must accept or judge to be so. That certainly gives those propositions a special, distinctive status in any thinker’s conception of the world. But Kant’s “deduction” seeks to establish and explain something stronger than that. It claims not only that certain propositions must be accepted or judged to be so by anyone who can think anything, but that what is said to be so by propositions that have that distinctive status is something all thinkers know a priori to be true of the world they live in. That is a strong epistemological conclusion about the knowledge human beings must have of the world they live in; they know a priori that those propositions that Kant shows have that distinctive status are true in the world. And that implies corresponding metaphysical conclusions about what the world those thinkers know about must be like. What is remarkable is that Kant appears to reach those conclusions about the world and our knowledge of it from nothing more than the necessity with which all thinkers must think of and experience things in certain ways. This calls for closer attention because the fact that someone thinks or judges that such-and-such is so, even if his thinking it is required for his being able to think of anything at all, does not on its own seem to imply that what that thinker thereby thinks is true. Nor does it imply that what that thinker thereby thinks is something he knows to be true either. Everyone’s having to think it is one thing; its being true, or its being known to be true, is something else. Kant holds not only that the distinctive propositions he has identified are known by all thinkers to be true, but that they are known a priori to be true. That is what the “transcendental deduction” is meant to establish and explain. But it is not established, and certainly not explained, solely by the fact (even if it is a fact) that thinkers must think all such propositions are true if they can think and experience anything at all. We saw that for Kant necessity is a “sure criterion” of the a priori: “if we have a proposition which in being thought is thought as necessary, it is an a priori judgment” (B3). But if Kant were relying on that principle to support the view that the propositions he claims all thinkers must accept are known a priori, it would mean that he thinks those propositions, in being thought, are thought as necessary. If that means that those propositions are thought to hold necessarily in the sense of there being no possibility of their being or having been false, it would mean that Kant thinks that


bar ry st roud

the distinctive general propositions he identifies – e.g., ‘Objects and their properties endure in space and time independently of whether they are perceived or thought about’ or ‘Objects enter into events that are connected by causal laws with other objects and events’ – hold necessarily and could not possibly have been false under any circumstances. Since Kant holds that those distinctive general propositions must be accepted by anyone who can think, perhaps he means only that those propositions are “necessary” in the sense that any thinker must accept them or judge them to be true. They would be “necessary” in the sense of being necessary-for-thought; they must be held to be true by anyone who can think. But understanding their “necessity” in that way does not imply that in being thought, those propositions are themselves thought to hold with necessity. So that way of understanding their “necessity” would not in itself support the idea that those propositions are even known to be true, let alone that they are known a priori. If we do know many general propositions of that kind to be true, as I think we do, and if Kant is right about the distinctive status his arguments show those propositions to have in our conception of the world, it still does not follow that we know them a priori. Kant explicitly argues for something he thinks does hold with necessity, and so could not possibly be or have been false. He claims in each case that if there is any thought and experience at all, then thinkers accept those distinctive propositions that he argues are required for the possibility of thought and experience. If Kant is right about that, and if he is also right that necessity is a “sure criterion” of the a priori, then those claims he makes about the necessary conditions of thought are something Kant knows a priori. All the rest of us would also be able to know a priori that those propositions Kant identifies must be accepted by anyone who thinks. But what is thereby known or knowable a priori is a conditional claim. It does not imply that any of those propositions that must be accepted by thinkers are themselves known a priori, even if they are known. Kant thinks those propositions that must be accepted by anyone who can think involve the application of concepts that could not be acquired simply by finding instances of them in experience; they involve what he calls a priori concepts. That is one mark of the distinctive status of those propositions. But not every proposition involving the application of an a priori concept enjoys the distinctive status Kant identifies. Nor is every proposition involving application of an a priori concept knowable or known a priori. For instance, to know that an object of a certain kind caused such-andsuch changes in an object of another kind at a certain place and time one must rely on observation and experience. Such truths about the world are

Kant’s “Transcendental Deduction”


knowable only a posteriori, not a priori, although what is thereby known involves application of a priori concepts like ‘enduring object’ and ‘causation.’ Kant’s arguments establish only that a capacity for the application of such concepts to objects of experience in general is required of every thinker. He takes that to show that they are a priori concepts. But the “transcendental deduction” aspires to more than that. It seeks to establish that all thinkers know a priori that those a priori concepts truly apply to objects in the world they think about and experience, and then to explain how that a priori knowledge is possible. Knowing such completely general truths about the world a priori is compatible with knowing a posteriori that those fundamental concepts apply in particular to this or that item or kind of item in the world as things are. Kant’s “deduction” does not aim to show that all knowledge expressed in terms of those fundamental concepts is a priori. And the fact that relatively specific propositions about enduring objects or causal connections, for instance, can be known only a posteriori might suggest (as it has to many philosophers) that we could come to know that those concepts apply to things in the world in general simply by generalizing from what we can find in particular cases. After all, if we have found some objects standing in causal relations with others, we know that, in general, there are objects standing in causal relations in the world. But Kant’s arguments rule out the possibility of knowing such completely general truths only a posteriori. Let us suppose that it is a necessary condition of thinking of and experiencing anything at all that one accepts and makes sense of one’s experience in general in terms of enduring objects and causal connections, for instance. Then one could not come to know by experience that more specific versions of those concepts apply to this or that kind of object or cause without already possessing the more general concepts of enduring object and causal connection. For Kant that necessity shows that they are a priori concepts. ‘There are enduring objects and causal connections in the world’ is a more general statement than any statement about particular kinds of objects or causes, but one could not come to know that more general statement a posteriori by generalizing from the more particular instances one finds to be true in one’s experience. What this shows, strictly speaking, is that the completely general propositions Kant argues must be accepted by anyone who can think and experience anything could not be known a posteriori. That is an important step in the overall strategy. But it does not alone achieve the declared goal of Kant’s “transcendental deduction”: to establish that such general structural propositions are known a priori to be true of the world, and to explain how


bar ry st roud

that synthetic a priori knowledge is possible. To succeed in that ambitious task it must be shown not only, as Kant has argued, that those distinctive general propositions must be understood and accepted by anyone who can think and experience anything at all, and not only that for that reason they cannot be known a posteriori. It must be shown that those distinctive general propositions are known a priori by all thinkers and perceivers to be true of the world they live in. Holding that anyone who can think at all must employ certain distinctive (a priori) concepts, and that the conditions of thought expressed in those concepts apply as well to the conditions of all possible objects of experience, puts Kant in position to explain how a posteriori knowledge of the world is possible in general. If those very general a priori concepts are true of things in the world, and if they hold as well of all objects of experience, perceivers could often know a posteriori that more determinate specifications of those general concepts apply to particular objects they are aware of in their experience. And since those conditions hold as well for the possibility of experiences of objects, empirical knowledge of those features of independent “outer objects” could be gained by “immediate perception” of them. That would give us direct knowledge of “a reality which does not permit of being inferred, but is immediately perceived” (A371). Anything less than that as an account of our relation to the things around us would leave us vulnerable, in general, to what Kant calls “problematic idealism,” which he finds, for instance, in Descartes. On that view, the existence of external objects can be at best only inferred from one’s “inner perceptions,” and so remains always subject to doubt or uncertainty. Kant recognizes that all such – as he calls them, “idealist” – conceptions of experience and of perceptual knowledge of an independent world must be seen through and rejected. In his preface to the second edition of the Critique he draws special attention to the question. However harmless idealism may be considered in respect of the essential aims of metaphysics (though, in fact, it is not thus harmless), it still remains a scandal to philosophy and to human reason in general that the existence of things outside us (from which we derive the whole material of knowledge, even for our inner sense) must be accepted merely on faith, and that if anyone thinks good to doubt their existence, we are unable to counter his doubts by any satisfactory proof. (Bxl)

For Kant the way to “counter the doubts” of those who find the existence of “things outside us” doubtful or epistemically problematic in general would not be to try to “prove,” somehow, that the “outer objects” we perceive

Kant’s “Transcendental Deduction”


really do exist after all. How could such a thing be proved? And would whatever could be proved really silence the doubts of those who find the existence of such objects problematic? The strategy of Kant’s “transcendental deduction” is rather to: show that we have experience, and not merely imagination of outer things; and that this, it would seem, cannot be achieved save through proof that even our inner experience, which for Descartes is indubitable, is possible only on the assumption of outer experience. (B275)

Such a “proof” would completely turn the tables on “problematic idealism.” On that view, “the only immediate experience is inner experience,” so “outer things” can at best be known only by inference from what is experienced. But the “proof” embodied in Kant’s “deduction” is meant to show that “outer experience is really immediate, and that only by means of it is inner experience . . . possible” (B276–77). “Outer objects” of experience as Kant’s “deduction” explains them are required for the possibility of any experience of anything at all. What we experience directly are objects, causal relations between objects, etc., in the world as it is independently of our perceiving them. We could not even have the kind of “inner experience” “problematic idealism” takes for granted without being capable of “immediate perception” of “outer objects” that are as they are independently of us in that way. For knowing a posteriori how things are in the world there would be no need, in general, to rely on inference from what we experience to something else. With respect to a posteriori knowledge of the world around us, then, our position in the world must be understood as what Kant calls “empirical realism.” It is “realism” in the metaphysical sense that objects and events and their properties exist and are as they are independently of us and our perceptions of them. And it is “realism” in the epistemological sense that our knowledge of those “outer” or independent objects and their properties is possible by direct perception of them. Perceptual experience is not restricted to what “problematic idealism” calls “mere appearances” or “inner experiences” of objects, all of which depend for their existence on being perceived. Kant’s “empirical realism” says we can know by experience that we are in direct perceptual and cognitive touch with the independent objects around us, with no need to infer their existence from something we are said to be even more immediately aware of. This, of course, is meant to hold only in general. Whether this or that particular experience is an experience of how things independently are is something that can be doubted on this or that occasion, and so enquired into and settled empirically. But the


bar ry st roud

completely general “proof” of “empirical realism” would suffice to counter all “problematic” doubts about the possibility of a posteriori knowledge of an independent world in general. The possibility of a posteriori knowledge of the world of nature would be successfully accounted for in this way only if it is true, as Kant claims, that all thinkers who employ the required a priori concepts know a priori that those concepts are true of objects in the world they live in. Everyone’s knowing a priori that that is the way the world is, at least in general, is what would account for the possibility of their knowing by experience that relatively specific determinations of those same general a priori concepts apply to certain objects within their experience. But has that guarantee of the truth of “empirical realism” been secured? Is it true that everyone who employs a priori concepts knows a priori that those concepts have “objective reality”; that they truly apply to objects in the world? Has that been established by Kant’s “deduction” as developed so far? And if so, has it also been explained how that a priori knowledge is possible, as a successful “deduction” requires? Here we come at last to the sense in which Kant’s “deduction” is to be understood as “transcendental.” I entitle transcendental all knowledge which is occupied not so much with objects as with the mode of our knowledge of objects in so far as this mode of knowledge is to be possible a priori. (A11–12/B25)

To have “transcendental” knowledge, or to engage in a “transcendental” inquiry, is to know or inquire into how a priori knowledge of certain kinds is possible: the conditions that must be fulfilled for there to be knowledge of that kind. Any knowledge or understanding yielded by such an inquiry must itself therefore be a priori, concerned as it is with the conditions and the possibility of a priori knowledge. The point of such a “transcendental” inquiry or understanding is not simply to gain knowledge of the world but to achieve what turns out to be a distinctively philosophical understanding of how human knowledge of the world is possible in general. “The distinction between the transcendental and the empirical belongs therefore only to the critique of knowledge; it does not concern the relation of that knowledge to its objects” (A57/B81). We have seen that for Kant “realism” is the only satisfactory explanation or “critique” of a posteriori knowledge of things in the world around us. We can determine by empirical investigation that most of the objects we perceive exist and are as they are independently of us and our perceptions of them, and we can gain immediate, underivative knowledge of them and

Kant’s “Transcendental Deduction”


their properties by perceiving them. By contrast, any form of “idealism” concerning this knowledge of the world would leave us at best only in indirect perceptual and cognitive touch with objects in the independent world, and so always in “problematic” or “skeptical” doubt, with no reliable knowledge of the way things are. But if we ask now in this same “transcendental” spirit how the a priori knowledge of the general structure of the world that makes this “empirical realism” possible is itself to be explained, Kant finds an apparently parallel “realism” at this “transcendental” level completely unacceptable. To accept, as part of a “transcendental” explanation of our a priori knowledge, that the objects we perceive and know about exist and are as they are independently of us and our capacities for experiences of them, would leave that allegedly a priori knowledge completely inexplicable to us. Kant’s original problem was how those concepts that are marked out for a priori employment “can relate to objects which they yet do not obtain from any experience” (A95/B117). Here Kant finds that if the objects we could have knowledge of were all completely independent of us and our capacities for knowing them, we could have no a priori concepts that we had any reason to suppose applied to those objects, and so no a priori knowledge of those objects at all. This would leave us with no explanation of the a priori knowledge Kant thinks every thinker has of the “objective reality” of those a priori concepts he must employ if he can think at all. And that would mean that the “empirical realism” just now thought to have been vindicated by that a priori knowledge would be left unexplained. To accept “realism” as a “transcendental” explanation of the possibility of a priori knowledge would leave both a priori and a posteriori knowledge of the world unintelligible to us. For Kant the only acceptable “transcendental” doctrine – “the only refuge left open” – is therefore idealism. “Transcendental idealism” is the only way of avoiding “problematic” or “skeptical” idealism about the world of nature, and so the only way of guaranteeing the truth of “empirical realism.” We can be sure in general of reliable a posteriori knowledge of a world of nature that is “empirically” independent of us because – as “transcendental idealism” implies – the objects we are aware of in that world are one and all “appearances,” speaking “transcendentally.” They are not objects that are as they are independently of their fulfilling the conditions of being thought of and experienced by us. They are not, in the “transcendental” sense, independent of their fulfillment of the necessary conditions of their being thought of and experienced. Kant’s “transcendental idealism” explains the truth of “empirical realism” only because it implies that the necessary conditions of thought and


bar ry st roud

experience are “supplied by us” to the objects we think of and experience. We thinkers “contribute” to what we think of and experience those very conditions without which we could not think of or experience anything at all. It is because we “supply” or “contribute” those conditions to whatever we think of or experience that we cannot find anywhere in our thought or experience anything that fails to conform to them. This same understanding of idealism as our “contributing” to the objects we know about is also involved in accounting for our knowing a priori that the fundamental categories of thought apply to anything we can think of or experience. That a priori knowledge must be accounted for if the truth of “empirical realism” is to be explained. So the “transcendental idealism” that explains both a priori and a posteriori knowledge really is a robust form of idealism; whatever we think about or experience is not something that is as it is independently of its fulfilling the conditions of our thinking of or experiencing it. What we think about or experience is dependent on something about us; it is in that sense not fully independent of “us” and our “contribution” to it. For Kant this is “the only refuge left open” for explaining “transcendentally” our knowing what we know a priori and our knowing what we know a posteriori. I think Kant must understand his “transcendental idealism” in some such way if he is to regard his “deduction” as successful and as having explained the possibility of our a priori and our a posteriori knowledge “transcendentally.” He does not appear to envisage or even find intelligible any form of idealism in which we do not “contribute” something to our knowing or experiencing the things we do. He rejects “transcendental realism” from the outset as even a candidate for a possible explanation of our a priori knowledge of objects that are understood as independent in every way of us and our “contribution” to them. He appears to find the possibility of a priori knowledge of such objects or states of affairs as beyond explanation. For him a priori concepts can be understood to “relate to objects which they yet do not obtain from experience” only because we “supply” the conditions that make possible the objects which we take those concepts to apply to. In calling “transcendental idealism” a robust form of idealism I mean it is directly opposed to “realism.” It implies that we are never in cognitive or experiential contact with anything that is as it is independently of its fulfilling the conditions that “we” “supply” for being in cognitive or experiential contact with it. The tables and trees and other apparently “outer” objects we are aware of in experience are, “transcendentally speaking,” only “appearances” which depend for their nature on the “contribution” we perceivers make to our being aware of them. As a “critique” of knowledge, then, or as a

Kant’s “Transcendental Deduction”


philosophical understanding of how our knowledge is possible, the objects we experience cannot be understood as being as they are independently of whether we experience them or not. That same understanding of “idealism” also means that the objects we take our purely a priori concepts to apply to are not objects that are as they are independently of whether we think of them in those a priori ways either. Kant finds no explanation of a priori knowledge possible along any such strictly “realist” lines. He thinks the complete independence of such objects would leave no way to account for what he regards as the “objective reality” of our a priori concepts. So if Kant’s “deduction” is to explain the possibility of a priori knowledge of the world – as he requires – it must therefore see the necessary conditions of the possibility of thought and experience as something “supplied” or “contributed” by us. The explanation must lie in those “a priori possessions of the understanding” which “have not to be sought for without,” and “cannot remain hidden from us” because “I come upon them in my own self.” This account of the conditions of success of Kant’s “deduction” leaves two lingering questions worth pondering. (1) Has Kant accounted “transcendentally” for a priori knowledge of the “objective reality” of our a priori concepts – our knowing a priori that such concepts apply truly to things in the objective world? Or has he explained only the impossibility of our ever finding anything within our thought or experience to which those concepts do not apply? (2) If the necessary conditions of thought and experience of things in the world are “supplied” or “contributed” by “us,” how is that known? If we have found those conditions to be necessary for the possibility of all thought and experience, as Kant claims, can we hope to explain their holding necessarily by appeal to something which does not appear to be necessary at all? To insist that we know a priori that those conditions must be “supplied” or “contributed” by us – that is “the only refuge left open” – because otherwise we would have no “transcendental” explanation of a priori knowledge of the world, seems to leave us at best with an uncomfortable disjunction: either we “supply” the conditions that make thought and experience of the world possible or we have no satisfactory explanation of a priori knowledge at all.

c h a p ter 7

Kant’s Critique of the Layer-Cake Conception of Human Mindedness in the B Deduction James Conant∗

7.1 The Layer-Cake Conception, Three Exegetical Puzzles, and the Aim of the B Deduction The aim of this essay is to suggest that Kant should be read as seeking to take aim at a deeply rooted assumption – one that has controlled much modern philosophical thought about the nature of human cognition. It is this: our nature as sensibly receptive beings, in so far as it makes a contribution to cognition, represents a self-standingly intelligible aspect of our nature. According to this assumption, to claim that the sort of knowledge that animals like us – rational animals – have requires “something more” than our merely sensible nature is to claim that there are capacities which must be “added on” to our “merely animal” capacities for sensation and desire. To understand human cognitive functioning in this way is to picture it as a layer-cake: the bottom-level of the cake is the layer of our merely animal capacities for interacting with the world. The layer that sits on top of that is the upper layer of human cognitive functioning: the layer of our (more or less) distinctively human (so-called rational) capacities. What is crucial to the assumption is the following idea: that the internal character ∗

The present essay is a selection from a longer monograph that I hope to publish soon. The interpretation of Kant presented here is the product of a collaborative effort. It was my privilege to read through the First Critique and discuss it in detail, line by line, over a period of six years, from 1993 to 1999, with Haugeland and McDowell, who were at the time my colleagues in the Philosophy Department at the University of Pittsburgh. A fairly detailed description of the modus operandi of our collaboration may be found in McDowell’s paper “Notes on the B Deduction” (forthcoming). John Haugeland is sadly no longer with us and I am thus no longer able to discuss these topics with him. I have, however, on any number of occasions continued to discuss both the exegetical details and philosophical implications of this reading of Kant with John McDowell and am no less indebted to these further conversations with him. Though the reading of Kant’s First Critique presented in these pages is for these reasons a product of a collaborative effort, some of the wrinkles in formulation introduced here are due to me. If I could, I would attempt clearly to mark off what is my own development of the reading in question from what originally belonged to our joint understanding of Kant. However, I am very far from being in a position to do so.


Kant’s Critique of the Layer-Cake Conception of Human Mindedness 121 of the manifold constituting the bottom layer remains unaffected by the introduction of the upper layer. The concerns of the essay involve an intertwining set of systematic and exegetical concerns. Let us begin with the latter. Here are three central exegetical puzzles with which any satisfactory reading of the first Critique must come to terms: First puzzle: What is the relation of the doctrine of the formal conditions of sensibility set forth in the Transcendental Aesthetic and the doctrine of the formal conditions of understanding set forth in the Transcendental Analytic? Second puzzle: What is the relation of the versions of the Transcendental Deduction offered in the A and B editions of the first Critique? Third puzzle: What is the relation between the first half and the second half of the Transcendental Deduction in B? The main exegetical claim of this essay may be summed up as follows: the proper resolution of each of the three puzzles depends upon the proper resolution of the other two. This means that we must answer the first question properly if we hope to make genuinely satisfactory progress on the other two. In its relation to contemporary Anglophone Kant commentary, the main polemical claim of the talk may therefore be put as follows: most accepted solutions to the first puzzle render the second and third puzzles insoluble. The aim of what is to be shown by the end of the B Deduction may be summed up as coming in the following three steps: (1) What has already been shown in the Transcendental Aesthetic What we intuit through our senses has, as such, a certain form: namely, that of space and time. We represent what we intuit as spatial and temporal merely in virtue of intuiting it. Space and time are the forms of our intuition. (2) What is shown in the first half of the B Deduction That which is given through the senses can only exhibit unity of intuition – regardless of what the specific character of the form of intuition in question is – if it exhibits the unity of thought – categorical unity – those forms of unity which characterize any finite discursive intellect. (3) What remains to be shown in the second half of the B Deduction What is given through the senses exhibits the form of thought if and only if the categories do not prescribe a unity which is simply


j ames conant other than that which something exhibits insofar as it is spatial and temporal.

The crucial question is this: what is accomplished in the progress from the second to the third of these steps? Only if we properly understand this are we in a position to appreciate the true nature of the progress from the first to the second of these steps. In order to clarify what the philosophical stakes are here, I will first engage in a brief discussion of the issues that dominate Anglophone Kant commentary. I will present these as four (as I will call them) choice-points in a reading the first Critique.

7.2 Four Choice-Points in a Reading of the First Critique I will begin by presenting these choice-points separately, as if they represented fully independent exegetical issues – but, in fact, I do not think they do: I think that the commitments incurred in an attempt to take (what we might call) the left-hand or the right-hand fork at any one of these choicepoints are intimately related to the commitments incurred in any attempt to take the left-hand or the right-hand fork at any of the other three. 7.2.1

A First Choice-Point in Reading the Deduction: Restrictive versus Nonrestrictive Conceptions of Subjectivity

It is clear that Kant’s aim in the first Critique is to elucidate the concept of finite knowledge. But much depends upon how the concept of finitude or limitation is elaborated here. There is a tendency to elaborate it in (what I will call) restrictive terms – in accordance with a conception according to which the finite knower is pictured as if he were sealed into a delimited sphere. According to this picture, there is an area, within which our reason operates, but the price of our being able to enjoy its satisfactory operation within that domain is that we must pay the price of “running up against a limit” – a limit the other side of which we can dimly make out, although we cannot have knowledge of what goes on there. So, on this picture our form of knowledge is restrictive because it debars us from being able to enjoy a kind of knowledge it would make sense for us to hanker after but which, alas, we cannot have.1 Kant is often read as if his conception of the finitude of our faculty of knowledge is to be explicated in accordance with such a picture. A nonrestrictive conception of knowledge is one which rejects 1

I discuss this picture in more detail in Conant (1991).

Kant’s Critique of the Layer-Cake Conception of Human Mindedness 123 such an interpretation of wherein the finitude of our cognitive capacity lies. To turn left with respect to this choice-point is to endorse a restrictive conception; to turn right is to endorse a nonrestrictive conception.2 Here then is a pair of formulations of what it might mean to read Kant as having a restrictive and a nonrestrictive conception respectively: (1) A fairly standard reading of Kant: Conditions of experience as restrictions All claims of necessity were, in Kant’s view, subject to conditions. For instance, particular claims of, for example, causal connection are always conditional on some prior state of affairs . . . , and general claims that objects conform to the requirements of the possibility of our experience are also subject to a condition – the condition, presumably, that we do in fact experience them. Kant’s general position on necessity would thus . . . suggest . . . that the conditions of the possibility of experience are restrictions on what we can experience. . . . The principles of thought which Kant’s argument will finally produce . . . have to be regarded as “restrictions” on those cases of consciousness which do in fact count as cognitive judgments. (Guyer 1989: 60–61; 1987: 153–54) (2) An alternative reading of Kant: To show that the subjective conditions of experience are not merely subjective Kant . . . aims to show that the requirements of the understanding are not just subjective requirements but requirements on objects themselves. (McDowell 2009b: 74) Either of these two ways of reading Kant can often agree with the letter of the other’s formulations of Kantian thoughts, while completely differing in its understanding of the underlying spirit of the letter. They can agree that Kant saw the very survival of philosophy as resting upon our being able to make sense of the possibility of there being subjective but necessary conditions for the possibility of experience, while differing on how to understand the idea of “forms of subjectivity” with which Kant works. On the restrictive conception, the dimension of subjectivity hereby introduced into an account of knowledge restricts or otherwise compromises the character of the objectivity to which such a form of knowledge can lay claim; whereas on a nonrestrictive conception, the relevant dimension of subjectivity that thereby comes into view is properly understood only once it has been appreciated that it is nothing other than a condition 2

No political overtones are intended in my description of the turn which I oppose with respect to each of these choice-points as a “left” turn and the one I condone as a “right” one.


j ames conant

on the possibility of the only conception of objectivity we are coherently able to think through and make sense of. Robert Pippin observes about such a conception of the conditions of knowledge that, if “the only way to make sense of such a subjective contribution was to accept that their status as subjective [conditions] was also a restriction and that therefore we were restricted to possible objects of ‘our’ finite experience,” then this immediately raises ”the issue of what sort of subjective restrictions these could be if not psychological” (Pippin 2005: 16). This raises the concern that any way of understanding what a restrictive conception of such conditions amounts to will in the end be faced with the threat of devolving into a philosophically hopeless form of subjectivism. More crucially, it gives rise to an entire tradition of Kant interpretation which turns on the assumption that the aim of the transcendental deduction is to move from some broadly psychological “facts” about the nature of our mindedness (how we “must” think) to a claim about the nature of reality (and how it “must” be). If one begins by understanding Kant’s conception of finitude to be a restrictive one, then it is almost impossible to avoid eventually sliding into (what I call) an impositionist reading of the first Critique – a reading according to which the categories of the understanding are taken to impose certain forms of unity on an exogenous matter; or, to put the same point the other way around: that the forms of the understanding are taken to be exogenous to and thus to leave unaffected the internal nature of that which is given to us in sensibility. Rather than exploring the problems with impositionism itself as a reading of Kant, what I shall do in the next section (Section 7.2.2) of this essay, instead, is first to bring out some of the further assumptions which contribute to making those problems urgent for most readings of Kant. 7.2.2 A Second Choice-Point in Reading the Deduction: Two-Stage versus Anti-Two-Stage Readings of the Relation between the Aesthetic and the Analytic Of the four choice-points that I will discuss here, the one we shall discuss next is the one which most immediately and obviously requires one to take a stand for or against a layer-cake conception of human mindedness. This second choice-point has to do with whether one should accept what I will call a two-stage reading of the first Critique. Most Anglophone readings of the work involve some version or other of a two-stage reading. What I want to bring out, however, is how they all participate in a common assumption

Kant’s Critique of the Layer-Cake Conception of Human Mindedness 125 and how they all make a common left turn with respect to a common choice-point. Here are three versions of such a reading: (1) The standard variant of the two-stage reading: Two temporally discrete stages in apperceptive consciousness: a first apperceptive stage in which a manifold of bare sensory consciousness is constituted, followed by a second stage in which it is then synthesized and brought into accord with the unity prescribed by the categories of the understanding. (2) The unconscious/conscious variant of the two-stage reading: Two temporally discrete stages the first of which is subapperceptive: in which forms of sensory input are processed but not yet brought to the level of consciousness, followed by a second stage in which they are brought to consciousness through the involvement of the categories. (3) The “logically but not temporally distinct” variant of the twostage reading: Two logically distinct and self-standingly intelligible moments of cognition which co-occur in actual sensory consciousness: a merely receptive moment of sensory uptake of that which is given and an intellectual moment in which what is given is apprehended as displaying forms of categorical unity. In my following discussion, I will focus mostly on (3), for it is the one which brings out most clearly what is the crucial and fateful presupposition shared by all two-stage readings. Option (1), however, is the one most widely explored in the Anglophone secondary literature on Kant. This requires that one read the Transcendental Aesthetic as treating of a form of conscious awareness of an object – the sort given to us in the sort of immediate singular representations which Kant calls intuitions – and as considering that form of awareness to be something we can enjoy through a self-standing exercise of our faculty of sensibility. Such a form of sheer sensory awareness is taken, on the first option, to be what is given to us first in the process of cognition. A proponent of the first option will generally then go on to read the Transcendental Analytic as introducing a further requirement on genuinely objectively valid representations of objects – one that comes into play when these elements of an episode of sensory consciousness are “brought under” concepts – where this business of bringing things under concepts is taken to be both a temporally and logically posterior stage in the cognitive process. Option (2) notices that the first option runs into a great many problems – both of a systematic as well as of an exegetical sort. It seeks to remedy


j ames conant

these by kicking the first stage of the cognitive process below the threshold of apperceptive consciousness. This allows one to preserve the letter of a great many things that Kant says in the Transcendental Analytic about how essential the categories are to our so much as being able to enjoy an intuition of an object, while preserving most of the spirit of the first reading. It allows one to praise Kant for being the father of cognitive science in his concern to describe the nature of the subpersonal levels of processing which are antecedent to the possibility of genuinely apperceptive consciousness of objects. It also, however, requires that one read a great many passages where Kant seems to be discussing forms of self-consciousness as actually being about forms of processing which fall outside the purview of self-consciousness. Option (3) involves an appreciation that pushing the first stage of a twostage temporal picture of cognitive processing underground does not really solve any of the problems which arise for the first reading, either exegetically or philosophically, while creating quite a few new ones of its own. Perhaps the most clear-sighted exponent of this subtlest variant of a twostage reading within the Anglophone tradition of Kant commentary is C. I. Lewis.3 Experience always comes to us as unity, Lewis insists – a unity in which the contributions of our sensible and intellectual faculties are inextricably intertwined. Moreover, according to him, the common error of both traditional empiricism and traditional rationalism is to attempt an impossible separation of the two. All such factorizing analyses of the phenomenology of experience – ones which attempt to isolate two temporally successive contributions in experience of each of these faculties – end up failing to do justice to the actual thoroughgoing unity of our experience. In all of these respects, Lewis is a forceful opponent of both of the first two variants of the two-stage reading. On the other hand, he insists that the categories (qua pure forms of understanding) and the given (qua sheer deliverances of sensory manifolds) must be independent of one another – that neither limit the other. This is a conclusion that he thinks is forced upon us through an act of transcendental reflection in which we consider what is required in order for our concepts about the world to be subject to a form of a genuinely external constraint which comes through our actual interaction with the world. Lewis’s opening moves are recognizably Kantian. He begins from the thought that all conceptual activity must in some sense come from us – that 3

This discussion of Lewis is indebted to McDowell (2015).

Kant’s Critique of the Layer-Cake Conception of Human Mindedness 127 it is a form of activity. He is here following Kant in taking it that all cognition must involve an element of spontaneity. But, according to Lewis, this requires that we also recognize a given element in experience – something that operates as a constraint from outside the sphere of conceptual activity – the element in knowledge which we, as knowers, must be able to take in without bringing something to it. Lewis’s thought here is that in order for knowledge of the world to be something more than the contemplation of our own reflection it must come from something genuinely outside us; and he takes himself simply to be echoing Kant in taking it that that must mean that cognition involves an element of passivity. For this element of passivity to make a genuinely independent contribution, it must be logically distinct – even if it is not temporally distinct – from that which is made by our faculty of spontaneity. What Lewis calls “the given” is that which is thus delivered up to the mind, furnishing it with content upon which to operate, thus rescuing its operations from emptiness and arbitrariness. What makes Lewis a proponent of the two-stage reading – his rejection of the standard variants of such a reading notwithstanding – is the manner in which he construes the aforementioned second condition on knowledge. Lewis not only declares something Kant would be happy to declare: “If there be no datum given to the mind then knowledge must be contentless and arbitrary; there would be nothing which it must be true to” (Lewis 1929: 38–39). But he also goes on to declare: “The pure concept and the content of the given are mutually independent; neither limits the other” (Lewis 1929: 37). He also insists that the given, qua merely given, must remain utterly uncorrupted by the concepts which we bring to it, on pain of our falling into a vicious form of idealism, in which we are no longer able to see our conceptual activity as constrained by anything from outside its own sphere. Thus he is committed to regarding the initial form of the manifold of sensory receptivity, considered as the sort of logical moment it is in the constitution of knowledge, as one which is independent of the form of conceptual knowledge. I will refer to any reading of Kant which rejects this common presupposition of the three variants of a two-stage reading sketched above an antitwo-stage reading of the first Critique. The initial characterization of what makes a particular reading of that work qualify as an anti-two-stage reading is purely negative. What all such readings have in common is merely this: that they think that the three ways of reading Kant sketched above are not only mistaken, but are ultimately mistaken for the same reason.


j ames conant


A Third Choice-Point in Reading the Deduction: Two Senses of the Term “Intuition” (and Related Terms)?

Consider the following quotation from Henry Allison: [A] tension, if not outright contradiction, has often been noted between the official definition of ‘intuition’ as a “singular representation” and the account of sensible intuition. The problem is that, according to Kant’s theory of sensibility, sensible intuition provides the mind with only the raw data for conceptualization, not with the determinate knowledge of objects. Such knowledge requires not only that the data be given in intuition, but also that it be taken under some general description or “recognized in a concept.” Only then can we speak of “representation of an object.” Kant gives clear expression to this central tenet of his epistemology in the famous formula, “Intuitions and concepts constitute, therefore, the elements of all our knowledge, so that neither concepts without an intuition in some way corresponding to them, nor intuition without concepts, can yield knowledge” (A50/B74). (Allison 1983: 67)

Allison is noting a problem which must arise for any proponent of a two-stage reading. Consider now the form of a solution to this problem which Allison himself endorses: The key to the resolution of this tension is well expressed by W. H. Walsh, who remarks that a Kantian sensible intuition is only “proleptically” the awareness of a particular. The point here is simply that, although intuitions do not in fact represent or refer to objects apart from being “bought under concepts” in a judgment, they can be brought under concepts, and when they are they do represent particular objects. In this respect, they differ from purely subjective or aesthetic “representations,” such as feelings, which can have no representative function. Thus . . . it is really necessary to draw a distinction between determinate or conceptualized and indeterminate or unconceptualized intuitions. (Allison 1983: 67–68)

Some version of this solution has become very popular in Kant interpretation. The solution to the problem is, in effect, to insist that one must distinguish two sorts of intuition – those that figure in the first layer of the layer-cake picture of sensory cognition and those which figure in the second layer, once those which figure in the first layer have come to be reshaped in the light of their interaction with our higher cognitive capacities. The first sorts of intuitions are nonconceptual modes of apprehending an object which require no involvement of the understanding. The second sorts of intuitions are those which only come into view for us once they have been informed by the categories. Put this vaguely, any of the three

Kant’s Critique of the Layer-Cake Conception of Human Mindedness 129 variants of a two-stage reading spelled out above could help itself to such a distinction between two fundamentally different sorts of intuitions and then spell out the details of how they differ from one another in accordance with the demands of each variant. Once one makes this move, one starts to see that it requires one to begin to make all sorts of further local distinctions that the text itself did not originally force on one, but which are now taken to flow from which sense of the term “intuition” is supposedly at issue in a given passage. So every time one encounters the term ‘intuition,” one is, in effect, now obliged to see it as involving an implicit subscript, since Kant’s topic must either be intuitions of the first sort or intuitions of the second sort, and to such an interpreter Kant is bound often to seem not to be very clear which of the two is at issue. It looks as if the task of the responsible commentator requires of him that he introduce distinctions all over the place where Kant failed to. For the commentator will now also need to distinguish two senses of “form of intuition” (hence two senses of “space,” two senses of “time”), two senses of “synthesis,” two senses of “manifold,” and so on, for virtually every major term of art in Kant’s epistemological vocabulary. Part of what creates this appearance of constant contradiction is the assumption that one can treat the initial introduction of a supposed term of art, such as “intuition,” as a definition of the term. A crucial exegetical assumption underlying two-stage readings of the text may thus come to light: the assumption that the prosecution of the task of the first Critique is one that is compatible with the work in question assuming the form of an ordinary treatise. The form of the traditional philosophical treatise comes from mathematics: one begins with definitions and unassailable propositions, and one proceeds to demonstrate further truths that follow from them. At any point in the traditional treatise, the truth of what has been shown up until that point does not depend upon what comes later and the truth of what comes later does depend on what has come before. This, in turn, requires that our understanding of the crucial terms which figure in such a demonstration be fixed once and for all at the start and that what is shown over the course of the demonstration be irrelevant to a proper understanding of their use. Consider the following passage from Kant: Whereas . . . mathematical definitions make their concepts, in philosophical definitions concepts are only explained. From this it follows: . . . That in philosophy we must not imitate mathematics by beginning with definitions, unless it be by way of experiment. . . . In short, the definition in all its precision and clarity ought, in philosophy, to come rather at the end than at the beginning of our enquiries. (A730/B758)


j ames conant

To take the right turn at the third set of choice-points is to reject the idea that a work of critique has the form of a mathematical demonstration and to take seriously that we fully understand the concepts that we deploy in philosophical reflection only once we have completed our task of critique. Thus, to stick to the example at hand, the notion of an intuition is initially glossed by Kant as an immediate singular representation. But this is then shown not to be a self-standingly intelligible form of representation – that its very possibility requires the involvement of a capacity whose exercise cannot be restricted to the production of such representation. 7.2.4

A Fourth Choice-Point in Reading the Deduction: The Relation between Subjective and Objective Unity of Consciousness

The tendency among many commenters on Kant is to think that Kant’s criticism of Hume is of such a sort that it is compatible with Hume and Kant agreeing on the following point – let’s call it the point of putative agreement between Hume and Kant: the form of consciousness which Hume takes to be initially available to us through the exercise of our sensory faculty is not sufficient to give us objectively valid representations of objects, but it is a form of consciousness that we could enjoy even if objectively valid representations of objects were impossible for us – merely subjectively unified consciousness is a self-standingly intelligible form of consciousness. (The difference between Hume and Kant is then taken to be the following: Hume thinks that what we add to what is thus initially given is a projection; whereas Kant thinks it is a form of representation which is objectively valid.) The fourth choice-point has to do with whether there is such a point of putative agreement between Hume and Kant. Lewis White Beck is a particularly lucid exponent of this way of reading Kant. I will therefore use him to illustrate what it means to turn left at this (fourth) choice-point. Beck distinguishes between two senses of the term “experience” in the first Critique: The opening sentences of the Introductions to both editions use the word ‘experience’ equivocally. In B we read: There can be no doubt that all our knowledge begins with experience. For [otherwise] how should our faculty of knowledge . . . work up the raw material of sensible impressions into that knowledge of objects which is entitled experience? In the first sentence, ‘experience’ means “the raw material of sensible impressions,” the manifold of apprehensions or Lockean ideas without the conceptual or interpretative activities of the mind. In the second sentence ‘experience’ means “knowledge of objects.” . . . Let us call these two meanings “Lockean experience” and “Kantian experience,” or, for short, L-experience and K-experience. (Beck 1978: 40–41)

Kant’s Critique of the Layer-Cake Conception of Human Mindedness 131 This leads him then to propose a very particular characterization of the role of the categories in carrying out the overall project of the Critique of Pure Reason: the task of the categories is to transform L-experience into Kexperience. Or as he puts it: “One way of reading the Critique of Pure Reason is to see it as an answer to the question: how do we move from L-experience to K-experience?” (Beck 1978: 41). Beck introduces, somewhat like Allison, two different conceptions of intuition to go with his two different concepts of experience: an inspectional conception and a functional conception. Let us consider Beck on the inspectional conception of intuition: The Critique begins with an inspectional conception of intuition and ends with a functional conception. According to this first conception, an intuition is a passively received inspectable sensory datum giving consciousness of an individual object independently of all categorization. It is given to consciousness ready-made and labeled. (Beck 1978: 41)

We can begin to see here how a left turn on this fourth choice-point is related to a left turn at each of the preceding three. As with each of the other left turns, this one implicitly commits the commentator to attributing the layer-cake conception to Kant. Beck says: “The inspectional conception of intuition is presupposed in the ‘difficulty’ raised in §13” of the Deduction (A84–92/B116–24). Beck therefore takes himself to have a crucial bit of text that shows that Kant himself is committed to this way of understanding his own text. This is what Beck takes to be crucial about the inspectional conception of intuition: “Given this conception of intuition, it is obvious that there could be intuitions which would not be tractable to categorial rules” (Beck 1978: 41). Beck takes §13 to be committed to this claim; this is why he thinks it supports his reading. Let us look at that section: That objects of sensible intuition must conform to the formal conditions of sensibility . . . is evident, because otherwise they would not be objects for us. But that they must likewise conform to the conditions which the understanding requires for the synthetic unity of thought, is a conclusion the grounds of which are by no means so obvious. Appearances might very well be so constituted that the understanding should not find them to be in accordance with the conditions of its unity. Everything might be in such confusion that, for instance, in the series of appearances nothing presented itself which might yield a rule of synthesis and so answer to the concept of cause and effect. This concept would then be altogether empty, null, and meaningless. But since intuition stands in no need whatsoever of the functions of thought, appearances would none the less present objects to our intuition. (CPR, A90–91/B122–23)

The question of the modality of the thought hypothesis here is crucial. It is worth noting in passing that the English translation does not do full justice


j ames conant

to the subjunctivity of the mood in which this entire possibility is entertained here. A difference between a right turn and a left turn with respect to our fourth choice-point can helpfully be summarized simply in terms of the issue how one understands the status of this possibility. There are two options here. The first option is to take the possibility under consideration here to be a fully intelligible one. Then Kant’s task is to show that, though things might be as this passage suggests – though the formal conditions of sensibility and the formal conditions of understanding might well be entirely orthogonal to one another in the manner described in this passage (so that in the series of appearances nothing presented itself which might yield a rule of synthesis) – fortunately, that is not how things stand. There is an argument – call it the transcendental deduction – which shows that this possibility does not in fact obtain. On this reading of Kant, there is a genuine gap between the conditions required for something to figure for us as an item in a series of sensory appearances and the conditions required for it to yield to a rule of synthesis – all that is required for it to meet the first set of conditions is that it be in space and time, whereas what is required for it to meet the second set of conditions is something entirely different – but it turns out that Kant can deliver an argument that guarantees that the items on either side of this gap are appropriately in accord with one another. I will call this reading of §13 the “Phew!” reading of the text. It turns on the idea that it makes sense to suppose that we might have been screwed and left with a mere blooming buzzing confusion of sensory appearances, but Kant succeeds in showing that – Phew! – this turns out not to be the case! To turn right at this choice-point is to reject the “Phew!” reading. It is to read Kant as seeking to show that the possibility that is entertained here in §13 is to be unmasked as a merely apparent possibility. The consequences of this right turn are explored in the next section.

7.3 Sections §20 and §21: Halfway through the B Deduction When we read the conclusion of the first half of the Deduction in the Bedition (§20), if we read the text the way Allison and Beck do, the following should be an urgent question: why doesn’t what Kant says at that point suffice to secure the conclusion of the Deduction as a whole? After all he takes himself to be able to conclude the following at that point: “All sensible intuitions are subject to the categories, as conditions under which alone their manifold can come together in one consciousness” (B143). What is missing? This looks to say that in order for something to meet the conditions on being an intuition it must be subject to the categories. But what does this mean? That is precisely the question that Kant himself, in effect,

Kant’s Critique of the Layer-Cake Conception of Human Mindedness 133 goes on to raise in section §21, which is a commentary on the previous section. Section §21 is titled an “Anmerkung” – an “observation.” Why is this commentary, this “observation,” there in the text? I take this question to be related to our second exegetical puzzle, pertaining to the relation between the A and B Deductions. There is an objection the B Deduction is structured so as to avoid. The objection in question is one to which the Deduction in A, on the standard reading offered of it in Kant’s time, appears to be open. The entire structure of the B Deduction reflects its effort to make perspicuous how a proper understanding of it allows it to avoid this objection. I do not take this to reflect any sort of retraction of the A Deduction on Kant’s part, but merely an attempt to rewrite it entirely, from beginning to end, with his focus on presenting it in such a manner that it is now made clear, once and for all, that if one’s reading of the Deduction continues to leave it looking vulnerable to the objection in question then one has utterly misunderstood its entire point. This is how John McDowell has put the objection in question: Kant wants to establish that experience has its objective purport in virtue of being informed by the pure concepts of the understanding. The objection is that that ensures only their thinkability. But a condition for objects to be thinkable is not thereby a condition for them to be capable of being given to our senses. Indeed (the objection goes on) the Transcendental Aesthetic has already supplied an independent condition for objects to be able to be given to our senses: they must be spatially and temporally ordered. For all Kant can show, objects could satisfy that condition for being present to our senses without conforming to the requirements of the understanding. (McDowell 2009b: 73)

Such an understanding of the Deduction naturally encourages an impositionist reading of its point: our initial access to objects has nothing to do with the forms of the understanding and these are then viewed as subsequently imposed on that form of access so as to allow that which we experience to be amenable to the conditions of thought. The requirement that intuitions have categorial unity looks to be something that the structure of our minds brings to experience in order to turn its deliverances into the sorts of things that can be true or false, but that (due to its thus merely “imposing” such a structure or unity on experience) its claim to genuine objective validity appears dubious at best. For it looks as if the unity here in question comes merely from the mind and has nothing to do with the nature of the objects which it enables us to think about. The B Deduction is rewritten precisely to invite such a reading and then to reject it.


j ames conant

From Kant’s point of view, if this objection were to go through, if it were true that the unity that the categories prescribe has nothing to do with the form of sensory experience as such, then the claim that Kant is seeking to vindicate through a Transcendental Deduction of the Categories of the Understanding – namely, that the pure concepts of the understanding have genuinely objective validity – would not only turn out not to have been vindicated at all, but, worse still, and utterly contrary to Kant’s most cherished intention in this work, the Deduction would have succeeded in showing precisely the opposite of what it sets out to show. For if the objection goes through, then the categories are shown not be valid for that which is given to us in sensory consciousness as such. If that is the case, then all the categories would represent are mere conditions on the thinkability for us of that which is given to us in such a self-standing form of consciousness. Kant organizes the way the entire B Deduction is written precisely so as to be able, first, to thematize this objection, and then to address it: to invite and then to repudiate it. The essential move, in rebutting this objection, is to deny the central assumption of the layer-cake conception of human mindedness. Or, to translate the assumption in question into terms that allow us to see how such a denial figures at the heart of the Deduction, we can put our point as follows: the Deduction is rewritten in such a way as to make it as clear as Kant possibly can that the Transcendental Aesthetic does not present us with a separate and independent condition for objects to be given to our senses. Here is the formulation – the Hauptsatz – that stands at the heading of §20 of the B Deduction: “All Sensible Intuitions are subject to the Categories, as Conditions under which alone their Manifold can come together in one Consciousness.” In commenting on this proposition in §21, Kant insists he is by no means done by this point: Thus in the above proposition a beginning is made of a deduction of the pure concepts of understanding; and in this deduction, since the categories have their source in the understanding alone, independently of sensibility, I must abstract from the mode in which the manifold for an empirical intuition is given and must direct attention solely to the unity which, in terms of the category, and by means of the understanding, enters into the intuition. (B144)

Dieter Henrich (1969) has been taken by many to show that a careful reading of the text requires us to accept that what this commentary aims to show us is that the overall form of the B Deduction is one in which we have a single proof, but that it requires two separate steps to carry it out. If the alternative to holding this is to claim that Kant takes himself to have

Kant’s Critique of the Layer-Cake Conception of Human Mindedness 135 already mostly completed his task by Section §20, and that what he does in the second half is simply to reprove the same proposition a second way or to prove some minor corollary in addition to it, then I suppose I agree with Henrich. However, the reading I favor is one for which the language of “steps in a proof” (and a privileging of the question of how to count how many of them there are) is misleading. The aim of the second half of the Deduction is to clarify what it is properly to understand the force of what is claimed at the end of the first half. Kant has placed this issue at the center of our attention by allowing a certain abstraction to be in place in the first half of the Deduction which is now lifted. Until now, in the first half of the B Deduction, we have considered the understanding in relation to a manifold of intuition, abstracting from the particular formedness of that manifold of intuition. Now we lift that abstraction and ask ourselves the question: how does this form of unity (that which belongs to our forms of sensibility) relate to that considered in the first half of the Deduction – that required for objectively valid judgment? Once the abstraction is lifted, the form of unity which had previously been treated in the Transcendental Aesthetic is to be reconsidered, now with regard to the question of its degree of accord or lack thereof with the form of unity which the categories prescribe. In particular, once the abstraction is lifted, we are supposed to come to see the pure intuitions of space and time (treated in the Aesthetic) in a new and proper light. In coming to see them in this light, what we come to see is that the crucial assumption of all two-stage readings is false. The second half of the B Deduction aims to show that the formedness of our sensibility, treated in the Aesthetic, cannot be in view independently of the form of apperceptive spontaneity, treated in the Analytic – even if the initial treatment of it had not yet revealed that its very possibility was subject to such a further condition.4 Here is how Kant himself puts what remains to be shown if a Deduction of the Categories is fully to attain its purpose: In what follows . . . it will be shown, from the mode in which the empirical intuition is given in sensibility, that its unity is no other than that which the category . . . prescribes to the manifold of a given intuition in general. (B144–45) 4

More specifically, Kant’s argument strategy in the second half of the B Deduction goes through the formal intuitions. The relation between the categories and the formal intuitions is clarified, and in light of the relation between any objects of our senses to the formal intuitions, the validity of the categories for all objects or our senses is demonstrated. So the crucial step involves showing that the unity constituted by conformity to the requirements of our form of sensibility, which is the unity of the pure formal intuitions of space and time, is not an utterly separate form of unity of manifold – one that could be in place altogether independently of the sort of unity of manifold that consists in our perceptual experience’s being informed by the categories.


j ames conant

This is Kant’s most precise statement of what the transcendental deduction of the categories is supposed to show. A full understanding of this passage involves seeing that it requires us to forego any of the aforementioned left turns at each of the choice-points canvassed above. On the other hand, it also leaves open a construal of it which does not require that we take a hard right at each of these choice-points. The point at issue here is perhaps most easily clarified by considering it in relation to the second choice-point reviewed above – that which involves embracing a two-stage reading or rejecting such a reading. What I will call a hard right turn inverts the fundamental claim of the two-stage reading and argues where the former holds there are two unities – sensible and intellectual – there is only one, that any distinction between these two forms of unity is merely notional. I take it that such a conclusion would also involve a misreading of the Deduction. Let us therefore distinguish a hard right turn from (what I call) a soft right turn. The soft right turn rejects the fundamental premise of two-stage readings – the self-standingly independent character of the unity of our forms of intuition from those of the understanding, while refraining from simply turning it on its head and thereby insisting on some form of thoroughgoing identity between these two forms of unity. Kant says that the former unity (the unity of the manner in which objects are given to us) “is no other than that which the categories prescribe.” We could try to encapsulate his moral here by summarizing it the form of the following catchy slogan: “There is only one unity!” This will not be false, but it does threaten a misunderstanding – one which might encourage a hard right turn. If considered at the appropriate level of abstraction – that which is in play in the first half of the B Deduction – there is a single form of unity with which both that treated in the Aesthetic and that treated in the Analytic are in accord. Kant’s term for this unity, considered at this level of abstraction, is the original synthetic unity of the understanding. This admits of two forms of further determination, one sensible and one intellectual. This form of unity – categorical unity – characterizes both the manner in which objects are given to us in intuition and the manner in which concepts are combined in judgments. But to say that it can be in act in these two different ways is not merely to identify the two forms of synthesis here at issue.5 There is one form of unity here which can show up in two different ways. I take this to be the point of the following famous passage: “The same function which gives unity to the various representations in a judgment also gives unity to the mere synthesis of various representations in 5

For further discussion of this point, see Land (2006).

Kant’s Critique of the Layer-Cake Conception of Human Mindedness 137 an intuition” (A79/B104–5). This unity grasped in its most abstract form is the original synthetic unity of apperception – this is Kant’s most abstract characterization of the unity of the understanding – it exhibits the structural feature which became so important to Hegel: any determination of it is one in which the unity of the whole remains prior to the unity of the parts. A synthesis of concepts in a judgment is one way of making this highest form of unity more determinate in cognition; a synthesis of a manifold into an intuition is another way of making this highest form of unity more determinate in cognition – both presuppose the involvement of the understanding. The threat of taking an overly hard right turn shows up in the overarching slogan with which McDowell summarizes the reading (which he and I both favor) of the B Deduction: “There is only one unity, common to the Aesthetic and the Analytic; not two separate unities” (McDowell 2009b: 74n10). This way of putting the moral is compatible with a central claim of McDowell’s earlier work Mind and World (1994).6 A central thesis of that book was the following: if we want to hold on to the idea that our conceptual capacities inform the exercise of our sensory capacities (an idea McDowell tries to show is obligatory if we wish to avoid the two horns of the dilemma outlined in that book), then this requires of us that we see that both sensory consciousness and judgment share one and the same form; this, in turn, requires (since we already know that the full-blown form of an exercise of our conceptual capacities in judgment is propositional in nature) that we conclude that the form of sensory consciousness is propositional. The emphasis on the propositional reflects a post–linguistic turn way of putting matters. The corresponding misreading of Kant which I am concerned to ward off here is the following: since the categories inform the exercise of our sensory capacities, this requires of us that we see that both a synthesis of a manifold of intuition and a synthesis of concepts into a judgment involve the same form and thus that we conclude that the form of our sensory deliverances is judgmental. One can take on board the point which Kant aims to demonstrate in the second half of the B Deduction (that the unity of the manner in which objects are given to us is no other than that which the categories prescribe) without losing a handle 6

McDowell himself has since retracted precisely this feature of the doctrine of Mind and World. See McDowell (2009c: 260) for his more recent account of perceptual knowledge: “what we need is an idea of content that is not propositional but intuitional, in what I take to be a Kantian sense.” (Note that intuitions in this sense, for McDowell, still draw on our conceptual capacities.) McDowell’s concerns in that essay are primarily systematic and not exegetical, but – as the final clause of that quotation indicates – he would now also retract the ascription to Kant of this feature of the doctrine of Mind and World.


j ames conant

in this way on the difference between the manner in which the categories are involved in sensible and intellectual synthesis respectively. Now let’s look at what else Kant says in Section §21 about what it would mean for a Deduction of the Categories to have fully attained its purpose – and not merely to have stopped at a point when it was only halfway completed. Here is Kant’s own alternative formulation of what remains to be shown in the second half: Only thus, by demonstration of the a priori validity of the categories in respect of all objects of our senses, will the purpose of the deduction be fully attained. (B145)

Notice: “all objects of our senses.” Thus Kant offers two formulations of what needs to be shown in the second half of the Deduction: (1) the unity of our mode of intuition is no other than that which the categories prescribe, and (2) the categories are valid for all objects of our senses. Part of what needs to be grasped is how these two ways of formulating what needs to be shown in the second half of the Deduction come to the same thing. A full appreciation of this point is incompatible with a reading of Kant which attributes to him a commitment to the layer-cake conception. If Kant can show that the categories are valid for all objects of our senses then the objection which he seeks to forestall no longer can arise. Hence Kant’s insistence on the second way of putting his point here. The aim of the second half of the Deduction comes to be slightly reformulated in Section §26, when Kant looks back on what he has accomplished. Putting things slightly differently, there he says that what needs to be shown is that he is entitled to the claim that the categories apply to “whatever objects may present themselves to our senses.” If he can show this, then he will have averted the risk that figured in the objection – he will have shown that categorial requirements, which appear to be mere subjective impositions on the independent form of our senses are in fact partially constitutive of the possibility of sensory consciousness in the first place.

7.4 Conclusion In the central early modern controversy, the empiricist and the rationalist disagree as to which of two cognitive faculties – sensibility and understanding – should be given logical priority in an account of the epistemic credentials of knowledge. As against both the empiricist and the rationalist, Kant wants to argue that the terms of their debate rests on a shared common assumption: that the capacities here in question – qua cognitive

Kant’s Critique of the Layer-Cake Conception of Human Mindedness 139 capacities – are self-standingly intelligible. I have focused on Kant’s argument here against the empiricist. A full account of his argument strategy here, however, requires that one come to see how a reciprocal moral is to be drawn from his critique of the rationalist. The aim of the Deduction is one of making sense of each capacity (sensibility and understanding) in the light of the other. For the front of the argument which is directed against the empiricist, this means coming to see how the standard assumption of the two-stage reading which takes the Transcendental Aesthetic to give us the full story about the nature of our faculty for sensory apprehension is mistaken. For the front of the argument which is directed against the rationalist, this requires coming to see how a mere inversion of the central claim of such a reading would be equally wrong. It would require seeing how a discursive faculty of understanding able to traffic in nothing more than empty concepts would no more amount to a genuinely cognitive power than would a faculty of intuition able to traffic in nothing more than blind intuitions. That is, it requires seeing how each of these faculties depends on its relation to the other to be the sort of faculty that it is in a finite rational being.

c h a p ter 8

The Critical and “Empty” Representation “I Think” Patricia Kitcher


The Puzzle

The Critique of Pure Reason is a study of the necessary conditions for, and inevitable limitations of, knowledge. Kant casts the issue in terms of delineating the necessary features of creatures that, like humans, cognize through concepts. Much of the obscurity of the book stems from his difficulties in presenting his novel and complex theory of a concept using cognizer. The goal of this chapter is to remove some of the barriers to understanding that theory, by trying to resolve a tension in his claims about the expression “I think.” Kant presented his theory of the cognizer twice, once in the “A Deduction” of the 1781 edition and a second time in the “B Deduction” of 1787. Although the A Deduction does not use the expression “I think,” it discusses the representation “I.” The context is the introduction of the thesis that all discursive cognition requires the unity of a cognitive subject. He appends a note elaborating the thesis and connecting it to “I”: All empirical consciousness, however, has a necessary relation to a transcendental consciousness . . . namely the consciousness of myself, as original apperception. . . . The synthetic proposition that every different empirical consciousness must be combined into a single self-consciousness is the absolutely first and synthetic principle of our thinking in general. But it should not go unnoticed that the mere representation I in relation to all others (the collective unity of which it makes possible [or whose collective unity make it possible]) is the transcendental consciousness. (A117n)

What is the relation between the principle that “any representation that can represent must belong with others to a single consciousness” and the “mere representation” “I”? Kant’s view seems to be that “I” stands for the condition, “transcendental consciousness,” that the principle claims is required for cognition to be possible. This is a striking claim. It implies that if someone says, e.g., “I see a dog,” then she is not just reporting on a passing 140

The Critical and “Empty” Representation “I Think”


event in her mental life. Rather, she is asserting an objective state of affairs – there is a dog that she is aware of through her visual system – and asserting that the visual image(s) through which she is aware of the dog belongs with others to a single cognitive subject or mind. The B Deduction presentation of the necessary conditions for cognition highlights the expression, “I think.” The I think must be able to accompany all my representations. . . . That representation that can be given prior to all thinking is called intuition. Thus all manifold of intuition has a necessary relation to the I think in the same subject in which this manifold is to be encountered. But this representation is an act of spontaneity; i.e., it cannot be regarded as belonging to sensibility. (B131–32)

The last sentence alludes to the origin of the representation “I think.” It is not received through the senses and so a posteriori, but produced spontaneously by the subject’s (mental) activities and thus a priori (B1–2). In A, Kant claims that different (conscious) representations must belong to a single “transcendental consciousness” that is indicated by “I.” He weakens the claim in B: all representations that could represent anything must be able to be accompanied by the expression, “I think.” Since this necessary reference of the representations is to a possible “I think” in the same subject in which the varied representations are found, the connection of “I” to a unitary consciousness is preserved in B. Furthermore, the A-edition doctrine that the highest principle of thought is that varied empirical consciousness(es) must be combined in one single self-consciousness is reiterated in its weaker form in B. All manifold representations of intuition . . . must be capable of being combined in one consciousness. (B136–37)

The relation between the expression “I think” and this “supreme” principle of the understanding seems to be the same as the relation between “I” and the “absolutely first” principle of thought. “I think” in B indicates the consciousness whose possibility the principle proclaims to be necessary for any cognition, viz., a unitary consciousness in which different representations are or can be combined. On this account, the use of the expression “I” or “I think” would be governed by the supreme principle of thought: “I see a dog” or “I think that is a dog” should be asserted only if a dogish perception or a judgment about a dog belongs to, is, or can be combined with, varied representations in a common subject. The view that “I think” must be used according to


pat r i c i a k i tc h e r

this rule fits with Kant’s general position that concepts are associated with rules for their use (e.g., A106). The Logic Lectures offer a more extensive account of concepts than the Critique. Here Kant maintains that conceptual cognition takes place through representations that make “marks” (characteristics) that are common to many things the ground of cognition (Jäsche Logic 9:58, LL 564; Refl 2288, 16:300 NF 41) and that cognizers regard concepts in this way (Refl 2285, 16:299). Although a layperson would not use terms such as “concept” or “representation,” she understands that in saying that something is “golden,” for example, she is saying that it shares a feature with other things that she has called by that name. For simple concepts, the rule governing use is similarity. Concept users can also use some (partial) concepts as the basis or ground for applying other (complex) concepts (Jäsche Logic 9:58, LL 564). So a concept user might call something a piece of “gold” on the grounds that it is golden, has a certain weight and is malleable. Again, ordinary cognizers would not speak of “grounds of cognition,” but they understand that color, weight, etc. are reasons for calling something “gold.” The rules associated with (complex) empirical concepts are not definitions, but “expositions,” i.e., incomplete analyses that give some marks of the concept, e.g., gold has a certain weight and color. Although rules can change, e.g., when new characteristics are found for gold, concepts must always be used in accord with some rule (A728–29/ B756–57). At this point, we are in a position to appreciate the importance of the expression “I think” to Kant’s system. Its use is governed by the principle that is the highest principle of discursive cognition, the principle that varied representations must be able to be combined in a unitary selfconsciousness. Given the relation between the expression and the principle, the status of “I think” as a necessary representation seems secure. It is thus somewhat shocking when the “Paralogisms of Pure Reason” chapter maintains that this critical (with both a capital and lower case “c”) representation may not even be a concept, but is in content for itself [a] wholly empty representation. (A345–46/B403–4)

How can a representation whose use is governed by the highest principle of cognition be empty of content? The “Paralogisms” chapter presents a critique of claims made by a group of philosophers Kant calls “Rational Psychologists.” According to Rational Psychology, the human subject who thinks is a simple substance that retains its identity through time and that can exist independently of any bodies

The Critical and “Empty” Representation “I Think”


(material things) (A344/B402). His diagnosis of the errors of Rational Psychology revolves around the emptiness of the representation “I think” that is the sole text of rational psychology, from which it is to develop its entire wisdom. (A343/B401)

That is the puzzle: both Deductions present “I” or “I think” as indicating a unitary consciousness whose possibility is required for all thought; yet the “Paralogisms” chapter claims that the emptiness of this representation is the key to explaining how Rational Psychologists mistakenly thought they could demonstrate the simplicity and immortality of the soul. To unravel the puzzle, we need an interpretation of “the transcendental unity of apperception” (“TUA”) that is required by thought and is indicated by the “I think” and an interpretation of the emptiness claim that illuminates the errors of Rational Psychology such that the two interpretations are not merely consistent, but mutually illuminating. These issues have been the subject of intense study by Kant scholars and scholars with deep Kantian affinities. I present three possible solutions. Section 8.2 lays out a line of interpretation that began with P. F. Strawson (1966) and has attracted many distinguished adherents. On this approach, the thinker who is the referent of “I think” is an embodied subject, a “person” whose attributes are characterized by both mental and physical predicates. In 1983, partly in response to Strawson, Henry Allison argued that the Critique’s investigation of the “transcendental conditions” for cognition is legitimate and that one of those conditions is an active or spontaneous thinker – a combiner or synthesizer of representations, who is conscious of her identity across different thoughts (1983: 142). The “thinker as active combiner” model has also been widely influential and Section 8.3 will consider a sophisticated version recently developed by Béatrice Longuenesse (2008, 2012). In the last section I present my own model of Kant’s thinker – not as an embodied person, nor as a power of combining, but as a set of representations united (or capable of being united) by necessary connections produced through the activity of cognizing through representations (Kitcher 2011). Longuenesse (2008: 30) suggests that the real work of exposing the errors of Rational Psychology is carried out by the original theory of the Deduction(s), with Kant simply applying the results of his analyses to the claims of Rational Psychology. I agree, so most of the following sections will be devoted to trying to understand what Kant indicates by “I think” in those difficult texts.


pat r i c i a k i tc h e r


The Self-Ascription of Mental States and the Embodied Thinker

How might the unity of self-consciousness be a necessary condition for cognition? Strawson begins by explaining what he understands by a single consciousness. Unity of consciousness to which a series of experiences belong implies . . . the possibility of the self-ascription of experiences on the part of the subject of those experiences. (1966: 98)

He takes Kant’s argument for the necessity for cognition of the possibility of self-ascription of states of consciousness to a subject to revolve around two distinctions that were fundamental to the Critique: that between concepts and intuitions and that between subjective and objective. Strawson invites us to consider a state of the subject – a momentary tickling sensation. Since a sensation does not exist apart from the experience of it, how is it possible to make sense of an experience recognized through the concept “tickling sensation” as opposed to merely a tickling kind of awareness? How can the “recognitional component” avoid being “absorbed into” the item recognized? He thinks that Kant saw the way through the problem: The recognition component, necessary to experience, can be present in experience only because of the possibility of referring different experiences to one identical subject of them all. (1966: 101)

As I understand Strawson’s view, it is that the required separation between an experience and the recognition of it through a concept can be guaranteed only if it is possible to refer the experience to a single subject of varied experiences. The possibility of referring experiences to a single subject implies, in turn, that at least some of the concepts under which particular experienced items are recognized as falling should be such that the experiences themselves contain the basis for certain allied distinctions: individually, the distinction of a subjective component within a judgment . . . collectively the distinction between the subjective order and arrangement of a series of such experiences on the one hand and the objective order and arrangement of the items of which they are experiences on the other. (Strawson 1966: 101)

Strawson’s idea seems to be that cognition of purely mental states such as sensations is possible only because the intuition/concept distinction can

The Critical and “Empty” Representation “I Think”


be preserved in their cases through the use of other concepts – concepts of objects – that enable subjects to recognize judgments as including a subjective side. Through their experiences of objects, subjects would have a notion of a subject of varied experiences to which purely mental experiences, such as sensations, could be ascribed. In Strawson’s (1966: 102) view, this is the force of Kant’s claim that the “I think” must be able to accompany all representations of a single subject. The intuition/concept requirement for cognition of sensations can be met only by appealing to the possibility of the self-ascription of sensations to a subject – which is possible only through the fact that human experience includes experience of objects. He expects critics to protest that Kant’s solution was too weak, because there is no determinate intuition of an “abiding self” of transcendental apperception through which the concept “subject of experiences” could be applied. To meet this condition, it would be necessary to identify a subject of experience as a man, which Kant clearly did not do (102–3). Still Strawson thinks that Kant took the most fundamental step in solving the problem, which is to insist on the self-reflexiveness of experience. Experience must be such that it provides room for the thought of experience itself (107). Strawson is clear that understanding the “I think” as referring to an embodied thinker represents an extension of Kant’s views. Still, it has been an extremely popular way to look at Kant’s contributions to the theory of cognitive subjects. Strawson’s reading of the “Paralogisms” chapter has been equally influential. Having provided an interpretation of the transcendental unity of apperception – it involves the possibility of the self-ascription of representations to a single subject – he uses that interpretation to lay out the precise criticism that Kant makes of Rational Psychology. Strawson takes Kant to hold that any use of the concept of a numerically identical subject of experiences persisting through time must have empirically applicable criteria of identity (163–64). This is not problematic if the subject of experience is a person. Confusion enters because, in the case of oneself, there are no criteria of identity invoked and that is the “root of the Cartesian illusion”: When a . . . subject of experience ascribes a current or directly remembered state of consciousness to himself, no use whatever of any criteria of personal identity is required to justify his use of the pronoun “I” to refer to the subject of experience. (164–65)

The idea of “criterionless self-ascription” of mental states derives from Wittgenstein’s (1958: 66–67) remarks about “I” in the Blue and Brown


pat r i c i a k i tc h e r

Books. Wittgenstein’s claim was that in the self-ascription of mental states, it is impossible to make an error through “misidentification of the subject” (Shoemaker 1968). In Strawson’s (1966: 165) example, it makes no sense for someone to think or say: “This feeling is anger, but is it I who am feeling it?” Given the impossibility of error, it makes no sense to talk of criteria of identity that might be used by the subject to ensure correct application of “I.” On Strawson’s interpretation, Kant understood this fact about the use of “I” and recognized that his predecessors’ failure to grasp it left them grasping at straws. When the ordinary criteria of personal identity are missing, then there arises a certain illusion: the illusion of a purely inner and yet subjectivereferring use for “I” . . . [and] it will appear that the object of this reference must be an object of singular purity and simplicity – a pure, individual, immaterial substance. (166)

The theme of the impossibility of error due to misidentification has been elaborated by Strawson’s successors. Gareth Evans (1982: 222) made a seminal contribution in arguing that the impossibility of error due to misidentification also occurs in the ascription of bodily states. Seemingly I cannot know directly and immediately that someone is moving and wonder if it is I. Assimilating bodily cases to the mental lends further support to the neo-Kantian view that the subject of experience should be understood as an embodied thinker. Longuenesse questions whether Evans’s (2012: 85ff.) extension of immunity to error through misidentification to bodily conditions does not show, instead, that there are two different uses of “I.” As she notes, bodily cases of immunity to error through misidentification depend on special sources of information that subjects have about their bodies that permit them to say, e.g., “I am in pain” without any possibility of failing to refer to themselves (95–96). But she thinks this is not a view that was on Kant’s radar. Conversely she thinks that Evans and this tradition are missing the use of “I” in “I think,” resting on that kind of self-consciousness which Kant claims – correctly in my [BL’s] view – is a condition for any use of “I” – the consciousness of myself as an agent of thinking and judging. (95)

Despite the wide reach of the view that Kant’s fundamental insight in the Paralogisms critique turns on the immunity to error through misidentification of the self-ascription of mental states, I think that Longuenesse is right. Many of the examples that Evans cites, “I am moving,” “Someone is feeling

The Critical and “Empty” Representation “I Think”


a piece of cloth, is it I?” etc. would be cases that Kant would count as knowledge of oneself as object (Longuenesse 2012: 95). Furthermore, Strawson (1966: 166–67) acknowledges that his interpretation of Kant’s insights in the Paralogisms is somewhat at odds with the text, which focuses on the necessary unity of consciousness rather than the self-ascription of thoughts.

8.3 The Thinker as Combiner Where Strawson (1966: 32) deliberately avoids Kant’s many discussions of the synthesizing or combining activities required for cognition, Allison (1983) takes combining to be central to the argument for the necessity of a single consciousness. He makes two points about the relation between a single consciousness and acts of synthesizing: synthesizing requires a single subject (of the synthesized representations) and cognizers can recognize themselves as single cognizers only through engaging in acts of synthesis. He invites us to consider the simplest possible case: where a subject has two representations, A and B, each of which is accompanied by a distinct awareness of “empirical consciousness.” . . . Clearly in order for the subject of both these thoughts to become reflectively aware of its identity, it must combine A and B in a single consciousness. (142)

Allison’s two claims are related, because the reflective awareness of the necessary identity of the I that thought A and the I that thought B is presumably a result of the recognition that the two thoughts could be combined in a third only if there is one consciousness that has access to both. We can see a difficulty with this approach by looking at Eric Watkins’s discussion, where he takes himself to follow Allison (2005: 277n52). Kant thinks that we can be aware of the self and its identity indirectly, that is, by being aware of the activity of the self when it connects its various representations and by then inferring that it is one and the same self that does the connecting. (278)

If Watkins is right about the upshot of Allison’s reading, then it attributes to Kant the implausible view that cognitive subjects’ knowledge of their identity is not direct, but inferential. Longuenesse has also made the synthesizing activities required by cognition central to her account of TUA. She draws on her detailed analyses of the syntheses that Kant took to be required for cognition to offer a fuller account of the unifying activities of the mind: all modes of cognition from perception to the principles of mathematical physics


pat r i c i a k i tc h e r depend on a unifying activity by which all mental contents (intentional objects of mental states) are bound together in one whole and become directed at a world of objects recognized under concepts. . . . The fact that we may eventually accompany the state of a mental content with the proposition “I think” is just the expression of this unified, reason-giving activity. (Longuenesse 2012: 89)

The function of the “I” in the Deduction is to express the identity of the subject that thinks a variety of thoughts about the objects of perceptual experience and commits himself to the consistency of his thoughts about those objects. (Longuenesse 2008: 17; see also 2012: 91, 94)

On this reading, TUA expressed by “I” is a complex condition required for cognition: particular unifying activities subject to the general constraint of a commitment to a consistent picture of an objective world and of a cognizer who occupies a standpoint on that world (Longuenesse 2008: 17). Longuenesse’s reading of TUA in terms of the unifying activities of apperception enjoys significant textual support. When Kant introduces the term “apperception,” he presents it along with “sense” and “imagination” as indicating a power of synthesis (A94). There are also passages where not just “apperception” but “TUA” refers to a power of combination: This transcendental unity of apperception . . . makes out of all possible appearances that can ever come together in one experience a connection of all these representations in accordance with laws. (A108)

On the other hand there are also passages suggesting that TUA is not a unifying power, but a condition that is achieved through the exercise of a unifying power.1 In [original apperception] everything is necessarily in agreement with the conditions of the thoroughgoing unity of self-consciousness, i.e., must stand under universal functions of synthesis, namely of the synthesis in accordance with concepts as that in which alone apperception can demonstrate a priori its thoroughgoing and necessary identity. (A112)

Kant’s use of “i.e.” makes it seem as though he is identifying TUA with combining according to universal (categorial) functions, but he is simply getting ahead of himself. That is the conclusion to be established: the necessary unity or identity of self-consciousness will be shown to be possible only because the manifold of intuition is synthesized ways that make the categories applicable to all objects of sense. 1

Longuenesse (1998: 68) notes some of these texts.

The Critical and “Empty” Representation “I Think”


The same point is made in B: As my representations . . . they must yet necessarily be in accord with the condition under which alone they can stand together in one universal selfconsciousness, because otherwise they would not throughout belong to me. (B132)

Categories must be used if different representations are to be able to stand together in one self-consciousness. In such passages, TUA is presented as something to be achieved, not as a power that might be required to achieve that unity. The idea that TUA needs to be created is also evident in the final paragraph of §16, where the point is made twice: Synthetic unity of the manifold of intuitions, as given a priori, is thus the ground of the identity of apperception which precedes2 a priori all my determinate thinking. (B134–35)

The final thought of the section repeats its crucial lesson: All representations given to me stand [under the original synthetic unity of apperception], but under which they must be brought through a synthesis. (B135–36)

If TUA were a unifying power, then it would not need to be grounded in the special character of intuitions nor created by the activity of synthesizing. Still, Longuenesse is surely right that the unifying activities of the faculty of apperception are a crucial part of the account of how objects of the senses can fall under the categories – and of how the necessary identity of apperception is possible. Longuenesse’s account avoids collapsing TUA into the synthesizing activities in accord with categorial functions that are supposed to produce it by appealing to a commitment to consistency. TUA involves a commitment to the consistency of one’s representations – which consistency is possible through only through categorial syntheses. She tries to bring out the importance of consistency in combining to the use of “I” by considering a logical proof. Using “I” here is making explicit the presence, throughout the conduct of the proof, of oneself as the agent generating and connecting together each step of the proof, and committed to providing more justification if challenged. . . . To think of the proof as one sequence of interdependent 2

The claim that the “I think” is a priori and so “precedes” actual thought is consistent with the claim that the unity of apperception is made possible through the unity [or combinability] of representations. Space is also a priori, yet Kant is clear that the representation of space does not precede sensory data, but is originally acquired through interaction with objects (see Discov 8.221–22, TP2 311–12).


pat r i c i a k i tc h e r steps, I must implicitly or explicitly be aware of myself as one and the same agent through, accountable for the sequence and connection of these steps. (Longuenesse 2012: 91)

For this reason the capacity to think of myself and refer to myself via the concept and word “I,” and the process of connecting my thoughts and representations in such a way that I can account for my reasons, are mutually conditioning. (92)

Although I agree with Longuenesse that the availability of the concept “I” and the activity of combining representations in, e.g., carrying out a logical proof, are “mutually conditioning,” what explains the connection is not a commitment to consistency, but something else. To see this, consider producing a logical proof. Longuenesse suggests a difficult proof, but simple proofs would also require the unity of apperception and it is unclear why there should be a difference in principle. If length is important, we can think of a proof that is an extended syllogism in Barbara running from the As to the Zs, where the cognizer concludes that all As are Zs. Having constructed the proof, the cognizer would know why she believes that all As are Zs – because she believed that all As are Bs, . . . and that all Ys are Zs. So she could give her reasons. But how does a commitment to consistency play any role? Her reasoning progresses, because she grasps the successive logical relations and those graspings bring her to the successive steps. A commitment to consistency would not help her see the logical relations if she does not, nor would its lack impede her reasoning if she does grasp them. Furthermore, it is not a commitment to consistency that enables the logician to give her reasons, but the fact that she consciously combined various representations, e.g., inferring all As are Hs, from all As are Gs and all Gs are Hs. I agree with Longuenesse that the availability of the representation “I” is tied to an activity that enables cognizers to give their reasons – but what matters is that the activity is a matter of conscious combining, not a commitment to consistency. Consistency with oneself seems unavoidable in each inference in the chain, since either you grasp the logical relations or you do not. In long stretches of reasoning, where the proof is not all held in the mind or remembered, and checking for consistency might be useful, Kant does not insist only on consistency with oneself. He presents “Always think consistently with oneself” as a maxim for thinking (Anthr 7:228, cf. CJ 5:293), but he also maintains that only the “logical egoist” fails to test his judgment against the understanding of others – because that is an indispensable touchstone for truth (Anthr 7:128). A cognizer might renounce

The Critical and “Empty” Representation “I Think”


the belief that all Gs are Hs (and hence that all As are Zs) either because she has recognized that it is inconsistent with other things she believes – or because it is inconsistent with things that many other apparently reliable cognizers believe. Longuenesse’s (2008) account of Kant’s diagnosis of the errors of Rational Psychology makes several points that are common to virtually all interpretations: Given the role of the “I think” in reflecting the act by which we bring about overall unity and consistency of our representations . . . there are ways in which we necessarily think of ourselves. . . . The features are described by Kant as . . . “merely logical” because they are merely the ways in which we think of ourselves or form a concept of ourselves . . . just by relating our representation to one and the same “I,” and thus independently of any corresponding intuition. (21)

She goes on to spell these out: since different representations are attributed to the same subject, and each of the representations is thought of as inseparable from the others, and there is just one identical subject through time, Rational Psychologists mistakenly think that the subject is a simple identical substance (22–23). All this seems right, but there is no special role for a commitment to consistency. As Longuenesse recognizes, and as will be clearer below, what enables subjects to understand themselves as having diverse representations is the (conscious) activity of connecting the representations.

8.4 The Thinker as Representations Connected by “Real Bonds” I draw on one piece of historical context and one feature of the introduction of TUA in A to offer a different reading of the doctrine and of the light it sheds on the errors of Rational Psychology. For years, Kemp Smith’s (1923/1962: 207) great commentary led scholars to believe that Kant had no access to Hume’s discussion of personal identity. That view has now been refuted, though it is an open question whether Kant’s knowledge of Hume’s view had a significant influence on his theory of the unity of the self (Kuehn 2001: 178–90). In Hume’s view, the legitimacy of the idea of a continuing self could be defended in two ways, either by tracing it back to a constant impression of a self or by finding a “real bond” or necessary connection across mental states (Hume 1739/1978, I.iv.2: 250, 252, 259). It is widely agreed that, like Hume, Kant denies any intuition of a constant self (see, e.g., A350) and also that the


pat r i c i a k i tc h e r

absence of an intuition is a reason why he characterizes the representation “I think” as empty (see A51/B75). If he accepted the Humean options for the legitimacy of the concept “I,”3 then he could defend its use only by showing, contra Hume, that diverse representations stand in relations of real or necessary connection to each other (where such a connection is understood in terms of existential dependence: B is really connected to A, just in case it is impossible for B to exist in the absence of A [Hume 1739/1978: Appendix 634; A178/B220–21, Prol 4.257]). In that case, the Deduction argument for TUA would involve two necessities: it is necessary for cognition that representations are necessarily connected to each other. The passage that introduces the expression “transcendental apperception” indicates a double necessity in the argument to come: The conscious of oneself in accordance with the determinations of our state in internal perception is merely empirical, forever variable. . . . That which should necessarily be represented as numerically identical cannot be thought as such through empirical data. There must be a condition that precedes all experience and makes the latter itself possible, which should make such a transcendental presupposition valid. (A107)

The last two sentences explain that what is to be shown about transcendental apperception is that it involves a necessary identity and that how that status is to be established is by showing that necessary identity is necessary for the possibility of experience or empirical cognition (B147). I also take the context of the introduction of TUA to be significant. In A, Kant has already discussed two syntheses, but he argues for the necessity of a continuing consciousness only in relation to the third synthesis, that of recognition in a concept. His introductory discussion of synthesis is confusing, because he presents it as blind (unconscious) (A77–78/B103). In that account, however, he distinguishes the synthesis that is involved in concept use as not belonging to the [blind] imagination, but to the understanding. A striking feature of his discussion of the third synthesis is his insistence that the synthesizing involved in concept application is conscious. Kant argues that conscious synthesizing is necessary for judging through concepts by the use of an extended example, that of counting.4 If, in counting, I forget that the units that now hover before my senses were successively added together by me I would not cognize . . . the number. 3 4

Kant disagrees with Hume about the possible sources of representations, see below. Although helpful, the counting case is unusual in that cognizers do not have to wait upon perceptions supplied by the world. They can simply construct the set of units in accord with the counting rule.

The Critical and “Empty” Representation “I Think”


. . . The word “concept” itself could already lead us to this remark. For it is this one consciousness that unifies the manifold that has been successively intuited, and then also reproduced, into one representation. This consciousness may often only be weak, so that we connect it with the generation of the representation only in effect, but not in the act itself, i.e., immediately; but regardless of these differences one consciousness must always be found, even if it lacks conspicuous clarity, and without that consciousness, concepts, and with them cognition of objects, would be entirely impossible. (A103–4)

Kant makes two points (relevant to my current project). The first is that a subject could not count if she keeps losing information. She must recall the first unit, the second, and so forth. The second point is the key for his claim about a single consciousness. It is not enough that the units are preserved and accessible. In producing the representation, e.g., judging “nine,” the subject must be conscious in doing so on the basis of the counting. If she could not be so conscious, then she would be unable to use concepts and so would lack all cognition. Absent conscious acts of producing representations through combining other representations, the counter would not be a rational cognizer, because she would not know the reason for judging “nine” – namely, the counting. As should now be obvious, Longuenesse’s example of logical proof is analogous to Kant’s counting example. That is why I claimed above that what enables the cognizer to give her reasons has nothing to do with a commitment to consistency and everything to do with the conscious activity of carrying out the proof. If rational cognition requires act-consciousness, consciousness in combining, however, then it also requires the unity of apperception, as I understand it. The subject is conscious in combining the counted off units in judging “nine.” That is how she knows the reason. In that case, however, she also understands her judging “nine” as dependent on her counting. Putting the point in philosophical terms that the ordinary person would not use, she recognizes her mental states of judging and of counting each unit as standing in the relation of “necessary connection,” because one state depends on the others for its existence. A layperson might say only “how could I know the number if I hadn’t counted?” It is important to see that this question is rhetorical. Someone who has mastered the numbers knows that there are other ways to determine the number of items. She might remember it, hear about it, or just estimate. Those alternatives are not to the point, however, because she used none of them in producing the judgment. The force of her question is that, in making this judgment she counted, and since that was all that was going


pat r i c i a k i tc h e r

on that was relevant to it, this judgment would not have happened had she not counted. Since Kant takes concept users to be rational – to know the reason for their judgment – he maintains that conceptual cognition requires conscious syntheses and that mental states stand in recognized relations of real or necessary connection, i.e., it requires the unity of apperception. On this account of cognition, recognition of the unity of apperception is not an inference from the (conscious) activity of rational cognizing. It is a necessary part of that activity – because it amounts to neither more nor less than understanding one’s states as necessarily connected. On this approach, we can see more clearly why having the concept “I” available and being engaged in mental activity where one can give the reason are, as Longuenesse maintains, mutually conditioning. I have considered only the A Deduction, but the doctrine is basically the same in B, albeit with the slight, but as we’ll see, important, weakening already noted. To repeat, the claim that all empirical consciousness must be combined5 in one single self-consciousness is weakened to the claim that all representations must be capable of being combined in one consciousness. (B136–37, my italics)

The B presentation is also different, because Kant argues from the other direction: not that concept application requires an identical cognizer, but that it is only through combining representations in cognition that subjects can be aware of their identities. The empirical consciousness that accompanies different representation is by itself dispersed and without relation to the identity of the subject. The latter relation therefore does not yet come about by my accompanying each representation with consciousness, but rather, by my adding one representation to the other and being conscious of their synthesis. Therefore it is only because I can combine a manifold of given representations in one consciousness that it is possible for me to represent the identity of the consciousness in these representations. (B133, my italics)

Again, through engaging in cognition a subject forges and recognizes relations of necessary connection across mental states that make them states of a single subject. 5

But see A116: “any such representations . . . must be at least capable of being connected in it [one consciousness].” So it is not clear whether this is a change or a clarification.

The Critical and “Empty” Representation “I Think”


If that is the Deduction argument for TUA, how does that illuminate the “emptiness” of the representation “I think” and so the errors of Rational Psychologists? Kant begins the “Paralogisms” chapter with an unusual claim about the “I think.” Although it is like the categories in being a priori and transcendental, it has no special “title” (A341/B399–400). Paul Guyer (1987: 41ff.) explains that the “titles” of understanding in Kant’s Nachlaß refer to rules that govern the concept of an object and are forerunners of the categorial principles. So if “I think” has no special title, it would have no principle governing its use. Seemingly, however, it has such a principle, viz., the principle that every representation must be combinable with others in a single self-consciousness. Why should this highest principle not be considered a “title”? My hypothesis is that Kant rejects a special “title” for the “I think” in light of his novel Deduction argument about the way in which the expression is used. It is not applied on the basis of an intuition of the “I.” Furthermore, it is not applied on the basis of any marks contained in representations, which is why it involves no content in the sense of “common characteristics” at all. Rather, “I think” can be used only when the characteristics contained in representational states fall under some rule for an object concept (or, in the case of logical proof, fall under principles concerning the form of the contents), thereby enabling the cognizer to make a judgment – and to forge and recognize a relation of necessary connection across the states. Despite being the highest principle of cognition, the rule governing the use of “I think” not an independent rule, but piggy-backs on the use of other rules. I take it that that is why the principle of the necessary unity of self-consciousness is not a proper “title,” why the “I think” is not a regular concept, but the “vehicle” of concepts, and why the conformity of a set of states to the principle does not disclose anything about the nature of thinkers (A341–42/B399–400, A345–46/B403–4). Although the unique way in which the expression “I think” is used does not make the Paralogistic inferences inevitable, it shows why they are likely. Only a philosopher who considers the necessary conditions for empirical cognition is in a position to see the true source of the ubiquity of the representation. And only a philosopher who sees the combination of sensory materials by the activities of the mind as a source of relational content of representations (the representations as necessarily connected) is in a position to agree with Hume that there is no intuition of an “I,” without falling into his mistake of denying that humans can legitimately use the representation “I.” Without this background, but with some appreciation that the “I” must always be present, it is natural to think that the ubiquitous “I”


pat r i c i a k i tc h er

is a permanent feature of representations, and further that it has a special simplicity – because no multiplicity [of characteristics] is contained in it (A354). I conclude by considering two lines of objection to interpreting Kant’s thinker as a set of states connected by real bonds, and a further objection that might arise from one of my replies. One problem with the interpretation is that it seems to imply that selves are “short” – e.g., the self that works through a logical proof or that counts (cf. Anscombe 1975: 58). Those stretches of states would stand in relations of real connection, but what about the representations of a single thinker who both counts and constructs a proof? As we have seen, Kant shifts in the B-edition from the claim that different representations must be combined in a single self to position that they must be able to be connected in a single self. Short selves are only a problem for the A-edition theory; with the emendation in B, the problem does not arise. States of judging “There are 9 units” and “All As are Zs” would belong to a single self if their contents can be combined in some further judgment. Still, the weakening from ‘must be connected’ to ‘must be able to be connected’ may seem to raise a different problem.6 Diverse states are able to be combined only if the same consciousness can access them both – or in Kant’s terms, only if the same consciousness can reproduce them both (A100–103). But now the account seems circular, since states would be accessible (reproducible) by one consciousness only if they already belong to the same consciousness. Alternatively, it is not combinability, but accessibility, that is doing the work in assigning mental states to the same or to different ‘I thinks.’ I address the two halves of the objection (since it presupposes accessibility, the combinability criterion is either circular or otiose) separately. Kant’s term ‘reproduced,’ which indicates ‘recalled,’ may raise worries about circularity in exactly the way that Locke’s appeal to ‘memory’ inspired charges of circularity. Since a person can only access, reproduce her own states, it is not possible to use accessibility or reproducibility as part of the explication of belonging to a single ‘I.’ As Sydney Shoemaker showed, however, this sort of problem with circularity can be gotten around. If ‘x can access state S’ presupposes that state S belongs to x, then define a relation of quasi-accessing (as Shoemaker [1970: 271–73] defined quasi-remembering) that is just like accessing, except that it does not presuppose the identity 6

I’m grateful to Peter Sperber and Timmy deGoeij for pressing me on this issue when I visited Paul Ziche’s research group at Utrecht.

The Critical and “Empty” Representation “I Think”


of the consciousness who originally had the state with the consciousness now accessing or reproducing it. My guess is that in appealing to ‘memory’ and ‘reproduction’ neither Locke nor Kant is using a notion that they take to presuppose identity of consciousness, but from a philosophical perspective, all that is needed is that this solution counters the charge of circularity.7 One way to see why combinability is not otiose is to recognize that if reproducibility were all that were needed, then Kant’s view would reduce to Locke’s, or offer a variation on Locke’s: the identity of a consciousness extends as far back as states which are accessible by the present consciousness. Kant’s argument is, however, that reproduction is not enough for cognition by concepts – for thinking. So a consciousness that could merely access and reproduce different mental states would not be a thinker; a fortiori, the states would not belong to a single ‘I think.’ To bring out this feature of his account, ‘combinability’ needs to be understood as going beyond accessibility: states A and B are combinable/connectable by real relations just in case there is one consciousness that can access both A and B and that can consciously combine their contents in some state C in such a way that it recognizes the dependence of the existence of state C on the existence of states A and B. A second line of objection concerns intuitions. How could intuitions be understood as consciously produced by the mind on the basis of other intuitions and so recognized as necessarily connected to them? We can begin to see how this prima facie counterintuitive claim – or a close relative – might be plausible by recalling Kant’s catalog of different types of representations. Intuitions are not produced by sensory encounters with objects, sensations are (A19/B34). Furthermore, intuitions are conscious cognitions (A320/B376) and only conscious representations must be united in a single self-consciousness (A117n). Although Kant stresses the given-ness of intuitions, he also seems to think that they must be received into consciousness. Anticipating arguments to come, passages in each Deduction claim that unless the categories 7

Shoemaker observed that the notion of ‘quasi-memory’ raises a question about criterionless selfascription. He wondered about ‘branching’ persons, where physically different humans could quasiremember the same event. In that case, quasi-remembering an event would not be immune to error through misidentification of the participant in the event. Shoemaker’s (1970: 284) reply to this challenge – that in such a world the notion of a single person would not have the significance it has in our world – would apply equally to a situation where some consciousnesses could access the states of many consciousnesses. Since that is not true in our world, a person could self-ascribe the states that she quasi-accesses – and connects – without any possibility of error due to misidentification in our world.


pat r i c i a k i tc h er

could be applied to objects of sense or unless the subject could unite different representations in a single self-consciousness, then putative intuitions or perceptions would not be allowed as such. Without the universal [i.e., categorial] functions of synthesis . . . [no] thoroughgoing and universal, and hence necessary unity of consciousness would be encountered in the manifold perceptions . . . these perceptions would then belong to no experience . . . [they] would be nothing but a bold play of representations, i.e. less than a dream. (A112, also partially cited above) The thought that these representations given in intuition all together belong to me means, accordingly, the same as that I unite them in one selfconsciousness, or at least can unite them therein . . . otherwise I would have a self as multicolored, diverse a self as I have representations of which I am conscious. (B134)

Dennett’s (1969) familiar distinction between subpersonal and personal levels of cognition can clarify Kant’s doctrine.8 At the personal level, human mental life is not “less than a dream,” nor do humans have selves as numerous as their representations. His claim is that the normal coherence of human consciousness is possible only because of subpersonal level activity of combining sensations in accord with functions of synthesis governed by the categories. The sort of personal level argument that is supposed to vindicate this claim is that, e.g., humans can understand their perceptions to be temporally coherent, to follow one another in a uniform time sequence, only if the contents of their perceptions fall under causal laws. I do not believe those arguments are successful, but my point here is interpretive. Kant accepts the arguments, so his claim in these passages is that personal level coherence of conscious life is possible only through such unconscious combining. By definition, ordinary cognizers have no (direct) knowledge of “subpersonal” combining. But that does not mean that they are clueless about relations of coherence across their perceptions or intuitions that such combining is supposed to enable. Consider a case of perceiving a ship moving down stream. Even where the perceiver makes no judgment, she is still a rational cognizer and consciousness is still crucial, because it provides her with some knowledge of why her perception is at it is. It is an acceptable perceiving, because it seems fine given other perceptions. It follows that a putative representation being allowable as a perception depends on the 8

I am grateful to Hannah Ginsborg (2013) for suggesting the utility of Dennett’s distinction in this context. I am not using it precisely as he did, but to flag the distinction between conscious and unconscious processes.

The Critical and “Empty” Representation “I Think”


existence of other perceptions, those that enable it to be received as a coherent perception. Furthermore, earlier (conscious) perceptions depend on the existence of later ones that render them coherent in order to avoid being rejected as real perceptions. As with concept application, we need to be careful about what is being claimed to be necessary. Consider drawing the winning number out of a bin. It seems obvious that you could see any of the numbers purchased. That does not mean that your perception of the number, say 708, is not necessary, however, given your preceding perceptions, and the fact that you did not draw out one of the possible alternatives. Some coherent continuation of your perceptions is necessary and, given that you did not take out any other ticket, it is necessary that you perceive the number written on that ticket, which is 708.9 Ordinary cognizers would not express themselves in terms of necessary connection or existential dependence. As with concept application, however, that does not mean that they do not recognize relations of dependence across their perceptual states. If challenged as to why they thought they perceived ticket 708, they might reply only “nothing was wrong.” They accept the perception at face value, because it fits seamlessly with others in a spatiotemporally continuous flow of conscious perceptions that are also coherent in other ways (e.g., no daggers vanishing into thin air). The simple reply “nothing was wrong” implies a great deal: had the perception not made sense to them, then they would have rejected it, so they understand, even if only in a general way, that what they unquestioningly perceive depends on what else they have been perceiving. In perception, cognizers do not recognize relations of dependence across states through consciously producing them, but through being conscious of a continuous flow of perceptions, which they understand would not be continuous, but halt, unless later states “made sense” in light of earlier states. In this way they recognize that the later and earlier perceptions must all fit together, be together. Kant does not spell out how cognizers think of their perceptions as more than a dream or of themselves as not just being a series of perceptions, but gives only an abstract philosophical account: they think of perceptions as necessarily belonging to a single self-consciousness. Given later arguments about spatiotemporal and other kinds of coherence (no vanishing substances), however, the above ways of filling in the positive view seem 9

Complications may be added about the soundness of the perceptual system, etc., but they do not alter the basic point that, as things were, cognizers would have to perceive ‘708.’


pat r ic ia k itc h er

plausible. With perceptions as with thoughts, the ability to see them as states of a single self-consciousness, and so to attach “I think” to them, rests on something else, viz. seeing each in relation to the others as a necessary part of a “normal” flow of perceptions. Let me add that the B-edition solves the “short self” problem for intuitions in a different way. “Short perceivers” belong to a human lifespan sized cognitive self, not because the contents of perceptual states could be combined, but because temporally separated flows of perceptions could be linked together in overlapping chains. This reply to the query about intuitions may prompt a further objection. Reading TUA in terms of relations of recognized real connection across intuitions adds unnecessary complexity. Why think that coherence (spatiotemporal and otherwise) of perceptions is the basis on which cognizers think their perceptions are fine – and thereby grasp that later perceptions depend on earlier ones and so are, as philosophers would say, “really connected”? It is simpler to agree with the Strawsonians that what makes different representations belong to a single self just are their relations of spatiotemporal and other kinds of coherence. Or, why not agree with Longuenesse that bringing representations to the TUA only involves a commitment to holding representations that are consistent and coherent with the unicity of space and time and the categorial principles? Since consistency and spatiotemporal coherence also play a role in the necessary connection account, what is gained by the further condition that the representations must be recognized as really connected with each other? As noted, several texts suggest that TUA is something that is achieved, made possible by human engagement with cognizable objects. Neither a Strawsonian person nor a unifying faculty is brought into being through the activity of cognizing. Furthermore, as Longuenesse (2008: 25) recognizes, Kant was no less committed to the existence of a thinker than Descartes. If a thinker is understood as a Strawsonian person, then she enjoys a robust existence. As Longuenesse argues, however, Kantian subjects are capable of thought independently of whether they are embodied. So what is the existing thinker? An essential teaching of the “Paralogisms” chapter is that thought provides no insight into the nature of a thinker. But thinkers know that they exist and have states, so how do they think about the relation between the two? Either they think of the states as belonging to something else on a substance model or they understand them as necessarily connected to each other or they have some other understanding of the relation. It is not clear that the third option is viable, however, if the goal is to explain how humans understand the relation between thinkers and their

The Critical and “Empty” Representation “I Think”


states. Karl Ameriks (2000: 64ff.) argues that Kant’s view is that cognizers understand themselves as a kind of substance, just not a spatiotemporal substance. That possibility remains open and has many attractions. If it is not embraced, however, then there seems no alternative to the necessary connection reading. Otherwise, cognizers would not only not know their natures as thinkers, they would not be able to think about themselves as thinkers with diverse states. Finally, the extra complexity of interpreting TUA in terms of necessary connections across conceptual and intuitive states enables the model to capture something essential to Kant’s view of cognition and cognizers. He often presents humans as rational in ways that other creatures with representations are not. When the B Deduction introduces the combination that is required by cognition and is “the subject’s self-activity” (B130),10 the implicit contrast is with the blind play of representations brought about through association. A 1789 letter to Marcus Herz explains that without the categories and the unity of apperception they [representations] could still (if I imagine myself to be an animal) carry on their play in an orderly fashion, as cognitive states connected according to empirical laws of association, and thus even have an influence on my feeling and desire, without my being aware of them. . . . This might be so without my knowing the slightest thing thereby, not even what my own condition is. (11.52, Corr 314)

Animals are led blindly by their representations, but humans are not. It is easy to dismiss such claims, because Kant knew little about animal cognition. If we abstract from that issue, however, we can focus on the contrast that he wants to draw. Rational cognizers escape the fate of merely being led around by their representations, not just because they have conscious representations, but because they consciously combine them and can be conscious of things going wrong in perception. In these ways, they can know the reasons for their judgments and know, in a general way, why they have the perceptions that provide the grounds for those judgments (and so are necessarily connected to them) – their perceptions are flowing normally. But knowing the reasons for their judgments and for the acceptability of their perceptions requires recognizing necessary connections across their representations. Real connections across representations 10

As in his introductory discussion of synthesis, this passage deals with both conscious and unconscious syntheses, because his claim is that in human cognition perceptual materials are combined by categorial functions – not association.


pat r i c i a k i tc h e r

are central to Kant’s theory of the self-consciousness required for cognition, because they are what permit humans to have some rational insight into the beliefs that enable them to negotiate the world. On the interpretation of TUA just presented, the relations of recognized necessary connection that enable humans to be rational cognizers are also the relations that constitute them as single cognizers.

c h a p ter 9

Kant’s Mathematical Principles of Pure Understanding Lisa Shabel∗

Kant’s “mathematical principles of pure understanding,” the Axioms of Intuition and the Anticipations of Perception, are the synthetic a priori judgments that establish the conditions for applying the categories of quantity and quality to the objects of possible experience. The principles of pure mathematics, by contrast, are the fundamental truths of the mathematical sciences. Both sets of principles comprise cognitions that are, according to Kant, intuitively certain, synthetic, and a priori, but the former mathematical principles of pure understanding are transcendental principles that ground the “possibility and objective a priori validity” of the latter scientific principles of pure mathematics. Moreover, the former principles of pure understanding “proceed from concepts to the intuition” while the latter principles of pure mathematics proceed from “the intuition to concepts” (A160/B199). My modest aim in this essay is to explore the relation that the mathematical principles of pure understanding bear to the principles of mathematics proper, while also exploring Kant’s very notion of a “principle,” whether of pure understanding, mathematics, or sensibility.

9.1 Introduction The positive part of Kant’s argument in the Critique of Pure Reason systematically describes our collective cognitive power to sense, perceive, represent, understand, judge, and thereby experience mind-independent objects. The phenomenal objects of our experience are fundamentally spatiotemporal, unified, substantial, real existences that stand in causal relations to one another, and to us. Crucially, such experiential objects are mathematically describable and subject to mathematical investigation and reasoning. Kant’s system for describing our power to experience and mathematically ∗

I am grateful to Emily Carson, Gary Hatfield, Dai Heide, James O’Shea, Eric Watkins and audiences at Loyola University Chicago, Haverford College, the University of Wisconsin and the University of Texas for helpful comments on this essay.



l isa sh abel

codify such objects is notoriously complex; I will describe the overall system just to the extent necessary to situate the relation mentioned above – between the mathematical principles of pure understanding and the principles of the scientific discipline of mathematics – within the context of Kant’s larger project. In the Transcendental Aesthetic, Kant presents the science of a priori sensibility, and establishes that space and time are the pure forms of sensible intuition. He shows that, as such, our original representations of space and time determine any and everything that can appear to us as an object of possible experience (but also that they determine nothing that is aptly described as independent of our receptive sensible faculty, i.e., things in themselves).1 In presenting this transcendental “science,” Kant identifies “principles of sensibility” and makes suggestive claims about how the principles of the pure but non-transcendental science of mathematics depend on space, time, and sensibility. (These dependencies will be discussed in more detail below.) In the Transcendental Analytic, Kant isolates understanding and “expounds the elements of the pure cognition of the understanding and the principles without which no object can be thought at all,” recognizing that such pure cognition of the understanding must be directed at objects already “given” in intuition (A62/B87). He uses the notion of “understanding” in at least two senses, as both the faculty of concepts, or rules, and more broadly as the power of judgment. The transcendental principles that this faculty of understanding delivers are the “synthetic judgments that flow a priori from pure concepts of the understanding”; that “ground all other cognitions a priori” (A136/B175); and that “are not themselves grounded in higher and more general cognitions” (A148/B188). Moreover, such a priori principles originate only under the special sensible conditions and constraints established by the schematism of the pure understanding. Kant’s system thus conceives us to be perceptually receptive to given objects by means of sensible intuition, while we are active thinkers and judgers by means of the twelve categories and the principles that “flow” therefrom. The first two parts of the four-part table of categories, Quantity and Quality, correspond to the “mathematical” principles of pure understanding, the Axioms of Intuition and the Anticipations of Perception. So, the principles included in the Axioms of Intuition “flow a priori” from the (schematized) categories of Quantity, and the principles included in the 1

For an overview of the arguments of the Transcendental Aesthetic, see Shabel (2010) and the many sources cited there.

Kant’s Mathematical Principles of Pure Understanding


Anticipations of Perception “flow a priori” from the (schematized) categories of Quality.2 It is interesting to note that Kant also takes the principles and judgments of mathematics proper to “flow” a priori from the representations of space and time conceived as principles of sensibility (B40). But this means that the “flow” metaphor, which is invoked to explain the dependence of a set of principles on a set of mental representations, operates differently in sensibility and in understanding: in the case of sensibility, and the Transcendental Aesthetic, the non-transcendental principles of mathematics3 flow from pure intuitions of space and time, conceived as transcendental principles of sensibility, whereas in the case of understanding, and the Transcendental Analytic, the transcendental principles of pure understanding flow from (schematized) pure concepts. This contrast highlights at least two different ways that Kant uses the notion of a “principle,” which will be discussed further. Kant claims that he calls the first two sets of the principles of pure understanding “mathematical” (thereby distinguishing them from the “dynamical” Analogies and Postulates) for a precise reason. The Axioms and Anticipations are “capable of an intuitive [as opposed to a discursive] certainty” and are, moreover, the principles through which the principles of mathematics proper “acquire their possibility”; they are thus titled “more with respect to their application than on account of their content” (A162/B202, emphasis added). This claim, when taken together with the results of the Transcendental Aesthetic, suggests that the principles of sensibility and the mathematical principles of pure understanding are both presupposed by any explanation of the possibility of the principles of mathematics proper. Understanding these transcendental presuppositions on the possibility of the principles of mathematics will help to explicate the value of transcendental principles. But note that the discipline of mathematics, including its foundations, comprises a body of decidedly a priori truths, which does not, according to Kant, have to “beg philosophy for any certification of the pure and lawful pedigree of its fundamental concept[s]” (A87/B120). One wants to know exactly how Kant thinks that the possibility of such fundamental mathematical truths, such as the axioms of geometry, depends on transcendental principles of sensibility and pure understanding, in 2


A discussion of the Schematism is beyond the scope of this essay, but note that the mathematical principles of pure understanding flow from a temporally determinate conception of the categories of Quantity and Quality. The principles of mathematics are not transcendental principles because, though synthetic and a priori, qua mathematical truths they do not count as necessary conditions on possible experience.


l isa sh abel

particular on the “mathematical” principles of pure understanding. Our route toward an explication of the Axioms and Anticipations will therefore be taken via this question. At this stage, two perhaps obvious points deserve special emphasis. First, Kant is very clear that the mathematical principles of pure understanding are fully distinct from the principles of pure mathematics. And second, that neither such principles of mathematics nor any principle of sensibility belongs, strictly speaking, in the system of Transcendental Analytic with the principles of pure understanding. In the introduction to the system of all principles of pure understanding, he writes: The principles of the transcendental aesthetic, therefore, according to which space and time are the conditions of the possibility of all things as appearances, as well as the restriction of these principles, namely that they cannot be related to things in themselves, do not belong within our confined field of investigation. Likewise the principles of mathematics4 do not constitute any part of this system, since they are drawn only from intuition, not from the pure concept of the understanding; yet their possibility, since they are likewise synthetic a priori judgments, necessarily finds a place here, not in order to prove their correctness and apodictic certainty, which is not at all necessary, but only to make comprehensible and to deduce the possibility of such evident cognitions a priori. (A149/B188–89)

Beginning with the latter half of the passage: the “principles of mathematics” are nonidentical to the “mathematical” principles of pure understanding and, indeed, are excluded from the internal system of the Transcendental Analytic. The principles of mathematics include such claims as that space has three dimensions; that between two points lies one (and only one) straight line, and that it is the shortest line between those points; and that the whole is greater than the part. So, Kant wants to say that the content of such principles is “drawn only from intuition,” indeed that such claims “flow” from a priori intuition in a principled way, and so are dependent in some primary sense on sensibility and the results of the Transcendental Aesthetic. Such claims are not themselves, then, principles of pure understanding flowing from the categories, and this allows us to distinguish principles of mathematics as a scientific discipline from the principles of pure understanding that Kant labels “mathematical.” Nevertheless, the mathematical principles of pure understanding do serve to “make comprehensible” and demonstrate the “possibility” of claims that are specific to the discipline of 4

Here Kant uses the phrase “die mathematischen Grundsätze,” which Guyer and Wood naturally translate as “the mathematical principles.” But it is clear from context that he is here referring to what he calls elsewhere “die Grundsätze der Mathematik,” so I have amended the translation, accordingly.

Kant’s Mathematical Principles of Pure Understanding


mathematics. So, part of what we want to understand is the precise sense in which the principles of mathematics are grounded in the mathematical principles of pure understanding. Proceeding to the first part of the passage, Kant demarcates a domain for the principles of pure understanding that excludes “principles of the transcendental aesthetic” or principles of pure sensibility, thus preserving the mutual exclusivity of the faculties at the level of cognitive, transcendental principles. Though he does not itemize a set of synthetic a priori judgments that count as principles of sensibility (as he does for the principles of pure understanding), Kant does indicate in this passage that space and time as forms of intuition provide sensible “principles” for perceiving all (but only) the appearances.5 As suggested before, it is something of a puzzle to describe the dependence of the principles of mathematics on both the principles of sensibility and the principles of pure understanding, seeing now that the two sorts of transcendental principles are systematically and mutually exclusive. The task, then, is to explore three kinds of principles – principles of mathematics, principles of sensibility, and principles of pure understanding – in order to achieve a better grip on what Kant thinks grounds and explains the very possibility of a priori cognition, and, in a narrower sense, on the role of the mathematical principles of pure understanding in his overarching system. I will begin with the principles of mathematics proper, which stand outside (or, at least, alongside) the complete system of transcendental philosophy. I will describe how these principles are understood to provide a foundation for mathematical reasoning, codifying our most basic mathematical representations. Next, I will discuss the sense in which such principles of pure mathematics might be thought to presuppose space and time as principles of pure sensibility. In these discussions, the sense of “principle” must be taken broadly to connote an origin or source for cognition, but not necessarily a fundamental truth expressed propositionally. The final task will be to specify the role that the mathematical principles of pure understanding play in grounding and explaining the possibility of the principles of mathematics, especially since the cognitive source or origin for the latter will already have been located in sensibility. If this can be achieved, then some insight will have been gained into the transcendental judgments known as the Axioms of Intuition and the Anticipations of Perception. 5

In the introduction to the Transcendental Aesthetic, Kant makes this explicit, saying that “there are two pure forms of sensible intuition as principles of a priori cognition, namely space and time” (A22/B36).


l isa sh abel


Principles of Pure Mathematics

Following Kant, I will discuss geometry as the paradigm example of a pure mathematical discipline.6 We tend to think of the first “principles” or axioms of geometry before the nineteenth century as comprising the five postulates of Euclid,7 or the five postulates together with the “common notions.”8 Kant himself cites various other examples of principles of mathematics: e.g., that between two points only one straight line is possible; two straight lines do not enclose a space; three points always line in a plane; a straight line is the shortest between two points; and space has three dimensions.9 Kant’s examples highlight the fact that Euclid’s geometry as we know it from the ancient Elements was augmented and amplified in early modern translations and editions: the “principles” of geometry were not taken to demarcate the smallest set of foundational claims from which all other mathematical truths could be logically derived. Parsimony and irredundancy were not virtues in identifying the self-evident “principles” of mathematics in Kant’s time. But the examples Kant cites can also be explained by the relationship that Kant takes to obtain between the principles of mathematics and the transcendental principles of sensibility, which I will discuss below. For Kant, the “principles” of mathematics include the origins or starting point of the practice of, in this case, geometry, if not also the propositionally expressed axioms for the strict logical derivation of its truths. And these origins or foundations of mathematical practice originate in the transcendental principles of pure sensibility. In order to see how principles of pure sensibility are conceived to provide a transcendental source of and explanation for the possibility of mathematics, conceived as a unique body of synthetic a priori cognition, we must take a rather expansive view of the originating “principles” of mathematical practice.



8 9

For a discussion of the mathematical landscape that Kant engaged, see Shabel (2003: Part 2). Note that, according to Kant, arithmetic does not have axioms or first principles, but might nevertheless be understood to depend on a geometric representation of discreteness, and also on the principles of sensibility that describe temporal relations. The Postulates are as follows: to draw a straight line from any point to any point; to produce a finite straight line continuously in a straight line; to describe a circle with any center and radius; that all right angles are equal; the parallel postulate. The common notions include the logical claims such as that things which equal the same thing also equal each other, and the whole is greater than the part. See A10/B14, A163/B205, and A239/B299. Also note that in reference to such first principles, Kant uses the terms “Grundsätze” and “Axiome,” whereas in reference to derived geometric propositions or theorems, he uses “Sätze” (A718/B746).

Kant’s Mathematical Principles of Pure Understanding


Notably, Kant claims that the axioms of geometry depend on the imagination’s ability to generate shapes.10 That is, the representation of shapedness in imagination – the capacity to imagine circles and squares and triangles as bits or regions of space – is a necessary condition on codifying and describing the features of extension that are grasped as self-evidently true of space. So, there is a sense in which Kant conceived the practice of representing shapes in space, of actively diagramming the parts and relations of space itself, as the fundamental act of geometric reasoning; the codification and description of these acts together with the geometric “elements” they produce are then counted among the principles of a theory of geometry. This way of thinking helps to clarify the connection Kant establishes between mathematical concepts and intuitions, when he says that mathematical cognition depends on the “construction” of its concepts: “to construct a concept means to exhibit a priori the intuition corresponding to it” (A713/B741). Constructibility is a special feature of the a priori representations that align in this distinctive way with sensibility: only pure sensible concepts, such as mathematical concepts, are constructible, since one can display the content of such a concept in a given pure intuition. One cannot so exhibit the content of a category, and so constructibility is not a feature of the pure concepts of the understanding. But, notice that the idea of constructing a mathematical concept is dual. According to Kant’s definition, to construct the concept one exhibits a priori the intuition corresponding to it. This might mean that one starts with a discursive definition of the concept (e.g., a rectilinear figure contained by three lines) and then “constructs” the content of the concept by exhibiting a triangle. But, Kant also means to suggest that one can literally construct or produce the mathematical concept itself by imagining a three-sided region of space. In other words, the conceptual representation or thought of a rectilinear figure contained by three lines can be conceived to originate in an act of imagining a triangular bit of space. Thus, the constructed concept or thought of an element of mathematics results from an act of imagination and a presentation in the sensible space of intuition. And, on this way of thinking, both the exhibited spatial diagram and the general concept that it exhibits are aptly described as having been “constructed.” This brief discussion of constructibility is meant to support the idea that the “principles” of mathematics, the very foundations of the discipline of 10

“On this successive synthesis of the productive imagination, in the generation of shapes, is grounded the mathematics of extension (geometry) with its axioms, which express the conditions of sensible intuition a priori” (A163/B204).


l i sa sh abel

mathematics that Kant aimed to explain, are not exhausted by propositionally expressed axiomatic truths, nor by concept-driven acts of instantiation and determination, but rather must be expansively conceived to include the original generation of basic mathematical concepts via acts of imagination that display intuitive content.11 The spatiotemporal form that things take – the shaped-ness of things, both spatially and temporally speaking – is imaginable a priori, and the acts of imagining such forms provide us with the “principles” of pure mathematical thought. On this way of thinking, there is for Kant a distinctively nonlogical aspect of the foundations of mathematics, the possibility of which will be guaranteed by transcendental principles of pure sensibility.


Principles of Pure Sensibility

This brings us to the question of how the principles of mathematics relate to the grounding principles of transcendental philosophy. Once we have identified how the principles of mathematics are related to the principles of pure sensibility, in particular, we will be in a position to understand the special role that the mathematical principles of pure understanding play in explaining the possibility of mathematics. Though Kant does not label any particular judgments the “principles of sensibility,” he refers explicitly to space and time as “principles,” and also makes clear in context that the status that space and time share as forms of intuition provides a “principle of sensibility” (A22/B36). He also claims that the principles of mathematics, such as that three straight lines enclose a space, “flow from” space and time in their capacity as principles of sensibility: the representation of space is exposited as a “principle from which insight into the possibility of other synthetic a priori cognitions can be gained . . . such cognitions actually flow from the given concept [of space]” (B40).12 One can interpret Kant literally when he says that space and time are forms of intuition that function as “principles” of sensibility. This would be to think of space in the sense of “principle” that connotes an origin, or 11


Interlocutors have helpfully noted that an opponent of this way of thinking might concede that the principles of mathematics depend on pure constructions but maintain that such imaginings are necessarily concept-governed. I wish to argue, by contrast, that the relevant imaginings are conceptgenerative in the mathematical case. This line of thought will need to be developed elsewhere in more detail. Thanks especially to James O’Shea for pressing me on this issue. Likewise, the “a priori necessity” of a representation of time is identified as the source of “axioms of time in general” (B47).

Kant’s Mathematical Principles of Pure Understanding


fundamental source, but not necessarily as one that is expressed in propositional form, or as a judgment about objects. In this sense, the principles of sensibility, space and time, are the origin for, or source of, the principles of mathematics (just as the principles of mathematics are the origin for, or source of, the body of derived truths that the discipline of mathematics comprises). Space and time provide the grounds for our receptivity to the elements of mathematical thought by yielding representations of determinate parts of space and time, or formal intuition13 of shapes and sizes. As the infinite and single wholes that provide the form for all sensible intuition, space and time serve both to warrant and constrain the representations of finite spaces and times, regions that are diagrammable and serve to exhibit and construct mathematical concepts. In other words, space and time as forms of intuition govern, in a principled way, the production of spaces and times: any sort of diagram that is used in mathematical reasoning, like a triangle or a set of strokes, is exhibited in a delimited space and time and provides a form for empirical intuition. Notice that these ideas can be accessed without using Kant’s technical terminology: conceive of some primitive (original, nonderivative) ability to represent bits of space as shaped, or to count off ticks of time, in a way that is conceptgenerative. It is these sorts of primitive acts that, for Kant, i) originate in space and time as forms of intuition; and, ii) serve as the origin for (or even themselves count among) the principles of the practice and discipline of mathematics. Consistently, Kant’s talk of space and time as principles of sensibility can be interpreted as identifying a fundamental judgment defended in the Aesthetic, which explains the possibility of derivative judgments of mathematics. On this interpretation, it is a principle of sensibility (and Aesthetic) that space and time are the only two forms of intuition, to which all sensible representation necessarily conforms. Principles of mathematics depend on the principles of sensibility understood in this sense in whatever way that a fundamental axiomatic truth of geometry expressed as a judgment depends on the transcendental truth that space is the form or structure of outer sense; the truth of the axiom depends on a construction or exhibition in the given pure intuition of space.14 13 14

See Onof and Schulting (2015) and the sources cited there for a discussion of the interpretation of Kant on “formal intuition.” For Kant, the truth of self-evident axioms is displayed or exhibited in pure intuition, though such an exhibition would not count as a “proof.” See B16, where Kant says, with respect to the example


l i sa sh abel

So, we can conceive space and time as “principles of sensibility,” and as such as a source of the principles of mathematics, either by thinking of space and time as the literal representational origin of the elements of mathematical thought (e.g., triangles are parts of space) or by thinking of the transcendental fact that space and time as the forms of all sensible intuition are presupposed by any and all mathematical truths. Either way of thinking of the “principles of sensibility” as a source of the principles of mathematics links the original intuitions of space and time to shaped and sized things via the most basic elements of mathematical thought and reasoning. In order to transition to a discussion of the mathematical principles of pure understanding, and their role in grounding the principles of mathematics, consider the following passage, from the “Systematic representation of all synthetic principles of pure understanding”: There are . . . pure principles a priori that I may nevertheless not properly ascribe to the pure understanding, since they are not derived from pure concepts but rather from pure intuitions (although by means of the understanding). . . . Mathematics has principles of this sort, but their application to experience, thus their objective validity, indeed the possibility of such synthetic a priori cognition (its deduction) still always rests on the pure understanding. (A160/B199)

Here Kant reiterates that the principles of mathematics are “derived from” pure intuitions of space and time, which can be understood as the source of those principles in the two senses described above.15 But, as Kant says in the latter part of the passage, for such principles to count as delivering objectively16 valid synthetic a priori cognition, and not merely formal descriptions of the space and time of pure intuition alone, the principles must be conceived also to “rest on the pure understanding.” That is, while



of the geometric principle that the straight line between two points is the shortest, that “help must be gotten from intuition, by means of which alone the synthesis is possible.” I am leaving it open exactly what additional “means” the understanding provides in the derivation of principles of mathematics. Minimally, one must generate mathematical concepts from intuition in order to form mathematical judgments. But note that to remain consistent with the view defended above, I am forced to emphasize Kant’s claim that what rests on the pure understanding is the application and objective validity of the principles of mathematics and deemphasize his parenthetical remark that understanding also directs their original derivation. Thank you to Emily Carson, Rachel Zuckert and James O’Shea for helpful discussion of this passage. (This point is related to the issue raised in footnote 12.) According to the ‘Stufenleiter’ passage (A319–20/B376–77), intuitions (unlike sensations) are objective, as representations. But such intuitions are not the constituents of objectively valid judgments about experiential objects unless they have a thinkable relation to the appearances, and so can succeed in referring to particular things that are thinkable in a general sense, such as when one judges that a particular given quantum has quantitas.

Kant’s Mathematical Principles of Pure Understanding


pure sensibility provides the cognitive basis for deriving the principles of pure mathematics as axioms that hold within the domain of formal intuition, pure understanding will make it possible to give such principles substantive, experiential content. If we can explain the reliance that principles of mathematics have on the conceptual apparatus of the Transcendental Analytic, and say what role the understanding plays in applying such principles, we will have addressed the question of the relation between the “mathematical” principles of pure understanding and the principles of mathematics, proper.17

9.4 Mathematical Principles of Pure Understanding The Axioms of Intuition and the Anticipations of Perception comprise “those [principles of pure understanding] on which the possibility and objective a priori validity of the [principles of mathematics] are grounded, and which are thus to be regarded as the principle of these principles” (A160/B199). Thus, Kant says explicitly that the principles of mathematics are themselves governed by transcendental principles, and so the possibility of the principles of mathematics is explained, at least in part, by the “mathematical” principles of pure understanding. Accordingly, the success of mathematical reasoning, and its basis in the original foundations of mathematics, depends on and presupposes the results of both the Transcendental Aesthetic, as discussed above, and the Transcendental Analytic. The relevant question is: in what sense do the mathematical principles of pure understanding ground and support the principles of mathematics, thereby making possible the practice and applicability of the mathematical sciences? In a passage that closes the introduction to the Analytic of Principles, Kant explains what transcendental philosophy and mathematics have in common. This passage gives us entrée into the role that the former plays for the latter. But the peculiar thing about transcendental philosophy is this: that in addition to the rule (or rather the general condition for rules), which is given in the pure concept of the understanding, it can at the same time indicate a priori the case to which the rules ought to be applied. The cause of the advantage that it has in this regard over all other didactic sciences (except for mathematics) lies just here: that it deals with concepts that are to be 17

A longer version of this story would include a detailed discussion of the threefold synthesis and the relations among intuitability, imaginability and thinkability; as well as a discussion of how Kant distinguishes the “synopsis of sense” from the deliverances of imagination and understanding.


l i sa sh abel related to their objects a priori. . . . [Transcendental philosophy] must at the same time offer a general but sufficient characterization of the conditions under which objects in harmony with those concepts can be given. (A135– 36/B174–75, emphasis added)

The mathematician is in possession of the concept as a product of her own constructive act in pure intuition, and so can determine a priori the things to which the concept applies, by comparing the relevant features of other intuitions to the content of the constructed concept.18 What Kant claims here is that the transcendental philosopher in possession of the categories of quantity and quality has a priori access to the things that can be conceived as quanta with qualities. Moreover, the Axioms and the Anticipations provide the explicit principles for determining the cases to which the “rules” given in the categories can (and “ought to”) be applied, and the “conditions” under which such determinations are made. The role that the mathematical principles of pure understanding play in grounding the principles of mathematics is informed by this parallel between mathematics and transcendental philosophy: while both mathematics and transcendental philosophy are equipped a priori to identify objects that are “in harmony” with their respective pure concepts, the mathematical principles of pure understanding establish meta-conditions on the application of mathematical concepts to their objects. That is, the mathematical principles of pure understanding will establish the transcendental conditions under which the principles of mathematics – conceived expansively to include fundamental mathematical representations as well as axiomatically true mathematical judgments – codify and describe sensible objects of experience. Kant says that despite not themselves being grounded in any “higher or more general cognitions” these mathematical principles of pure understanding are not “beyond all proof.” But they are demonstrated subjectively as the only possible source of objective cognition (A149/B188) and are thereby principles of the form of all experience and “general rules of unity in the synthesis of appearances” (A157/B196). As “mathematical” (and not “dynamical”) principles they are unconditionally necessary because they pertain immediately to the intuition of objects of possible experience (A160/B199). Putting these claims together, we see that there is evidence for the truth of the mathematical principles of pure understanding, and this comes via their indispensability as subjective sources of objective cognition, but that this evidence is immediate and not governed by further 18

It is nontrivial to say how this comparison is meant to work, and beyond the scope of this essay.

Kant’s Mathematical Principles of Pure Understanding


conditions. Moreover, though the mathematical principles of pure understanding “proceed from concepts to the intuition” and not “from the intuition to concepts,” as do the principles of mathematics proper (A160/B199), nevertheless the truth of the mathematical principles of pure understanding depends on their putting the categories of quantity and quality in an immediate and noncontingent relation to that which is given in intuition, in a sense analogous to the relation that holds between the pure intuitions of space and time and the possible objects that can be represented as taking such a form. The principle that flows from the category of quantity, and which yields the “Axioms of Intuition” is: “All intuitions are extensive magnitudes.” That which flows from the category of quality, and which yields the “Anticipations of Perception” is: “In all appearances the real, which is an object of the sensation, has intensive magnitude, i.e., a degree.”19 One way to think about these claims is as expressing the minimal formal conditions on an intuition’s being a constituent of an objective possible experience, and on the object of an intuition’s being a conceptualized object of thought. I will return to this way of understanding the claims after looking at Kant’s arguments. Kant’s “proof” of the principle of the Axioms of Intuition happens in the first paragraph. There he claims that an appearance cannot be apprehended “except through the synthesis of the manifold through which the representations of a determinate space or time are generated, i.e., through the composition of that which is homogeneous and the consciousness of the synthetic unity of this manifold (of the homogeneous)” (A161/B202). That is, the apprehension of an appearance requires that it be represented as circumscribed within spatiotemporal limits, which is to intuit it as filling “determinate space or time.” The act of synthesis that “generates” a determinate region of space and time a priori likewise serves to delimit the bounds of what is sensibly given and apprehended as appearance. Formally speaking, the parts of what is apprehended as appearance are homogeneous and so that which is apprehended has the same structure, or takes the same spatiotemporal “shape,” as the a priori representation of the determinate part of space or time that it fills. But, Kant also claims that representing an appearance with such a structure requires “consciousness” that its homogeneous parts are united as a whole, and further that this “consciousness” 19

These are the formulations from the B-edition, at B202 and B207, respectively. The A-Edition formulations are: “All appearances are, as regards their intuition, extensive magnitudes” (A162) and “In all appearances the sensation, and the real, which corresponds to it in the object (realitas phaenomenon), has an intensive magnitude, i.e., a degree” (A165).


l i sa sh abel

is achieved via the pure “concept of a magnitude (Quanti).” Sensible intuition of the determinate spatiotemporal form of an appearance corresponds to the conscious thought of that appearance conceptualized as a magnitude; this is what is implied by Kant’s claim that the “synthetic unity of the manifold of given sensible intuition” is the same as the “unity of the composition of the homogenous manifold” that is “thought in the concept of a magnitude” (A161/B203). The act of synthesis that delivers an intuition of a spatiotemporally determinate whole thus also delivers the thought of such a thing as unified under the pure concept of magnitude. So, all intuitions come to be conceptualized as magnitudes and, in particular, as extensive magnitudes, namely, those “in which the representation of the parts makes possible the representation of the whole” (A162/B203). From here Kant proceeds to make the case that the principle given in the Axioms of Intuition, that all intuitions are extensive magnitudes, serves to ground and explain the principles of mathematics, proper. First, he explains the role that mathematical representation plays in cognizing extensive magnitudes in space and time, using the example that one cannot represent a line, no matter how small, without drawing it, “i.e., successively generating all its parts from one point, and thereby first sketching this intuition”; he remarks that time is generated and represented in the same way (A162/B203). This allows him to reiterate the conclusion that anything that can be intuited in space and time necessarily is represented and apprehended as extensive, that is, as an aggregate of parts cognized in succession. Moreover, the axioms of the “mathematics of extension (geometry)” depend on the representation and apprehension of such spatiotemporal forms: geometry is founded on the imagination’s capacity for generating determinate parts or regions of space and thereby on the capacity for representing and classifying shapes, such as lines, triangles, and circles. So, the practice of geometry and the apprehension of appearances in intuition are ultimately based on the same fundamental act or operation of mind, namely, on the formal representation of wholes as aggregates of parts. This is why the axioms of the “mathematics of extension,” the science of geometry, just are what Kant means to identify with the phrase “Axioms of Intuition.”20 The principle that is set out in this section, that all intuitions are extensive magnitudes, is not itself an “axiom of intuition.” Rather, it is the principle that explains the possibility that there are axioms of intuition, that is, axioms of magnitude or quanta “as such.” These axioms 20

Despite the fact that arithmetic does not, according to Kant, have axioms, numerical formulas also play a role in grounding a procedure for cognizing extensive magnitude since one can only count up parts of an aggregate if one can represent the whole as a sum of its enumerable parts.

Kant’s Mathematical Principles of Pure Understanding


of intuition/geometry express the fundamental features of space and time that condition any and all sensible representation and apprehension, but that are themselves likewise governed by the principle stating that such axioms apply only to extensive magnitudes. So, the axioms of intuition are themselves the principles of mathematics proper, which flow from pure sensible intuition and ground the practice of the discipline of mathematics, while the principle of the axioms of intuition is a synthetic and a priori judgment of transcendental philosophy that flows from the (schematized) pure concept of magnitude and specifies the domain of (mathematically describable) things that the mind is prepared to apprehend and represent. Finally, Kant argues that this particular judgment of transcendental philosophy is what alone “makes pure mathematics in its complete precision applicable to objects of experience,” which he glosses with the claim that what geometry says about pure intuition is “undeniably valid” of empirical intuition (A165/B206). He gives two arguments. The first is negative: if empirically intuitable objects of the senses were not “in agreement with the [a priori] rules of construction in space,” then mathematics would not be objective. The second is positive: because the synthesis of pure spaces and times is the very procedure by which appearances are apprehended and outer objects intuited, whatever pure mathematics proves true of such pure spaces and times is necessarily also true of that which appears in space and time. Kant closes the section by raising his familiar retort against an objector who posits that outer objects of experience are things in themselves, and so not constrained by “the formal condition of our sensibility”: such a transcendental realist must give up a priori knowledge of mathematics in order to release the objects of understanding from their sensible constraints. The minimal condition expressed by the principle of the Axioms of Intuition thus seems to be that any intuitive representation that is to feature in a judgment of experience must be conceptualized as a quantum that is extended in a determinate bit of space and time. That is, any thing that we can intuit is necessarily thought to be sized. More specifically, any thing that we intuit must be represented and apprehended as a sized whole by virtue of a representational aggregate of its parts. For example, if I intuit an ice cube, it is by virtue of intuiting its six sides individually and composing them in my representation of the whole. But notice that the activity of the geometer is just the same, but with respect to imagining the pure shape of a cube (if not an actual cube of ice). That is, the geometer’s task is to determine the features of space that warrant and constrain the construction of shaped figures like cubes, and this constructive activity and the


l i sa sh abel

principles of mathematics that underwrite such activity is, as we have seen, further underwritten by the transcendental principles of sensibility. Kant now adds the idea that the principles of mathematics must also be underwritten by a transcendental principle of understanding if, for example, the geometry of cubes is to yield thoughts and judgments about actual cubic things: in our experience, things like ice cubes are represented as taking the cubic form delineated by the geometer because (and only because) all appearances are represented in intuition as extensive magnitudes. When conceived as judgments that express basic truths about the objects of experience that take spatiotemporal form, the principles of mathematics must be taken to depend on the claim that we necessarily intuit those objects as the very sorts of things, namely, extensive magnitudes, that can be treated in a mathematical science. Else, the principles of mathematics risk vacuity. This is as much to say that a principle of pure understanding as expressed by the principle of the Axioms of Intuition is a principle of the principles of mathematics (cf. A160/B199). For there to be a priori axioms of intuition, namely, pure principles of mathematics, those principles must ultimately be about the extensive magnitudes that appear to us. In the Anticipations of Perception, Kant makes a remark that helps to clarify the connection between the two “mathematical” principles of pure understanding, and reveal how the Anticipations of Perception will complement the Axioms of Intuition. He writes: All appearances whatsoever are accordingly continuous magnitudes, either in their intuition, as extensive magnitudes, or in their mere perception (sensation and thus reality), as intensive ones . . . appearance as unity is a quantum, and is as such always a continuum. (A170–71/B212)

The Axioms and Anticipations together make a claim about our cognition of objects by, in part, distinguishing between two ways that an appearance can be represented as a continuous magnitude, that is, as a unified, whole quantum. The principle of the Axioms of Intuition specifies the sense in which an appearance is intuited as a unity, namely, by being represented and apprehended as an extensive magnitude the form of which is describable by the truths of pure geometry. The principle of the Anticipations of Perception will specify the sense in which an appearance is perceived or sensed as a unity, namely, by being represented and apprehended as having an intensive magnitude. It is this claim, that “in all appearances the real, which is an object of the sensation, has intensive magnitude, i.e., a degree” that we now seek to explicate, in particular as it relates to the principles of mathematics, proper.

Kant’s Mathematical Principles of Pure Understanding


Kant opens the Anticipations with the reminder that the objects that we perceive and sense empirically are not exhausted by their formal spatial and temporal characteristics; it follows that the axioms of their intuition are clearly not sufficient to characterize them as objects of possible experience. As objects of perception, the appearances “also contain in addition to the intuition the materials for some object in general (through which something existing in space or time is represented), i.e., the real of the sensation” (A165/B207). He further explains that the real that is given in sensation is subjectively represented as a flowing magnitude, in the sense that any sensation can be conceived to start from nothing and gain intensity, and then diminish again to nothing. A subjective sensation thus has a qualitative magnitude that provides the material for a perception; Kant describes such a magnitude as “intensive” because it has a certain degree of intensity over a certain period of time. When such a subjective sensation is taken to correspond to a particular object of perception, which is apprehended and represented as extensively contained within formal and objective spatiotemporal bounds, the intensive magnitude of the sensation is attributed to the empirically real object of perception, which Kant thereby understands as having a particular “degree of influence on sense” (A165/B208). Kant’s argument for the principle begins with the observation that sensations are instantaneous and so cannot have extensive magnitude, since they are not apprehended by any synthesis of parts and also can be entirely absent in a given moment of time. But the impact of a sensation can nevertheless be described as having a size. In the Transcendental Aesthetic, Kant had defined sensation as “the effect of an object on the capacity for representation, insofar as we are affected by it” and specified further that sensation is the mechanism by which we empirically intuit appearances (A20/B34). He notes here that sensations can gradually increase and decrease, even diminishing to nothing, and he adds that “reality” is that aspect of the appearance that corresponds to what is delivered sensationally. Kant’s defense of the principle that the real has intensive magnitude depends on his positing a symmetry between a sliding scale of individual sensations (each of which can be subjectively experienced to have a certain power) and a correlated real perceptual object that is represented via empirical intuition.21 He writes: Hence between reality in appearance and negation there is a continuous nexus of many possible intermediate sensations, whose difference from one 21

This correlation is crucial to Kant’s argument, but exploring his justification for it is beyond the scope of this essay.


l isa sh abel another is always smaller than the difference between the given one and zero, or complete negation. That is, the real in appearance always has a magnitude . . . but not an extensive one. (A168/B210)

The “continuous nexus” of sensations describes a flowing sequence of possible subjective states that directly mirrors the intensive magnitude of its empirically real correlate. More generally, the degree to which an empirically real thing influences and affects us perceptually is quantitatively describable. The minimal condition expressed by the principle of the Anticipations of Perception thus seems to be that any empirically real object of perception must be conceptualized as having a measurable qualitative impact or degree of influence. That is, any thing that we can perceive as real is necessarily thought to have a certain quantitatively describable (if also fluid and variable) power to affect us. So, we can anticipate that any perception – any empirical consciousness that includes a subjective sensation – will have an intensity, or size of impact. For example, one can imagine the subjective sensation that results from touching an ice cube. The sensation of cold that one feels is not itself an extensive magnitude, as is the perceived shape of the ice, since as a singular sensation it is instantaneous and not objectively locatable between fixed spatiotemporal boundaries. But the sensation of coldness nevertheless has a quantifiable size, an intensive magnitude, the degree of which can be perceived to change as the ice melts upon being handled. Moreover, intensive magnitudes succumb to mathematical representation and treatment in accordance with a theory of analysis such as Newton’s method of fluxions: the sensational contents of our perceptual states – and thus also their corresponding empirical objects – take the form of flowing, continuous quantities and so are subject to the same mathematical analysis as the flowing, continuous quantities that are given in our original representations of space and time. As in the case of the Axioms, pure mathematical cognition ultimately depends on the features of space and time as quanta continua being quantifiable properties of “the real”; if not, the pure principles of mathematical analysis would have no purchase on the empirically real objects of perception.

9.5 Conclusion Kant’s discussion of exactly why the principle that the real has intensive magnitude counts as an “anticipation,” in the sense of an expectation, or prediction, illuminates the question of how the principles of pure understanding can be thought to ground the principles of mathematics. He

Kant’s Mathematical Principles of Pure Understanding


says that an “anticipation” is any cognition “through which I can cognize and determine a priori what belongs to empirical cognition” (A166/B208). Now, specific subjective sensations are strictly empirical, and so cannot be anticipated a priori at all. But, the “real, which is an object of the sensation,” namely, the material aspect of a perceptual object conceived in a general sense, can be a priori anticipated to possess a degree of influence on a perceiver, lest such perceptual objects be inert and unaffecting things. Kant thus thinks that we have an a priori justifiable expectation that perceptual objects have such intensive magnitudes, or degrees of sensational influence; such magnitudes can be manipulated mathematically via numbers or symbols that represent a continuous series of sensation-producing forces. Likewise, “pure determinations in space and time” serve as a priori representations of the shapes and sizes of the objects of experience that are given a posteriori. Such a priori representations of the elements of geometry thereby provide a priori anticipations of the forms that objects of experience will take. This parallelism, between the two ways in which experiential objects are represented as having magnitude, is why Kant suggests that the principle of the “Axioms of Intuition” might have been called the “Anticipations of Appearances”:22 the Anticipations of Appearances and the Anticipations of Perception would then together a priori anticipate that everything that is given to sensibility objectively as a spatiotemporally appearing perceptual object is represented as having both extensive and intensive magnitude. The former “anticipations” would then be thought to ground “axioms of intuition” so-called, the principles of mathematics that treat extended magnitudes, while the latter “anticipations” ground whatever principles of mathematics treat intensive magnitudes, or continuous quantities qua flowing quantities, presumably the principles of a theory like Newton’s calculus. In the Doctrine of Method Kant writes: To decide about everything that exists (a thing in space or time) whether and how far it is or is not a quantum, whether existence or the lack thereof must be represented in it . . . all of this belongs to rational cognition from concepts, which is called philosophical. But to determine an intuition a priori in space (shape), to divide time (duration), or merely to cognize the universal in the synthesis of one and the same thing in time and space and the magnitude of an intuition in general (number) which arises from that: that is a concern of reason through construction of concepts, and is called mathematical. (A724/B752) 22

Kant mentions this at A167/B209.


l isa sh abel

In the latter part of this passage, Kant explains that mathematical cognition through the construction of concepts is grounded in pure sensibility, via a determination of an intuition of space, time, or magnitude. This is what it means to say suggestively that the principles or elements of mathematics “flow from” the principles of sensibility; pure sensibility thus provides a transcendental warrant (and constraint) on what can be constructed, and thereby perceived and understood, in a mathematical sense. At this stage of explanation, Kant is concerned with the relation between pure mathematical cognition and the forms of sensible intuition, but not yet with the empirical things that take such form. The principles of mathematics at this stage describe our sensible capacity to represent pure parts of space and time with shapes, moments, and numbers, all of which are presented to the self as particulars, diagrammable in a priori intuition and conditioned by space and time themselves. In the first part of the passage, he discusses the stage of explanation that addresses the sorts of sensible and thinkable possible things that take up space and time, that take shape, have duration, and can be conceived to have size, and which have a certain potential impact on sensation. Here the principles of mathematics are ultimately descriptive of empirically real spatiotemporal objects about which we can think and make judgments; these judgments are themselves conditioned by rational, philosophical, transcendental principles, as are all synthetic and a priori judgments. While a principle of pure sensibility, space as pure intuition, can warrant and constrain our representations of pure shapes and sizes, a principle of pure understanding is required to warrant and constrain our thoughts of the things that are empirically given to sensibility as shaped and sized, as having extensive and intensive magnitudes. So, while we can mathematically diagram a cube as a solid figure contained by six equal squares, a particular experiential object like an ice cube is, say, 1 cubic inch and freezing cold. The judgment that the ice cube is 1 cubic inch and freezing cold is governed, transcendentally speaking, by the principles given in the Axioms and the Anticipations. That is, in order for the principles of mathematics to be about empirically real things given to sensibility via sensation, such things must be conceptualized as qualitatively given magnitudes, or magnitudes with qualities that can be sensibly perceived. They must therefore be thought under the pure concepts of quantity and quality, and conform to the rules specified in the Axioms and Anticipations. One common way to think about the “mathematical” principles of pure understanding is as uber-rules for the application of pure mathematics to what is given in sensibility. On this way of thinking, the

Kant’s Mathematical Principles of Pure Understanding


Transcendental Aesthetic demonstrates the possibility of pure mathematics, while the Analytic secures its application.23 What I have proposed can be reconciled with this common interpretation, but I mean to have complicated the picture somewhat. When we distinguish the roles that the Aesthetic and the Analytic play in grounding the principles of pure or applied mathematics, the emphasis ought to be on the role that the mathematical elements play in, first, structuring the sensible as given to intuition and, second, structuring the domain of conceptually describable “objects.” In a passage from the Analogies of Experience, Kant claims that the Axioms and the Anticipations are “constitutive” because they explain how intuitions and the real in their perception are “generated in accordance with rules of a mathematical synthesis, hence how . . . numerical magnitudes . . . could be used” (A179/B221). Thus, the Axioms and the Anticipations are a priori guides for the determination and conceptualization of the appearances as magnitudes. So it is apt to conclude that these principles do serve to justify the application of mathematics to appearances, but in the special sense of making it possible to deploy the basic elements of mathematics, our sensible representations of pure shapes and sizes, as proper concepts of material objects in a transcendentally principled way. The mathematizablity of the domain of objects to which natural science applies thus depends on both the “principles of sensibility” and the “mathematical principles of pure understanding.” 23

This traditional view has been challenged by Daniel Sutherland, who argues rather that the interpretive use of a pure/applied distinction to distinguish the role of the Aesthetic from the Analytic diminishes the importance of the Axioms of Intuition, which secure the “possibility of any mathematical cognition whatsoever, whether pure or applied, general or specific” (Sutherland 2005, emphasis added).

ch a p ter 1 0

Kant’s Dynamical Principles The Analogies of Experience Kenneth R. Westphal

Hume interrupted Kant’s “dogmatic slumbers” (Prol. 4:260), i.e., his unCritical presumption that our most fundamental a priori concepts properly apply to spatiotemporal particulars. Sir Peter Strawson (1966: 29) declared Kant’s reanalysis of those issues made: “very great and novel gains in epistemology, so great and so novel that, nearly two hundred years after they were made, they have still not been fully absorbed into the philosophical consciousness.” Anti-Cartesianism, mental content externalism, and “broad” notions of mental or semantic content have been vigorously developed since, yet Kant’s Kritik der reinen Vernunft is still ahead of our time on all three counts. Although Kant claimed that his transcendental analyses and proofs regarding necessary a priori conditions of our commonsense, selfconscious experience require transcendental idealism (Bxvi–xix, A369–70), this is mistaken (Westphal 2004). Kant’s analysis of our necessary, legitimate use of causal concepts is sufficiently justified by his Critical account of cognitive judgment. I begin with an overview of Kant’s cognitive semantics in Section 10.1, followed by an analysis of Kant’s Analogies in Section 10.2.


Kant’s Semantics of Singular Cognitive Reference

Hume’s followers neglect five central points Kant examined. First, Hume recognized that we have and use a range of what may be called merely determinable concepts, as well as linguistic tags for various concepts. However, Hume’s official “copy theory” of sensory impressions and ideas, and his three official principles of psychological association (resemblance, contiguity and 1:1 correlation, supposedly causal), can account only for determinate classificatory concepts of sensed qualities, as fine-grained as one can regularly discriminate. Hume’s official mechanisms of the mind do not suffice to account for merely determinable concepts, such as ‘space,’ ‘region of space,’ ‘time,’ ‘period of time,’ ‘cause,’ 184

Kant’s Dynamical Principles: The Analogies of Experience


‘substance,’ ‘number,’ ‘color’ or ‘word,’ nor for linguistic tags (names), in contrast to meaningless sounds. For determinable concepts and for words only Hume’s ever-capacious “imagination” can account, but for the imagination and its prodigious results Hume provides no specifically empiricist account. Hume’s specifically empiricist principles consist only in his official copy theory of impressions and ideas and three principles of psychological association (Westphal 2013a). Second, in explaining our mere though ineradicable belief in the existence of physical objects (“body”), Hume found that his official empiricist principles require three additional “propensities” of the mind to form various beliefs in response to various characteristic sequences of impressions. Hume’s focus upon the occasioning causes of these beliefs occludes how these beliefs each require concepts which cannot be defined in accord with concept empiricism; if they could be, Hume would not need to introduce these propensities: (1) to believe that any unchanging impression, which occurs during any other series of impressions, is a physical object; (2) to believe that any series of closely resembling impressions is an experience of some one physical object; (3) to believe that any closely resembling series of impressions which occur in different, discontinuous periods, are experiences of some one physical object which continues to exist during the (seeming) interruption in our experience of it. The conceptual content of each of these beliefs defies concept empiricism, and can only result (on Hume’s view) from our fertile imaginations. Third, the concept ‘cause,’ even as mere 1:1 contiguity, can be neither learned nor defined on the basis of our typical human experiences, because – as Hume recognized – we so very often experience either a purported cause or a purported effect without its purported partner. Consequently, by the official empiricist mechanisms of the copy theory and the three principles of association, we should only form very few, very weak beliefs (if any) in particular causal relations, which would never suffice to define, to learn, or to prompt the thought of the general concept of cause invoked in the general causal principle, ‘every event has a cause’ (Beck 1975: 121–29). Any sorting of our experiences to select only those favorable cases in which we happen to observe both the purported cause and its purported effect presupposes the concept of cause as 1:1 contiguity, which is required to form any expectation that we should meet with such 1:1 correlations in whatever series of impressions happen upon us, or likewise that we should sort our impression-experiences to select only the relevant paired instances. Fourth, when sitting before the fire in his study, Hume received a letter by porter (T–25). His delivery requires the continued existence of


k e n n e t h r. we s t ph a l

the stairs Hume no longer perceives, so that the porter can reach Hume’s door. His recognizing the knock at the door requires believing that the door continues to exist, and that very likely someone outside knocks. Recognizing a squeak as the sound of that door opening requires believing that door continues to exist when otherwise unperceived by Hume, seated facing the fire. Neither the content nor the justification of such beliefs can be accounted for by Hume’s official empiricist principles. Yet without the belief in the continuing, mind-independent existence of physical objects, our commonsense beliefs lose all coherence. This much Hume recognized, though he denies it justifies such beliefs. Fifth, Hume recognized not only these difficulties regarding our understanding and identification of particular physical objects over time (their diachronic identity), but also a key problem regarding their synchronic identity at any time: any physical object has a variety of characteristics or properties, although it is one single object. This identity, however, is not simply numerical: neither ‘unity’ nor ‘plurality,’ as numerical concepts, suffice to define the singularity of any one physical object with manifold characteristics (T 1.4.3). By rigorously developing the implications of his official concept empiricism and verification empiricism, together with the sensory atomism apparently endorsed by his predecessors, Hume verged upon recognizing a core problem in the Modern “new way of ideas” and the sense data tradition, only recently recognized by analytic epistemologists (cf. Cleremanns 2003): the host of problems called “the binding problem.” Kant expressly recognized these problems concerning how any plurality of sensations become integrated into some one percept of some single object; and likewise, how any plurality of sensory information about the characteristics of any one perceived object becomes integrated into their identification as characteristics of some one perceived, recognized individual. These problems arise both synchronically and diachronically, and they arise both within and across each of our sensory modalities. Kant recognized that none of these problems can be solved simply by adding further sensations to any such series or concurrent plurality of sensations. The integration of sensations into percepts at any time, and the integration of any series of percepts over time into the continuing perception of any particular, requires nonsensory functions guided by relevant principles. This point holds generally; it requires neither sensory atomism, nor that sensations themselves be objects of our self-conscious awareness. Kant held that the principles guiding perceptual synthesis derive from the same basic logical functions of judgment from which derive our most

Kant’s Dynamical Principles: The Analogies of Experience


fundamental a priori concepts, the categories. These functions of judgment are used twice over, Kant argues, first at a subpersonal level by transcendental imagination in perceptual synthesis; and second at an explicit, cognitive level within experience by understanding in making even putative cognitive judgments (A79/B105–6, B152, B162n). Michael Wolff has shown that Kant has excellent grounds for maintaining that his table of judgments is complete, and that his general logic is indeed more general and more basic than Frege’s mathematical logic: the greater power of first-order predicate calculus is obtained by additional semantic postulates which also make it less general. Kant’s cognitive psychology is complex and sophisticated, and may only be touched upon here. That some such transcendental cognitive psychology must hold true of us homo sapiens, such that we can be Homo sapiens sapiens, is indicated by a further important insight of Kant’s, won by Critically rethinking Hume’s empiricist epistemology. Insofar as we perceive our surroundings via our sensory channels, we can sense and perceive physical objects and events. However, we can sense neither space nor time as such (A172–73/B214, A188/B231, A214/B261, A487/B515). Consequently, we cannot localize physical particulars simply by sensing the regions they each occupy. Additionally, our sensory experiences are continually successive; however, no mere succession of sensations, nor of sensory percepts, nor of perceptions, qua successive sequence suffices to discriminate whether the features of objects or events so sensed are themselves sequential, or instead exist concurrently. This Hume failed to note, except to the (insufficient) extent that the porter temporarily interrupted his studied empiricist repose. A central tenet of Cartesian skepticism, and of analytical epistemology prior to Gettier (1963), is that justification sufficient for knowledge entails the truth of what is known, so that any logical possibility of error must be excluded; otherwise what is putatively known could be false. This tenet requires epistemologists to demonstrate that our cognitive capacities are sufficiently reliable for cognition in any possible environment before trusting our actual cognitive capacities within our actual environment. This tenet is not merely Cartesian, it is Roman Catholic, asserted upon authority of the Pope by Étienne Tempier, Bishop of Paris, in 1277, when condemning 220 neo-Aristotelian theses as heretical (Boulter 2011). His condemnation asserted and implied that knowledge (scientia) requires eliminating all logically possible alternatives. That converted Aristotle’s flexible model of scientia into infallibilist deductivism, and directly fostered Descartes’s merely possible nemesis, the malin genie, since according to Tempier, the divine omnipotence can bring about any event regardless of its typical


k e n n e t h r. we s t ph a l

natural causes, including those events we regard as our experiences of our natural and social environs. Descartes generalized this into possible global perceptual skepticism. Long-standing rejection of issues about cognitive judgment within analytic epistemology resulted in part from the seeking to avoid “psychologism” (of whatever varieties), though also by the implicit though fundamentally Cartesian aspiration to refute the epistemological nightmare of global perceptual skepticism. All of Gettier’s counterexamples centrally involve what became known as ‘externalist’ factors bearing upon the justificatory status of Someone’s beliefs, factors such that S/he neither was, nor could easily become, aware by simple reflection. These may be environmental, somatic, or habitual, as in S’s inference patterns. Skeptics stress that all of our experiences and beliefs could be as they are, although as a simple matter of logic they could all be false (Stroud 1994: 241–42, 245). This instead demonstrates that cognitive justification is not reducible to logical deduction. Kant recognized this in his distinction between general logic and a specifically “transcendental logic” (B170), which considers the various possible and necessary roles of a priori concepts and principles within human experience and knowledge, their respective domains, and the conditions under which their use can be legitimate (or not). Kant understood that understanding human knowledge requires understanding how knowledge is possible for beings like us. So doing requires a basic inventory of our characteristically human cognitive capacities; Kant provided the necessary minimum inventory. Guyer (1989) showed that Kant’s analysis of the subpersonal cognitive processing effected by transcendental power of imagination is necessary for any cognizant being who synthesizes sensory information over time (in response to stimulation by spatiotemporal objects and events; cf. B178, B298). These programmatic points help elucidate and assess Kant’s problem, analysis, and justification of the principles guiding causal judgment in the “Analogies of Experience.” Kant appeals directly to several of these points when introducing the Analogies, at the beginning the Transcendental Analytic, Book 2: “The Analytic of Principles,” which details the transcendental power of judgment as such. Kant notes that using concepts, principles, or classifications within cognition requires judgment to ascribe relevant characteristics to particulars by subsuming that (or those) particular(s) under the conceptual classifications used in our judgments. This holds regardless of how specific our rules or classifications may be (in nonformal domains; B172–74). Conversely, Kant notes that merely possessing or using a priori concepts or principles does

Kant’s Dynamical Principles: The Analogies of Experience


not suffice for knowledge using those concepts or principles. Knowledge requires applying those concepts or principles to particulars which purportedly instantiate them. In our human case, locating such particulars and subsuming them under our classifications or principles (whether a priori or empirical; both are involved in any empirical judgment) are localizable only via sensation, whether directly by sensory perception or by using observational instruments. This marks Kant’s decisive semantic and epistemological critique of pre-Critical metaphysics: that we possess a priori concepts shows nothing about whether we can use them in justifiable cognitive judgments. Kant’s a priori justification of synthetic principles brooks no transcendent metaphysics. To the contrary, Kant adopted Tetens’s use of the term ‘realisieren’ (B185–86, B187) to stress that demonstrating the legitimate, justifiable use of any concept, especially any a priori concept, requires demonstrating that we can locate actual particulars to which we can correctly apply that concept. Kant calls this demonstrating the “objective reality” of various a priori concepts and principles (B288, B300–303, B314); or conversely: showing that some a priori concepts are such that we cannot provide them any objective reality, in which case such concepts are cognitively transcendent. Avant la lettre, Kant’s semantics of singular cognitive reference embeds Evans’s thesis about predication within a much richer epistemological analysis. Evans (1985: 36, cf. 34–37) argued that “the line tracing the area of [ascriptive] relevance delimits that area in relation to which one or the other, but not both, of a pair of contradictory predicates may be chosen. And that is what it is for a line to be a boundary, marking something off from other things.” Specifying the relevant boundary for the use of any member of a pair (or set) of mutually exclusive predicates is only possible by specifying the region relevant to the manifest characteristic in question, and vice versa, and (for reasons Evans provides, concerning the mastery of the relevant predicates of a language) this region will be either co-extensive with or included within the spatiotemporal region occupied by some particular object or event. Predication requires conjointly specifying the relevant spatiotemporal region and some manifest characteristics of any particular we self-consciously experience or identify within that region. These conjoint specifications may be approximate; the key point is that spatiotemporal localization and ascription of manifest characteristics are conjoint, mutually interdependent cognitive achievements (B162). This conjoint designation of any particular’s region and at least some of its manifest characteristics requires thorough integration of sensibility and understanding: sensibility is required (though not sufficient) for sensing


k e n n e t h r. we s t ph a l

various manifest characteristics of the sensed particular, and in directing us to its location; Understanding is required (though not sufficient) for explicitly delineating its region and identifying its manifest characteristics as its characteristics, thus enabling Someone to be self-consciously aware of this particular. This point about predication is justified by Kant’s semantics of singular, specifically cognitive reference.1 ‘Cognitive’ reference concerns our reference to (putatively) known individuals, as instances of our (putatively cognitive) judgments or assertions. Kant’s point is that knowledge, justified belief, error, or indeed experience (whether veridical or not) of or about particulars require satisfying further conditions of reference than those implicit or explicit within conceptual content or linguistic meaning alone. According to Kant, concepts have ‘meaning’ or content as predicates of possible judgments, though no concept has specifically cognitive significance unless and until it is incorporated into a candidate cognitive judgment which is referred to some actual particular(s) localized within space and time by the presumptive judge, some cognizant person S. The relevant particulars are located within space and time; I use the term ‘localized’ to stress that S identifies (at least approximately) where and when (putatively) known or experienced particulars are located. Kant analyses the first stage of conceptual meaning in the derivation of the Table of Categories from the Table of Judgments and in the Schematism of the Categories; he analyses the second stage of cognitive significance in the Transcendental Aesthetic, the Amphiboly of the Concepts of Reflection and in the Analytic of Principles. To have any possible significance for theoretical cognition, the categories – and likewise for all of our concepts – require applicability to objects we can experience. (This is the task of Kant’s Schematism, augmented in the Analytic of Principles.) However, to have actual cognitive significance, the categories and our other concepts must be “applied to objects” which we experience. Through his critique of Leibniz (in the “Amphiboly of the Concepts of Reflection,” in his Appendix to the “Analytic of Principles”) Kant identified the cognitive insufficiency of the descriptions theory of reference. According to the descriptions theory of reference, our statements refer to whatever is described when we analyze the meanings of our terms or statement into explicit descriptions. The problem with this approach within epistemology is that, no matter how specific or extensive a description may be, no description by itself determines whether it is (logically) empty, determinate, 1

Westphal (2004, esp. §§7–9, 33, 62–63.2).

Kant’s Dynamical Principles: The Analogies of Experience


or ambiguous because it describes no, only one, or instead several individuals. Which may be the case is not simply a function of the description: it is equally a function of what there is. The inclusion of definite pronouns (such as ‘the’ or ‘the one and only’) within an attributive phrase cannot settle this issue because no definite article insures that the phrase in which it occurs is neither empty nor ambiguous; this was Russell’s problem (ca. 1905) about ‘the present King of France.’ To know any one spatiotemporal particular (even putatively) requires both correctly ascribing characteristics to it and locating it in space and time. Integrating both of these is required for predication, and also for knowledge of (or even error about) that individual: predication (even putative predication) is a cognitive achievement; it is not merely a grammatical or judgmental form. Only through singular sensory presentation and competent use of conceptions of time, times, space, spaces, individual, and individuation, Kant argues, can we localize any object or event in space and time (even putatively). Only through ostensive designation can we ascribe the predicates used in our (perhaps implicit) description to any one, putatively known particular. Therefore, predication is required for singular, specifically cognitive reference to any spatiotemporal particular(s). Only through predication as this kind of cognitive achievement can anyone specify (even approximately) the relevant spatiotemporal region (putatively) containing the particular(s) one purports to designate ostensively – by specifying its occupant, the (putatively) known particular(s). Only in this way can one note, specify, or determine precisely which spatiotemporal region to designate, in order to grasp this (intended, ostended, presented) particular, and to ascribe to it any manifest characteristics, all of which is required to achieve any knowledge (whether presumptive or actual) of that particular (B162). Thus, in brief, does Kant show that determinate cognitive judgments are possible for us only through conjoint spatiotemporal designation of, and predicative ascription of characteristics to, any experienced particular(s).2 As important as predication is to philosophy of language, analyzing the meanings of our terms or the contents of our concepts, descriptive phrases, or psychological “attitudes,” does not because it cannot suffice for epistemology. Kant’s semantic thesis can be formulated in terms of claims, beliefs, statements, assertions, or judgments. Put in terms of judgments, Kant’s Thesis of Singular Cognitive Reference is as follows: terms or phrases have 2

Kant’s semantics of singular cognitive reference provides for scientific reference to indirectly observed entities or forces, e.g., the magnetism of the loadstone responsible for the stone’s observed effects upon iron filings (B273).


k e n n e t h r. we s t ph a l

‘meaning,’ and concepts have (classificatory) content, as predicates of possible judgments, though (in nonformal, substantive domains) none of these has specifically cognitive significance unless and until it is incorporated into a candidate cognitive judgment which is referred to some actual particular(s) localized (at least putatively) by the presumptive judge, S, within space and time. Cognitive significance, so defined, is required for cognitive status (even as merely putative knowledge) in any nonformal, substantive domain. Kant’s cognitive semantics does not rule out second-hand “knowledge by description” based upon reliable testimony; it establishes basic cognitive conditions upon the acquisition of empirical knowledge, by identifying basic conditions under which alone synthetic statements have specifically cognitive status within any nonformal domain. Kant’s cognitive semantics founds an important quadruple distinction between description, ascription (i.e., attribution of the predicates contained in S’s description to some particular(s) localized by S, or: predication), sufficiently accurate or true ascription, and sufficiently justified accurate or true description. Only the latter counts as empirical knowledge. Consequently, Kant’s analysis of specifically cognitive reference shows why philosophy of language or philosophy of mind may augment epistemology, but cannot supplant it, insofar as neither cognitive justification nor singular cognitive reference can be reduced to, nor substituted by, analysis of linguistic meaning or mental content.3 Kant’s cognitive semantics secures the key aim of verification empiricism, without invoking verification empiricism! Kant’s point holds regardless of whether the concepts we use in cognitive judgments (in nonformal, substantive domains) are a priori, a posteriori, or mixed. His cognitivesemantic point is that, whatever may be the conceptual content or linguistic meaning of our claims, judgments, or propositions, they have no cognitive status unless and until they are referred to particulars we have (presumptively) localized within space and time. This requirement is a necessary condition for the truth-evaluability of our claims (etc.); it is necessary for us to know enough about our claims and whatever about which we make those claims to discover and thereby to determine their truth value, their accuracy, or their adequate approximation. It is also necessary (though not sufficient) for our assessing the justification of our cognitive claims about those particulars. This is the nerve of Kant’s critique of prior, cognitively 3

These important features of Kant’s semantics of singular cognitive reference, and indeed of Evans’ analysis of predication, are neglected by McDowell; see Westphal (2008b).

Kant’s Dynamical Principles: The Analogies of Experience


transcendent metaphysics. Kant’s a priori justification of some central synthetic claims provides no solace for transcendent, rationalist metaphysics (nor for its contemporary echoes within analytical metaphysics).4 Kant’s cognitive semantics also shows that justificatory infallibilism is in principle irrelevant to the nonformal domain of empirical knowledge. Strictly speaking, formal domains are those which involve no existence postulates. Strictly speaking, the one purely formal domain is a careful reconstruction of Aristotle’s Square of Opposition (Wolff 2000, 2009a, 2013). All further logical or mathematical domains involve various existence postulates, including semantic postulates. We may define ‘formal domains’ more broadly to include all formally defined logistic systems (Lewis 1930 [1970]: 10). Whether we construe formal domains narrowly or broadly, deduction suffices for justification within any formal domain because deduction constitutes justification within any formal domain. Indeed, a domain is formal only insofar as deduction constitutes justification within it. Only within formal domains is justification constituted by provability. The relevance of any such logistic system to any nonformal, substantive domain rests, however, not upon formal considerations alone, but also upon substantive considerations of how useful a specific logistic system may be within a nonformal, substantive domain (Lewis 1929: 298; cf. Carnap 1950). The use of any specified logistic system within any nonformal domain does not suffice for justification within that domain; justification within that domain also requires assessment of the adequacy, accuracy, and specific use of, inter alia, the semantic and existence postulates which partially constitute and delimit that domain. Consequently, within any substantive domain, fallibilism is no skeptical capitulation, not because infallibilist standards of justification are too stringent, but because in principle they are inappropriate to – indeed, they are insufficient for – any and all substantive domains. Conversely, within any substantive domain, mere logical possibilities as such have no cognitive status and cannot serve to “defeat” or to undermine (refute) an otherwise well-grounded line of justificatory reasoning within that domain. The domain of (putative) empirical knowledge includes spatiotemporal objects and events; accordingly, empirical knowledge is a nonformal domain. Consequently, Kant’s 4

Kant’s epistemology is (in these regards) sound; his semantics is much more sophisticated than Coffa (1991) recognized. See Melnick (1989), Hanna (2001), Westphal (2004), Bird (2006), and Haag (2007). Melnick’s (1989) unjustly neglected masterpiece first made Kant’s semantics evident to me, including Kant’s understanding of the pitfalls of both causal and descriptions theories of reference. To Arthur Melnick I gratefully dedicate this chapter for all I have learned from his deeply instructive writings over these many years.


k e n n e t h r. we s t ph a l

analysis of singular cognitive reference rules out the ideal of infallible justification (scientia) within the entire nonformal domain of empirical knowledge. Recognizing that only fallibilist accounts of justification are tenable within the nonformal domain of empirical knowledge concedes nothing to skepticism. Kant’s semantics of singular cognitive reference underscores that empirical knowledge is discriminatory; it involves discriminating particulars both spatiotemporally and by their manifest or measurable characteristics. The discriminatory character of our empirical knowledge is greatly augmented and underscored by Kant’s analysis of the basic principles of causal judgment in the “Analogies of Experience.”


The Principles of Causal Judgment in Kant’s Analogies

Discussion of Kant’s Analogies has focused almost exclusively upon the Second, where Kant allegedly replied to Hume’s causal skepticism. That cannot be true; in the Second Analogy Kant’s model of causality is Leibnizian (Beck 1975: 149n); it only concerns rule-governed causal changes of state within any one substance, whereas Hume’s skepticism concerns causal relations between two or more particulars. Kant’s First Analogy concerns the persistence of any one substance through causal changes of its states. Only in the Third Analogy does Kant defend a principle of causal judgment regarding causal interactions between any two or more substances. Recent literature has paid more attention to Kant’s Third Analogy, yet even leading research on Kant’s Analogies of Experience neglects Guyer’s decisive finding, that Kant’s principles of causal judgment in the Analogies form an integrated set, because no one of these principles can be used without conjoint use of the other two.5 Furthermore, these three principles of causal judgment provide an integrated, incremental justification of judgments about transeunt causal interactions. (A cause is ‘transeunt’ if it extends beyond any one substance in order to effect a change in another; O.E.D.) Kant’s main examples in the Third Analogy are astronomical, but his analysis is general and holds of all forms of causal interaction between physical particulars, of whatever kinds, at whatever scale. Following Caird and Paton, Guyer notes that Kant’s defense of causal interaction counters Leibniz as well as Hume. Once again I gladly summarize Guyer’s findings: The three Analogies form a tightly integrated set of mutually supporting principles. The empirical criterion of succession is lack of reversibility 5

Guyer (1987: 168, 212–14, 224–25, 228, 239, 246, 274–75).

Kant’s Dynamical Principles: The Analogies of Experience


of the type of sequence of appearances produced by one or more objects; the empirical criterion of co-existence is the reversibility of the type of sequence of appearances produced by one or more objects. Determining that either co-existence or succession occurs requires determining that the other does not, and both determinations require that we identify objects which persist through both the real and the apparent changes involved in the relevant sequence of appearances. We directly perceive neither time nor space, and the mere order in which we apprehend appearances determines no objective order of objects or events: our ever-successive perceptions may be perceptions of concurrently coexisting particulars or features of some one particular. Consequently, the only condition under which we can determine which states of affairs precede, and which coexist with, which others is if there are enduring perceptible substances which interact causally, thereby producing changes of state in one another, including changes in location or motion. Enduring substances are necessary for us to determine the variety of spatial locations, to determine changes of place, and to determine nonspatial changes objects undergo. To ascertain whether a change of appearances is a function of one object, previously in view, moving out of view when displaced by another; or instead is a function of one object rotating to reveal a different aspect; or instead is a function of one spatially stable object undergoing a nonspatial change of state, requires that we are able to identify places, changes of state, and objects which change place or state, and that we are able to distinguish these different kinds of causal scenario. To identify any one such scenario requires conjoint, discriminatory use of all three principles defended in the Analogies. The principles of causal judgment defended in the Analogies all stand together, or not at all. Defending transeunt causality is thus central to Kant’s Analogies as a whole, and not only to the Third Analogy. Both the valid and the possible use of Kant’s causal principles requires that changes in material substances we identify are produced, directly or indirectly (via their “relatively inner” determinations), by external transeunt causes. Kant states these principles of causal judgment in the three Analogies: (1) Principle of the Persistence of Substance: In all change of appearances substance persists, and in nature its quantum is neither increased nor diminished. (B223) (2) Principle of Temporal Sequence According to the Law of Causality: All alterations occur in accord with the law of the connection of cause and effect. (B231)


k e n n e t h r. we s t ph a l

(3) Principle of Concurrent Existence According to the Law of Interaction, or Community: All substances, insofar as they can be perceived in space as concurrent, are in thoroughgoing interaction. (B256)6 The differences in Kant’s formulations between the two editions are far less significant than the central cognitive-semantic point they share. Kant’s Principles are stated in categorical, universal form, indeed in ways resembling conservation laws in physics (Weizsäcker 1971). Kant’s Principles conjointly state the universal causal principle, that every event has a (perhaps jointly) sufficient efficient cause (or causes). Rationalists, occasionalists, idealists, and even Hume in some moods agree. Between the two editions of the first Critique Kant introduced (MF 4:543) an important distinction between the universal causal principle and this specific causal principle: each spatiotemporal, physical event has a (perhaps jointly) sufficient, external spatiotemporal, physical efficient cause (or causes). Kant nowhere in the first Critique states this specific causal principle, though later (CJ 5:181) he reiterates its distinction to the universal causal principle. Because Kant aims in the Analogies to justify causal interaction between physical substances, he must justify this specific causal principle regarding external physical causes, and not merely the universal causal principle. Kant has two strategies to meet this challenge; the most important builds upon Kant’s cognitive semantics of singular reference. Kant’s critique of cognitive judgment, including his cognitive semantics of determinate singular reference, requires distinguishing the literal and full meaning (intension) of these causal principles as formulated, and the legitimate, justifiable cognitive significance of any judgments we can make using those principles. This accords with Kant’s calling his analyses and justification of these principles “Analogies,” insofar as these causal principles regulate our causal judgments by guiding our identifying efficient causes of observed spatiotemporal events. How fully or precisely we may identify causes and effects is a matter for empirical inquiry, whether commonsense, diagnostic, forensic, or natural-scientific. Because our causal judgments are (as indicated) discriminatory, we can only discriminate apparent from real changes of objects’ states, locations, or motions insofar as we identify – sufficiently to recognize them at all – other physical events which cause those changes, so as to distinguish those objective, physical changes from merely apparent 6

I have slightly revised Guyer’s translation; Kant’s term ‘Zugleichsein’ concerns perceptible concurrence during observable intervals of time, nothing so exact as the relativization of instantaneous ‘simultaneity’ in General Relativity theory; see Westphal (2007b: 740–41).

Kant’s Dynamical Principles: The Analogies of Experience


changes resulting from our contingent observations, including our bodily comportment; e.g., the directions in which we gaze co-determines the sequence in which we happen to observe the concurrently existing aspects of any building (B162). In contrast, when observing, e.g., a ship navigating a river, our own changes in viewing perspective (or period) alters nothing about the observable locations of the ship (B237). In making such discriminatory judgments, Kant expressly notes, we cannot possibly refer in any determinate way whatever to any transcendent (“foreign”) cause of the sorts alleged by occasionalists (B251–52). Between the two editions Kant further notes, in explaining why there can be no natural (causal) science of psychology (MF 4:471, cf. A381), that we can only make specific, justifiable causal judgments about spatiotemporal events. Within the sole temporal dimension of inner sense we cannot discriminate any substance(s) and so cannot specify any supposed causal action of (putative) psychological substance(s) or states of affairs. This finding is directly implied by the integrity of the three principles of causal judgment Kant defends in the Analogies; the Third Analogy is expressly and rightly restricted to spatial objects and events (B231, 257–58, 275–78, 291, 340). Because the first two causal principles can only be used in justifiable causal judgments in conjunction with the third, all three principles are jointly restricted to guiding our judgments about spatio-temporal objects and events.


Kant’s Transcendental Proof of Realism

Kant’s analysis of the necessary conditions of our inherently discriminatory causal judgments provides a transcendental proof for (not ‘from’) mental content externalism. Briefly, it is this: (1) Each of us is conscious of our own existence as determined in time; i.e., we are aware of ourselves as being aware of some things (in a broad, noncommittal sense of ‘things’) as appearing to occur before, during or after others (B275). Hume’s experience of the porter delivering his letter (T 1.4.2) commits him to this premise. (Any presumed skeptic who refuses to answer the question, ‘What did you just say?,’ poses no philosophical challenge.) (2) The kind of awareness indicated in (1) is self-conscious (apperceptive) experience, and the ‘things’ experienced are objects (in a broad, noncommittal sense of ‘accusatives’) of our experience. (3) One can be self-conscious only if one can distinguish oneself from something of which one is conscious. One may have sensations


k e n n e t h r. we s t ph a l

without apperception (A90–91, 111–12, 116–17/B122–23), but one can’t be self-consciously aware of anything through them unless one can identify oneself as the conscious subject to whom it at least appears that S/he senses something. Being self-conscious requires distinguishing oneself from the contents of one’s experiences in order to be able to think that you have experience of it (self-ascription). Here the ‘something’ of which one is aware is understood as an object of thought, per (1). (4) One can distinguish between oneself and something of which one is aware only if one can identify that of which one is aware. (5) One can identify something (as an object of self-conscious awareness) only if one both correctly characterizes it and correctly locates it in space and time. ‘Correctly’ does not require ‘precisely’; (5) is supported by Hume’s problems in his study (T–25) and by Kant’s semantics of cognitive reference and his account of perceptual synthesis. (Recall Kant’s warning against conflating sensation and conception, and his recognition of “binding problems,” noted above.) (6) Correctly characterizing something and locating it in space and time requires being able to correctly characterize and locate it within (apparent) space and time. (7) The order of apprehension of the objects of experience is always successive, regardless of whether the objects experienced or their features are concurrent or successive. On Humean grounds alone we cannot distinguish between these three accounts of the experience of a blue dot on a white field being succeeded by a red dot on a white field: (a) A blue impression being replaced by a red impression; (b) A ball, blue on one side, being instantaneously rotated to reveal its red side; (c) A blue ball transforming into a red disk. Therefore, (8) our apprehension does not, of itself, reveal the objective order of events; the temporal order of the objects (broadly speaking) of experience is not indicated simply by our successive apprehensions of experience (from (6), (7); cf. A182, A194/B225; B219, 226, 243, 257). (9) Time itself is not an object of possible experience. (A172–73, A188/B214, B231) Therefore, (10) Temporal order cannot be determined by reference to time itself (from (9); cf. A182/B255, A183/B226, A215/B263; B219, 233, 277). (11) Space itself is not an object of possible experience (A172–73/B214, A214/B261, A487/B515). Therefore, (12) spatial order cannot be determined by reference to space itself (from (11)).

Kant’s Dynamical Principles: The Analogies of Experience


Therefore, (13) to be able to identify changes of state, local motions, translational motions, and radical transformations of substance, one must be able to discriminate each of these potential kinds of change from each other in any particular case (from (8), (10), (12)). Therefore, (14) to be able to discriminate among the potential kinds of change indicated in (7) in any particular case, one must be able to reconstruct the spatial order of events by distinguishing it from apparent spatial locations, i.e., from one’s subjective order of spatial apprehension (from (8), (13)). Therefore, (15) to be able to locate objects of experience in time (as occurring before, concurrently with, or after others), one must be able to distinguish one’s subjective order of apprehension of the objects of experience from the objective, spatial and temporal order of the world (from (5), (7), (8), (14)). (16) Subargument by reductio ad absurdum to show that there are rule-governed relations among appearances, to support (17) and the antecedents of (18), (19); (A194–95/B239–40, cf. A112): (i) Suppose: there is nothing antecedent to an event appearance, upon which it follows according to a rule. Then: (ii) all succession of perception would be only in apprehension (i.e., it would be merely subjective), and would disable us from ever determining objectively which perceptions really precede, which follow, and which are concurrent. Thus: (iii) the relations between any two appearances (which would only be distinguishable, if at all, on the basis of apparent sense-content) would be the same (i.e., equally arbitrary). Thus: (iv) we would lack criteria, guides, or indications for grouping our sensory representations together. Thus: (v) we would lack usable criteria for identifying objects and events. Thus: (vi) the succession in our apprehension would always be the same; and: (vii) There would be nothing in the appearances which so determine it that a certain sequence is rendered objectively necessary. Thus: (viii) we would have an apparently random play of representations relating to no object. Therefore: (ix) that we are so much as putatively aware of objects or events entails that there are rule-governed relations among appearances ((i)). Therefore: (17) if one can distinguish between one’s subjective (spatial and temporal) order of apprehension and the objective order of events


k e n n e t h r. we s t ph a l

in space and time, then there is an objective order of events in space and time. (18) If there is an objective order of events then things determine their own sequence in time and their own locations and motions in space (cf. A199/B244). Therefore: (19) if things determine their own order in time and space then those things are causally related, so that the antecedent of an event contains the condition of a rule upon which necessarily follows the event (from (17), (18); A144/B183, A189–91/B236, A195/B240, A198/B243). (20) If something contains such a condition of such a rule then that thing is a substance, i.e., an enduring thing having properties, some of which are dispositional and causal (cf. A183–84/B226–27, A204/B249). Here the ‘things’ in a very broad, noncommittal sense used in (1) are given a very committed interpretation, in opposition to Hume’s impressions and their collections. (21) If one can identify events as occurring before, during, or after others, then one is able to recognize the objective order of events, that is, to construct knowledge of the objective order of events on the basis of one’s experience of them (from (4)–(6), (15), (20)). Therefore: (22) one can distinguish the subjective order of apprehension of things from the objective order of the world only if one can correctly use object concepts to identify what one experiences; that is, only if one can distinguish the three different kinds of accounts of the experience described in (7) by using concepts of substance, cause, and event (rule-governed causal succession) to judge what one experiences (from (21); cf. A195/B240, A199–200/B244–45). Therefore: (23) the conditions for the possibility of self-conscious human experience (of identifying our self-consciousness “as determined in time” (1)) are likewise the conditions for the possibility of knowledge of objects, including that perceptible, spatiotemporal, causally interacting physical objects exist and that one perceives and identifies at least some of them (from (3)–(6), (22); cf. A111, B275, B276.). Therefore: (24) if a human being is self-consciously aware of him- or herself as minimally determined in time, then S/he perceives and has a least some knowledge of spatiotemporal, causally active substances in his or her environs (from (1), (23)).7 7

Notice further that all of this must hold, i.e., we must in fact be aware of at least some physical particulars in our surroundings, to formulate, consider, or investigate those judgments which Kant

Kant’s Dynamical Principles: The Analogies of Experience



Kant contra Global Perceptual Skepticism

The strategy of Kant’s analysis of causal judgment is to show that the fact of occasional perceptual error warrants no generalization to the possibility of global perceptual skepticism. Instead, for beings like us, apperceptive (selfconscious) awareness of so little as some events appearing to us before, during, or after others is only possible for us if in fact we successfully identify at least some perceptible physical objects and events in our surroundings. If so, then the cognitive question is which perceptions involve perceptual knowledge, not whether any do. Philosophers are prone to some characteristic responses to Kant’s analysis. Many take the form, ‘But might it not possibly happen that?’ Here speaks Cartesian infallibilism: Kant’s cognitive semantics shows that mere logical possibilities – expressed by merely logically consistent descriptions – have no cognitive standing unless and until they are referred by someone to some purportedly relevant particulars (Bxxvin, B175, 242–43, 267–68, 270, 304, 309–10, 799). Hence mere logical possibilities do not undermine cognitive justification within the nonformal domain of empirical knowledge. Another response is incredulity at the thesis that only in, through, and by identifying physical objects and events can or do we concurrently identify and differentiate (at least approximately) the regions they occupy, some of the manifest characteristics they exhibit and some of their causal interactions. How is any such achievement possible? Kant’s answer is that this commonsense achievement involves exercising a host of integrated, subpersonal cognitive capacities, principles, concepts, and judgments.8 Modern philosophers often ask how we do this in any particular case, and have often supposed that if we cannot answer that specific ‘how’ question, then we do not – or we do not know that we do – achieve all that in any particular case. This is Cartesian internalism speaking, as the expectation of Cartesian selftransparency, according to which we ought to be able to identify by introspective reflection all (or at least most) of the main elements involved in any instance of genuine perceptual knowledge. Kant’s entire transcendental analysis rejects the ‘KK’ principle (that to know that x, S must know that S/he knows that x), arguing on these – and on other, independent grounds – that human apperception is parasitic upon perception of one’s


in the Prolegomena (§§17–20) calls ‘judgments of perception’; to wonder whether the sun warms the stone, one must identify the sun and identify the stone and at least conjecture about their causal relation. The alleged inconsistency between Kant’s doctrines in the two works is spurious (cf. B162, 240). These are diagrammed in Westphal (2017).


k e n n e t h r. we s t ph a l

surroundings, and that awareness of oneself as (apparently) percipient is parasitic upon at least some veridical perceptions of one’s surroundings.9 However catholic such challenges remain today are so many indications of how pervasively contemporary philosophy remains pre-Kantian. Through his analysis of these principles of causal judgment Kant also purports to prove stronger claims, specifically: Every spatiotemporal event has a (perhaps complex, jointly) sufficient cause or causes; Causal determinism holds universally within space and time; There is some one constant quantum of substance within space and time; Hylozoism is necessarily false; i.e., matter is intrinsically lifeless, though not causally inert. These claims cannot be assessed in detail here, but basic flaws in Kant’s proofs may be indicated; they underscore the strengths of his successful proofs (summarized above). Laplace prominently espoused universal causal determinism. However, causal determinism is not entailed by Newtonian mechanics. Causal determinism requires a causally closed system; nothing in Newtonian mechanics requires or entails a closed physical universe (Earman 1986: 4–54; cf. Lighthill 1986). Laplace (1847 [1820], 7: vi–vii; Nagel 1961: 281n4) appears to state universal causal determinism in his famous image of an omniscient mind who could calculate the current physical state of the universe using perfected Newtonian mechanics to predict or retrodict and thereby know all events throughout time and space. However, Laplace’s formulation is doubly subjunctive: he states that we “ought to regard” (envisager) the current state of the universe as effected by its predecessor and as effecting its successor. Laplace’s counterfactual demon in fact expresses a regulative principle of causal and (he supposed) of probabilistic inquiry. A cardinal tenet of transcendental idealism is that the matter of experience is provided us ab extra; we only generate the form of experience. An indirect though ineluctable consequence of this is that Kant cannot rule out occasional odd experiences such as a random flash of color (with no further effects) within a room one occupies. We could locate and date that event approximately, though precisely enough, yet we could not at all explain it. Kant claims that the occurrence of any uncaused event within experience, or likewise any increase or reduction of the total quantity of 9

For Kant’s second transcendental proof of mental content externalism, see Westphal (2005, 2010).

Kant’s Dynamical Principles: The Analogies of Experience


substance in nature, would disrupt all humanly possible time determination (B231–32, B251). That is his transcendental premise for rejecting such possibilities. However, Kant asserts rather than demonstrates these alleged implications for the very possibility of our determining temporal (and also spatial) sequences (cf. Harper 2007). When he distinguished the universal from the specific causal principle, that every physical event has a sufficient external physical cause (or causes) (MF 4:543), Kant recognized that this latter cannot be proven on transcendental grounds alone, but requires Critical metaphysical explication of the concept of matter as the movable in space (MF 4:470). However, Kant’s argument for this specific causal principle ultimately rests neither on transcendental, nor upon Critical metaphysical grounds, but solely upon our de facto empirical ignorance of any instance of living matter, i.e., a purely material entity which causes some of its own changes, e.g., motions (MF 4:544). Kant is correct about this empirical claim, but no empirical claim provides a legitimate premise for Kant’s transcendental or metaphysical analyses or proofs. That Kant fails to demonstrate these very strong claims is salutary. His cognitive semantics reinforces the regulative status of the three key principles of causal judgment, so that the specific causal principle (that every physical event has an external physical cause) is indeed a regulative principle of all causal inquiry, whether commonsense, diagnostic, forensic, or scientific. However, his cognitive semantics also entails that causal knowledge is only obtained by successful causal explanation of specific events, or specific classes of events. Consequently, the key premise of the debate about causal determinism and human freedom, that each human action is fully causally determined physicalistically, is not known because it is not cognitively justified; nor is sufficient cognitive justification of this premise remotely in the offing (Westphal 2016). Nevertheless, the principles of causal judgment Kant justifies in the Analogies of Experience retain their constitutive role in this Critical regard: if we human beings failed to make any approximately correct causal judgments at all, we could not distinguish ourselves from anything we experience, nor could we be aware of various events so much as appearing before, during, or after others. Accordingly, we would fail to achieve apperception. How extensive such judgments must be cannot be determined by transcendental analysis. Once we achieve apperception by causally discriminating at least some objects and events in our surroundings, it is then a regulative issue of causal inquiry into nature to determine how extensive may be the causal connections among the phenomena we experience.


k e n n e t h r. we s t ph a l

Kant’s cognitive semantics directly and strongly supports Newton’s methodological Rule 4 of scientific explanation, thus undergirding Newton’s causal realism about gravitational force (Harper 2011; Westphal 2013b). Standard objections to scientific realism appeal to mere logical possibilities, an appeal sanctioned by Tempier (1277), but rightly ruled out by Kant’s cognitive semantics, which strongly supports Kant’s fallibilist account of empirical judgments and knowledge.

c h a p ter 1 1

The Refutation of Idealism Ralf M. Bader∗

11.1 Introduction The Refutation of Idealism is an addition to the B-edition of the Critique of Pure Reason that replaces Kant’s discussion of problematic idealism in the Fourth Paralogism of the A-edition.1 Kant was much occupied by this form of idealism, considering it “a scandal to philosophy and to common human reason that the existence of things outside us . . . must be accepted merely on faith” (Bxxxix fn), and describing it as “a kind of cancer in metaphysics” (Metaphysik Dohna 28:681, LM 382). He returned to this topic at various points after the publication of the B-edition2 and already made a number of changes and clarifications to the Refutation in the B-Preface, not being entirely satisfied with the presentation of the argument. Kant aims to establish in the Refutation that problematic idealism is an untenable theory. This form of idealism, which he also calls skeptical idealism (cf. Prol 4:375) and psychological idealism (cf. Bxxxix fn), is an epistemological theory, more precisely a form of external world skepticism, that considers the existence of outer objects to be “doubtful and indemonstrable” (B274). While doubting outer experience, problematic idealism accepts inner experience, considering it to be unproblematic. In this way, it involves a privileging of our epistemic access to the inner over that to the outer, treating the former as immediate and unproblematic, whereas the latter is considered to be mediate and doubtful. ∗ 1 2

For helpful comments, I would like to thank Corey Dyck, Tal Glezer, Colin Marshall, James Messina, Ian Proops, Andrew Stephenson, and Rob Watt. For an account of the relation between the Refutation and the Fourth Paralogism, as well as of the Refutation’s role and location within the Critique more generally, cf. Bader (2012). Cf. R5653–5 (18:306–16, NF 281–88), R6311–17 (18:607–24, NF 355–67), R6319 (18:633–34, NF 374), R6323 (18:641–45, NF 375–77), and the Leningrad fragment (Loses Blatt Leningrad 1, in NF 364– 66, originally published in Brandt and Stark [1987: 18–21]). All translations from Kant are my own unless otherwise noted. References to ‘R’ are to the numbered ‘Reflexionen’ in Kant’s handwritten remains, with references to the Akademie and Notes and Fragments (NF) Cambridge volume provided as well. (The Ak. and NF page references will not be repeated for further references to the range of Reflexionen included in this footnote.)



r alf m . bader

In particular, the problematic idealist holds that, though we are immediately consciousness of our own existence and mental states, the external world is not immediately accessible to us but can only be established on the basis of a doubtful inference from outer representations to the existence of outer objects (cf. R5653, R5709 (18:332, NF 294), and R6315). This inference is considered to be problematic on the basis of the subjective indistinguishability of outer experience and outer imagination. Given that any outer representation could just as well have been the product of the imagination rather than of intuition, it would seem to follow that the existence of the external world is doubtful and cannot be demonstrated.3 The fact that outer imagination and outer intuition are indistinguishable from the subject’s point of view implies that one cannot distinguish them in terms of their determinations, but instead has to appeal to their origin, showing that outer intuitions are required and have to be admitted by the problematic idealist.4 This is precisely what Kant does in the Refutation, trying to show that something that the problematic idealist accepts, namely, inner experience, in fact presupposes outer experience. By showing that inner experience requires outer experience, Kant can refute problematic idealism and establish that “we have experience and not merely imagination of outer things” (B275). In particular, Kant argues that the conditions of the determinability of one’s existence in time involve the existence of an external world, since the objective ordering of inner states is parasitic on that of outer states. Kant establishes this on the basis that time determination requires something permanent in perception together with the claim that the relevant permanent cannot be inner but has to be outer. Accordingly, one can only have inner experience if one has outer experience. That is, an objective ordering of inner states presupposes an objective ordering of outer states. In this way, Kant turns the tables on the idealist, showing that, instead of inner experience being privileged over outer experience, the former actually presupposes the latter.5 3



Cf. “From an effect I can indeed infer to a cause, but not to a determinate cause. A representation of things outside us can have its cause (1) in the imagination, (2) in the presence of the thing” (Metaphysik K2 28:771, LM 410). The indistinguishability on which problematic idealism is based is present even within the framework of transcendental idealism, where outer and inner sense are equally immediate and only give us access to appearances. This explains why the argument provided in the Fourth Paralogism is insufficient for refuting problematic idealism and why the Refutation is required. Kant also provides a somewhat dogmatic argument in the first note following the Refutation (cf. B276–77n), that was already found in the Fourth Paralogism (cf. A373–75), based on the idea that the imagination needs to have material provided by outer sense, given that it is merely a reproductive

The Refutation of Idealism


The core of Kant’s argument, which will be explicated in this chapter, is presented at B275 (together with the emendation in the B-Preface): (1) I am conscious of my existence as determined in time. (2) All time determination presupposes something permanent in perception. (3) But this permanent something cannot be an intuition within me. —————  Hence determination of my existence in time is possible only through the existence of actual things that I perceive outside me.

11.2 Determinability in Time I am conscious of my existence as determined in time.

Kant’s argument begins with a claim about time determination. What exactly this claim consists in is a rather contentious matter. In particular, it is unclear how robust this starting point is meant to be. What kind of temporal determinations does this claim attribute to one’s existence and one’s mental states? Does it merely amount to the consciousness that one exists now? Or does ‘my existence’ refer to the subject’s entire life (up to the present moment)? Does this claim involve any reliance on memory, on how past states are ordered? Does it only assert that one is conscious that one’s existence is determined in time, or also that one is conscious of its time determinations (= duration, succession, and simultaneity), i.e., conscious not just that it is determined in time but how it is determined in time? The robustness of the starting point of the Refutation has far-reaching consequences for the interpretation of this argument, since it affects both its plausibility and its significance, as well as how we are to understand its role within the Critique of Pure Reason, especially its relation to the Transcendental Deduction. The more robust this initial claim is, the more plausible it is that a successful argument can be given. After all, it is much easier to establish that a robust claim has substantive presuppositions than it is to identify substantive requirements for a more minimal claim. Yet, any robust starting point will also be more easily susceptible to skeptical challenges and, accordingly, cannot be used in a non-question-begging manner to refute those who are not willing to grant this initial claim. By contrast, the less robust and more minimal the starting point is, the fewer resources faculty that cannot generate a manifold of its own. This chapter is only concerned with the time determination argument.


r alf m . bader

Kant will have at his disposal and the more difficult it will be to establish any substantive conclusions. Yet, any conclusions that are established from such a basis will be highly significant, given that the minimality of the starting point makes it much more difficult for a skeptic to call it into doubt. Kant is quite clear that the Refutation is based on a robust starting point. What is at issue is not the mere ‘I’ of transcendental apperception (cf. B277 and Bxl). As Kant notes, we are concerned with the conditions for inner experience, for empirical cognition of the self. It is for this reason that the upshot of the Refutation amounts to the claim that inner experience (which involves the determination of the manifold of inner intuitions in time) presupposes outer experience (cf. B275 and B278). In short, the self the existence of which is to be determined in time is the self that is inwardly intuited rather than just transcendentally apperceived (cf. R5653, R6311, R6313). Kant is asserting not just the existence of a bare subject of consciousness, i.e., the determining self, but the existence of the determinable self and therewith the existence of a manifold of inner states that exist in time. This is important because the argument is based on identifying the preconditions for the possibility of objectively ordering these various states in time. Without a manifold of inner states the argument would, consequently, not be able to get off the ground. The fact that Kant is operating with a robust starting point makes his argument susceptible to skeptical doubts. Along with denying knowledge of the external world, the problematic idealist can also deny that the relevant knowledge of inner states is possible. In particular, he can doubt the existence of the manifold of inner states that is to be determined in time, on the grounds that our access to past mental states is inferential and subject to a type of subjective indistinguishability between imagination and memory that is analogous to that in the outer case. In this way, Kant’s Refutation can be circumvented by simply rejecting its substantive starting point. This more radical kind of skeptic countenances only a more minimal starting point consisting in an indubitable basis. The question is thus: why should we suppose that we do have inner experience, or at least suppose that this type of inner experience is possible? On what grounds is Kant allowed to claim (in arguing against the problematic idealist) that inner states can be objectively ordered in time? And what is the significance of an argument based on this claim about time determination? In particular, what is its antiskeptical upshot? Kant’s argument puts substantive pressure on moderate forms of skepticism that are based on an asymmetry between inner and outer sense.

The Refutation of Idealism


These types of skeptics (or problematic idealists) consider inner sense to be unproblematic but cast doubt on outer sense. They accept that inner states exist in time and undergo various changes, but deny that we can know the same of outer states. It is precisely this type of view on which Mendelssohn, Lambert, and Schultz based their criticisms of Kant’s claims about time merely being a form of intuition, which Kant addresses in §7 of the Transcendental Aesthetic (cf. A38/B55 and Bader 2013).6 The idea that the inner is epistemically privileged over the outer is refuted by showing that inner experience presupposes outer experience. However, the argument does not affect more radical forms of skepticism that are willing to accept a wholesale rejection of all beliefs that are not indubitable, doubting inner experience as much as outer experience. Those skeptics who are only willing to countenance the indubitable Cartesian ‘I think’ will simply reject the starting point of the Refutation. Accordingly, it is important to distinguish the project of turning the tables on the problematic idealist (= epistemological skepticism regarding the objects of outer sense) from the project of refuting hyperbolic doubt. The Refutation is merely concerned with the former, attempting to refute the problematic idealist who privileges the inner over the outer and who adopts a form of skepticism that is restricted to the external world due to this supposed inner/outer asymmetry.7 While the robust starting point can be doubted by the radical skeptic, it cannot be called into question by the problematic idealist. Someone who privileges inner over outer sense is precisely someone who doubts the possibility of outer experience while considering the possibility of inner experience to be unproblematic. “Alterations of inner sense or inner experience is [sic] thus admitted by the idealist, and if one wants to refute him then this cannot happen otherwise than by showing him that this inner experience, or which is the same the empirical consciousness of my existence, 6


Dyck has argued that “Kant’s Refutation of Idealism is intended (at least in part) to undermine the Cartesian starting-point Mendelssohn had presumed throughout his campaign against Kantian idealism” (Dyck 2011: 161). It might be objected that Kant identifies problematic idealism as that of Descartes at B274 (likewise at Prol 4:375 and in R6311) and that his target is thus after all the radical skeptic embracing hyperbolic doubt. Kant, however, claims at B275 that Descartes does not doubt inner experience. Given that inner experience can be put into doubt by the hyperbolic skeptic, this suggests that Kant considered Descartes to restrict his skepticism to the outer. Moreover, it is unclear whether Descartes did in fact consistently implement hyperbolic doubt, in particular whether he consistently applied this method to the question of the reliability of memory (cf. Chignell 2010a: 492). Given that even the Refutation is not meant to refute hyperbolic Cartesian skepticism, we can see that interpretations of the Deduction that consider it as an anti-skeptical argument addressed at the radical Cartesian skeptic are a nonstarter (e.g., Strawson 1966).


r alf m . bader

presupposes outer perception” (R6311). The fact that only the outer but not the inner is put into doubt, after all, is precisely the reason why Kant characterizes the type of skeptic that he is addressing as a (problematic) idealist. Thus, although Kant is operating with a robust starting point, this is dialectically unproblematic, given that his refutation is only targeted at a restricted form of skepticism that doubts outer but not inner experience. Accordingly, we can see that worries about the reliability of memory (raised, for instance, by Allison [1983 [2004]: 290], who favors a thin conception of self-knowledge on the basis of these concerns) are misplaced and are not relevant to the project that Kant is engaged in. The target of Kant’s Refutation is not the radical skeptic who is willing to doubt his own memory and countenance the suggestion that he has only come into existence at this very moment, but the person who privileges inner sense over outer sense, considering the former to be immediate and the latter to be mediate and doubtful. Put differently, the issue of contention is not the epistemic standing of inner experience or the reliability of memory, but the question whether the inner is privileged over the outer. Moreover, the reliability of memory is beside the point since Kant is not concerned with a judgment of time determination that is indubitable or immune to error (contra Allison 1983 [2004]: 290). This is because what is at issue is not actual time determination but determinability in time (cf. Bxl). The Refutation is based on a claim about the possibility of objective time determination, that is, about the determinability of one’s existence in time. It is concerned with the necessary preconditions for one’s existence to be objectively determinable in time, identifying the presuppositions that must be satisfied for inner experience to be possible. Put differently, the Refutation concerns the conditions that make inner experience possible rather than those that make it actual. What is at issue is the possibility rather than actuality of inner experience.8 This implies that all that the argument requires is the claim that inner experience is possible, i.e., that inner states can be objectively determined in time. There is no need to appeal to the idea that the subject has in fact objectively determined them or that the subject is conscious of their objective determination. There is hence no need to bring in knowledge of the history of the subject’s mental states.9 Because of this, Kant does not need to claim that one has infallible access to one’s past inner states. More generally, Kant can allow that we are fallible when it comes to determining our 8 9

Cf. “In this way the possibility of inner experience presupposes the reality of outer sense” (R6311). Contra, e.g., Bennett (1966: 205), Gochnauer (1974: 198), and Emundts (2010: 170).

The Refutation of Idealism


existence in time, i.e., that we can be wrong about the ordering of inner states.10 All that the Refutation requires is that it is in principle possible to establish the objective ordering. Fallibility is thus not a problem. Instead, the only problem is the hyperbolic skeptic who even doubts that one existed in the past and who is only willing to grant the existence of one’s present mental states. Such a skeptic does not countenance any diachronic manifold, which implies that there is nothing that can be ordered and determined in time and thereby ensures that the Refutation cannot get off the ground. That determinability has to be at issue can be seen from the fact that the Refutation attempts to establish that inner experience is parasitic on outer experience, such that we only have inner experience if we have outer experience. This means that any actual objective ordering of inner states requires an objective ordering of outer states. The ordering of outer states, however, is not something that is simply given, but that has to be established by us, whereby we have to proceed on the basis of the principles outlined in the Analogies and made determinate in the Metaphysical Foundations of Natural Science (MF). This time determination presupposes the general claim that there is outer experience.11 Once the possibility of outer experience has been established, we can proceed from possibility to actuality and inquire as to which particular perceptions belong to it, i.e., which are due to imagination and which due to outer sense. As Kant notes, the project of classifying particular outer representations presupposes that there is outer experience as established by the Refutation, i.e., “the proposition that there actually is outer experience must always lie at the basis” (cf. Bxli). This means that the rules of experience (specified in the Analogies) only apply once it has already been established that there is outer experience.12 The Refutation thus cannot be based on an actual ordering. This is because an inner ordering presupposes a corresponding outer ordering. Yet, 10 11


Since the inner ordering is parasitic on the outer ordering, and since we can clearly be wrong about the outer ordering, it follows that we can also be wrong about the inner ordering. In addition, objective time determination can only be achieved in the limit, because it depends, among other things, on empirical laws which are in an important sense holistic, since they are constitutive of the whole of experience. It is in this way that idealism poses a threat to the applicability of these rules by questioning their underlying presupposition of the actuality of outer experience (cf. B274), which explains the location of the Refutation in the B-edition after the Second Postulate, which proceeds from what is actual. This also means that one cannot refute idealism by simply appealing to the rules of experience and that we, consequently, have to reject the view that the Refutation does not go beyond the Analogies, contra Abela (2002: Section 3.2) and Emundts (2010: 182). In order to apply these rules, one already needs to have established ‘that there actually is outer experience,’ that is, the conclusion of the Refutation must already be established.


r alf m . bader

such an outer ordering, in turn, presupposes the general claim that there is outer experience, where this is established via the general dependence claim that the inner depends on the outer (together with the possibility of inner experience) which is the conclusion of the Refutation. Any attempt to base the Refutation on an actual ordering would consequently be circular, presupposing precisely that which the Refutation is meant to establish. Instead, the Refutation is based on a determinability claim (i.e., it proceeds from the possibility of inner experience rather than from any actual time determination), on the basis of which it establishes a general connection between the inner and the outer. Rather than starting with the claim that one actually has inner experience, the argument is based on the claim that one’s existence in time is objectively determinable, i.e., that inner experience is possible, and that this possibility presupposes the existence of something permanent in perception, where one can then show that this permanent has to be outer. To be conscious of one’s existence as determined in time is thus to be conscious that one is a temporal entity, i.e., that the manifold of inner states that constitutes one’s existence (i.e., that constitutes the determinable self ) is a temporal manifold. The various inner states making up this manifold are in time and change in time. This manifold of states when objectively ordered constitutes inner experience. Kant does not make any claim to the effect that one is conscious of this objective ordering. Instead, it is merely claimed that this ordering can be objectively determined, i.e., that inner experience is possible. Determinability of this ordering is meant to follow straightforwardly: “consciousness of my existence in time is necessarily linked with consciousness of the possibility of this time determination” (B276). How exactly this is supposed to follow is not made clear by Kant (the idea that what is in time can be determined in time could well be based on Kant’s commitments regarding the ideality of time). Yet, the details do not matter all that much since the existence, as well as the objective determinability, of one’s various inner states that are in time and that change in time is, at any rate, admitted by the problematic idealist. (After all, this type of idealist is only concerned about outer sense being problematic and doubtful.) It is from this starting point that the Refutation begins, i.e., from the claim that there are various inner states that can be objectively ordered in time. What Kant wishes to establish then is that, instead of the inner being privileged over the outer, the outer is prior to the inner, such that it is not coherent to grant the possibility of inner experience without also granting the possibility of outer experience. This is argued on the basis that the ordering of inner states can only be objectively determined with reference

The Refutation of Idealism


to an objective ordering of outer states, which implies that the objective determinability of inner states (= the possibility of inner experience) presupposes that there are outer states that can be objectively ordered (= the possibility of outer experience).


The Need for Permanence

All time determination presupposes something permanent in perception.

Having set out his starting point to the effect that there are inner states that can be objectively determined in time, Kant now turns to the presuppositions of objective time determination, with the goal of showing that these presuppositions turn out to require the existence of objects of outer sense, thereby refuting the problematic idealist who doubts their existence while nonetheless accepting the objective determinability of inner states in time. The presupposition of time determination to which Kant appeals in the Refutation of Idealism is the need for the existence of something permanent in perception.13 This requirement was established in the First Analogy, where it was argued that substance, which is schematized as the permanent (cf. A144/B183), has to underlie all time determination and has to satisfy a conservation principle insofar as the quantum of substance can neither increase nor diminish. In the A-edition the First Analogy is stated as follows: “All appearances contain the permanent (substance) as the object itself and the changeable as its mere determination, i.e. as a way in which the object exists” (A182). Here substance is understood as that which underlies all changes, so that changing appearances are understood as mere determinations of the underlying substance. When applied to inner appearances problems would seem to arise. This is because inner appearances would have to reside in an inner substance and would have to be mere determinations of this substance. Yet, the argument of the Refutation relies precisely on the fact that inner permanence and hence substantiality cannot be established and that applying the category of substance to the self gives rise to problems. There would thus seem to be an inconsistency between the Refutation and the First Analogy as stated in the A-edition.14 13 14

Permanence is to be understood in the sense of absolute not relative permanence, i.e., not just something that persists through some stretch of time but through all time (cf. A182/B225–26). Cf. Guyer (1987: 284) for a somewhat related concern regarding what he calls the ‘analytical argument’ interpretation of the First Analogy, according to which it is analytic of an alteration that it is a change in the mode of an enduring substance.


r alf m . bader

This problem is circumvented by the B-edition version of the statement of the principle of the permanence of substance: “In all change of appearances substance is permanent and its quantum in nature is neither increased nor diminished” (B224). The difference between the two editions involves a reconceptualization of the role of substance in time determination, in particular a switch from considering substance as that which underlies all changes to that which underlies all time determination, thereby allowing not only for time determination that is direct insofar as that which is determined in time inheres in the substance but also for indirect time determination. It thus seems plausible to suppose that this reformulation was connected to Kant’s work on the Refutation.15 This is supported by the marginal notes attached to A182–83 that Kant made in his own copy of the A-edition.16 “Here the proof must be conducted so that it applies only to substances as phenomena of outer senses” (R LXXX E 32, 23:30). Likewise, according to R LXXVII (E 31, 23:30), it needs to be shown that the First Analogy pertains only to substances the alterations of which are brought about by moving forces and consist in nothing but motions (also cf. R LXXXI E 32, 23:30– 31). Most notably, R LXXXIII explicitly states that the proof has restricted applicability and that time determination is parasitic in those cases where the First Analogy does not apply, insofar as one needs to estimate duration by reference to outer objects, whereby he includes one’s own existence as such a case. Kant concludes this marginal note as follows: “my permanence is therefore not proven” (E 32–33, 23:31).17 These notes suggest that Kant was attentive to the problem that the A-edition version was too broad in scope and conflicted with the Refutation.18 15 16



Together with his argument for the conservation principle in the Metaphysical Foundations of Natural Science, cf. Friedman (2013: 315n61). Kant’s notes in his copy of the first Critique (A-edition) were published in Benno Erdmann’s Nachträge zu Kants Kritik der reinen Vernunft in 1881 (Kiel: Lipsius & Tischer) and later in volume 23 of the Akademie edition (citing “E” for pages in Erdmann). They are also included as footnotes to the relevant passages in the Cambridge Edition of the Critique of Pure Reason. Kant also notes the seeming conflict between the First Analogy and the claim that the permanence of the soul cannot be established in Metaphysik K2 after outlining his refutation of Mendelssohn’s proof of the permanence of the soul. “It seems contradictory to this way of presenting things that in all alterations in nature substance persists and only the accidents change. But here we are talking merely of bodily substances which we cognize, yet in the case of the human soul we do not cognize anything permanent” (Metaphysik K2 28:764, LM 404). Kant came to realize more generally that the A-edition versions of the three Analogies were problematic since they had troublesome implications when applied to the inner and that all of them consequently had to be restricted to outer objects. The Second Analogy in the A-edition implies that there are laws of inner sense governing the changes of inner states, given that it is phrased in terms of ‘everything that happens’ which is unrestricted and applies equally to inner happenings.

The Refutation of Idealism


The existence of permanent substance is thus a precondition of the possibility of objective time determination in accordance with the First Analogy, whereby the determination of the temporal ordering of inner states does not require these states to be mere determinations of a permanent underlying substance but can proceed indirectly by correlating inner states with outer states, thereby rendering the ordering of inner states parasitic on the objective ordering of outer states (cf. B156 and B293). In this way, the principle of the First Analogy is compatible with a situation in which time determination of inner states is possible even though permanence is restricted to the outer realm. In such a situation, outer appearances are nothing but determinations of the underlying material substance, while inner appearances need not be thought of as inhering in a substance.

11.4 A Critical Gap? But this permanent something cannot be an intuition within me.

Time determination thus requires something permanent in perception. The next step of the argument consists in showing that the requisite permanent cannot be inner but has to be outer. If this can be established, then the problematic idealist is refuted, since he will have to posit an outer permanent in order to make possible the determinability in time of inner states that he admits. This step consequently turns out to be crucial. Yet, it is also highly controversial. In fact, one of the most common criticisms of the Refutation of Idealism is that Kant has not managed to establish that an outer permanent is required. In particular, it is claimed that he has not succeeded in excluding the possibility that the self is the only existing permanent and that we can determine the temporal ordering of our inner states by reference to this permanent self. Guyer, for instance claims that “there is no obvious reason why such a substratum, even if required, would have to be a substance which is either spatial in form or ontologically independent from the self” (Guyer 1983: 333). Likewise, Walker asks “why should I not regard my various The B-edition statement, by contrast, is restricted to alterations, i.e., to changes of determinations of substances, requiring only these to be law-governed. Similarly, whereas the A-edition version of the Third Analogy is meant to hold for all substances that are simultaneous (which is problematic since space is the medium of interaction yet inner states are only in time cf. R LXXXVI E 34, 23:32– 33), the B-edition version is restricted to all substances insofar as they can be perceived in space as simultaneous. There is thus a general asymmetry between the inner and the outer that ensures that duration, succession and simultaneity of inner states are not determined directly by means of the three Analogies but only indirectly.


r alf m . bader

representations as all modifications of one permanent phenomenal self, not unknowable but known through them?” (Walker 1978: 114). Vogel goes so far as to call this “the critical gap in Kant’s argument” (Vogel 1993: 876). The First Analogy establishes that substance is permanent and is necessary for objective time determination. The question now is: what can be subsumed under the (schematized) category of substance? To which objects can we apply this concept? In particular, the question is whether it can equally be applied to inner as to outer objects. Kant’s argument in the Refutation is based on the idea that only the latter but not former is possible. If this category can only be applied to outer objects, then the need for permanence will amount to a need for outer substance, which is what the Refutation is trying to establish. To see whether the ‘critical gap’ in Kant’s argument can be filled, we thus need to find out whether inner objects can be subsumed under the category of substance. The crucial point now is that one needs to establish that something is permanent in order to apply the category of substance. Although we need something permanent in perception, this permanent is not perceived as being permanent (“permanence is not obtained from outer experience” B278), i.e., perception cannot inform us that the sensible conditions for the application of the schematized category of substance are satisfied.19 Since the permanence of substance is not perceived, but is an a priori necessary condition of experience, permanence has to be proven. In short, permanence is not revealed in perception but is proven. This means that, rather than determining that something is a substance and hence inferring its permanence, we need to prove that it is permanent in order to subsume it under the category of substance. As Kant states at A403, the condition of the application of the category of substance in concreto is that one first needs to establish (“voraus festzusetzen”) the permanence of the object in question, and that permanence needs to be laid at the basis (“zum Grunde zu legen”) of the application (also cf. A365). Kant also makes this clear when considering whether the self can be called a substance in §47 of the Prolegomena. This thinking self (the soul) may now however, as the ultimate subject of thinking, which itself cannot be represented as predicate of another thing, be called substance: yet this concept remains in this way completely empty and without any consequences, if permanence, as that which makes the concept of substances fertile in experience, cannot be proven of it. (Prol 4:334) 19

This shows that the First Analogy is not based on the claim that something permanent (what Allison calls a permanent backdrop) is required in order to be able to perceive change, contra, e.g., Paton (1936: 196) and Allison (1983 [2004]: 239).

The Refutation of Idealism


The crux of the argument then is that permanence can only be proven for outer but not for inner objects. It is this difference between the inner and the outer that allows Kant to fill the so-called critical gap. In particular, it can be filled without having to deny the possibility of inner permanence and without having to assert that the self is not a substance.20 Instead, all that needs to be done is to establish that one cannot prove the permanence of the self, thereby ensuring that one cannot be justified in applying the category of substance to the inner. As Kant notes at A403, applying the category of substance to the self amounts to making an inadmissible application (“unzulässiger Gebrauch”) of this category, since permanence cannot be established of the self. By ruling out a valid application of this category to the self, one undermines the possibility of treating it as a substance for the purposes of objective time determination. In order for the permanent to play the requisite role in time determination, it must be known to be permanent and hence subsumable under the category of substance, where this can only be established by means of a proof of permanence. In other words, Kant’s claim is not that the self cannot be permanent, but only that we cannot prove its permanence and that it, accordingly, cannot play a role in time determination, which suffices for the Refutation to succeed.21 The difference between the inner and the outer on which the Refutation relies is that one can prove permanence for outer objects (for matter) and hence subsume them under the schematized category of substance, but that one cannot likewise prove permanence for inner objects (for the mind). This difference is due to the fact that any decrease in the intensive magnitude (reality) of outer objects goes together with a compensating adjustment elsewhere in space, such that if the intensive magnitude of some objects is diminished, then that of others is increased in a way that conserves the overall quantity of matter.22 There is thus a necessary 20



If a denial of inner permanence were required for the Refutation to succeed, then there would be a conflict with the postulate of the immortality of the soul established in the Critique of Practical Reason, as well as with the regulative use of the ideas of reason insofar as the idea of the soul would not be able to play its regulative role and function as a focus imaginarius, i.e., a denial of inner permanence would contradict Kant’s assertion that “there is not the slightest thing which hinders us from assuming these ideas also as objective and hypostatic” (A673/B701). If the rational psychologist’s proof of the substantiality of the soul were to succeed, then permanence of the soul would follow (cf. B417). One would thereby remove the resources for the only possible refutation of problematic idealism, insofar as outer objects would then no longer be required for time determination. It is for this reason that “idealism, at least problematic idealism, is inevitable in this rationalist system, and that when the existence of outer objects are not at all required for the determination of one’s existence in time, then it can only be assumed entirely gratuitously, without ever being able to give a proof thereof” (B418). A decrease in intensive magnitude consists in a reduction in density resulting from expansion, i.e., a given quantity of matter takes up a larger volume of space, and this expansion leads to a compression


r alf m . bader

interdependence between intensive and extensive magnitudes (cf. MF 4:539–40) that allows Kant to prove a conservation principle, namely: “First Law of Mechanics. In all changes of corporeal nature the total quantity of matter remains the same, neither increased nor diminished” (MF 4:541). Material substance thus turns out to be permanent (cf. Friedman 2013: 311–35). This can be established because the quantity of matter is given by parts external to one another in space, which allows for the requisite interdependence between the extensive and intensive magnitudes, such that it is not possible to have a diminution in intensive magnitude in one region without a corresponding increase elsewhere. An analogous argument cannot be given in the case of the self. This is because the self lacks the kind of extensive structure that matter has, due to the fact that it is not in space but only in time, and thus does not have the requisite quantitative structure to establish a conservation principle and prove the permanence of the self (cf. MF 4:542–43). “In the soul no quantum of substance is possible” (R LXXXIV E 33; 23:31). Instead, the self is only characterized by an intensive magnitude at any moment in time, making it possible for it to go out of existence as a result of this magnitude continuously fading out until it reaches zero, as argued in the Remark to the Proof of the First Law of Mechanics and in the Refutation of Mendelssohn’s Proof of the Permanence of the Soul in the B-Paralogisms (cf. B413–15).23 In this way, the one-dimensionality of time is responsible for the fact that no inner permanence can be proven (cf. MF 4:471). Space but not time has the relevant mathematical structure to allow for a conservation principle to be established, ensuring that we only have a proof of outer permanence but not of inner permanence.24



of matter elsewhere and hence to a corresponding increase in intensive magnitude, i.e., an increase in density. Expansion and compression in this way go together. Since Mendelssohn recognized that a simple soul could go out of existence, not by disintegration (i.e., by removal of parts), but by vanishing, his proof of the permanence of the soul was based on the claim that this kind of vanishing would violate the law of continuity. Accordingly, it is crucial for Kant to show that this magnitude can fade out continuously. Kant clearly was aware of this: cf. Metaphysik Mrongovius 29:912, LM 277–78 (also cf. Metaphysik K2 28:763–64, LM 404), contra Falkenstein (1998: appendix). The proof of the permanence of material substance is based on the interdependence between the intensive and extensive magnitudes of matter. Falkenstein (1998) has raised an important problem for this argument. His objection is based on the fact that there are two ways in which objects can differ in terms of their intensive magnitudes, namely, (1) in terms of compression/expansion and (2) in terms of their specific densities (cf. MF 4:533–34). The problem is that the conservation argument is restricted to the former case and does not rule out the possibility of there being a diminution in terms of the latter quantity. Accordingly, it would appear that, in the same way that the self can fade out of existence, matter can go out of existence by a diminution of its specific density. Friedman has attempted to respond to this objection by claiming that specific densities cannot change. “Kant’s

The Refutation of Idealism


The Refutation is, accordingly, not based on the claim that inner permanence is not possible, nor on the claim that inner permanence is not perceived, given that permanence cannot be perceived at all (not even in the case of the outer). Rather, it is based on the claim that permanence can only be proven for matter, which is why Kant holds that matter is the only thing in perception that is known to be permanent and that is subsumable under the concept of substance (cf. B278 and B291). The relevant difference is thus not at the level of what is perceived but at level of what can be proven.25 As a result, the category of substance is applicable to the outer but not the inner. For this reason, time determination can only proceed by reference to the outer. As a result, the temporal ordering of inner states becomes parasitic on the objective ordering of outer states, and inner experience turns out to presuppose the existence of an outer permanent.


The Tables Turned

Hence determination of my existence in time is possible only through the existence of actual things that I perceive outside me.

Determining one’s existence in time, i.e., objectively ordering the manifold of inner states in time, presupposes an outer permanent, given that permanence cannot be proven of anything inner, thereby rendering the inner ordering parasitic on the ordering of outer states. Accordingly, if inner experience is to be possible, there have to be actual outer things to which one stands in perceptual relations such that one can establish the objective outer ordering.


First Law of Mechanics does not allow for changes of intrinsic specific densities – changes that would transform one type of matter into another – as physically possible changes of matter. The only physically possible density changes, therefore, result from compressions or expansions of a given type of matter” (Friedman 2013: 322n72). This response appears to be problematic. First, it is not explained how this law of mechanics is meant to preclude the possibility of such changes. Second, it implies that there is a necessary distribution of different types of matter, which seems somewhat implausible. A better response would be to find an analogous connection between intensive and extensive magnitudes in the case of specific densities, such that the decrease in the specific density of one object is necessarily accompanied by a corresponding increase in the specific and/or compression densities of other objects, which would allow for necessary conservation without the requirement of a necessary distribution of types of matter. (Tal Glezer has pointed out to me that the necessary distribution of types of matter might not be all that problematic, on the basis that the variety in types of matter simply consists in the variety in fundamental repulsive forces. Since these forces are fundamental, corresponding to fundamental causal laws, the forces themselves cannot change, such that density can only vary due to compression and expansion.) We can thus see, contra Vogel, that it is not the case that “[t]o complete the Refutation, Kant needs to establish some disparity between inner and outer sense, such that outer sense gives us direct knowledge of enduring objects, while inner sense does not” (Vogel 1993: 878).


ralf m. bader

This conclusion is a general claim regarding the presuppositions that must be satisfied for inner experience to be possible. The upshot of the Refutation is that the determinability of one’s existence in time turns out to require the existence of outer objects. In particular, there must be material substance, of which one can prove permanence. Accordingly, what one can establish is that there is material substance. However, although one can in this way establish the existence of outer objects to which one is perceptually related (assuming that inner experience is in fact possible, which is granted by the problematic idealist), one cannot establish which outer representations are in fact the result of outer intuition (i.e., which supposed outer objects are real and which ones merely imaginary). That is, we cannot establish that any particular outer representation does constitute an outer intuition rather than an outer imagination. This means that, while we can establish that outer objects do exist, we are not able to establish the existence of any particular outer object. As Kant notes: From the fact that the existence of outer objects is required for the possibility of a determinate consciousness of ourselves, it does not follow that every intuitive representation of outer objects at the same time implies the existence thereof, because it may well be the mere effect of the power of imagination (in dreams as well as in madness). . . . What was here to be proved is only that inner experience in general is possible only through outer experience in general. (B278–79)

As regards particular objects, all we can do is to appeal to the rules of experience to determine which representations can be integrated into a systematic whole of experience. “For which given intuitions there are corresponding objects that actually exist outside me, and that accordingly belong to outer sense, which are to be ascribed to it and not to the faculty imagination, has to be established in each particular case in accordance with the rules on the basis of which experience in general (even inner experience) is distinguished from imagination” (Bxli; also cf. Loses Blatt Leningrad 1, NF 364–66). Yet, this kind of procedure is inferential and only provides defeasible evidence that is not beyond doubt. This means that we are not able to address the concerns of the problematic idealist when it comes to particular outer objects. With respect to any particular outer representation we cannot rule out the possibility that it is the result of the imagination, making the existence of particular outer objects doubtful. The existence of outer objects in general, however, has been conclusively established. Moreover, analogous problems also arise when it comes to ordering inner states

The Refutation of Idealism


and integrating them into inner experience, undermining any claim to that effect that inner experience is privileged over outer experience.26



Thus, Kant’s Refutation of Idealism is concerned with identifying the preconditions that must be satisfied for inner experience to be possible. The argument begins with a relatively robust starting point consisting in the determinability of one’s existence in time. Although it does not require any reliance on memory or the like, it is nonetheless open to skeptical doubt. This, however, is unproblematic, given that it is not addressed at the radical Cartesian skeptic who is prepared to doubt inner sense but only against the person who privileges the inner over the outer. Since something permanent is necessary for objective time determination (as was established in the First Analogy), the determinability of one’s existence in time presupposes something permanent in perception. The requisite permanent has to be outer and cannot be inner because what is at issue is proving permanence, which can only be done for material substance. In this way, the supposed critical gap in Kant’s Refutation can be filled. As a result, inner experience turns out to presuppose outer experience and to be only possible given the existence of material substance. A noteworthy upshot of this line of argument is that Kant could have equally refuted problematic idealism by appealing to the Second Analogy (and likewise for the Third Analogy27 ). Instead of relying on the claim that something permanent is required for objective time determination to be possible, where this permanent has to be outer rather than inner, he could have appealed to the claim, established in the Second Analogy, that causal laws are required if objective time determination is to be possible. Inner experience can then be seen to presuppose outer experience on the basis that causal laws are required for time determination, together with 26


There is some asymmetry between inner and outer sense, insofar as it is beyond doubt that any present representation is part of inner experience. That is, no subjective indistinguishability argument can be raised in this case, since there is no such thing as inner imagination when it comes to present representations. (The imagination only gives rise to difficulties insofar as it can produce false memories of past mental states, which ensures that claims regarding past states are inferential in nature and can be put into doubt.) Accordingly, given that the inner representation is the very inner object at issue, there is no need for a problematic inference from (present) inner representation to inner object that is analogous to that from outer representation to outer object. Nevertheless, even in the case of present representations, the issue as to how precisely they fit into the whole of inner experience is analogous to that in the case of outer experience. In R6312 Kant sketches an argument based on simultaneity.


ralf m. bader

the claim that there are only laws of outer sense but not of inner sense, which follows from Kant’s claim that no science of inner sense is possible (MF 4:470–71). The need for causal laws for objective time determination corresponds to the need for something permanent, while the impossibility of establishing laws of inner sense corresponds to the impossibility of proving inner permanence, i.e., in the same way that we can only prove outer but not inner permanence, we can only establish laws of outer but not of inner sense. This suggestion is supported by the fact that the General Note on the System of the Principles, which Kant considers to be a confirmation of his Refutation of Idealism (cf. B293), exhibits the dependence of inner on outer sense with respect to all three relational categories. In this way, objective time determination of inner states as such, independently of whether it is concerned with duration (substance), succession (causality), or simultaneity (reciprocity), turns out to presuppose outer experience.28 28

Kant notes in the B-Preface that his refutation is the only possible proof (Bxxxix fn). The suggestion that Kant could have devised variants based on the Second and Third Analogy does not conflict with this claim, since these variants all essentially involve the same proof strategy based on identifying necessary conditions on objective time determination of one’s own existence, i.e., conditions on inner experience. They all refute problematic idealism by showing that inner experience presupposes outer experience – they simply differ in terms of whether they are concerned with the conditions on time determination treated in the First, Second, or Third Analogy.

ch a p ter 1 2

The Antinomies An Entirely Natural Antithetic of Human Reason Graham Bird

Kant refers, in his own terminology, to the traditional conflicts outlined in the Antinomies as an “entirely natural antithetic of human reason” (B 433). The terminology reflects his central aim in the Antinomies to resolve issues where reason inevitably “comes into conflict with itself” (A xii–xiii) and “precipitates itself into darkness and contradictions” (A viii). That project is an essential part of his wish to reform philosophy by laying bare the underlying errors which have encouraged the futile pursuit of these apparently insoluble conflicts. The upshot of Kant’s account, however, is not negatively to reject reason but only to restrict it by recognizing more positively its legitimate function. That conclusion is captured in his claim that reason in this context has only a “regulative,” but not a “constitutive,” role. I make the assumption that the aim is to make philosophical sense of Kant’s account of the Antinomies in the Critique of Pure Reason. I therefore say little about the historical antecedents to Kant’s discussion, which have been carefully outlined in such work as Laywine (1993), Hinske (1998), and Grier (2001). In what follows I first offer in Section 12.1 a general account of the structure of Kant’s account, which includes both the arguments given for the theses and antitheses and his response in resolving the conflicts. In Section 12.2 I examine the first Antinomy and in Section 12.3 the third Antinomy; finally, in Section 12.4, I draw some general conclusions about the character of Kant’s resolution of the conflicts, and his appeal to “human reason.”


The Structure of Kant’s Account

Formally the structure of Kant’s discussion is that arguments are given to represent the grounds for each of the opposed conclusions (in the first Antinomy that the universe is limited (finite) or unlimited (infinite) in space and time; or in the third that freedom does, or does not, exist as well as 223


graham bird

natural causality). Kant then makes some “observations” on each of the given arguments, and in his final discussions offers a resolution of the conflicts. Whether in the end he regards the opposed conclusions as both false, or as both true, his resolution depends in every case on demonstrating an ambiguity in the thesis and antithesis arguments. The ambiguity in question, between treating items as phenomena (appearances) or as noumena (things in themselves) forms a fundamental part of Kant’s Critical philosophy.1 Kant claims that once the ambiguity is acknowledged the apparent conflict disappears, and the whole “experiment” on which the Critique of Pure Reason has embarked is shown to have succeeded (Bxviii ff.). The point and the terminology are made clear in the following: If on the supposition that our empirical knowledge conforms to objects as things in themselves we find that the unconditioned cannot be thought without contradiction, and that when on the other hand we suppose that our representation of things, as they are given to us, does not conform to these things as they are in themselves but that these objects, as appearances, conform to our mode of representation, the contradiction vanishes . . . we are justified in concluding that what we first assumed for the purposes of the experiment is now definitely confirmed. (Bxx)

The resolution of the Antinomies is evidently central to Kant’s whole project in the first Critique. Even though Kant claims to have established the same conclusions (that there can be knowledge a priori; that we can never transcend the limits of possible experience (Bxviii)) in the earlier parts of the Critique, nevertheless this result from the Dialectic (that we can never have knowledge of a “supersensible unconditioned,” and must restrict our knowledge to “possible experience”) is fundamentally important in his argument. I make the following summary comments. (1) Kant’s reference to the “unconditioned” in that Preface passage is elaborated later in the introduction to the Dialectic (B364–65). There it is used as a central conception, lying behind all the dialectical illusions, linked to reason’s unavoidable drive toward what transcends the merely conditioned aspects of our human experience. To recognize those conditions in experience is, Kant suggests, to be drawn to a conception of what is unconditioned, just as our limited human, and sensible nature is inevitably drawn to a conception of what goes 1

Note: I write “phenomena” for “appearance” and “noumena” for “things in themselves” throughout, and have used Kemp Smith’s (1929) translation of CPR.

The Antinomies


beyond such limitations (see Bird 2006: 551, 758, 763; see also 411, 611–14, 686). But the quoted passage makes it quite clear that we have no hope of attaining knowledge of any such transcendent unconditioned. We have an inevitable conception of it, are drawn toward a wish to know it, and may even be persuaded that philosophy can sometimes attain such knowledge with suitable thought and logical argument; but the primary lesson of the Critique is that we cannot fulfill that wish to transcend possible experience except conceptually, or “in Idea” as Kant puts it. The Dialectic as a whole in its negative part, including the Antinomies, is an extended catalogue of the errors philosophers have made in their doomed attempts to transcend our experience. The Dialectic’s positive conclusion outlines a more modest, legitimate role which the conception of the unconditioned in its various specific forms may achieve. In Kant’s terminology the conception of the unconditioned has no “constitutive” role, but can fulfill a less ambitious “regulative” role. (2) Two questions arise about the thesis and antithesis arguments used by Kant to represent the opposed claims. A more general question asks how far Kant actually endorses the competing arguments, but it might also be asked more specifically how he evaluates each of the arguments. The given arguments for the thesis and antithesis are evidently ‘Kant’s arguments’ in that he has chosen them to represent each of the positions, and to do so, as his Observations make clear (B459– 61), as fairly as he can. At B459, for example, he explains that he avoids a formulation of the thesis argument which he regards as “sophistry”; and at B468–69 he similarly explains his rejection of a formulation for the antithesis of the second Antinomy as “playing tricks with mathematics.” Evidently Kant’s chosen formulations of the arguments are designed to avoid what he regards as sophistry (metaphysical juggling (B88)). But the given arguments are not ‘Kant’s own arguments’ in the stronger sense that he is committed to endorsing them (Wood 2010: 252). Rather they are arguments which can properly represent a serious view prompted by reason and held, for example, by “dogmatists” (B459) or “monadologists” (B470). Moreover it is plain that although in his formulation the arguments offer a serious case for their respective conclusions Kant believes that they are in some way fatally flawed. They all rest, as he explains in his resolution, on a confusion or ambiguity between phenomena and noumena. (3) The more specific question about Kant’s evaluation of the given thesis and antithesis arguments is less easy to answer, for his comments


graham bird cover a wide variety of such evaluations. I offer a list, each item of which should be qualified by the general recognition that something goes wrong in all the arguments: (a) (B449) Both opposed claims have grounds that are just as valid and necessary. (b) (B529) . . . although they have failed to support their contentions by valid grounds of proof, nothing seems to be clearer than that since one asserts that the world has a beginning and the other that it has no beginning . . . one of the two must be right. ((B531–32): Kant distinguishes “analytic” and “dialectical” opposites, and this allows the dialectically opposite conclusions of the mathematical Antinomies to be both false since they rest on the same erroneous assumption (given in (d)).) (c) (B529) The very fact of their being able so admirably to refute one another is evidence that they are really quarreling about nothing, and that a certain transcendental illusion has mocked them with a reality where none is to be found. (d) (B535) . . . the proofs are not merely baseless deceptions. On the supposition that appearances and the sensible world . . . are things in themselves these proofs are well grounded. (e) (Prol. 4:339) . . . because thesis and antithesis can be supported by equally transparently clear and irresistible proofs – for I give my pledge for the correctness (Richtigkeit) of the arguments.

These comments may seem inconsistent, but some of the apparent inconsistencies can be easily set aside. When Kant claims that the proofs are “well grounded” and offers his pledge that the arguments are “correct” it is reasonable to accept that this holds just for the proofs on the supposition in (d), which Kant rejects, that “appearances and the sensible world are things in themselves.” This allows him to separate the correctness of the arguments on that supposition from the ambiguity which for Kant makes them fallacious. (c) is a reminder of the form of both arguments in which their positive conclusions rest on a prior negative claim that the opposition view is mistaken, a result that Kant thinks both sides “achieve admirably.” This brings to light another formal assumption underlying both proofs, namely, that there are just two alternative options between which we must choose: either the world is (spatiotemporally) limited or else it is not so limited; either freedom exists or it does not. It may seem more difficult to reconcile these claims about the qualified correctness

The Antinomies


of the arguments with the claim that both are “equally valid and necessary,” but perhaps this is to say no more than that the proofs are correct on the contested supposition. These comments show that in writing out the positions of each protagonist several assumptions are needed to preface the given arguments. The given arguments themselves need to make clear that their conclusions have the form: the only alternative can be rejected, and therefore the original claim must be true. These features will be explicitly included in the examples discussed later of the first and third Antinomies. (4) I add a further comment about the basis for Kant’s response to the given thesis and antithesis, which is already made clear before the Dialectic. It should be remembered that the Analytic and Dialectic are each subsections of what Kant calls “transcendental logic” introduced at B74–88. In that introduction Kant makes two distinctions with an explicit link to his treatment of the dialectical illusions. First he distinguishes between “general,” that is, formal, logic, which provides the basic discipline for “forms of thought” (B79), and a “transcendental logic” which is directed specifically at the concepts, claims, and inferences involved in a “transcendental” philosophy. General logic “abstracts from all content of knowledge” in examining forms of thought and their rules, and does not consider the “origins” (or we might say the “status” as a priori or empirical) of any representations involved with forms of thought. Transcendental logic, on the other hand, does not abstract from content and does consider the origins, or status, of any representations (e.g., as a priori) occurring within its scope. Second Kant also distinguishes between two ways of understanding, or using, logic, namely, as a legitimate, formal “canon” or as an illegitimate, material “organon” (B84–86). The former, correct, use is confined to establishing the consistency or inconsistency of some conceptual connection, for example in inference, but the latter is intended to establish actual, material truth. For Kant that illegitimate use is always a “logic of illusion” and “dialectical” (B86), and he insists that his use of that latter term is always as a “critique” of that illusion. It should be noted that this contrast between a correct use of logic as a canon and its incorrect use as an organon is applied both to general and to transcendental logic. In a brief passage at B87–88 Kant outlines the general fallacy of treating transcendental logic as an organon:


graham bird But since it is very tempting to use these pure modes of the understanding and these principles by themselves, and even beyond the limits of experience, which alone can yield the matter (objects) to which these pure concepts of understanding can be applied, the understanding is led to incur the risk of making, with a mere show of rationality, a material use of its pure and merely formal principles, and of passing judgements upon objects without distinction – upon objects which are not given to us, and perhaps cannot be given in any way. (B87–88)

Kant goes on to associate this error with the later Dialectic. Put briefly the error consists of inferring material claims about truth from a purely formal, logical appeal to consistency, or inferring material truth solely from the logic of concepts; it is powerfully illustrated in the sphere of general logic by Kant’s trenchant criticisms of Leibniz’s monadology in the Amphiboly (B316–49, especially B320–23 and B336–37). In the transcendental context those logical fallacies are made worse by the risk of claiming material truth for claims about objects, noumena, which are strictly beyond our cognitive reach. Kant makes clear again that his goal is not to indulge in such an erroneous dialectic but only to criticize it, and so to “expose the false, illusory character of those groundless pretensions” (B88). The quoted passage provides the best short summary of dialectical error, as Kant understands, and later extensively documents, it. No account of his discussion of the errors in the Antinomies can be adequate without such a reference.

12.2 12.2.1

The First Antinomy The First Antinomy Proofs

The thesis and antithesis arguments concern the limits, or limitlessness, of the physical universe (PU) (Kant uses “Welt” / “world”) in time and space. I consider only the arguments for and against temporal limits, that is whether the universe has a beginning in time or has no such beginning. I preface the given, abbreviated, arguments with two assumptions, noted above as implicit, or explicit in those arguments. There is a further general assumption in both proofs, namely, that “the PU” (or “PU as a whole”) is unequivocal, but that is not explicitly recognized, or questioned, in the proofs themselves but only in Kant’s resolution of the issue. Implicit Assumption (1): The conception of “the PU” (or “PU as a whole”) is unequivocal.

The Antinomies


Explicit Assumption (2): There are only two options – either the world is limited in time or it is not. If it can be demonstrated that one alternative is impossible, then the other must be true. Thesis Proof

Antithesis Proof

(3) Assume the world has no beginning (4) In that case at any given moment an time infinite time has elapsed (been completed) (5) But an infinite completed synthesis is not possible; hence the assumption (3) is impossible; and by (2) the world must have had a beginning.

(3’) Assume the world has a beginning. (4’) In that case there must have been a beforehand when there was nothing – an empty time. (5’) But from an empty time nothing could have arisen.

(6’) Assumption (3’) must be false; by (2) the world must have no beginning. These formulations make clear that the form is that of an indirect proof to argue, along with assumption (2), that the initial hypothesis must in each case be rejected and the preferred alternative accepted. In the thesis the primary issue arises from (4) and (5) and their references to an elapsed infinite time and a completed synthesis. In the antithesis the central issue concerns the status of the claim at (5’) that an empty time cannot explain an occurrence such as the beginning of the PU. I consider each in turn. The argument for the thesis is too brief and inexplicit to give confidence in its validity. It turns first (4/5) on an alleged conceptual incoherence in the notion of a “time elapsed” in the PU on the assumption (3); and second on the inadequacy of any required testing procedure for (3). It might be argued that it is incoherent to talk of a time elapsed to some point now from an infinite past, on the ground that the time elapsed should be, but cannot be, determinate. But it is questionable whether we should assume that “determinate” means “finite,” and Kant himself denies this (B546na). Even if the notion of a “time elapsed” in these circumstances is problematic this does not provide a decisive objection to (3). These inadequacies suggest a need for the supplementary argument in (5) which rests on the outcome of any, actual or notional, testing procedure for (3). Such a procedure would end either in a failure to find any terminus, which would be insufficient


graham bird

to establish (3), or else would falsify (3). Kant himself indicates such an argument at B545–46. These points raise questions about (3), but if Kant requires that the arguments should be demonstrably valid he does not meet that requirement; nor did he expect the disputes to stop (B529–30). The crucial claim in the antithesis is evidently (5’). This has an intuitive appeal and reflects the conceptual problems which even contemporary cosmologists face in advocating a beginning of the PU in a “big bang.” When it is objected that such an event has no antecedent cause to initiate the bang it is replied that the objection makes no sense because before the bang there was no antecedent time to accommodate such a prior cause. If time begins with the bang then we cannot talk of, or ask questions about, a temporally prior trigger to initiate the event. The antithesis argument might be answered in a similar way by rejecting any appeal to the notion of a time before the beginning of the PU, but such answers, and the situation which provokes them, undoubtedly remain problematic. Both thesis and antithesis present problems for their opponents, but are indecisive as they stand. 12.2.2

Kant’s Resolution of the Conflict – Three Factors: Comparative and Absolute “World as a Whole”

Kant’s resolution of the first Antinomy is complex and perhaps made more so by the added claim that it serves to confirm his transcendental idealism. That latter claim, however, turns essentially on Kant’s belief that the resolution requires his transcendental idealist distinction between phenomena and noumena. That distinction is the main point of the following interpretation of Kant’s resolution of the issue. I focus on three related aspects, which are not always noticed or sufficiently emphasized; namely, (i) the explicit ambiguities in conceptions of “the PU as a whole,” (ii) the transition from appearances to a relevant Idea of reason (e.g., in the principle of reason), and (iii) the relevance of Kant’s criticisms of logic (see earlier §1, and quote B88–89). (i) Commentators rightly recognize that Kant regards references to the PU, or to the PU as a whole, as ambiguous. At the heart of his resolution of the Antinomy is the belief that in some way what starts as a discussion of the PU in our experience as a phenomenon turns into a supersensible reference to the PU, or the PU as a whole, as a noumenon. Sometimes that transition is thought to be effected directly in a move from observable features of the PU to an unobservable “PU as a whole,” that is, from, say, some specific claim or physical law in the PU to a claim about the different category of the cosmos

The Antinomies


as a whole. If that were Kant’s belief, then he would be committed to the consequence that no claims of the latter kind are knowable by us, since they will all refer to unknowable noumena. It would amount to the questionable claim that while we can legitimately discover specific truths within the PU related to our experience, we cannot do so when we change the category from such specific items in the PU to the PU itself as a whole. Such an account would be open to the objections that such category shifts are at least not always mistaken and that in any case current scientific cosmology shows that claims about the PU as a whole can be properly made, discussed, and revised within a standard scientific paradigm (see Swinburne 1996; Moore 2011; Bird 1966, 2006, 2011). What is less often noticed is that Kant also further explicitly qualifies our understanding of such an expression as “the PU as a whole” at B510–12. In that passage Kant underlines the difference between empirical knowledge of items in the PU, which can be presented to intuition and the “absolute totality” involved in references to the “unconditioned” in, for example, the Idea of a beginning of the PU. This prompts him to distinguish the empirical, comparative meaning of the term “whole” from a notion of an absolute whole, when it is only the latter which is involved in the present “transcendental problems of reason,” that is in our Idea of an “unconditioned”: In its empirical meaning the term “whole” is always only comparative. The absolute whole of quantity (the universe) . . . with all questions as to whether it is brought about through finite synthesis or through a synthesis requiring infinite extension, have nothing to do with any possible experience. . . . Appearances demand explanation only so far as the conditions of their explanation are given in perception; but all that may be given in this way, when taken together in an absolute whole is not itself a perception. Yet it is just the explanation of this very whole which is demanded in the transcendental problems of reason. (B510–12, emphasis added)

Kant evidently has no reservations about the PU as a whole in its comparative meaning, and this allows him to recognize a scientific explanation of the PU as a whole in that empirical context. It is a further issue whether Kant not only allows such a scientific enquiry but actually accepts it, and later I argue that he does as part of the “regulative” use of the cosmological Ideas at B436 and B544–45. This distinction between a “comparative” and an “absolute” meaning is used elsewhere (in the Amphiboly for example) and throughout the discussion of the Antinomy (e.g., B434–35, B443, B380–85, B 442–43 B445n). The issue in the Antinomy, therefore, is not


graham bird

merely the transition from empirical items in the PU to a new categorylevel in an empirical account of the PU as a whole. It is instead a transition from either of those empirical accounts to the “absolute totality” of the PU as a whole in the notion of the unconditioned. (ii) Even if that further qualification about “the PU as a whole” is noted it can still be asked how that transition is supposed to be made, for example in the arguments of thesis and antithesis. Kant provides his answer in §7 “Critical Solution of the Cosmological Conflict of Reason with itself” (B435–36). There Kant recalls the principle of reason from B364–65, namely “If the conditioned is given then the entire series of all its conditions is likewise given.” This is regarded as the most general form of inference in the background of all the dialectical inferences. It relates to, but is a more general expression of, the specific arguments in the first Antinomy. Kant’s principal point in his initial resolution is to draw a distinction between the principle’s application to noumena and to phenomena. In the former case, the claim is, the principle holds while in the latter case it does not. In that latter case, to anticipate the later constitutive/regulative distinction, the most that can be said is that if a condition is given then the entire series is “set as a task,” which is to say that once we investigate conditions we are led to enquire further back without limit into their antecedent conditions. Kant makes it plain that this difference (between the application to noumena and to phenomena) is associated with his other distinctions, respectively between an “absolute” and a “comparative” totality, and between what can, and what cannot, be derived from concepts alone without reference to possible experience. Kant’s consistent attitude to all the dialectical inferences is that they rest on the ambiguity between noumena and phenomena (B528). So the major premise (in the principle of reason) uses its terms in the “transcendental sense of a pure concept,” while the minor premise uses its terms in the “empirical sense of concepts of understanding applied to appearances.” In a similar way the major premise makes no reference to time while the minor makes that reference essential. Without reference to time, the suggestion is, the step from condition to unconditioned, to the totality of the entire series, is governed by reason, that is by logic; we are led to think that we can deduce the character of the unconditioned from the conditioned members of the series (“if the conditioned is given then the entire series of all its conditions is likewise given”). Such a step reflects an over-ambitious rationalist attempt to establish truths about reality from pure reason, concepts, and

The Antinomies


logic, alone. It is in this way that Kant envisages the slide in the given arguments from an interest in the PU, exemplified in our experience, to a reference to an unconditioned totality of the PU as an unknowable noumenon, beyond any possible experience and amenable only to logical, conceptual argument. (iii) Once that step is made clear then Kant’s central criticism of the thesis and antithesis arguments can be directly related to Kant’s earlier criticisms of potential misuses of both general and transcendental logic (B86–89). At the heart of the conflicts is an appeal to reason, to reasoning and to pure concepts alone, which involves potentially a standard misuse of logic, namely one which infers material truth from formal consistency, “from mere concepts” (B444); or beliefs about the experienced PU from pure reason and reasoning alone. If we ask why or how the thesis and antithesis arguments commit this fallacy of ambiguity, the answers are: first that the fallacy naturally moves toward an ambitious desire for completeness and totality (B445–46); second that it does so by starting with the experienced PU but then offers claims about its absolute totality which rest only on logic and reason. That step compels the concept of such a totality to signify a noumenon inaccessible to possible experience. If such an account is right, then we have an immediate answer to the question why these issues are somehow endemic to reason. They belong to reason just in so far as logic and its forms belong to reason. Jonathan Bennett (1974: 3) says that the Dialectic problems “have nothing to do with reason,” but these extensive appeals to logic and its misuse show that that view is mistaken. Earlier it was noted that Kant’s distinction between comparative and absolute uses of “the PU as a whole” allowed him to recognize a legitimate empirical, scientific cosmology. Now his resolution of the first Antinomy, in which Ideas of reason have only a regulative but no constitutive role, commits him to such a scientific cosmology (Section 9, B543ff. “The Empirical Use of the Regulative Principle of Reason”). This is no more than we should expect, since in other areas, such as psychology, Kant similarly recognizes both a legitimate empirical psychology and an illegitimate pseudo-rational psychology (B346). Empirical cosmology is able to pursue the temporal/causal and phenomenal antecedents of the comparative PU as a whole, which is “set as a task” in the regulative use of the principle of reason. But neither a scientific cosmology nor philosophy can expect to identify any absolute terminus in such an enquiry. Kant is careful to keep


graham bird

the two options open of either continuing seamlessly to extend the enquiry among phenomena, or of finding some current discontinuity in the series, a comparative terminus which may bring the enquiry temporarily to a halt (B550). In either case such outcomes are in principle revisable, and that is the cash value of the regulative injunction.


The Third Antinomy: Freedom and Determinism

The structure of the third Antinomy follows the pattern of the first Antinomy outlined above. I outline the two proofs, comment on them, and then identify the central elements in Kant’s discussion, often unnoticed or disregarded, which I use to represent Kant’s position. 12.3.1

The Two Proofs

Assumptions (1) and (2) as in First Antinomy proofs, with “the natural world” replacing “the PU.” Thesis

Natural causality requires another kind of causality, namely, freedom.


There is no freedom: everything in the world is governed only by laws of nature.



(3) Assume there are only natural causes.

(3’) Assume freedom in a transcendental sense that is, an absolute beginning (4’) Such a beginning (by def.) can have no antecedent. But it presupposes both a prior state and no causal connection to an antecedent. (5’) So, as opposed to the law of causality, it renders all unity of experience impossible – and is an empty thought-entity.

(4) So every event presupposes a prior cause

(5) So there can never be a first beginning and so no completeness in the causal series (6) So (3), taken in its unlimited universality, is inconsistent.

The thesis proof seems clearer in that it turns on a charge of inconsistency in the opposition, while the antithesis proof rests instead on a conflict

The Antinomies


between the opposition view and a “unity of experience,” that is a claim that experience and its unity must rely on universal natural causality and cannot permit any exceptions. The antithesis, however, might be understood in terms of direct inconsistency. For the proof turns on the two claims that an “absolute beginning” both requires, and lacks, some antecedent causality. Both proofs, as in the first Antinomy, rely on notions of universality, totality, or completeness. The thesis proof might be queried on the grounds that the determinist is not strictly committed to a “completeness” in the causal series which would require a “first beginning.” But determinism can turn essentially on the necessity of natural causal laws and a determinate, necessary, outcome from some initial starting point. It is arguable that without such an assumption the determinist cannot insist on the strict necessity or predictability of all events in the world (B577–78). But it is noteworthy that the appeal in both proofs to causal completeness reflects both the general reference to an unconditioned completeness, and Kant’s claim that the arguments threaten not only freedom and morality but causality and science as well “if we yield to the illusion of transcendental realism” (B571). That illusion is effectively a failure to draw Kant’s distinction between phenomena and noumena. In the thesis proof that threat is shown in the requirement of the determinist for nothing other than natural causality and a resulting incoherent completeness. In the antithesis proof the indeterminist side is again charged with either direct inconsistency or conflict with the requirement of a unified experience. Both proofs throw doubt on the coherence of the concept of a “complete explanation” whether it involves causality or an indeterminist freedom. Whether the “first beginning” is purely causal, or just “in time” (B479), or demonstrates a noncausal spontaneity it is open to attack, in the proofs on conceptual grounds, as incoherent. 12.3.2

Kant’s Resolution of the Antinomy

I outline an account of Kant’s resolution by summarily rejecting a standard, traditionalist interpretation and by identifying key elements in that resolution which are often either not noticed or disregarded. In the conclusion I pull these threads together to formulate a defensible position on Kant’s behalf. A traditionalist account emphasizes the presence in Kant’s resolution both of an indeterminist conception of freedom, predicated of noumenal agents, and an extensive picture of what such noumenal (transcendental) freedom is like (B566–86). The account simply ascribes freedom to the


graham bird

noumenal realm and then claims that this is compatible with causality in the realm of phenomena. Both sides in the dispute can apparently be satisfied in such an acceptance of both natural causality and human free agency. The account provokes immediately two decisive objections: first that it is inconsistent with Kant’s belief that we have no knowledge of noumena, and second that compatibility cannot be secured merely by assigning the conflicting concepts to different realms, one knowable the other unknowable. That account, however, is too naive to be accepted as Kant’s position, and in particular fails to note the following complexities which conflict with it. In his summary of what the extensive picture of transcendental freedom has established he insists on the following provisos (B585–86). He now says that he has not established (has not even intended to establish) “the reality of freedom as one of the faculties which contain the cause of the appearances of our sensible world.” Clearly this disclaimer applies to the special case of the reality of “absolute” (transcendental) freedom, and to its supposed activities beyond the sensible world of phenomena in initiating effects in that latter world; so understood it neutralizes the standard objections noted above. Kant goes on to say that he has not intended to prove even the possibility of such freedom. Even that limited aim is said to be beyond what is provable, for “we cannot from mere concepts know the possibility of any real ground and its causality” (B586). Kant is evidently well aware of the objections to the traditionalist account and tries hard to distance his own view from it. Finally what he says we have been able and concerned to establish is that “the antinomy rests on sheer illusion, and that causality through freedom is at least not incompatible with nature.” Kant has repudiated the traditional view, but it may still be thought that his positive goal remains unclear. Accordingly I offer an alternative account of Kant’s resolution which highlights the following points: (1) a determinist commitment to universal causation with a reference to noumena; (2) Kant’s objections to determinism in his construction of a fantasy about noumenal agents; (3) the outcome in a purely defensive position against determinism; and (4) further connections between that outcome and our experience of morality. (1) It was shown earlier that the thesis and antithesis arguments represent determinism as a threat to morality because of its commitment to a strictly necessary universal causality. In both arguments it is assumed that determinism leaves no room for the ascription of a freedom which conflicts with that universal causality, and that assumption is further supported in

The Antinomies


Kant’s Observations (B477–78). This makes clear what kind of determinism is the target of Kant’s own resolution of the issue. But it also explains why Kant thinks that the issue is a threat not only to freedom and morality but also to causal explanation as well, when he claims that in “transcendental realism neither nature nor freedom would remain” (B571). The transcendental realist view recognizes no Kantian distinction between phenomena and noumena; its universal causality covers everything, including noumena. Kant thinks such a commitment is vulnerable at this stage not because it disregards transcendental idealism (which would beg the question) but because it relies on the idea of an absolute beginning in the causal series from which all events can be deduced by means of causal laws. That conception of an absolute beginning was regarded at various points in the arguments of the first Antinomy as incoherent, and the given determinism is vulnerable to the same objections. Although determinism is identified in terms of the predictability of every event from a starting point via natural causal laws its concepts of a beginning and of a totality are recognized as vulnerable in the earlier arguments. (2) In his discussion, however, Kant gives a more positive reason for rejecting such determinism by constructing a fantasy about intelligible agents with noumenal (transcendental) freedom. Kant recognizes the character of his construction when he says: These requirements [of universal causality among natural phenomena] are not in any way infringed if we assume, even though the assumption should be a mere fiction, that some among natural causes have a faculty which is intelligible only and never rests on empirical conditions but solely on grounds of understanding. (B573, emphasis added)

Kant’s fiction does not concern merely a description of such an intelligible agency among noumena, but also a description of the way in which that appeal to noumena can be connected to the natural causes of the associated phenomenon in the empirical agent. It was this connection, and its description of the compatibility of freedom and causality in the same individual which Kant recognized as the central point of his resolution. No other path remains than to ascribe the existence of a thing . . . and its causality in accordance with the law of natural necessity only to appearance and to ascribe freedom to the same being as a thing in itself. This is certainly unavoidable if one wants to maintain both of these mutually repellent concepts together; but in application, when one wants to explain them as united in the same action, and so to explain this union itself, great difficulties come forward which seem to make such a unification unfeasible. (CPrR 5.95)


graham bird

These passages show both that Kant recognized the primary issue of the unification and compatibility of cause and freedom in the same being, and also that the resolution of the issue turned on the description of intelligible agency, even if it is only a fiction. Whatever role such a fiction can legitimately play Kant evidently did not regard himself as committed to the constitutive truth of claims about a noumenal world or, inconsistently, to any knowledge of such a world. All that is required in his argument is the consistency of noumenal (transcendental) freedom, and its compatibility with phenomenal, natural, causality. (3) That conclusion is further confirmed when Kant represents his position as purely “defensive,” that is not as asserting supposed truths about noumena but only as rejecting the determinist’s right to deny such claims. In the Grundlegung Kant expresses that position: No insight can be had into the possibility of [transcendental] freedom as an efficient cause, [but only] a sufficient assurance that there can be no proof of its impossibility. Now where determination by laws of nature ceases there all explanation ceases as well, and nothing is left to us but defence, that is to repel the objections of those who pretend to have seen deeper into the essence of things and therefore boldly declare that [such] freedom is impossible. (G 4.459)

Later in that passage Kant asks the determinist to admit that even though the notion of “appearance” (phenomenon) presupposes that of noumena we have no ground to make assertions about their character, for example that they operate under the same causal laws as phenomena. We can get up to the notion of a noumenon in this context but not beyond (B585). The same lessons are underlined at B804 where Kant talks of a correct polemical, as opposed to an unjustifiable dogmatic use of reason. For Kant speculative issues, such as that of (transcendental) freedom (B831), can be approached only in that defensive way and not by dogmatic assertion. But, although in dealing with the merely speculative questions of pure reason hypotheses are not available for the purpose of basing propositions on them, they are entirely permissible for the purpose of defending propositions; that is, they cannot be employed dogmatically but only in a polemical fashion. By the defence of propositions I do not mean the addition of fresh grounds for their assertion but merely the nullifying of sophistical arguments by which our opponent professes to invalidate this assertion. Now all synthetic propositions of pure reason have this peculiarity that while in asserting the reality of this or that Idea we can never have knowledge sufficient to give certainty to our proposition our opponent is just as little able to assert the opposite. (B804, emphasis added)

The Antinomies


Kant continues in that passage to note how these provisos apply especially to the practical sphere. In that sphere he says “reason does not require to offer proof . . . and could not supply [it]. The burden of proof rests upon the opponent” (B805). (4) That identification of the determinist position, and Kant’s “defensive” response to it, demonstrate the structure of his resolution of the third Antinomy, but it leaves some residual questions. Kant’s end position is evidently compatibilist. Although the central point of his fantasy is to describe noumenal freedom at the transcendental level it nevertheless includes a Humean compatibilism at the empirical level. We might ask why Kant did not just accept that Humean empirical compatibilism instead of supplementing it with his transcendental fantasy. The answer evidently lies in Kant’s identification of the targeted determinism as committed through mere concepts to the impossibility of transcendental freedom (see Bird 2006: Chapter 26, 700–704). Beyond that query are other questions about the connection between the noumenal fantasy and the empirical account of moral acts and properties. Some of these have been answered already; the general appeal to noumena rests on the logical presupposition between phenomena and noumena. That forms an essential part of what I had elsewhere called a “conceptual shadow” (Bird 2006: 551, 758, 763), that is a necessary but illusory logical reflection from our concepts which is both inevitable and potentially misleading. The concept of a phenomenal experience implies the necessary possibility of items, noumena, which lie beyond our experience, so that we can draw a boundary between what is for us knowable and what is not. But in drawing that boundary, “up to but not beyond” that limit (B585), we are enabled, wrongly, to move from a formal to a material mode, and as a consequence to make illegitimate claims about noumena (B88–89). That same point is exemplified in Kant’s account of the relation between practical (empirical) freedom (which “can be proved through experience,” B830) and transcendental freedom. The correct way to express that relation is to say, on Kant’s behalf, that if transcendental freedom were (logically) impossible, then we could make no use of practical freedom (B562); and the central point of his resolution is precisely to reject that antecedent. That rejection does not allow us to infer validly that practical freedom is possible, but there is no need to do that, since it can be “proved through experience” (B 831). But it enables Kant to set aside the primary obstacle to practical freedom from the determinist’s argument. Kant’s appeal throughout the argument (and the Critique) to noumena might be taken, as it is in the traditional account, as a strong commitment


g r a h a m bi rd

to the existence of noumena through moral properties, but when Kant comes to look more closely at the reference to moral properties he takes a different view: The Idea of a moral world has, therefore, objective reality, not as referring to an object of an intellectual intuition (for we are quite unable to think any such object), but as referring to the sensible world viewed as being an object of pure reason in its practical employment. (B836) This law is to furnish the sensible world, as a sensible nature (in what concerns rational beings) with the form of a world of understanding, that is of a supersensible nature, though without infringing on the mechanism of the former. (CPrR 5.43)

These remarks provide a striking contrast to William James’s “crass supernaturalism” in religion, committed to the separate existence of a sensible and a moral world: The world interpreted religiously is not the material world over again with an altered expression . . . it must have a natural constitution different from that which a materialist world would have. (James 1902: 518)

Kant’s account of nonnatural moral properties, unlike James’s crass supernaturalism, allows him to identify and characterize such properties in the world of sense, of phenomena. A further, more specific and central, link between the two, phenomenal and noumenal, aspects to morality arises from the nonnatural appeal in the fantasy to agents’ reasons for acting, for that appeal to reason and logic exemplifies, as Kant says, an order quite different from the causal and temporal order of phenomena. Man, however, who knows all the rest of nature solely through the senses, knows himself also through pure apperception; and this indeed in acts and inner determinations which he cannot regard as impressions of the senses. . . . Reason does not here follow the order of things as they present themselves in appearance, but frames for itself with perfect spontaneity an order of its own according to Ideas, to which it adapts the empirical conditions, and according to which it declares actions to be necessary, even though they have never taken and perhaps never will take place. (B574–76)

Kant refers in this passage to an order of reasoning associated with the noumenal fantasy, but that same order applies to moral reasoning in empirical actions. Kant holds that such an order is compatible with causal explanation, and with reason’s being causally effective in action. Kant is not required to assert the knowable existence of noumena, or to regard

The Antinomies


noumena as actual ingredients in empirical agency. All that is required is that the noumenal fantasy is coherent and conceivable so that the threat from determinism, as Kant understands it, can be turned aside in his defensive argument. Nor does the reference to pure apperception commit Kant in those ways, for in the reference to B152–53, B157–59 Kant allows that although we have a concept, a consciousness, of our existence as pure appercception, it must be related to inner sense if it counts as genuine knowledge (Bird 2006: 301–3, 378–86). Nor is he required to regard empirical moral judgments as made without appeal to dates and times. That is true of the noumenal fantasy which conceptually isolates the new order of moral reasoning from temporal and causal experience. But in the empirical application of moral principles, and imputation of moral praise and blame, both times and dates are indispensable (B582–83).



Kant’s resolution of the Antinomies guides both scientific enquiry and moral judgment in the regulative use of the relevant Ideas. In cosmology it encourages endless empirical exploration of the PU as a whole, but recognizes comparative termini arising, for example, from limits to theory or observation. But it disallows constitutive claims about an “absolute” terminus, argued for on purely logical or conceptual grounds. The third Antinomy reinforces that scientific guidance by recognizing the same explanations from natural causes even for agents’ behavior. But if Kant’s argument is correct that causal explanation of behavior is compatible with the ascription of nonnatural moral values. So long as his defensive rebuttal of determinism holds that ascription of moral properties can continue legitimately and “leave aside the speculative issues of transcendental freedom” (B831). On the negative side Kant recognizes two factors in reason which together encourage systematic error. First is the inevitable motive to understand its intellectual systems, in science or morality, in their totality or completeness. Second is the equally natural, but misguided, belief that reason’s own resources, from logic and concepts, can achieve that understanding. The primary mistake arises from that belief, in which those logical or conceptual resources are misused and lead to opposed untestable, transcendent claims and to the futile disputes of the Antinomies. The errors anticipated in the Analytic at B88–89 are exemplified in the details of those disputes. Logic and its misuses are implicated at every stage in the debates. It is the basis for the construction of the concept of transcendence from that of immanence, of the concept of a noumenon from that of a


g r a h a m bi rd

phenomenon, of the (transcendent) concept of a “thing that appears” from the concept of an “appearance” (Bxxvi–vii), or of the unconditioned from the concept of a conditioned. It is the form in which in the principle of reason we may think that an unconditioned can be deduced from what is conditioned. In that case a legitimate formal claim that the concept of what is conditioned implies an understanding of what is unconditioned is mistakenly interpreted as the material claim that from a conditioned we can logically infer the character and existence of the unconditioned. Those logical, conceptual, arguments are appropriate for the realm of noumena for which we have no other investigative resources, but they are not appropriate for areas which require sensory, empirical enquiry. Kant’s central transcendental idealist distinction between noumena and phenomena captures that crucial contrast and is the basis for his resolution of the issues.

ch a p ter 1 3

The Ideal of Reason John J. Callanan

13.1 Critique and Transcendental Theology In his submission for the 1763 Berlin Academy Prize Essay competition Moses Mendelssohn offered a defense of rationalism. Descartes had already secured two incontrovertible proofs with it, he claimed, those of the cogito and the ontological proof of God’s existence (Mendelssohn 1997: 275–76). The actuality of one’s own self and the divine being could be proven from the mere analysis of the concepts of the self and God respectively. The runner-up in the competition was Kant, who had offered a more negative picture: not only had no metaphysical results yet been secured, Kant claimed that “no metaphysics has yet been written” (Inquiry 2:283, TP1 255). When Mendelssohn returned to the topic of God’s existence in the Morning Hours in 1785, the situation had changed, and he had to acknowledge the enormous influence that the “all-quashing Kant” had wielded in the intervening years, especially with regard to the ontological proof.1 In his view, Kant’s overall influence had not obviously been a positive one (Mendelssohn 2011: xix). As he saw it, the opposed trends of materialism and religious fanaticism had each continued to grow since the introduction of the Critical philosophy. Mendelssohn’s hope was that there might yet emerge a “Kant who will hopefully build up again with the same spirit with which he has torn down” (Mendelssohn 2011: xx). Kant himself might reasonably have been disappointed with this characterization of his ambitions. His goal had never been to tear down religious faith, but rather only to usher in “the genuine age of criticism, to which everything must submit” (Axin). Kant had explicitly included 1

In the “Ideal of Pure Reason” chapter of the Critique of Pure Reason, Kant claimed, first, that there were only three kinds of proof (ontological, cosmological, and physicotheological) for the existence of God (A590–91/B618–89); second, that they all had a dependency upon the ontological proof (A607/B635, A629/B657); and third, that the ontological proof is irrevocably flawed (A592/B620ff.). For critical discussion of the dependency relation claim see Grier (2010), Smith (2003), and Wood (1978).



j o h n j . c al l an an

religion within the scope of “everything” here, but so as to protect rather than undermine religious belief. In Kant’s opinion, it was the lack of critical self-appraisal among those who defend religion that meant that “they excite a just suspicion against themselves” (ibid.) By deploying the method of criticism he had hoped to undermine only the notion that God’s existence was a proper subject matter for theoretical proof or disproof (see A641/B669). This negative result was itself secured only so as to validate a distinct conceptual space for faith (Bxxx). Moreover, Kant repeatedly argued in the first Critique and elsewhere not just for the theoretical space within which one could maintain religious belief but also for the practical demand that one must so believe.2 Yet it seems that such was the power of Kant’s negative arguments regarding the impotence of philosophical proof that an image of Kant as explicitly or implicitly hostile to religious belief persisted. Heine, perhaps echoing Mendelssohn, described Kant as having “destructive, world-annihilating thoughts” (Heine 1986: 109). Considered in the context of Heine’s claims, there is something ironic about Kant’s expressed aim of protecting religious believers from suspicion. Ricoeur characterized Marx, Nietzsche and Freud as the three masters of the “school of suspicion” (Ricoeur 1970: 32) – one that undermined our trust in our ordinary conscious experiences and rationalizations as transparent to ourselves – but Heine had himself already identified Kant as one such master. For Heine, Kant had a “talent for suspicion” and that was most apparent when “manifested in the direction of thought and was called criticism” (Heine 1986: 109). While Kant viewed critique as the means for protecting religious belief from suspicion, his method for protecting it involved laying bare the cognitive mechanisms through which we come to form such belief. These mechanisms – ones certainly not always transparent to common consciousness – show that both the belief in the existence of God and our characterization of Him are both “natural” demands of human beings’ rational capacities.3 This approach threatens to raise as much suspicion as it allays. The famous first line of the first edition of the Critique, whereby Kant attributed to reason itself the “peculiar fate” (Avii) of being burdened with ideas that press questions which reason can neither answer nor dismiss, would have seemed to any ordinary reader a strikingly negative opening statement. 2 3

E.g., A809–11/B837–39, CPrR 5:124–32, CJ 5:484–85. For discussion, see Beiser (2006), Gardner (2011), and Wood (1992). For Kant’s characterization of reason’s demands as “natural” to it, see Axvii, Bxxxi, B21–22, A298/B354, A314/B371, A407/B433–34, A421–22/B449–50, A581/B609, A583/B611n, A642/B670, A669/B697, A813/B841.

The Ideal of Reason


Moreover, Kant is clear that the idea of God is one such notion. Kant’s manner of defending the naturalness of the idea of God flirted with an account whereby that idea was presented as a by-product of reason’s more ordinary truth-generating operations. For some Kant’s talent for suspicion with regard to the proof of the existence of God threatened to tip over into a suspicion regarding the very idea of God itself. It is perhaps unsurprising that such an image of Kant should have emerged. The anthropocentric re-orientation of the Copernican turn must have appeared as recommending replacing God with human beings as the source of the laws of nature.4 It is Kant who tells us, for example, that “human reason has a natural propensity to overstep all these boundaries [of possible experience]” and that its “ideas effect a mere, but irresistible, illusion, deception by which one can hardly resist even through the most acute criticism” (A642/B670). Similarly, it is Kant who claims that in order to protect one’s position from attack, one “must always seek the enemy here in ourselves. For speculative reason in its transcendental use is dialectical in itself” (A777/B805). Yet it is surely just as obvious that Kant also had a thoroughly positive outlook with regard to the faculty of reason and its ideas. While acknowledging that the ideas of reason can generate illusions, Kant does not lay the blame with the ideas, which he claims “can never be dialectical in themselves,” but rather with our use of them. The ideas themselves “have their good and purposive vocation in regard to the natural predisposition of our reason” (A699/B697). That a representation that emerges naturally from reason – which is the source of our ability to discriminate truth from falsehood (A699/B697) – might itself be inherently illusory, is a possibility that Kant cannot countenance. As such, Kant is committed to the view that if a rational representation is a natural one, then it must have some positive epistemic function within human experience. Contrary to Heine’s image, Kant’s attitude toward reason’s ideas appears strikingly optimistic, perhaps even Pollyannaish. My focus in this chapter will not be upon the familiar negative onslaught upon the proofs for the existence of God;5 instead I focus upon the positive theoretical picture that remains despite that onslaught, of Kant’s account of the origin of the concept of God. As mentioned Kant holds that there is a crucial practical interest in and warrant for our use of the concept of God. My focus here though will be upon some of the elements of Kant’s account of the concept’s theoretical role that aroused suspicion. Kant complains that 4 5

E.g., see A127. For a mitigating impression, see Watkins (2013). For a small selection of the relevant literature, see Allison (1983 [2004]), Bennett (1974), Grier (2001, 2010), Plantinga (1966), Proops (2014, 2015), Shaffer (1962), Van Cleve (1999), and Wood (1978).


j o h n j. c al l an an

in the past no one has “taken trouble . . . to understand whether and how one could so much as think of a thing of this kind as rather to prove its existence” (A592/B620). This is the subject matter of “transcendental theology.” As Kant characterizes its subject matter, “the thing that contains the supreme condition of the possibility of everything that can be thought (the being of all beings) is the object of theology.”6 While knowledge of God’s existence is denied, Kant claims that pure reason itself provides the idea that allows for a “transcendental cognition of God (theologia transcendentalis)”(A334/B391). Kant’s positive account is found in the Third Chapter of the Second Book of the Transcendental Dialectic, entitled “The ideal of pure reason,” but also in the second half of the Appendix to the Transcendental Dialectic, entitled “On the final aim of the natural dialectic of human reason.” The primary goal of the former section is to reveal a special kind of “sophistical inference,” such that “from the totality of conditions for thinking objects in general insofar as they can be given to me I infer the absolute synthetic unity of all conditions for the possibility of things in general” (A340/B398). This fallacy – that of inferring the existence of things from reflections on the conditions of thinking about them – in this instance produces a claim to cognize “a being of all beings.”7 In these sections Kant also lays out how the positive conception of God is warranted just because it is a natural demand of human rationality. The account engages with the denunciation of anthropomorphism in religious matters criticized by Montaigne, Spinoza, Hume, and others. Kant was aware in particular of Hume’s claim in the Dialogues Concerning Natural Religion regarding the attribution of the property of intelligence to the explanatory ground of the universe. Hume’s claim was that such an attribution is merely an instance of the unwarranted anthropomorphic projection of rationality onto that thing, performed for the sole function merely of serving human interests. Kant’s response is striking just because it does not deny many of the essentials of Hume’s account. His position accepts there is such an anthropomorphic projection. Second, he acknowledges that such projections do literally mischaracterize the “being of all beings.” Third, he is explicit that the projections are made for the purposes of serving human interests. Nevertheless he claims that our characterizations of God are warranted. 6


It is worth noting that Kant’s metaphysics textbook, Baumgarten’s Metaphysics, devotes Chapter I of Part III (“Natural Theology”) to the concept of God, but does not address the question of the origin of the concept. In a letter to Schütz of 1785, Kant refers to Mendelssohn’s Morning Hours as a “masterpiece of the self-deception of our reason” in its failure to attend to the fallacy (10:427–28, Corr. 237).

The Ideal of Reason


The chapter will be structured as follows. In the following section I briefly consider one aspect of the historical context with which Kant was concerned, namely, that of the conditions under which the concept of God is originally formed. In Section 13.3 I set out Kant’s own account of the origin of that idea as the transcendental ideal of reason.8 In Section 13.4 I outline the analogical reasoning that is deployed in defense of the characterization of God. I conclude in Section 13.5 with a consideration of Kant’s use of the notion of an archetype of rationality. Kant in essence upholds the Leibnizian view of human rationality as partaking in divine rationality but has to justify this claim within the constraints of the Critical system. The peculiar position developed is that it is one of the “interests” of human rationality that it projects its characteristics onto the idea of a divine being, yet only for the purpose of subsequently viewing human reason as a copy of that original divine reason. It is such aspects of Kant’s thinking that plausibly aroused some of the suspicion in its reception.


The Origin of the Idea of God

In the Discourse Descartes complained that only the unwarranted presuppositions of empiricists could lead one to think that there is “some difficulty in knowing God.” In general, their error is that “they never raise their minds above things which can be perceived by the senses” (AT vi. 37, CSM i. 129).9 While it might be the case that sensory representations must be deployed to aid the expression of the concept of God, it does not follow that the content of that concept reduces to such representations.10 One of Descartes’s targets here is surely Montaigne, who had maintained a far more skeptical attitude toward the possibility of knowledge of the divine. For Montaigne our conceptualization of God is inherently problematic, since the predicates we might apply are only comprehensible to us from our familiarity with them in the context of their imperfect actualization in the terrestrial realm. This imperfect context of content determination then precludes their apt predication of a perfect being: 8



My analysis here focuses on the initial picture offered in the first Critique alone, considered in relative isolation from the subsequent development of Kant’s thought in the Critique of Practical Reason, the Critique of the Power of Judgment and Religion within the Boundaries of Mere Reason. References to Descartes are to the Cambridge edition of his writings cited in the Bibliography (‘CSM,’ volume and page number), including the standard marginal ‘AT’ page references to the volumes of the Charles Adam and Paul Tannery edition, Oeuvres de Descartes (Paris: 1974–86). See also Descartes AT vi. 37 CSM i. 129 and AT vii. 305, CSM ii. 212. Arnauld and Nicole follow Descartes’s line of argument (targeting Gassendi) in the Port-Royal Logic (Arnauld and Nicole 1996: 29).


j o h n j. c al l an an Nothing of ours can be compared or associated with the Nature of God, in any way whatsoever, without smudging it and staining it with a degree of imperfection. How can infinite Beauty, Power and Goodness ever suffer any juxtaposition or comparison with a thing as abject as we are, without experiencing extreme harm and derogating from divine Greatness? (Montaigne 2003: 585)

The inappropriate ambition to cognize the divine is itself motivated by the “natural, original distemper of Man,” namely, the “presumption” that human beings might share something in common with the divine (Montaigne 2003: 505). For Montaigne, the only available means of representation of God available to us is that of analogical representation, but the anthropomorphic projections it invariably produces reveals more about human hubris than anything about a divine being. For example, he describes virtue as that which is particular to human beings, on the grounds of it being a status achieved in the face of the temptations and frailties of our nature. It follows, he claims, that a perfect being, lacking temptations and imperfections as it does, cannot achieve that particular status: Take Prudence; that consists in a choice between good and evil; how can that apply to God? No evil can touch him. Or take Reason and Intelligence, by which we seek to attain clarity amidst obscurity; there is nothing obscure to God. Or Justice, which distributes to each his due and which was begotten for the good of society and communities of men; how can that exist in God? (Montaigne 2003: 556)

In the Port-Royal Logic, Arnauld and Nicole attack similar approaches to the issue of God’s moral attributes. They quote Cicero’s reporting of Cotta’s argument in De Natura Deorum: [S]hall we then assign to God that prudence which distinguishes things good, things evil, and things neither good nor evil? But if a being does not and cannot partake of evil, what need has he to make choice between good things and evil things, and what need has he or reason and understanding? We apply these faculties to advance from what is revealed to what is hidden, but nothing can be hidden from God. (Arnauld and Nicole 1996: Part III, Chapter 19, 200)

Arnauld and Nicole are outraged by this kind of “impertinent” argument, glossing it as the claim that since God could have no virtues similar to those found in humans, that therefore God must lack virtue altogether (Arnauld and Nicole 1996: 201). But they are unfair to the argument here: the claim by Cotta and Montaigne is not that we cannot conceive of a being having a perfect instantiation of that which is imperfectly instantiated in human beings; rather, it questions how an attribute whose core content

The Ideal of Reason


is originally defined relative to a context of instantiation in finite beings could be commensurably manifested in an infinite and perfect being at all.

13.3 The Notion of an Ideal The question of the origin of the concept of God is then crucial for any account of how true – and perhaps even truth-evaluable – judgments about God can be made. To attribute the origin of the content to idiosyncratic acts of an individual subject’s cognition would be to denigrate the content of that concept. While Descartes’s solution involves positing an innate status for the concept, Kant’s response is that the only way to secure the nonarbitrariness of the concept’s content is to assign it an origin in the representational capacities of the subject, not as an innate idea but as a natural product of our rational cognitive functions.11 Kant’s approach involves finding an appropriate middle status for the origin of the idea of God. The idea cannot be generated in the same way that the categories are formed, i.e., in the particular context of the possible experience of spatiotemporal particulars. Moreover, the idea is not of a general rule, but is rather that of a particular thing, which Kant calls the “original image” or archetype [Urbild] (A569/B597). It also cannot be that the idea emerges as a merely arbitrarily formed concept, a hodgepodge of conceptual contents that reflect more the particular expressive capacities of the individual agent than the object itself. Kant warns against trying to express the notion of the archetype through any kind of example: the example will always sully the notion and repeated failed attempts might “render even what is good in the idea suspect by making it similar to a mere fiction.” Kant’s worry that the idea must be distinguishable from mere “creatures of the imagination” is similarly reminiscent of Cartesian complaints (A570/B598). In the Lectures on Philosophical Doctrine of Religion Kant expresses the more general worry that “we will have to take the materials for the concept of God from empirical principles and empirical knowledge.” This risks that one might be picking “bad predicates” and thereby engaging in an act of self-deception whereby we “let ourselves be blinded by a mere show and ascribe predicates to God which can only be true of objects of sense” (LRel 28:1021–23, RRT 365–66). Kant’s key claim is that reason itself, considered as a transcendental faculty, is the source of the generation of content (A299/B355). More specifically, as Wood puts it, Kant argues that “the idea of God takes its rational origin from the fact that it is presupposed by any attempt to think of 11

For discussion of Kant’s relation to nativism, see Callanan (2013).


j o h n j . c al l an an

individual things in general as thoroughly determined, and hence as absolutely possible” (Wood 1978: 62). It is in this way that Kant thinks he has shown that the ideal is “grounded on a natural and not a merely arbitrary idea” (A581/B609). Just as the forms of judgment provide a purely logical analogue to the categories, so too can the purely logical operations of reason give a guide to the conceptual contents that reason generates out of itself. It is important that it is reason and not understanding that generates this content. Transcendental idealism has it that it is only through the nondiscursive representational capacity of intuition that objects can be given (see A92–93/B125, B146, B148, A139/B178). The categories considered on their own represent no particular object determinately, but are merely the conditions for the representation of an object in general (A51/B75, A93/B125–26, A290/B346). Kant relates the categories to the possibility of experience of objects and claims that they generate cognition of an object only upon receiving a spatiotemporal schematization (A139/B178, A145– 46/B185). The unschematized categories are insufficient then to represent a particular thing, while the schematized categories would render a representation of God infected with that sensible content and would risk “falling into that anthropomorphism which transfers predicates from the sensible world onto a being wholly distinct from the world” (Prol. 4:358). Kant seeks to locate the content as removed from our possible experience of objects though not so removed as to render it an arbitrary or dispensable product of rationality. He introduces the notion of an ideal in Section 1 of Chapter 3 by appealing to the idea of something being represented in concreto.12 Something is represented in concreto when it is represented by appeal to a particular instance of the thing (see A713/B741; also A241–42n, A711/B739). The categories can represent something in concreto, Kant claims, when they have sensible appearances as their “proper material.” This is to transform them into “concepts of experience.” They do not contain within themselves the conditions of their own application to things and to that extent they represent “no objects at all.” However, ideas of reason are “still more remote from objective reality” since these concepts cannot be even possibly satisfied by the presentation of a particular object, as “no appearances can be found in which they may be represented in concreto” (A567/B595).13 12 13

For an overview of the themes of Kant’s Chapter 3 on “the Ideal,” see Buroker (2006), Gardner (1999), Grier (2010), O’Shea (2014), and Wood (1978). As we shall see, they do however relate indirectly to such objects in relation to the infinite task of securing systematic empirical unity (A568/B596).

The Ideal of Reason


These aspects of Kant’s model of cognition risk presenting the ideal as an “empty” concept, i.e., as one lacking any possible object as a referent (A51/B75, A62/B87, A77/B103, A155–56/B194–95). However, Kant introduces the notion of the ideal which is “the idea not merely in concreto but in individuo, i.e., as an individual thing which is determinable, or even determined, through the idea alone” (A568/B596). Although sensibility allows particular things to be given, and although reason is the capacity farthest removed from sensibility, the ideal nevertheless is the notion of a particular thing. More specifically, the ideal is the notion of a particular thing that is somehow given in the idea itself. This is the connotation of something being an idea in individuo, whereby the metaphysical gap between a representation and its object is occluded, and the idea itself somehow provides its own exemplification.14 Kant’s claim is that the particular content is generated by the structure of human rationality itself, in this case by certain natural structures regarding our capacity for syllogistic reasoning.15 The concept of God as the ens realissimum is generated by the “principle of thoroughgoing determination” (A571/B599), which states that “among all possible predicates of things, insofar as they are compared with their opposites, one must apply to it” (A572/B600). The principle generates the thought of “every thing as deriving its own possibility from the share it has in that whole of possibility” (A572/B600). From this notion we secure the idea of a “sum total of all possibility” (A573/B01). Kant describes the transition, involving two distinct steps, whereby one moves from a principle of reason to the notion of an individual that is the “being of all beings.” The first step starts from the claim that the principle of thoroughgoing determination is a principle of the transcendental use of reason (A571–72/B599–600). Kant argues that this principle itself generates a concept, which is the idea of the sum total of all possibility.16 14



Kant uses a characterization of Plato’s theory of the forms to illustrate the notion of an idea that is itself an individual object which serves as the metaphysical original and the “original ground of all its copies in appearances” (A568/B596). The reference is possibly to the Timaeus and the cosmology of the “archetype” in the divine mind (Timaeus 28a). Kant is notoriously sanguine regarding the connection between ideas of reason and forms of inference, claiming that the movement from “the cognition of oneself (of the soul) to cognition of the world and, by means of this, to the original being, is so natural that the progression appears similar to the logical advance of reason from premises to a conclusion” (A337/B394–95). In the case of the concept of God, Kant claims that the idea emerges from the mere form of the disjunctive syllogism where the inferences begin from an instantiation of the law of excluded middle (A576–7/B603–4). The claim is that in order to think of a thing as thoroughly determined, one must presuppose something else, namely, the idea of the sum total of all possible predicates. Each object either bears a positive or negative relation to every possible predicate, depending on whether it instantiates that


j o h n j . c al l an an

The second step claims that one can “refine” the concept of the sum total of all possibility into a different concept, that of an “individual object that is thoroughly determined merely through the idea, and then must be called an ideal of pure reason” (A574/B602). This refinement involves the transformation of the former idea into the notion of an “individual thing.” The key move is that one is not considering here merely logical relations such as affirmation and negation, but rather transcendental predications, i.e., the real presence (or absence) of predicates in individual things (A574/B602). Kant claims that the real presence of predicates in things must presuppose the idea of the whole totality of predicates itself considered as a “Something” [Etwas] (A574/B602). By the end of Section 2 Kant claims to have shown that the nature of reason itself generates the notion of a single original being, and that the concept of God just is that notion of a “being that is singular, single, allsufficient, eternal, etc.” (A580/B608). The account that justifies the production of that content as natural and nonarbitrary involves denying that it is directly related to the conditions under which it might gain sense and significance. As such the idea is warranted, but only if it is understood as offering “nothing other than a regulative principle of reason, to regard all combination in the world as if it arose from an all-sufficient necessary cause” (A619/B647). The “object” of God is given, Kant claims, but “only as an object in the idea.” An “object in the idea” does not serve to present its actual object to the subject, but “serves only to represent other objects to us, in accordance with their systematic unity” (A670/B698). Thus as far as the theoretical interest of reason is concerned, the positive epistemic function of the idea of God is not to represent a distinct divine object. Rather it functions indirectly to aid our representation of empirical objects.17 The way in which this function operates is that we are provided with a warranted regulative assumption for the inference that characterizes the design argument, that the “things in the world must be considered as if they had gotten their existence from a highest intelligence” (A671/B699). The ideas of reason in general function to provide a “schema of the regulative principle for the systematic unity of all cognitions of nature” (A674/B702). That they can perform this function is only possible because of their character


predicate or not. The idea of a thing as thoroughly determined is then the idea of the totality of positive and negative values that thing bears to the entire set of possible predicates. See also Prol., 4:348. That the concept then has a primary role in relation to our knowledge of the natural world is emphasized with Kant’s claim that the concepts of God and nature can be substituted salva veritate in such judgments (A699/B727), a claim perhaps unwisely made by one who took himself to be fundamentally at odds with Spinoza.

The Ideal of Reason


of not being conditioned by sensibility. It is the conditioning of sensibility, however, that allows for concepts to represent a determinate object in the first place (see Prol. 4:355–56). Attempts to characterize further the “Something” that the Ideal represents as a determinate object must inevitably fail, or can succeed only by unwarrantedly attributing features of sensible reality to them. Thus while these ideas must then be related to necessary rules for the representation of empirical things, they themselves “should be grounded only as analogues of real things” (A674/B702).


Reason and Analogy

Kant nevertheless thinks that this notion of the original being can be supplemented in various rationally warranted ways. While the preceding attributes are those that are predicated of the being by a natural feature of the essential functions of reason itself, other attributes may be predicated on weaker epistemic grounds. For example, Kant clearly thinks that we are rationally warranted in characterizing the being as an intelligence (A583/B611n, A640/B668, A670/B698, A697/B725). This might appear to be just the kind of anthropocentric reasoning that Philo warns against in Hume’s Dialogues Concerning Natural Religion, whereby we grant a “peculiar privilege” to the “little agitation of the brain which we call thought” such that “we must thus make it the model of the whole universe” (Hume 2007: 2.19, 24). For Hume, the selection of rational intelligence as the crucial aspect of the divine being requires a selection procedure and one that can overcome the natural anthropomorphic bias (Hume 2007: 2.18, 24). Kant recognizes that Hume’s “dangerous arguments relate wholly to anthropomorphism” (Prol. 4: 356) yet is nevertheless confident that such projection can be warrantedly made. He claims that the “personification” of the ideal of reason is again a “natural progress of reason in the completion of its unity” (A583/B611n). This personification is explicitly that characterization of the ideal as “an intelligence.” So here it seems Kant will take issue specifically with Hume’s claim that the attribution of intelligence as a characteristic of the “being of all beings” is an anthropomorphic projection. Kant does not deny this but rather claims that it is a warranted anthropomorphic projection. In the Prolegomena Kant distinguishes between “dogmatic” and “symbolic” anthropomorphisms (Prol. 4:357).18 The former makes attributions to a thing in itself, i.e., God. The latter makes no such attribution – here 18

See Kant’s Refl 6056, 18:439, in NF 329.


j o h n j . c al l an an

instead “we attribute those properties, nonetheless, to the relation of this being to the world.” Such an attribution “concerns only language and not the object itself” (Prol. 4:357) and should be understood as revealing only the interests of the agent making the analogy. It is unclear how the distinction is supposed to be drawn in a satisfactory way. Some assistance is offered though by Kant’s characterization in §58 of the Prolegomena of symbolic anthropomorphism as a product of analogical reasoning. In particular Kant holds that some analogical characterization of the divine being is warranted just because such claims merely express the equivalence of two relations, rather than making a claim about the character of the relata themselves. This feature allows for one of the relata to be something entirely unknown (or even unknowable), as in the analogical claim that “the promotion of the happiness of the children = a is to the love of parents = b as the welfare of humankind = c is to the unknown in God = x, which we call love” (Prol. 4:358n). The mathematical form of analogical reasoning (a:b::c:x) is one Kant identified earlier in the Critique when introducing the Analogies of Experience (A179/B222).19 Here the same schema is suggested as providing an epistemically weaker rational basis for the characterization of God. Several different challenges to Kant’s use of analogy have been identified however.20 Since any analogical predication will have the form a:b::c:x, in order to make such a predication one must have familiarity at least with the elements a, b, and c, and the relation that exists between a and b. With these elements in hand, one can analogically predicate that same relation as holding between the third known something c and the fourth unknown member x. More than this is required however. Also required is an understanding that a:b is in general the appropriate kind of material from which to form analogies with x in question. Second, it must be known that a and c in particular are similar enough in kind so as to think that they can be plausibly linked for the purposes of analogically predicating the relation from the first pair to the second. Why does Kant think that he can evade the worry that we are unwarrantedly predicating properties of things when he claims that he is merely predicating the relation that holds between things? On Kant’s view this move shows that the intentional object of our predication is not a property of things but in fact merely a property of human beings’ way of representing, specifically a claim about only about “language” (Prol. 4:357). Just as human beings represent the relation between a and b as thus 19 20

For discussion, see Callanan (2008). For some of the following analysis, see Gill (1984), Logan (1998), and Wood (1978: 89–90).

The Ideal of Reason


and so, we might analogically predicate that human beings can represent the relation between c and x as thus and so. Kant holds here that analogy is the proper method for ascribing predicates to God in a way that is epistemically respectable and doesn’t fall prey to the risk that we are engaged in a “mere show.” He claims that in the model of analogy (a:b:: c:x) we are not claiming an “imperfect similarity between two things, but rather a perfect similarity between two relations in wholly dissimilar things” (Prol. 4:357–58).21 Kant suggests that the mathematical model of analogy can allow us to “form a concept of God and of his predicates which will be so sufficient that we never need anything more” (LRel 28:1023, RRT 366). It is hard to see how this specification of the form of inference removes the previous concerns. In this form, the claim is of a similarity between two relations, each of which connects a pair of relata. However the claim that there is any similarity between the relations presupposes that the things connected by each relation are at least minimally alike such as to make sense of the comparison. A proposed similarity between the relations then presupposes at least a minimal affinity between the relevant pairs of relata. But this is just the assumption that Kant denies in criticizing the dogmatic form of analogy. A vindication of our analogical characterization of God would surely show that the identification of a ground of analogy is one shared by humans and the thing targeted for characterization. But this would be to assume what is being inferred for the possibility of making the inference. Kant’s defense here can then seem more like an apology for anthropomorphism than a vindication of it.

13.5 The Archetype of Reason For Kant the power of discursive representation is not among with the capacities of a divine being. In a letter to Schultz regarding the latter’s work on the Critical philosophy, Kant complains that at a certain point “the divine understanding appears to have a sort of thinking ascribed to it” (10:557, Corr. 285). When Kant models an epistemically perfect agent, he does not characterize it as one who has perfect discursive (i.e., conceptwielding) capacities but rather as one who has a perfect non-discursive capacity of intellectual intuition (B145, A252/B309, B307). Human understanding is the capacity to bring particulars under rules, the latter which indicate a general “mark” that can allow us to group a plurality of objects 21

See also LRel 28: 1023, RRT 366–67; and Kant’s lectures on metaphysics, Metaphysik K1 28:1471, LM 558n14.


j o h n j . c al l an an

together. The position is reminiscent of that of Cotta and Montaigne: a divine being, lacking any epistemic imperfection, has no need of any such mediating and grouping representations, since such a being just immediately gives itself the thoroughly determinate object in the same act of representing it. Thinking, considered as a representational kind, is a manifestation of human beings’ cognitive finitude. Despite the gulf between our representational capacities and that of a divine being, Kant nevertheless posits an intimate connection, again at least as a regulative principle. This relationship is that between the archetype and ectype, of the original image and the copy.22 One of the more striking occurrences of the distinction would have been in Hume’s Dialogues, where Philo claims that Cleanthes’s anthropomorphism leads him to an egregious fallacy in formulating arguments from analogy. In experience one attempts to move from ectypes to the archetype, Philo claims, by thinking a discursive representation that has been copied from our experience of individual things. Cleanthes then holds that the archetype is somehow thought itself, but Philo points out that this is just to conflate our mode of representing something with the ontological character of the thing represented: In all instances which we have ever seen, ideas are copied from real objects, and are ectypal, not archetypal, to express myself in learned terms: You reverse this order, and give thought the precedence. (Hume 2007: 62)

For Kant, this order of explanation is on the contrary entirely valid. It is a regulative demand of reason that we view human reason as a copy of divine reason. The notion of archetype and ectype was central to the discussion in the Ideal of Reason. There Kant described how the notion of an archetype was itself the product of human reason: For reason the ideal is the original image (prototypon) of all things, which all together, as defective copies (ectypa), take from it the matter for their possibility, and yet although they approach more or less nearly to it, they always fall infinitely short of reaching it. (A578/B606)

In the Appendix to the Dialectic, Kant argues that the fact that the archetype is demanded by reason recommends that it should be characterized as a perfected form of reason: 22

Apart from Plato, the notion has many potential sources: Kepler’s The Harmony of the World hinges its entire cosmology on the notion of archetypes (Kepler 1997); Malebranche, with whom Kant acknowledged some affinity in the Inaugural Dissertation (ID 2:410, TP1 405), appealed to “intelligible extension” as a divine archetype (Malebranche 1997: 138); Locke makes explicit use of the archetype/ectype distinction (Locke 1975: III.iii.18), which is then picked up in Leibniz’s New Essays (Leibniz 1997: Book II, Chapter xxx, 263, 268).

The Ideal of Reason


The greatest systematic unity, consequently also purposive unity, is the school and even the ground of the possibility of the greatest use of human reason. Hence the idea of it is inseparably bound up with the essence of our reason. The very same idea, therefore, is legislative for us, and thus it is very natural to assume a corresponding legislative reason (intellectus archetypus) from which all systematic unity of nature, as the object of our reason, is to be derived. (A694–95/B722–23)23

Kant moves between the claims that we must view all the objects of nature as the products of a divine reason and that we must view human reason as itself also a product or copy of divine reason. We think of a divine being “according to the analogy of realities in the world,” Kant claims, for the purposes of unity in our empirical investigations, by “seeing all combinations as if they were ordained by a highest reason of which our reason is only a weak copy” (A678/B706). The picture that Kant is arguing for here would have seemed to many as philosophically and theologically orthodox. Human beings after all were thought to have been made in God’s likeness. A central claim of the natural law tradition is that this likeness is expressed in both beings’ rational nature. The imago dei doctrine is moreover a cornerstone of Leibniz’s way of thinking.24 However committed he is to it, Kant cannot assume the imago dei doctrine as a metaphysical fact – for it too is a claim that must be submitted to the “fiery test of critique” (A406/B433) – but he does nonetheless, as we have seen, ground it within the resources of his system. This restriction demands that the imago dei doctrine be itself justified internally as one of the interests of human reason. This results in the peculiar claim that the relation of human to divine reason is itself demanded by the analogical attribution of the categories “of substance, causality and necessity,” concepts that Kant himself acknowledges only get their sense and significance as forms of thought relatable to sensible intuition. The peculiarity is supposed to be allayed however by the merely “relative,” i.e., regulative and analogical, use to which we are putting these concepts (A678/B706). Kant grants that these concepts “lose all meaning” and are “without any content” when actually applied beyond the “field of sense.” On the other hand the 23


Cf. Inaugural Dissertation, ID 2:397, TP1 389. The claim is buttressed by the considerations regarding the regulative employment of reason in the sciences. We must view the empirical world “as if” it had a “single supreme and all-sufficient ground . . . in relation to which we direct every empirical use of our reason it its greatest extension as if the objects themselves had arisen from that original image of all reason” (A672–3/B700–701). For discussion of the regulative employment of reason in science, see Buchdahl (1969) and O’Shea (1997). For a single example, see §28 of the Discourse on Metaphysics (Leibniz 1989). For discussion, see Hillman (2010).


j o h n j . c al l an an

Ideal of Reason requires them since Kant claims that “we cannot think [the determinate concept of God] except in accordance with the analogy of an actual substance that is the cause of all things according to laws of reason” (A675/B703). The categories do not lose their meaning though when put to the purpose of expressing a relation that human reason demands: I think only the relation [Relation] which a being, in itself unknown to me, has to the greatest systematic unity of the world-whole, and this is solely in order to make it into the schema of a regulative principle for the greatest possible empirical use of my reason. (A679/B707)

There is then again a paradoxical feel to Kant’s defense of the original being as an intelligence. Human reason has a need to seek unity, which in turn demands that we think of the sum total of all appearances as a unity grounded in a rational being. However, the rational being that we then envisage is characterized as the archetype of human rationality. Thus human reason generates from itself a model of rationality which it then projects onto God. Then however the order of explanation is reversed and human rationality is viewed as a derivative copy of that perfected rationality. Thus divine reason is modeled upon human reason only so that human reason can subsequently be viewed as a copy of divine reason. On Kant’s analyses, ordinary agents are unaware of the proper function of their judgments about God, since what human beings are doing when they make such judgments is simply not characterizing an experiencetranscendent object. Rather, such judgments should be properly understood as being “about our language” or about the sense-making practices that human beings require for themselves in navigating the empirical world.25 It is surely these aspects of Kant’s approach that led to Heine’s thought of Kant as a master of suspicion. Kant’s maneuvers were made for the aims of combating the illnesses of religious indifference and distrust in our own reason. The thought that the cure might be more damaging than the illnesses is the legacy of those maneuvers. 25

This element is surely central to the contentious interpretive trend of taking Kant’s philosophy of religion as marking a “shift to an anti-realist mode, or an allegorical mode, or an apophatic mode” (Chignell 2009: 121) or – even more contentiously – as making the claim that “[t]alk that had seemed to be about God turns out, on inspection, to be, literally, about us” (Godlove 2014: 160). For a small sample discussion, see Byrne (2007), DiCenso (2011), Pasternack (2011, 2014b), Wolterstorff (1991, 1998), and Wood (1970).

ch a p ter 1 4

Knowledge, Discipline, System, Hope The Fate of Metaphysics in the Doctrine of Method Andrew Chignell∗

We always return to metaphysics as a beloved from whom we have been estranged. – A850/B878


Kant’s Plan

Despite being dwarfed by “The Transcendental Doctrine of Elements” in both size and influence, “The Transcendental Doctrine of Method” is officially the second main part of the Critique of Pure Reason. It starts with an architectural metaphor: in the first part of the book, Kant says, we “made an estimate of the building materials and determined for what sort of edifice, with what height and strength, they would suffice” (A707/B735). Those materials came from both sensory intuition (Transcendental Aesthetic) and conceptual understanding (Transcendental Analytic). But although “we had in mind a tower that would reach the heavens,” it turned out that speculative a priori reasoning did not offer legitimate materials for the purposes of either knowledge or science (Transcendental Dialectic). In this slender second part of the Critique Kant turns from estimating materials to developing “the plan.” The goal is to avoid the fate of the architects at Babel by constructing “an edifice that is in proportion to the supplies given to us and at the same time suited to our needs.” In other words, the goal is to rejoin the general human project of trying to understand the world – a project “from which we are not able to abstain” – while taking into account what we have discovered about the nature and limits of our cognitive materials (A707/B735). ∗

For feedback on earlier drafts, I’m grateful to the people mentioned in the notes to follow, as well as to Karen Bennett, Angela Breitenbach, Susan Brower-Toland, Alix Cohen, Everett Fulmer, Don Garrett, John Greco, Desmond Hogan, Anja Jauernig, Béatrice Longuenesse, Michela Massimi, Tyke Nunez, Michael Oberst, Jim O’Shea, Konstantin Pollok, Erica Shumener, Ted Sider, and Eric Watkins.



an d rew c h i g n e l l

This way of characterizing the plan – i.e., both negatively and positively – is typical of the entire Critique, but especially of this second part. On the one hand, we have to keep in mind the newfound limits of our materials, and thus “discipline,” “censure,” “humiliate,” “caution against,” “constrain,” “compel,” and even “extirpate” the illicit urge (often arising from reason itself ) to construct an edifice for which our building materials are not adequate (see e.g., A708–11/B736–40). On the other hand, we have to acknowledge our needs as rational, inquiring beings who “lust after knowledge” and “speculative expansion” (A708/B736), who seek “the therefore to every wherefore” (zu allem Warum das Darum) (A585/B613) and thus always “return to metaphysics as to a beloved from whom we have been estranged” (A850/B878). We also have to acknowledge our needs as sensory, moral beings: for most of us, virtue does not suffice as its own reward: we need some reasonable way to hope for individual happiness and collective justice if we are to avoid becoming demoralized as agents. The Doctrine of Method is divided into four chapters: “The Discipline of Pure Reason,” “The Canon of Pure Reason,” “The Architectonic of Pure Reason,” and “The History of Pure Reason.” Each is about half the size of its predecessor, and reading through all four gives the impression of an author who was running out of time, steam, or both. In this paper, I will set aside the “History” chapter, which comprises a three-page sketch of the developments leading up to the critical philosophy, and focus instead on the first three chapters. After providing Kant’s positive account of knowledge (Wissen) in the Canon, I go on to discuss the prohibition (in the Discipline) against synthetic knowledge claims regarding “supersensibles” – i.e., those things of which we cannot have any sensible experience per se. My main proposal is that, from the 1760s onward, Kant held that objects have a modal status – their “real” modal status – that is more restrictive than their merely “logical” modal status. But by 1781, Kant had also come to reject the Cartesian-rationalist idea that we have a faculty of clear and distinct perception that can tell us what is really, metaphysically possible. Naturally we can still think up all sorts of nonactual beings, but such thought merely tracks the wider domain of logical possibility. As a result, we cannot rule out the concern that in speculative contexts we are thinking up concepts of objects – souls, God, freedom, or monads, for instance – that are logically possible but still really impossible. Without the ability to rule this out, Kant concludes, we must restrict knowledge claims to propositions whose really possible truth we can in some way “justify” (rechtfertigen) (A259/B315).

Knowledge, Discipline, System, Hope


If the proposal is accurate, then the modal condition that we be able to “justify” the real possibility of the objects we claim to know is more fundamental than, and even motivates, the empiricistic demand that an object of knowledge be “given” to the mind via some sort of relation to intuitional experience. In the Discipline, Kant argues that analytic judgments, mathematical judgments, and the transcendental judgments of the critical philosophy itself can satisfy this modal condition, but that the synthetic claims of “dogmatic” metaphysics cannot. In the Architectonic, Kant gives us a further clue about what it would be to meet the condition: we must be able to “prove” (in a way to be discussed below) that a proposition’s truth systematically coheres with our background knowledge. In effect, then, I submit that Kant embraces what contemporary epistemologists call a “coherentist” constraint on knowledge, one that is based in our antecedent grasp of nature and its laws. At the end of the paper, I return to the Canon and look briefly at the way in which Kantian “hope” (Hoffnung) satisfies the rational needs that the Discipline quashes in the epistemic sphere. The word “hope” appears far more frequently in the Doctrine of Method than in other parts of the book. Kant’s use of it highlights his two-part message: dogmatic metaphysicians should abandon all hope, but the new critical “metaphysics of experience” offers hope for a new way forward. It also leaves room for rational hope and even belief (Glaube), though not knowledge, regarding at least some of the paradigmatic objects of traditional metaphysics.

14.2 Knowledge and the Canon In order to grasp the account of knowledge (Wissen) in the Canon, it is important to recognize that Kant’s conception of the fundamental positive propositional attitude differs from our contemporary Anglophone concept of belief and its contemporary German equivalent, Überzeugung. For Kant the fundamental attitude is that of holding-for-true (Fürwahrhalten) – a term often translated into English as “assent.” Kant describes Fürwahrhalten in the Canon as a state that “may rest on objective grounds (Gründen), but that also requires subjective sources (Ursachen) in the mind of him who judges” (A820/B848, my emphasis). The character of a particular assent, in other words, is determined both by its subjective features (how firmly the assent is held, what the subject takes it to be based on) and by its objective features (how probable the proposition is on those grounds, whether the assent was appropriately caused,


an d rew c h i g n e l l

etc.). We might hold a proposition weakly as a mere opinion (Meinung), for example, if we acknowledge that we have some limited grounds for it (most hypotheses, for instance, count as opinions). We can also hold a proposition for true in a firm way but for self-consciously “subjective” reasons as a matter of belief (Glaube). Finally, we can hold a proposition for true in a “sufficient” (zureichend) manner from both a subjective and an objective point of view – in that case, it counts as knowledge (Wissen) (A822/B850). I’ve analyzed the notions of objective and subjective “sufficiency” elsewhere and won’t try to reproduce the account again here (see Chignell 2007b).1 The core set of epistemic conditions that results, however, is this: Knowledge (Wissen): S’s assent that p counts as knowledge only if (g) such that (i) g is an objectively sufficient ground that S has, (ii) S’s assent that p is based on g, (iii) p is true, and (iv) g is subjectively sufficient – i.e., on reflection, S would cite g as her objective ground for the assent that p.

Conditions (i), (ii), and (iii) are “external” constraints: the assent must be based on what are in fact sufficient grounds that S has in her possession, and the assent has to be true.2 By contrast, (iv) articulates an “internal” constraint, albeit a very weak one. It says that, for example, if S knows that the rabbit is presently in the garden, S has to be such that, if she were asked why she holds that, she would reference (or “cite”) her visual experience of the rabbit and the garden (or testimony, or induction, or some other objectively good ground for that assent).3 S does not have to be able to say why that experience is a sufficiently good objective ground for the assent, how likely the ground renders the assent, and so forth. That’s what makes the constraint both internal and weak (and plausible!). People familiar with the Critique may wonder at this point whether there is an important condition missing, one that expresses Kant’s conviction, in the Doctrine of Method and throughout, that substantive knowledge must be grounded in intuition: 1 2


See also Stevenson (2003) and Pasternack (2014a). I favor a “fallibilist” or “defeasibilist” picture according to which all the other conditions for knowledge could be met apart from (iii) and the assent still turn out to be false (see Jäsche Logic 9:72 LL 575–76, for instance). For indefeasibilist readings, see Makkreel (2003) and Pasternack (2014a). See, for example, Kant’s talk of a subject being “in a position to make a supposition” about whether a given ground is an objectively good one (Blomberg Logic 24:87–88, LL 66–67). For Kant on testimony, see Gelfert (2006).

Knowledge, Discipline, System, Hope


All of our cognition is in the end related to possible intuitions: for through these alone is an object given. (A719/B747; compare A286/B342)

The proposed condition would be something like (v) g bears an appropriate relation to a possible intuition (pure or empirical).

As stated, this is too vague: if the ground of assent, g, just is an intuition, or is partly constituted by intuition, then the condition is presumably met. But what other relations to intuition count as “appropriate”? Can we make knowledge-preserving inferences from what we do intuit to what we could intuit? Or to unobservables? What about inferences from a structural feature of intuition to its transcendental conditions – do those relations count as appropriate? What does “givenness” amount to here, and why is intuition so important – isn’t the relevant thing just the character of the justifying ground, no matter what its source? In other words, why do we need a condition like (v) if we’ve already got (i)? I will return to these questions below. For now, it is worth noting that (v) as stated won’t do as a necessary condition on knowledge simpliciter. For Kant’s claim in the quotation here is about cognition (Erkenntnis) rather than knowledge (Wissen), and elsewhere he explicitly allows that there is some knowledge that is not based in cognition or related to possible intuition in any interesting way. For example, well-formed analytic judgments allow us “to know what lies in the concept” (wissen, was in seinem Begriffe liegt) (A259/B315), but aren’t necessarily based in intuitions of the objects of the concepts (I don’t have to perceive any bachelors in order to know that they’re all male). In speculative contexts, moreover, some of the concepts will pick out things-in-themselves (God, the soul, free wills, monads, etc.) – and thus supersensibles of which we can’t have intuition. But surely Kant would allow (under threat of performative contradiction, given all those lectures on metaphysics and philosophical theology) that we have some analytic knowledge of their contents. Similarly, we might know about a domain of things in a wholly negative fashion without having intuition of those things – consider here the assent that things-in-themselves are not in space and time. The objectively sufficient grounds of such knowledge might be Kant’s transcendental arguments establishing that space and time are the mere forms of our receptive sensible intuition, together with the fact that things-in-themselves are, by definition, supersensible (see B307). As we will see, Kant suggests in places that we have positive but very general knowledge about such things – that


an d rew c h i g n e l l

things-in-themselves exist and ground appearances, for example (Prol. 4:314– 15, TP2 107–8). But any such negative or very general positive knowledge about things-in-themselves will not be based in cognition or intuition of those things (see B149). There is more to be said about these cases, but this suggests that, for Kant, there is some knowledge that is not related to actual or possible intuition in an epistemically significant way. It also emphasizes the need to find a revised version of (v) that is adequate to such cases and yet allows us to preserve a unified account of Kantian knowledge. Ideally, the revised version would also explain Kant’s claims about the importance of intuition as well as his prohibition on synthetic knowledge claims about the positive characteristics of specific things-in-themselves.

14.3 Ignorance and the Discipline A familiar thought comes to mind at this juncture: Kant often says that our concepts of supersensibles are “useless” in epistemic contexts because “concepts without intuitions are empty.” In other words, it’s not that the inferences in speculative arguments are always invalid (though some of them clearly are), or that one of the premises must be false, but rather that some of the concepts lack the right kind of content (see, e.g., A62/B87; B148; A220/B267).4 Without a connection to possible intuition, Kant says in the Discipline, we risk “basing our reasoning on empty figments of the brain rather than concepts of real things” (A770/B798). He often uses hylomorphic images to describe what he has in mind: the “form” of our concepts has to be connected to the “matter” of intuition in order for us to be sure we’re not merely groping among “thought-entities” (Gedankenwesen) or “playing with fancies instead of concepts and words instead of things” (A710/B738; see also A723/B751, A771/B799, and A239/B298). This talk of emptiness, brain-figments, matter, and groping is clearly metaphorical. No serious rationalists – old- or new-fangled – will find it persuasive on its own, nor will they accept a bald stipulation that concepts without a connection to possible intuitions have no epistemic use. Some 4

Regarding his own theistic proof from 1763, for instance, the Critical Kant says that “this proof can in no way be refuted (allein widerleget kann er auf keine Weise werden), because it has its ground in the nature of human reason. For my reason makes it absolutely necessary for me to accept (annehmen) a being which is the ground of everything possible, because otherwise I would be unable to cognize (erkennen) what in general the possibility of something consists in” (LRel 28:1034, RRT 375). The “acceptance” here is what he calls, in the Canon, “doctrinal belief” (see below for more discussion).

Knowledge, Discipline, System, Hope


commentators seem willing to take the demand for intuitional “matter” here as indicative of a broadly empiricist commitment that Kant adopted after Hume woke him from his slumbers, and leave it at that.5 But why did he go in that direction? What was the root concern? Why must our concepts, as well as the assents that involve them, be related in some way to intuition? And, again, which kinds of relation are appropriate? A related line of thought is this: in order to claim knowledge about any object at all, we have to have some sort of mental grasp of or reference to that object. Perhaps this is what Kant means when he repeatedly demands that objects be “given” to the mind and that concepts without intuitions are “without sense or reference” (A155–56/B194–95; A721/B729). But again, why must the “giving” occur by way of a relation to actual or possible intuitions, rather than by simply generating and entertaining concepts? When Descartes entertains the idea of God or an immaterial soul in Meditations, he surely has some sort of mental grasp of what he is discussing. But then why aren’t those things “given” to him in the relevant sense? To insist that mental “givenness” just has to go by way of a connection to intuition again looks merely stipulative.6 I think we can go beyond metaphors and stipulations by viewing Kant’s prohibition on synthetic knowledge of particular supersensibles as arising more organically out of his lifelong reckoning with his rationalist predecessors. Here’s the story in brief:7 in the 1760s, Kant spied a metaphysical difference between “logical” possibility and real possibility; he also saw the related point that there can be “real opposition” (reale Entgegensetzung) or “real repugnance” (Realrepugnanz) between logically consistent positive predicates. In the “Negative Magnitudes” (NM) essay of 1762, for instance, he cites the example of two equally powerful winds blowing from opposed directions on a sail: they “cancel out” (aufheben) one another, and the ship remains at rest (NM 2:171, TP1 211). In that essay as well as in The Only Possible Argument (OPA) of the following year, Kant also mentions cases of subject-canceling rather than merely predicate-canceling real opposition. This form of real opposition doesn’t merely cancel out the effects of the predicates involved; rather, it “cancels” the subject altogether. One of Kant’s 5 6 7

A. W. Moore points out that Kant first uses the “awakening” metaphor in connection with Hume here in the Discipline at A764/B792 (Moore 2010). For more on “givenness” and cognition, see Watkins and Willaschek (2017), Grüne (2017), and Chignell (2017). See Chignell (2011, 2014a) for a longer version of the story, and Abaci (2014) and Yong (2014) for objections to it.


an d rew c h i g n e l l

examples is that “the impenetrability of bodies, extension, and the like, cannot (können nicht) be properties of what has understanding and will.” It’s not that being impenetrable and having understanding are logically inconsistent: there is no way to generate a contradiction from their conjunction using standard rules and definitions. Rather, it’s that “these predicates can by no means co-exist together as determinations in a single subject” (nimmermehr in einem einzigen Subject als Bestimmungen neben einander können statt finden) (OPA 2:85, 130, my emphasis). The ‘cannot’ and ‘can’ in these sentences clearly express real rather than logical modalities: Kant thinks there simply cannot be a subject that is both a body and has understanding. By the time of the Critique, Kant regarded the rationalists’ alleged neglect of these nonlogical constraints on possibility as a serious mistake. In the “Phenomena/Noumena” chapter, for example, he points out that the real possibility of something cannot be established by mere thinking: That the not-being of a thing does not contradict itself is a lame appeal to a logical condition, which is certainly necessary for the concept but far from sufficient for real possibility. (A244/B302, my emphasis)

He laments in the Amphiboly that with respect to the concept of a supreme being, for instance, the rationalists find it not merely possible but also natural to unite all reality in one being without any worry about opposition, since they do not recognize any opposition except that of contradiction. (A273–74/B229–30)

The mistake here is modal: when Leibniz and Wolff seek to demonstrate the existence of a supreme being through armchair speculation, they just presume, as Kant puts it in a lecture, that they have “insight (Einsicht) into whether all realities could be united together in one object (Objekt), and hence into how God is possible” (LRel 28:1025–27, RRT 368–69, my emphasis).8 But, again, if there are nonlogical, real constraints on possibility – constraints that can’t reliably be tracked via mere thought – then that presumption looks hasty. For all we know, in such a speculative 8

Other cases of predicate-canceling real impossibility are found in the Metaphysical Foundations of Natural Science (MF). For instance, “two motions” that are “combined in precisely opposite directions in one and the same point” are such that they cancel the entire subject to which they are ascribed: “[R]epresenting two such motions at the same time in exactly the same point within one and the same space would be impossible, and thus so would the case of such a composition of motions itself” (MF 4:491, TP2 203–4). Another example: a material being “is impossible if it has mere attractive forces without repulsive forces,” and this impossibility has its basis in “the essence of matter” rather than in a logical contradiction (MF 4:511, TP2 222).

Knowledge, Discipline, System, Hope


context, the concept includes predicates that are really opposed in such a way as to cancel out the subject itself – that is, to make it really impossible. It is in direct response to these metaphysical and epistemological concerns about rationalist modal theory, I submit, that Kant develops his fifth epistemic condition. “I can think whatever I like,” he says in the B-Preface, as long as I do not contradict myself, i.e., as long as my concept is a possible thought, even if I cannot give any assurance as to whether or not there is a corresponding object [Objekt] somewhere within the sum total of all [real] possibilities. . . . [But] to cognize an object, it is required that I be able to prove its [real] possibility (whether by the testimony of experience from its actuality or a priori through reason)” (Bxxvin).

This passage is focused on cognition, but the few kinds of knowledge that are not based in cognition can easily meet this modal condition. In analytic contexts, the relevant “objects” are just the concepts we are analyzing,9 so if we can “prove” that the concepts themselves are actual (by being aware that we have them), then we can a fortiori prove that they (though not necessarily their objects) are really possible.10 If the synthetic but very general assent that things-in-themselves exist and ground appearances counts as knowledge, then we can find “proof” of their (or its) real possibility in the fact that their (its) actual existence is a condition of the existence of any appearances at all. In the Preface, Kant seems to go this direction: he says it would be “absurd” for there to be appearances without there also being something that appears, i.e., something that grounds those appearances (Bxxvi). In the Prolegomena, he says, similarly, The understanding, just by the fact that it accepts appearances, also admits to the existence of things in themselves, and to that extent we can say that the representation of such beings as underlie the appearances, hence of mere intelligible beings, is not merely permitted but also unavoidable. (Prol. 4:315, TP2 107; see also Prol. 4:355, TP2 144)

Commentators disagree about whether this is supposed to be self-evident, or a conceptual truth, or a quick deductive inference, but many follow Erich Adickes in holding that it provides a “proof” of the actuality (and thus real possibility) of noumenal grounds taken collectively, though not 9


“For an analytic assertion takes the understanding no further, and since it is occupied only with that which is already thought in the concept, it leaves it undecided whether the concept even has any relation to objects . . . ; it is enough for [the subject] to know (wissen) what lies in the concept; he is indifferent to what the concept might pertain to” (A 258–59/B 314). “That the concept (thought) is possible is not an issue; the issue is rather whether it relates to an object and therefore signifies anything” (B 302–3n).


an d rew c h i g n e l l

of any specific, determinate thing (monad, deity, soul, etc.) among them.11 Any negative knowledge about the same objects, such as that things-inthemselves are not in space and time, would presumably be based on the same “proof” as well as Kant’s transcendental reflections about the nature of space and time. Since the modal condition here obviously applies to all knowledge that is based in cognition and, as we have seen, to the few kinds of noncognitional knowledge that Kant seems to allow, it is a nice candidate for a revised version of the fifth condition on knowledge. Here is a first stab: (v∗ ): S is in a position to prove the real possibility of the objects referred to in p.

A remaining problem with this is that we clearly know, about impossible objects, that they are impossible. For instance, we know that intersecting parallel lines cannot exist. Kant says that in such cases “the impossibility rests not on the concept itself but on its construction in space, i.e., on the conditions of space and its determinations” (A220–21/B268; cf. A224/B271). This indicates that the modality in such a case is real rather than merely logical, and that the “proof of possibility” that Kant has in mind is something more like proof of whether or not the objects referred to are really possible (think of phrases like “prove your mettle” or “proof of the pudding” – we test for the presence of a certain property rather than proving that the property is positively there). This suggests the following revision: (v∗∗ ) for any object referred to in p, if it is really possible then S is in a position to prove its real possibility, and if it is really impossible then S is in a position to prove its real impossibility.

More work would be required to go beyond propositions with atomic structure,12 but for now I propose to take (v∗∗ ) as a good approximation of Kant’s fifth condition on knowledge. 11


Adickes took this belief in the existence of things in themselves to be simply “self-evident” for Kant (see Adickes [1924] and the extensive discussion of it in Bird [2006: Chapter 23]). Some commentators want to avoid interpreting this or any other text as licensing a synthetic existence-claim about things-in-themselves (Bird 2006: 42–44, 559–63; cf. O’Shea 2014: 106–15). If they are right, then the modal condition I’m developing here would be in even better shape, since cognition is clearly governed by a modal condition for Kant. We would need to know, for instance, about knowledge of (a) conditionals (I can know that “if there are unicorns, then there are horns” is true without proving that unicorns are really possible), (b) negations (I can know that “it’s not the case that there is a golden mountain in the room” without proving that golden mountains are really possible or impossible), and (c) disjunctions (I can know that “I am writing a paper or God is a deceiver” without proving that God is really possible). My sense is that (a) can be assimilated to conceptual or broadly analytic knowledge that is about the concepts rather than the objects of those concepts, (b) can be interpreted as about the room and the items in it, rather than about the golden mountain, and in (c) only the disjunct that makes the entire disjunction true has to meet the modal condition.

Knowledge, Discipline, System, Hope


14.4 Possibility and the Postulates At this point, the obvious question is: well, what sort of “proof” or “justification” of real possibility is available? Three different strategies emerge in Kant’s discussions in the Discipline and, more expansively, in the “Postulates of Empirical Thought” (the chapter on modality). 14.4.1 First Strategy: Appeal to Actuality According to Locke, if we propose to combine some qualities in the complex idea of some substance or kind, we must actually experience the individual qualities and their combination.Without such actual experience, according to Locke, we can’t be sure that the qualities are not “in-co-existent” (Locke’s term: Essay 4.3.12). This is part of what leads him to the doctrine that things may have real essences that ground the “strict union” of all sorts of qualities that we haven’t experienced at all – or haven’t experienced together – but that a good empiricist will restrict her knowledge claims to propositions about nominal essences that contain qualities which “we can be sure are, or are not, inconsistent in Nature.” And the only way she can be sure of that is by appeal to “Experience and sensible Observation” (Essay 4.4.12). In places, Kant sounds downright Lockean on this question: In a word, all of these concepts could not be vouched for (belegen) and their real possibility thereby established (dartun), if all sensible intuition (the only one we have) were taken away, and there then remained only logical possibility. (B302–3n) Now, however, the possibility of a thing can never be proved (bewiesen werden) merely through the non-contradictoriness of a concept of it, but only if is vouched for (belegt) by an intuition corresponding to the concept. (B308)

The appeal to intuition works in some a priori contexts as well: in the Discipline, Kant emphasizes that many mathematical judgments satisfy the modal condition by appealing to intuitive “constructions” of their objects in pure intuition. Such construction proves their actuality, and that trivially entails their real possibility (see A718/B746). 14.4.2

Second Strategy: Appeal to Formal Possibility

Elsewhere in the Discipline, though, Kant seems willing to allow that we can prove real possibility in ways that do appeal to actual experience. One strategy involves appealing to the “conditions of possible experience”:


andrew chignell In a word: our reason is only able to use the conditions of possible experience as conditions of the possibility of things (Sachen); but it is by no means possible for us as it were to create new ones independent of those conditions, for concepts of the latter sort, although free of contradiction, would nevertheless also be without any object (Gegenstand). (A771/B799; see also A602/B630, A610/B638)

In order to understand this claim, we need to look at the account of the category of possibility that Kant provides in the “Postulates” chapter: [Formal Possibility:] That which agrees (übereinkommt) with the formal conditions of experience (according to intuition and concept) is possible. (A218 /B265)

What is formally possible, in other words, is what agrees with the axioms of space and time (“according to intuition”) and the principles derived from the categories (“according to concept”) – in particular the Analogies of Experience. The latter principles state that all objects of our experience must be persisting substances whose states are nomologically determined and which stand in dynamic relations with all other such substances.13 Although this definition of formal possibility is relatively clear, Kant’s illustrations of how it operates are obscure. With Swedenborg and other enthusiasts in the background, he mentions concepts like that of [Ghostly Matter]: a substance that is persistently present in space yet without filling it, or [Soothsaying]: a special fundamental power of our mind to intuit the future (not merely, say, to deduce it), or [Telepathy]: an ability of the mind to stand in a community of thoughts with other men (no matter how distant they may be).

These concepts may be logically consistent, Kant says, but they are concepts the possibility of which [i.e., the real possibility of whose objects] is completely baseless, because it cannot be grounded upon experience and the laws [of experience] with which we are acquainted (auf Erfahrung und deren bekannte Gesetze gegründet werden kann), and without this is an arbitrary combination of thoughts that, although it contains no contradiction, still can make no claim to objective reality, thus to the [real] possibility of the sort of object that one would think here. (A223/B270, my emphasis)14 13 14

For versions of the taxonomy of kinds of modality here, see Chignell (2010b, 2011), Abaci (2016), Chignell and Stang (2015), Kannisto (2013), Leech (2014), and Stang (2016). This quotation is from the Postulates, but Kant cites the same examples in the Discipline at A770/B798, and goes on to say that they are “mere thought-entities, the [real] possibility of which is not demonstrable, and which therefore cannot be used to ground the explanation of actual appearances” (A771/B799).

Knowledge, Discipline, System, Hope


It is not obvious what is meant here by “grounded upon the laws of experience with which we are acquainted.” Given the definition of possibility just offered, it seems that Kant might mean in agreement with the formal conditions of experience. But then it is unclear how all of the ghostly phenomena Kant mentions would fail this test. Telepathy, in particular, seems like a hard case: the existence of such a faculty is clearly compatible with the formal axioms of space and time, and it is hard to see how it would be ruled out by the principles of the understanding (such as the Analogies). Building more of the laws of nature into the formal conditions might be thought to help here. Perhaps, for instance, we can take the first Analogy to demonstrate that all outer objects of experience are composed of persisting extended, space-filling matter – matter that is not, thus, ghostly.15 In Metaphysical Foundations of Natural Science (1786, MF), Kant himself seems to try to underwrite more substantive claims about matter and mechanical laws by appeal to formal contributions of the mind as well as a priori “construction.” But this strategy won’t provide proof of real possibility in all cases. For Kant makes it clear in the B-edition Critique, written after the Metaphysical Foundations, that most of the laws of nature are “particular,” i.e., not part of the formal “constitutive a priori” contribution of the mind: Particular laws, because they concern empirically determined appearances, cannot be completely derived from the categories, though they all stand under them. Experience must be added in order to know particular laws at all. (B165; see also A206/B 252; A 222/B 269; CJ 5:179–80)

So even if we can prove a priori that ghostly matter is formally impossible by appealing to the robust principles of “the metaphysics of corporeal nature” (MF 4:472, TP2 187), telepathy as well as many other events and objects may “agree” with the formal principles but still be really impossible given how nature actually is. This leads to a third sort of strategy, one that falls between the two just considered. 14.4.3 Third Strategy: Appeal to Empirical Possibility The third strategy invokes a modal notion that can be derived from the definition of necessity that Kant offers in the same “Postulates” chapter: 15

This reflects an amendment to the position in Chignell (2014b), from which I have drawn some of the present discussion. I am grateful to the editor of that journal for permission to reuse this material here.


an drew c h ig n el l [Empirical necessity:] That whose connection with the actual is determined according to universal conditions (bestimmt nach allgemeinen Bedingungen) on experience is (i.e. exists) necessarily. (A218/B266)

The “universal conditions on experience” are thicker or more “particular” than the merely formal conditions referred to in the definition of possibility, as Kant’s discussion goes on to indicate. Any object (or change of state) that is connected to actual events via the formal conditions or the much more particular “empirical laws of causality” counts as necessary in this sense (see A 227/B 280). It thus seems appropriate to call this kind of necessity empirical. We can then define a counterpart conception of possibility: Empirical possibility: Something is empirically possible iff it is not empirically necessary that it is not the case. In other words, an object or event is empirically possible iff its existence agrees with the universal conditions on experience – i.e. the formal conditions plus the particular laws and preceding actual events.

Figure 14.1 provides a depiction of the various concentric modal domains, from the very broad domain of logical possibility all the way down to empirical actuality. The telepathy example is what motivates the need to make the rather ugly space for things that are formally possible but not

Constrained by empirical laws Constrained by forms of intuition, categorial structure Constrained by metaphysical facts about natures, compossibility, repugnance, etc.

Empirically Actual Empirically Poss

Constrained by logical laws

Formally Possible Really Possible Logically Possible

Figure 14.1

Knowledge, Discipline, System, Hope


really possible. And any existing things-in-themselves are logically and really possible but not empirically possible or empirically actual. So the diagram would have to be more complex (and three-dimensional) in order to represent their modal status.16 With this picture of the modal situation in the background, consider now the following account of what it would be to be able to “prove” real possibility: Third strategy: S is in a position to prove the real possibility of an object if S is in a position to prove its empirical possibility.17

This specifies the correct modal domain, I think, but we still haven’t said much about what it would mean to “prove” such a thing, especially since our understanding of what is empirically possible is often inductive and provisional. The best way to go here is to return to the Postulates passage where Kant appeals to the “laws of experience with which we are acquainted.” He elaborates as follows: If one wanted to make entirely new concepts of substances, of forces, and of interactions from the material that perception offers us, but without borrowing (entlehnen) the example of their connectedness from experience itself, then one would end up with nothing but brain-figments for the [real] possibility of which there would be no indications at all, since in their case one did not accept experience as instructress nor borrow these concepts from it. (A222/B269)

The appeal to our “experience as instructress” suggests that the real possibility of the items referred to must itself agree with what the subject already knows about the world and its workings. Kant says something similar in the Discipline: For the explanation of given appearances no other things and grounds of explanation can be adduced than those which are connected to the given appearances by already known laws of appearances. (A772/B800, my emphasis)

In other words, even if S is not able to prove that an object’s possible existence agrees with what in fact is true of nature and its laws, she might be 16 17

Thanks to Iakovos Vasiliou, Rachel Cristy, and Avi Appel for helpful discussion of this diagram. I leave it as a sufficient condition here, since as we have seen there may be other ways to prove the real possibility of objects of analytic knowledge, negative knowledge, and very general positive pieces of synthetic knowledge.


andrew chignell

able to show that its possibility agrees with her own background knowledge of nature and its (formal and particular) laws. Let’s revisit the telepathy case with this in mind. Even if I do have experiences whose best explanation appeals to such a relation (the professional mind-reader constantly tells me precisely what I am indeed thinking), and thus even if my assent is highly probable on the grounds I possess and would cite, I am not in a position to prove that the possibility of such a phenomenon agrees with my existing background knowledge of nature and its laws. So a case like this will not satisfy the modal condition in (v∗∗ ), even if it does satisfy conditions (i)–(iv). In the Discipline, Kant goes so far as to say that we can’t even form mere hypotheses or opinions for or against such things without satisfying something like this modal condition: Merely intelligible beings or merely intelligible properties of the things of the sensible world cannot be assumed even in opinions with any well-founded authority of reason, although (since one has no concept of either their possibility or their impossibility) they also cannot be dogmatically denied on the basis of any supposedly better insight. (A772/B800, my emphasis)

The strategy revised in light of all this, then, would be: Third Strategy∗ : S is in a position to prove the real possibility of an object if S is in a position to prove that its possible existence agrees with S’s background knowledge of nature and its laws.

This is getting close. There is still a lingering question, however, about the notion of “agreement” involved. Kant uses “übereinkommen” or “zusammenhängen.” But does this mean that the possible existence of the relevant objects must be provably consistent with our background knowledge of nature? If so, then (v∗∗ ) would be largely impotent: logical consistency with the laws is pretty easy to come by (almost all the supersensibles and ghostly phenomena would achieve it, for instance). But then does it mean that the existence of the relevant objects must be provably compossible with what we know about nature? If so, then the condition in (v∗∗ ) would smuggle in the presumption that we’re able to discern, from the armchair, what is really compossible with what. But that was the presumption that Kant found problematic at the start! So does “übereinkommt” mean something like follows from or is entailed by? If so, then it would be hard to see how some new assents could satisfy the condition and count as knowledge. For in many cases we would presumably posit the existence of new kinds of objects or forces along with the laws governing their relations. And so their

Knowledge, Discipline, System, Hope


empirical possibility could not be entailed by what we knew about nature antecedently.18 It’s often wise to go for a middle way in Kant interpretation. Here an attractive thing to say (I submit) is that Kant has in mind what we now call a positive coherence relation when he speaks of “agreement with” or “conformity to” experience and its known laws. It’s not merely that the possibility of the objects referred to is consistent with our background knowledge of nature, but it’s not that it is entailed by it either. Rather, there are positive coherence relations between the claim that such items are possible and our19 background knowledge of the way the world works. As Kant says at A537/B565 “appearances . . . are . . . mere representations, which cohere according to empirical laws” [Erscheinungen . . . sind . . . bloße Vorstellungen, die nach empirischen Gesetzen zusammenhängen]. In other words: Third Strategy∗∗ : S is in a position to prove the real possibility of an object if S is in a position to prove that its possible existence positively coheres with S’s background knowledge of nature and its laws.

Talk of “proof” may sound too ambitious here, but recall that elsewhere Kant is willing to use weaker expressions such as “justify the possibility” (A259/B315). For the vast majority of our synthetic knowledge, I think this strategy describes the way the modal condition in (v∗∗ ) is satisfied.

14.5 Coherence and the Architectonic It is not unusual for coherentist accounts of a particular item of knowledge to make reference to other pieces of a subject’s background knowledge. Still, in order to flesh out the proposal just sketched, we’d need to say more about what Kant takes positive coherence to consist in. My suggestion, left undeveloped here, is that this is where Kant’s discussion of the importance of “systematicity” in the Architectonic (as well as the Appendix to the Dialectic and the third Critique) plays a crucial role. To know a proposition involves not just having probabilistic grounds for it and being able to cite those grounds when one assents to it. Rather, one also has to be justified in 18 19

Consider in this connection Kant’s discussion of why we can cognize the existence of unobservable “magnetic matter” at A226/B273. I leave the scope of “our” vague here, but note that this may make it harder for experts in a domain to acquire knowledge of certain truths than it is for others, since experts have so much more background knowledge against which to test the new proposition.


andrew chignell

holding that the real possibility of the objects it refers to positively coheres with one’s background knowledge.20 The goal is to form an edifice (here Kant returns to his architectural metaphor) in which all the assents are purposively united with each other as members of a whole in a system of human cognition, and allow for an architectonic to all human knowledge. (A835/B863)

This is not a mere scientific ideal – as we have seen, Kant’s view is that without at least some sense of the background system, we can have “no coherent use of the understanding, and, lacking that, no sufficient mark of empirical truth; thus in regards to the latter we must presuppose the systematic unity of nature as objectively valid and necessary” (A651/B679). I have argued that the modal condition on knowledge, interpreted in the manner of (v∗∗ ), underwrites Kant’s frequent appeals to the need for a relation to possible intuition, for objectively real (as opposed to “empty”) concepts, and for the “givenness” of objects. I have also suggested that (v∗∗ ) should be read along the lines of (Third Strategy∗∗ ): for S to be in a position to “prove” that an object is really possible is for S to be able to justifiably claim that its real possibility positively coheres with her background knowledge of nature and its laws. Here’s a final case to consider: suppose you infer from the harmonious and fecund character of the natural laws as a whole to the existence of a supersensible world-author. Kant seems to regard such inferences as sound, and as providing probabilistic grounds for the conclusion (see A624/B652; A826/B854; Prol 4: 278, TP2 74). So (i)–(iv) are satisfied in this case. So why wouldn’t this count as knowledge?21 The question indicates that (Third Strategy∗∗ ) has to be read as invoking our background knowledge of the content of the laws that take the world from one state to the next. In other words, when Kant says we should look to the “universal conditions” on experience as our instructress in systematizing our assents, he is talking not about second-order features of the laws themselves (their elegance, simplicity, etc., taken as a set), but about what the laws say regarding which event-types follow other event-types. Assent to the existence of the world-author fails to satisfy (v∗∗ ) in this way, then, because we are never in a position to show that there are formal or particular laws whose specific content makes the possibility of a spiritual author of the world seem more or less likely. We are simply ignorant of the 20 21

For good discussions of this, see Geiger (2003), Bird (2006: 744–52), and Gava (2014). Thanks to Michael Friedman for raising this case in conversation.

Knowledge, Discipline, System, Hope


real modal status of such a being, and so the strongest attitude we can take toward it is, as Kant explicitly tells us in the Canon, a form of belief – i.e., “doctrinal belief” (doktrinale Glaube) (A826/B854). For similar reasons, we also can’t establish positive coherence between our background knowledge of the world and the possibility of an immaterial substance, or a zombie, or a free will, or an ens realissimum, even if we have sufficient objective grounds for positing their existence. So speculative arguments that begin or end with such things (Descartes’s Real Distinction argument, contemporary conceivability arguments for dualism, the ontological argument, and even Kant’s own possibility proof ) are ruled out of epistemic bounds, even if they satisfy (i)–(iv). This is the result we wanted, and it indicates that condition (v∗∗ ) interpreted as a positive coherence condition is a prime candidate for inclusion in a unified account of Kantian knowledge.

14.6 Hope, Belief, and the Canon Again In the Discipline, as we have seen, Kant’s main goal is to dash the hopes of traditional metaphysicians by showing that their speculations about thingsin-themselves cannot result in knowledge. He also argues that opponents of traditional metaphysics overreach when they claim, equally dogmatically, that a disproof of speculative theses is in the offing. The proper attitude, in a theoretical-epistemic context at least, is suspension of assent, together with a little bravado: Thus, think up for yourself the objections which have not yet even occurred to your opponent, and even lend him the weapons or concede him the most favorable position that he can desire. There is nothing in this to fear, and much to hope, namely that you will come into a possession that can never be attacked in the future. (A778/B806)

Kant goes on to say more about that secure “possession” in the Canon and the second Critique. Because the needs to which metaphysical speculation respond are rational and legitimate, we can reasonably hope that at least some of them will be fulfilled in a nonepistemic context. In the end Kant is famously willing to endorse both hope and full-blown “belief” (Vernunftglaube) that is based on “subjective” grounds. Most of these subjective grounds are moral: Thus without a God and a world that is now not visible to us but is hoped for, the majestic ideas of morality are, to be sure, objects of approbation


and rew chignell and admiration but not incentives for resolve and realization, because they would not fulfill the whole end that is natural for every rational being and determined a priori and necessarily through the very same pure reason. (A813/B841)

But as mentioned earlier Kant also speaks in the Canon of “doctrinal” forms of belief (Glaube). The latter “must not be called practical” but rather “theoretical,” and it is often directed toward traditional objects of speculation: the existence of God and “the future life of the human soul” (A826– 27/B854–55). Moreover, when doctrinal belief is formed for the right “subjective” reasons – in this case as a response to our speculative need to find ultimate explanations – it is fully rational.22 This is presumably why Kant says, even in the Discipline, that “as far as the critique of the grounds of proof of the dogmatic affirmations is concerned, one can very well concede it all without thereby giving up those propositions [about the existence of God and the soul], which still have at least the interest of reason in their behalf, to which their opponent cannot appeal at all” (A741/B769). Interestingly, Kant thinks doctrinal belief is rational even though it doesn’t meet the modal condition that he places on both opinion and knowledge. The same thing goes for moral belief: [T]here is a ground of assent that is, in comparison with speculative reason, merely subjective but that is yet objectively valid for a reason equally pure but practical . . . objective reality is given to the ideas of God and immortality and a warrant [Befugnis], indeed a subjective necessity (a need of pure reason) is provided to accept [anzunehmen] them, although reason is not thereby extended in theoretical cognition and, instead, all that is given is that their [real] possibility, which was hitherto only a problem, here becomes an assertion and so the practical use of reason is connected with the elements of the theoretical. (CPrR 5:4–5; compare A818/B846)

It is fitting that here at the end of the first Critique and the beginning of the second, Kant delivers what he promised back in the Preface – namely, an account of what knowledge is, of why we must deny knowledge of thingsin-themselves, and of how this still leaves room for both practical and theoretical varieties of belief (Glaube) regarding some such things. He reiterates this point in the Discipline: “What is in dispute here is not the topic [i.e., the doctrine] but the tone. For enough remains left to you to speak the language, justified by the sharpest reason, of a firm belief, even though you must surrender that of knowledge” (A744–45/B772–73). The Doctrine of 22

For more on “doctrinal belief,” see Chignell (2007a), Pasternack (2010), Gava (forthcoming), Pickering (2016).

Knowledge, Discipline, System, Hope


Method thus offers a sophisticated account of how the “elements” of cognition that were estimated in the first part of the Critique can be used, in conjunction with a disciplined “plan,” to build a secure edifice of systematic knowledge, and yet still leave room for beloved metaphysics in the mode of both hope and belief.


Abaci, Uygar 2014. “Kant’s Only Possible Argument and Chignell’s Real Harmony,” Kantian Review 19.1: 1–25. 2016. “The Coextensiveness Thesis and Kant’s Modal Agnosticism in the ‘Postulates,’” European Journal of Philosophy 24.1: 129–58. Abela, Paul 2002. Kant’s Empirical Realism. Oxford: Clarendon Press. Adams, Robert Merrihew 1994. Leibniz: Determinist, Theist, Idealist. Oxford: Oxford University Press. Adickes, Erich 1924. Kant und das Ding an Sich. Berlin: R. Heise. Allais, Lucy 2004. “Kant’s One World,” British Journal for the History of Philosophy 12.4: 655–84. 2007. “Kant’s Idealism and the Secondary Quality Analogy,” Journal of the History of Philosophy 45.3: 459–84. 2015. Manifest Reality: Kant’s Idealism and His Realism. Oxford: Oxford University Press. Allison, H. 1983. Kant’s Transcendental Idealism: An Interpretation and Defense. New Haven, CT: Yale University Press. 2004. Kant’s Transcendental Idealism: An Interpretation and Defense. Rev. and enlarged ed. New Haven, CT: Yale University Press. Ameriks, Karl 2000. Kant’s Theory of Mind. 2nd ed. Oxford: Clarendon Press. 2003. Interpreting Kant’s Critiques. Oxford: Clarendon Press. 2006. Kant and the Historical Turn. Oxford: Clarendon Press. 2012. Kant’s Elliptical Path. Oxford: Oxford University Press. Anscombe, G. E. M. 1975. “The First Person,” in Samuel Guttenplan (ed.), Mind and Language. Oxford: Clarendon Press, pp. 45–65. Aquinas, St. Thomas 1948. Summa Theologica. Fathers of the English Dominican Province (trans.). Rev. ed. New York: Benziger Brothers. 1952. Truth, vol. 1. R. W. Mulligan (trans.). Chicago: Henry Regnery. Aristotle 1831a. Opera, ed. I. Bekker, vol. I. Berlin. 1831b. Analytica priora I, in 1831a: 1:24–52. Arnauld, Antoine, and Pierre Nicole 1996. Logic, or, The Art of Thinking. Jill Vance Buroker (trans.). Cambridge: Cambridge University Press. Bader, Ralf M. 2012. “The Role of Kant’s Refutation of Idealism,” Archiv für Geschichte der Philosophie 94.1: 53–73. 280



2013. “Self-Knowledge in §7 of the Transcendental Aesthetic,” in S. Bacin (ed.), Proceedings of the XIth International Kant Kongress, vol. 2. New York: de Gruyter, pp. 531–40. Baumgarten, A. (2013 [1739/1757]). Metaphysica. 4th ed. English trans. Metaphysics, C. D. Fugate and J. Hymers (eds.). New York: Bloomsbury. Beck, Lewis White 1975. Essays on Kant and Hume. New Haven, CT: Yale University Press. 1978. “Did the Sage of Königsberg Have No Dreams?” in Essays on Kant and Hume. New Haven, CT: Yale University Press, pp. 38–61. Beiser, Frederick 2006. “Moral Faith and the Highest Good,” in Paul Guyer (ed.), Cambridge Companion to Kant and Modern Philosophy. Cambridge: Cambridge University Press, pp. 588–629. Bennett, Jonathan 1966. Kant’s Analytic. Cambridge: Cambridge University Press. 1974. Kant’s Dialectic. Cambridge: Cambridge University Press. Bird, Graham 1966. “The Beginning of the Universe,” Aristotelian Society XL(Suppl.): 139–50. 2006. The Revolutionary Kant: A Commentary on the Critique of Pure Reason. Chicago: Open Court. 2011. “Replies to My Critics,” Kantian Review 16.2: 274–82. Boulter, Stephen 2011. “The Mediaeval Origin of Conceivability Arguments,” Metaphilosophy 42.5: 617–41. Brandt, Reinhard, and Werner Stark (eds.) 1987. Neue Autographen und Dokumente zu Kants Leben, Schriften und Vorlesungen. Hamburg: Felix Meiner. Brewer, K., and E. Watkins 2012. “A Difficulty Still Awaits: Kant, Spinoza, and the Threat of Theological Determinism,” Kant-Studien 103: S163–87. Brook, Andrew 1994. Kant and the Mind. Cambridge: Cambridge University Press. Buchdahl, Gerd 1969. Metaphysics and the Philosophy of Science the Classical Origins: Descartes to Kant, vol. 46. Buroker, Jill Vance 2006. Kant’s Critique of Pure Reason: An Introduction. Cambridge: Cambridge University Press. Byrne, Peter 2007. Kant on God. New York: Ashgate. Caird, Edward 1889. The Critical Philosophy of Immanuel Kant. 2 vols. Glasgow: Maclehose. Callanan, John J. 2008. “Kant on Analogy,” British Journal for the History of Philosophy 16.4: 747–72. 2013. “Kant on Nativism, Scepticism and Necessity,” Kantian Review 18.1: 1–27. doi:10.1017/S136941541200026X. Carnap, Rudolf 1950. “Empiricism, Semantics, and Ontology,” Revue International de Philosophie 4: 20–40; 2nd rev. ed. in: idem., (1956), 205–21. 1956. Meaning and Necessity. 2nd rev. ed. Chicago: University of Chicago Press. Chignell, Andrew 2007a. “Belief in Kant,” Philosophical Review 116.3: 72–101. 2007b. “Kant’s Concepts of Justification,” Noûs 41.1: 33–63. 2009. “‘As Kant Has Shown . . . ’: Analytic Theology and the Critical Philosophy,” in Analytic Theology. Oxford: Oxford University Press, pp. 117–35. 2010a. “Causal Refutations of Idealism,” Philosophical Quarterly 60.240: 487– 507.



2010b. “Real Repugnance and Belief about Things-in-Themselves: A Problem and Kant’s Three Solutions,” in Kant’s Moral Metaphysics, James Krueger and Benjamin Bruxvoort Lipscomb (eds.). Berlin: Walter De Gruyter, pp. 177– 210. 2011. “Real Repugnance and Our Ignorance of Things-in-Themselves: A Lockean Problem in Kant and Hegel,” Internationales Jahrbuch Des Deutschen Idealismus 7: 135–59. 2014a. “Kant and the ‘Monstrous’ Ground of Possibility,” Kantian Review 19.1: 53–69. 2014b. “Modal Motivations for Noumenal Ignorance: Knowledge, Cognition, and Coherence,” Kant-Studien 105.4: 573–97. 2017. “Kant on Cognition, Givenness, and Ignorance,” Journal of the History of Philosophy 55: 131–42. Chignell, Andrew, and Nicholas Stang. 2015. “Postulate des Empirischen Denkens,” in Georg Mohr, Juergen Stolzenberg, and Marcus Willascheck (eds.), Kant-Lexicon. Berlin: Walter de Gruyter. Cleeremans, Axel (ed.) 2003. The Unity of Consciousness: Binding, Integration and Dissociation. Oxford: Oxford University Press. Coffa, Alberto 1991. The Semantic Tradition from Kant to Carnap: To the Vienna Station. New York: Cambridge University Press. Conant, James 1991. “The Search for Logically Alien Thought: Descartes, Kant, Frege, and the Tractatus,” Philosophical Topics 20.1: 115–80. Dennett, D. C. 1969. Content and Consciousness. New York: Routledge and Kegan Paul. Descartes, René 1985 [CSM]. The Philosophical Writings of Descartes, vol. 1, J. Cottingham, R. Stoothoff, and D. Murdoch (ed. and trans.). Cambridge: Cambridge University Press. 1991. The Philosophical Writings of Descartes, vol. 3, J. Cottingham, R. Stoothoff, D. Murdoch, and A. Kenny (ed. and trans.). Cambridge: Cambridge University Press. DiCenso, James 2011. Kant, Religion, and Politics. Cambridge: Cambridge University Press. Dummett, Michael 1993. “Realism and Anti-realism,” in Dummett (ed.), The Seas of Language. New York: Oxford University Press, pp. 462–78. Dyck, Corey W. 2011. “Turning the Game against the Idealist: Mendelssohn’s Refutation of Idealism in the Morgenstunden and Kant’s Replies,” Studies in German Idealism 13: 159–82. Earman, John 1986. A Primer on Determinism. Dordrecht, Netherlands: Kluwer. Emundts, Dina 2010. “The Refutation of Idealism and the Distinction between Phenomena and Noumena,” in P. Guyer (ed.), The Cambridge Companion to Kant’s ‘Critique of Pure Reason.’ Cambridge: Cambridge University Press, pp. 168–89. Engstrom, Stephen 2006. “Understanding and Sensibility,” Inquiry 49: 2–25. 2013. “Unity of Apperception,” Studi Kantiani 26: 37–54. Erdmann, Benno 1881. Nachträge zu Kants Kritik der reinen Vernunft in 1881. Kiel: Lipsius & Tischer.



Evans, Gareth 1975. “Identity and Predication,” in Evans (ed.), Collected Papers. Oxford: Oxford University Press, pp. 25–48. 1982. The Varieties of Reference, J. McDowell (ed.). Oxford: Clarendon Press. Falkenstein, Lorne 1998. “A Double Edged Sword? Kant’s Refutation of Mendelssohn’s Proof of the Immortality of the Soul and Its Implications for His Theory of Matter,” Studies in History and Philosophy of Science 29.4: 561–88. 2006. “Kant’s Transcendental Aesthetic,” in Graham Bird (ed.), A Companion to Kant. Oxford: Blackwell, pp. 140–53. Förster, Eckart 2012. Twenty-Five Years of Philosophy. Cambridge, MA: Harvard University Press. Friedman, Michael 1992. Kant and the Exact Sciences. Cambridge, MA: Harvard University Press. 2013. Kant’s Construction of Nature. Cambridge: Cambridge University Press. Gardner, Sebastian 1999. Routledge Philosophy Guidebook to Kant and the Critique of Pure Reason. London: Routledge. 2011. “Kant’s Practical Postulates and the Limits of the Critical System,” Bulletin of the Hegel Society of Great Britain 63: 187–215. Gava, Gabriele 2014. “Kant’s Definition of Science in the Architectonic of Pure Reason and the Essential Ends of Reason,” Kant-Studien 105.3: 372–93. 2016. “The Fallibilism of Kant’s Architectonic,” in G. Gava and R. Stern (eds.), Pragmatism, Kant and Transcendental Philosophy. London: Routledge, pp. 46– 66. Forthcoming. “Sind die regulativen Ideen ein doktrinaler Glaube? Über die Rechtfertigung des regulativen Gebrauchs der Ideen im Anhang zur transzendentalen Dialektik” [Are the regulative ideas a doctrinal belief? On the justification of the regulative use of the ideas in the appendix to the transcendental dialectic], in V. L. Waibel and M. Ruffing (eds.), Akten des 12. internationalen Kant-Kongresses, Berlin: De Gruyter. Geiger, Ido 2003. “Is the Assumption of a Systematic Whole of Empirical Concepts a Necessary Condition of Knowledge?” Kant-Studien 94.3: 273–98. Gelfert, Axel 2006. “Kant on Testimony,” British Journal for the History of Philosophy 14.4: 627–52. Gettier, Edmund 1963. “Is Justified True Belief Knowledge?” Analysis 23: 121–23. Gill, Jerry H. 1984. “Kant, Analogy, and Natural Theology,” International Journal for Philosophy of Religion 16.1: 19–28. Ginsborg, Hannah 2006. “Empirical Concepts and the Content of Experience,” European Journal of Philosophy 14.3: 349–72. 2013. “Kant’s Perceiver,” Philosophy and Phenomenological Research 87.1: 221–28. Gloy, Karen (ed.) 1990. “Die Differenz von Verstand und Sinnlichkeit,” in Studien zur theoretischen Philosophie Kants. Würzburg, pp. 1–41. Gochnauer, Myron 1974. “Kant’s Refutation of Idealism,” Journal of the History of Philosophy 12: 195–206. Godlove, Terry 2014. Kant and the Meaning of Religion. New York: Columbia University Press. Greenberg, Robert 2001. Kant’s Theory of A Priori Knowledge. University Park: Pennsylvania State University Press.



Grier, Michelle 2001. Kant’s Doctrine of Transcendental Illusion. Cambridge: Cambridge University Press. 2010. “The Ideal of Pure Reason,” in Paul Guyer (ed.), The Cambridge Companion to Kant’s Critique of Pure Reason. Cambridge: Cambridge University Press, pp. 266–89. Griffith, Aaron 2010. “Perception and the Categories: A Conceptualist Reading of Kant’s Critique of Pure Reason,” European Journal of Philosophy 20.2: 193–222. Grüne, Stefanie 2017. “Givenness, Objective Reality, and A Priori Intuitions,” Journal of the History of Philosophy 55.1: 113–30. Guyer, Paul 1983. “Kant’s Intentions in the Refutation of Idealism,” The Philosophical Review 92.3: 329–83. 1987. Kant and the Claims of Knowledge. Cambridge: Cambridge University Press. 1989. “Psychology and the Transcendental Deduction,” in E. Förster (ed.), Kant’s Transcendental Deductions. Stanford, CA: Stanford University Press, pp. 47–68. 2006. Kant. London: Routledge. 2007. “Debating Allison on Transcendental Idealism,” Kantian Review 12.2: 10– 24. Haag, Johannes 2007. Erfahrung und Gegenstand. Frankfurt am Main, Germany: Klostermann. Hanna, Robert 2001. Kant and the Foundations of Analytic Philosophy. Oxford: Clarendon Press. 2006. Kant, Science and Human Nature. Oxford: Clarendon Press. Harper, William 2007. “Comments on Westphal,” Dialogue: Canadian Journal of Philosophy/Revue canadienne de philosophie 46.4: 729–36. 2011. Isaac Newton’s Scientific Method: Turning Data into Evidence about Gravity and Cosmology. New York: Oxford University Press. Hatfield, Gary 2006. “Kant on the Perception of Space (and Time),” in Paul Guyer (ed.), The Cambridge Companion to Kant and Modern Philosophy. Cambridge: Cambridge University Press, pp. 61–93. Heidemann, Dietmar 2002. “Anschauung und Begriff. Ein Begründungsversuch des Stämme-Dualisms in Kants Erkenntnistheorie,” in K. Engelhard (ed.), Aufklärungen. Festschrift für Klaus Düsing zum 60. Geburtstag. Berlin: de Gruyter, pp. 65–90. Heine, Heinrich 1986. Religion and Philosophy in Germany: A Fragment. John Snodgrass (trans.). New York: State University of New York Press. Henrich, Dieter 1969. “The Proof-Structure of Kant’s Transcendental Deduction,” Review of Metaphysics 22.4: 640–59. Hillman, T. Allan 2010. “Leibniz on the Imago Dei,” in Daniel Garber and Steven M. Nadler (eds.), Oxford Studies in Early Modern Philosophy, vol. V. Oxford: Oxford University Press, pp. 171–213. Hinske, Norbert 1998. Zwischen Aufklarung und Vernunftkritik. Stuttgart: frommann-holzboog. Hume, David 1739/1978. A Treatise of Human Nature. P. H. Nidditch (ed.) (revision of L. A. Selby-Bigge). 2nd ed. Oxford: Oxford University Press.



2000. A Treatise of Human Nature. D. F. Norton and M. J. Norton (eds.). Oxford: Oxford University Press (designated ‘T,’ cited by Book.Part.§.¶ numbers). 2007. Dialogues Concerning Natural Religion and Other Writings. Dorothy Coleman (ed.). Cambridge: Cambridge University Press. James, William 1902. The Varieties of Religious Experience. London: Longmans Green. Janiak, A., and E. Schliesser 2012. Interpreting Newton: Critical Essays. Cambridge: Cambridge University Press. Kannisto, Toni 2013. “Modality and Metaphysics in Kant,” in Stefano Bacin, Alfredo Ferrarin, Claudio La Rocca, and Margit Ruffing (eds.), Kant Und Die Philosophie in Weltbürgerlicher Absicht. Akten Des XI. Kant-Kongresses 2010. Berlin: Walter de Gruyter, pp. 633–46. Kant, Immanuel 1900–. Kants gesammelte Schriften. Berlin: Reimer/de Gruyter. 1987. “Vom inneren Sinne (Loses Blatt Leningrad 1; Leningrad Fragment I, NF 364–6),” in R. Brandt and W. Stark (eds.), Neue Autographen und Dokumente zu Kants Leben, Schriften und Vorlesungen. Berlin: Felix Meiner, pp. 18– 21. Kemp Smith, Norman 1923/1962. A Commentary to Kant’s “Critique of Pure Reason.” New York: Humanities Press. 1929. Immanuel Kant’s Critique of Pure Reason. London: Macmillan. Kepler, Johannes 1997 [1619]. The Harmony of the World. E. J. Aiton, A. M. Duncan and J. V. Field (eds., trans.). Memoirs of the American Philosophical Society 209. Philadelphia: American Philosophical Society. Kitcher, Patricia 2011. Kant’s Thinker. Oxford: Oxford University Press. Kneale, William, and Martha Kneale 1975. The Development of Logic. Oxford: Clarendon Press. Kuehn, Manfred 2001. Kant: A Biography. Cambridge: Cambridge University Press. Lambert, Johann Heinrich 1764. Neues Organon. 2 vols. Leipzig: Wendler. Land, Thomas 2006. “Kant’s Spontaneity Thesis,” Philosophical Topics: Analytic Kantianism 34: 189–220. Langton, Rae 1998. Kantian Humility: Our Ignorance of Things in Themselves. Oxford: Clarendon Press. Laplace, Pierre Simon de 1847 [1820]. Theorie analylique des probabilites. Paris: V. Courcier. In Oevres de LaPlace, vol. 7. Paris: Imprimerie Royale, 1847. Laywine, Alison 1993. Kant’s Early Metaphysics and the Origins of the Critical Philosophy. North American Kant Society Studies in Philosophy. Atrascadero, CA: Ridgeview. Leech, Jessica 2014. “Making Modal Distinctions: Kant on the Possible, the Actual, and the Intuitive Understanding,” Kantian Review 19.3: 339–65. Leibniz, Gottfried Wilhelm 1989. Philosophical Essays. Roger Ariew and Daniel Garber (eds.). Indianapolis, IN: Hackett. 1997. New Essays on Human Understanding. Peter Remnant and Jonathan Bennett (eds.). Cambridge: Cambridge University Press.



Lewis, Clarence Irving 1929 [1956]. Mind and the World Order: Outline of a Theory of Knowledge. New York: Dover. 1970. Collected Papers of Clarence Irving Lewis. J. D. Goheen and J. L. Mothershead Jr. (eds.). Palo Alto, CA: Stanford University Press. Lighthill, Sir James 1986. “The Recently Recognized Failure of Predictability in Newtonian Dynamics,” Proceedings of the Royal Society of London, Series A 407: 33–55. Locke, John 1975. An Essay Concerning Human Understanding. P. H. Nidditch (ed.). Oxford: Clarendon Press. Logan, Beryl 1998. “Hume and Kant on Knowing the Deity,” International Journal for Philosophy of Religion 43.3: 133–48. Longuenesse, Béatrice 1998/2001. Kant and the Capacity to Judge: Sensibility and Discursivity in the Transcendental Analytic of the Critique of Pure Reason. Princeton, NJ: Princeton University Press. 2008. “Kant’s ‘I Think’ versus Descartes’ ‘I Am a Thing That Thinks,’” in Kant and the Early Moderns, Daniel Garber and Béatrice Longuenesse (eds.). Princeton, NJ: Princeton University Press, pp. 9–31. 2012. “Two Uses of ‘I’ as Subject?” in Immunity to Error through Misidentification, Simon Prosser and François Recanati (eds.). Cambridge: Cambridge University Press, pp. 81–103. Makkreel, Rudolf A. 2003. “The Cognition–Knowledge Distinction in Kant and Dilthey and the Implications for Psychology and Self-Understanding,” Studies in History and Philosophy of Science Part A 34.1: 149–64. Malebranche, Nicolas 1997. Dialogues on Metaphysics and on Religion. Nicholas Jolley and David Scott (eds.). Cambridge: Cambridge University Press. Massimi, Michela 2014. “Prescribing Laws to Nature,” Kant-Studien 105.4: 491– 508. 2017. “The Legacy of Newton for the Pre-Critical Kant,” in The Oxford Handbook on Newton, E. Schliesser and C. Smeenk (eds.). Oxford: Oxford University Press. McDowell, John 1994. Mind and World. Cambridge, MA: Harvard University Press. 1998. “Having the World in View: Sellars, Kant, and Intentionality,” Journal of Philosophy 95.9: 431–92. 2009a. Having the World in View: Essays on Kant, Hegel, and Sellars. Cambridge, MA: Harvard University Press. 2009b. “Hegel’s Idealism as Radicalization of Kant,” in Having the World in View: Essays on Kant, Hegel, and Sellars. Cambridge, MA: Harvard University Press, pp. 69–89. 2009c. “Avoiding the Myth of the Given,” in Having the World in View: Essays on Kant, Hegel, and Sellars. Cambridge, MA: Harvard University Press, pp. 256–72. 2015. “Putnam on Natural Realism,” in R. E. Auxier, D. R. Anderson, and L. E. Hahn (eds.), The Philosophy of Hilary Putnam. Chicago: Open Court, pp. 643–58.



Meier, George Friedrich 1752. Auszug aus der Vernuftlehre. Halle, Germany: Gebauer. Melnick, Arthur 1989. Space, Time, and Thought in Kant. Dordrecht, Netherlands: Kluwer. Mendelssohn, Moses 1997. Moses Mendelssohn: Philosophical Writings. Daniel O. Dahlstrom (trans.). Cambridge: Cambridge University Press. 2011. Morning Hours, or Lectures on God’s Existence. Daniel Dahlstrom and Corey W. Dyck (trans.). Berlin: Springer. Milmed, Bella K. 1969. “‘Possible Experience’ and Recent Interpretations of Kant,” in L. W. Beck (ed.), Kant Studies Today. LaSalle, IL: Open Court, pp. 301–21. Montaigne, Michel de 2003 [1580]. “An Apology for Raymond Sebond,” in Essays, M. A. Screech (trans.). London: Penguin Books, pp. 489–683. Moore, A. W. 2010. “The Transcendental Doctrine of Method,” in Paul Guyer (ed.), Cambridge Companion to Kant’s Critique of Pure Reason. Cambridge: Cambridge University Press, pp. 310–26. 2011. “Bird on Kant’s Mathematical Antinomies,” Kantian Review 16.2: 235–43. Nagel, Ernest 1961. The Structure of Science. New York: Harcourt, Brace, and World. Newton, Isaac 1999. The Principia: Mathematical Principles of Natural Philosophy. Bernard Cohen, Anne Whitman, and Julia Budenz (trans.). Berkeley: University of California Press. 2004. “De Gravitatione,” in A. Janiak (ed.), Philosophical Writings. Cambridge: Cambridge University Press, pp. 12–39. Onof, Christian, and Schulting, Dennis 2015. “Space as Form of Intuition and as Formal Intuition: On the Note to B160 in Kant’s Critique of Pure Reason,” The Philosophical Review 124.1: 1–58. O’Shea, James R. 1997. “The Needs of Understanding: Kant on Empirical Laws and Regulative Ideals,” International Journal of Philosophical Studies 5.2: 216– 54. 2014. Kant’s Critique of Pure Reason: An Introduction and Interpretation. London: Routledge. First published by Acumen, 2012. Parsons, Charles 1992. “The Transcendental Aesthetic,” in Paul Guyer (ed.), The Cambridge Companion to Kant. Cambridge: Cambridge University Press, pp. 62–99. Pasternack, Lawrence 2010. “Kant’s Doctrinal Belief in God,” in Oliver Thorndike (ed.), Rethinking Kant, vol. 3. Newcastle upon Tyne, UK: Cambridge Scholars Press, pp. 200–18. 2011. “Kant and Theology at the Boundaries of Reason,” Kantian Review 16.3: 495–98. 2014a. “Kant on Opinion: Assent, Hypothesis, and the Norms of General Applied Logic,” Kant-Studien 105.1: 41–82. 2014b. Kant’s Religion within the Boundaries of Mere Reason: An Interpretation and Defense. London: Routledge. Paton, H. J. 1936. Kant’s Metaphysic of Experience. London: George Allen and Unwin.



Piché, David 1999. La Condamnation parisienne de 1277. Paris: Vrin. Pickering, Mark 2016. “Kant’s Theoretical Reasons for Belief in Things in Themselves,” Kant-Studien 107.4: 589–616. Pippin, Robert B. 2005. The Persistence of Subjectivity: On the Kantian Aftermath. Cambridge: Cambridge University Press. Plantinga, Alvin 1966. “Kant’s Objection to the Ontological Argument,” Journal of Philosophy 63.19: 537–46. Prauss, Gerold 1969. “Zum Warheitsproblem bei Kant,” Kant-Studien 60: 166–82. 1971. Erscheinung bei Kant. Berlin: Walter de Gruyter. 1974. Kant und das Problem der Dinge an sich. Bonn: Bouvier Verlag Herbert Grundmann. Proops, Ian 2014. “Kant on the Cosmological Argument,” Philosophers’ Imprint 14.12: 1–21. 2015. “Kant on the Ontological Argument,” Noûs 49.1: 1–27. Reich, Klaus 2001. Gesammelte Schriften. With Introduction and annotations. M. Baum, U. Rameil, K. Reisinger, and G. Scholz (eds.). Hamburg, Germany: Meiner. Ricoeur, Paul 1970. Freud and Philosophy: An Essay on Interpretation. Denis Savage (trans.). New Haven, CT: Yale University Press. Rosenberg, Jay F. 2005. Accessing Kant: A Relaxed Introduction to the Critique of Pure Reason. Oxford: Clarendon. Rosenkoetter, Timothy 2010. “Absolute Positing, the Frege Anticipation Thesis, and Kant’s Definitions of Judgment,” European Journal of Philosophy 18: 539– 66. Schaffer, Karl Forthcoming. “Kant’s Conception of Cognition and Our Knowledge of Things-in-Themselves,” in Nick Stang (ed.), The Sensible and Intelligible Worlds: New Essays on Kant’s Metaphysics and Epistemology. Oxford: Oxford University Press. Schliesser, Eric 2013. “On Reading Newton as an Epicurean: Kant, Spinozism and the Changes to the Principia,” Studies in History and Philosophy of Science 44: 416–28. Senderowicz, Yaron M. 2005. The Coherence of Kant’s Transcendental Idealism. Dordrecht, Netherlands: Springer. Setiya, Kieran 2004. “Transcendental Idealism in the Aesthetic,” Philosophy and Phenomenological Research 68.1: 63–88. Shabel, Lisa 2003. Mathematics in Kant’s Critical Philosophy: Reflections on Mathematical Practice. London: Routledge. 2010. “The Transcendental Aesthetic,” in Paul Guyer (ed.), The Cambridge Companion to Kant’s Critique of Pure Reason. Cambridge: Cambridge University Press, pp. 93–117. Shaffer, Jerome 1962. “Existence, Predication, and the Ontological Argument,” Mind 71.283: 307–25. Shoemaker, Sydney 1968. “Self-Reference and Self-Awareness,” Journal of Philosophy 65: 555–67. 1970. “Persons and Their Pasts,” American Philosophical Quarterly 7.4: 269–85.



Smith, Donald P. 2003. “Kant on the Dependency of the Cosmological Argument on the Ontological Argument,” European Journal of Philosophy 11.2: 206– 18. Stang, Nicholas 2016. Kant’s Modal Metaphysics. New York: Oxford University Press. Stevenson, Leslie 2003. “Opinion, Belief or Faith, and Knowledge,” Kantian Review 7.1: 72–101. Strawson, P. F. 1966. The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason. London: Methuen. 1974. “Imagination and Perception,” in Freedom and Resentment and Other Essays. London: Methuen, pp. 45–65. 1979. “Perception and Its Objects,” in G. F. MacDonald (ed.), Perception and Identity. Ithaca, NY: Cornell University Press, pp. 41–60. 1989. “Sensibility, Understanding, and the Doctrine of Synthesis: Comments on Henrich and Guyer,” in E. Förster (ed.), Kant’s Transcendental Deductions. Palo Alto, CA: Stanford University Press, pp. 69–77. 1997a. “Kant’s New Foundations of Metaphysics,” in Entity and Identity and Other Essays. Oxford: Oxford University Press, pp. 233–44. 1997b. “The Problem of Realism and the A Priori,” in Entity and Identity and Other Essays. Oxford: Oxford University Press, pp. 246–51. Stroud, Barry 1994. “Kantian Argument, Conceptual Capacities, and Invulnerability,” in P. Parrini (ed.), Kant and Contemporary Epistemology. Dordrecht, Netherlands: Kluwer, pp. 231–51. Sutherland, Daniel 2005. “The Point of the Axioms of Intuition,” Pacific Philosophical Quarterly 86.1: 135–59. Swinburne, Richard 1996. “The Beginning of the Universe,” Aristotelian Society XL(Suppl.): 139–50. Tempier, Étienne 1277. Opiniones ducentae undeviginti Sigeri de Brabantia, Boetii de Dacia aliorumque, a Stephano episcopo Parisiensi de consilio doctorum Sacrae Scripturae condemnatae 1277. Paris. http://hiphi.ubbcluj.ro/fam/texte/ tempier_opiniones_219.htm. 1999. Critical edition of Tempier (1277), in David Piché, La Condamnation parisienne de 1277. Paris: Vrin, pp. 17–62. Tolley, Clinton 2013. “The Non-Conceptuality of the Content of Intuitions: A New Approach,” Kantian Review 18: 107–36. Treisman, Anne 2003. “Consciousness and Perceptual Binding,” in Axel Cleeremans (ed.), The Unity of Consciousness: Binding, Integration and Dissociation. Oxford: Oxford University Press, pp. 95–113. Van Cleve, James 1999. Problems from Kant. Oxford: Oxford University Press. Vogel, Jonathan 1993. “The Problem of Self-Knowledge in Kant’s ‘Refutation of Idealism’: Two Recent Views,” Philosophy and Phenomenological Research 53.4: 875–87. Walker, Ralph C. S. 1978. Kant. London: Routledge and Kegan Paul. 1995. “Verificationism, Anti-Realism, and Idealism,” European Journal of Philosophy, 3.3: 257–72.



Warren, Daniel 1998. “Kant and the A Priority of Space,” Philosophical Review 107: 179–224. Watkins, Eric 2000. “Kant on Rational Cosmology,” in E. Watkins (ed.), Kant and the Sciences. New York: Oxford University Press, pp. 70–89. 2005. Kant and the Metaphysics of Causality. Cambridge: Cambridge University Press. 2006. “On the Necessity and Nature of Simples: Leibniz, Wolff, Baumgarten, and the Pre-Critical Kant,” in D. Garber (ed.), Oxford Studies in Early Modern Philosophy, vol. 3. Oxford: Oxford University Press, pp. 261–314. (ed., trans.) 2009. Kant’s “Critique of Pure Reason”: Background Source Materials. Cambridge: Cambridge University Press. 2013. “Kant on the Natural, Moral, Human, and Divine Orders,” in The Divine Order, the Human Order, and the Order of Nature. New York: Oxford University Press, pp. 219–36. Forthcoming. “Breaking with Rationalism: Kant, Crusius, and the Priority of Existence,” in B. Look (ed.), Leibniz and Kant. Oxford: Oxford University Press. Watkins, Eric, and Marcus Willaschek 2017. “Kant’s Account of Cognition,” Journal of the History of Philosophy 55: 83–112. Watson, John 1881. Kant and His English Critics: A Comparison of Critical and Empirical Philosophy. New York: Macmillan. Weizsäcker, C. F., von 1971. “Kant’s ‘First Analogy of Experience’ and Conservation Principles of Physics,” Synthese 23: 75–95. Westphal, Kenneth R. 2004. Kant’s Transcendental Proof of Realism. Cambridge: Cambridge University Press. 2005. “Kant, Wittgenstein and Transcendental Chaos,” Philosophical Investigations 28.4: 303–23. 2007a. “Kant’s Anti-Cartesianism,” Dialogue: Canadian Journal of Philosophy/Revue canadienne de philosophie 46.4: 709–15. 2007b. “Proving Realism Transcendentally: Replies to Rolf George and William Harper,” Dialogue: Canadian Journal of Philosophy/Revue canadienne de philosophie 46.4: 737–50. 2007c. “Consciousness and Its Transcendental Conditions: Kant’s AntiCartesian Revolt,” in S. Heinämaa, V. Lähteenmäki, and P. Remes (eds.), Consciousness: From Perception to Reflection in the History of Philosophy. Dordrecht, Netherlands: Springer, pp. 223–43. 2008. “Contemporary Epistemology: Kant, Hegel, McDowell,” in J. Lindgaard (ed.), John McDowell: Experience, Norm and Nature. Oxford: Blackwell, pp. 124–51. 2010. “Kant’s Critique of Pure Reason and Analytic Philosophy,” in P. Guyer (ed.), The Cambridge Companion to Kant’s Critique of Pure Reason. Cambridge: Cambridge University Press, pp. 401–30. 2013a. “Hume, Empiricism and the Generality of Thought,” Dialogue: Canadian Journal of Philosophy/Revue canadienne de philosophie 52.2: 233–70.



2013b. “Kant’s Cognitive Semantics, Newton’s Rule Four of Natural Philosophy and Scientific Realism Today,” in Kant and Contemporary Theory of Knowledge, Kant Yearbook 5:127–68. 2016. “Mind, Language and Behaviour: Kant’s Critical Cautions contra Contemporary Internal and Causal Naturalism,” in S. Babür (ed.), Felsefede Yöntem/Method in Philosophy, special issue of Yeditepe’de Felsefe/Philosophy at Yeditepe, vol. 10. Istanbul: Yedite Üniversi Press, pp. 102–49. 2017. “Kant’s Analytic of Principles,” in M. Timmons and S. Baiasu (eds.), Kant. London: Routledge. Wittgenstein, Ludwig 1958. The Blue and the Brown Books. New York: Harper. Wolff, Christian 1740. Philosophia rationalis sive logica, Pars I–III. Frankfurt und Leipzig. Reprinted in Gesammelte Werke (Hildesheim: Olms, 1983), 2. Abt., vols. 1.1, 1.2, 1.3. Wolff, Michael 1995. Die Vollständigkeit der kantischen Urteilstafel. Mit einem Essay über Freges “Begriffsschrift.” Frankfurt am Main, Germany: Klostermann. 2000. “Kantische Urteilstafel und vollständige Induktion: Nachtrag zu meiner Kontroverse mit Ulrich Nortmann.” Zeitschrift für Philosophische Forschung 54.1: 86–94. 2009a. Abhandlung über die Prinzipien der Logik. Mit einer Rekonstruktion der aristotelischen Syllogistik. 2nd rev. ed. Frankfurt am Main, Germany: Klostermann. 2009b. “Vollkommene Syllogismen und reine Vernunftschlüsse: Aristoteles und Kant,” Journal for General Philosophy of Science 40: 341–55. 2010a. “Vollkommene Syllogismen und reine Vernunftschlüsse: Aristoteles und Kant. Eine Stellungnahme zu Theodor Eberts Gegeneinwänden,” Part 1. Journal for General Philosophy of Science 41: 199–213. 2010b. “Vollkommene Syllogismen und reine Vernunftschlüsse: Aristoteles und Kant. Eine Stellungnahme zu Theodor Eberts Gegeneinwänden,” Part 2. Journal for General Philosophy of Science 41: 359–71. 2013. “Viele Logiken – Eine Vernunft. Warum der logische Pluralismus ein Irrtum ist,” Methodus: International Journal for Modern Philosophy 7: 73–128. Forthcoming. “Kant’s Table of Judgments: Frege’s Critique and Kant’s Rejoinder,” K. R. Westphal (trans.), Dialogue: Canadian Journal of Philosophy/Revue canadienne de philosophie. Wolterstorff, Nicholas 1991. “Conundrums in Kant’s Rational Religion,” in P. J. Rossi and M. J. Wreen (eds.), Kant’s Philosophy of Religion Reconsidered. Bloomington: Indiana University Press, pp. 40–53. 1998. “Is It Possible and Desirable for Theologians to Recover from Kant?” Modern Theology 14.1: 1–18. Wood, Allen W. 1970. Kant’s Moral Religion. Ithaca, NY: Cornell University Press. 1978. Kant’s Rational Theology. Ithaca, NY: Cornell University Press. 1992. “Rational Theology, Moral Faith, and Religion,” in Paul Guyer (ed.), Cambridge Companion to Kant. Cambridge: Cambridge University Press, pp. 394–416.



2005. Kant. Oxford: Blackwell. 2010. “The Antinomies of Pure Reason,” in Paul Guyer (ed.), The Cambridge Companion to Kant’s Critique of Pure Reason. Cambridge: Cambridge University Press, pp. 245–65. Wright, Crispin 1992. Truth and Objectivity. Cambridge, MA: Harvard University Press. 1996. Realism, Meaning and Truth. Cambridge: Wiley-Blackwell/Polity. Yong, Peter 2014. “God, Totality and Possibility in Kant’s Only Possible Argument,” Kantian Review 19.1: 27–51.


a posteriori knowledge, 31, 113, 114–18, 141, 181 a priori, 83, 84, 103, 173, 271 cognition, 20, 22–23, 167, 172 concepts, 50, 107, 109, 112–14, 116, 117, 118, 119, 155, 164, 172, 184, 187, 188, 192 intuition, 46–58, 60, 61, 62, 164, 166, 169, 178, 181–82, 269 knowledge, 28, 30–31, 32, 35, 37, 42, 107, 109, 111, 112–14, 116–19, 141, 163, 177, 224 representation, 31, 66, 175, 181, 227 synthetic, 48, 54–57, 63, 163, 164, 165, 166–67, 168, 170, 177, 182, 189, 193 truth, 165 absolute space, 64–68, 70–74, 75, 76, 77, 78, 80, 81 Adickes, Erich, 267, 268 Aesthetic, Transcendental, 21, 22, 46–63, 64–66, 121, 125, 133–34, 135, 139, 164, 165–67, 171, 173, 179, 183, 190, 209, 259 affection (external/sensory), 18, 19, 24, 37, 39, 43–44, 66 Allison, Henry, 128, 132, 143, 147, 210 Ameriks, Karl, 48, 161 Amphiboly, 190, 228, 231, 266 Analogies of Experience, 188, 197, 203, 254, 270 Analytic of Principles, 109–11, 173, 188, 190 Analytic, Transcendental, 121, 125, 135–37, 164–66, 173, 183, 188, 227, 241, 259 analytic/synthetic distinction, 23, 107 Anticipations of Perception, 163, 164–66, 167, 173, 174–75, 178, 180, 181, 182, 183 Antinomies, 23, 32, 53, 223–27, 228, 230, 231, 233–39, 241 anti-realism, 49, 60–61, 62, 63 appearances, 2, 9–10, 20, 48, 51, 58, 60, 66, 172, 175–79, 181, 183, 195, 199, 206, 213, 214, 215, 224, 226, 231, 250, 264, 267, 273, 275 apperception, 137, 140, 145, 148, 152, 198, 201, 203, 208, 240–41 transcendental unity of, 143, 145, 147–55, 160–62

apprehension, 130, 139, 175–77 order of, 198–200 Aquinas, Thomas, 29, 34–35 Aristotle, 33–35, 40, 85, 97, 98, 100, 101, 187, 193 Arnauld, Antoine, 247, 248–49 Baumgarten, Alexander Gottlieb, 12, 65, 73–74, 75, 76, 78, 81, 246 Beck, Lewis White, 130–31, 132 Bennett, Jonathan, 233 Berkeley, George, 48–49, 60, 61 categories, 47, 108, 109, 118, 121, 124–25, 126, 128, 131, 132, 134–38, 148–49, 155, 157–58, 161, 163, 164, 165, 166, 174–75, 187, 190, 222, 249–50, 257, 270, 271 schematized, 164–65, 177, 213, 216, 217, 250 table of, 83, 84–85, 164, 190 unschematized, 250 causality, 69, 72, 82, 194–97, 222, 224, 234–38, 257, 272, see also judgments, causal Coffa, Alberto, 193 cognitive semantics, 184, 192–93, 196, 201, 203–04 concepts, 10, 13, 16, 18, 19–21, 23, 26, 31, 38, 43, 45, 46, 47, 49–51, 52, 54, 85–94, 99, 100, 102, 103, 104, 106, 108, 112, 117, 119, 125, 127, 128, 129, 136, 140, 142, 145, 148, 152–53, 155, 157, 163, 164–65, 173, 175, 181, 184, 188–92, 200, 232, 239, 241, 250, 253, 257, 260, 263, 264, 267, 268, 269–70, 273 determinable, 184 mathematical, 169–71, 172, 174 of God, 11, 13, 72, 243, 245–53, 255, 258, 260, 263 pure, 83–84, 113, 114, 116, 133–34, 169, 174, 182, 228 without intuitions, 128, 139, 264 constructibility, 169 Copernican turn, 28, 30, 31, 33, 35, 39–40, 43, 44, 245




cosmology, 74, 231, 233, 241, 251, 256 Crusius, Christian August, 11–13, 26, 73, 75, 78, 79, 81 deduction. See Metaphysical Deduction, Transcendental Deduction Dennett, Daniel, 158 Descartes, René, 12, 30, 51, 59, 74, 79–81, 114–15, 145, 160, 184, 187, 188, 201, 209, 221, 243, 247, 249, 260, 265, 277 determinism, 69, 202–03, 235, 236–39, 241 Dialectic. See Transcendental Dialectic dialectical illusions, 224, 227 Doctrine of Method, 47, 54, 181, 259–61, 262, 279 duration, 67, 181–82, 207, 214, 215, 222 empirical deduction, 106 empirical realism, 30, 31, 58–59, 115–18 empiricism, 10, 14, 27, 30, 126, 138–39, 185–87, 192, 247, 261, 265, 269 ensrealissimum, 251, 277 essence, 11–12, 75, 238, 247, 257, 266 Euclid, 56, 168 Evans, Gareth, 146, 189, 192 existence as not a predicate, 12, 13, 26 of God, 67, 69–70, 243–46, 266, 276–78 of noumena, 69, 240, 242, 267, 268 of objects, 10, 15, 25–27, 40, 43, 56, 69, 76, 114–15, 163, 181, 185–86, 205–07, 213, 217, 219–20, 252, 272, 273–75 of substance, 72–73, 76, 78, 80, 82, 213, 215, 221 of the self, 197, 206–12, 214, 217, 218, 219, 221, 222, 241 Principle of Concurrent, 196 experience, 10, 28, 31, 32, 33–34, 35, 59, 60, 67, 106–07, 110, 115, 126, 135, 144, 145, 185, 187–88, 190, 197, 200, 202, 209, 224, 231–33, 239, 249, 260–61, 273, see also Analogies of Experience conditions of possibility of, 55, 56, 60, 109, 110–12, 114, 118, 119, 123, 133, 152, 165, 174, 216, 269, 270–72, 273 given element in, 127 inner as presupposing outer, 206, 208–13, 219–22 judgements of, 39 of objects, 28, 31, 50, 61, 63, 108–09, 113–18, 145, 148, 163, 174, 177, 179, 181, 185, 190, 197–98, 205, 250, 271 two concepts of, 130 unity of, 234, 235

Falkenstein, Lorne, 9, 47, 218 fallibilism/infallibilism, 6, 193–94, 201 finitude, 122, 123, 124, 256 force, 76, 191, 204, 214, 219, 266, 274 freedom, 20, 69, 70–71, 72, 80–81, 82, 203, 223, 226, 234–39, 241, 260 Frege, Gottlob, 27, 83, 187 Friedman, Michael, 64, 214, 218, 276 Gettier, Edmund, 187–88 God. See Concepts of God, Existence of God Guyer, Paul, 48, 56, 123, 155, 166, 188, 194, 196, 213, 215 Hegel, G.W.F., 137 Heine, Heinrich, 244, 245, 258 Henrich, Dieter, 134, 135 Herz, Marcus, 30, 83, 161 Hume, David, 130, 151, 152, 155, 184–87, 194, 196, 197–98, 200, 239, 246, 253, 256, 265 Ideal of Pure Reason, 243, 246, 252 idealism, 49, 51, 55–56, 62, 65, 74, 114, 118, 127, 211 about space, 53, 66, 68–74, 81, 82, 212 metaphysical, 49 moderate, 58 phenomenalist, 48, 58, 60 problematic, 114–17, 205–06, 208–10, 217, 221, 222 psychological, 205 Refutation of, 205–22 Transcendental, 9, 28, 30, 43, 46, 48, 58, 60, 117–18, 184, 202, 206 ideas cosmological, 231 empirical, 10, 61, 106 Humean, 184, 256 innate, 10 new way of, 186 of God, 278 of Reason, 59, 64, 217, 230, 233, 241, 245, 250, 251, 252 identity of the Self, 142, 145–54, 156, 157 illusion. See Dialectical Illusion imagination, 6, 25, 43, 115, 148, 152, 169–70, 173, 176, 185, 187, 188, 206, 208, 211, 220, 221, 249 immortality of the soul, 20, 217, 278 of the soul, 143 inner sense, 114, 197, 206, 209–10, 214, 219, 221, 222, 241

Index intuitions, 9, 10, 15, 18, 19, 23, 25, 28, 31, 43, 45, 46–47, 49–52, 54–55, 57, 59, 61–63, 66, 88–90, 95, 125–26, 128–39, 141, 144, 149, 151, 155, 157, 160, 169, 172, 175–77, 179, 206, 220, 250, 259, 262–65, 269, see also A priori intuition Axioms of, 163, 164, 167, 173, 175–78, 181, 183 forms of, 9, 52–53, 56, 58, 66, 121, 136, 164, 167, 170–72, 182, 209 intellectual, 19, 240, 255 manifold of, 43, 44, 135, 137, 148, 149 original, 42 pure, 17–18, 19, 21, 23–24, 54, 55, 135, 165, 169, 171, 172–74, 177, 182, 269 two conceptions of, 131 judgments affirmative, 97–98 causal, 188, 194–97, 201–03 infinite, 97–99 negative, 98 table of, 83–85, 87, 94, 95–100, 104, 105, 187, 190 Kemp Smith, Norman, 33, 151 Langton, Rae, 49, 66 Laplace, Pierre-Simon, 202 laws empirical, 17 moral, 69, 72 natural, 16, 72, 76, 78, 79, 234, 238, 245, 271 universal, 79 Leibniz, Gottfried Wilhelm, 10, 11, 12, 22, 24, 26–27, 51, 59, 65, 67, 69, 73–75, 79–80, 81–82, 190, 194, 228, 247, 256, 257, 266 Lewis, Clarence Irving, 126–27 Locke, John, 10, 106, 130, 156, 157, 256, 269 logic general, 85, 91, 98, 187, 188, 227–28 Port Royal, 248 Transcendental, 21, 188, 227, 233 Longuenesse, Béatrice, 143, 147–48, 149–51, 153–54, 160 Malebranche, Nicolas, 65, 73, 79, 81, 256 mathematics, 5, 24, 32, 62–63, 129–30, 163–83, 225 matter, 41, 42, 71, 75, 82, 202–03, 217–19, 266, 271 ghostly, 270 magnetic, 275 of intuitions, 25, 28, 43, 265 of judgments, 103 vs form, 14, 15, 21, 40–45


McDowell, John, 120, 133, 137, 192 Melnick, Arthur, 193 Mendelssohn, Moses, 70, 209, 214, 218, 243–44, 246, 287 Metaphysical Deduction, 4, 20, 21, 43 Metaphysical Exposition, 51, 52–53, 62 metaphysics, 13, 18, 20, 22, 29, 32–34, 36, 48, 63, 107, 114, 189, 193, 205, 243, 260–61, 277, 279 Montaigne, Michel de, 246, 247–48, 256 Moore, Adrian, 265 morality, 235, 236, 240, 241, 277 nature, 28, 31, 39, 116, 117, 203, 214, 237, 252, 269, 274 causality in, 72, 76, 203, 236 God’s omnipresence in, 72, 76 laws of. See laws, natural mathematical investigators of, 66, 67, 68–69 metaphysicians of, 67 unity of, 79, 257, 276 necessity, 74, 103, 123 concept of, 20, 39, 257 empirical, 272 natural, 69, 72, 235, 237 of a priori knowledge, 35, 39, 51, 107, 111–14, 170 of consciousness, 144, 147, 152 of space, 66 of the categories, 109 subjective, 278 Newton, Isaac, 64–82, 180, 181, 202, 204 Nicole, Pierre, 247, 248 noumena, 82, 224, 225, 228, 230–42, 266, see also existence of noumena; things in themselves noumenal world, 80–82, 238 O’Shea, James R., 62 object representation of the, 25–26 object, 11, 17, 25–27, 28, 51, 56, 59–63, 66, 88, 89, 92, 98, 110, 112, 131, 175, 181, 183, 186, 191, 193, 200, 214, 251, 257, 267, 270, 272, see also existence of objects concept of, 38, 42, 155, 183, 200 divine, 252 external, 10 givenness of the, 10, 18, 19, 21, 22–24, 25–26, 35, 37, 43, 47, 50, 51–52, 54, 57, 61, 128, 133, 136, 137, 164, 250, 261, 263, 265, 276 inner and outer, 214, 216–17, 220, 221 knowledge of, 28–45, 48, 54, 56, 88, 130, 224 knowledge of, 116–17 mathematical, 19, 25



object (cont.) representation of the, 12, 15, 19, 29, 30, 38, 47, 49, 52–53, 58, 125, 128, 130 undetermined, 92, 101 objective validity, 31, 39, 133, 134, 172 objectivity, 124 occasionalism, 74, 79–81, 196–97 Paralogisms, 142–43, 145, 146, 147, 155, 160, 218 personal identity. See identity of the Self phenomena, 65, 66, 74, 214, 224, 225, 230, 232, 234–42, 266, 271, see also appereances phenomenal, 80–82 Pippin, Robert, 124 Plato, 251, 256 Postulates, 165, 269–73 pre-Critical period, 9–13, 26, 75, 189 principle of reason, 230, 232, 233, 242, 251–52 Principles of Pure Understanding, 163–83 Pure Concepts. See concepts, pure and categories pure intuition. See intuition, pure quid juris, 106, 108, 109 Rational Psychology, 142–43, 145, 151, 233 rationalism, 10, 13, 14, 18, 22, 26, 27, 126, 138–39, 193, 196, 217, 232, 243, 260, 264–67 realism. See anti-realism, empirical realism, scientific realism, transcendental realism reason, 21, 22, 30, 36, 66, 82, 114, 205, 223, 224, 232–33, 238, 240, 241, 244, 246, 248, 251, 252, 256, 257, 258, 264, 267, 270, 274, see also idea of reason, Ideal of Pure Reason, principle of reason concern of, 181 divine, 247, 256–58 faculty of, 20, 21, 32, 91, 95, 102, 122, 249, 253 practical use of, 278 pure, 31, 33, 59, 69, 83, 108, 232–33, 238, 240, 246, 252, 278 speculative, 245 transcendental ideal of, 247 transcendental problems of, 231 tribunal of, 32 receptivity, 14, 16, 21, 37, 42, 43, 66, 127, 171 Refutation of Idealism. See idealism, refutation of regulative demand of Reason, 256 principle, 202–03, 252, 256, 258 use of concepts, 257 use of Ideas, 217, 231, 241 use of Reason, 257 vs constitutive, 223, 225, 232–33 Ricoeur, Paul, 244 Rosenberg, Jay, 54, 62 Russell, Bertrand, 27, 191

Schematism, 164, 165, 190 Schliesser, Eric, 68 science, 16, 30, 76, 86, 183, 235, 241, 257 cognitive, 126 mathematical, 163, 173, 176, 178 metaphysics as a, 29, 32, 34, 108, 164 scientific realism, 204 sensations, 10, 11, 14–18, 19, 21–22, 23, 25–26, 37, 43, 50, 51–52, 120, 144–45, 157–58, 175, 182, 186–87, 189, 197, 198 sensibility, see also receptivity faculty of, 9–10, 11, 13–16, 17, 18, 23, 24, 28, 31, 37, 43, 66, 125, 128, 134, 138, 141, 164, 166, 169, 173, 181–82, 189, 251, 253 formal conditions of, 55, 58, 59, 66, 72, 81, 109, 121, 131–32, 135, 177 principles of, 168, 172, 178, 182, 183 Shoemaker, Sydney, 156, 157 simultaneity, 196, 207, 215, 221, 222 skepticism, 117, 187–88, 194, 201, 205, 209, 211, 221 soul. See immortality of the soul space, 9, 11, 17–18, 20, 21–22, 31, 42, 46–49, 52, 53, 54, 56, 62, 66, 69, 70, 75, 82, 110, 112, 121, 129, 132, 135, 149, 160, 164, 165–72, 175–82, 184, 187, 190–203, 215, 217–18, 223, 228, 263, 268–71, see also intuition, forms of, idealism, about space; necessity, of space as phenomenon, 77–82 Spinoza, Baruch, 64–66, 68–75, 76, 79, 81–82, 246, 252 spontaneity, 21, 42, 43, 127, 135, 141, 235, 240 Strawson, Peter Frederick, 28, 66, 143–47, 160, 184, 209 substance, 20, 31, 42, 65, 66, 68, 75, 146, 160, 185, 194–97, 199, 200, 202–03, 213–22, 257, 269, 270, 273, 277, see also matter; existence of substance and God, 68, 69–82 Principle of the Persistence of, 195, 214 succession, 176, 187, 194, 199–200, 207, 215, 222 syllogism, 16, 92, 101, 150, 251 syllogistic inference, 92–93, 95, 97, 99–102 logic, 85, 96, 97, 98 reasoning, 251 use of concepts, 85 synthesis, 38, 41, 43, 47, 51–52, 54, 63, 129, 131, 132, 136–38, 147, 148–49, 152, 154, 158, 161, 169, 172, 173, 174, 175–77, 179, 181, 183, 186–87, 229, 231 Tempier, Étienne, 187, 204 Tetens, Johannes Nikolaus, 189

Index things in themselves, 9–10, 20, 22, 26, 31, 48–49, 55, 58–59, 66, 69–71, 79, 81–82, 164, 166, 177, 224, 226, 267, 268, see also noumena time, 9, 11, 18, 20, 21, 22, 46, 48–49, 56, 60, 62, 66–70, 72, 75, 81, 110, 112, 121, 129, 132, 135, 158, 160, 164, 165, 167, 170–72, 175–77, 179, 180–82, 184, 187, 190–92, 195, 197–200, 202–03, 213, 223, 228–30, 232, 235, 263, 268, 270, 271 determination of the Self in, 206–22 totality absolute, 231–33, 235, 237 of predicates, 252 of substances (world), 4, 65, 73, 76–78, 79, 80 Transcendental Deduction, 20, 21, 43, 47, 107–19, 136–39, 207 transcendental realism, 118, 235, 237 unconditioned, the, 224–25, 231–32, 242 understanding, see also Pure Concepts, Principles of Pure Understanding, Spontaneity


concepts of the, 84, 104, 109, 119, 124–25, 126, 133–34, 169, 232 faculty of, 9–11, 13–14, 16–27, 28, 31, 37, 39, 43, 77, 78, 83–96, 104, 108–09, 121, 123, 128, 131, 136, 138–39, 152, 172, 189, 228, 231, 237, 248, 250, 255, 259, 267, 276 principles of the, 83, 141, 271 unity of the, 136–37 universality, 39, 51, 94, 234–35 unknowability of things in themselves, 66, 70 Vaihinger, Hans, 9 Van Cleve, James, 48 Vogel, Jonathan, 216, 219 Watkins, Eric, 69, 147 Wittgenstein, Ludwig, 145 Wolff, Christian, 11–12, 22, 26, 67, 98, 99, 104, 266 Wood, Allen W., 166, 249 world. See totality of substances (world)

cambrid ge c ritical g uides Titles published in this series: Hegel’s Phenomenology of Spirit edited by dean moyar and michael quante Mill’s On Liberty edited by c. l. te n Kant’s Idea for a Universal History with a Cosmopolitan Aim edited by a m é l i e o k s e n b e rg ro rt y and j a m e s s c h m i d t Kant’s Groundwork of the Metaphysics of Morals edited by je ns t i m m e rm an n Kant’s Critique of Practical Reason edited by and rews re at h and je ns tim m er m an n Wittgenstein’s Philosophical Investigations edited by arif ahme d Kierkegaard’s Concluding Unscientific Postscript edited by rick an tho n y f u rtak Plato’s Republic edited by mark l. mcph erran Plato’s Laws edited by c h r i s to ph e r b o b o n i c h Spinoza’s Theological-Political Treatise edited by y i t z hak y. m e l am e d and m i c hael a. ro sen t h al Aristotle’s Nicomachean Ethics edited by j o n m i l l e r Kant’s Metaphysics of Morals edited by l ara de n is Nietzsche’s On the Genealogy of Morality edited by s imo n may Kant’s Observations and Remarks edited by ri c hard ve l k l ey and s u san sh ell Augustine’s City of God edited by jame s we tze l Descartes’ Meditations edited by k are n de tle f s e n Kant’s Religion within the Boundaries of Mere Reason edited by g o rdo n m i c h a l s o n Kant’s Lectures on Anthropology edited by alix co he n Kierkegaard’s Fear and Trembling edited by dan ie l co n way Kant’s Lectures on Ethics edited by l ar a d e n i s and o l i ve r s e n s e n Aristotle’s Physics edited by m ari s k a l e u ni s s en Aristotle’s Politics edited by t ho rnto n lo c k wo o d and than assis sam ar as Aquinas’s Disputed Questions on Evil edited by m. v. do u g he rt y Plato’s Symposium edited by pie r re d e s t r é e and z i n a g i a n n o p o u lo u Spinoza’s Ethics edited by yitzhak y. mel amed