The World According to Kant: Appearances and Things in Themselves in Critical Idealism 9780199695386, 0199695385

Citation preview

The World According to Kant

The World According to Kant Appearances and Things in Themselves in Critical Idealism ANJA JAUERNIG

1

3

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Anja Jauernig 2021 The moral rights of the author have been asserted First Edition published in 2021 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2020945711 ISBN 978–0–19–969538–6 DOI: 10.1093/oso/9780199695386.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

für meine Eltern

Preface and Acknowledgments Things in themselves play a central role in Kant’s critical philosophy, that is, his mature philosophy that he presented to the world in his publications from 1781 onwards starting with the Critique of Pure Reason. As allegedly unknowable grounds of all knowable objects, things in themselves are as important in Kant’s ontology as they are perplexing in his epistemology. The present book grew out of the question of how to think about things in themselves, according to Kant. This question can be understood in two ways. First, one can take it as a broadly ontological question about Kant’s conception of things in themselves. What sort of things are they? What are their properties? What is their ontological status in Kant’s critical philosophy? Second, one can regard it as a question about whether, and if so how, we can have meaningful thoughts about or even cognize things in themselves, according to Kant’s theory of thinking and cognition. In virtue of what can our concepts refer to them? How can we manage to cognitively access them through our thinking? What sort of cognition of them, if any, is possible for us? More than a decade ago, in my (then) youthful naïveté and optimism, I decided to select this double-sided question as the topic for a short essay intended as a brief restorative diversion from my ‘real’ project at the time (which was to sort out the precise relation between Kant’s philosophy, the philosophy of Leibniz, and the philosophy of Wolff and his followers). The diversion turned out to be neither brief nor restorative. Kant is a truly systematic thinker; most of his doctrines and arguments are intricately connected with one another. As a result, in order to make genuine headway on any particular question about Kant’s philosophy, one is forced to cast one’s net rather widely. The question of how to think about things in themselves is no exception. In order to properly understand Kant’s conception of things in themselves, one also has to understand his conception of appearances and his views about how things in themselves and appearances are related. One also must take into account Kant’s evaluation of the conceptions of things in themselves of other philosophers, in particular, the Leibniz-Wolffian conception, which represents one of the central reference points for Kant’s own thinking about this topic. In short, a satisfactory explanation of Kant’s conception of things in themselves requires, more or less, a comprehensive explanation of his ontological views in general. Similarly, making real progress with respect to the question of whether, how, and to what extent, we can think about and cognize things in themselves on Kant’s view requires giving a detailed account of his theory of meaning, reference, and thought and cognition in general. This is how my envisioned brief essay on how to think about things in themselves, by and by, transmogrified into a long essay, a book, and, finally, two books. I should have published these books a long while ago—the original manuscript (before the ‘fission’) was completed in 2009—and I confess to feeling considerable embarrassment about having failed to do so. But the devil in the details continued to beckon without mercy, and every ‘final’ round of revisions kicked up yet other lines of questioning that deserved to be pursued. Not all of these lines of questioning yielded fruits worthy to be included in the books, but they all kept me busy and caused delay over delay. Although this exorcism is by no means completed, I finally got

xii

  

up the nerve to let at least the first book go. This book, now before you, is devoted to an examination of Kant’s critical idealism, understood as an ontological position. Less technically put, it is about Kant’s account of what there is in the world, understood as the sum total of everything that has reality, including, in particular, his account of appearances and things in themselves and their relation to one another. The second book, Thought and Cognition according to Kant—Our Cognitive Access to Things in Themselves and Appearances in Kant’s Critical Philosophy, investigates Kant’s theory of thinking and cognition, with a special emphasis on the implications of this theory for the question of whether and how we can think about and cognize things in themselves. I hope it will be released into the world before long, but, for the moment, I have to admit that taking a little break from endlessly curating ever-expanding footnotes seems very attractive, as much as I admire Sisyphus and imagine him happy. These two books are conceived of as companion volumes—one ontological and one epistemological, broadly speaking—but each one can also be read on its own. I presented different parts of the material included here at Yale University, the University of Michigan, the University of Pittsburgh, the University of Oxford, the University of North Carolina at Chapel Hill, the University of Miami, the North American Kant Society at the Eastern APA, Harvard University, the University of California at Berkeley, the University of Chicago, New York University, the New School for Social Research, Wake Forest University, Brown University, Syracuse University, and Cornell University. I thank the audiences at these occasions for helpful comments and questions, in particular, Desmond Hogan who was my commentator at Chapel Hill and Julian Wuerth who was my commentator at the APA. I also gratefully acknowledge valuable feedback from Banafsheh Beizaei, Andrew Chignell, Don Garrett, Christian Johnson, Nick Stang, and Eric Watkins. Peter Momtchiloff from Oxford University Press deserves a medal for his patience and has my sincere gratitude for his unwavering encouragement. Last but certainly not least, I want to express my deep appreciation to Robert Adams, Karl Ameriks, and Béatrice Longuenesse for their kind support over the years and their penetrating and incredibly useful comments on various parts of this book.

Abbreviations All quotations from Kant are cited according to the Academy Edition, in the format ‘Abbreviated Title, volume:page number(s),’ except for quotations from the Critique of Pure Reason, which are cited according to the paginations of the original first (A) and second (B) edition, and for quotations from the anthropology lecture transcript by Dohna-Wundlacken, which is cited as ‘V-Anth/Dohna, Kowalewski’ according to the pagination of the first publication of the lecture transcript by Kowalewski from 1924. All other titles are referenced by the author’s name and the year in which the work was first published or written (in the case of posthumous works), as specified in the following bibliography. Some of the bibliography entries contain several publication dates. The date listed first is the one that I use for purposes of reference; dates marked with an asterisk indicate publication dates of the first edition. Unless an English translation is referenced, all translations are my own, from Kant and everybody else not writing in English. I have aimed for as literal translations as possible, grammatical errors, infelicitous word choices, and ambiguities included. I also would like to state explicitly that my interpretation of Kant is based on the German text, not the English translation. My efforts to provide a faithful translation notwithstanding, I do think that this makes a difference. In general, I urge all readers to consult Kant’s original German text as much as they can stomach while reading this book. (The entire Academy Edition is conveniently available online for free at http://www.korpora.org/kant/). A Anth BDG

B BJ

EEKU EF FM

FM/LB GMS Kä KpV KU LB Log MAN

Kritik der reinen Vernuft, erste Auflage von 1781 (Critique of Pure Reason, first edition from 1781) Anthropologie in pragmatischer Hinsicht (Anthropology from a Pragmatic Standpoint) Der einzig mögliche Beweisgrund zu einer Demonstration des Daseins Gottes (The only possible ground of proof for a demonstration of the existence of God) Kritik der reinen Vernunft, zweite Auflage von 1787 (Critique of Pure Reason, second edition from 1787) Einige Bemerkungen zu Ludwig Heinrich Jakob’s Prüfung der Mendelssohn’schen Morgenstunden (Some comments on Ludwig Heinrich Jakob’s Examination of Mendelssohn’s Morning Hours) Erste Einleitung in die Kritik der Urteilskraft (First Introduction to the Critique of Judgment) Erklärung in Beziehung auf Fichtes Wissenschaftslehre (Declaration with respect to Fichte’s Doctrine of Science) Welches sind die wirklichen Fortschritte, die die Metaphysik seit Leibnizens und Wolff ’s Zeiten in Deutschland gemacht hat? (What is the real progress that has been made in metaphysics in Germany since the time of Leibniz and Wolff?) Lose Blätter zu FM Grundlegung zur Metaphysik der Sitten (Groundwork for the Metaphysics of Morals) Über Kästners Abhandlungen (On Kästner’s Treatises) Kritik der praktischen Vernunft (Critique of Practical Reason) Kritik der Urteilskraft (Critique of Judgment) Lose Blätter aus Kants Nachlaß (Loose sheets from Kant’s literary estate, listed using Reicke’s numbering) Logik nach Jäsche (Logic) Metaphysische Anfangsgründe der Naturwissenschaft (Metaphysical Foundations of Natural Science)

xiv



MS MSI

NG NTH OP PND Prol R SF Söm TG ÜE

V-Anth/Dohna VA-Prol VA-Söm V-Lo/Pölitz V-Lo/Wiener V-Met/Dohna V-Met/Heinze V-Met/Herder V-Met-K2/Heinze V-Met-K3/Arnoldt V-Met-L1/Pölitz V-Met-L2/Pölitz V-Met/Mron V-Met/Schön V-Met/Volckmann V-Mo/Collins V-Phil-Th/Pölitz VT V-Th/Baumbach V-Th/Volckmann WDO WSLK

Die Metaphysik der Sitten (Metaphysics of Morals) De mundi sensibilis atque intelligibilis forma et principiis (On the Form and Principles of the Sensible and Intelligible world; Kant’s Inaugural Dissertation) Versuch, den Begriff der negativen Größen in die Weltweisheit einzuführen (Attempt to Introduce the Concept of Negative Magnitudes into Metaphysics) Allgemeine Naturgeschichte und Theorie des Himmels (Universal History of Nature and Theory of the Heavens) Opus Postumum Principiorum primorum cognitionis metaphysicae nova dilucidatio (New Elucidation of the First Principles of Metaphysical Cognition) Prolegomena zu einer jeden künftigen Metaphysik (Prolegomena to Any Future Metaphysics) Reflexion (Reflection); R followed by a Roman numeral indicates a reflection in Kant’s personal copy of the A-edition of the Critique of Pure Reason Der Streit der Fakultäten (The Battle of the Faculties) Aus Sömmering: Über das Organ der Seele (From Sömmering: On the Organ of the Soul) Träume eines Geistersehers erläutert durch Träume der Metaphysik (Dreams of a Spirit-Seer Elucidated through Dreams of Metaphysics) Über eine Entdeckung, nach der alle neue Kritik der reinen Vernunft durch eine ältere entbehrlich gemacht werden soll (On a Discovery, according to which all new critique of pure reason is supposed to be made dispensable by an older one) Anthropologie Vorlesung Dohna-Wundlacken (Anthropology Lecture) Vorarbeit für die Prolegomena (preparatory notes for the Prolegomena) Vorarbeit für den Sömmering Aufsatz (preparatory notes for the essay on Sömmering) Logik Vorlesung Pölitz (Logic Lecture) Wiener Logik Vorlesung Metaphysik Vorlesung Dohna (Metaphysics Lecture) Metaphysik Vorlesung L₁ Heinze Metaphysik Vorlesung Herder Metaphysik Vorlesung K₂ (Heinze, Schlapp) Metaphysik Vorlesung K₃ (Arnoldt, Schlapp) Metaphysik Vorlesung L₁ Pölitz Metaphysik Vorlesung L₂ Pölitz Metaphysik Vorlesung Mrongovius Metaphysik Vorlesung von Schön Metaphysik Vorlesung Volckmann Moralphilosophie Vorlesung Collins (Moral Philosophy Lecture) Philosophische Religionslehre Vorlesung Pölitz (Philosophical Religion Lecture) Von einem neuerdings erhobenen vornehmen Ton in der Philosophie (On a Recently Adopted Pretentious Tone in Philosophy) Danziger Rationaltheologie Vorlesung nach Baumbach (Rational Theology Lecture) Natürliche Theologie Vorlesung Volckmann nach Baumbach (Natural Theology Lecture) Was heißt sich im Denken orientiren? (What Does it Mean to Orient Oneself in Thinking?) Wahre Schätzung der lebendigen Kräfte (True Estimation of Living Forces)

1 Introduction—the Plan for this Book and the Lay of the Land 1.1 General Project Description: Understanding Kant’s Critical Idealism The project of this book is to develop a comprehensive interpretation of Kant’s ontological views in his theoretical critical philosophy, or, perhaps more accurately, of the ontological implications of his theoretical critical philosophy. More specifically, my aim will be to sort out Kant’s views on the nature of appearances, the nature of things in themselves, and the relation between them, as presented and hinted at in the Critique of Pure Reason (‘Critique’ for short from now on) and associated writings. Following Kant, I will refer to the position defined by these ontological views as ‘critical idealism.’¹ Accordingly, one could also characterize this book as aiming at a comprehensive interpretation of Kant’s critical idealism in his theoretical philosophy as an ontological position. To be sure, since Kant is an idealist and holds that certain parts of reality are mind-dependent, investigating his ontological views will bring with it an examination of at least some aspects of his theory of cognition and his theory of the human mind. But for the purpose of this book, this examination will be quite selective and restricted to those elements of said theories that have ontological implications or shed light on the nature and ontological status of appearances. This project description will be met with skepticism, or worse, by readers who subscribe to the assessment that Kant’s philosophy, in particular, his theoretical philosophy, is thoroughly anti-metaphysical and self-consciously refrains from making any ontological claims at all. Throughout the history of the reception of Kant’s philosophy, the fortunes of anti-metaphysical and metaphysical interpretations have waxed and waned several times; each group had its moments in the sun but also several periods in the shade, so to speak. Proponents of anti-metaphysical, anti-ontological interpretations read Kant as a philosopher who has laid metaphysics to rest once and for all, and who takes worthwhile theoretical philosophy to be coextensive with logic and epistemology, broadly conceived, but has no time for ontology. This kind of interpretation was prominently defended by leading neo-Kantians in the late nineteenth and early twentieth century, for example, who saw Kant as having effectively destroyed the metaphysical edifices of his predecessors, and who celebrated as one of Kant’s main achievements the substitution of Geltungstheorie (‘validity-theory’) for ontology, that is, the substitution of an examination of the conditions for the objective validity of judgments, in particular, scientific judgments, for the

¹ See Prol, 4:293–294. Throughout this book, I will be using single quotation marks as scare quotes and to talk about names, words, concepts, predicates, and sentences. Double quotation marks will be used exclusively to indicate quotations and for essay titles. The World According to Kant: Appearances and Things in Themselves in Critical Idealism. Anja Jauernig, Oxford University Press (2021). © Anja Jauernig. DOI: 10.1093/oso/9780199695386.003.0001

2



examination of what there is.² Anti-metaphysical readings have also been dominant in Anglo-American Kant scholarship in the second half of the twentieth century ever since the publication of Strawson’s seminal The Bounds of Sense and have retained their popularity to the present day. For example, Henry Allison, who is one of the most prominent living Kant interpreters today, asserts that Kant’s transcendental idealism “provides a radical alternative to ontology” (Allison 2006, 2).³ Proponents of metaphysical interpretations, by contrast, read Kant as the founder of a new kind of, or a new way of doing, metaphysics and see him as taking a stand on ontological matters, while (for the most part) acknowledging that his methodology differs markedly from the methodology of the traditional metaphysicians who are the target of his critique in the Critique. Apart from several contemporaries of Kant, this group includes later eminent Kant interpreters such as Friedrich Paulsen, Heinz Heimsoeth, Max Wundt, and Martin Heidegger, who were trying to stem the anti-metaphysical tide unleashed by the neo-Kantians.⁴ Although the majority of present day Kant scholars still appears to favor anti-metaphysical, anti-ontological interpretations, if I read the signs correctly, the ray of prominence and popularity has started to move past them and is about to shine its light on the metaphysics- and ontologyfriendly group again. There is a growing number of scholars who read Kant as a metaphysician of sorts and as very much interested in ontology, and I count myself among them.⁵ As I see it, the indicated anti-metaphysical, anti-ontological kind of Kant interpretation is misguided. To start with the question of metaphysics, Kant tells us in no uncertain terms that his ultimate philosophical concern, like the ultimate philosophical concern of all of us, is metaphysics, which he understands as a science that aims to provide us with a priori cognitions, in particular, a priori cognitions of the supersensible.⁶ And, far from being on a mission to put metaphysics out of business, he is on a mission to put metaphysics “onto the secure path of a science,” as he impresses on his readers in the preface to the Critique.⁷ In a letter, Kant explains his philosophical intentions by clarifying his often misunderstood position vis-à-vis Leibniz-Wolffian metaphysics as follows: At the same time, I may be permitted to explain that my efforts, which so far have been directed at critique, are by no means intended to work against the Leibniz-Wolffian philosophy, as it might have appeared, (for I have been finding it neglected for quite some time now). Rather they are intended to lead through a roundabout route that, it seems to me, these great men judged superfluous, onto the same track of a schooled procedure and through it to the very same destination, but only through the combination of the theoretical philosophy with the practical—an intention that will become clearer if I live long enough to present metaphysics in form of a coherent system, as I am planning to do. (Letter to Kästner, August [?] 5, 1790, 11:186)

² See Cohen 1885; Natorp 1912; Bauch 1917. ³ Another example of a contemporary self-consciously anti-ontological, epistemic reading of Kant’s transcendental idealism is Patricia Kitcher’s, in Kitcher 2011. ⁴ See Paulsen 1899; Paulsen 1900; Wundt 1924; Heimsoeth 1924; Heidegger 1929. Also see Martin 1948; Martin 1951. ⁵ Among others, this group includes Karl Ameriks (Ameriks 2003), Rae Langton (Langton 1998), and Eric Watkins (Watkins 2005). Also see Lipscomb and Krueger 2010. ⁶ See FM, 20:293–301; B6–7; B21–22; Prol, 4:353. ⁷ See Bxiv–xxxi.

  :   ’   

3

While Kant employs a new methodology and offers a new, more sophisticated analysis of our cognitive attitudes toward many traditional metaphysical propositions of the LeibnizWolffians―classifying them no longer as a kind of theoretical but as a kind of practical cognition―his ultimate destination is the same as the destination of the Leibniz-Wolffians, namely, the cognition of the supersensible. Indeed, turning to the question of ontology, Kant arguably not only shares the aim of cognizing the supersensible with the LeibnizWolffians but also much of their conception of the supersensible as a realm populated by simple substances or souls under the legislation of God, substances on whose perceptions the phenomenal world depends. To be sure, Kant’s theoretical philosophy by itself does not warrant the ascription of all elements of this Leibniz-Wolffian ontological picture to him. Moreover, it is also undisputed that in his theoretical philosophy Kant spends much of his time and effort on issues in logic and epistemology, broadly conceived. Nevertheless, it seems undeniable to me that, although these theoretical inquiries are not directly targeting ontological questions, they have substantive ontological implications, implications that Kant takes to be of crucial importance for metaphysics and for its subsequent progress in the context of his practical philosophy. This is part of his new methodology, to arrive at ontological results by way of epistemological examinations. As we will see, chief among these ontological implications of his theory of experience and of his analysis of the workings of the human cognitive faculties that are employed in the construction of experience are that space and time are nothing but forms of one of these faculties, namely, sensibility, that empirical objects, such as tables, trees, and cats, are appearances and not things in themselves, and that appearances are grounded in things in themselves. These theses lie at the heart of the new ontological position that is Kant’s critical idealism, which is a crucial element of his envisioned revamped form of metaphysics. Although Kant holds that the most appropriate name for the ontological views developed in his theoretical critical writings is ‘critical idealism,’ the name that he introduced initially and continued to use most frequently, and that most people associate with him, is ‘transcendental idealism.’ So, why do I not describe my project in more familiar terms as aiming to provide an interpretation of Kant’s transcendental idealism as an ontological position? In order to answer this question, I need to anticipate a terminological suggestion that will be spelled out in section 3.2. Kant uses the expression ‘transcendental idealism’ in a broader and a narrower sense. In the narrower sense, transcendental idealism can be understood as defined by the two theses about the ontological status of empirical objects and of space and time mentioned in the previous paragraph, that is, that empirical objects are appearances and not things in themselves, and that space and time are nothing but forms of sensibility, plus a corresponding thesis about (what I will call) empirical selves, namely, that they are also appearances and not things in themselves. In the broader sense, transcendental idealism can be understood as a more comprehensive position that comprises transcendental idealism in the narrow sense but also several other elements. These other elements include empirical realism (in a sense to be explicated in chapter 3) and several additional claims, concerning things in themselves and their relation to appearances as well as the nature of our cognitive faculties, that are usually not mentioned in Kant’s explicit characterizations of transcendental idealism but still play an important role in defining his ontological position overall, as for example the claim that things in themselves ground appearances, or that sensibility is passive and can bring about sensations only if it is affected by things in themselves. For the sake of terminological clarity, it is

4



desirable to have two different terms available to signify transcendental idealism in the broad and the narrow sense respectively. This is why I propose to use ‘critical idealism’ to stand for Kant’s overall theoretical ontological position and ‘transcendental idealism’ to stand for a core element of critical idealism, namely, the three stated theses about the ontological status of empirical objects, empirical selves, and space and time. Given this terminology, saying that my purpose in this book is to provide an interpretation of Kant’s transcendental idealism would not quite capture the full scope of the present study. Getting to the bottom of Kant’s transcendental idealism is only part of my purpose, albeit an important and central one.

1.2 Mapping Different Interpretations of Kant’s Critical Idealism At the core of Kant’s transcendental idealism and thus of his critical idealism is the claim that the objects of human experience are not things in themselves but appearances. The question of how to understand transcendental idealism and, especially, how to spell out the crucial distinction between things in themselves and appearances has puzzled many a reader of Kant, and the secondary literature is full of competing interpretations.⁸ For ease of communication, I will refer to this distinction as ‘the transcendental distinction’ for short from now on. Recent discussions of the transcendental distinction have been dominated by the debate between partisans of what has come to be called the ‘two-world’ or ‘two-object’ view and partisans of what has come to be called the ‘one-world,’ ‘one-object,’ or ‘two-aspect’ view.⁹ The latter regard appearances and things in themselves as, in some sense, ‘the same things,’ while the former deny that they are the same things. Unfortunately, despite (or maybe because of) the prevalence of the two-world versus two-aspect terminology, there does not seem to be any clear agreement among commentators about the exact conditions that an interpretation has to satisfy in order to qualify as one or the other of these kinds of views. Indeed, one is hard pressed to find any precise characterizations of these conditions in the literature at all. As a result, the battle lines in the two-world versus two-aspect wars that have been raging for several decades now are rather blurry, and the fighting often appears to be more akin to shadow boxing, or windmill sparring, than to a well-disciplined exchange of attacks and defense maneuvers among real-life opponents. I want to emphasize at the outset that I do not conceive of this book as yet another campaign in this endless war. I agree that there are several exegetical questions haunting this battlefield that are worth clarifying and addressing; and part of my task in the following will be to do precisely that. But those are not the only questions nor, in my view, the most important questions, that deserve close attention if one wants to properly understand Kant’s critical idealism, or even his transcendental distinction. For the purpose of surveying the interpretative landscape, it is helpful to identify a number of classification questions with respect to which different interpretations of Kant’s ⁸ For a useful review of the literature of the final decades of the last century, see Ameriks 2003b, 69–84; Ameriks 2003c. For a helpful review of the literature that extends into the twenty-first century, see Schulting 2011. ⁹ I suppose the grammatically more appropriate versions of these labels would be ‘two-worlds,’ ‘two-objects,’ and ‘two-aspects’ views. But the s-less versions admittedly sound much better, which is presumably why they dominate in the literature. I will follow the practice of choosing sound over grammar in these cases.

     ’   

5

critical idealism can be distinguished and categorized. The first classification question will help to set up how I propose to chart the two-world versus two-aspect minefield. It asks whether appearances and things in themselves are the same things, where ‘same’ is to be understood in the sense of numerical identity. More precisely put, the question is whether every appearance is numerically identical to a thing in itself.¹⁰ Readings on which appearances and things in themselves are numerically distinct commonly present Kant’s transcendental distinction as a distinction between different entities, different things, different objects, or different domains of discourse. Interpretations of this kind are usually called ‘two-world’ or ‘two-object’ views. In the following, I will adopt the name ‘multiple-object view’ for all interpretations that answer our first classification question in the negative, but, for reasons that will become clear shortly, I will reserve the name ‘two-world view’ for a specific version of the multiple-object view. I prefer ‘multiple-object view’ to ‘two-object view’ because the latter might be taken to suggest that proponents of this view, although denying that every appearance is numerically identical to a thing in itself, still hold that there is a one-to-one mapping between appearances and the things in themselves that ground them. Some people who answer the first classification question in the negative might want to assert that such a one-to-one mapping obtains, but going with the negative answer certainly does not commit one to doing that; hence, the greater suitability of the ‘multiple-object’ label.¹¹ Readings on which every appearance is numerically identical to a thing in itself commonly present the transcendental distinction as a distinction between different aspects of, or different ways of considering, the same thing. The thing (considered) as it appears to us is the appearance, the same thing (considered) as it is in itself is the thing in itself. In the following, interpretations of this kind will be called ‘oneobject views.’ I will use the label ‘two-aspect view’ more broadly than ‘one-object view.’ One-object views are a kind of two-aspect interpretation; indeed, they are the most popular kind.¹² But, in my usage of the term, one-object views are not the only kind of two-aspect interpretation, as will be explained shortly. It is worth emphasizing that the proposed characterization of the multiple-object view is fairly permissive. It requires only that appearances and things in themselves are numerically distinct but places no restrictions on the ontological categories into which they can fall. To a historian of philosophy who is familiar with Descartes’ theory of distinction, it may come naturally to read Kant’s transcendental distinction against this Cartesian background.¹³ According to Descartes, there are three kinds of distinctness: real distinctness that pertains to different substances, modal distinctness that pertains to a substance and its modes (or, derivatively, to different modes of the same substance), and distinctness by reason that pertains to a substance and its attributes (or, derivatively, to different ¹⁰ It is uncontroversial that Kant admits the possibility of things in themselves that do not appear to us, for example, God. That is, not every thing in itself is numerically identical with an appearance. Hence, the need for the more precise formulation. ¹¹ None of these labels are perfect, but we need some way to refer to the different kinds of interpretations, and since the ‘one-object,’ ‘two-aspect,’ and ‘two-world’ labels have become common currency in the literature (despite not being used in a uniform way) I will work with them but not without clarifying how exactly they are supposed to be understood in my usage. The name ‘multiple-object view’ is not perfect either, among other things, because, strictly speaking, things in themselves are not proper objects, as we will see in section 2.5.2. So ‘object’ in ‘multiple-object view’ must be understood in a broad sense. ¹² Most of the two-aspect commentators mentioned in notes 32 34, 36 39, and 70, endorse a one-object view. ¹³ Thanks to Don Garrett for prompting me to think more about the connection of Kant’s transcendental distinction and Descartes’ theory of distinction.

6



attributes of the same substance).¹⁴ The historian may be tempted to assert that the ‘multiple-object’ label is appropriate only for interpretations that present appearances and things in themselves as really distinct in (roughly) Descartes’ sense, that is, as belonging to the same, most fundamental ontological category such as ‘substance,’ or ‘thing’ (understood in an ontologically robust sense), and as being ontologically independent in a modally robust sense.¹⁵ According to our characterization, however, interpretations that present appearances and things in themselves as modally distinct in (roughly) Descartes’ sense also count as multiple-object readings, as for example the reading that appearances are complexes of mind-dependent properties of things, and things in themselves are those very things but without any mind-dependent properties.¹⁶ Similarly, some contemporary commentators express reservations about being classified as proponents of the two-world view because, even though, on their reading, the transcendental distinction is a distinction between numerically distinct entities, only one kind of these entities deserves to be called a kind of things, strictly speaking. Something like this appears to be Rae Langton’s contention. According to her reading, “things in themselves are substances that have intrinsic properties,” and appearances are identified “with the relational properties of substances, and with whatever things are constituted by those relational properties” (Langton 1998, 20).¹⁷ Since a substance that has intrinsic properties and something that is constituted by relational properties arguably are numerically distinct, it is a fair description that, on her view, things in themselves and appearances are numerically distinct entities. But with respect to the question of whether, on Kant’s view, there is one world or two, Langton replies that “there is one world, one set of things, but two kinds of properties” (Langton 1998, 12 13).¹⁸ In a related vein, other commentators express reservations about being classified as proponents of the two-world view because, although, on their reading, appearances and things in themselves belong to different domains of discourse and, thus, are numerically distinct, they are not distinct existents and, hence, also cannot properly be said to make up distinct worlds, simply because, strictly speaking, appearances do not genuinely exist. James Van Cleve holds a view of this kind. He emphasizes that his virtualobject theory of appearances, according to which appearances are understood as logical constructions out of states of (noumenal) perceivers, does not give us a dualism of two sorts of existents. The only existents, on Van Cleve’s view, are things in themselves. But he acknowledges that his interpretation is dualistic in that the distinction between appearances and things in themselves is understood as a distinction between two separate

¹⁴ See Descartes 1644, §60, AT 8:28–30. For an illuminating discussion of Descartes’ theory of distinction, see Hoffman 2002. ¹⁵ On Descartes’s view, x and y are really distinct if, and only if, x and y are substances and it is possible that x exists and y does not exist, and vice versa, which, for him, is equivalent to saying that x can be clearly and distinctly understood apart from y, and vice versa. See Descartes 1644, §60, AT 8:28. ¹⁶ Interpretations on which appearances and things in themselves are distinct only ‘by reason’ in (roughly) Descartes’ sense do not count as multiple-object, though, since this kind of distinctness is not sufficient for numerical distinctness (on my view). (A substance and its attributes are distinct by reason but, arguably, still numerically identical.) The methodological one-object view, to be described below, is an interpretation of this kind. ¹⁷ Also see Langton 1998, 12–21, 34–40. ¹⁸ I should acknowledge that Langton herself does not explicitly say that appearances are entities. But I do see no reason why “the relational properties of substances, or whatever things are constituted by those relational properties” should not count as entities. (Note that she even says, somewhat misleadingly, that appearances are “whatever things are constituted by those relational properties” (Langton 1998, 20, my emphasis). ‘Things’ must be understood in a broader sense here, I suppose.)

     ’   

7

universes of discourse, which is why he calls his reading a “qualified two-worlds view” (Van Cleve 1999, 150).¹⁹ I agree that the ‘two-world’ label is prone to eliciting various misleading associations, and so I sympathize with Langton’s and Van Cleve’s reluctance to affix it to their interpretations. But I want to make a few comments with respect to the question of whether appearances should be classified as things or existents. There is a certain danger here of sliding into a mostly terminological dispute—how do we want to use the words ‘thing’ and ‘existent’?—and of imposing upon Kant’s view our own understanding of what should count as a thing or existent. (This statement is meant as a general note of caution, not as a specific complaint about Langton or Van Cleve.) As I see it, insofar as we want to engage in this kind of debate at all, we should try to settle it based on Kant’s text. That is, we should try to determine, based on Kant’s own characterizations of appearances, things, and existents, whether he is prepared to classify appearances as existents and, more specifically, whether he counts them as things of some kind or in some sense. To be sure, there is no question (in my mind) that Kant regards things in themselves to be ontologically more fundamental than appearances, but by itself this is compatible with holding that appearances also exist and are also some sort of things. The upshot of these considerations is that commentators who want to explain Kant’s transcendental distinction should not only aim to determine whether appearances and things in themselves are numerically distinct or not but also try to identify to which more specific ontological categories they belong, according to Kant’s own understanding of these categories, and exactly what kind of ontological dependence relation obtains between them. But as important as these further questions are, there is no need to over-complicate our classification scheme by introducing special sub-categories for different varieties of the multipleobject view, depending on which ontological categories appearances and things in themselves are assigned to. We will stick with the generously inclusive characterization that anybody who holds that things in themselves and appearances are not numerically identical is to count as a proponent of the multiple-object view. There are some commentators who see themselves as proponents of a version of the two-aspect interpretation, on account of the fact that they endorse the view that appearances and things in themselves are, in some sense, the same things, but who refrain from asserting that every appearance is numerically identical to a thing in itself.²⁰ Some versions of the reading that the transcendental distinction is supposed to be understood as a distinction between the ‘world’ of appearances and the ‘world’ of things in themselves or between appearances considered collectively and things in themselves considered collectively, rather than a distinction between individual appearances and individual things in themselves, can be regarded as falling into this category as well.²¹ There is one world, one reality, which (considered) as it appears to us is a collection of appearances and

¹⁹ See Van Cleve 1999, 150: “If there is a sense in which I believe in one world only, it is not a world containing objects that are things in themselves or appearances depending on how one considers them. It is, instead, a world whose only denizens are things in themselves.” ²⁰ Thanks to Banafsheh Beizaei for pushing me to think more about this kind of version of the two-aspect view. ²¹ This kind of reading is proposed in Matthews 1969, for example, esp. 208–11. It is worth flagging that, of all the versions of the two-aspect view, this version has the least direct support in Kant’s text. As we will see, the twoaspect interpretation does not have a very strong textual basis in general, but most of the alleged proof passages that are commonly cited in support of this interpretation are much more reasonably read as concerned with a relation between individual appearances and individual things in themselves than with a relation between all appearances taken collectively, and all things in themselves that ground them taken collectively.

8



(considered) as it is in itself is a collection of things in themselves, but there may not be a one-to-one mapping between the members of these collections.²² Unfortunately, it is rare to find any detailed proposals in the literature about how this kind of two-aspect view might be spelled out.²³ A promising start would be to question the stipulation that ‘same’ in our first classification question is to be understood as ‘numerically identical,’ and to argue that sameness is a broader notion than numerical identity. While all things that are numerically identical are the same, not all things that are the same are numerically identical. Accordingly, there is room for interpretations of the transcendental distinction that (a) hold that every appearance is the same as a thing in itself or several things in themselves taken collectively, or that the ‘world’ of appearances is the same as the ‘world’ of things in themselves, but (b) remain agnostic about whether, or even deny that every appearance is numerically identical to a thing in itself. On account of their ability to accommodate the claim that appearances and things in themselves are the same things or aspects of the same things, so the proposal goes, interpretations of this kind should still count as two-aspect interpretations. This is another point in the debate where we must be careful not to get mired in a mostly verbal dispute. Everybody, including all multiple-object interpreters of whom I am aware, agrees that appearances are closely related to things in themselves, on Kant’s view.²⁴ But instead of debating whether the relation between appearances and things in themselves is ‘close’ enough to count as a form of sameness, we are better served, it seems to me, by focusing our efforts on trying to specify as precisely as possible what exactly this relation consists in. In order to forestall this kind of verbal dispute, I will thus refrain from opening the floodgates by counting any view as two-aspect on which appearances and things in themselves are the same in some, however tenuous, sense. Having said that, there still are some possible more specific readings of this kind that, arguably, embody a true two-aspect spirit and also provide a helpful contrast to views that are commonly classified as

²² See Allison 2004, 459, note 19: “It is one thing to distinguish between things (taken collectively) as they are for us in virtue of the sensible conditions of human cognition and as they might be for some putative pure understanding, unburdened by such conditions, and quite another to affirm a one-to-one correspondence or isomorphism between the members of the two domains.” Also see Allison 1987, 168; Allison 1996, 12, 15–16. I cannot refrain from complaining that Allison made it quite difficult for his readers to understand that he does not want to ascribe the view to Kant that every appearance is numerically identical to a thing in itself. In his writings, formulations abound that, in the absence of explicit instructions to the contrary, anybody would read as expressing such a numerical identity claim, for example, when he says that “in this case also [in the case of the transcendental distinction], what we have is the distinction between a thing considered in a certain relation, in virtue of which if falls under a certain description, and the same thing considered in abstraction from this relation, and therefore not falling under this description” (Allison 2004, 43), or that “it is rather the object that appears that has two sides, one of which is the way in which it appears (under the conditions of sensibility) and the other the way in which it is thought in itself (independently of these conditions)” (Allison 2004, 62). ²³ A notable exception is Marshall 2013. In her earlier work, Lucy Allais presents her reading of the transcendental distinction in terms that quite clearly express a one-object view. See Allais 2004 and Allais 2007. But in her more recent book, she seems to want to distance herself from the claim that every appearance is numerically identical to a thing in itself, while holding on to the claim that appearances and things in themselves are, in some sense, aspects of the same things. See Allais 2015, 72 83. Unfortunately, Allais does not spell out how exactly we are supposed to think about the things that, according to her, have intrinsic natures and relational properties, so that these things as they appear to us are not numerically identical to these things as they are in themselves. Without any additional help in that department, the reader will most naturally continue to spell out her view as saying that the thing as bearer of intrinsic properties is the thing in itself and the very same thing as bearer of relational properties is the appearance, which, in turn, is difficult not to understand as a one-object view. ²⁴ On many multiple-object readings, this close relation is understood as a relation of grounding: things in themselves are transcendental grounds of appearances.

     ’   

9

two-world, so that it would be reasonable and useful to include them under our official two-aspect classification. First, one could think of the relation between appearances and things in themselves as (what one might call) compositional sameness. Suppose a cabin is composed of logs, or has logs as its parts, that is, the cabin is the mereological sum or fusion of the logs. It is innocent to say that the cabin is numerically identical to the mereological fusion of the logs. It is far from innocent, however, to assert that the cabin is numerically identical to the logs, and many people will want to deny that the latter assertion is defensible. The cabin is one but the logs are many. How could one thing be numerically identical to many things? But one could argue that, while not numerically identical to the logs, the cabin stands in a relation to the logs that bears a striking family resemblance to numerical identity, precisely because the cabin is numerically identical to the logs’ mereological fusion. Call this relation ‘compositional sameness.’ On the imagined reading, appearances and things in themselves are the same things in the sense that every appearance is compositionally the same as several things in themselves, but there is no one-to-one mapping between appearances and the things in themselves that ground them. Second, one could conceive of the relation between appearances and things in themselves as (what one might call) genetic sameness. We will say that an appearance E and a thing in itself D are genetically the same if, and only if, (1) there is a thing T such that (a) E is derived from T by means of an ‘as it appears’ operation, (b) D is derived from T by means of an ‘as it is in itself ’ operation, and (2) E and D share an ontological ingredient. The ontological ingredients of a thing are elements, constituents, or building blocks, that, in proper combination, make up the thing. If, when, and where the thing is, there are its ontological ingredients as well. Intuitively, one can think of the ontological ingredients of a thing as entries in a recipe for making the thing in question in an ontology ‘cookbook,’ so to speak. Possible ontological ingredients include bare particulars, substances, attribute instances, property instances, and various kinds of parts or elements (temporal, spatial, etc.). Appearances and things in themselves as conceived on the one-object view are genetically the same in the sense just defined. But not all views on which appearances and things in themselves are genetically the same are committed to classifying them as numerically identical. A popular kind of interpretation on which things in themselves are things as bearers of mind-independent properties, and appearances are the same things as bearers of mind-dependent properties, can be spelled out as a view on which they turn out to be genetically the same but numerically distinct. One way of doing so would be to work with a thing-conception according to which a thing is a compound of a (nonindividuating) bare particular, that is, a property-less particular, and various properties, and to explicate a thing in itself as a compound of the bare particular and all mindindependent properties of a given thing, and the corresponding appearance as a compound of the same bare particular and the mind-dependent properties of the same thing.²⁵ While the appearance and the thing in itself so conceived are numerically distinct, they are genetically the same in that they are derived from the same thing by means of an ‘as it

²⁵ The bare particular in question must be assumed to be non-individuating, that is, it must be assumed not to be, by itself, what makes the thing in question the individual it is, because otherwise all things that include the bare particular would be numerically identical to the thing and to each other, contrary to our present assumption that the thing in itself and the appearance are not numerically identical.

10



appears’ operation and an ‘as it is in itself ’ operation, respectively, and in that they share an ontological ingredient, namely, the bare particular. The ‘as it appears’ operation in this case consists in deleting all mind-independent properties of the original thing to arrive at the appearance; the ‘as it is in itself ’ operation consists in deleting all mind-dependent properties of the original thing to arrive at the thing in itself.²⁶ On views on which the relation between appearances and things in themselves is conceived as genetic sameness, appearances and things in themselves are the same things in the sense that every appearance is genetically the same as a thing in itself or several things in themselves, but there may be no one-to-one mapping between appearances and the things in themselves that ground them. Third, one could conceive of the relation between appearances and things in themselves as (what one might call) collective sameness. Suppose that there are four swimming clubs in Waterville with ten members each. The swimmers in Waterville taken collectively can be considered as persons or as clubs. The swimmers (considered) as persons are a collection of forty and (considered) as clubs they are a collection of four. Since these collections contain different numbers of elements, their members are not related by a one-to-one mapping, and many people will want to deny that they are numerically identical.²⁷ But one could argue that, while not numerically identical, the collection of swimmers (considered) as persons and the collection of swimmers (considered) as clubs stand in a relation that bears a striking family resemblance to numerical identity, precisely because in both cases we are talking about the very same collection of swimmers, or because both collections share the same basic ontological ingredients. Call this relation ‘collective sameness.’ On the imagined view, appearances and things in themselves are the same things in the sense that the ‘world’ of appearances and the ‘world’ of things in themselves are collectively the same, but there may be no one-to-one mapping between appearances and the things in themselves that ground them. Despite the fact that none of these readings is a one-object view, they all strike me as deserving of the two-aspect label. They have the similarity that, on all of them, either appearances and things in themselves share some ontological ingredients or things in themselves are ontological ingredients of appearances. For future reference, I will say that A and B ontologically overlap if, and only if, (1) A is an ontological ingredient of B, (2) B is in an ontological ingredient of A, or (3) A and B share at least one ontological ingredient. That things in themselves and appearances ontologically overlap on the sketched readings is a large part of the reason why it is reasonable to count them as two-aspect views. Accordingly, I will record as our second classification question the question whether appearances and some things in themselves are the same things in some reasonable sense without being numerically identical, where being the same things in some reasonable sense minimally requires that they ontologically overlap. I will call views that answer this

²⁶ A proposal for how to understand the transcendental distinction along these lines can be found in Marshall 2013, which is recommended for further discussion. ²⁷ Arguably, how many elements a collection contains is part of its identity conditions. So, if collection A and collection B contain different numbers of elements, it directly follows that A and B are numerically distinct, even if they share the same basic ontological ingredients, as in our example (the swimmers). It would not help to protest that, in the case of things in themselves and appearances, what we are considering in two ways is not really a collection of things but just ‘reality,’ which is not individuated or divided into individuals. Indeed, far from helping, it would make matters even worse. If reality is non-individuated goo, the claim that the collection of appearances and the collection of things in themselves are numerically identical because, at bottom, they are the same goo, just ‘cut-up’ in different ways, is not even meaningful because, by assumption, goo is non-individuated.

     ’   

11

question in the affirmative ‘same-things views.’²⁸ Moreover, I will say that an interpretation of Kant’s transcendental distinction is a two-aspect view if, and only if, it classifies appearances and some things in themselves as the same things in some reasonable sense, where the latter minimally requires that they ontologically overlap. So, every two-aspect view is either a one-object view―being numerically identical is one way of being the same things―or a same-things view. Readings that are commonly classified as two-world views deny that appearances and things in themselves ontologically overlap, and thus qualify neither as one-object views nor as same-things views. So, by allowing that some two-aspect views are multiple-object views, I am not blurring the line that divides two-aspect interpretations from interpretations that are commonly classified as two-world. In the following, I will call readings that answer both the first and the second classification question in the negative ‘two-world views.’²⁹ All two aspect-views, by contrast, answer one of these questions in the affirmative.³⁰ Moving on to our third classification question, it asks whether the distinction between appearances and things in themselves is ontological or merely methodological/epistemological. All two-world interpreters answer this question in the affirmative,³¹ but the twoaspect view is compatible with either answer. On a methodological two-aspect reading, the transcendental distinction is understood as a distinction solely between different ways of considering, or different perspectives from which to consider, things. The most wellknown contemporary interpretation of this kind has been developed by Henry Allison, who cashes out the two perspectives in the following terms: appearances are things considered as dependent on the conditions of human sensibility, things in themselves are things considered as independent of these conditions.³² As already mentioned, Allison insists that Kant’s transcendental idealism, in particular, his claim that space and time are ideal and do not pertain to things in themselves, does not constitute an ontological position but provides a radical alternative to ontology.³³ According to a similar proposal in the methodological two-aspect tradition, the relevant contrast in the transcendental distinction is between considering things from the human perspective, on the one hand, and from

²⁸ I am leaving the question open whether there are any other possible same-things views, in addition to the three possible versions just sketched. I am also leaving open what else might be required, in addition to ontologically overlapping, for appearances and things in themselves to count as the same things in some reasonable sense. ²⁹ Versions of a two-world view have been defended by Sellars 1968; Aquila 1979; Aquila 1983, ch. 4; Van Cleve 1999, ch. 1, esp. sects D and E. Also see Walker 2010 and Stang 2014. Earlier two-world interpretations will be referenced in section 1.4. ³⁰ What about views on which appearances and things in themselves are derived from the same things by means of ‘as it appears’ and ‘as it is in itself ’ operations but do not ontologically overlap and thus are not the same things in any reasonable sense? According to the proposed classification scheme, they count as two-world views— and well they should. Even on the classic two-world view, according to which appearances are fully minddependent and things in themselves are mind-independent grounds of appearances, one could legitimately say that appearances and things in themselves are derived from the same things by means of ‘as it appears’ and ‘as it is in itself ’ operations. The relevant ‘as it appears’ operation could be defined as the affection of sensibility by the thing in question followed by the processing of the sensible material provided through these affections by our cognitive machinery. The relevant ‘as it is in itself ’ operation could be defined as leaving the thing in question as it is. ³¹ Perhaps Van Cleve would want to take exception? In my assessment, even on his qualified two-world view the distinction between appearances and things in themselves is ontological. Whether logical constructions out of perceptual states are existents or not, it seems undeniable that they are ontologically distinct from things in themselves. ³² See Allison 1983; Allison 1987; Allison 1996; Allison 2004; Allison 2006; Allison2007. Other proponents of a methodological two-aspect view are Herring 1953, 77, 80–2; and Prauss 1974. Also see Bird 1962, 28 9. ³³ See Allison 2004, 2, 98.

12



God’s perspective, on the other.³⁴ Proponents of ontological one-object readings like to use formulations similar to those favored by their methodological counterparts—in particular, expressions such as ‘thing as . . . ’ or ‘object qua . . . ’ but they emphasize that the transcendental distinction tracks an ontological distinction between different aspects of the same thing.³⁵ On most readings of this stripe, the relevant aspects consist in different kinds of properties, as for instance, relational and intrinsic properties,³⁶ mind-dependent and mind-independent properties,³⁷ and, here, more specifically, between secondary and primary qualities (where the former are often understood as specific kinds of dispositional properties),³⁸ or properties that are knowable by us and properties that are not knowable by us.³⁹ Appearances, then, are said to be things qua bearers of relational or minddependent or knowable properties, while things in themselves are identified with the very same things qua bearers of intrinsic or mind-independent or unknowable properties. Or similarly, relational, mind-dependent, or knowable properties are attributed to things as they appear to us, while intrinsic, mind-independent, or unknowable properties are attributed to the very same things as they are in themselves. A fourth classification question, which does not lead us far from the preceding, concerns the status of empirical objects, the objects of our experience, such as tables, trees, cats, and bodies in general. How exactly are empirical objects related to things in themselves, and how exactly are they related to appearances? To my mind, answering this question is ³⁴ See Adickes 1924, esp. sect. II; Matthews 1969, 208–14; Robinson 1994, sects 4–5; also see Allison 2004, 16 17. It should be noted that, although Robinson himself does not say so, in the end his two-perspective view might be classified more aptly as a version of fictionalism (to be discussed below), given that, on his view, “the divine perspective lies behind the human one, as a perspective we conjure up to meet human needs” (Robinson 1994, 439). ³⁵ I should add that not all of the one-object interpretations in the literature are as clear on this point as one could wish. Some of them leave it largely indeterminate whether the distinction between ways of considering objects is underwritten by a distinction between ontologically different aspects in the objects. ³⁶ See Hanna 2006, 195–6, 425–33. If Allais accepted that the appearance and the thing in itself are numerically identical—as opposed to merely being aspects of the same thing, in some sense of ‘same’ that falls short of numerical identity—the view articulated in her book would be an example as well; see Allais, 2015, 207–68. The indicated version of the one-object view is also discussed by both Langton and Van Cleve, but neither one endorses it. See Langton 1998, 20; Van Cleve 1999, 150–5. Langton says about the proposal to understand things in themselves as substances qua bearers of intrinsic properties and appearances as substances qua bearers of relational properties that it “would be basically right,” but adds that “it would also be potentially misleading, since it would bring with it a temptation to think that substances are somehow in the phenomenal world” (Langton 1998, 20). I find her acknowledgment that the proposal would be basically right a bit puzzling. Langton identifies appearances with relational properties of substances or with whatever things are constituted by those relational properties, and things in themselves with substances that have intrinsic properties. This characterization is quite different from saying that appearances are substances qua bearers of relational properties and things in themselves are substances qua bearers of intrinsic properties. The former expresses a multiple-object view; the latter expresses a one-object view (at least on the most natural way of construing it); and not both of them can be right as interpretations of Kant—unless the idea is that the text leaves room for both readings, which would be a reasonable thing to say. Heimsoeth seems to endorse the view that things in themselves are substances qua bearers of intrinsic properties, while appearances are substances qua bearers of relational properties. See Heimsoeth 1924, esp. 124–36. (I say that he seems to endorse it because he is less explicit than one might wish about whether, on his view, appearances are to be identified with things considered as they relate to us or with the relations of things. In the latter case, Heimsoeth’s view would be identical to Langton’s, and he would count as a multiple-object theorist after all.) ³⁷ See Westphal 1997, sect. IX; Setiya 2004; Westphal 2004, 56–61. ³⁸ See Collins 1999, 11–17; Allais 2004; Friebe 2007; Allais 2007; Rosefeldt 2007; Allais 2015, 101–44. Also see Roche 2013. ³⁹ See Rogerson 1999, 10–16. Obviously, the indicated various possible ways of conceiving of the two kinds of properties that mark the difference between appearances and things in themselves are not mutually exclusive. For example, one could hold that appearances are things considered as bearers of properties that are relational, knowable, and mind-dependent, while things in themselves are things considered as bearers of properties that are intrinsic, unknowable, and mind-independent.

     ’   

13

essential for an adequate account of Kant’s ontology, but, unfortunately, not all available readings achieve a satisfactory degree of clarity on this point. The classic two-world view, according to which appearances are mind-dependent and things in themselves are mindindependent grounds of appearances, does give a clear answer: empirical objects are appearances. Most proponents of the two-aspect interpretation talk about empirical objects as if they are the things that are considered in the relevant two ways or have the relevant two aspects, that is, the in-itself aspect and the appearance-to-us aspect.⁴⁰ One might argue, however, that the option of identifying them with things as they appear to us, or the appearance aspects of things, is favored by the text. A fifth classification question is whether Kant is some kind of idealist with respect to empirical objects, and if he is, how exactly this idealism is to be understood, and what exactly he takes its scope to be. As just indicated, the classic two-world view characterizes appearances as mind-dependent and empirical objects as appearances. More specifically, on most variants of this view, appearances are understood to be fully mind-dependent in the sense that they do not have any mind-independent determinations, aspects, features, or ontological ingredients. That is, Kant is portrayed as endorsing a robust form of idealism about empirical objects. By contrast, on a Langton-style two-world interpretation, on which things in themselves are substances with intrinsic properties and appearances are the relational properties of these substances, or whatever things are constituted by these relational properties, there is nothing specifically idealist about Kant’s ontology at all. Unless some further specifications are added to the effect that all properties that pertain to appearances are not only relational but also mind-dependent in some way, we are left with a reading on which Kant’s ontological position does not amount to any form of idealism. There are also both idealist and realist versions of the two-aspect view. A two-aspect view on which appearances are identified with substances qua bearers of mind-independent relational properties is committed to a form of realism. But a two-aspect view on which appearances are identified with substances qua bearers of mind-dependent properties incorporates a form of idealism with respect to empirical objects, whose strength depends on whether empirical objects are identified with the things that have the two aspects or with their appearance aspects. Even Allison, who insists on the strictly methodological character of the transcendental distinction, claims that his interpretation presents Kant as a kind of idealist in that, on his reading, the concept of an object is relativized to human cognition and the conditions of its representation, in particular, to its sensible conditions. According to Allison, this means that things that transcend our epistemic conditions cannot count as objects for us, that things have spatiotemporal properties only insofar as ⁴⁰ See Allison 1987, 155: “ . . . a number of commentators, myself included, have advanced various versions of what has been called the ‘two aspect’ view. Proponents of this view contend that Kant’s transcendental distinction is between the ways in which things (empirical objects) can be ‘considered’ at the metalevel of philosophical reflection rather than between the kinds of things that are considered in such reflection.” Also see Allison 2004, 51 (quoted in note 126, chapter 5). Actually, Allison’s position seems to be somewhat unstable on this point. At a later stage of his discussion, he suggests that “Kant’s best, perhaps his only answer to this question [the question of what the thing is that is considered in two ways] is that it is the transcendental object that is considered from two points of view” (Allison 2004, 62). But a little further down on the very same page we are also being told that “the distinction between the empirical and the transcendental object, like that between things as they appear and as they are in themselves, is not between two ontologically distinct entities but between two perspectives from which ordinary empirical objects might be considered” (Allison 2004, 62). So, at that point it looks again as if the things that are considered from two points of view are ordinary empirical objects, while the transcendental object corresponds to one of these points of view. Also see Allais 2004; Allais 2007; Rosefeldt 2007.

14



they are considered in their epistemic relation to the human mind and its form of sensibility, and that objects that are represented as spatiotemporal entities cannot be said to exist mind-independently in the way in which they are represented to us.⁴¹ Finally, there are also two-aspect readings that ascribe to Kant a robust kind of idealism about empirical objects similar to the classic two-world view. For example, as I understand him, the version of the two-aspect view defended by Gerold Prauss, who is standardly celebrated as one of the ‘heroes’ of the two-aspect tradition, falls into this category, which seems to have gone largely unnoticed by the celebrants. For Prauss, ‘thing in itself ’ and ‘appearance’ are concepts belonging to a second level of transcendental reflection. These concepts allow us to express the results of the first level of transcendental reflection, namely, that (a) empirical objects depend for their existence on the subject who interprets or projects (erdeutet) them, and (b) this projection includes an objective side, namely, the transcendental object, which Prauss understands, roughly, as a projected ground that the subject represents as distinct from itself. When considering an empirical object as it appears to us we focus on the fact that it depends on us; when considering an empirical object in itself we focus on the fact that it is represented as distinct from us.⁴² A sixth and final classification question that helps to introduce some order into the unruly multitude of interpretations of Kant’s critical idealism is whether Kant is committed to the actual existence of things in themselves or whether he regards them as some type of useful fictional objects.⁴³ The classic two-world view and most contemporary two-aspect views present Kant as a realist about things in themselves. There are various versions of the fictionalist reading that differ with respect to how they characterize the fictional nature of things in themselves. A weaker version would say that the concept ‘thing in itself ’ is a theoretical construct of which we avail ourselves in our philosophical efforts to make sense of the world and our experience, but of which we do not know whether there is anything actually existing and extra-mental that corresponds to it. Stronger versions of the fictionalist reading result from modifying the latter claim to say that we are ignorant not only about the actual existence of things in themselves but even about whether they are really possible or, still more strongly, that we know that there are no things in themselves or even that we know that they are impossible.⁴⁴ Different versions of the fictionalist reading can also vary with respect to how they spell out the fiction that determines the nature and properties of the feigned things in themselves. For example, the relevant fiction could

⁴¹ See Allison 2004, 12, 36; Allison 2006, 11–12. ⁴² See Prauss 1974, 62–191, esp. 106–46. It is almost commonplace for two-aspect commentators to give a deferential nod to Prauss, claim him as a defender of the two-aspect view, and dutifully praise his ground-breaking textual analysis of Kant’s use of the qualifier ‘in itself,’ which, Prauss argues, should not be understood as adnominal to ‘thing’ but as adverbial to an often merely implicit ‘considered,’ an analysis, I might add, that occurs in the very first section of his monograph. But almost nobody ever talks about the (very interesting) rest of the book. As far as I can see, the reading developed by Prauss is reminiscent of Fichte and stands in stark contrast to the realism about empirical objects that is intended by most contemporary defenders of the two-aspect view— which might be part of the explanation for the telling silence about pretty much everything that Prauss says after page 23. Although I do not think that Prauss’s interpretation is defensible as an account of Kant’s transcendental distinction between things in themselves and appearances, I agree with much of it if it is understood as an account of Kant’s theory of the constitution of empirical objects, on which more in sections 2.5–2.8 and 3.5. ⁴³ Versions of a fictionalist interpretation have been advocated or sympathetically discussed by Bird 1962, ch. 2; Schaper 1966; Melnick 1973; Rescher 1981; and Pogge 1991. For earlier examples of fictionalist readings, see notes 62 and 68. ⁴⁴ For example, Hans Vaihinger defends a version of the strongest type of fictionalist reading; see Vaihinger 1918.

     

15

characterize things in themselves in classic two-world fashion as supersensible, mindindependent entities that are distinct from and grounds of appearances, or it could describe them in two-aspect fashion as the mind-independent aspects of things.⁴⁵ It should also be noted that on readings that endorse both (a) one of the stronger kinds of fictionalism, according to which the existence of things in themselves is not only unknown or unknowable but denied or declared to be impossible, and (b) a conception of appearances as fully mind-dependent, Kant turns out to be an absolute idealist, as one might put it, in the sense of holding that the subject by itself—that is, without being affected by mindindependent, external things—brings about all of its representations and, with them, the empirical world.⁴⁶ It should also be explicitly acknowledged that, in addition to mixing and matching twoaspect or two-world, methodological or ontological, realist or idealist (about appearances), and realist or fictionalist (about things in themselves) elements in one’s interpretation of Kant’s transcendental distinction, one could introduce a further level of mixing by adopting different mixtures for different kinds of entities, yielding even more interpretative options. For example, one could be a multiple-object theorist with respect to empirical objects or bodies but a one-object theorist with respect to the self. Or one could be a realist about things in themselves that appear to us as empirical objects but a fictionalist about things in themselves that do not appear to us, such as God.⁴⁷

1.3 Putting My Reading on the Map It is not my ambition to provide a detailed discussion of the virtues and vices of the various different kinds of interpretations of Kant’s conception of the relation between appearances, things in themselves, and empirical objects, introduced in section 1.2. To be sure, we will have occasion to talk about a number of problems for and objections to many of them, but my primary objective is to present and defend my own interpretation.⁴⁸ To put my cards on the table, here are the answers to our six classification questions according to my reading: (1) Appearances and things in themselves are not numerically identical. More

⁴⁵ Prauss’s reading, as just described, is an example of the latter type of view. ⁴⁶ All strong versions of fictionalism about things in themselves in the literature that I am aware of are of this kind. Given that for Kant things in themselves are quite clearly ontologically more fundamental than appearances, it would be quite awkward, to say the least, to hold both that things in themselves are fictional objects and that appearances are mind-independent. ⁴⁷ Robert Adams proposes a mixed view that comprises a two-aspect view with respect to ourselves, as agents or minds, agnosticism or a tentative two-world view with respect to bodies, and the acknowledgment that God is in a special category, being only a noumenon. See Adams 1997, 821–5. ⁴⁸ There is no shortage in literature on critical discussions. For Allison’s version of the two-aspect view, see Aquila 1979; Aquila 1983, ch. 4; Guyer 1987, 336–44; Guyer 1989; Robinson 1994, 419–28; Van Cleve 1999, 146–50; Ameriks 2003c; Guyer 2007; Wood 2007. For Langton’s interpretation, see Falkenstein 2001; Ameriks 2003d; Breitenbach 2004; Allais 2006. For Allais’s interpretation, see Roche 2011; Schulting 2011, 9–16; Walker 2016; Guyer 2016. For one-object views in general, see Walker 2010, and Stang 2014. It is also worth flagging up front that my critical engagement with the two-aspect interpretation on systematic grounds in the following chapters will be tilted heavily toward ontological versions of this view. As I see it, the textual evidence conclusively establishes that the transcendental distinction is, and is intended by Kant to be an ontological distinction. Also, I find myself sharing the sentiment expressed by James Van Cleve that the methodological two-aspect view is either “unfathomably mysterious,” namely, if it maintains that the properties of things truly vary according to how we consider them, or disappointingly modest, namely, if it maintains that considering things as they are in themselves amounts to no more than abstracting from some of their properties. See Van Cleve 1999, 8.

16



specifically, they are distinct existents. Also, both things in themselves and appearances are things, albeit not in exactly the same sense. (2) Appearances and things in themselves do not ontologically overlap and thus are not the same things in any reasonable sense. Still, they are closely related: appearances are grounded in things in themselves. (3) The transcendental distinction is an ontological distinction. (4) Empirical objects are appearances. (5) Appearances and, hence, empirical objects are fully mind-dependent. That is, Kant is a genuine idealist about empirical objects. (6) Things in themselves, which are mind-independent, actually exist. Since my reading of the transcendental distinction answers the first two questions in the negative, it counts as a two-world view according to the characterization spelled out in the previous section. And since it views appearances as fully mind-dependent and things in themselves as mind-independent grounds of appearances and asserts the actual existence of things in themselves, it counts as a version of the classic two-world view. I am happy to call my reading a two-world view, but I want to add two notes of caution about what not to infer from my readiness to embrace this label. First, it should not be taken as an endorsement of the view that the realm of things in themselves and the realm of appearances are worlds in the strict sense. For Kant, a world, strictly speaking, is a unified whole of substances that stand in mutual interactions.⁴⁹ But we have no way of knowing whether all things in themselves mutually interact (or whether they are substances).⁵⁰ Having said that, it seems perfectly acceptable to me to say that the realm of appearances and the realm of things in themselves are two worlds broadly understood. These realms are inhabited by different kinds of entities, are governed by different laws, and are cognitively accessible to us in different ways.⁵¹ Kant himself frequently uses the expressions ‘world of sense’ (‘Sinnenwelt’) or ‘sensible world’ to describe the dwelling place of appearances and contrasts it with the ‘intelligible world’ or the ‘world of the understanding’ (‘Verstandeswelt’) which is populated by things in themselves.⁵² Second, my gladly accepting the two-world label should not be construed as an indication that I regard the line that separates the camp of the two-world interpreters from the camp of the two-aspect interpreters as the most important battle line between all the various readings of Kant’s transcendental distinction. As already noted, although I do agree that our first two classification questions are worthwhile questions to address, I believe that they have been overemphasized in recent decades and present day Kant scholarship would be in better shape if the two-aspect versus two-world debate had not influenced the general ⁴⁹ See V-Met-L1/Pölitz, 28:211–12: “The world is a whole of substances that are in mutual connection and thereby make up a unity, a whole.” Also see MSI, 2:407: “ . . . from this it is supposed to follow how it is possible that multiple substances stand in mutual interactions and therefore belong to the same whole, which is called the world . . . ” ⁵⁰ In fact, it is not even clear whether all appearances mutually interact, on Kant’s view. It is certainly clear that he holds that all outer appearances, that is, all empirical objects, which are in space and time, mutually interact. That is the content of the Third Analogy; see B256/A211. But it is much less clear at all if inner appearances, that is, empirical selves, which are only in time, mutually interact with each other or with outer appearances. But empirical selves are also part of the ‘world’ of appearances. ⁵¹ See KpV, 5:43: “But the law of this autonomy is the moral law, which is also the foundational law of a supersensible nature and of a pure world of the understanding, whose pendant [Gegenbild] is supposed to exist in the world of sense but at the same time without interfering with its laws.” ⁵² See B480/A452: “The world of sense, as the whole of all appearances . . . ” See GMS, 4:453: “The rational being counts itself as intelligence as belonging to the world of the understanding. . . . From the other side, it is also conscious of itself as a part of the world of sense . . . ” See Prol, 4:354. See GMS, 4:451; KPV, 5:132; B566–570/ A538–542.

     

17

conversation to quite the extent that it did. The related point that I want to add now is that the other classification questions are, not only at least as important as the first two but also at least as useful in defining plausible battle lines. In the end, every particular interpretation must stand alone, but if we agree to admit temporary alliances and camps, my preferred allies would be, first, proponents of the reading that the transcendental distinction is an ontological distinction, second, proponents of the reading that Kant is an idealist about appearances and, hence, about empirical objects, and, third, proponents of the reading that Kant is a realist about things in themselves—all of whom include two-aspect commentators among their ranks. Also, if we ever get to a point where all of these battles have been decided and all there is left to fight about is the meaning of ‘thing,’ ‘existent,’ ‘entity,’ or ‘same thing,’ we should probably declare a truce with respect to these large-scale classification questions and either go our merry separate ways or pour even more energy into developing the details of our respective interpretations so as to engage the remaining competitors in precisely circumscribed local battles over specific questions concerning the transcendental distinction such as how exactly to understand the transcendental ideality of appearance and their dependence on things in themselves. My reading is a version of the classic two-world interpretation but it is not a mere rehash of it; it improves upon and goes beyond extant readings of this type. The main improvements consist in a precise and detailed account of the mind-dependence, ontological status, and nature of appearances, as well as of their grounding in things in themselves. The main expansions and additions consist in a novel explanation of how to understand Kant’s empirical realism in relation to his transcendental idealism by reading him as committed to the view that reality comprises different ontological levels, a comprehensive collection of all core theses that define critical idealism, an original account of Kant’s views about space and time, in particular, of his conception of pure space and time as forms of sensibility and their relation to empirical space and time, an explicit story of how to spell out the twoworld view for the case of the self, a careful analysis of the differences between critical idealism and ordinary idealism such as Berkeley’s, and original proposals for how to make sense of Kant’s fictionalist sounding pronouncements and his seeming fondness for the Leibniz-Wolffian conception of things in themselves. Moreover, I will not only offer an interpretation of the meaning of critical idealism but also of Kant’s arguments for it, where, again, my reading in part improves upon and in part goes beyond what is already available in the literature. The improvements include comprehensive and detailed reconstructions of Kant’s arguments for the theses that space and time are nothing but forms of sensibility, and that empirical objects are appearances; the additions and expansions include a patiently worked-out account of the foundational structure of the core theses of critical idealism and the logical relations obtaining between them, as well as novel reconstructions of Kant’s arguments for the theses that things in themselves exist, affect us, and ground empirical objects. The main recommendation for my interpretation of Kant’s critical idealism, I take it, is that it presents a coherent, comprehensive view that makes excellent sense of Kant’s text overall. It has a large amount of direct and indirect textual evidence behind it, is wellintegrated with many other important Kantian doctrines—which we will have occasion to consider in the course of our investigation—and can accommodate pretty much all passages that allegedly support competing readings. All of the other interpretations have some basis in the text—although some certainly have less support than others—but they

18



struggle if confronted with the whole body of relevant textual evidence and with the full range of relevant Kantian teachings. Their proponents are forced to disregard many passages and important elements of Kant’s system or at least read them in a way that is highly strained or outright implausible.⁵³ Granted, there are also two or three passages that my interpretation can accommodate only if we allow that Kant was a little sloppy in his formulation. Kant is only human, and so he is not always as careful and precise as one would wish. But in all of these cases, the assumption of sloppiness is quite plausible, and it is easy to see how he could have ended up expressing himself in the way he does, even though the formulation does not exactly capture what he means.⁵⁴ Overall, my interpretation is a near perfect fit, or so I hope to show. In this context, it is also worth noting that two-aspect formulations are often quicker and less cumbersome than the corresponding two-world formulations and roll much more easily off the tongue since the relation between things and their appearances with which we are familiar from our ordinary lives, such as the relation between a bunch of water droplets and their appearance as a rainbow or between a fat cat and its appearance as a raccoon in a dark yard, lends itself to being understood in two-aspect terms, as we will see in section 5.3. More generally, as James Van Cleve points out citing excerpts from Locke and Berkeley, it is not uncommon for writers in the modern period to use two-aspect language to express what is clearly a two-world view.⁵⁵ So, it would neither be surprising nor unprecedented if,

⁵³ For an example, see note 12, chapter 2. ⁵⁴ The most prominent example of a passage like this is Bxxvi–xxvii, which is one of the main pieces of textual evidence that proponents of the two-aspect view cite in support of their interpretation: “At the same time, the reservation is always made, which must be duly noted, that we must at least be able to think, if not cognize, the very same objects also as things in themselves. For otherwise, the paradoxical sentence would follow that appearance is without anything that appears there. Now if we were to assume that the distinction, made necessary by our Critique, of things, as objects of experience, from the very same things, as things in themselves, were not made at all, the principle of causality, and, thus, the natural mechanism in their determination, would have to be valid for all things in general as efficient causes.” The last sentence can be explained relatively straightforwardly. As comes out clearly in the following sentences, the things that Kant has in mind here are human beings, and the problem that he is thinking about is that without the transcendental distinction we would have no way to save human freedom, given that all appearances, including human bodies, are determined by universal causal laws. As will be explicated in section 5.8, in the case of human beings, two-aspect formulations are apt and legitimate even on a two-world reading of the transcendental distinction because human beings are complex entities that are composed of distinct parts, some of which are appearances and at least one of which is a thing in itself. The more problematic part of the passage is the sentence that “we must at least be able to think, if not cognize, the very same objects also as things in themselves,” where, as the context makes clear, ‘the very same objects’ refers to the objects of experience mentioned in the previous sentence. One way to explain this sentence would be to say that Kant is already thinking about the special case of human beings. Alternatively, one could conclude that Kant’s formulation is a bit sloppy. What he means to say, more precisely, is that for every object of experience we must be able to think a thing in itself that appears as this appearance, that is, that grounds this appearance, “for otherwise the paradoxical sentence would follow that appearance is without something that appears there.” Other passages that are best explained by taking Kant to be engaged in loose talk are passages in which he seemingly identifies things in themselves with the objects of the senses. See Letter to Tieftrunk, December 11, 1797, 12:224: “ . . . we cannot cognize objects of the senses (of outer as well as inner sense) in any other way than merely as they appear to us, not according to what they are in themselves.” Also see MAN, 4:507: “ . . . space is merely the subjective form of our sensibility, under which objects of outer senses appear to us that we do not know as they are in themselves, which appearance we then call matter.” What Kant means here, more precisely, is that we cannot cognize or know the things in themselves that ground the objects of our senses but only their appearances, which are the objects of our senses. As Robert Adams observed (in correspondence), these scattered prima facie problematic passages also do not reveal any systematic reasons in Kant’s philosophy for why the objects of the senses ought to be regarded as numerically identical to things in themselves. Textual evidence in support of a given interpretation that makes contact with systematic reasons for the interpretation is much more weighty than isolated formulations that speak for an interpretation but have no connection to any systematic grounds. ⁵⁵ See Van Cleve 1999, 145: “My point in citing these passages, of course, is to raise the possibility that for Kant, too, double aspect language is a rhetorical device for expressing what is actually a double-object view.”

     -  ’  

19

on a few occasions, we found Kant guilty of the same kind of imprecision. But a few scattered two-aspect formulations do not a two-aspect view make. Another general consideration in favor of my reading, which, for reasons of space, I will not have time to develop here but at least would like to mention, has to do with Kant’s philosophical development. Compared to other readings, my interpretation of Kant’s critical idealism fits much better with the trajectory of his thinking in his pre-critical writings that culminates in the formulation of the critical philosophy. The Inaugural Dissertation from 1770, Kant’s last major publication before the Critique, famously even sports the terminology of two worlds in its title “On the forms and principles of the sensible and the intelligible world.”⁵⁶ Telling the story of Kant’s philosophical development as a guide to a proper understanding of the nature and meaning of his critical idealism is a big project that will have to wait for another day. Accordingly, the present investigation is focused primarily on Kant’s views as presented in his critical writings.

1.4 The Vagaries of the Classic Two-World View’s Fortune In preparation for the more detailed presentation of my version of the classic two-world view in the following chapters, it will be useful for us to briefly reflect on the history of this kind of interpretation and its undeserved marginalization over the course of the second half of the twentieth century. During the first one hundred years or so after the publication of the Critique, the main focus of the debate about how to properly understand Kant’s conception of things in themselves was not on the question of whether he regards things in themselves and appearances as numerically identical or not but on the question of whether he really holds that things in themselves actually exist and affect sensibility and thereby produce sensations. In fact, the latter question was one of the central interpretative questions about Kant’s theoretical philosophy overall on which the philosophical luminaries of the day

⁵⁶ To give credit where credit is due, there are some strands in Kant’s thinking in the late 1790s that sound fictionalist; to be precise, they sound much like Prauss’s two-aspect version of fictionalism (see note 42). In the relevant passages, Kant calls the thing in itself a Gedankending (literally a ‘thought-thing’) and describes its function in a way that is strikingly similar to the way in which he describes the function of the transcendental object in the Critique. So, one might say that at least one version of the fictionalist interpretation has the distinction of fitting well with one of the multiple trajectories of Kant’s philosophical development that may be discerned in his ‘post-critical’ period, as some commentators call it. See OP, 22:37: “Space and time are not apprehensible objects but merely modification of the power of representation in which the concept of a thing in itself is merely a thought-thing (ens rationis) and serves as thing = x in order to represent the object of the intuition by contrast as appearance.” Also see OP, 22:23: “Every representation as appearance is thought as something that is distinct from what the object is in itself (the sensible from an intelligible); but the latter = x is not a special object that exists outside my representation but merely the idea of the abstraction from the sensible which is recognized to be necessary.” Also see OP, 21:4; OP, 22:26–9, 31–2, 36, 37, 42, 46. I should like to add, though, that I am not convinced that Kant actually endorses this version of the fictionalist view in his later years. There are two other options for how to read these and related passages. First, it might be that Kant is merely playing around with a fictionalist conception of things in themselves in an exploratory fashion for the purposes of clarifying his own position vis-à-vis the new bold systems of the German Idealists. Given that the Opus Postumum is a mere collection of largely unedited notes and sketches, this hypothesis is not implausible. Second, it might be that he is only talking about the transcendental object and its role in the constitution of empirical objects and not about things in themselves as they figure in his transcendental distinction. For the transcendental object can be classified as an empirical object as it is in itself, as we will see in section 3.5; also see note 131, chapter 3.

20



focused their attention.⁵⁷ According to one of the most discussed objections to Kant at the time, the claim that things in themselves actually exist and affect sensibility is at best unjustified or at worst inconsistent with the main epistemological result of the Critique that our substantive theoretical cognition is restricted to the empirical realm and, more specifically, with the result that the categories of actuality and causality can be meaningfully applied only to appearances and objects of possible experience. A version of this objection was first raised by Hermann Andreas Pistorius and forcefully reiterated by Friedrich Heinrich Jacobi and Gottlob Ernst ‘Aenesidemus’ Schulze; their early discussions of this objection were enormously influential and set the agenda for much of the subsequent debate.⁵⁸ On the view of many early critics, the problematic doctrine of the actual existence of things in themselves and of their affection of sensibility ought to be removed from the critical philosophy, much like a cancerous tumor, if the patient is supposed to have any chance of survival. What we are left with after such an operation―and appropriate corresponding changes in the rest of the system in order to avoid inconsistencies and contradictions―is a fictionalist view about things in themselves and an idealist view about appearances and empirical objects, which adds up to absolute idealism, in the sense indicated in the previous section.⁵⁹ The affection of sensibility that is said to produce sensations could then be understood either as part of the fiction about things in themselves that we tell ourselves in order to make sense of our experience or as an empirical affection by appearances. (Whether this operation is indeed the panacea it was hoped to be is a question for another time.⁶⁰) During this first phase of the reception of the Critique, the boundaries between interpreting the text, interpreting the ‘spirit,’ and developing revisionist extensions, of Kant’s philosophy, were somewhat fluid. While many early readers regarded the fictionalist conception of things in themselves as a ‘fix’ for Kant’s theoretical philosophy in response to the problems of the existence and affection of things in themselves, there are also many early readers who claim that a version of this conception is, in fact, Kant’s own.⁶¹ Johann ⁵⁷ Other central questions of the early debate concern the relation of Kant’s idealism to skepticism and the question of the ultimate foundation, the ‘first principle,’ of the critical philosophy. ⁵⁸ See Pistorius 1784; Pistorius 1788, 444 6; Jacobi 1787, Appendix “On Transcendental Idealism,” 209 30; Schulze 1792, esp. 259 75, 294 9; 374 82. Also see Schwab 1796, 119, 123–5, 143; Schelling 1836/37, 81–5; Schelling 1842/43, 49 50. For useful discussions of the early debate of the problem of the existence of things in themselves and their affection of sensibility, see Vaihinger 1881, 172–5; Vaihinger 1892, 35–50; Erdmann 1878, chs 2 and 3, 80–128; also see Herring 1953, §§1–5. ⁵⁹ See Jacobi 1787, 229: “The transcendental idealist must have the courage to assert the strongest idealism that has ever been taught and cannot even be afraid of the accusation of speculative egoism, since he cannot possibly remain in his system even if he merely wants to drive this last accusation away from himself.” As already indicated in note 42, there is room in logical space for a position that is fictionalist about things in themselves but realist about appearances in asserting their mind-independent existence. But the text quite clearly disqualifies this position as an interpretation of Kant, and (to my knowledge) nobody in the early scholarly debate advocates it. ⁶⁰ We will talk about some of the relevant issues in section 5.10.1. ⁶¹ After surveying the literature about the Critique published before its second edition, Erdmann concludes that, apart from Schultz and Reinhold, everybody else reads Kant as an absolute idealist. “Without exception, Kant’s inquiries concerning the noumenon and the transcendental object are interpreted idealistically. But it is not only claimed that the denial of things in themselves is the main purpose of the work, it is also everywhere assumed, even if in multiply hidden ways, that this solution is impossible. The opponents find in it direct contradictions, the mediators identify irresolvable difficulties, the followers look for extensions―but all in a realistic sense” (Erdmann 1878, 128). As becomes clear from the final sentence of this passage, Erdmann’s judgment that Kant is read mostly as an absolute idealist should be understood as restricted to those scholars who critically examined the Critique in print during the first few years after its initial publication. The default interpretation among the early Kantians, that is, among those early readers who saw themselves as followers of Kant, was to read him as a realist about things in themselves. Compare in this sense Fichte 1797b, 1:480–1: “The (merely historical) question is the

     -  ’  

21

Gottlieb Fichte’s ‘interpretation’ of Kant is a prominent example here.⁶² I put scare quotes around ‘interpretation’ in the previous sentence because one may very well question whether Fichte’s Wissenschaftslehre only makes explicit what Kant himself had in mind all along, as its author insists.⁶³ Notably, Kant, for one, does not think that Fichte has offered a faithful interpretation of the text or the spirit of the Critique. In a public declaration from 1799, in response to a demand in a literary journal to position himself with respect to Fichte’s philosophy, Kant declares that he regards “Fichte’s Wissenschaftslehre as a completely untenable system” (EF, 12:370). More specifically, with respect to the claim that “the Critique is not to be taken literally with respect to what it explicitly teaches about sensibility,” namely, that sensibility delivers sensations in response to being affected by actually existing, mind-independent, supersensible things in themselves, he protests that “the Critique must, indeed, be understood according to the letter” (EF, 12:371).⁶⁴ Kant sums up the lesson that he has learned from his experience with Fichte―and other supposed allies such as Jacob Sigismund Beck―with the proverb “God save us only from our friends, of our enemies we will beware ourselves” (EF, 12:371).⁶⁵ In contrast to the question about the status of things in themselves as real or fictional, the question about whether things in themselves and appearances are the same things, or following: Did Kant really ground experience, according to its empirical content, in something that is distinct from the I? I know very well that all Kantians understood Kant in this way, with the sole exception of Mr. Beck, whose work, the ‘Standpunkt,’ which is crucial here, appeared after the Wissenschaftslehre. (Note: I do not count Mr. Schelling among Kant’s exegetes . . . ) He is understood in this way by his exegete who recently was approved by him, Mr. Schulz, whom I bring up here because of this circumstance. . . . How Reinhold still reads Kant to this very hour we have just seen.” ⁶² See Fichte 1792; Fichte 1794, 1:174–8, 186; Fichte 1795, esp. 2: 444–5, note; Fichte 1797a, 1:427–9; Fichte 1797b, 2: 480–91; Letter from Fichte to Reinhold, July 4, 1797, Briefwechsel, 1:562–4. Also see Beck 1796a, esp. 23–30; 43–5; Maimon 1790, 203–5, 419–20; Maimon 1794, 351–4, 377–8; Maimon 1797, esp. 191; Jakob 1786, esp. 33, 130–2; Brastberger 1790, 3–7; Brastberger 1792, esp. 397–405. Also see Letter from Tieftrunk to Kant, November 5, 1797, 12:215 17. ⁶³ See, for example, Fichte 1794, 1:186, note. At least, Fichte is aware that one might accuse him of not having understood Kant, which, however, he hastens to add, is “truly not an objection,” because, on his view, “Kant’s writings [are] completely unintelligible for anybody who does not already know what can be contained in them” (Letter to Reinhold, July 4, 1797, Briefwechsel, 1:563). ⁶⁴ Also see Letter to Tieftrunk, April 5, 1798, 12:241: “What do you think about Mr. Fichte’s general Wissenschaftslehre? . . . For the moment I do not have leisure to take it up; but the review of Fichte (which is written with a lot of sympathy on the part of the reviewer) looks to me like a kind of ghost, which [sic], once one believes to have grasped it, one finds no object in front of oneself but always only oneself and here also only the hand that grasps for it.” ⁶⁵ Beck, a former student of Kant’s, started out as a faithful follower of the critical philosophy and first made a name for himself as its able expositor; see Beck 1793 and 1794. (Kant himself had invited and encouraged Beck to write a book explicating his critical works; see Letter to Beck, September 27, 1791, 11:289–92; Letter to Beck, November 2, 1791, 11:304–5). Kant was even initially sympathetic to Beck’s Standpunkt-project, in which he wanted to present the critical philosophy in a different way and from a different starting point compared to Kant’s own presentation in the Critique, namely, from the standpoint of the original synthetic unity of apperception; see Letter from Beck to Kant, June 17, 1794, 11:509–11; Letter to Beck, July 1, 1794, 11:514–16; Letter from Beck to Kant, September 16, 1794. But the final product turned out to be rather un-Kantian, and Kant expressed the hope that Beck could be moved “to change his standpoint again and rectify it” (Letter to Tieftrunk, July 12, 1797, 12:183). Among other things, Beck ended up denying the existence of things in themselves and claims that sensations are due to affections of sensibility by appearances; see Beck 1796a, esp. 23–30; 43–5, 156–9, 163, 172–5, 247–8, 265–7, 369–70; also see Beck 1796b, 13–14, 44–6, 66–7; Letter from Beck to Kant, June 20, 1797, 12:165–6, 168. [Appearances of closeness to Fichte notwithstanding, Beck assures Kant, as an “honest man,” that he is “infinitely far away from this Fichtean nonsense” (Letter from Beck to Kant, June 24, 1797, 12:174).] Reinhold went through a similar progression away from Kant, from faithful follower and expositor (see Reinhold 1786–1787), to defender, improver, and extender (see Reinhold 1789; Reinhold 1790; Reinhold 1791), and, eventually, serious revisionist à la Fichte (see Reinhold 1798; Reinhold 1799). Reinhold’s Fichte-phase also did not last for very long, however.

22



aspects of the same things, tended not to be explicitly addressed in the first phase of the reception of Kant’s philosophy. There is a prevalence of two-world formulations in the early literature, but in the case of many early writers it remains ultimately unclear whether they read Kant as advocating a two-world or a two-aspect view, an unclarity that is aided by the wide-spread practice to present Kant’s views through close paraphrase and lengthy verbatim quotations.⁶⁶ The question whether Kant should be read as a realist or fictionalist about things in themselves, and thus as a realist or absolute idealist, continued to play a central role during the second phase of the reception of his philosophy, which may be dated, roughly, from about 1865 to the late-1920s. This period saw the rise of neo-Kantianism and the beginning of ‘serious’ Kant scholarship that is based on a close reading of Kant’s texts, mostly in Germany but subsequently and to a lesser extent also in England and other European countries.⁶⁷ Both of these developments can be seen, to some degree, as part of the backlash to the metaphysical excesses of the intervening German Idealist movement. Fictionalism, coupled with an understanding of the affection of sensibility as an empirical affection by appearances, was popular in neo-Kantian circles, although many of these fictionalist readings appear to have been intended, not so much as accurate interpretations of the text but as revisionist reconstructions in order to make Kant consistent with himself and to capture the much invoked ‘true spirit’ of the critical philosophy.⁶⁸ Many of the ‘serious’ Kant scholars, by contrast, made a point of explicitly arguing against the fictionalist interpretation, insisting that Kant’s texts unequivocally commit their author to the view that things in themselves actually exist.⁶⁹ It is also in this group of scholars, with their focus on exegesis and faithful textual analysis, that the question of the sameness or distinctness of ⁶⁶ This also holds, in particular, for the widely read reviews and summaries of Kant’s work by Pistorius, Schultz, Jakob, and Schmid, through which his philosophy became known to a broader philosophical audience. See Pistorius 1784; Schultz 1785; Jakob 1786, esp. 29–36; Schmid 1788, esp. 135–9; 167–70. Also see Schulze, 1801, volume 1, esp. 236–48, 375–410. In his review of Jakob, ever insightful Pistorius notes that Kant’s text leaves it unclear how he conceives of appearances and their relation to things in themselves, see Pistorius 1788, 429–32, 446–7. It should be noted that in his later Examinations of the Kantian Critique, Part II, Schultz at one place provides what sounds like a fairly clear articulation of a two-aspect reading, see Schultz 1792, 279. But the subsequent discussion quickly reverts to the usual level of unclarity regarding the two-world versus two-aspect issue, see in particular Schultz 1792, 286–8. ⁶⁷ It was also during this time, namely, in 1894, that the Academy Edition of Kant’s works was brought under way under the leadership of Wilhelm Dilthey. (The first volume appeared in 1900.) The list of editors reads like a who-is-who of ‘old style’ German Kant scholarship. Apart from Dilthey, the list includes, among others, Erich Adickes, Benno Erdmann, Paul Natorp, Wilhelm Windelband, and Karl Vorländer. An early British Kant interpreter who followed in their interpretative footsteps is Norman Kemp Smith, see Kemp Smith 1918. ⁶⁸ See Liebmann 1865, 25–69, 205–15; Lange 1866, esp. 267–9; Cohen 1871, 15–17; 239–53; Cohen 1885, 107–8; 501–26, 548–50, 605–16; Drobisch 1885, esp. 1–16; Natorp 1912; Bauch 1917, esp. 163–6. Windelband argues that Kant’s conception of things in themselves went through several phases, the two main phases being a fictionalist conception and a two-world conception, where the former is philosophically most important, but the latter is the one most central to Kant’s own presentation of his views; see Windelband 1877. For a brief overview of the debate about the problems of the existence of, and affection through things in themselves in neo-Kantianism, see Vaihinger 1892, 50–1. ⁶⁹ This group includes Ueberweg 1872, 195–6, 199; Riehl 1876, 311–15, 432–43; Riehl 1908, 371 2, 571 2; Fischer 1869, 435–45; Hartmann 1875; Erdmann 1878, esp. 41–2; Paulsen 1898, 155–9; Vaihinger 1892, 20–1, 52–3; and Adickes 1924, 4–19, 28–37. (Vaihinger joined the fictionalist camp later on; see Vaihinger 1918.) Of course, not all of the scholars who understand Kant as a realist about things in themselves take this aspect of his view to be unproblematic. In particular, the objection that Kant illegitimately applies the categories of actuality and causality to things in themselves is still widely acknowledged, for example, see Trendelenburg 1862, 159–60. Also, not all of them share the same view with respect to Kant’s doctrine of affection. Vaihinger and Adickes, for example, famously read Kant as being committed to a double affection view, according to which we are affected both transcendentally by things in themselves and empirically by appearances in themselves; see Vaihinger 1892, 52–5; Adickes 1929, 27–94; also see note 145, chapter 2.

     -  ’  

23

things in themselves and appearances first appears on the radar screen as a specific point of possible contention. This question by no means occupied as central a place in the scholarly debate at the time as it does nowadays, but interpreters then were more careful in their formulations concerning the relation between appearances and things in themselves than Kant’s earliest readers had been, and one can find occasional explicit discussions of the issue. Against the background of the now customary characterization of the classic twoworld interpretation as ‘traditional’ and, in English speaking circles, as the reading of Kant that was prevalent before the two-aspect interpretation started to conquer the literature about fifty years ago, it is striking that many early Kant scholars in this second phase of the reception of Kant’s philosophy were proponents of the two-aspect reading.⁷⁰ This is not to deny that the two-world reading had its prominent champions. Kuno Fischer, who was one of the main driving forces behind the ‘back to Kant’ movement that rolled over Germany then, reads Kant as clearly advocating that things in themselves and appearances are numerically distinct and says of the two-aspect view that it is “how the Kantian Philosophy has been mostly understood and could not have been understood more wrongly.”⁷¹ For the last four decades or so, the two-aspect view, and, in particular, the one-object version of the two-aspect view, has been, by far, the most popular interpretation of Kant’s transcendental distinction, especially among Kant scholars working in the so-called analytic tradition. In order to understand how the classic two-world view came to be known as ‘traditional,’ how, in the 1970s, the one-object reading could have been perceived as a breath of fresh air, and how this reading could have risen to the status of dominant majority view, we must take a look at Peter Strawson’s pivotal role in the development of Kant scholarship in the English speaking world. The backlash to German idealism, as well as its British counterpart, that helped to renew interest in closely studying Kant’s texts and to allow neo-Kantianism to emerge in the final decades of the nineteenth and the early decades of the twentieth century, also contributed to the flourishing of logical empiricism in the 1920s and 1930s in Austria, Germany, and England, which significantly influenced the goals and methods of early analytic philosophy in the 1940s and 1950s in England and the US. By mid-century, the philosophical climate in the English speaking philosophical world was as anti-metaphysical as it was anti-idealist, and the study of Kant cannot be described as much more than a niche interest.⁷² The publication of Strawson’s The Bounds of Sense in 1966 marks a significant turning point for the fate of Kant’s philosophy in analytic circles. Strawson deserves great credit for making Kant fit for mainstream analytic philosophical society again, but he famously accomplished this feat by separating transcendental idealism from (what he took to be) the analytical argument of Kant’s theory of experience. Strawson is well aware that Kant’s text demands a two-world reading of the

⁷⁰ See Riehl 1876, 313, 424; Riehl 1908, 405–07; Erdmann 1878, 19, 69, 208; Adickes 1920, 255; Adickes 1924, esp. 20–7; Adickes 1929, 3; Heimsoeth 1924, 125 (but see note 36); Heidegger 1991 (*1928), esp. 31–5. The earliest English language interpretation that is clearly two-aspect of which I am aware is Paton 1936, vol. I, esp. 61–3. Prichard already discusses a version of the two-aspect view as one possible way of reading the transcendental distinction but appears to refrain from identifying it as the view that was intended by Kant himself; see Prichard 1909, 71–100. ⁷¹ Fischer 1869, 439. ⁷² This is not to deny that some good work was being done in the area then, such as the earlier work by Lewis White Beck or W. H. Walsh.

24



transcendental distinction, but he is equally aware that, in 1966, the two-world view would be a hard sell. The doctrine of transcendental idealism, and the associated picture of the receiving and ordering apparatus of the mind producing Nature as we know it out of the unknowable reality of things as they are in themselves, are undoubtedly the chief obstacle(s) to a sympathetic understanding of the Critique. (Strawson 1966, 22)

This assessment, Strawson’s two-world reading of transcendental idealism, and his claim that Kant’s more valuable philosophical insights can, fortunately, be “disentangled” from the doctrine of transcendental idealism,⁷³ inspired legions of Kant scholars to either follow him in setting transcendental idealism to one side,⁷⁴ or engage in the business of saving Kant from his bad reputation by devising interpretations that do not present him as a twoworld theorist after all―as Strawson did with his ‘traditional’ reading―nor as an absolute idealist, which would be at least as bad. The hour of the two-aspect view had struck. As I see it, the subsequent takeover of the interpretative landscape by the two-aspect interpretation, and, in particular, the initial take-over by the methodological version, has more to do with the perceived greater attractiveness of this view as a philosophical position than with its merits as a reading of Kant. It is no accident that the two-aspect view’s transformation into the new orthodoxy in analytic Kant scholarship occurred during a period in which both metaphysics and idealism were largely out of fashion and out of favor in the general philosophical population. I agree with Kant that the only saving he needs is from revisionist commentators who think of themselves as his friends, and I am hopeful that the stepmotherly attitude toward the classic two-world view in much of the literature since Strawson will prove to be merely an expression of the mainstream philosophical tastes of an era that is drawing to a close, rather than a stable consensus that can withstand the test of time.⁷⁵ Another factor that has likely contributed to the lingering unpopularity of the classic two-world view is that most of the versions of this reading that have been singled out for critical discussion in the more recent literature are of a particularly implausible variety and easily refutable.⁷⁶ This is bound to create the impression that the classic two-world view is a ⁷³ Strawson 1966, 42. Also see Strawson 1966, 253–73. ⁷⁴ See Guyer 1987, 335: “One can enter the critical philosophy, or at least the transcendental theory of experience, without the presupposition of the thing in itself, because none of Kant’s arguments for the nonspatiality and nontemporality of things in themselves, certainly none of his arguments from legitimate claims of the transcendental theory of experience, succeeds. Thus one can accept the transcendental theory of experience finally expounded in the analogies of experience and the refutation of idealism without any commitment to dogmatic transcendental idealism.” ⁷⁵ I like to think that the fact that most (all?) more recent two-aspect interpretations are ontological versions of the view and some of them are even multiple-object versions is a sign of the turning tides. ⁷⁶ For example, in a 2004 paper, Allais focuses her attack against the two-world view primarily on what she calls ‘phenomenalism,’ according to which appearances are to be understood à la Berkeley as ideas in the minds of perceivers. Allais takes phenomenalism thus understood to be an integral part of the two-world view, which, accordingly, can be rejected once phenomenalism is refuted. See Allais 2004, 660–5. In a later paper, Allais admits that her objections to phenomenalist readings of Kant only affect certain versions of the two-world view. See Allais 2007, 460. But this does not keep her from continuing to use phenomenalism (as just described), in her book, as the main representative of a genuinely idealist competitor to her own ‘essentially manifest view,’ according to which Kantian appearances are to be understood as things qua bearers of essentially manifest properties, that is, of ways of perceptually appearing to us. See Allais 2015, esp. chs 2, 5, and 6. If the choice were between reading Kant as a phenomenalist as characterized by Allais or as a proponent of the essentially manifest view, I might also go

 

25

dead-end, not worthy of serious consideration, and, hence, deservedly out of favor. It is my aim to rehabilitate the classic two-world interpretation as a viable contender in present day Kant scholarship by offering an interpretation that is, unambiguously and proudly, twoworld, but does not suffer from the defects that are usually presented as decisive blows against this kind of reading. Of course, whether Kant’s critical idealism thus understood can hold its own as a viable philosophical position in the contemporary arena is a question for another day. (I believe that it can.) But I hope that by presenting it in the best possible light―which will bring with it filling in some details that Kant’s own account left a bit vague, as well as thinking through some implications that Kant himself did not explicitly draw―and by bringing out its philosophical depth and richness, I can begin to drum up some interest and whet some appetites.

1.5 Concrete Plan Due to the genuinely systematic nature of Kant’s philosophy, our discussion has to cover a lot of ground. Chapter 2 deals with the nature and ontological status of appearances. We begin by looking at the transcendental distinction. It is shown that things in themselves and appearances are numerically distinct existents whose primary difference consists in that the former are mind-independent while the latter are mind-dependent, in a sense that is explicated in detail. On the proposed reading, the world, understood as the sum total of everything that has reality, comprises several levels of reality, most importantly, a mindindependent, transcendental level, at which things in themselves exist, and a minddependent, empirical level, at which appearances exist. Appearances are identified to be intentional objects of experience. The nature and ontological status of appearances is further investigated by way of an examination of Kant’s account of perception and his theory of experience, including a detailed consideration of the formal and material conditions of experience and of the implications of the mathematical antinomies for the specific flavor of Kant’s idealism about appearances. Chapter 3 examines the core claims of transcendental idealism that empirical objects and empirical selves are appearances and not things in themselves, and that pure space and time are nothing but forms of sensibility. Kant is shown to be a relationalist about empirical space and time in holding that empirical space and time are constituted by the spatial and temporal determinations of empirical objects. Furthermore, it is explicated how Kant can be both a transcendental idealist and an empirical realist about empirical objects, empirical selves, and empirical space and time, and how his idealism differs from transcendental realism, as well as from ordinary idealism such as Berkeley’s. The foundational structure of, and Kant’s arguments for his transcendental idealism and empirical realism are the topic of chapter 4, with special emphasis on the ‘master argument’ in the Transcendental Aesthetic of the Critique for the thesis that pure space and time are transcendentally ideal and nothing but forms of sensibility, which has as one of its central premises the claim that we have an a priori intuition of space and time. Chapter 5 contains a discussion of the theses that things in themselves exist and that

with the latter. But, of course, that is not the only choice we have. One can easily construct much more sophisticated broadly phenomenalist readings of Kant’s conception of appearances than the one Allais considers that do not suffer from the problems that she identifies with respect to phenomenalism as construed by her.

26



they ground appearances by affecting sensibility, which are important elements of Kant’s critical idealist position overall. In the context of the examination of the grounding thesis, the difference between an empirical and a transcendental version of the distinction between things in themselves and appearances is explicated, and the role in Kant’s explanation of critical idealism of the much discussed analogy between secondary qualities and spatiotemporal determinations is analyzed. Moreover, the relation between the transcendental and the empirical self is revisited, and the two-world reading of this relation is confirmed and integrated with an account of Kant’s conception of human beings as composed of various distinct parts, including a body, an empirical self, and a transcendental self. The account of how critical idealism differs from ordinary idealism that was begun in chapter 3 is further refined, and Kant’s arguments for the existence thesis and the grounding thesis with respect to things in themselves are reconstructed and shown to ultimately rely on the assumption that the human mind is essentially finite. Finally, two versions of critical idealism that differ in strength are distinguished—bold critical idealism and timid critical idealism—and it is argued that Kant is a bold critical idealist. Our journey ends, in chapter 6, with an account of how Kant’s apparent endorsement of fictionalism about things in themselves, as well as his apparent endorsement of the Leibniz-Wolffian conception of things in themselves, can be reconciled with the reading that he is a realist about things in themselves as characterized in critical idealism. In this context, the difference between Kantian things in themselves and noumena, that is, objects of a pure understanding, is explained as well. Furthermore, two additional arguments for transcendental idealism that are suggested by Kant are subject to scrutiny, both of which seem odd at first glance since they rely on premises about things in themselves to which he does not appear to be entitled within the framework of the critical philosophy. The appendix is intended to ease comprehension by providing a comprehensive summary of all foundational theses that define bold critical idealism, timid critical idealism, transcendental realism, and ordinary idealism, respectively, as well as a collection of various other theses, definitions of key Kantian terminology, and special terminology discussed and introduced along the way, and a diagram of the foundational structure of bold critical idealism and the main arguments on which it depends, as reconstructed in the previous chapters.

2 The Nature and Ontological Status of Appearances 2.1 The Full Mind-dependence of Appearances and their Distinctness from Things in Themselves Appearances and things in themselves are distinct existents; more precisely, appearances are mind-dependent existents, while things in themselves are mind-independent existents, and they do not share any ontological ingredients. The project of this chapter is to lay the foundation for explicating and defending this two-world reading of Kant’s transcendental distinction,¹ spell out the relevant sense of mind-dependence and mind-independence, and examine in more detail how exactly Kant conceives of the nature and ontological status of appearances. First, a couple of terminological matters. The term ‘appearance’ can be used in a general and a more specific sense. In the general sense, it means something that is mind-dependent in a sense to be characterized later in this section and in section 2.3. In the more specific sense, it refers to Kantian appearances, that is, appearances as conceived in Kant’s critical idealism. Like all appearances, Kantian appearances are mind-dependent but they also have several further specific features that set them apart from other kinds of appearances, for example, they depend on a special kind of representations that satisfy certain conditions of objectivity, as we will see. The term ‘thing in itself ’ can also be used in a general and a more specific sense. In the general sense, it means something that is mindindependent in a sense to be characterized later in this section and in section 2.3. In the more specific sense, it refers to Kantian things in themselves, that is, things in themselves as conceived in Kant’s critical idealism. Like all things in themselves, Kantian things in themselves are mind-independent but they also have several further specific features that set them apart from things in themselves as conceived by other philosophers. For example, Kantian things in themselves are supersensible and ground Kantian appearances, as we will discuss in chapter 5. Since this chapter is devoted to explicating more fully how Kant conceives of appearances, in the following I will not add the explicit qualification ‘Kantian’ whenever I use the term ‘appearance,’ but I should be understood as having Kantian appearances in mind unless otherwise indicated. This is also how Kant typically uses the term ‘appearance.’ By contrast, until the focus of our investigation shifts to a closer inspection of how Kant thinks about things in themselves, I will use the term ‘thing in itself ’ in the general sense. Kant himself often uses this term in this sense, for example, when he describes the difference between transcendental idealism and transcendental

¹ This explication and defense will be continued throughout the following chapters. The World According to Kant: Appearances and Things in Themselves in Critical Idealism. Anja Jauernig, Oxford University Press (2021). © Anja Jauernig. DOI: 10.1093/oso/9780199695386.003.0002

28

      

realism, which consists in that the latter affirms, while the former denies, that empirical objects are things in themselves.² The literal meaning of the word ‘entity’ is ‘being,’ which, in turn, is often used interchangeably with ‘existent.’ But in order to properly capture the nuances of Kant’s ontological views, it will prove to be useful to distinguish between entities, or beings, and being, or having reality, on the hand, and existents and existing, on the other hand. Anything that exists also has being, or reality, and, thus, is a being, or an entity. But not everything that has being also exists. Existing is one way of being, or having reality, but it is not the only way. For example, the hobbit Frodo, arguably, has being and, thus, is an entity in the sense just explicated—he is not nothing, he is different from, say, Sherlock Holmes, and he can causally affect us by, for example, making us laugh—but, as a fictional character, he does not exist and, thus, is not an existent. There are several reasons that speak for the reading that Kant regards appearances not just as beings, or entities, but as existents. First, he frequently states that appearances exist and talks about their existence.³ Second, he classifies appearances as empirically real and describes himself as an empirical realist about appearances.⁴ There is much to say about what exactly Kant means by these expressions—and we will talk about that in chapter 3—but, unless he uses them in a highly unusual sense, it is plausible to assume that by calling himself a kind of realist about appearances he is, at the very least, committing himself to their existence. Third, one of the categories that, in the Transcendental Analytic of the Critique, Kant deduces to be a priori valid for appearances is the category of actuality (Wirklichkeit).⁵ And it seems reasonable to hold that unless something exists it cannot be actual. There are also several reasons that speak for the reading that Kant regards things in themselves as existents. First, he frequently states that they exist.⁶ Second, he holds that things in themselves are ontologically more fundamental than appearances.⁷ Since existents are ontologically more fundamental than non-existents, the status of appearances as existents means that things in themselves are existents as well. For all of these reasons, we can safely conclude that Kant regards both things in themselves and appearances not only as beings or entities but as existents. But he does not regard them as the same kind of existents. They do not share the same mode of being. Things in themselves are mind-independent existents, appearances are mind-dependent existents. Appearances are, as Kant frequently puts it, “mere representations” or “nothing apart from representations.” As is customary, I use ‘representation’ to translate Kant’s ‘Vorstellung,’ which he employs as a somewhat loose cover term for all sorts of vehicles of content in the mind that bear, express, or represent some kind of content, including mental states such as sensations, intuitions, and thoughts, as well as more complex conceptions such as scientific or philosophical theories.⁸ And it should ² Obviously, Kant does not want to ascribe the view to the transcendental realist that empirical objects are Kantian things in themselves. ³ See B199/A160; B218–265/A177–218. ⁴ See A370; B44/A28; B52–54/A35–37. ⁵ See B266/A218; B272–274/A225–226. ⁶ See Prol, 4:314–315; GMS, 4:451. More passages will be provided later on in this section and in section 5.1; also see note 56, chapter 6. ⁷ This comes out, in particular, in his characterization of things in themselves as grounds of appearances. See B428; Prol, 4:345; KU, 5:345. Also see note 2, chapter 5. ⁸ See the famous Stufenleiter at B376–377/A320. In its multi-purpose nature, the expression ‘representation’ in Kant is comparable to ‘perception’ in Leibniz and ‘idea’ in Locke, although, arguably, Kant’s ‘Vorstellung’ is even more comprehensive than these other two expressions as used by Leibniz and Locke.

  -  

29

go without saying that, on Kant’s view, representations and all constituent parts of representations are internal to the mind and, in that sense, mental.⁹ I will have more to say about what exactly it means for appearances to be mere representations in section 2.2, and I will provide a precise definition of the relevant sense of mind-dependence in section 2.3. For now, I want to start by highlighting two important features of the minddependence of appearances that can be easily gleaned from the text right away. First, as anticipated above, the mind-dependence of appearances extends to their existence and being. Things in themselves and appearances differ with respect to their mode of being and kind of existence. The being and existence of things in themselves is mind-independent. The being and existence of appearances is mind-dependent; or, as Kant often says, appearances exist only “in us” or “in our representations.” In this context, it is worth pointing out—not least for the benefit of proponents of the two-aspect view who want to read the phrase ‘in itself ’ in the expression ‘thing in itself ’ as elliptical for ‘considered in itself ’ that when describing the difference between things in themselves and appearances, Kant not infrequently uses ‘in itself ’ adverbially to ‘exists,’ ‘subsists,’ and ‘is real.’¹⁰ Things in themselves are things that exists in themselves—an sich existierende Dinge—while appearances are things that exist “in representations,” or “in us.”¹¹ The methodological version of the two-aspect view, in particular, has no way of accounting for Kant’s distinction of appearances and things in themselves in terms of their ways of existing. A difference in the way of existing indisputably is an ontological difference. The second important feature of the mind-dependence of appearances that deserves to be highlighted as we embark on our journey is that it is all-encompassing, so to speak. Given that appearances are said to be nothing apart from representations, and representations are in the mind, appearances are mind-dependent through and through, or fully mind-dependent. Ontologically speaking, there is nothing in them or about them that is ⁹ He consistently describes representations as being in us. See A378: “If we regard outer objects as things in themselves, it is completely impossible to comprehend how we should attain cognition of their reality outside us in that we merely rely on the representation, which is in us.” See KU, 5:359: “ . . . the representation of things, since it is something in us, could be thought as proper and useful with respect to the inner purposive harmony of our faculties of cognition . . . ” See V-Met-K2/Heinze, 28:759: “All representations are something in us, and we cannot say that they are objects of outer senses.” Also see V-Lo/Blomberg, 29:40; KU 5:204. Also see Letter to Beck, December 4, 1792, 11:395: “For representation means a determination in us that we relate to something else (whose spot in us it occupies by proxy, as it were.)” If the non-mental object to which the representation is related were a constituent of the representation, it would not make any sense to say that the representation acts as a kind of proxy in us for the object. Kant also frequently describes representations as “modifications of the mind.” See A97; B242/A197. A modification of the mind cannot be partly constituted by something that is not mental, just as a modification of the body cannot be partly constituted by something that is not corporeal. Also see note 12 and the following quotations cited in the main text. ¹⁰ See A380: “But if the psychologist regards appearances as things in themselves, he may include in his doctrine only matter as a materialist or only thinking beings as a spiritualist . . . , or both as for themselves existing things as a dualist.” See B519/A491: “The realist in the transcendental sense turns these modifications of our sensibility into in themselves subsisting things, and thus mere representations into things in themselves.” See Bxx: “ . . . the result of this first assessment of our a priori cognition of reason, namely, that it only applies to appearances, but, by contrast, sets the thing in itself aside as for itself real but uncognized by us.” See Prol, 4:354: “The world of sense is nothing but a chain of appearances that are connected according to general laws, it thus has no subsistence for itself, it is not properly the thing in itself, and thus it is necessarily related to that which contains the ground of this appearance . . . .” Also see many of the quotations collected in the main text below and in the appended notes. This point is also emphasized by Richard Aquila, see Aquila 1983, 89f. ¹¹ The German ‘an sich existieren’ is much less harsh on the ears than the English ‘exist in themselves.’ Also note that by saying that ‘in itself ’ in the phrase ‘thing in itself ’ can be used and often is used by Kant as adverbial to ‘exists’ or ‘is,’ I do not mean to deny that it can also be used and is used by Kant as adverbial to an implicit or explicit ‘considered.’ But that is perfectly compatible with a two-world conception of the transcendental distinction. On occasion, it may be useful to consider things that exist in themselves as they are in themselves.

30

      

not mind-dependent.¹² By contrast, things in themselves are completely independent from us, or fully mind-independent. Turning to the textual evidence, there is a veritable flood of passages, including many passages from the allegedly more ‘realist’ B-edition of the Critique, in which Kant asserts, unambiguously and clearly, that appearances are fully mind-dependent and exist only “in us” or “in our representations,” in contrast to things in themselves, which are said to be completely independent from us. It thus follows that appearances in general are nothing apart from our representations, which is just what we wanted to signify by their transcendental ideality. (B535/A507) The world of sense contains nothing but appearances, but these are mere representations . . . (B591/A563) For laws exist as little in the appearances but only relative to the subject in which the appearances inhere insofar as it has understanding, as appearances do not exist in themselves but only relative to the same being insofar as it has senses. (B164) Thus we wanted to say: that all our intuition is nothing but the representation of appearance . . . and that if we take away the subject or even merely the subjective quality of the senses in general, all quality, all relations of objects in space and time, indeed even space and time themselves, would disappear and can as appearances not exist in themselves but only in us. (B59/A42) . . . if I take away the thinking subject, the whole corporeal world must disappear, which is nothing but appearance in the sensibility of our subject and a kind of representation of it. (A383) . . . space and time together with the appearances in them are nothing that exists in itself and outside of my representations, but [they are] themselves only kinds of representations, and it is obviously contradictory to say that a mere kind of representation exists also outside of our representations. (Prol, 4:341–342)

¹² Allais attempts to accommodate the many passages in which Kant identifies appearances with mere representations (Vorstellungen) on her anti-phenomenalist reading by claiming that ‘Vorstellung’ does not refer to something internal to the mind, or mental. See Allais 2015, 24 5. Never mind that, in many of the passages in question, Kant himself explicitly concludes from the fact that appearances are mere representations that they exist “only in us,” this reading is also highly questionable from a historical point of view. If one wants to ascertain the meaning of a term that is common currency in eighteenth-century German philosophical circles, Christian Wolff ’s German writings are the first place to look. (I do not think it is an exaggeration to say that they did for philosophical German what Luther’s bible translation did for German in general.) Given Kant’s lifelong entanglement with the Leibniz-Wolffian philosophy, those writings are an especially pertinent place to look if one is interested in the ‘default’ meaning of a certain term employed by Kant. Wolff uses ‘Vorstellung’ as a cover term that centrally includes what Leibniz (and Wolff himself in his Latin writings) calls ‘perceptio.’ Vorstellungen are modifications of the soul, which is characterized by one principal power, namely, the power to represent (Vorstellungskraft). See for example Wolf 1729, §§744 755. From the way Wolff describes Vorstellungen, I do not see how it would be possible to avoid the conclusion that he regards them as necessarily internal to the mind, or mental, just as perceptiones are necessarily internal to the mind/soul, or mental, for Leibniz. There is no indication anywhere in Kant’s writings that he deviates from this established usage of the term. In fact, the famous Stufenleiter-passage at B376 377/A320, in which Kant enumerates the different kinds of representation, lists exactly the same items that you would expect to find in such a list if it was prepared by Wolff; Kant even adds the corresponding standard Latin terms in brackets. Also see B61/A44: “By contrast, the representation of a body in intuition contains nothing at all that could pertain to an object in itself but merely the appearance of something and the manner in which we are affected by it . . . ” Also see the second part of note 66, chapter 4.

  -  

31

A thing in itself does not depend on our representations, and thus can be much larger than our representations reach. But appearances are themselves merely representations and their size, i.e., the idea of their generation through progress, cannot be larger than this progress . . . (R5902, 18:379)¹³

There might be readers who are inclined to dismiss the idea of a mind-dependent mode of being or a mind-dependent kind of existence as confused. They might hold that the assumption of different kinds of existence or different modes of being is implausible, and that the characterization that something exists only in representations should be understood as intended to mean that the thing does not really exist and thus cannot possibly be a genuine existent. Everybody is entitled to their own intuitions about what is plausible when it comes to matters of fundamental ontology. But I fear that clinging to the indicated view while attempting to do serious work in the history of philosophy, where the idea of different ways of being can be encountered rather often, will prove to be problematic.¹⁴ The question is not what we may or may not find plausible from our contemporary perspective but what is plausible from the point of view of one’s author and against the background of one’s author’s text. And, as

¹³ Also see B520/A492: “But this space itself, together with this time, and together with both, all appearances, are in themselves not things but nothing but representations and cannot exist outside our mind at all . . . ” B521–522/A493–494: “For that it [an appearance] exists in itself, without relation to our senses and possible experience, could indeed be said, if we were talking about a thing in itself. But we are merely talking about an appearance in space and time, both of which are not determinations of things in themselves but merely of our sensibility; therefore, what is in them, (appearances) are not in themselves something but mere representations, which, if they are not given in us (in perception), are nowhere to be found.” See B523/A494–495: “But the appearances are given, according to it [the transcendental object], not in themselves but merely in this experience, since they are mere representations . . . ”Also see A101: “One soon discovers this if one remembers that appearances are not things in themselves but the mere play of our representations that in the end amount to determinations of inner sense.” See A114: “But if one remembers that this nature in itself is nothing but a sum total [Inbegriff] of appearances, and thus not a thing in itself but merely a multitude of representations of the mind . . . ” See A120: “ . . . without the relation to an at least possible consciousness appearance could never become an object of cognition for us, and thus would be nothing for us and nothing anywhere, since it has in itself no objective reality and exists only in the cognition.” See A127: “ . . . for appearances, as such, cannot take place outside us but exist only in our sensibility.” See A129: “For as appearances they amount to an object that is merely in us, since a mere modification of our sensibility is not at all encountered outside us.” See A369: “The transcendental realist thus regards outer appearances (if one admits their reality) as things in themselves, which exist independently of us and our sensibility, and thus would be outside us also according to pure concepts of the understanding.” See A372: “ . . . those are merely appearances, i.e., mere kinds of representations, that are always only in us and whose reality is grounded in immediate consciousness, just as the consciousness of my own thoughts.” See A387: “But we should take into account that bodies are not objects in themselves that are present to us but a mere appearance of who knows which unknown object, that matter is not the effect of this unknown cause but merely the appearance of its influence on our senses, that, consequently, both are not something outside us but mere representations in us . . . ” See A391: “ . . . that surreptitious dualistic assumption that matter as such is not appearance, i.e., mere representation of the mind to which an unknown object corresponds, but the object in itself such as it exists outside us and independent of all sensibility.” Prol, 4:342: “Thus, the objects of the senses exist only in experience; on the other hand, to also give them an existence that subsists for itself without experience or before it means as much as to imagine that experience is actual without experience or before it.” ¹⁴ One can find the thought that there are different kinds of being in Aristotle, Aquinas, and Leibniz, for example, to name just a few especially prominent figures. According to Aristotle, “being can be said in many ways,” Aquinas holds that God exists in a different sense than creatures, and Leibniz distinguishes between the mind-dependent being of well-founded phenomena and the mind-independent being of monads. See, for example, Aristotle, Metaphysics Γ.2, 1003a33–1005b18; Aquinas 1947, Part I, Query 13, article 5; Leibniz, Letter to Remond, January 10, 1714, G 3:606; Letter to Remond, March 14, 1714, G 3:612. A noteworthy contemporary defense of the fragmentation of being is mounted in McDaniel 2017.

32

      

we have seen, our author clearly takes appearances to be existents, despite their minddependence. Appearances are “mere representations” and exist only “in us,” that is, they are fully mind-dependent, while things in themselves are fully mind-independent. In order for things to be numerically identical or the same in a reasonable sense, as defined in chapter 1, they must ontologically overlap. But something that is fully mind-dependent does not ontologically overlap with something that is fully mind-independent. Therefore, appearances and things in themselves neither are numerically identical nor the same things in any reasonable sense. In a similar vein, if x considered in one way is a fully mind-dependent entity and y considered in another way is a fully mind-independent entity, x and y are not the same thing. No mere change in perspective can transform a fully mind-independent thing into a fully mind-dependent thing. In sum, things in themselves and appearances are not aspects of the same things. Two-aspect readings face several other insurmountable problems, both textual and systematic, but the argument just given by itself is already sufficient to conclusively settle the dispute about whether Kant conceives of the transcendental distinction in two-world or two-aspect terms. No two-aspect reading can do justice to Kant’s full-blooded idealism about appearances. Therefore, the two-aspect interpretation must be rejected.¹⁵ There are many passages in which Kant quite clearly says that appearances and things in themselves are distinct or ontologically non-overlapping, as for instance, when he asserts that “these [appearances] really relate to something that is distinct from them (and thus completely dissimilar), insofar as appearances always presuppose a thing in itself and thus give an indication of it . . . ” (Prol, 4:355); or when he claims that “one has to admit and assume behind the appearances still something else that is not appearance, namely, things in themselves . . . ” (GMS, 4:451), and that one must concede “that behind the appearances there must still be the things in themselves as grounds (although hidden)” (GMS, 4:459); or when he states that “appearances are the only objects that can be given to us immediately” and in the very next sentence adds that “those appearances are not things in themselves but themselves only representations” (A108–109); or when he explains that sensibility is affected by “objects that are in themselves unknown to it and entirely distinct from those appearances” (Prol, 4:318).¹⁶

¹⁵ The other major, equally fatal, systematic problem is that the two-aspect interpretation cannot do justice to Kant’s empirical realism either. We will discuss this failing of the two-aspect view in section 3.5. ¹⁶ Also see the following passage from a letter to Mendelssohn in which Kant explicitly characterizes things in themselves as “other objects” in addition to empirical objects (which are appearances). He says there that his “final conclusion” in the Critique is “that all speculative cognition a priori that is possible for us reaches no farther than to the objects of an experience that is possible for us, only with the qualification that this field of possible experience does not comprise all things in themselves, and thus indeed still leaves over other objects, yes even presupposes them as necessary, without it being possible for us to determinately cognize anything of them.” (Letter to Mendelssohn, August 16, 1783, 10:346). Note in passing: the formulation that “this field of possible experience does not comprise all things in themselves” is somewhat sloppy and sounds a little odd, for it suggests that the field of possible experience comprises some things in themselves, just not all of them. But the context, in particular the following sentence, makes clear that what Kant means is that the field of possible experience does not comprise all things, namely, it does not comprise things in themselves, which are “left over” and “necessarily presupposed.” We find a similarly sloppy formulation in B43/A27: “Since we cannot turn the special conditions of sensibility into conditions of the possibility of the things but merely of their appearances, we can indeed say that space comprises all things that might outwardly appear to us, but not all things in themselves, whether they be intuited or not . . . ” The sloppy bit here is the formulation that “space does not comprise all things in themselves,” which suggests that it comprises some things in themselves. Obviously, what Kant means is that space does not comprise all things, since it comprises only appearances but not things in themselves.

     

33

These are exactly the kind of statements that one would expect from somebody who conceives of the transcendental distinction in two-world terms.

2.2 Appearances are Intentional Objects of Representations Kant distinguishes between two different kinds of appearances. Outer appearances are appearances of things in themselves that affect outer sense; they are in time and space. Inner appearances are due to affections of inner sense by ourselves; they are only in time.¹⁷ Both outer and inner appearances are fully mind-dependent and numerically distinct from the things in themselves that ground them by affecting us, but there are also some differences between them that subtly affect their ontological status. In the interest of clarity, we will examine these two kinds of appearances in turn, starting with and focusing on outer appearances, which are also Kant’s primary focus. In the following, ‘appearance’ is to be understood as ‘outer appearance’ unless otherwise indicated. We will turn to considering inner appearances in section 2.9. On my reading, appearances are intentional objects, which, utilizing scholastic terminology (in a slightly modified sense), may be said to ‘in-exist’ in our representations as part of the content that these representations present or disclose to the cognizer who is entertaining them.¹⁸ The indicated scholastic terminology is probably most familiar to contemporary ears from Franz Brentano.¹⁹ Closer in time to Kant than the scholastics, Descartes’ conception of the ‘objective reality’ of ideas also provides a helpful comparison to the mode of being of intentional objects that I take appearances to have.²⁰ As I see it, that appearances are representation-immanent intentional objects, or that appearances exhibit a special kind of mind-dependence characteristic of intentional objects, is a central part of what Kant expresses with his frequently repeated formulations, cited in the previous section, that appearances exist only “in our representations,” or “in us,” or “in our mind.” It is important to note that the term ‘Vorstellung’ is ambiguous. Kant mainly uses it to refer to what does the representing, that is, the content-bearing vehicle, for ¹⁷ See B37/A22: “By means of outer sense, (a property of our mind), we represent objects to us as outside us and them altogether in space. . . . The inner sense by means of which the mind intuits itself or its inner state does not provide an intuition of the soul itself, as an object; but it is still a determinate form under which alone the intuition of its inner state is possible, so that everything that belongs to the inner determinations is represented in relations of time.” See B43–44/A27–28: “All things, as outer appearances, are next to each other in space . . . ” See B49/A33: “Time is nothing but the form of inner sense, i.e., of the intuiting of ourselves and our inner state.” We will talk about and look at textual evidence for the claim that inner and outer sense are affected by things in themselves in sections 2.4, 2.5, and, in detail, in chapter 5. ¹⁸ Different versions of an intentional object reading of Kantian appearances are defended by Vaihinger 1892, 33–4, 58–72; Sellars 1968, ch. 2; Prauss 1971; Aquila 1983, ch. 4; Pereboom 1988; and Aquila 2003. ¹⁹ See Brentano 1874, vol. 1, 115–16: “Every psychological phenomenon is characterized by what the Scholastics of the Middle Ages called the intentional (or mental) inexistence of an object, and what we would call, albeit not entirely unambiguously, relation to a content, direction toward an object (which is here not to be understood as a reality), or immanent objectivity. Every psychological phenomenon contains something as an object within itself, although not each in the same way. In representation something is represented, in judgment something is acknowledged or rejected, in love loved, in hate hated, in desire desired and so on. This intentional inexistence is exclusively characteristic of psychological phenomena. No physical phenomenon exhibits anything similar. And so we can define psychological phenomena by saying that they are such phenomena that intentionally contain an object in themselves.” ²⁰ See Descartes 1649, AT 7:161: “III. By the ‘objective reality of an idea’ I understand the being of the thing represented by the idea, insofar as it exists in the idea. . . . For whatever we perceive in the objects of ideas exists objectively in these very ideas.” Also see Descartes 1649, AT 7:40.

34

      

example, a perception that represents a tree or a dream that represents a unicorn.²¹ But the term can also be employed to refer to the content that is represented in a certain mental state or conception, for example, a tree or a unicorn.²² When Kant says that “appearances are mere representations,” as in some of the passages just quoted, I want to suggest, he is using the term in the second sense and means to be referring to the intentional objects that are immanent to our mental states and conceptions (in a sense to be explicated later on in this section) and not to these mental states or conceptions themselves.²³ Of course, this does not change anything about the generally speaking mental nature of appearances, but it puts us into the position to allow some daylight, as it were, between appearances, which are mental contents, on the one hand, and particular representation tokens of particular human minds, on the other hand, which will be very useful in the following. Even though you and I cannot have the same particular representation token in our respective minds, under the right kind of circumstances we can represent the same content and thus cognitively relate to the very same intentional object. For example, suppose you and I have qualitatively identical memories of seeing Aachen Cathedral from a particular spot on the square in front of it. Even though the memory token in your mind is distinct from the memory token in my mind, our memories reveal the same content to us, and, thus, have the same intentional object, namely, Aachen Cathedral as seen from a particular vantage point. In order to avoid the indicated representing-versus-represented ambiguity, I will use ‘representation’ exclusively in the first sense, that is, as referring to what does the representing, the content-bearing vehicle. Also, for ease of communication, I will say of a representation R ‘in which’ an intentional object O exists, or thanks to which O has being and all of its determinations and other ontological ingredients, that R ontologically specifies O.²⁴ Similarly, I will say that representation R ontologically specifies a property P of object O if, and only if, O has P in virtue of R presenting O as having P. This is not the place to provide a general theory of intentional objects.²⁵ But there are several sources of possible confusion in this neighborhood that I would like to address directly at this early stage of our investigation. Some of what I am about to say will be fleshed out more fully only in the course of our later discussion, in particular in ²¹ There is another ambiguity here: the content bearing mental vehicle could be conceived of either as some kind of mental ‘item’ in our mind that represents a certain content, for example, the mental image of a tree or a unicorn, or as an act of representing a certain content, for example, the act of perceiving a tree or the act of dreaming a unicorn. Kant himself wavers between the two. ²² This kind of vehicle-content ambiguity is widespread in the early modern period. We also find it in Descartes’ and Berkeley’s use of the term ‘idea’ and in Leibniz’s use of the term ‘perception,’ for example. ²³ See Kant’s official definition of ‘appearance’ at B34/A20: “The undetermined object of an empirical intuition is called appearance.” That is, the appearance is not the empirical intuition itself but the object represented in it. Also see B235 236/A190 191: “But now as soon as I bring up my concept of an object to transcendental significance, the house is not a thing in itself but merely an appearance, i.e., representation whose transcendental objects is unknown; thus, what do I mean by the question how the manifold may be connected in the appearance itself (which after all is nothing in itself)? Here that which lies in the successive apprehension is regarded as representation but the appearance, even though it is nothing but the sum total [Inbegriff] of these representations, is regarded as their object with which my concept is supposed to cohere that I draw from the representations of apprehension.” ²⁴ I will say more about what a determination of an object is supposed to be in the following section. For now, some examples will be sufficient to illustrate the relevant contrast between determinations and other properties of an object that are not determinations. For instance, the properties of my cat of being cat-shaped, furry, and weighing about 15 pounds are among his determinations, while his properties of being self-identical, an appearance, and such that 2+2=4 are not. ²⁵ A lot of valuable material for such a general theory can be found in Alexius Meinong’s classic studies in Gegenstandstheorie (object-theory), see Meinong 1899; Meinong 1904; Meinong 1915.

     

35

sections 3.2–3.3 and 3.5. The expression ‘intentional object of representation R,’ and similarly the expression ‘the content of representation R,’ could be used and often is used in a more general sense to signify whatever object R intentionally relates to or refers to, even if this object is not captured by what we may call the presentational content of R, that is, by the mental content that R presents to the mind of the cognizer who entertains R, in that some of the properties presented in R in this way are not properties of the object or even if the object exists outside of our representations. In this use, ‘intentional object’ is understood in the sense of ‘ultimately intended object’ and ‘content’ is understood in the sense of ‘extra-representational referent’ or ‘extra-representational denotation.’ For example, if Descartes perceives a rainbow, he might say that the intentional object, and content, of his perception, in the sense of ‘ultimately intended object’ or ‘extra-representational referent,’ is a certain collection of water droplets that exists outside of his mind, even though its direct object, its presentational content, is an arch of seven bands of color, which has its being solely in his perception. This is not how the term ‘intentional object’ is used here; in my usage, intentional objects are special presentational contents, which, by nature, are immanent to representations. (As noted, I will say more about the relevant notion of immanence later in this section.) Just as I will use ‘intentional object’ as restricted to certain representation-immanent presentational contents, when I talk about the content of a representation without any further qualification, I will mean its presentational content and not its extra-representational referent or denotation. At the same time, I will not reserve the terms ‘referent’ and ‘denotation’ for extra-representational things only although I will restrict them to existents.²⁶ The terms ‘represent’ and ‘representation’ are related sources of possible confusion. The sentences ‘representation R represents X’ or ‘representation R is a representation of X’ could be used to express either the claim that X is part of R’s presentational content, as just characterized, or the claim that R is intentionally related to X, where X may not be captured by the presentational content of R or may exist outside of our representations. In the service of clarity, whenever the context calls for being specific and leaves room for doubt, I will use ‘R directly represents X’ to express the former sense and ‘R indirectly represents X’ to express that R is intentionally related to X, and X is not captured by the presentational content of R or exists outside of our mind. So, in terms of the terminology adopted here, Descartes should say that his rainbow perception directly represents and has as its intentional object an arch of seven bands of color but indirectly represents a collection of water droplets.²⁷ Similarly, the locutions that X is represented as Y in R, and that R presents X as Y, will be reserved exclusively to express that X, and X’s being Y, is part of R’s presentational content, while the phrase ‘X is represented by R as Y’ is to be understood as applicable both to cases where R directly represents X, and presents X as Y, and to cases where R indirectly represents X, and represents X as Y. ‘Represent’ and ‘representation’ without any added qualifications should be understood as ‘directly or indirectly represent’ ²⁶ So, in my usage, the name ‘Frodo’ is not a referring expression, strictly speaking, for example, since the putative referent, the hobbit Frodo, does not exist. We might say, though, that Frodo is the pseudo-referent of the name ‘Frodo,’ given that, as intentional object of The Lord of the Rings, he has some reality and exists in the ‘world’ of The Lord of the Rings, on which more presently. ²⁷ Obviously, there are different ways in which one could account for the intentional relation of representations to things other than their intentional objects. One possible account would be, for example, that representations are intentionally related to their proximate causes; another possible account would be that representations are intentionally related to all objects that have all of the properties that are represented in them.

36

      

and ‘direct or indirect representation,’ respectively. In much of our subsequent discussion, the context will make clear if a specific kind of representing is at issue and which kind it is; since we will be talking a lot about representations and their intentional objects, for the most part the direct kind of representing will be in the foreground. All intentional objects are presentational contents, but not all presentational contents are intentional objects. One might hold that even internally contradictory representations have some content, for example, the concept of a round square has a different content from the concept of a married bachelor, but there are no intentional objects corresponding to either one of these representations. But even if we restrict our attention to internally consistent representations, not all of their presentational contents are intentional objects either. To be sure, on a liberal understanding of the term ‘object,’ any representation that is not internally contradictory can be said to represent an object. For example, a red sensation may be said to have a certain red quale as its object, or the concept ‘justice’ may be said to have the abstract entity justice as its object. But there is also a more restrictive understanding, on which we have something more specific in mind when we talk about objects, and objects of representations in particular. Minimally, an object in this more specific sense is a particular something that is the bearer of certain determinations. Similarly, an intentional object of a representation, minimally, is presented as (a) a particular something that is distinct from both the representation and the cognizer who entertains the representation and (b) a bearer of certain determinations.²⁸ So, neither the presentational content of a red-sensation nor the presentational content of the concept justice count as intentional objects. As we will see, on Kant’s view, the concept of the transcendental object, which is the concept of a bare particular something = X that is (a) distinct from both our representations of it and the cognizer who entertains these representations and (b) a bearer of determinations, plays an important role in the constitution of appearances. On my reading, this is because the concept of the transcendental object is crucially involved in the constitution of all intentional objects, be it appearances or intentional objects of dreams, hallucinations, illusions, or fictions, to name a few other examples. Kant agrees with the common-sense view that empirical objects, or objects in space and time, persist over time and populate the public physical world that we all share. For ease of communication, I will call this shared public physical world ‘empirical world’ for short from now on. As we will discuss in detail later on, it is a central tenet of Kant’s transcendental idealism that empirical objects are appearances. So, his claim that appearances exist only ‘in our representations’ cannot be understood to imply that all appearances are private and that they ‘pop’ in and out of the empirical world depending on whether anybody currently represents them. My proposal to regard appearances as intentional objects is not to be understood as having these implications either, despite the representation-immanence of all intentional objects. To be sure, there are intentional objects that are private, for example, the oasis that Joe hallucinates just before passing out on his expensive adventure trip through the Sahara desert, the raccoon that Jane takes herself to be perceiving while being confronted with the neighbor’s overweight cat, or the

²⁸ There is one exception: the empirical self is also an intentional object, which, however, is not presented as distinct from the cognizer. On my reading, this difference is reflected in the fact that, in the construction of inner experience, we employ the concept of the transcendental subject instead of the concept of the transcendental object, on which more in sections 2.4, 2.5.1, and 3.5. Also see note 212.

     

37

friendly purple dragon on which I ride to my lectures in my favorite recurring dream. But there are also intentional objects that are public. For example, the hobbit Frodo, although an intentional object of The Lord of the Rings, is a public figure known to everybody who has read Tolkien’s books or watched the corresponding movies. Frodo differs in various ways from appearances, but they have in common that they are public intentional objects. Also, importantly, Frodo does not pop in and out of the world of The Lord of the Rings depending on when somebody reads or thinks about him or watches a movie in which he is a character. When and for how long Frodo exists in the world of The Lord of the Rings only depends on how Tolkien’s books represent him to exist. (This holds even if all copies of Tolkien’s books were to be destroyed someday.) Similarly, when and for how long a given appearance exists in the empirical world, for example, Aachen Cathedral, only depends on how the special kinds of representations that ontologically specify appearances (on which more in the following sections) represent it to exist and not on when and for how long it is actually represented. So, if the right kind of representations represent Aachen Cathedral as persisting in time—which they do—then it does persist in time in the empirical world even if there are periods during which nobody represents it. This is how Kant’s talk about the existence of appearances in our representations and my talk about the representationimmanence of intentional objects are to be understood, that is, not as implying that appearances and intentional objects are necessarily private and exist in the empirical world or in their respective intentional ‘world’ only when and for as long as somebody entertains a representation that represents them but as saying that their existence and the time and duration of their existence in their world are determined by certain representations that represent them in a way to be further specified in sections 2.5–2.8. As the foregoing remarks illustrate, when talking about the existence or non-existence of intentional objects in general, and the existence or non-existence of appearances in particular, it is important to distinguish between their existence in a certain intentional ‘world’ (which, as the case may be, could be quite small), for example, Frodo’s existence in the world of The Lord of the Rings, the purple dragon’s existence in the world of my favorite recurring dream, and the existence of Aachen Cathedral in the empirical world, on the one hand, and their existence or non-existence from a meta-perspective that is not restricted to the inside-view of a certain intentional ‘world’ but encompasses reality in general, on the other hand. I will call this wider meta-perspective ‘the point of view of fundamental ontology.’ When Kant claims that appearances exist ‘only in our representations,’ he is not speaking from the point of view of the empirical world but from the point of view of fundamental ontology. From this fundamental point of view, appearances as well as space and time and other intentional objects, such as Frodo, Joe’s oasis, Jane’s raccoon, and my dragon, are mind-dependent or ideal, in contrast to things in themselves, which are mindindependent and real.²⁹ Even though there are several similarities between appearances and other intentional objects such as fictional characters and the objects of dreams, hallucinations, and illusions, there are also important differences between them. On my reading of Kant, since appearances conform to certain special formal conditions of proper objecthood (to be discussed ²⁹ I am using the formulation that things in themselves are mind-independent and real (instead of ‘or’) because, I take it, saying of a certain object that it is real means more than that it is mind-independent; it also means that it exists. More details will follow in section 3.3.1.

38

      

in section 2.5.2) and are grounded in things in themselves (to be discussed in chapter 5) their way of being is a kind of existence even from the point of view of fundamental ontology, albeit a mind-dependent kind. By contrast, from this fundamental perspective, the way of being of fictional characters and of the intentional objects of dreams, illusions, and hallucinations is not a kind of existence, even though they are also real to some degree, and have a certain ‘amount’ of reality. For ease of communication, in the following, when speaking from the fundamental point of view, I will refer to the way of being of intentional objects that do not exist from this fundamental perspective as ‘pseudo-existence.’ Similarly, for emphasis and in order to avoid possible confusion, I will sometimes refer to the way of being of entities that exist from the point of view of fundamental ontology as ‘genuine existence,’ although, when speaking from this point of view, the addition of ‘genuine’ is redundant, strictly speaking. I submit that Kant—much like Leibniz—is committed to the view that reality comprises multiple ontological levels. The tiered character of reality according to Leibniz’s ontology is widely acknowledged in the literature. Most commentators favor a three-level scheme, which comprises a level of monads or substances, a phenomenal level of bodies or empirical objects, and an ideal level of abstract entities. The ‘amount’ of reality of the entities that exist at these levels decreases as we move up the hierarchy, but even abstract objects are still real to some degree and thus part of reality.³⁰ To my mind, appreciating that Kant, too, is committed to a division of reality into multiple ontological levels is the key to understanding his critical idealism. In Kant’s case, the world, understood as the sum total of everything that has reality, comprises several levels, most importantly (a) a fundamental level, which I will call the ‘transcendental level,’ and (b) a less fundamental level, which I will call the ‘empirical level.’³¹ The transcendental level of reality is the supersensible realm in which things in themselves exist; it can be characterized as a level of mind-independent reality. The empirical level is the sensible or empirical realm in which appearances exist; it can be characterized as a level of mind-dependent reality. All genuine existents exist at either one of these two levels (apart from God who, if he exists, exists at a level of his own, outside of the world). On my reading of Kant, the distinction between the transcendental and empirical level of reality underwrites his crucially important distinction between the transcendentally real and ideal, on the one hand, and the empirically real and ideal, on the other hand. That things in themselves are ³⁰ See McGuire 1976; Garber 1985; Hartz and Cover 1988; Adams 1994, 253–5; Garber 1995. On my reading, the indicated distinction into three levels is, basically, correct, but in order to do full justice to Leibniz, we need to add a few more sub-levels with respect to the phenomenal realm, which, however, is a complication that we need not pursue here. See Jauernig 2010. ³¹ Note that Kant uses ‘transcendental’ in several different senses. In the context of his theory of experience and cognition, the qualification that a certain cognition is transcendental means that it does not concern objects but our manner of cognizing objects a priori, or that it concerns the a priori conditions for the possibility of cognition in general; see B25, B80/A56. With ‘transcendental’ understood in this epistemological sense, broadly speaking, Kant’s description of his idealism as transcendental can be seen as indicating that the idealism in question allows us to account for the possibility of a priori cognitions, and that it is grounded in transcendental considerations, that is, considerations concerning the possibility of a priori cognitions, or the a priori conditions of cognition in general. But Kant also often uses the term ‘transcendental’ in ontological contexts, where it is usually contrasted with ‘empirical,’ as in ‘transcendental idealism’ versus ‘empirical idealism.’ (This ontological sense of the term is also emphasized by Martin: see Martin 1948, 324; Martin 1951, 47 9.) In these kinds of contexts, the qualification ‘transcendental’ indicates that a certain key ontological notion such as ‘object,’ ‘dualism,’ ‘outside us,’ ‘real,’ or ‘ideal,’ is to be understood in an ontologically more fundamental sense, as opposed to understanding it in the empirical sense, that is, in the sense that is appropriate when we are talking about ontological matters that concern the empirical realm. In this vein, Kant repeatedly uses ‘transcendentally real’ interchangeably with ‘absolutely real.’ See note 34, below, and note 12, chapter 3. These matters, including the relation of Kant’s transcendental idealism to his empirical realism, will be discussed in more detail in chapter 3.

     

39

transcendentally real can be spelled out as the claim that they are real with respect to the transcendental level of reality, that is, that they exist at the transcendental level and are mind-independent with respect to the transcendental level.³² Similarly, the characterization that appearances, space, time, Frodo, Joe’s oasis, Jane’s raccoon, and my dragon are transcendentally ideal can be cashed out as saying that they are ideal or mind-dependent with respect to the transcendental level of reality.³³ But with respect to the empirical level of reality, not all intentional objects are in the same boat anymore. Appearances and space and time are empirically real, that is, they are real with respect to the empirical level,³⁴ or, in other words, they exist at the empirical level and are mind-independent with respect to the empirical level.³⁵ But Frodo, Joe’s oasis, Jane’s illusion, and my dragon are not only transcendentally but also empirically ideal, that is, they are ideal or mind-dependent with respect to both the transcendental and the empirical level of reality. They do not exist from the point of view of fundamental ontology, that is, they are not genuine existents but mere pseudo-existents. Are there any less fundamental levels of reality in Kant’s ontology with respect to which Frodo, Joe’s oasis, Jane’s raccoon, and my dragon are real? Kant does not tell us explicitly, and it is not absolutely necessary to take a stand on this question. But since it is not implausible to think that, from the point of view of fundamental ontology, all intentional objects, even those that do not exist, do have some degree of reality—after all, they are not nothing—and that anything that has some degree of reality must be part of reality in some way, and given the Leibnizian precedent, it makes sense to add another level to Kant’s ontological scheme, even if this means going slightly beyond the textual evidence and what Kant himself might have explicitly had in mind. This level is similar to the level of abstract entities in Leibniz’s three-level scheme and is populated by intentional objects of the likes of Frodo, my dragon, Jane’s raccoon, and Joe’s oasis and possibly also several abstract entities such as properties or mathematical objects.³⁶ These intentional objects do not exist

³² This also means that things in themselves are real both from the point of view of fundamental ontology and from the point of view of the transcendental level, that is, a point of view that is located at the transcendental level and from which one can ‘see’ only the transcendental level. ³³ See B535/A507: “It thus follows that appearances in general are nothing apart from our representations, which is just what we wanted to signify by their transcendental ideality.” This also means that appearances, space, time, Frodo, Joe’s oasis, Jane’s raccoon, and my dragon are ideal both from the point of view of fundamental ontology and from the point of view of the transcendental level. ³⁴ See B44/A28: “Thus we assert the empirical reality of space (with respect to all possible outer experience) although at the same time its transcendental ideality, i.e., that it is nothing as soon as we leave out the condition for the possibility of all experience and regard it as something that grounds things in themselves.” See B52/A35: “Our claims thus teach the empirical reality of time, i.e., objective validity with respect to all objects that may ever be given to our senses. . . . On the other hand, we dispute all claim of time to absolute reality, according to which it inhered in the things as condition or property also without taking into account the form of our sensible intuition. Such properties that pertain to things in themselves also can never be given to us through the senses. In this thus consists the transcendental ideality of time, according to which, if one abstracts from the subjective conditions of intuition, it is nothing at all and cannot be attributed to the objects in themselves (without their relation to our intuition), neither as subsisting nor as inhering.” Further relevant textual evidence will be discussed in chapter 3. ³⁵ This also means that they are real from the empirical point of view, that is, from a point of view that is located at the empirical level and from which one can ‘see’ only the empirical level. So, appearances, and space and time, are transcendentally mind-dependent (that is mind-dependent with respect to the transcendental level of reality) but empirically mind-independent (that is mind-independent with respect to the empirical level of reality). How exactly the notion of empirical mind-independence should be understood will be explicated in the course of our subsequent discussion; see especially section 3.3.1 and 3.4.1. ³⁶ As I read Kant, he believes neither in the genuine existence of properties, understood as abstract universals, nor in the genuine existence of mathematical objects. All of them are pseudo-existing intentional objects, either of finite minds, like us, or of God.

40

      

from the point of view of fundamental ontology but are still real to some degree, in that they in-exist in representations and have some realitas or Sachgehalt, or, using medieval terminology, perfection. They are thus part of the world, understood as the sum total of everything that has reality. In this context, in order to forestall confusion, it is useful to add a few clarificatory comments about the terms ‘reality’ and ‘real,’ which can be used in several different senses. In the first two senses, ‘is real’ is used to describe the ontological status of a thing. First, a thing can be said to be real with respect to a certain level of reality L in the sense that the thing is mind-independent with respect to L and exists at L. This is the sense of ‘real’ that Kant has in mind when he talks about something’s being empirically real or transcendentally real. Second, ‘is real’ can be used to express that a thing is mindindependent with respect to some level of reality and exists from the point of view of fundamental ontology. This is the sense of ‘is real’ that a realist has in mind when he says of certain objects that they are real (although most realists only recognize one level of reality and, thus, would not explicitly relativize the thing’s mind-independence to a level of reality). It is also fair to say that this is the kind of reality that Kant’s category of actuality (Wirklichkeit) refers to. Third, ‘is real’ can also be used to describe the ontological value of a thing, as it were. In this this usage, ‘is real’ is usually combined with an indication of the amount or degree of reality that pertains to the thing in question. There are two dimensions along which the amount of reality of a thing can be specified. The first dimension concerns the way of being of a thing, which corresponds to its location in the hierarchy of ontological levels. Transcendentally real things exist mind-independently from the point of view of fundamental ontology and are located at the most fundamental level of reality. They have the most reality along this dimension. Empirically real things exist minddependently from the point of view of fundamental ontology and are located at the empirical level of reality. They have less reality than transcendentally real things but more reality than empirically ideal things, since the latter are not only mind-dependent but also do not exist from the point of view of fundamental ontology. The second dimension along which the amount of reality of a thing can be determined concerns what medieval and early modern philosophers liked to refer to as the ‘perfection’ of a thing, which is determined by the thing’s positive (as opposed to privative) properties. For example, a horse is more perfect and, hence, more real, than a tree, which, in turn, is more perfect and, hence, more real, than a stone. And, of course, the most real being (the ens realissimum, or omnitudo realitatis), which comprises all perfections, is God. This is the kind of reality that Kant’s category of reality (Realität) refers to. Note that, in the indicated tradition, existence is included among the perfections. On Kant’s view, by contrast, it is not included, as he famously explicates in connection with his critique of the ontological argument.³⁷ As I read him, Kant agrees that 100 existing Euros have more reality than 100 possible Euros with respect to the indicated first dimension along which amount of reality can be measured, but he denies that they have more reality with respect to the indicated second dimension. Also note that all really possible things are real and have reality along both these dimensions and, thus, are real in the third sense of being real just explicated, even if they are not actual or real in the second sense explicated above.

³⁷ See esp. B626–B629/A598–601.

     

41

There is another somewhat sticky, largely terminological question in this vicinity that we should briefly address, namely, whether appearances can be said to exist at the transcendental level of reality, albeit only in a mind-dependent way. The answer to this question, in turn, depends on how we want to respond to the following query. Given that the transcendental level is a level of mind-independent reality, it seems plausible to say that anything that exists at the transcendental level of reality also exists from the point of view of fundamental ontology, or, more briefly, genuinely exists. But how about the converse? That is, can anything that genuinely exists also be said to exist at the transcendental level of reality? As we have just learned, despite their transcendental ideality, appearances can be said to exist from the point of view of fundamental ontology (for reasons to be discussed). This means that if we were to opt for the response that anything that genuinely exists can be said to exist at the level of transcendental reality, we would be committed to the view that appearances exist at the level of transcendental reality, albeit only mind-dependently. There are three considerations that may be taken to speak against this view. First, one could argue that no entity can exist at multiple levels of reality. And we already know that appearances exist at the empirical level of reality. Second, one could balk at the oddness of saying that mind-dependent entities exist at the transcendental level, which after all is supposed to be a level of mind-independent reality. Third, one could hold that, since the transcendental level is the most fundamental level of reality, it is plausible to assume that only objects that are real from the point of view of fundamental ontology exist at this level. I take this debate to be mainly about matters of presentation rather than substance, and so I do not think that very much hangs on the answer. But I feel the pull of these considerations, and so I opt for the response that not everything that exists from the point of view of fundamental ontology can be said to exist at the transcendental level of reality.³⁸ Case in point: appearances.³⁹ Given the foregoing, we can also appreciate that it would be slightly misleading to explicate what it means to engage in considerations from the point of view of fundamental ontology by saying that it amounts to engaging in considerations from the transcendental point of view, that is, from a point of view that is located at the transcendental level of reality and from which one can ‘see’ only the transcendental level. Of course, there is quite a bit of overlap between these two kinds of considerations, since the transcendental level is the most fundamental level of reality. But, as anticipated above, a more apt explanation is that it means to engage in considerations of the nature and structure of reality in general from a meta-perspective, a perspective from which, among other things, reality is seen to comprise (at least) two levels of reality, the entities at the transcendental level are seen to be mind-independent existents, the entities at the empirical level are seen to be minddependent existents, and fictional characters and the intentional objects of hallucinations, illusions, and dreams are seen to be mind-dependent pseudo-existents.⁴⁰ From now on, ³⁸ This way of proceeding also coheres well with Kant’s explication that the description that space is transcendentally ideal means that space is “nothing” if “we regard it as something that grounds things in themselves” (B44/A28, my emphasis), that is, if we consider it as a possible inhabitant of the transcendental level of reality. ³⁹ So, where do appearances genuinely exist then? Answer: at the empirical level and only the empirical level. ⁴⁰ Could one say that considerations from the point of view of fundamental ontology amount to considerations of facts at the transcendental level of reality, which one could equally well consider from the transcendental point of view? After all, that reality comprises (at least) two levels and that appearances and other intentional objects are fully mind-dependent is not the case in virtue of our representing it to be so. These are mind-independent facts,

42

      

when talking about the mind-dependence of appearances, the mind-independence of things in themselves, and the existence of appearances and things in themselves versus the pseudo-existence of other intentional objects I will no longer add explicitly that I am speaking from the point of view of fundamental ontology, but this is how I should be understood unless indicated otherwise.

2.3 Capturing the Relevant Sense of Mind-dependence and Mind-independence Now, in what sense are things in themselves mind-independent and appearances minddependent exactly? In trying to answer this question, it is important to acknowledge that there are several senses in which things in themselves might be mind-dependent. For example, they might be mind-dependent in the sense that their creation and continued existence depends on God, who, one might argue, should be conceived of as an infinite mind. Similarly, things in themselves, or at least some things in themselves, might be minds. Such things in themselves would be mind-dependent in the sense that they cannot exist unless there is a mind. It could also be that some things in themselves are selfconscious minds that essentially represent themselves or minds that have some of their representations essentially. Such things in themselves would be mind-dependent in the sense that they cannot exist unless they are represented by a mind or unless there is a mind with certain specific representations.⁴¹ Bearing all of these constraints in mind, here is a first stab at making more precise in what sense the being of things in themselves is mind-independent and the being of appearances and of other intentional objects is mind-dependent. The being of an entity E is mind-dependent if, and only if, E has being, at least partly, in virtue of being represented by a finite mind; the being of an entity is mind-independent if, and only if, it is not mind-dependent.⁴² This characterization classifies the being of appearances and other intentional objects, such as Frodo, Joe’s oasis, Jane’s raccoon, and my blue dragon, as mind-dependent but the being of things in themselves as mind-independent, including the being of the possible kinds of things in themselves just considered. The suggested definition is only a first stab because, among other things, we still need to think more about whether we really want to say that an entity whose being is minddependent has being, at least partly, in virtue of the fact that (1) there is a finite mind who represents it, or whether it would be preferable to choose a weaker formulation, for example (2) there is a finite mind who could represent it, or (3) there could be a finite mind that represents it. The difference between (2) and (3) is that (3) allows for mindless

that is, facts at the transcendental level. It is correct that the indicated facts are mind-independent, but in order to see these facts one cannot just look at the transcendental level itself, that is, one cannot confine oneself to the transcendental point of view but has to look at reality in general. Moreover, mind-independent facts are not the only facts that we can consider from the point of view of fundamental ontology. There are facts at the empirical level of reality as well, which, from the point of view of fundamental ontology, are mind-dependent. See section 3.4.5. ⁴¹ Thanks to Andrew Chignell for helpful discussion of the points covered in this paragraph. ⁴² The qualification that E has being ‘at least partly’ in virtue of being represented by a finite mind is inserted in order to leave room for the possibility that, in addition to the mind representing E, there are other conditions in virtue of whose satisfaction E has being.

    -  -

43

worlds that contain entities that are mind-dependent in the relevant sense, while (2) rules them out. Based on Kant’s claim that “if we take away the thinking subject, the whole corporeal world must disappear” (A383), which he repeats in several variations, as cited above, I propose to set formulation (3) aside. For Kant, in a possible world in which there are no finite minds, there is no empirical level of reality and no appearances either.⁴³ With respect to the choice between (1) and (2), the crucial question in the background is whether appearances exist in virtue of being represented in actual representations or possible representations. This question is rather tricky, and we will be able to discuss it properly only after having learned more about what kind of representations ontologically specify appearances. To anticipate, the reading that, in my judgment, is ultimately most promising is a version of the first view, that is, the view that appearances exist in virtue of being represented in actual representations, which speaks for the strong formulation (1). But since the issue is complicated and bound to be controversial, and since I do not want to rule out that, even if appearances are ontologically specified by actual representations, there may be other intentional objects that depend on merely possible representations, I will refrain from choosing between options (1) and (2) for the purpose of settling on a basic definition of mind-dependence. In that way, readers who take appearances to be ontologically specified by possible representations can also remain on board, at least for the time being. We already saw above that whether, when, and for how long, appearances exist in the empirical world, or at the empirical level of reality, only depends on whether and how we represent them to exist there but not on when and for how long anybody actually represents them. But what about their existence from the point of view of fundamental ontology? Is it plausible to think that, from the point of view of fundamental ontology, appearances exist, and other intentional objects have being, only when and for as long as they are represented by a finite mind? As I see it, the answer to this question is no. For, as noted in section 2.2 and as will be discussed at length in chapters 4 and 5, Kant classifies time as transcendentally ideal and, hence, as ideal from the point of view of fundamental ontology. This makes it difficult to assign any meaning to the claim that, from the point of view of fundamental ontology, some entity exists or has being at a certain time or for a certain length of time. In general, due to the transcendental ideality of time, non-relativized existence or being claims, when articulated from the point of view of fundamental

⁴³ Lucy Allais and Tobias Rosefeldt understand appearances as things qua bearers of certain properties, which I will call ‘appearance properties.’ According to Allais’ account, appearance properties are ‘essentially manifest properties,’ that is, properties of appearing in certain ways to minds like ours, for example, the property of a rose of appearing extended to us; see Allais 2007, 469–77; Allais 2015, chs 5 and 6. According to Rosefeldt’s account, appearance properties are dispositions to cause certain representations in subjects with our forms of sensibility, for example, the property of a rose to have a property such that through the presence of this property, in appropriate circumstances, intuitions of it as extended are caused in subjects with the faculty of spatiotemporal intuition; see Rosefeldt 2007, 189–93. Appearances as characterized by Allais and Rosefeldt are mind-dependent in the weak sense that they cannot be described or conceived without reference to a certain kind of mind and maybe also in the sense that their existence depends on the possible existence of a mind of the relevant kind. But appearances as conceived by Allais and Rosefeldt do not depend on the actual existence of a mind in the same world in which they exist. In a mindless world with the same laws of nature as the actual world, a rose still has the property of appearing extended to human beings, and it also still has the disposition to cause intuitions of it as extended in beings like us. It just so happens that no mind is around to whom the rose actually appears and who actually intuits it as extended. This means that these readings admit mindless possible worlds in which appearances exist, which is a possibility that I take to be ruled out by Kant’s text. Further problems for the Allais–Rosefeldt reading are discussed in sections 5.3 and 5.4.

44

      

ontology, are always to be understood in a tenseless sense.⁴⁴ The transcendental ideality of time also raises the question of what to say, with respect to the issue of tense, about the representing in virtue of which mind-dependent entities have being. Given that ‘has being’ in our definition is to be read in a tenseless sense, it seems advisable to understand the relevant representing in the same way. There is one slight complication, though, namely, that many representations of finite minds that ontologically specify intentional objects actually are in time, including the special kind of representations that ontologically specify appearances, as we will see.⁴⁵ So, in order to prevent confusion, we will define that, from the point of view of fundamental ontology, a finite mind M has a representation R in a tenseless sense if, and only if, either R is outside time, or R is in time and M has R at some time in the history of the world at the empirical level of reality. This allows us to understand the representing in our definition, on which the being of mind-dependent entities depends, in a tenseless sense. Given all of these considerations, I propose the following more precise definition of a basic sense of mind-dependence that applies to intentional objects, including appearances. The being of an entity E is mind-dependent if, and only if, necessarily, if E has being, there is a finite mind M, and E has being, at least partly, in virtue of the fact that M has, or could have, a possibly complex representation whose presentational content includes E; the being of an entity is mind-independent if, and only if, it is not mind-dependent. This definition, as all of the subsequent definitions in this section, is to be understood as formulated from the point of view of fundamental ontology, and ‘has being,’ ‘there is,’ and ‘has’ are all to be understood in a tenseless sense as just explicated. An entity is mind-dependent in the basic sense if, and only if, its being is mind-dependent; an entity is mind-independent in the basic sense if, and only if, its being is mind-independent. For clarity’s sake, it is also useful to highlight explicitly that, since the way of being of an existent is existing, and since the way of being of a non-existent intentional object is pseudo-existing, the indicated definitions of what it means for the being of an entity to be mind-dependent/mind-independent, also characterize what it means for the existence of an entity to be mind-dependent/mindindependent and for the pseudo-existence of an entity to be mind-dependent.⁴⁶ These definitions are not fully explicit—in order to be fully explicit, the ‘in virtue of ’ relation would still have to be unpacked—but, together with our previous discussion, they serve to sufficiently fix our ideas and capture a basic sense of mind-dependence that we can work with going forward.

⁴⁴ Certain relativized existence or being claims, such as the claim that, from the point of view of fundamental ontology, my cat Franz exists at the empirical level of reality, or that Frodo has being in the ‘world’ of The Lord of The Rings, can, of course, be understood in a tensed sense without difficulty. Every appearance exists and has being at a certain time and for a certain period of time at the empirical level of reality, and every character in The Lord of the Rings exists and has being at a certain time and for a certain period of time in the ‘world’ of The Lord of the Rings. ⁴⁵ I say ‘many’ and not ‘all’ since there may be finite minds who have representations that are outside time and on which certain intentional objects depend. Whether human minds also have such representations, on Kant’s view, is a complicated issue that lies beyond the scope of the present investigation. ⁴⁶ Note that since we introduced the notion of pseudo-existence explicitly in order to have a term to refer to the way of being of intentional objects that do not exist (from the point of view of fundamental ontology), the pseudoexistence of any entity is necessarily and trivially mind-dependent. But the definition just formulated is still useful with respect to clarifying the notion of pseudo-existence since it makes explicit in what sense the pseudo-existence of intentional objects is mind-dependent, namely, in the sense that they pseudo-exist in virtue of the fact that there is a finite mind that represents (or could represent) them.

    -  -

45

Appearances are mind-dependent in the basic sense just defined, but they are also minddependent in a more specific sense, and more can be said to characterize it. First, it may turn out that appearances depend on actual representations of finite minds, as opposed to actual or possible representations, as stated in our definition. As already anticipated, this is, indeed, the reading that I will settle on by the end of this chapter. Second, the minddependence of appearances is not just a dependence on finite minds but a dependence on human minds or minds that have the same forms of sensibility and the same forms of thinking as human minds. (In the following, I will use ‘human mind’ in a broad sense to include the minds just described.) There might be entities that, although not dependent on human minds, are mind-dependent in the sense of our definition and count as appearances in a more general sense of the term ‘appearance,’ on account of their dependence on nonhuman finite minds, whose possibility Kant contemplates at several places.⁴⁷ But in Kant’s primary use, which I will follow, the term ‘appearance’ can be understood as an abbreviated form of ‘appearance to us,’ where ‘us’ means human beings and beings with minds like ours. Up to this point, we have focused on the mind-dependent being of appearances and other intentional objects. It is time say something about mind-dependent properties and other mind-independent ontological ingredients and how they fit into the story outlined so far in order to further illuminate the specific character of the mind-dependence of appearances and other intentional objects. What does it mean for a property or an ontological ingredient of an entity to be mind-dependent? Against the background of our previous considerations, we can offer the following definition. A property P/an ontological ingredient OI of an entity E is mind-dependent if, and only if, necessarily, if E has P/comprises OI, then there is a finite mind M, and E has P/comprises OI, at least partly, in virtue of the fact that M has, or could have, a possibly complex representation whose presentational content includes E having P/comprising OI; a property P/an ontological ingredient OI of an entity E is mind-independent if, and only if, it is not minddependent.⁴⁸ Again, ‘is,’ ‘has,’ and ‘comprises’ in this definition are to be understood in a tenseless sense, and the definition is formulated from the point of view of fundamental ontology. It is a noteworthy fact that the class of entities that are mind-dependent in the basic sense includes entities that have mind-independent properties or other mind-independent ontological ingredients, which might even be essential. How so? The being of any entity that has at least one mind-dependent essential property or comprises at least one other ⁴⁷ See GMS, 4:451: “This must suffice as a, though rough, distinction of a world of sense from the world of the understanding, of which the former can be very different according to the difference of sensibility in various world-onlookers, while the second that is its ground remains always the same.” See FM, 20:267: “For it may be that some world-beings might intuit the same objects under another form . . . ” ⁴⁸ My reason for formulating this definition in such a way that it covers not only properties but also other kinds of ontological ingredients of entities will become clear shortly. (For a reminder about how to think about ontological ingredients, revisit section 1.1.2) Also note that when I talk about properties of entities, I mean their particular property instances. To illustrate the notion of a ‘property instance,’ take the example of the property of being red, or redness, understood as an abstract universal. The redness of my coffee mug, the redness of my cat’s tongue, and the redness of the rose in the vase in front of me, are all particular instances of the property redness. This distinction between properties and property instances is useful to highlight because, on my view, it is in principle possible for the same property to have mind-independent and mind-dependent instances. See note 111, in chapter 3. Another potentially useful observation with respect to the definition is that pseudo-existents, by stipulation, do not have any mind-independent properties nor comprise any mind-independent ontological ingredients.

46

      

mind-dependent essential ontological ingredient counts as mind-dependent, according to our definition, even if this entity also has mind-independent properties or other mindindependent ontological ingredients, and even if these mind-independent properties and ingredients are essential to it. For since no entity can have being unless it has all of its essential properties and ingredients, any entity with a mind-dependent essential property or ingredient has being at least partly in virtue of the fact that there is a mind that represents it, or could represent it.⁴⁹ Conversely, this means that entities that count as mind-independent in the basic sense have no essential mind-dependent ontological ingredients or properties. This result also helps to pinpoint another respect in which the mind-dependence of appearances is stronger than the basic sense of mind-dependence defined above. As explicated in the previous section, statements such as that appearances “are nothing apart from our representations” (B535/A507), and that they “cannot exist outside of our mind at all” (B520/A492) do not leave room for doubt that Kant takes appearances to be fully mind-dependent, or mind-dependent through and through without any mind-independent remainder, so to speak. It is tempting to suggest that the relevant stronger sense of mind-dependence can be characterized by saying simply that not only the existence of appearances but also all of their properties are mind-dependent. This is a promising suggestion, but we need to be a bit more careful about how to phrase it and also specify more precisely what kind of properties we are talking about. Starting with the latter issue, there are the properties of appearances at the empirical level that determine them as the empirical objects they are, for example, the properties of my cat to be extended and cat-shaped and of being currently located on my bed. These properties are, indeed, mind-dependent (from the point of view of fundamental ontology). But it is not implausible to hold that all entities, be they mind-dependent or mind-independent, necessarily also have certain properties that one might call ‘logical properties,’ such as the property of being self-identical or of not being A and not A at the same time or of being such that 2 +2=4. And it is equally plausible to hold that these logical properties are mindindependent. Moreover, there are also the (non-logical) properties of appearances that are the target of our meta-investigations from the point of view of fundamental ontology, for example, the property of being mind-dependent or of being an intentional object or of being an appearance, properties that are arguably also mind-independent. Following Kant, I will use the term ‘determinations’ to refer to the particular property instances and particular relation instances of a thing that determine it as the entity that it is at the level

⁴⁹ Note that this is not a special feature of our definition of mind-dependence in the basic sense. It can be said in general that the being/existence of any entity that has at least one mind-dependent essential ontological ingredient is mind-dependent, even if the entity also has mind-independent ontological ingredients, be they essential or contingent. For since the entity cannot have being/exist unless it has all of its essential ingredients, its being/ existence depends, at least in part, on the mind on which its mind-dependent essential ingredients depend, regardless of how exactly this dependence is cashed out. As already noted, it is a popular strategy among twoaspect commentators to explicate the mind-dependence of appearances in terms of the mind-dependence of certain properties of empirical objects, for example, the property of appearing to human perceivers as extended. On this kind of two-aspect view, appearances are identified with empirical objects qua bearers of these minddependent properties. Some commentators present it as a decisive advantage of this reading that it only characterizes certain properties but not the existence of empirical objects as mind-dependent; see Allais 2004, Allais 2007. But since these mind-dependent properties are essential to empirical objects, on Kant’s view— empirical objects are essentially extended, for example—the alleged advantage is illusory. On the sketched conception of appearances, the existence of empirical objects is mind-dependent after all, even if we grant that empirical objects have mind-independent aspects.

    -  -

47

of reality at or in the intentional ‘world’ in which it exists.⁵⁰ Using this terminology, we can say that appearances and other intentional objects are mind-dependent in that their being and all of their determinations are mind-dependent. We are getting closer, but this description still does not go far enough to fully capture the special sense of minddependence that is characteristic of appearances and other intentional objects. For it allows that the objects so described comprise, in addition to their mind-dependent determinations, other ontological ingredients that are mind-independent. For example, one might argue that some objects are compounds of mind-dependent determinations and a mindindependent particular ingredient that acts as the bearer of these determinations while being completely ‘bare’ on its own.⁵¹ Such objects are mind-dependent in the stronger sense just explicated but they are still not fully mind-dependent, or mind-dependent through and through. Relatedly, appearances and other intentional objects have, or comprise, their ontological ingredients not only partly but solely in virtue of being represented (or possibly being represented) to have them. To be sure, a lot more can be said and will be said about various conditions on which the existence and ontological ingredients of appearances depend, for example, about the various kinds of faculties that minds need to have in order to be capable of the special kind of representations that ontologically specify appearances or about the sources from which such minds receive the materials out of which these special representations are constructed. But those are conditions for the existence and ontological ingredients of appearances only because they are conditions for the possibility of the special kind of representations that ontologically specify appearances and not because they add any further requirements in addition to the requirement that there must be a finite mind who has (or could have) the indicated special kind of representations. Taking all of this on board and noting that the determinations of a thing are among its ontological ingredients but its logical and meta-theoretical properties are not, we can propose the following definition of full mind-dependence, which applies to appearances and other intentional objects, such as Frodo, Joe’s oasis, Jane’s raccoon, and my purple transportation dragon, as well as possible intentional objects of other, non-human finite minds. An entity E is fully mind-dependent if, and only if, all of E’s determinations and other ontological ingredients are strongly mind-dependent, where a property P/an ontological ingredient OI of an entity E is strongly mind-dependent if, and only if, necessarily, if E has P/comprises OI, then there is a finite mind M, and E has P/comprises OI in virtue of the fact that M has, or could have, a possibly complex representation whose presentational content includes E having P/comprising OI.⁵² It follows from this definition of full mind-dependence that the being of a fully mind-dependent ⁵⁰ See B599–611/A571–583; B626–629/A598–601. In the following, I may on occasion refer to certain determinations as properties or relations. This should be understood as a measure of convenience; precisely speaking, determinations are particular instances of properties or relations. For an explanation of the notion of a ‘property instance,’ see note 48. ⁵¹ Note that one might argue that, on Kant’s view, appearances have such a bare particular ontological ingredient (albeit a mind-dependent one), namely, the ingredient that is specified by the concept of the transcendental object, that is, of an as yet completely undetermined particular something = X that is (a) distinct from both the representations of it and the cognizer entertaining these representations and (b) a bearer of determinations. ⁵² So, whether an ontological ingredient of E is strongly mind-dependent or just mind-dependent depends on whether E has it in virtue of (possibly) being represented to have it or whether E has it partly in virtue of (possibly) being represented to have it. For a slightly more inclusive conception of full mind-dependence that is particularly relevant with respect to Berkley’s idealism, see note 32, chapter 3.

48

      

entity is mind-dependent in the sense defined above. That is, being fully mind-dependent implies being mind-dependent in the basic sense, which is a welcome result. But the being of entities that are fully mind-dependent is also mind-dependent in a stronger sense, that is, it is strongly mind-dependent. The being of an entity E is strongly mind-dependent if, and only if, necessarily, if E has being, there is a finite mind M, and E has being in virtue of the fact that M has, or could have, a possibly complex representation whose presentational content includes E.⁵³ Now that we have a conceptually more precise grasp of the sense in which appearances are mind-dependent, let us briefly think a bit more about the sense in which things in themselves are mind-independent. To anticipate, while the definition of mindindependence in the basic sense provided above must also be slightly enhanced to capture the specific sense in which things in themselves are mind-independent, many fewer adjustments are required than we had to make to the definition of mind-dependence in the basic sense in order to capture the specific sense in which appearances are minddependent. No matter whether appearances are ontologically specified by actual representations or possible representations or whether there are any entities at all that are ontologically specified by merely possible representations, it is correct to say that things in themselves depend on neither actual nor possible representations of finite minds, which is also what the definition of mind-independence in the basic sense says. With respect to the question of what kinds of minds things in themselves depend on, one can also distinguish a stronger and a weaker sense of ‘mind-independent’ depending on whether the minds in question are any kind of finite minds, as in our definition of basic mindindependence, or human minds. Kant often emphasizes the independence of things in themselves from us, but the principal sense of the term ‘thing in itself ’ in Kant’s use, I submit, is the stronger sense, according which things in themselves are independent from any kind of finite minds.⁵⁴ Following Kant, this is also how I will use ‘thing in itself ’ in the following. Since the relevant minds in our definition of mind-independence in the basic sense are finite minds, no further specification needs to be added in this respect to capture the sense in which things in themselves are mind-independent. Regarding the question of what kind of properties or determinations things in themselves can have, we already noted that our definition of mind-dependence in the basic sense implies that entities that count as mind-independent according to it do not have any essential mind-dependent ontological ingredients or properties, which comports well with how Kant describes things in themselves. But this definition also classifies entities as mind-independent in the basic sense that have non-essential mind-dependent ontological ingredients or non-essential mind-dependent properties. So, the question is whether things in themselves can have such ingredients or properties. On my view, the answer is yes with respect to the non-essential mind-dependent properties but no with respect to the non-essential mind-dependent ontological ingredients. Consider the property of being self-reflectively thought about by me, which is mind-dependent according to our ⁵³ So, whether the being of E is strongly mind-dependent or just mind-dependent depends on whether E has being in virtue of (possibly) being represented or whether E has being partly in virtue of (possibly) being represented. ⁵⁴ This reading is supported by, for example, Kant’s assertion that while the world of sense can differ for different kinds of finite beings, the underlying intelligible world, the world of things in themselves, is the same for all of them; see GMS, 4:451, quoted in note 47.

    -  -

49

definition, because it pertains to a given thing at least partly in virtue of the fact that I present the thing to have it. (Self-reflectively thinking about a thing requires thinking about thinking about the thing.). Arguably, I can self-reflectively think about things in themselves,⁵⁵ and my self-reflectively thinking about them is neither essential to them nor a determination of them, and so things in themselves can have non-essential minddependent properties that are not among their ontological ingredients. Regarding the question of non-essential mind-dependent ontological ingredients, it is not clear if things in themselves can have non-essential ontological ingredients at all—Leibniz, for example, holds a ‘super-essentialist’ view, according to which all of the ontological ingredients of things in themselves are essential to them. Be this as it may, it at least seems a fairly safe bet that, on Kant’s view, if there are things in themselves with non-essential ontological ingredients, these ingredients must be mind-independent. Just as Kant regards appearances as mind-dependent through and through, or as fully mind-dependent, he regards things in themselves as mind-independent through and through, or as fully mindindependent, that is, as mind-independent in the basic sense and having only mindindependent ontological ingredients. For a while I had myself convinced that, for somewhat arcane reasons that we need not go into here, it is metaphysically impossible for an entity to be mind-independent in the basic sense and have non-essential mind-dependent ontological ingredients. In that case, the definition of mind-dependence in the basic sense by itself would already be sufficient to capture the specific sense in which things in themselves are mind-independent. But in the course of various internal debates with myself this conviction suffered some erosion, and so in order to be on the safe side I will add yet another definition. An entity E is mind-independent if, and only if, E is mindindependent in the basic sense, and all of E’s determinations and other ontological ingredients are mind-independent. The reason why I refer to the special kind of mindindependence that characterizes things in themselves simply as ‘mind-independence,’ while I refer to the special kind of mind-dependence that characterize appearances as ‘full mind-dependence’ is twofold. First, if I were to use corresponding labels for them, that is, ‘mind-independence’ and ‘mind-dependence’ or ‘full mind-independence’ and ‘full mind-dependence,’ respectively, we would end up with the terminologically potentially confusing situation that being (fully) mind-independent is not the same as not being (fully) mind-dependent and being (fully) mind-dependent is not the same as not being (fully) mind-independent.⁵⁶ Second, being mind-independent as just defined is not much stronger (if at all) than being mind-independent in the basic sense, and it is what most commentators have in mind when they characterize things in themselves as mindindependent. By contrast, being fully mind-dependent, as defined above, is quite a bit stronger than being mind-dependent in the basic sense, and it is also quite a bit stronger than what most proponents of the two-aspect interpretation have in mind when they characterize appearances as mind-dependent. So, it seems advisable to stick simply with ‘mind-independence’ as a label for the kind of mind-independence that applies to things in ⁵⁵ For a defense of the claim that we can think about things in themselves, see Jauernig (in preparation). Also see Bxxvi. ⁵⁶ While for an entity to be mind-dependent in the basic sense is equivalent for it not to be mind-independent in the basic sense, for an entity to be fully mind-dependent as defined above is not equivalent for it not to be mindindependent as just defined. For example, an entity that has some mind-dependent and some mind-independent essential determinations is not mind-independent, but it is also not fully mind-dependent.

50

      

themselves and to adopt a special label for the kind of mind-dependence that applies to appearances so as to keep readers mindful of the facts that the relevant sense of minddependence is very strong, and that being fully mind-dependent is not the same as not being mind-independent. To summarize the main results of the foregoing discussion, appearances are fully minddependent, that is, all of their determinations and other ontological ingredients, and, thus, their existence, are strongly mind-dependent as defined above, with the small proviso that the phrases ‘or could have’ and ‘or possibly being represented’ in the suggested definitions (in ‘M has, or could have, a representation whose presentational content includes . . . ’ and ‘ . . . in virtue of being represented, or possibly being represented, by a finite mind’) are to be deleted and with the added further specification that the relevant finite minds are human minds. (The argument for the deletion will be supplied in section 2.8.) In the following, I will no longer add explicitly that appearances are human-mind-dependent, but this is how I should be understood. Things in themselves are mind-independent in the sense that their existence and all of their determinations and other ontological ingredients are mind-independent as defined above, where the relevant minds are any kind of finite minds. As a kind of signpost, I should like to add that, from now on until chapter 5, we will be mostly concerned with Kant’s views about and arguments concerning appearances and the empirical level of reality. Readers who are eager to learn more about the transcendental distinction and, more generally, about how things in themselves and appearances are related are encouraged to skip ahead to chapter 5 to read sections 5.1–5.5 before returning here to continue with section 2.4.

2.4 Appearances and Perception A crucial task in order to specify in more detail what kind of intentional objects appearances are and how they differ from other intentional objects, such as fictional characters and the objects of dreams, illusions, and hallucinations, is to find out more about the special kind of representations that ontologically specify them.⁵⁷ This task will occupy us for the remainder of this chapter. Since appearances and their ontological ingredients―including their spatial and temporal determinations and with them space and time⁵⁸―is all there is at the empirical level of reality, our task can also be described as the search for the specific kind of representations that ontologically specify the ontological furniture of the

⁵⁷ As you will recall, by that I mean the representations in virtue of being represented in which a given intentional object has being and all of its determinations and other ontological ingredients. ⁵⁸ The spatial determinations of an appearance comprise (a) its extension, three-dimensionality, shape, and spatial size, where the latter is measured according to the Euclidean distance function (that is, the Pythagorean theorem) and specified with respect to a particular unit of measurement, as well as (b) its spatial locations with respect to and its spatial distances from other appearances, where the former comprise instances of the relations of being in between and being next to, and the latter are of the form of being a particular Euclidean distance away from X specified with respect to a particular unit of measurement. The temporal determinations of an appearance comprise (a) its duration and temporal size (that is, the length of its duration), where the latter is determined according to a measure derived from the distance function of Galilean spacetime and specified with respect to a particular unit of measurement, as well as (b) the temporal locations of its states with respect to and the temporal distances of its states from each other and the states of other appearances, where the former comprise instances of the relations of being before, being after, and being simultaneous with, and the latter are of the form of being a particular Galilean temporal distance before/after X specified with respect to a particular unit of measurement.

  

51

empirical level of reality, so to speak. As before, our main focus for the time being will be on outer appearances, that is, appearances in time and space that populate the public empirical world that we all share. (‘Appearance’ without qualification should thus still be understood as ‘outer appearance’ unless indicated otherwise.) In addition to examining to which species of representations the representations belong by which appearances are ontologically specified and which conditions a representation has to satisfy in order to be a member of this species, we will also investigate whether appearances are ontologically specified by actual or merely possible representations of the identified kind, a question that we left deliberately open in section 2.3. Before we start, two more preliminary remarks are in order, one of them terminological. In the context of our search for the special kind of representations that ontologically specify appearances, we will have ample occasion to talk about various cognitive faculties and various different kinds of representations. We already know that appearances depend on human minds. Accordingly, the faculties and representations that we will examine are mostly of the human variety, and so, for efficiency’s sake, I will not explicitly say that every time I mention them. Whenever I want to refer to cognitive faculties or representations of non-human minds, I will explicitly add an appropriate qualification. The second preliminary comment is that it is not true in general that, if an intentional object is represented in a certain representation, then the object is ontologically specified by this representation. Different representations and different kinds of representations can have the same presentational content and thus the same intentional object. But while it is legitimate to say that by representing a given intentional object each one of these representations engenders the object’s presence to the mind of the relevant cognizer, arguably only one of them can be responsible for ontologically specifying the object. For example, Frodo, the hobbit, is an intentional object of original imaginings in Tolkien’s mind, thousands of copies of a series of books, thousands of showings of various movies, countless conversations, newspaper articles, and so on. All of these representations make Frodo present to the mind of the cognizers who process them, but Frodo pseudo-exists and has its determinations in virtue of being represented in only one of them.⁵⁹ With respect to our project, this means that while it is a necessary condition for the kind of representations that we are looking for that they have appearances as their intentional objects, this condition is not sufficient. Appearances may also be represented in other kinds of representations. What can be said, though, is that it is part of the nature, or essence, of appearances to be intentional objects of the representations that we are looking for, and that appearances can be defined as intentional objects of these representations. Among commentators who read Kant as an idealist about appearances, it is a popular view that, for Kant, appearances exist in virtue of being perceived.⁶⁰ On this view, Kant’s idealism is cut from the same cloth as Berkeley’s in sharing the core thesis that for appearances to exist is to be perceived. This reading could be challenged by appealing to ⁵⁹ Which representation is the ontologically specifying one will depend on one’s theory of intentional objects, and, in this case more specifically, on one’s theory of fictional objects. ⁶⁰ For example, Van Cleve’s conception of appearances as logical constructions out of the perceptual states of perceivers is a version of this kind of view. See Van Cleve 1999, 9 11. The view in question is also frequently presented by saying that appearances exist in virtue of being experienced, where ‘experience’ is understood as synonymous with ‘perception’ or, at least, with ‘personal experience of individual perceivers.’ As we will see, I take appearances to exist in virtue of being experienced too, but it is crucially important on my reading that experience must be carefully distinguished from perception and from the personal experience of individual perceivers.

52

      

the fact that Kant allows that there are―or, for all we know, could be―appearances in the actual world that are never perceived by human beings, such as inhabitants of the moon and other planets, distant stars, ‘magnetic matter,’ and tiny insensible corpuscles.⁶¹ If appearances were ontologically specified by perceptions, none of these appearances would exist. While I agree with the bottom line that perceptions are not the kind of representations that ontologically specify appearances, I also think that the sketched challenge is not the last word about the proposal that they are. It will be instructive to consider how a defender of the proposal could respond to this challenge. Such a response could maintain that the indicated problematic examples can be accommodated by allowing that the relevant perceptions need not be actual but may be merely possible. Although nobody may ever actually see the inhabitants of Neptune, we could see them in the sense that we would see them if we figured out a way to get there (provided they exist). Similarly, if we assume that the restricted acuity and limited scope and range of our senses is not essential to human beings as such, we also could perceive tiny corpuscles and ‘magnetic matter’ (provided they exist) in the sense that we would perceive them if our senses were more acute and had a wider scope,⁶² and perhaps we could also perceive the inhabitants of Neptune (provided they exist) without leaving our home planet in the sense that we would perceive them if our senses had a greater range.⁶³ Another response to the allegedly problematic examples would be to argue that the perceptions that ontologically specify appearances may be unconscious. Even though no human being may ever consciously perceive tiny corpuscles, magnetic matter, or the inhabitants of Neptune, we do perceive them unconsciously (provided they exist). On first glance, this proposal may sound preposterous and without any basis in Kant’s texts, but on second glance it turns out to be surprisingly defensible, to a surprising extent— although the terminology in which it is expressed ought to be slightly adjusted, as we will see. Understanding why and to what extent the proposal is defensible requires paying attention to some underappreciated Leibnizian aspects of Kant’s account of perception. This will take a bit of work but since these Leibnizian aspects have largely gone unnoticed in the literature, and since appreciating them translates into a better understanding of Kant’s views about perception overall, and since we will have occasion to return to them throughout the course of this book, it is time and effort well spent. For all the innovation ⁶¹ See B521/A493: “It must indeed be admitted that there could be inhabitants on the moon, although no human being has ever perceived them . . . ” See B853/A825; B524/A496; B273/A226; ÜE, 8:205. ⁶² I am assuming, with Kant, that ‘magnetic matter’ is indeed a kind of subtle matter, that is, something that is perceptible in principle and not some kind of insensible force or energy field. ⁶³ This assumption is much less implausible than might appear at first glance. What seems to be essential to human minds, on Kant’s view, is that they comprise two distinct faculties that are employed in the cognition of objects, sensibility and the understanding, which both have specific forms, namely, space and time, on the one hand, and the logical forms of judgment or the pure categories, on the other hand. Kant also appears to take it as essential to human beings that we have a body and senses. But the acuteness, scope, and range of our senses could be viewed as contingent features. See B273/A226: “For in general, according to the laws of our sensibility and the context of our perceptions, we would also hit upon the immediate empirical intuition of it [magnetic matter], if our senses were more subtle, whose roughness does not at all concern the form of possible experience.” See KU, 5:455: “Thus matters of opinion are always objects of an empirical cognition (objects of the world of sense) that is at least in itself possible but that is impossible for us, according to the mere degree of the faculty that we possess. Accordingly, the ether of the newer physicists . . . , is a mere matter of opinion, but still of a kind that if the outer senses were sharpened in the highest degree it could be perceived . . . ” In this last quotation, ‘impossible for us, according to the mere degree of the faculty that we possess’ could be understood as ‘impossible for us, holding fixed the acuity and range of our senses,’ which, in turn, could be taken to suggest that the relevant perception is not impossible for us if we do not artificially hold fixed the acuity and range of our senses.

  

53

for which it is justly celebrated, Kant’s account of perception shares several important ingredients with Leibniz’s account, including the view that all perceptions comprise a multitude―indeed, an infinitude―of unconscious elements,⁶⁴ or elements that are not “clear,” to use Leibniz’s and Kant’s technical terminology. For both Kant and Leibniz, all conscious perceptions comprise infinitely many ‘smaller’ representations, not all of which are individually apprehended with consciousness. That is, all conscious perceptions are confused to a certain degree—or, as Kant prefers to put it, indistinct to a certain degree.⁶⁵ On Kant’s view, empirical intuitions of objects comprise smaller intuitions of all the objects’ parts, even of their minute parts, with the proviso that once a certain degree of minuteness has been exceeded the intuitions of the parts are too ‘little’ to be themselves consciously apprehended by us. If, on the other hand, the word sensible is used in its proper meaning, then it is obvious: that . . . if something is an object of the senses and of sensation all simple parts must be too, even though in their case the clearness of representation may be lacking. . . . No microscope has yet been able to discover Newton’s little particles of which the color particles of bodies consist, but the understanding recognizes (or suspects) not only their presence but also that they are actually contained in empirical intuition, even though without consciousness. (ÜE, 8:205)⁶⁶ We see a country house in the distance. If we are aware that the perceived object is a house, we must necessarily also have a representation of the different parts of this house, the windows, doors, etc. For if we did not see the parts, we would not see the house itself either. But we are not conscious of this representation of the manifold of its parts, and our representation of the thought object itself is an indistinct representation. (Log, 9:34)⁶⁷

Indeed, at least in his younger years, Kant even appears to have endorsed Leibniz’s view that we perceive the entire empirical world, albeit mostly unconsciously.

⁶⁴ There are various ways in which one might understand the terms ‘unconscious’ and ‘conscious.’ The sense of ‘unconscious’ that is primarily relevant for Leibniz’s and Kant’s accounts of perception pertains exclusively to representations and is that of not being consciously apprehended, or of not being apprehended with consciousness. There are other senses of ‘unconscious,’ for example, as developed in the psychoanalytic tradition following Freud, that we can set aside for our present concerns. ⁶⁵ More precisely put, a representation R is fully distinct or distinct simpliciter, or entertained or grasped distinctly, if, and only if, R is conscious and each one of its elements is individually apprehended with consciousness; otherwise R is confused (or indistinct, as Kant prefers to say), or entertained or grasped confusedly, at least in part. See Log, 9:33–34: “If I am conscious of a representation, it is clear; if I am not conscious of it, it is obscure. . . . All clear representations . . . can be distinguished with respect to distinctness and indistinctness. If we are conscious of the whole representation, but not of the manifold that is contained in it, it is indistinct.” See B414–415, note; V-Lo/Wiener, 24:805; V-Lo/Pölitz, 24:510–511, 536. See Leibniz 1714a, §13, G 6:604: “Each soul knows the infinite, knows everything, but confusedly. It is like walking on the shore of the ocean, and hearing the great noise that it makes, I hear the particular noises of each wave, of which the whole noise is composed, but without distinguishing them.” Also see Leibniz 1684, A VI.4A:585 588; Leibniz 1704, A VI.6:120. ⁶⁶ Note that the reason why Kant is talking about ‘simple’ parts in this passage is that he is responding to a claim by Eberhard to the effect that the simple parts of space and time are supersensible because we are not individually conscious of them in our perceptions. That is, Kant should not be read as committing himself to the view that appearances are composed of simple parts, a view that he explicitly rejects at other places. ⁶⁷ See ÜE, 8:210: “According to the Critique everything in an appearance is thus itself again appearance, as far as the understanding might divide it into its parts and prove the actuality of the parts, of whose clear perception the senses are no longer capable.”

54

       There is something great and, methinks, very correct in the thought of Mr. von Leibniz: the soul comprises the entire universe within its power of representation although only an infinitely small part of this representation is clear. (NG, 2:199)

How exactly should the confusion or indistinctness of perceptions be understood, on Kant’s view? Taking a brief look at some of the details of Leibniz’s account is helpful here, I believe. According to Leibniz’s doctrine of pre-established harmony, each monad or simple soul at every moment completely perceives the entire empirical world in perfect synchronicity with all other souls.⁶⁸ Since matter is infinitely divisible, and individuals are infinitely complex,⁶⁹ every perceptual state of a monad comprises infinitely many representations. But due to their limited cognitive abilities as finite creatures, and due to the ‘smallness’ or ‘weakness’ of many of these component representations, no monad, not even a rational monad, consciously apprehends each one of the infinitely many representations that are comprised in its perceptual states.⁷⁰ Despite these limitations, many monads do have conscious perceptions—although there are also some particularly dull fellows who spend their life in perpetual cognitive darkness. What happens in the case of the more perfect monads, including us, who are capable of conscious perceptions is that their ‘little’ unconscious representations blur together, or are con-fused, to yield emergent, or supervenient, perceptions that are conscious but have a partially different presentational content.⁷¹ For example, the ‘little’ unconscious representations that represent the water droplets that appear to us as a rainbow are con-fused into conscious perceptions that represent bands of color.⁷² The perceptual states of monads at a given time are thus best ⁶⁸ The doctrine of pre-established harmony is Leibniz’s account of how monads compose a world. The harmony is pre-established because monads do not interact, which means that there could not be any coordination between the states of different monads unless God had pre-established it at the moment of creation. See Leibniz 1686, §9, A VI.4B:1542: “Moreover, every substance is like an entire world, and like a mirror of God or of the whole universe, which each one expresses in its own manner, a bit like how the same city is represented differently depending on the different positions from which it is regarded. . . . It can even be said that every substance in some way bears the character of God’s infinite wisdom and omnipotence and imitates him as much as it is capable. For it expresses, however confusedly, everything that happens in the universe, past, present, or future; this resembles somewhat an infinite perception or an infinite knowledge.” See Leibniz 1686, §14, A VI.4B:1549 1551; Leibniz 1695, §§14 15, G 6:484 485; Letter to Arnauld, October 9, 1687, A II.2:244 245. ⁶⁹ See Leibniz 1704, A VI.6:289–290: “The most important thing in this is that individuality involves infinity, and only someone who is capable of grasping the infinite could have knowledge of the principle of individuation of a given thing.” Also see Letter to de Volder, 1704/1705, G 2:277. ⁷⁰ See Leibniz 1714b, §60, G 6:617: “Since the nature of the monad is representative, nothing can limit it to represent only a part of things. However, it is true that this representation is only confused as to the detail of the whole universe, and can only be distinct for a portion of things, that is, either for those that are closest, or for those that are greatest with respect to each monad, otherwise each monad would be a divinity. . . . Monads all go confusedly to infinity, to the whole; but they are limited and differentiated by the degrees of their distinct perceptions.” ⁷¹ See Leibniz 1684, A VI.4A:592: “When we perceive colors or smells, we certainly have no other perceptions than of shapes and motions, but so numerous and so very small that our mind cannot distinctly consider each individual one in this present state of itself, and thus does not notice that its perception is composed of perceptions of minute shapes and motions alone, just as when we perceive the color green in a mixture of yellow and blue powder, we sense only yellow and blue finely mixed, even though we do not notice this, but rather feign some new thing for ourselves.” I should also add that the con-fusion of ‘little’ perceptions happens in the case of the dull monads too but, due to their extreme dullness, they do not apprehend even these confused perceptions with consciousness. ⁷² More precisely, on my reading of Leibniz, whether a given representation can be consciously apprehended by a monad depends on both a special intrinsic quality of the representation, which I call its ‘strength of protoconsciousness’ or SPC, and on a special intrinsic quality of the monad, namely, its specific consciousness threshold, which corresponds to the particular SPC that a representation has to exceed in order to be consciously apprehended by it. Monads also have specifically different powers of ‘con-fusing’ representations, that is, of

  

55

understood as incorporating two components, on Leibniz’s account, a basic component and an emergent component. The basic component is unconscious and comprises infinitely many little representations that, taken together, completely represent the entire empirical world at the time in question in all of its rich detail from a God’s eye point of view, including all individuals existing in it, but are not ‘strong’ enough to be individually apprehended with consciousness. This component is the same for all monads that exist in the same world and is modeled on and represents the very same content as God’s complete, purely conceptual representation of the relevant time-slice of the empirical world. On my reading, it is these basic components of the perceptual states of all monads that are coordinated in the pre-established harmony and ontologically specify the ontological furniture of the empirical level of Leibniz’s tiered ontology. The emergent component of the perceptual states of monads at a given time supervenes on the basic component, consists of finitely many, partly confused emergent perceptions and depends on special features of the monad to which it belongs. In the case of rational monads, some of the confused emergent perceptions are conscious. The emergent component is peculiar to each monad―indeed, it is what distinguishes each monad from all of the others―and represents only part of the empirical world and does so incompletely, with various subjective distortions, and from a particular point of view.⁷³ Turning to Kant, the passages just cited make it plausible to ascribe a similar twocomponent model of the perceptual states of human minds to him. These perceptual states at a given time incorporate a basic component that is essentially unconscious and an emergent component that can and often does include conscious elements. The basic, unconscious component is infinitely complex in that it comprises a manifold of infinitely many ‘little’ empirical intuitions, and represents the same content for all human subjects that are affected by the same things in themselves.⁷⁴ The emergent component is finitely complex, partly indistinct (confused), and supervenes on the basic component. More precisely, it results from the basic component through (among other things) the confusion―or, to use a more recognizably Kantian expression, the synthesis―of the manifold of the little empirical intuitions making up the basic component during their apprehension.⁷⁵ The emergent component thus represents a partially different content compared to melting multiple ‘little’ representations together in such a way that a perception with a new content emerges whose SPC corresponds to the sum of the SPCs of all of the ‘little’ representations comprised in it. As a result, monads differ with respect to which of their ‘little’ representations are con-fused, how many conscious perceptions they have, what exact content is represented by their perceptions, and how confused their perceptions are. Incidentally, the particular distribution of confusion and distinctness over the perceptual state of a monad also determines where the monad’s point of view is located and which phenomenal body is its body. See Leibniz 1686, §33, A VI.4B:1582; Leibniz 1714b, §62, G 6:617; Letter to Arnauld, April 30, 1687, A II.2:175 176. ⁷³ For a fuller discussion and more textual evidence, see Jauernig 2019, 53 62; and Jauernig, forthcoming 2021a. ⁷⁴ That intuitions are infinitely complex, while concepts are finitely complex, is crucial for Kant’s fourth space argument in the Transcendental Aesthetic. See B40: “ . . . but no concept, as such, can be thought as if it contained an infinite amount of representations in it. Nevertheless, space is thought like this (for all parts of space to infinity are at the same time). Therefore, the original representation of space is an intuition a priori and not a concept.” This argument will be discussed in more detail in section 4.2.2.2. ⁷⁵ See B160: “At first I remark that I understand by the synthesis of apprehension the composition of the manifold in an empirical intuition, through which perception, i.e., empirical consciousness of it becomes possible.” See B162: “For example, if I thus turn the empirical intuition of a house through the apprehension of its manifold into perception . . . ” I say that the emergent component results ‘among other things’ from the confusion/synthesis of the ‘little’ intuitions since, arguably, certain other cognitive operations are involved as well at this stage on Kant’s view, such as the ordering of the newly generated content according to certain forms of sensibility and certain forms of the understanding that guide the imagination in its synthesizing activity.

56

      

the basic component, a content that tends to be peculiar to each observer and her particular circumstances. Lest I am accused of likening Kant’s account of perception too much to Leibniz’s account, it is worth briefly highlighting some of the more salient differences between them. The most obvious differences concern the origin and generation of the unconscious ‘little’ representations out of which the basic components of our perceptual states are composed, and the way in which the emergent components result from the basic components. On Leibniz’s account, the ‘little’ representations as well as the con-fused representations emerging from them arise from each monad’s own depth, so to speak, as a result of the action of the monad’s internal so-called ‘law-of-the-series’ that governs and drives its progression through the series of its perceptual states.⁷⁶ More precisely, the ‘little’ representations are produced and their evolution is guided and fueled by the monad’s primitive active force that makes up one component of its law-of-the-series, while the confusion is mainly due to the monad’s primitive passive force that makes up the other component of its law-of-the-series.⁷⁷ On Kant’s account, although the forms of sensibility and the imagination are involved in the generation of the ‘little’ representations, these representations do not arise from our own depth but are, at least in part, due to affections of our passive sensibility by things in themselves. The emergent perceptions then result by means of further spontaneous, or active, cognitive processing of the ‘little’ representations.⁷⁸ Another potential difference between Leibniz’s and Kant’s account emerges with respect to the question of the nature of the ‘little’ representation contained in the basic components of our perceptual states. Some philosophers in the Leibnizian tradition, in particular, Christian Wolff, characterize perceptions as confused concepts, and thus conceive of the ‘little’ representations as intellectual in nature.⁷⁹ This is how Kant reads Leibniz as well.⁸⁰ ⁷⁶ See Leibniz 1695, §14, G 4:484 485: “ . . . I was insensibly led to an opinion that surprised me. . . . This is that we must say that God has originally created the soul, or every other real unity of this sort, in such a way that everything in it must arise from its own depth by a perfect spontaneity with regard to itself, yet with a perfect conformity to things outside. And thus, since our internal sensations . . . are merely phenomena that follow upon external events, or better are really appearances or like well-ordered dreams, it follows that these perceptions internal to the soul itself come to it through its own original constitution, that is to say, through the representative nature (capable of expressing entities outside of itself in agreement with its organs) that has been given to it from its creation and makes up its individual character. It is this that makes each substance exactly represent the entire universe in its own way and according to a specific point of view, and the perceptions or expressions of external things arrive at the soul at the proper time by virtue of its own laws, as in a world apart, and as if there existed nothing but God and itself . . . ” See Letter to de Volder, January 21, 1704, G 2:264: “That succeeding substance will be considered the same as long as the same law of the series or of simple continuous transition persists, which makes us believe in the same subject of change, or the monad. That a certain law persists that involves all of the future states of that which we conceive to be the same, that is what I say constitutes the same substance.” Also see Leibniz 1686, §14, A VI.4B:1551; Leibniz 1698, G 4:518; Remarks on Arnauld’s letter, A II.2:53; Letter to Arnauld, July 14, 1686, A II.2:80 81. ⁷⁷ See Leibniz (mid-1680s), A VI.4B:1504: “Substances have metaphysical matter or passive power insofar as they express something confusedly, active power insofar as they express something distinctly.” Also see Letter to Remond, February 11, 1715, G 3:636: “Furthermore, since the monads (except the primitive one) are subject to passions, they are not pure forces; they are the foundation not only of actions but also of resistances and passivity, and their passions are in their confused perceptions. It is this that involves matter or the numerically infinite.” ⁷⁸ The cognitive processing involved in perception will be discussed in more detail later on in this section and in section 2.5.2, and the affection of sensibility by things in themselves will be discussed in more detail later on in this section and in sections 2.5.3 and 5.1. ⁷⁹ More precisely, according to Wolff, our soul has one fundamental power of representing, which underlies both our intellect and our senses such that they are distinguished by the distinctness and confusedness of their representations, respectively. See Wolff 1729, §§277, 282, 747, 753, 773. ⁸⁰ See FM, 20:285: “ . . . a kind of enchanted world, to the assumption of which the famous man [Leibniz] could only have been misled by taking sensible representations, as appearances, not, as it should be, for a kind of

  

57

Several Leibniz scholars hold that Kant is guilty of a misreading here and argue that Leibniz does not equate perception with a confused kind of intellection.⁸¹ For reasons that would lead us too far afield, I am with Kant on this point and believe that his ascription of the Wolffian view of perceptions as confused concepts to Leibniz is defensible.⁸² If this is correct, it is another difference between Leibniz’s and Kant’s account of perception that the former regards the unconscious ‘little’ representations comprised in our perceptual states as intellectual, while the latter regards them as sensible. This is a good place to emphasize, though, that Kant consistently refrains from speaking about unconscious perceptions. When he talks about unconscious sensible representations, he usually uses the term ‘empirical intuitions.’ Perception, by contrast, is characterized as consciousness of empirical intuitions, or conscious empirical intuition, or empirical consciousness.⁸³ That is, in Kant’s standard use, ‘perception’ refers to the conscious elements in our perceptual states. This is also how the term should be understood from now on, that is, as referring to representations that are essentially conscious. The term ‘empirical intuition’ is to be read in a more general sense; both conscious and unconscious empirical intuitive representations as well as our perceptual states in general are all empirical intuitions. So, the proper way of characterizing the unconscious basic components of our perceptual states, on the Kantian account, is to say, not that they comprise a manifold of infinitely many ‘little’ perceptions, but that they comprise a manifold of infinitely many ‘little’ empirical intuitions―which is, indeed, how I characterized them above. A final salient potential difference between Kant’s and Leibniz’s account of perception pertains to how they conceive of the presentational content of the unconscious basic components of our perceptual states. It is tricky to figure out how exactly Kant thinks about it, partly because it is not entirely obvious to what extent he takes other cognitive faculties apart from sensibility to be involved in the generation of the ‘little’ empirical intuitions contained in the basic components but also because it is not clear if, in the critical period, he still believes that there is something “very correct” in Leibniz’s view that we confusedly perceive the entire empirical world. If Kant no longer approved of this view, or if he took the ‘little’ empirical intuitions to depend on less than all of the cognitive operations that are required for the representation of appearances, the presentational content of the unconscious basic components of our perceptual states on his view would turn out to be poorer than the kind of content that Leibnizian little representations represent.⁸⁴ representation that is completely distinct from all concepts, namely, intuition, but for a cognition through concepts, albeit a confused cognition, which have their seat in the understanding, not sensibility.” Also see B61 62/A44; B326 327/A270 271; B332/A276; ÜE, 8:219–220; FM, 20:278. ⁸¹ See McRae 1976, esp. ch. 5; Brandom 1981; Parkinson 1982; Wilson 2005. ⁸² See Jauernig 2019, 59 60. That Leibniz on occasion uses the expression ‘little perceptions’ to refer to the unconscious representations from which our confused emergent perceptions result should not be taken as evidence that he conceives of them as perceptions, strictly speaking, that is, as sui generis sensible representations. He often uses ‘perception’ as a loose cover term for all kinds of representations, including concepts. ⁸³ See ÜE, 8:217: “The consciousness of an empirical intuition is called perception.” See B207: “Perception is empirical consciousness, i.e., one in which there is sensation at the same time.” See B220: “The general principle of the three analogies rests on the necessary unity of apperception with respect to all possible empirical consciousness (perception) at every time.” See A119 120: “The first that is given to us is appearance which, if it is connected with consciousness, is called perception.” See Prol, 4:300: “At the foundation is intuition of which I am conscious, i.e., perception (perceptio), which merely belongs to the senses.” Also see B376–377/A320; A371; FM, 20:266; FM, 20:274; R5661, 18:319; V-Lo/Dohna, 24:752. ⁸⁴ As will become clear soon, while I do believe that the critical Kant is still fond of Leibniz’s claim that we confusedly perceive the entire empirical universe, I also think that, on Kant’s view, the representational capabilities of empirical intuitions, including ‘little’ empirical intuitions, are limited in certain ways with the result that their

58

      

With this sketch before us, let us get back to the main issue at hand and see how a proponent of the proposal that appearances are ontologically specified by actual, as opposed to possible, perceptions might try to utilize the often neglected Leibnizian strand of Kant’s account of perception just described to account for the existence of unperceived tiny corpuscles, unperceived ‘magnetic matter,’ and unperceived Neptunians. Such a reader would have to first modify the formulation of his proposal slightly to say that appearances are ontologically specified by actual empirical intuitions, not perceptions, in order to accommodate Kant’s terminological choice of reserving the label ‘perception’ for conscious empirical intuitions. He could then try to neutralize the alleged counterexamples by arguing that whenever we perceive a body we also have unconscious empirical intuitions of all the tiny corpuscles that compose it; and whenever we perceive a magnetic object we also unconsciously empirically intuit the ‘magnetic matter’ permeating it (or, rather, we would do so if, indeed, there were such a thing as ‘magnetic matter’). And if Leibniz is indeed very correct, the reader could contend, then we even have unconscious empirical intuitions of Neptunians whenever we perceive anything at all (provided there are some Neptunians around at the time of our perception). Despite these valiant responses to the alleged counterexamples and despite my endorsement of much of the sketched Leibnizian reading of Kant’s account of perception, I still reject the proposal that the representations that ontologically specify appearances are empirical intuitions, be they conscious or unconscious. Before stating my reasons for this rejection, I want to acknowledge that perception undoubtedly plays an important role in Kant’s theory of appearances, though. I agree that Kant is quite clearly committed to the view that being sensible, or in principle perceivable, is a necessary condition for something to be an appearance. Even though the limited range or restricted scope of our senses or unbridgeable distances in space or time, may prohibit us from actually perceiving certain appearances, we would perceive them if we were appropriately located or had more acute or more powerful senses. More generally and more importantly, Kant’s idealism is not an absolute idealism, according to which appearances and the empirical world are ‘made’ by the human mind alone, so to speak, but merely a “formal” one, as he calls it.⁸⁵ Finite minds like ours are ontologically uncreative, on Kant’s view, which means, among other things, that we need ‘outside’ help in constituting appearances and the empirical world. This constitution is a joint venture, as it were, between our mind, which is responsible for supplying the ‘form’ of appearances, and things in themselves, which are responsible for supplying their ‘matter’ and underwriting their existence.⁸⁶ For example, that appearances have shapes that conform to the laws of Euclidean geometry is due to one of the forms of sensibility, but their particular shapes and existence depend, at least in part,

intentional objects are not appearances, even though they are intimately related to appearances. This also means that we do not confusedly perceive the entire empirical universe, on Kant’s critical view, although we may well perceive something that is closely related to the entire empirical universe. ⁸⁵ See B519, note: “I have also otherwise occasionally called it [the doctrine that all objects of possible experience are nothing but appearances] formal idealism in order to distinguish it from the material one, i.e., the common one, which doubts or denies the existence of outer things themselves.” ⁸⁶ I will continue to use scare quotes around ‘form’ and ‘matter’ when these terms are to be understood metaphorically as referring to the formal and material aspects of appearances that we are currently talking about. When used without scare quotes with respect to appearances, the terms are to be understood as referring to the shapes of appearances and the physical matter out of which they are composed.

  

59

on how things in themselves happen to be.⁸⁷ A lot can be said about the contribution of things in themselves to the constitution of appearances, as well as Kant’s use of the form– matter distinction in this context, and, in particular, about what exactly he means by ‘matter’ and I will say more about all that it in the following sections and chapters. For now, the main point to appreciate is that while the ‘form’ of appearances is contributed by us, namely, by our cognitive faculties, due to the limited nature of our mind, their ‘matter’ must be given to us and their existence must be underwritten by something distinct from us. And the way in which the ‘matter’ of appearances is given to us and their existence is underwritten is through perception, and, more specifically, sensation.⁸⁸ Apart from the forms of intuition and some categories, or rather the schemata of some categories,⁸⁹ sensations are the main ‘ingredients’ in perceptions, so to speak, and, in contrast to the other ingredients, which are provided by our cognitive faculties, result from affections of sensibility by things in themselves.⁹⁰ That perception furnishes us with the requisite ‘matter’ for the constitution of appearances and underwrites their existence is a central part of what Kant has in mind when he says, as he often does, that objects are given to us only through sensibility.⁹¹ In this sense, Kant’s statement can be cashed out as meaning that our ability to represent and cognitively access genuinely existing proper objects crucially depends on perception, since perception first supplies us with the ‘matter’ for their representation and provides a criterion for their existence. So, perception is undoubtedly an essential element in the constitution of appearances, on Kant’s account. Nevertheless, there are good reasons not to read Kant as holding that appearances exist and have all of their determinations in virtue of being perceived, or empirically intuited. One reason is that Kant’s standard way of describing appearances is, not as objects of ⁸⁷ It would have been fortunate if Kant had counted the Euclidean nature of the spatial determinations of appearances as part of their ‘matter’ rather than their ‘form.’ This would have made life much easier for the sympathetic commentator who wants to save Kant’s account of space in the face of the discovery of non-Euclidean geometries. It also would have preserved a nice parallel between the spatial properties of appearances and their other qualitative properties that reflect specific causal laws. That appearances are spatial and governed by the laws of some geometry would have been due to us, just as it is due to us that appearances are governed by some kind of causal laws; but which specific geometric laws obtain and which kind of spatial properties appearances possess would have depended on which things in themselves exist, just as which specific causal laws obtain and which kind of qualitative properties appearances possess depends on which things in themselves exist. ⁸⁸ See B609/A581: “The possibility of the objects of the senses is a relation of them to our thinking in which something (namely, the empirical form) can be thought a priori, but what constitutes the matter, the reality in the appearance (what corresponds to sensation) must be given, and without which it could not even be thought and its possibility thus could not be represented.” See B34/A20: “The effect of an object on the faculty of representation, insofar as we are affected by it, is called sensation. . . . I call that in the appearance which corresponds to sensation its matter, but that which makes that the manifold of the appearance can be ordered in certain relations the form of the appearance. Since that in which sensations alone can be ordered and placed in a certain form cannot again itself be sensation, it follows that, while the matter of all appearance is given to us only a posteriori, their form must altogether lie ready for them in the mind a priori and must be able to be considered apart from all sensation.” Note that saying that the existence of appearances must be underwritten by something distinct from us does not imply that this existence is mind-independent after all. It just means that it does not originate in us, and that the special representations that ontologically specify appearances depend on something that is mind-independent. ⁸⁹ The reason for the qualification ‘or rather the schemata of some categories’ will become clear below. ⁹⁰ See B34/A19–20: “The effect of an object on the capacity of representation insofar as we are affected by it is sensation. That intuition which relates to the object through sensation is called empirical.” See ÜE, 8:215: “After . . . asking ‘Who (what) gives sensibility its matter, i.e., the sensations?’ he [Eberhard] believes himself to have spoken against the Critique in saying: ‘We can choose what we want—we end up with things in themselves.’ Now, that is exactly the constant assertion of the Critique; except that it posits this ground of the matter of sensible representations not itself in things, as objects of the senses, but in something supersensible, which is the underlying ground of the former and of which we can have no cognition.” The doctrine of the affection of sensibility by things in themselves will be discussed in detail in chapter 5, where more textual evidence will be considered as well. ⁹¹ See B33/A19; B74 75/A50 51; Prol, 4:288.

60

      

empirical intuition or perception, but as objects of experience.⁹² Experience differs from both of these other kinds of representation; it goes beyond them in involving further cognitive operations. Experience is an empirical cognition, i.e., a cognition that determines an object through perceptions. It is thus a synthesis of perceptions that is not in itself contained in perception but contains the synthetic unity of the manifold of perception in a consciousness, which constitutes the essential part of the cognition of objects of the senses, i.e., of experience (not merely of intuition or sensation of the senses). (B218 219)⁹³

In addition to these direct textual grounds, other reasons for rejecting the proposal that appearances are ontologically specified by empirical intuitions or perceptions can be brought out through an examination of the representational capabilities of these kinds of representations. First, as we have just learned, perceptions, that is, conscious empirical intuitions, are partly indistinct and include many features, for example, color-sensations, that depend on the particular circumstances of the observer, such as the prevailing external circumstances and, possibly, peculiarities of his sense organs and cognitive faculties.⁹⁴ There is no guarantee that the perceptions of any two observers will ever be qualitatively identical, or have the same intentional object.⁹⁵ Of course, there is still a sense, to be explicated in section 2.5, in which both you and I can legitimately be described as perceiving the same appearance, for example, a certain rose. Our perceptions refer to the same public rose. But the specific presentational content that is disclosed to me when I see the rose, what the rose looks like to me, usually and typically is different from the specific presentational content that you are aware of when you see the rose. The intentional object ⁹² See B298/A238 239: “The transcendental use of a concept in any kind of principle is this: that it is related to things in general and in themselves, but the empirical use of it is merely related to appearances, i.e., objects of a possible experience.” See B232/A189: “Thus, accordingly persistence is a necessary condition under which alone appearances, as things or objects, are determinable in a possible experience.” That Kant also refers to the objects that he identifies as appearances as ‘objects of the senses’ should not be taken as evidence that he conceives of them as intentional objects of our perceptions after all. The characterization of something as an object of the senses is compatible both with it being an object of perception and with it being an object of experience; it means as much as that the object is sensible or perceivable, in contrast to supersensible or intelligible. ⁹³ See B195/A156: “The possibility of experience thus is what gives all of our cognitions a priori objective reality. Now experience rests on the synthetic unity of the appearances, i.e., on a synthesis according to concepts of the object of appearances in general, without which it would not even be cognition but a rhapsody of perceptions . . . ” See Prol, 4:298: “We thus must first note that, although all judgments of experience are empirical, i.e., have their ground in the immediate perception of the senses, on the other hand not all empirical judgments are therefore judgments of experience. Rather, apart from the empirical and, in general, apart from what is given in sensible intuition, special concepts still have to be added, concepts that have their origin completely a priori in the understanding and under which each perception first [must be] subsumed and by means of which it then can be transformed into experience.” See Anth, 8:144: “The perceptions of the senses (empirical representations with consciousness) can only be called inner experiences. The understanding that is added and connects them under a rule of thinking (brings order into the manifold) first makes empirical cognition, i.e., experience, out of it.” See B218: “Their principle [the principle of the analogies of experience] is: experience is possible only through the representation of a necessary connection of perceptions.” See R5661, 18:318: “An empirical representation of which I am conscious is perception; that which I think with the representation of the imagination by means of the apprehension and comprehension (comprehensio aesthetica) of the manifold of perception is empirical cognition of the object, and the judgment that expresses an empirical cognition is experience.” ⁹⁴ Kant also explicitly stresses that the colors of the objects of perception depend on the specific circumstances and peculiarities of the observes, see B45/A29–30. ⁹⁵ To be sure, it could happen that distinct cognizers have qualitatively identical perceptions. But the odds are very low, and, even if it were to happen often, this would not take away from the fact that if appearances were ontologically specified by perceptions, the existence of a common world would be a matter of coincidence.

  

61

of either one of our perceptions is not the public rose that exists in the empirical world shared by us but something more private. And since appearances are supposed to be public, they cannot be defined as intentional objects of perceptions. Second, due to the full mind-dependence of appearances, the representations that ontologically specify them must be capable of representing all of their determinations. But, arguably, not all determinations that Kant ascribes to appearances are empirically intuitable or perceivable. This issue is complicated, mostly because of the difficulty to determine with certainty what kind of cognitive operations Kant takes to be involved in the generation of empirical intuitions and perceptions. (I have the suspicion that Kant himself may have wavered on this question.) But here is a possible argument that tracks at least one prominent strand of thinking displayed in Kant’s texts. Insofar as empirical intuitions and perceptions represent objects―albeit in a ‘lightweight’ sense, compared to the objects of experience, as we will see―they depend not only on the forms of sensibility, space and time, and the sensible manifold given to us through the affections of sensibility by things in themselves but also on certain synthesizing and unifying operations by the imagination under the direction of the understanding. I use the circumspect formulation ‘under the direction of the understanding’ because we are not talking about an explicit application of concepts in judgments or an explicit subsumption of something sensibly given under a concept; rather, we are talking about the fact that, in its intuition-generating activity, the imagination brings the same functions of unity to bear on the sensibly given material that are at work in certain basic kinds of judgments and that constitute the core of certain pure concepts of objects⁹⁶ (that is, of certain a priori, non-sensible, or unschematized concepts).⁹⁷ In its synthesizing activity, the imagination thus follows certain rules, rules that Kant calls “schemata” of the concepts that embody the same function of unity.⁹⁸ Now, as I understand Kant, the generation of empirical intuitions and perceptions depends (1) on the concept of the transcendental object, that is, of a bare particular something = X that is (a) distinct from both our representations of it and the cognizer who entertains them and (b) a bearer of determinations, and (2) on the schemata of only two kinds of pure concepts

⁹⁶ See B104 105: “The same function that gives unity to the different representations in a judgment, also gives unity to the mere synthesis of different representations in an intuition, [a function of unity] which, generally expressed, is called the pure concept of the understanding.” Also see B153 154: “It [the understanding], under the name of a transcendental synthesis of the imagination, thus exerts that action on the passive subject whose faculty it is of which we rightfully say that inner sense is thereby affected. Apperception and its synthetic unity is so far from being identical with inner sense that it, as the source of all combination, rather applies to the manifold of intuitions in general, under the name of the categories before all sensible intuition to objects in general; on the other hand, inner sense contains the mere form of intuition but without combination of the manifold in it, and thus not yet any kind of determinate intuition, which is possible only through the consciousness of the determination of inner sense through the transcendental synthesis of the imagination . . . , which I have called the figurative synthesis.” ⁹⁷ A quick terminological comment is in order about the term ‘pure concept,’ which Kant uses in two senses. Pure concepts in the broad sense are a priori concepts. Pure concepts in the narrow sense are a priori concepts that are purely intellectual, that is, that are non-sensible and unschematized. The concepts of Euclidean geometry, such as the concept of a Euclidean triangle, on Kant’s view, are a priori but sensible, since they include their schemata or sensible application conditions. The categories, by contrast, are a priori and purely intellectual, for example, or, at least, the unschematized categories are. In order to avoid confusion, I will use ‘pure concept’ exclusively in the narrow sense, that is, as referring to a priori, purely intellectual, non-sensible, non-schematized concepts, which is also Kant’s primary use. This choice also coheres well with the fact that Kant calls the representations that are produced by and employed by the pure understanding ‘pure concepts,’ and that the pure understanding, by itself, cannot produce any sensible concepts or schemata. ⁹⁸ For a helpful, detailed account of the relation between concepts and their schemata along somewhat similar lines, see Longuenesse 1998, esp. chs 8–11.

62

      

or categories, namely, the categories of quantity and the categories of quality, but not on any relational or modal categories.⁹⁹ On my reading, one can distinguish several logical stages in the constitution of appearances, where the schemata of different groups of categories enter the process at different junctures. Each stage contributes to the constitution of a different aspect of appearances, leaving us with ever more determinate intentional objects. The first stage is the constitution of objects of intuition in general; this stage involves, apart from the sensibly given material and the forms of sensibility, only the concept of the transcendental object and the schemata of the categories of quantity and consists in representing objects with quantitative determinations, broadly speaking, including shape, size, and spatial relations to other objects.¹⁰⁰ This is why Kant chooses the title “axioms of intuition” for the section of the principles of the pure understanding where he discusses the synthetic a priori proposition that describes the most important general feature of appearances that is due to the categories of quantity, namely, that “all appearances, with respect to their intuition, are extensive magnitudes” (A162). The second stage is the constitution of objects of perception or empirical intuition; this stage additionally involves the schemata of the categories of quality and consists in representing objects that have not only quantitative but also certain qualitative determinations. This is why Kant chooses the title “anticipations of perception” for the section of the principles where he discusses the synthetic a priori proposition that describes the most important general feature of appearances that is due to the categories of quality, namely, that “in all appearances the real, which is an object of sensation, has an intensive magnitude, i.e., a degree” (B207). The third stage is the constitution of objects of experience. I will have much more to say about this stage in the following sections―a stage that itself includes several sub-stages―but, to anticipate, in addition to the previously mentioned ‘ingredients,’ it incorporates the schemata of the categories of relation and consists in representing objects that have quantitative, qualitative, and relational determinations, including, in particular, causal and temporal relations.¹⁰¹ This is why Kant chooses the title “analogies of experience” for the section of the principles where he discusses the synthetic a priori propositions that describe the most important general features of appearances that are due to the categories of relation, namely, that “in all change of appearances the substance persists, and its quantum in nature is neither increased nor diminished” (B224), that “all alterations happen according to the law of the connection of cause and effect” (B232), and that “all substances, insofar as they can be perceived in space as simultaneous, stand in thoroughgoing mutual interaction” (B256). For completeness’s sake, the categories of modality and their schemata serve a somewhat different function compared to the other categories; they do not specify any ⁹⁹ Kant identifies twelve basic functions of unity of the understanding, which are listed in the so-called table of judgment (B95/A70), and, correspondingly, twelve basic pure concepts or categories, which are listed in the table of categories (B106/A80). The functions of judgment and with them the categories are sorted into four groups―called ‘quantity,’ ‘quality,’ ‘relation,’ and ‘modality’—each containing three functions or categories, respectively. The function of the concept of the transcendental object in the constitution of the intentional objects of perception and experience will be further examined in sections 2.5.2 and 3.5. ¹⁰⁰ Note that the temporal duration of appearances, which, arguably, is also a quantitative determination, can only be specified at a later stage, which involves additional conceptual machinery. It may also be useful to note that our conscious empirical intuitions only represent those spatial relations of an object that it bears with respect to objects in its close spatial vicinity. ¹⁰¹ These determinations also include spatial relations to objects that are too far away to be perceived, temporal durations, and more determinate qualitative properties (than the ones specified in the previous stage), which are governed by special empirical causal laws.

  

63

kind of determinations of objects but are employed in determining the modal status of objects whose determinations have already been specified.¹⁰² Kant chooses the title “the postulates of empirical thinking in general” for the section of the principles where he discusses the synthetic a priori propositions that express conditions for appearances to be possible, actual, or (hypothetically) necessary.¹⁰³ What all of this means with respect to the main issue on the table right now, that is, the representational capabilities of empirical intuitions and perceptions, is that these kinds of representations do not and cannot represent causal relations or causal powers, which also means that they cannot represent objective temporal relations either.¹⁰⁴ Doing so would require the relational categories, which, however, are not involved in empirical intuition, not even in the unconscious variety. Since appearances have causal powers and stand in causal and objective temporal relations, according to Kant, the proposal that they are ontologically specified by empirical intuitions or perceptions will not do.¹⁰⁵ Note that the claim that we cannot perceive causal relations or causal powers is neither implausible nor idiosyncratic to Kant. Several modern thinkers before Kant, most prominently Hume, already made the same point. We can only perceive spatial and temporal contiguities between events, but we cannot perceive any ‘causal glue’ between them, so to speak. For instance, using an example by Kant, we can perceive sunshine followed by the warming of a stone, but we cannot perceive that the sun warms the stone―and yet, the sun has the causal power to warm stones.¹⁰⁶ In a related vein, it is not implausible to think that the full mind-dependence of appearances furthermore entails that the representations that ontologically specify them must also specify which of their determinations belong to them necessarily or essentially. At least, this thought is not implausible for those of us who, like Kant, hold that appearances have necessary determinations. And, again, this is a desideratum that, most people would grant, perceptions and empirical intuitions do not satisfy. For example, we can perceive of a triangular object that any two of its sides taken together are longer than its third side, but we cannot perceive that this is necessarily so—yet it is essential to all triangular objects that the lengths of their sides are related in this way.¹⁰⁷ It, again, follows ¹⁰² See B266/A219: “The categories of modality are special in that they do not in any way augment the concept, to which they are added as predicates, as determination of the object, but represent only the relation to the faculty of cognition. Even if the concept of a thing is already entirely complete, I can still ask of this object if it is merely possible or also actual or, if it is the latter, whether it is even necessary? Through this no other determinations in the object itself are thought, but it is merely asked how it (with all of its determinations) is related to the understanding and its empirical use, to the empirical capacity to judge, and to reason (in its application to experience)?” ¹⁰³ See B265 266/A218. ¹⁰⁴ Perceptions themselves occur one after the other and, thus, stand in temporal relations. But a temporal succession of perceptions is not the same as a perception of temporal succession. ¹⁰⁵ See B233–234: “ . . . in other words, through mere perception the objective relation of the appearances following each other remains undetermined. In order for this to be cognized as determinate the relation between the two states must be thought in such a way that it is thereby determined as necessary which one of them must be posited before, which one afterwards, and not the other way around. However, the concept that carries with it a necessity of the synthetic unity can only be a pure concept of the understanding, which does not lie in perception, and this is here the concept of the relation of cause and effect, of which the former determines the latter in time, as the consequent, and not as something that could precede merely in the imagination. . . . Thus only through our subsumption of the sequence of the appearances, and thus of all alterations, under the law of causality is experience itself, i.e., empirical cognition of them, possible; hence, they themselves, as objects of experience, are possible only according to that law.” ¹⁰⁶ See Prol, 4:301, note. Of course, in order to come to know specific causal connections we need perceptions, but no perception, by itself, amounts to a representation of a causal connection. ¹⁰⁷ See Letter to Herz, May 26, 1789, 11:51: “ . . . even if were capable of an intellectual intuition . . . the necessity of such judgments [synthetic a priori judgments] . . . would not obtain. For it would always be mere perception,

64

      

that appearances cannot be defined as the intentional objects of empirical intuitions or perceptions.

2.5 Appearances, Perception, and Experience 2.5.1 Appearances are Intentional Objects of Experience Appearances exist and have all of their determinations and other ontological ingredients in virtue of being represented in experience, which, like ‘appearance,’ is to be understood as relativized to human beings in the sense of ‘experience for us’ or ‘human experience.’ But this [that the determinate size of the world lies in itself] contradicts the concept of a sensible world, which is nothing but a sum total [Inbegriff] of appearance, whose existence and connection takes place only in representation, namely, in experience, since it is not a thing in itself but itself nothing but a kind of representation . . . To assume that an appearance . . . contains all parts in itself before all experience means to ascribe to a mere appearance, which can exist only in experience, at the same time a special existence that precedes experience, or to say that mere representations are there before they are found in the power of representation, which is contradictory . . . . (Prol, 4:342)

Of course, the general claim that, for Kant, appearances somehow depend on experience is not news. But there are many different ways of spelling out what exactly this is supposed to mean. The remainder of this chapter further spells out my reading. We will start by taking a closer look at the formal and material conditions of experience proposed by Kant and by saying more about the relation between experience and perception. (As indicated earlier, for now we are primarily concerned with outer appearances and hence outer experience. So, ‘experience’ without any qualifications should be understood as ‘outer experience’ unless indicated otherwise.)

2.5.2 The Formal Conditions of Experience As noted already, not all presentational contents are intentional objects. Minimally, in order for the presentational content of a representation to count as an intentional object, the representation must present a particular something that is (a) distinct from both the representation and the cognizer entertaining the representation and (b) a bearer of certain determinations. But there is yet an even more restrictive use of ‘object’ where it is understood as what Strawson calls ‘object in the weighty sense.’¹⁰⁸ Empirical objects such as tables, trees, and planets, are examples of objects in this more restrictive, weighty sense; illusory raccoons, dreamed dragons, and hallucinated oases are not. I will refer to objects in the more restrictive, weighty sense as ‘proper objects.’ One of Kant’s main e.g., that in a triangle two sides taken together are greater than the third, but not that this property must necessarily pertain to it.” Also see A24. ¹⁰⁸ See Strawson 1966, 73–4.

, ,  

65

questions in the Critique is how we (directly) represent such proper objects, or what kind of representations have proper objects as their intentional objects and, thus, can be called objective, strictly speaking. The answer to this question is provided by his theory of experience, whose center piece consists in the identification of conditions for the possibility of us representing proper objects. Appearances are proper objects, and we represent proper objects in experience. It is widely acknowledged that one of Kant’s most seminal insights is the recognition that the way in which we represent proper objects reflects the nature of our mind―although, of course, there are many different interpretations of how exactly Kant’s account of our representation of proper objects is supposed to be understood and what it implies for his conception of appearances. In the following, I will briefly sketch my reading.¹⁰⁹ The forms of our cognitive faculties―space, time, and the logical forms of judgments, or, more specifically, the pure categories, and, more generally, the original synthetic unity of apperception―are the ultimate a priori conditions for the possibility of us representing proper objects, conditions that are better known as the formal conditions for the possibility of experience.¹¹⁰ According to Kant’s analysis, these formal conditions enable us to represent objects in certain ways that are necessary for them to count as proper objects for us. In order for a representation to (directly) represent a proper object, and, thus, be objective, strictly speaking, it must present its object as (a) unified or one, (b) outside us, that is, as mind-independent and distinct from both itself (the representation) and our mind, and (c) part of a unified system of objects. That the concept of the transcendental object, that is, of a bare particular something = X that is (i) distinct from both the representations of it and the cognizer entertaining them and (ii) a bearer of determinations, arises as an objective reflection of the transcendental unity of apperception, contributes to our ability to have representations that accomplish (a) and, together with space being a form of sensibility, allows us to present objects as being in space, which makes us capable of representations that accomplish (b).¹¹¹ That, in addition to space, time is also a form of sensibility and the categories are forms of the understanding allows us to present objects as distinct from one another, as being necessarily governed by mathematical laws, and as occupying determinate positions in a system of objects in space and time that is necessarily governed by universal causal laws, which also contributes to our ability to have representations that accomplish (a) and enables us to have representations that accomplish ¹⁰⁹ A fuller account of Kant’s theory of cognition in general, of which his theory of experience is a central part, can be found in Jauernig (in preparation). ¹¹⁰ For the terminology of ‘formal conditions for the possibility of experience,’ see B271/A224: “That space is a formal condition a priori of outer experiences . . . ” Also see B283/A230: “The understanding gives a priori to experience in general only the rule according to the subjective and formal conditions of sensibility as well as of apperception, which alone make it possible.” ¹¹¹ See A250: “All our representations are in fact related to some object by the understanding, and since appearances are nothing but representations, the understanding relates them to something, as the object of sensible intuition: but this something is to that extent only the transcendental object. But this signifies a something = x, of which we do not know anything at all nor in general can know (according to the current constitution of our understanding) but which only serves as a correlate of the unity of apperception for the unity of the manifold in the sensible intuition, by means of which the understanding unifies the manifold in the concept of an object.” See B37/A22: “By means of outer sense, (a quality of our mind,) we represent objects as outside us and these [objects] altogether as in space.” See Prol, 4:336: “Empirically outside of me is that which is intuited in space.” See B38/A23: “For in order to relate certain sensations to something outside of me (that is, to something in a different place of space than the one in which I am), and similarly in order to represent them as outside and next to one another and, hence, not merely as different but as in different places, the representation of space must already be there as a ground.”

66

      

(c),¹¹² largely thanks to the fact that the categories and their schemata embody the original synthetic unity of apperception―or the transcendental unity of consciousness, as Kant also calls it―which functions as a kind of ur-unity or ultimate template of unity for all unifying activities of our cognitive faculties.¹¹³ If we fill in the details by paying attention to the particular unifying functions at work in the different categories and their schemata, we thus arrive at the following list of necessary conditions that a representation of ours must satisfy in order to (directly) represent a proper object and count as a piece of experience, or, more precisely, outer experience, conditions that are direct manifestations of the forms of our cognitive faculties. A representation (directly) represents a proper object, counts as a piece of outer experience, and is objective, strictly speaking, only if it presents its object as: (1) a unified bearer of determinations, or a bearer of determinations that is one; (2) extended and in space; (3) an extensive magnitude and, hence, necessarily governed by the laws of mathematics, in particular, Euclidean geometry and standard arithmetic;¹¹⁴ (4) having determinations that have intensive magnitudes and, as such, are also necessarily governed by the laws of mathematics;¹¹⁵ (5) persisting through time and remaining unchanged with respect to its quantum;¹¹⁶ (6) having states that are part of an objective temporal order and change only in response to the object’s being empirically affected by an external cause;¹¹⁷ and (7) mutually causally interacting with all objects that are simultaneous with it.¹¹⁸ In the following, I will refer to these conditions as the ‘formal conditions of experience’ or ‘formal conditions of objectivity.’ It is also useful to note that all representations that are pieces of the same (outer) experience must be coordinated in such a way that they add up to an overall coherent story that describes a possible empirical world, which requires, among other things, that they all present their objects as having states that belong to the same objective temporal order and as being governed by the same causal laws, and that they do not present distinct objects as occupying the same place at the same time. I will call this requirement ‘global formal condition of experience’ and capture it by saying that a collection of representations of which each conforms to the formal conditions of experience amounts to outer experience only if all of their intentional objects are compossible. Just as any representation of ours that (directly) represents a proper object must conform to the formal conditions of experience and objectivity, any object that is a proper object for ¹¹² See B125–126/A93: “Now the question is whether concepts not also precede a priori as conditions under which alone something is, if not intuited, then still thought as an object in general, for in that case all empirical cognition of objects necessarily conforms to such concepts since, without their presupposition, nothing is possible as an object of experience.” See B244–245/A199: “The understanding belongs to all experience and its possibility, and the first thing that it contributes to it is, not that it makes the representation of objects distinct, but that it makes the representation of an object possible at all.” ¹¹³ See B137: “An object is that in whose concept the manifold of a given intuition is united. Therefore, the unity of consciousness is that which alone constitutes the relation of the representations to an object, and thus their objective validity, and, accordingly, that they become cognitions, and on which thus rests the possibility of the understanding.” See B138: “The synthetic unity of consciousness is thus an objective condition of all cognition, not one that I myself need merely to cognize an object, but one under which each intuition must stand in order to become an object for me, since in another way, and without this synthesis, the manifold would not unite itself in a consciousness.” Also see B103 105/A78 79; B131 139; B150 152; B159 161. ¹¹⁴ See B202–207/A162–166. ¹¹⁵ See B207–218/A166–176. Also see B221/A178–179. ¹¹⁶ See B224–232/A182–189. ¹¹⁷ See B232–256/A189–211. The qualification that the affections in question are empirical is added in order to distinguish them from the transcendental affections that things in themselves exert on sensibility. We will take a closer look at these two kinds of affections in section 2.5.3. ¹¹⁸ See B258: “Hence, the simultaneity of substances in space cannot be cognized in experience in any other way than by assuming a mutual interaction among them; this [mutual interaction] is thus also a condition for the possibility of the things themselves as objects of experience.” See B258–62/A211–15.

, ,  

67

us must conform to corresponding ‘formal conditions of proper objecthood,’ as we may call them.¹¹⁹ For example, all objects that are proper objects for us must be in space and all changes in their states must be necessary effects of empirical affections by external causes. This is also why and how we can know a priori that the synthetic principles expressing the formal conditions of experience and objectivity are valid for appearances, the intentional objects of experience. In Kant’s famous words: The conditions for the possibility of experience in general are at the same time conditions for the possibility of the objects of experience, and thus have objective validity in a synthetic judgment a priori. (B197/A158)¹²⁰

Incidentally, as perceptive readers will have noticed, this also means that things in themselves are not proper objects for us.¹²¹ For, as just explicated, Kant’s formal conditions of proper objecthood include that all objects that are proper objects for us are in space and time. But, as we will discuss in chapters 3 and 4, things in themselves are demonstrably not in space and time. We simply do not have any conception of an equivalent of a proper object at the transcendental level of reality.¹²² Empirical intuitions and perceptions conform to some of Kant’s formal conditions of objectivity but not to all of them. This is not surprising, given that these representations do not involve the full range of categories. For example, since they do not involve the category of causality, they do not conform to the condition that objective representations in the ¹¹⁹ Kant himself does not terminologically distinguish between the formal conditions for the possibility of experience (the forms of our cognitive faculties), the formal conditions of experience/objectivity (the necessary conditions listed above that any representation of ours that (directly) represents a proper object and is a piece of experience must conform to), and the formal conditions of proper objecthood (the necessary conditions for an object to be a proper object for us); he tends to lump all of these conditions together under the label ‘formal conditions (for the possibility) of experience.’ But for the sake of clarity and precision, it is desirable to work with the indicated more fine-grained terminology, even though all of these conditions are very closely related. Similar remarks apply with respect to the material conditions (for the possibility) of experience, to be discussed in section 2.5.3. ¹²⁰ Also see Prol, 4:296: “For the subjective laws under which alone an empirical cognition of things is possible, also hold for these things as objects of a possible experience.” ¹²¹ This is part of the reason why Kant says that we cannot cognize things in themselves, strictly speaking. Cognitions, strictly speaking, are reserved for proper objects. ¹²² I want to stress explicitly that, on my interpretation, the dependence of our conception of a proper object on the forms of our cognitive faculties is not what Kant’s idealism about empirical objects amounts to. By contrast, this is a core element of Henry Allison’s explication of Kant’s idealism. In his own words: “The concept of an epistemic condition brings with it an idealistic commitment . . . , because it involves the relativization of the concept of an object to human cognition and the conditions of its representation of objects. In other words, the claim is not that things transcending the conditions of human cognition cannot exist (this would make these conditions ontological rather than epistemic) but merely that such things cannot count as objects for us. This also appears to be the sense of Kant’s famous ‘Copernican hypothesis’ that objects must ‘conform to our cognition’ . . . this means that objects must conform to the conditions of their representation; not that they exist in the mind in the manner of Berkeleian ideas or the sense data of phenomenalism” (Allison 2004, 12). In response, I agree that the dependence of the concept of a proper object on the conditions for our representation of proper objects does not mean that things in themselves do not exist but merely that they do not count as proper objects for us. But I emphatically do not agree that this is the sense of Kant’s famous Copernican hypothesis (which we will discuss in section 3.1), nor do I agree that it captures the meaning of Kant’s idealism. To be sure, the recognition that empirical objects must conform to the conditions of their representation is important on Kant’s account. But what this recognition entails, via an additional important argument, is precisely that empirical objects exist ‘in the mind,’ that is, that they are fully mind-dependent. It is this ontological thesis, that empirical objects are fully minddependent, that constitutes Kant’s idealism, a thesis that is part and parcel of the Copernican hypothesis, as we will see. The indicated additional important argument that underwrites Kant’s idealism is that anything that conforms to the sensible conditions of experience, that is, space and time, is fully mind-dependent because space and time are nothing but forms of sensibility and, hence, fully mind-dependent. We will consider this argument in detail in chapter 4.

68

      

strict sense must present the changes in the states of their objects as necessary effects of affections of these objects by external causes. This also means that they cannot present their objects as belonging to an objective temporal order, or as part of a unified system of objects in mutual causal interactions, which, again, would be required for these representations to be objective, strictly speaking. Empirical intuitions and perceptions thus differ from experience with respect to their degree of objectivity, as we might say, that is, their degree of conformity to Kant’s formal conditions of objectivity. The former kinds of representations are objective to some degree but much less so than experience.¹²³ This contrast is reflected in Kant’s distinction in the Prolegomena between judgments of perception and judgments of experience, that is, between empirical judgments that are “merely subjectively valid” and “do not require a pure concept of the understanding but merely the logical connection of perceptions in a thinking subject,” on the one hand, and empirical judgments that have “objective validity” and “still always require, in addition to the representations of sensible intuition, special concepts that are originally generated in the understanding, which bring it about that the judgment of experience is objectively valid ” (Prol, 4:298).¹²⁴ Thanks to its conformity to the formal conditions of objectivity, experience satisfies the demands identified at the end of section 2.4 that any representation that ontologically specifies appearances must be capable of (a) representing all of the determinations of appearances, including their causal relations and causal powers, and (b) specifying which of these determinations are essential, or necessary, to them. Experience is capable of representing that the sun has the causal power to warm stones because it presents its objects as governed by universal causal laws, in this case, a thermodynamic law that regulates heat transfer between objects. Similarly, experience is capable of representing that every triangular object necessarily has the property of being such that any two of its sides taken together are longer than the third side because it presents its objects as governed by the laws of Euclidean geometry, and every Euclidean triangle essentially has the indicated property.¹²⁵ The conformity of experience to the ¹²³ Conversely, perceptions and experience are both subjective to some degree, but perceptions are much more subjective than experience. Experience is subjective to some extent in that its content depends on the forms of the cognitive faculties of human subjects. This is the sense in which Kant also often describes space as a “subjective representation” or a “subjective form of our intuition.” See B44/A28; FM 20:268. ¹²⁴ See Prol, 4:297–302. Also see the references in note 93. The cited characterization of the difference between judgments of perception and judgments of experience raises the question of how it fits together with Kant’s apparent commitment to the view that at least some of the categories are involved in perception. How can this view be reconciled with the Prolegomena claim that judgments of perception ‘do not require a pure concept of the understanding’? Here are two options for how to respond. First, one could say that in the Prolegomena passage Kant is really only talking about the categories of relation, in particular, the category of causality, which, indeed, is not involved in perception. It speaks for this reading that, in the passage, Kant continues to explain that judgments of perception require “merely the logical connection of perceptions in a subject,” that is, a connection by means of the logical function of hypothetical judgment, as opposed to a connection by means of the categories of relation. Second, one could hold that in the Prolegomena passage Kant should be understood as saying that, in contrast to judgments of experience, judgments of perception do not require the explicit employment of the categories in the form of concepts under which the objects of our empirical intuitions are subsumed, although perception does require the implicit employment of the categories in the form of their schemata, as briefly explicated in section 2.4. My own view is that both of these options for how to answer the question at hand track Kant’s actual view. That is, I read Kant as holding that (1) perception only involves the categories of quantity and quality (2) these categories are only implicitly employed in perception in the form of their schemata, and (3) in contrast to judgments of perception, judgments of experience depend on the explicit application of the categories, including, in particular, the category of causality. ¹²⁵ But, one might wonder, do the claims that experience can represent that the sun’s heat is necessarily connected with the warming of a stone, and that a triangular object necessarily has the property that any two of its sides taken together are longer than its third side, not conflict with Kant’s insistence that necessity is a mark of the

, ,  

69

formal conditions of objectivity also begins to shed light on why and how its intentional objects are public, in contrast to the intentional objects of perceptions, which are a more private, context-dependent affair. Finally, it is also worth making explicit that, in contrast to empirical intuitions, which can be unconscious, experience is essentially conscious; and, in contrast to perceptions, which supervene on a manifold of infinitely many unconscious ‘little’ empirical intuitions and are generated with the help of the mysterious schemaguided figurative synthesis of the imagination, “a blind, albeit indispensable function of the soul . . . of which we are rarely conscious” (B103/A78), experience is conscious through and through, so to speak, and is constructed by means of the conscious, explicit application of concepts in judgments and the conscious, explicit subsumption of objects under concepts, on which more in sections 2.3. and 2.6.

2.5.3 The Material Conditions of Experience Experience must be carefully distinguished from perception and empirical intuition. But, as already noted, Kant also holds that the constitution of appearances is an enterprise with two main contributors, namely, our cognitive faculties, which provide the ‘form’ of appearances, and things in themselves, which provide the ‘matter’ and underwrite the existence of appearances by affecting sensibility and bringing about sensations. Appearances are not purely formal but also sport some material flesh on their formal bones, as it were.¹²⁶ They are not merely spatial in general; they have particular shapes and stand in particular spatial relations to particular other appearances. Similarly, appearances are not governed merely by the general law that all changes in their states are due to external causes; they have particular qualitative determinations, are governed by corresponding specific empirical causal laws that determine which changes of which of their qualitative determinations are due to what kind of causal influences, and are causally connected by particular causal relations to particular other appearances. Moreover, appearances exist not only in their special intentional ‘world’ as Frodo exists in the world of The Lord of the Rings―but also simpliciter, from the point of view of fundamental ontology. And, according to Kant, due to the limited, ontologically uncreative nature of our mind, the ‘matter’ of appearances must be given to us and their existence must be underwritten through perceptions, and, more specifically, sensations. This means that perception plays a crucial role in the construction of experience. Perceptions, and, more

a priori, and that empirical judgments can only inform us about how things are but not how they have to be? See B3–B6. This worry is based on a confusion of empirical judgments understood as judgments that are based solely on the information provided by perception and judgments of experience understood as judgments that express our experience in the sense presently under consideration. As just explained, experience is based on both sense perception and certain a priori concepts and laws. And it is thanks to those a priori concepts and laws that it can represent the indicated necessary features. See the famous line at B1: “Although all our cognition begins with experience, it does not follow that it all originates in experience. For it could very well be that even our cognition of experience is a compound out of that which we receive through impressions, and that which our own faculty of cognition (merely occasioned by sensible impressions) provides from itself . . . ” ¹²⁶ See B207 208: “Appearances, as objects of perception, are not pure (merely formal) intuitions, like space and time (for those cannot at all be perceived by themselves). Thus, they contain in themselves, in addition to intuition, also the matter for some object in general (by means of which something existing in space or time is represented), i.e., the real of the sensation as a merely subjective representation, of which one can only become conscious that the subject is affected and which one relates to an object in general.”

70

      

precisely, sensations, are a posteriori, material conditions for the possibility of experience.¹²⁷ And since sensations are the result of affections of sensibility by things in themselves, things in themselves can be identified as the ultimate material conditions for the possibility of experience. We will now take a closer look at this role of perceptions and sensations as the material conditions for the possibility of experience. A good place to start is with a brief examination of Kant’s ‘matter’ metaphor. To my mind, one must distinguish a more literal and a less literal sense of his characterization that the ‘matter’ of appearances corresponds to sensation.¹²⁸ In a more literal sense, this characterization applies, not to the intentional objects of experience, but to the intentional objects of perceptions, which, due to their status as intentional objects, can also be called ‘appearances’ in a general sense. More specifically, these objects are appearances of intentional objects of experience, that is, they are appearances of (Kantian) appearances, or empirical appearances.¹²⁹ Sensations—or more precisely the qualia corresponding to sensations—such as particular color-sensations or touch-sensations, are literally the material out of which the intentional objects of perceptions are constituted, a constitution that is effected by us intuiting these qualia as arranged in particular shapes in accordance with a certain general scheme of order, namely, Euclidean space, which is a form of sensibility, under the guidance of the schemata of the categories of quantity and quality. But when we turn to the intentional objects of experience, the characterization that the ‘matter’ of appearances corresponds to sensation is to be taken in a less literal sense. In this case, the characterization must be read against the background of Kant’s conception of appearances as depending both on our cognitive faculties and on things in themselves, and ‘matter’ must be understood as a sort of black-box placeholder for all those aspects of appearances that are not due to the nature of our cognitive faculties. The claim that the ‘matter’ of appearances corresponds to sensation, then, is to be read as a short-hand expression of the general thesis that whatever features of appearances are not due to the nature of our cognitive faculties are ultimately supplied to us by way of perceptions, which include sensations as one of their central ‘ingredients.’ It is also worth noting the artificiality involved in the neat separation between the formal features of appearances, which are due to our cognitive faculties, and their material features, which are due to things in themselves that affect us, that is suggested by the ‘form’–‘matter’ distinction. While we can, indeed, isolate and explicitly describe the formal aspects of appearances by analyzing the nature of our mind and identifying the forms of our cognitive faculties, their ‘matter’ is available to us only in form of ‘informed matter’ or ‘materialized (and hence particularized) form,’ as it were. For example, take the characteristic rose-shape of the rose in the vase in front of me, which is due to both the nature of our cognitive faculties and the things in themselves that affect us. As far as the contribution of our cognitive faculties is concerned, we can say more precisely that the nature of ¹²⁷ For the terminology of ‘material conditions for the possibility of experience,’ see for example B266/A218: “That which is connected with the material conditions of experience (sensation) is actual.” ¹²⁸ For textual evidence, see notes 88 and 90. ¹²⁹ As we will discuss in chapters 3 and 6, in addition to the transcendental distinction between things in themselves and appearances, Kant also recognizes an empirical version of the distinction. In terms of the latter, the intentional objects of our perceptions can be described as empirical appearances of the intentional objects of experience, which, in turn, are empirical things in themselves but transcendental appearances. I will continue to use ‘appearance’ without any qualification in the transcendental sense, as referring to the intentional objects of experience; if the empirical sense is intended, I will explicitly say so.

, ,  

71

sensibility is responsible for the rose’s spatiality and the conformity of its shape to the postulates of Euclidean geometry. But all that we can say about the contribution of things in themselves is that they are somehow involved in determining its shape to be the particular rose-shape it is. Or take the warming of a stone by the sun, which, again, depends both on the nature of our cognitive faculties and the things in themselves that affect us. With respect to the contribution of our cognitive faculties, we can elaborate that the sun shining and the stone warming are connected by a causal law because of the nature of the understanding, but all that can be said more specifically about the contribution of things in themselves is that they are somehow involved in determining one of the many special empirical laws of nature that are grounded in the analogies of experience to be a particular thermodynamic law that governs heat transfers between objects, as well as in determining the sun and the stone to have certain qualitative determinations and certain relative spatial positions in virtue of which said law applies to them. Finally, it is also useful to point out that the ‘matter’ of the intentional objects of experience depends only on what things in themselves affect us, while the ‘matter’ of the intentional objects of perceptions also depends on special features of the cognizer entertaining the perception, for example, certain idiosyncracies of their sense organs or special features of the empirical circumstances in which the perception is made, for example, the prevailing lighting conditions.¹³⁰ For instance, the specific rose-shape of the rose in the empirical world to which my current rose-perception refers, in a sense to be specified below, depends only on the thing in itself (or things in themselves) that affect us and the nature of our cognitive faculties in general, while the slightly foreshortened rose-shape of the rose that is the intentional object of my perception in addition depends on the special perspective from which I happen to observe the rose. This independence of the ‘matter’ of appearances from special empirical circumstances or peculiarities of individual observers constitutes another element in the explanation of the public nature of appearances, in contrast to the more private nature of the intentional objects of perceptions. So, then, how does perception give us the ‘matter’ and underwrite the existence of appearances? What exactly is the connection between perception and experience such that the ‘matter’ of appearances is derived from and their existence is established through perception? As anticipated, the construction of experience involves various stages. Also, ¹³⁰ On the assumption that appearances are ontologically specified by actual experience, in order for this claim to be true as stated, the development and progress of science and technology, which, as we will see, is part and parcel of the construction of actual experience, cannot depend on special intrinsic features of individual cognizers or special empirical circumstances. To see this, assume that the discovery of Coulomb’s law was due to Coulomb’s special creativity so that, if it had not been for him, the law would never have been discovered. In that case, the ‘informed matter’ or ‘materialized form’ of a given appearance of having a certain electrical charge and of attracting or repelling other charged appearances in accordance with Coulomb’s law would depend not only on what things in themselves affect us and the nature of our cognitive faculties in general but also on Coulomb’s special creativity. There is a more careful way of characterizing the difference of the grounds of the ‘matter’ of the intentional objects of perception and the grounds of the ‘matter’ of the intentional objects of experience that takes this complication into account, which exploits the fact that the way in which the indicated ‘informed matter’ depends on Coulomb’s special creativity is quite different from the way in which, say, the particular shade of red of the intentional object of my rose-perception depends on the special make-up of my eyes and the prevailing lighting conditions. But since spelling out this more careful characterization turns out to be a rather tedious and complex affair, I will simply ignore the indicated complication and just help myself to the not implausible assumption that, although on a local scale the development and progress of science and technology may depend on special features of individual cognizers or special empirical circumstances, over time this influence is ‘washed out, as it were, and becomes negligible. Coulomb’s law was bound to be discovered sooner or later; it was only a question of time.

72

      

experience is preceded and its construction is continually fueled by perception. More specifically, an important, recurring step in the constitution of appearances and the construction of experience is the transition from intentional objects of perception to intentional objects of experience and from judgments of perception to judgments of experience. In this step, we move from subjective representations of private objects and their spatial and subjective temporal relations, to objective representations of public objects and their spatial, causal, and objective temporal relations. As noted, this transition is brought about primarily by means of the explicit application of concepts, including, in particular, concepts that are grounded in the relational categories. On my reading, a crucial part of this transition consists in presenting, for each private intentional object of a perception, a public object as the proximate cause of the sensations contained in the perception. For example, the transition from my current perception that represents a private, slightly foreshortened red rose to a representation that is a piece of experience and represents a public rose that exists in the empirical world, crucially involves presenting the public rose as the proximate cause of my current red-sensations that are arranged in a slightly foreshortened rose-shape.¹³¹ Appreciating this transition in the construction of experience, from the private intentional objects of perception to public objects that are presented as the proximate causes of the sensations contained in these perceptions, is the key to seeing how perception contributes the ‘matter’ and underwrites the existence of appearances. More precisely, regarding the ‘matter’-giving function of perception, presenting objects as proximate causes of certain sensations and their particular arrangement requires ascribing various quantitative and qualitative determinations to them, as well as specifying empirical causal laws that govern their causal powers, including, in particular, the causal power to cause the relevant sensations in human observers. But this is a major part of what it means to give a certain kind of ‘matter’ to these objects. Moreover, once the indicated connection between appearances and the intentional objects of perceptions has been established, the observation of regularities in the behavior of the latter allows us to further refine and expand the ‘matter’ of the former, by way of ascribing additional quantitative and qualitative determinations to them and specifying additional empirical causal laws, so as to account for the observed regularities. Turning to the existence-underwriting function of perception, according to Kant’s analysis, in “empirical thinking in general,” that is, in thinking that involves empirical concepts and conforms to the formal conditions of experience, we present objects as actually existing by presenting them as causally connected to our sensations, in accordance with the postulate that “that which is connected with the material conditions of experience (sensation) is actual” (B266/A218).¹³² Given the described role of perception in the ¹³¹ This application of the category of causality in the constitution of appearances is also emphasized by Schopenhauer. But, in contrast to Kant, Schopenhauer takes this application to be part of the genesis of our empirical intuitions, and, thus, presents it as an objection to Kant’s conception of empirical intuition as independent from the category of causality. See Schopenhauer 1847, §21, Werke, 3:159 92; Schopenhauer 1859, §4, Werke, 1:13 15. ¹³² Since the intentional objects of experience actually exist, Kant cannot call the postulates that describe the application of the modal categories, which include the category of possibility, ‘postulates of experience.’ Accordingly, he chooses the more general title ‘postulates of empirical thinking in general.’ As indicated in the main text, on my reading, ‘empirical thinking in general’ is thinking that involves empirical concepts and conforms to the formal conditions of experience. Objects of empirical thinking in general are or could be intentional objects of experience, holding fixed the nature of our cognitive faculties.

, ,  

73

construction of experience as explicated so far, and given that experience presents its objects as causally connected to each other, this postulate allows us to make explicit two perception-based conditions for the presentation of objects in empirical thinking as actually existing, conditions that are individually sufficient and disjunctively necessary,¹³³ and that can be expanded into necessary and sufficient conditions for the existence of intentional objects of empirical thinking at the empirical level of reality.¹³⁴ An object O of a piece of empirical thinking ET is presented as actually existing if, and only if, there is a perception P, and (a) ET presents O as the proximate cause of the sensations contained in P, or (b) ET presents O as causally connected to O*, which it presents as the proximate cause of the sensations contained in P.¹³⁵ For example, my current empirical thinking presents a public, undistorted rose as actually existing by presenting it as the proximate cause of my current rose-perception that represents a red, slightly foreshortened rose. Or an astronaut’s empirical thinking presents Neptunians as actually existing by presenting, (i) a public square photograph of a group of smirking Neptunians as the proximate cause of his perception that represents a slightly foreshortened photograph of a group of smirking Neptunians, and, (ii) a group of smirking Neptunians as causally connected to that very photograph. In the following passages, Kant primarily has the second of these two forms of indirect-perception-based representation of actual existence in mind, but they can serve as excellent expressions of both of them. The postulate to cognize the actuality of things requires perception, and thus sensation, of which one is conscious, albeit not immediately of the object itself, whose existence is supposed to be cognized, but still its connection with some actual perception, in accordance with the analogies of experience, which represent all real connection in an experience in general. (B272/A225)

¹³³ These conditions are disjunctively necessary because (a) presenting an object as causally connected to sensations is the only way identified by Kant in which we can represent an object’s actual existence in empirical thinking, and (b) the two ways in which objects of outer empirical thinking can be causally connected to sensations that are featured in these two conditions respectively appear to be exhaustive. ¹³⁴ The conditions need to be expanded slightly in order to function as conditions for the existence of intentional objects of empirical thinking at the empirical level of reality because, arguably, there could be other representations apart from experience that both conform to the unexpanded version of these conditions and the formal conditions of experience, but we do not want to say that the intentional objects of these other representations exist at the empirical level of reality. In this context, it is relevant that some of Kant’s pronouncements about his formal conditions of experience suggest that he thinks of them as sufficient for distinguishing experience from all other kinds of representations. (Some of what he says in the version of the refutation of idealism in the Prolegomena, to be discussed in section 4.1.2, could be heard in this way. Also see B479/AA451: “ . . . in that case the connection of appearances that necessarily determine one another in accordance with general laws, which is called nature, would mostly disappear, and with it the criterion of empirical truth, which distinguishes experience from a dream.”) If that were true, we would not have to worry about expanding the conditions presently under discussion before being able to use them as necessary and sufficient conditions for the actual existence of the intentional objects of empirical thinking at the empirical level of reality. I take it that this is why Kant himself does not worry about it. ¹³⁵ We have not officially decided yet whether appearances are ontologically specified by possible or actual experience. As anticipated, my ultimate view is that appearances are ontologically specified by actual experience, a view for which I will argue in section 2.8. The stated conditions for actual existence are formulated with that ultimate view in mind. Proponents of the view that appearances are ontologically specified by possible experience should add the clarification that an object O of a piece of possible empirical thinking is presented as actually existing if, and only if, there could be a perception P, and (a) ET presents O as the proximate cause of the sensations contained in P, or (b) ET presents O as causally connected to O*, which it presents as the proximate cause of the sensations contained in P.

74

       The only thing that can be added to my understanding is something beyond the conformity to the formal conditions of experience, namely, the connection with some perception; but whatever is connected with this perception according to empirical laws is actual, even if it is not immediately perceived. (B284/A231)¹³⁶

Since appearances would not exist if they were not presented as existing, their existence is thus closely tied to perception. We will return to the question of how to formulate conditions for the actual existence of intentional objects of empirical thinking at the empirical level of reality later on in this section. Kant’s formal conditions of experience are direct reflections of the formal conditions for the possibility of experience, that is, the forms of our cognitive faculties. Armed with the results of our foregoing discussion of the role of perception in the construction of experience, we can attempt to offer a similar translation of the material conditions for the possibility of experience, that is, the conditions that, due to the ontologically uncreative nature of our mind, we need sensations and perceptions for the construction of experience, into material conditions of experience, that is, material conditions that any representation must satisfy that is to count as a piece of experience. So, here is a possible, albeit still somewhat ‘hand-wavy,’ formulation of such material conditions of experience, or, more precisely, outer experience, conditions that, just like the formal conditions, are to be understood as necessary. A representation R counts as a piece of outer experience only if (1) (a) R presents its object O as having some ‘matter,’ and (b) this ‘matter’ does not depend on special features of individual cognizers or special empirical circumstances, and (2) there is a perception P, and (a) R presents O as the proximate cause of the sensations contained in P, or (b) R presents O as causally connected to O*, which, in turn, it presents as the proximate cause of the sensations contained in P.¹³⁷ It is also useful to note that, in addition to the functions of perception in the construction of experience already identified, Kant quite clearly also regards it as providing global constraints on this construction. In the construction of experience, we are not only required to find a perception for each object that we want to represent in it so that we can present this object as causally connected to the perception, we are also required to incorporate into experience all perceptions, past and present, that are available to us at that moment—be it directly or through records of any kind—by presenting, for each available perception, an object as its proximate cause. I will

¹³⁶ Also see B273/A225: “For that the concept precedes perception signifies its mere possibility; but perception, which supplies the material for the concept, is the only character of actuality.” See B521/A493: “For in themselves appearances, as mere representations, are actual only in perception, which, indeed, is nothing other than the actuality of an empirical representation, i.e., appearance.” See V-Met-L2/Pölitz, 28:557: “I cannot cognize the existence of a thing entirely a priori. . . . Something apart from thinking has to be added and this is the intuition of something actual, or perception. Perception is the representation of the actual.” ¹³⁷ Are the stated material conditions also conditions of objectivity, that is, conditions for a representation to be objective, strictly speaking, or (directly) represent a proper object? Some of them clearly are, for example, condition (1-a), according to which the representation must present its object as having some ‘matter.’ No proper object lacks all ‘matter’ and is merely ‘formal.’ But since Kant’s interest with respect to objectivity is mainly focused on its formal conditions, we will also not pursue this question any further. Also, again, proponents of the view that appearances are to be regarded as ontologically specified by possible experience should add the clarification to the stated conditions that a representation R counts as a piece of possible outer experience only if (1) (a) R presents its object O as having some ‘matter,’ and (b) this ‘matter’ does not depend on special features of individual cognizers or special empirical circumstances, and (2) there could be a perception P, and (a) R presents O as the proximate cause of the sensations contained in P, or (b) R presents O as causally connected to O*, which, in turn, it presents as the proximate cause of the sensations contained in P.

, ,  

75

call this requirement ‘global material condition of experience’ and capture it by saying that a collection of representations of which each conforms to the material conditions of experience amounts to outer experience at time t only if, for all perceptions P available at t, one of the representations presents its object as the proximate cause of the sensations contained in P. Since it is through sensations that all ‘matter’ of objects is ultimately given to us, and since sensations are the result of affections of sensibility by things in themselves, the material conditions of experience, in effect, also characterize how experience depends on things in themselves, and, thus, how appearances, the intentional objects of experience, are grounded in things in themselves. The formal conditions of experience distinguish experience from most typical kinds of dreams, illusions, and hallucinations. Typical hallucinations, illusions, and dreams are quasi-perceptual, that is, from the ‘inside’ they look and feel just like perceptions, and thus, arguably, involve the same basic cognitive machinery as perceptions. This, in turn, means that they are as incapable as perceptions of representing causal relations or causal powers, which puts them in violation of some of the formal conditions of experience. But it seems not too difficult to come up with the description of an imaginary universe populated by all sorts of fairytale creatures that does conform to the formal conditions and even the global formal condition of experience.¹³⁸ So, the question arises whether there could be any representations other than experience that conform to both Kant’s formal conditions of experience and the material conditions just stated. The answer appears to be no. After setting aside all fictions that do not conform to the formal conditions, the vast majority of remaining fictions will run afoul of the second material condition and the global material condition by failing to properly connect their fictional objects to the available perceptions. But even if we grant that, unlikely as it may seem, an especially inventive fiction author could manage to establish such a connection by ascribing some complicated, baroque causal laws to his imaginary universe, the fiction in question would violate the b-part of the first material condition in that the ‘matter’ of his fictional objects—which includes the particular nature of their causal powers, which are characterized by the baroque causal laws—would depend, to a large extent, on his special ingenuity and his idiosyncratic creative ideas. That is, the fictional objects of his story, although grounded in things in themselves in some sense, would not be grounded in them in the direct kind of way required by the material conditions of experience. Thus, the conjunction of all formal and material conditions of experience, including the global ones, appears to be necessary and sufficient for a representation or a collection of representations to count as (outer) experience.¹³⁹ This also puts us into the position to state necessary and sufficient conditions for the actual existence of an intentional object of empirical thinking at the empirical level of reality, conditions that incorporate the conditions for the presentation of intentional objects in empirical thinking as actually existing, as identified above, but also go beyond

¹³⁸ As indicated in note 134, Kant sometimes sounds as if he thinks that his formal conditions of experience are sufficient to distinguish experience from all other kinds of representations. In Kant’s defense, like most other philosophers, then and now, in the relevant passages he is mainly focused on distinguishing experience from typical illusions, hallucinations, and dreams, a task that is indeed already successfully completed with the specification of the formal conditions of experience. ¹³⁹ The epistemic problem of how we can tell whether what we take to be experience is, indeed, experience, is a further question, which we will address in section 4.1.2.

76

      

them. An intentional object of a piece of empirical thinking ET actually exists at the empirical level of reality if, and only if, ET is a member of a collection of representations such that (i) each member of the collection conforms to the formal and material conditions of experience and (ii) the collection conforms to the global formal and global material conditions of experience. In other words, an intentional object of a piece of empirical thinking ET actually exists at the empirical level of reality if, and only if, ET is a member of a collection of representations that amounts to experience—which is exactly what we want.¹⁴⁰ The intentional objects of experience are the only intentional objects that both conform to Kant’s formal conditions of proper objecthood and are grounded in things in themselves in a properly direct way. This grounding ‘anchors’ the empirical level of reality in the fundamental transcendental level and lends appearances the required ontological weight, so to speak, to count as existents, even from the point of view of fundamental ontology. That is, thanks to the dependence of sensations on things in themselves, the existence conditions just formulated are conditions for the existence of the intentional objects of empirical thinking at the empirical level of reality, as well as for their existence from the point of view of fundamental ontology. On the account of the construction of experience just sketched, sensations turn out to have two different kinds of causes, namely, appearances that produce sensations by, among other things, affecting our sense organs, on the one hand, and things in themselves that produce sensations by affecting sensibility, on the other hand. This distinction between two different kinds of affection is a special case of a general distinction between two different kinds of causality—and, with it, two different kinds of causation and causes—in Kant’s ontological scheme. I will refer to them as ‘empirical causality’ and ‘transcendental causality,’ respectively. Empirical causality is the kind of causality that takes place at the empirical level of reality; it is exercised by appearances on appearances and connects all appearances in a deterministic system that is governed by the laws of nature. Appearances that empirically affect appearances exercise empirical causality and are empirical causes. Transcendental causality is the kind of causality that takes place at the transcendental level of reality; it is exercised by things in themselves on things in themselves. Things in themselves that transcendentally affect things in themselves exercise transcendental causality and are transcendental causes. Causality through freedom and the affection of sensibility by things in themselves that brings about sensations are instances of transcendental causation.¹⁴¹ Note that transcendental causes can and often do have empirical effects, that is, effects at the empirical level of reality, for example, actions in the case of causality through freedom and sensations in the case of transcendental affections of sensibility. But the things that are acted upon by these transcendental causes must be things in themselves

¹⁴⁰ Proponents of the view that appearances are ontologically specified by possible experience will have to slightly modify these claims in appropriate ways, as before. Since this is not my view, I will not bother with these modifications here. ¹⁴¹ See KpV, 5:47: “The moral law is in fact a law of causality through freedom and thus of the possibility of supersensible nature, just as the metaphysical law of the events in the world of sense is a law of causality of sensible nature . . . ” See B472/A44: “Thesis. The causality according to laws of nature is not the only one from which the appearances of the world can be altogether derived. To explain them it is still necessary to assume a causality through freedom.” (With respect to the third and fourth antinomy, Kant takes both thesis and anti-thesis to be true. See note 171, below, and note 2, chapter 3.) Also see B566/A538. The worry that, according to Kant’s theory of cognition, the application of the category of causality is restricted to appearances and objects of possible experience will be addressed in section 5.10.2.

, ,  

77

and cannot be appearances.¹⁴² Case in point: sensibility is not an appearance but a mindindependent cognitive faculty (in the sense defined in section 2.3.)¹⁴³ The result that sensations have two kinds of causes, transcendental and empirical ones, might raise the worry that this is one kind too many and amounts to a problematic case of causal overdetermination. One might also worry that if appearances indeed caused sensations, they would be among their very own existence conditions, which, however, is impossible. For, as we just learned, sensations and perceptions are the material conditions for the possibility of experience, and appearances essentially are intentional objects of experience. These worries are worth raising, but they can be answered. First, the apparent causal overdetermination that results from taking sensations to have both transcendental and empirical causes is harmless. One could hold that (a) there is nothing problematic about causal overdeterminations in general; or (b) there is nothing problematic about this particular causal overdetermination, precisely because the involved two kinds of causes belong to different levels of reality; or (c) the situation in question is not really a case of causal overdetermination; or (d) the seeming causal overdetermination is standard fare in Kant’s philosophy, and thus does not provide any grounds for rejecting the proposed account of the construction of experience. I endorse responses (c) and (d). Regarding (c), arguably, given the make-up of our cognitive faculties, it is impossible for us to have sensations without there being both a transcendental and an empirical cause for them, and so we are not dealing with a case of causal overdetermination but joint causation. Regarding (d), the causal overdetermination, or, rather, joint causation, with respect to sensations is of a piece with the joint causation that obtains with respect to all appearances and events at the empirical level. Their empirical causes are other appearances and events at the empirical level of reality, which are connected to them through empirical causal laws in accordance with the analogies of experience. Their transcendental causes are the things in themselves at the transcendental level of reality that ground them. A particularly striking case in point is the case of human actions. According to Kant, every human action has an empirical cause that operates ‘horizontally’ according to the laws of deterministic causality and a transcendental cause that operates ‘vertically’ according to the laws of freedom.¹⁴⁴ Second, the worry that appearances would be among their very own existence conditions can be countered by reminding ourselves that appearances empirically cause sensations, that is, that they cause sensations at the empirical level of reality. But at the empirical level, sensations are not conditions for the existence of appearances. From the point of view of the empirical level, the existence of appearances depends on other appearances. For example, the existence of my desk depends on the material out of which it is composed and the carpenter who put it together. Appearances would be among their own existence conditions from the point of view of fundamental ontology only if they were transcendental causes of sensations, which, however, they are not. That is, that appearances empirically

¹⁴² Similarly, the things that are transcendental causes must be things in themselves and cannot be appearances. We will think about the reasons why appearances could not be transcendental causes in section 5.10.2. ¹⁴³ Sensibility is mind-dependent in another sense, of course, namely, in the sense that it is a cognitive faculty of a mind. ¹⁴⁴ See Prol, 4:345: “Now I can say without contradiction: all actions of rational beings, insofar they are appearances (are encountered in some experience), stand under natural necessity; but the very same actions, but merely with respect to the rational subject and its capacity to act according to pure reason, are free.” See KpV, 5:97–98, 114–115; B566–586/A538–558.

78

      

cause sensations does not make them part of either their empirical or their transcendental existence conditions, and so there is no potential vicious circle to worry about.¹⁴⁵

2.5.4 Perceiving Appearances Before moving on to examine Kant’s conception of experience in more detail, there is one more loose thread with respect to his account of perception that needs to be tied up, namely, the question of what it means for an appearance to be perceived. Corresponding to the terminology of ‘directly representing’ introduced in section 2.2, we will say that an object is directly perceived if, and only if, the object is directly represented in a perception, that is, the object is an intentional object of a perception, and thus is part of the perception’s presentational content. Since perceptions are incapable of directly representing appearances, appearances are never directly perceived. So, the perception of appearances is essentially indirect.¹⁴⁶ Against the background of the above sketch of the transition from perception to experience in the construction of experience, which crucially depends on presenting appearances as the proximate causes of the sensations contained in our perceptions, it is tempting to conceive of the indirect intentional relation of perceptions to appearances on a causal model. For example, one might define that an appearance A is causally-mediately perceived if, and only if, there is a perception P, and experience presents A as the proximate cause of the sensations contained in P, or, equivalently, if, and only if, there is a perception P, and A is the proximate empirical cause of the sensations contained in P.¹⁴⁷ This proposal is attractive, both because Kant quite clearly believes that the empirical affections of our sense organs by appearances bring about sensations and because causal theories of reference are quite popular, especially in the contemporary context. I do not think, however, that Kant takes causal relations between perceptions and appearances to play a role in establishing intentional relations between them.

¹⁴⁵ Also note that although the claim that sensations have both transcendental and empirical causes can be described as a kind of doctrine of double affection in that the production of sensations is understood to involve the affection of sensibility by both things in themselves and appearances, the claim is to be distinguished from the theory that is known, or at least traditionally used to be known, under the name ‘doctrine of double affection.’ This doctrine is the one that Hans Vaihinger and Erich Adickes ascribe to Kant. See Vaihinger, 1884, 140–64; Adickes 1929; also see Vaihinger 1892, 52–5. The differences between my reading and the traditional doctrine of double affection are subtle but important. According to this doctrine, transcendental affections of our transcendental self by things in themselves produce unconscious transcendental representations, which are “appearances in themselves” (Adickes 1929, 36). Empirical affections consist in affections of our empirical self by these appearances in themselves, by means of which conscious sensations and representations are produced in us, which thus represent appearances of appearances. On my reading, transcendental affections of sensibility by things in themselves produce unconscious representations that provide the ultimate basis for the construction of perceptions and, eventually, experience. But this construction (as described in the main text in the present and foregoing sections) does not involve any affection of our empirical self by unconscious representations or by the objects represented in these representations. Rather, it involves various cognitive operations, including various acts of ordering and synthesizing and, eventually, the presentation of objects as affecting our sense organs and thereby producing sensations, which are states of our empirical self. ¹⁴⁶ But does Kant not say that intuitions relate immediately to objects, and does this not suggest that he holds that we directly perceive appearances after all? Yes to the former but no to the latter, on which more at the end of this section. ¹⁴⁷ The formulation that A is the proximate empirical cause of the sensations contained in P is a more efficient way of saying that, at the empirical level, or with respect to the empirical level, or from the point of view of the empirical level, A is the proximate cause of the sensations contained in P.

, ,  

79

Without going into too much detail, as I see it, Kant subscribes to a kind of visual resemblance theory about the intentional relation of intuitions to objects, as one might put it. For ease of communication, let us use the label ‘intuitive property’ or ‘intuitive determination’ for any property or determination of an object whose representation involves, apart from the sensibly given material, only the forms of sensibility and concepts or schemata that are grounded in the categories of quantity and quality. Roughly put, intuitive properties/determinations are properties/determinations that can be represented in intuition. For example, being square is an intuitive property but being a cause is not. It is essential to outer intuitions that they present their content to the cognizer entertaining them in a particular presentational format or mode, as one might say: the presentational content of an outer intuition is a pictorial content or ‘look.’ The German word for intuition, ‘Anschauung,’ reflects this feature by highlighting that we are dealing with a mode of presentation that necessitates a certain mode of uptake, namely, the mode of looking at, or anschauen.¹⁴⁸ Intuitive properties are properties that have a certain look. For example, being square looks a certain way, namely, like this: □; but there is no way that being a cause looks like. Now, on my reading, the intentional relation of intuitions to objects is determined by the resemblance, in terms of looks, between the presentational content disclosed in the intuitions and the intuitive determinations of the objects. An intuition intentionally relates to an object in virtue of the fact that the object’s look is completely captured in the presentational content disclosed in the intuition. Since the intentional object of an intuition is part of the intuition’s presentational content, there is a necessary resemblance between the intentional object and the presentational content, and so the intentional relation of intuitions to their intentional objects is easily accounted for. But what about the reference of intuitions to objects that are not their intentional objects? On the sketched visual resemblance account, it might seem that my current perception that represents a red, slightly foreshortened rose could not possibly intentionally relate to any public rose in the empirical world simply because public roses are neither red nor foreshortened in any way but non-colored and non-distortedly rose-shaped.¹⁴⁹ At this seeming impasse, the Leibnizian elements in Kant’s account of perception that we discussed in section 2.4 offer a way out by providing the basis for an account of the indirect perception of appearances within the framework of a visual resemblance theory of the intentional relations of intuitions. I will call the relevant kind of indirect perception ‘confused perception.’ The basic idea may be summarized by saying that while appearances are not directly perceived, they, or, more precisely, something intimately related to them, is directly represented in the unconscious basic components of perceptions. To make this idea more precise, it is useful to start by observing that, even though, in contrast to perceptions, the ‘little’ empirical intuitions, on which perceptions supervene through con-fusion, have public intentional objects,¹⁵⁰ they cannot directly represent appearances ¹⁴⁸ That concepts by themselves do not present us with any looks is part of the reason why they relate only mediately to objects, while intuitions do so immediately. In order for concepts to intentionally relate to objects they must be schematized because their schemata first provide a translation of their content into a pictorial format, and thus function as intermediaries between them and objects. See B179–181/A140–142. ¹⁴⁹ Colors are secondary qualities, which are not to be attributed to empirical objects but to the (empirical) appearances of empirical objects. We will consider Kant’s account of secondary qualities in detail in sections 5.3 and 5.4. For the claim that empirical objects are non-colored, in particular, see note 13, chapter 5. ¹⁵⁰ As you will recall, the ‘little’ empirical intuitions of observers that are affected by the same things in themselves are all qualitatively identical.

80

      

either. They are as incapable as perceptions of representing determinations that depend on the relational categories. But the ‘little’ empirical intuitions can represent the intuitive determinations of appearances, which means that their presentational content can completely capture the look of appearances. In other words, unconscious ‘little’ empirical intuitions are capable of directly representing objects that completely resemble appearances in terms of their looks. For ease of reference, I will call the intentional objects of the unconscious components of our perceptual states ‘unconscious appearances’ and say that intentional object B is the intuitive counterpart of intentional object A if, and only if, A and B both have intuitive properties, and B has all of the intuitive properties that A has. Using this terminology, on my Leibnizian reading of Kant’s account of perception, at any given time, for every appearance A there is exactly one unconscious appearance B that is A’s intuitive counterpart.¹⁵¹ And since Kant appears to be committed to the view that all appearances are distinguished from one another by their intuitive determinations, in particular, their spatial relations to one another,¹⁵² it also follows that, at any given time, for each unconscious appearance there is at most one appearance whose intuitive counterpart it is.¹⁵³ So, even though unconscious ‘little’ empirical intuitions cannot directly represent appearances, they can and do directly represent their unique intuitive counterparts, counterparts that resemble the appearances, in terms of looks, with respect to all of the appearances’ intuitive determinations. This is all we need to define, in compliance with Kant’s visual resemblance theory of the intentional relations of intuitions, what it means for an appearance to be confusedly perceived. An appearance A is confusedly perceived by means of perception P if, and only if, the unconscious appearance of the unconscious ‘little’ empirical intuitions on which P supervenes is A’s intuitive counterpart. While perceptions do not directly represent appearances, they refer to them in that the presentational contents of the perceptions’ unconscious basic components resemble, in terms of looks, all of the intuitive determinations of appearances. For example, my present perception that has a red, slightly foreshortened rose as its intentional object is a confused perception of and thus refers to a public rose in the empirical world because the public rose’s unique intuitive counterpart, which looks like the public rose, is directly represented in the unconscious ‘little’ empirical intuitions on which my present perception supervenes through con-fusion. In the following, when I talk about the perception of appearances or the reference of perceptions to appearances, I should be understood as having this kind of confused perception and reference through confused perception in mind, respectively. Before leaving the topic of the intentional relation of perceptions to appearances, it is also worth highlighting that, even though this relation is indirect in the sense that appearances are not the intentional objects of perceptions, it is still direct or immediate ¹⁵¹ Full disclosure: this claim rests on the assumption that the critical Kant still agrees with Leibniz that, at every moment, we confusedly perceive the entire empirical world. Also note that, if appearances are ontologically specified by possible experience, then A and B have all of their intuitive properties in common; if appearances are ontologically specified by actual experience, then A and B differ with respect to their intuitive properties only in that B is intuitively more determinate than A by having not only all intuitive properties that actual experience presents A to have but also all intuitive properties that possible experience presents A to have, holding fixed the transcendental level of reality and the nature of our cognitive faculties. Conversely (on either account of appearances), A has some properties that depend on the relational categories, with respect to which the unconscious ‘little’ empirical intuitions leave it undetermined whether B has them or not. ¹⁵² See B319 320/A263 264; B327 328/A271 272; FM, 20:282. See already PND, 1:409. ¹⁵³ I say ‘at most one’ because there may be none, namely, if appearances are ontologically specified by actual experience, and actual experience does not represent the relevant appearance.

   

81

in another sense. This is important because one of the features of intuitions that Kant tends to emphasize when he contrasts them with concepts is that they relate immediately to objects, in contrast to concepts, which relate to objects only mediately.¹⁵⁴ Since it is plausible to assume that in this context by ‘objects’ Kant means appearances, or at least a class of objects that includes appearances, any defensible interpretation of his account of perception must be able to offer a plausible explication of the sense in which perceptions (which are empirical intuitions) relate immediately to appearances. On my reading, by saying that intuitions relate immediately to objects, Kant means that they intentionally relate to objects without the mediation of any other representation.¹⁵⁵ Perceptions, as interpreted here, are immediately related to appearances in this sense. They intentionally relate to appearances without the mediation of any other representations because this relation is established by the presentational content of the ‘little’ empirical intuitions that are comprised in them, namely, in their unconscious basic components.

2.6 The Oneness of Experience Against the background of the general account of experience sketched in section 2.5, we can now turn to addressing a few complications and further questions that will help to fill in some more details. Although in our normal conversations it is natural to talk about experiences, plural, and although there is room for this kind of talk in Kant’s theory of experience—namely, as a way of speaking somewhat loosely of perceptions or of specific judgments of experience made by different subjects or on different occasions—when it comes to the kind of experience that ontologically specifies the ontological furniture of the empirical world, strictly speaking, there is only one experience. There is only one experience, in which all perceptions are represented as in thoroughgoing and lawful connection: just as there is only one space and time in which all forms of appearance and all relation of being and non-being take place. If one talks about different experiences, those are just as many perceptions insofar as they belong to one and the same general experience. The thoroughgoing and synthetic unity of perceptions namely precisely constitutes the form of experience, and it is nothing other than the synthetic unity of appearances according to concepts. (A110)¹⁵⁶

¹⁵⁴ See B376–377/A320: “ . . . an objective perception is cognition (cognitio). This is either intuition or concept (intuitus vel conceptus). The former relates immediately to the object and is singular; the latter relates mediately to the object, by means of a mark that can be common to several things.” ¹⁵⁵ This reading is confirmed by the opening sentence of the Transcendental Aesthetic: “In whatever way and through whatever means a cognition may be related to objects, the one through which it relates to them immediately and toward which all thinking aims as a means, is intuition” (B33/A19). This sentence makes clear that the objective for the sake of which intuitions are employed as a means is for concepts to be intentionally related to objects. Concepts intentionally relate to objects only with the help of other representations, namely, other concepts, their schemata, and ultimately intuitions, while intuitions do not require any such assistance. Also see B93/A68: “Since no representation goes immediately to the object apart from merely intuition, a concept is never immediately related to an object but to some other representation of it (be it intuition or itself concept).” Also see B41; B376–377/A320, FM, 20:325. We will return to Kant’s conception of intuitions in sections 4.2.2.2 and 4.2.4. Also see Jauernig 2019 and Jauernig (in preparation). ¹⁵⁶ See B610/A582: “ . . . since that in which the real of all appearances is given is the unified all-encompassing experience . . . ” The oneness of experience is an especially prominent theme in the Opus Postumum, where Kant also stresses that talk about experiences is to be understood as loose talk about (conceptually determined)

82

      

As a first pass, this one experience, I submit, is to be understood as a science-assisted, empirically and transcendentally grounded account of the empirical world that is shared, or at least sharable, by all properly functioning members of the community of human minds. This account includes an inventory of all appearances and their determinations that have been perceived by human subjects, as well as of various inferred appearances that are represented as causally connected to perceived appearances. It also includes statements of various general laws of nature that govern the behavior and interactions of all appearances as well as of various special empirical laws that govern the behavior and interaction of all appearances of a certain kind. The account is science-assisted in that the precise determination of the empirical laws that are included in it is the business of natural science. It is empirically grounded in that it is based on observations and experiments. It is transcendentally grounded in that it incorporates judgments of experience and in that the general laws of nature and the special empirical laws that it puts forward are (in some sense) derivable, with the help of certain auxiliary, partly empirical assumptions, from the transcendental principles of the understanding that Kant demonstrates in the Critique, in particular, the analogies of experience.¹⁵⁷ What makes the account sharable is that all of its elements are in principle communicable through language. Even though no single human mind ever entertains the account in its entirety―no single human mind has a representation of all appearances that are or have been perceived by human subjects, for example, and no single human mind knows absolutely all of science―each one of the elements of the account has been entertained by at least one human mind at some point and can in principle be conveyed to and entertained by every properly functioning human mind. This also completes the explanation, begun in section 2.5, of why and in what sense the intentional objects of experience are public. Since experience is shared, or at least sharable, among all human minds, it is itself public, which means that its intentional objects are public as well. Given that science progresses over time, and that we continue to perceive new appearances as time goes on, an obvious, tempting next move would be to say more precisely that the one experience that ultimately ontologically specifies appearances is the final sharable, science-assisted, empirically and transcendentally grounded account of the empirical world that stands at the end of all scientific and observational progress or maybe even at an imaginary ideal end of all scientific and observational progress. The thought would be that this final account amounts to a completely determinate, comprehensive, thoroughly unified, systematic representation of the empirical world. Such an account would include a description of all appearances that are possible in principle given the nature of our cognitive faculties and the prevailing ultimate material conditions for the possibility of perceptions or pieces of experience. See OP, 21:247: “Considered objectively, there is only One experience, and if subjectively one speaks about experiences, those are nothing but parts and lawfully connected aggregates of a synthetically-general experience.” See OP, 21:564: “Experience (objectively is only One) is an absolute unity, and if one talks about experiences those are merely perceptions that presuppose this unity of form.” Also see OP, 21:571, 576, 580, 582, 595, 597, 601; OP, 22:552. ¹⁵⁷ I say that these laws are ‘in some sense’ derivable from the transcendental principles because it is far from clear how exactly Kant conceives of the relation between them. There are only three general laws of nature, namely, Kant’s three laws of mechanics, and, arguably, one special empirical law, namely, the universal law of gravitation, for which Kant himself attempts such a derivation, in the Metaphysical Foundations of Natural Science. See B165: “Special laws, since they concern empirically determined appearances, cannot be completely derived from them [the transcendental principles of the understanding] although they all together stand under them.”

   

83

experience, specify precise determinations for all determinables of the included appearances,¹⁵⁸ and explain all changes in these determinations over time by subsuming them under general and special empirical laws of nature that are systematically related and govern the behavior and interactions of all appearances and thereby combine them into a unified system. For ease of reference, I will call this completely determinate, comprehensive, thoroughly unified, systematic account of the empirical world ‘Experience’ with a capital ‘E.’ So, the proposal would be that appearances exist and have all of their determinations and other ontological ingredients in virtue of being represented in Experience. Alas, as tempting as it is, I do not believe that this reading adequately captures how Kant thinks of the one experience that ontologically specifies the ontological furniture of the empirical world. For, as it turns out, Experience is impossible for us to attain. But the sketched reading provides a useful foil for our discussion insofar as seeing that and how Experience eludes us on Kant’s view will give us valuable clues about how the one experience that ontologically specifies appearances should be understood instead.

2.7 The Impossibility of Experience 2.7.1 Possibly Non-cooperating Things in Themselves There are several reasons to think that we cannot ever attain Experience. The first two of these reasons do not decisively rule out the possibility that we could develop such a completely determinate, comprehensive, thoroughly unified, systematic account of the empirical world but they raise serious doubts about it. Both of them have to do with a potential lack of cooperation, so to speak, on the part of things in themselves, which, it will be recalled, are in charge of contributing the ‘matter’ to the constitution of appearances, as well as underwriting their existence, by way of affecting sensibility and thereby bringing about sensations. It is not entirely clear whether Kant himself ever reached sufficient clarity with respect to the first of the reasons to be discussed that speak against us ever attaining Experience—which, incidentally, became a focus of attention in neo-Kantian circles in the early twentieth century—but he did explicitly worry about the second. The first reason to be skeptical about the possibility of us having Experience is that it might well be that not all of the material that is or will be given to us through affections of sensibility by things in themselves can be fitted into our intuitive and conceptual scheme, as described by Kant, to yield one comprehensive, unified account. For example, we might come up against certain observations of the movements of celestial bodies, for example, the precession of the perihelion of Mercury, that cannot be explained by appeal to Kant’s ¹⁵⁸ Each particular determination of an object belongs to a ‘determinable’ of the object. For example, the determination of my cat of currently weighing 8 kg belongs to the determinable of having some weight or other at every instant; his determination of currently being located on the highest platform of his favorite jungle gym belongs to the determinable of being located somewhere or other at every instant; and his determination of currently being asleep belongs to the determinable of being awake or asleep at every instant. Ontologically distinct kinds of objects differ with respect to their determinables. All appearances have the same determinables, but they have different determinables compared to things in themselves. For example, things in themselves are not in space and time, and so being located somewhere or other at every instant is not one of their determinables. ‘Determining’ an object consists in specifying a particular determination for one or more of its determinables; and a completely determined object is an object for all of whose determinables a particular determination has been specified. See B599–611/A571.

84

      

version of the laws of Newtonian science, even though Kant takes them to be transcendentally grounded laws that govern all mechanical and gravitational interactions and, thus, should be able to explain all planetary motions. Of course, this is exactly what transpired on the eve of the nineteenth century and what led to the development of scientific theories that do not conform to Kant’s mechanical principles, most prominently, Einstein’s theories of relativity—a development, which, in turn, prompted the discussion among neoKantians referenced above.¹⁵⁹ The second reason for skepticism is that, even if we grant that the material that is or will be given to us through sensibility could in principle be worked into a unified account of the empirical world by relying on laws that are grounded in the forms of our cognitive faculties, although maybe not the particular Newtonian laws that Kant himself identified, there remains the problem that due to our limited cognitive capabilities we may still not be able to do it. It might be that the variety of genera and species and of qualitative properties and special empirical laws is so great that the envisioned unified, systematic account is impossible for us, even though it is possible in principle.¹⁶⁰ This theme surfaces in Kant’s writings most explicitly in the Critique of Judgment and associated texts. For it can be well imagined that, despite all homogeneity of the objects of nature according to general laws without which the form of a cognition of experience would not take place at all, the specific difference of the empirical laws of nature together with their effects could be so great that it would be impossible for our understanding to discover in it a graspable order, to sort its products into genera and species, in order to use the principles of explanation and of intelligibility of the one also for the explanation and understanding of the other and to make a connected experience out of a material that is so confused for us (or really only infinitely varied, not appropriate for our cognitive capability). (KU, 5:185)¹⁶¹

So, the problem is that, even if things in themselves supply the right kind of material, that is, material that is amenable to be ‘harnessed’ in our conceptual and intuitive scheme, there could simply be too much variation in the material for us to be able to put the harness on, so to speak, in a way that covers all of the material in all of its richness.

2.7.2 Experience and the Mathematical Antinomies The third reason for thinking that we cannot have Experience not only casts doubt on our ability to come up with a completely determinate, comprehensive, unified, systematic account of the empirical world but effectively shows such an account to be impossible for us. This reason does not depend on any possible non-cooperation on the part of things in themselves in the construction of experience but reflects certain irremediable ¹⁵⁹ See Reichenbach 1920 and Cassirer 1921. ¹⁶⁰ That is, even though such an account is possible for a being with cognitive faculties that have the same forms as ours but with more intellectual fire power and more processing capacity. ¹⁶¹ Also see EEKU, esp. “IV. Of Experience as a System for the Power of Judgment”, 20:208–211; and KU, Introduction, “V. The Principle of Formal Purposiveness is a Transcendental Principle of the Power of Judgment”, 5:181–185.

   

85

peculiarities of our cognitive faculties and their influence on the construction process. Even if we were not confronted with any outlier phenomena that do not fit into our cognitive framework, and even if there was sufficiently little variety in nature so that we could come up with a comprehensive genera and species classification scheme and a systematic account of special empirical laws such that none of the behavior or interactions of appearances remained unexplained, we still could not ever attain Experience because of the way in which our cognitive faculties operate. It will take some time to explain this reason in detail, but here is a brief précis. Experience (with a capital E) is supposed to include not only a systematic collection of general and special empirical laws of nature but also a comprehensive inventory and completely determinate description of all appearances that are possible in principle, given the nature of our cognitive faculties and the prevailing ultimate material conditions for the possibility of experience. The construction of Experience thus includes the discovery of appearances at increasingly larger spatial and temporal distances from us, in order to fill the inventory, as well as the decomposition of appearances into increasingly smaller parts, again in order to fill the inventory but also to settle certain determinables of appearances that concern their inner constitution. And, as the case may be, accounting for some of these newly discovered appearances might well require the formulation of additional empirical laws as well. Now, it turns out that, due to the nature of our cognitive faculties, this kind of extension of experience beyond the familiar middle-sized objects in fairly close spatial and temporal proximity to us is, of necessity, a continual work in progress. But if the construction of experience is necessarily a continual work in progress, it directly follows that Experience, understood as a final, completely determinate, comprehensive account of the empirical world, is impossible. In order to see exactly why and how the construction of experience necessarily is a continual work in progress, on Kant’s account, we need to take a closer look at the mathematical antinomies. Of all sections in the Critique, none is more telling about Kant’s conception of experience and the nature of his idealism about appearances than this one. There is controversy in the literature about how best to understand the general dialectical structure of the mathematical antinomies and Kant’s solution to them. There is no need for us to get caught up in this debate right now. For we are mainly interested in two specific claims that play a crucial role in Kant’s solution, namely, that the extension and duration of the empirical world—understood as the sum total of all empirical objects, or all objects in space and time—is indeterminate, and that the compositional structure of empirical objects is likewise indeterminate, in particular, that it is indeterminate how many parts they contain.¹⁶² Since Kant explicitly endorses these claims a number of ¹⁶² I have included the explicit specification that the first antinomy concerns the extension and duration of the empirical world understood as the sum total of all empirical objects to make clear that we are not dealing with the question whether space and time are finite or infinite. Kant himself does not explicitly highlight that the first antinomy is about the extension and duration of the empirical world thus understood but his presentation of the antinomy and the proofs of the thesis and antithesis make clear that this is what he has in mind. For example, in stating the first antinomy, he says that the world has/does not have a beginning in time and has/does not have boundaries in space; see B454/A426, B455/A427. Similarly, in the proof of the thesis he says that the assumption that the world has no beginning in time would mean that “an infinite series of successive states of the things in the world has elapsed” (B454/A426), and that the assumption that the world has no boundaries must be rejected because “an infinite aggregate of real things cannot be regarded as a given whole” (B457/A429). Similarly, in the proof of the antithesis he states that the assumption that the world has a beginning in time means that “a time must have preceded [the beginning of the world] in which there was no world” (B455/A427), and that the assumption that the world has boundaries in space would mean that “it [the world] is in an empty space that is not

86

      

times, there should be no debate over whether he is indeed committed to them. So, for our present purposes it is sufficient for me to sketch, without further ado, how I understand the general structure of Kant’s reasoning in the section on the mathematical antinomies before focusing on a more detailed consideration of the indicated central claims.¹⁶³ The mathematical antinomies consist of two pairs of a thesis and an antithesis: first, that the empirical world is finite in space and time and, thus, has a first beginning in time and ultimate boundaries in space, versus that it is infinite in space and time and, thus, has no first beginning in time and no boundaries in space; and, second, that every composite substance in the empirical world is composed of finitely many parts and, thus, consists of simple parts, versus that it is composed of infinitely many parts and, thus, does not consist of simple parts.¹⁶⁴ We will say more about the genesis of these antinomial propositions in section 6.3. For now, it is enough to note that, on Kant’s analysis, both the idea of an unconditioned first element of a series of conditions, which underlies the thesis of each antinomy, and the idea of an infinite series of conditions, which underlies the antithesis of each antinomy, are so-called transcendental ideas of reason, which reason is led to reify as a result of having transformed its logical maxim “to find the unconditioned for the conditioned cognition of the understanding whereby its [the understanding’s] unity is completed” into the principle that “if the conditioned is given, the entire series of conditions subordinated to one another, which is thus itself unconditioned, is also given” (B364/ A307–308).¹⁶⁵ Now, according to Kant, for both mathematical antinomies, we can prove of both the thesis and the antithesis that they are false. This leads to a problem if empirical objects and with them the empirical world are assumed to be things in themselves (understood in the general sense), that is, things that are mind-independent. For in this case, Kant argues, we are committed to saying that the two claims in each pair are

bounded” (B455/A427). All of this shows clearly that by ‘world’ Kant means the sum total of all empirical objects. What are Kant’s views about the extension of space and time? Kant clearly thinks that pure space and pure time, the forms of sensibility, are infinitely extended. What he thinks about the extension of empirical space and time, that is, the space and time that exist at the empirical level of reality, is less clear. We will return to this question in section 3.4.3. ¹⁶³ For more discussion, see Jauernig (forthcoming 2021b) and section 6.6. ¹⁶⁴ In his own formulations of the mathematical antinomies—“the world has a beginning in time and is also confined into boundaries according to space” versus “the world has no beginning and no boundaries in space but is infinite with respect to both time and space,” and “every composite substance in the world consists of simple parts, and nothing exists anywhere except the simple or what is composed of it” versus “no composite thing in the world consists of simple parts, and there is nowhere anything simple in the world”—Kant is characteristically sloppy, I am afraid. The formulations are appropriate to express the contradictory views to which his transcendental realist opponent is committed (according to Kant’s assessment), but they are not appropriate given Kant’s own views on the matter. As we will see, a crucial element in his solution to the mathematical antinomies is that, on the assumption of transcendental idealism, the thesis and the antithesis both turn out to be false. But, given transcendental idealism, it is true that the world has no first beginning in time and no outer boundary in space, as Kant himself explicitly acknowledges at B548/A520, just as it is also true that no composite empirical substances consists of simple parts. What is false, on the assumption of transcendental idealism, is that the world is infinite in space and time, and that every composite empirical substance is composed of infinitely many parts. So, Kant should have used the more careful formulations of the first and second antinomy just proposed in the main text. Despite the sloppy formulations, Kant’s discussion leaves no doubt, though, that he well understands that the crucial claims at the heart of the antinomies concern the finite versus infinite extension and duration of the empirical world, in the case of the first, and the finite versus infinite number of parts of empirical substances, in the case of the second antinomy. ¹⁶⁵ Also see B525/A497: “The entire antinomy of pure reason rests on the dialectical argument: if the conditioned is given, then the entire series of all its conditions is also given. Now, objects of the senses are given to us as conditioned, therefore . . . ”

   

87

genuinely contradictory. And if these claims are contradictory, the proof of the falsity of one amounts to a proof of the truth of the other. This means that, if we assume that empirical objects are things in themselves, we can prove all of the claims in question— which is precisely what Kant does in the early parts of the antinomy chapter—leaving us with the disaster of two provable contradictions. Kant’s solution is to reject the assumption that empirical objects are things in themselves and assert instead that they are appearances. This allows him to deny that the two claims in each pair are contradictory, which, in turn, permits him to regard all of these claims as provably false without thereby committing himself to the existence of reductio proofs for all of them. More specifically, in each of the two mathematical antinomies, the thesis and antithesis rely on a common presupposition. These presuppositions are that the empirical world has a determinate extension and duration in the case of the first antinomy, and that the compositional structure of a complex empirical substance is determinate in the case of the second antinomy. On Kant’s analysis, whatever has a determinate extension and duration must be either finite or infinite in space and time; and whatever has a determinate compositional structure must either have finitely many parts or infinitely many parts. So, if we were committed to the indicated presuppositions, we also would be committed to acknowledging that the empirical world must be either finite or infinite in space and time, and that every composite substance in the empirical world must be composed of either finitely or infinitely many parts. In other words, we would be committed to saying that the thesis and antithesis of both antinomies are, indeed, contradictorily opposed to each other, leaving us no way to escape from the debacle of a provable contradiction. On Kant’s view, if we regard empirical objects and with them the empirical world as things in themselves, we are in fact committed to the problematic presuppositions. But if we regard empirical objects and with them the empirical world as appearances, he argues, we can reject the presuppositions, which, in turn, allows us to reject both the thesis and the antithesis of each mathematical antinomy and emerge as happy, contradiction-free transcendental idealists.¹⁶⁶ So much for the sketch of the general structure of Kant’s reasoning concerning the mathematical antinomies. The crux of Kant’s solution are the claims that the extension and duration of the empirical world, as well as the compositional structure of composite empirical substances, are all indeterminate, provided we assume that empirical objects are appearances and not things in themselves. (For brevity’s sake, from now on in place of ‘extension and duration’ I will use the shorter ‘size.’) Figuring out Kant’s basis for these claims is an important key to understanding his conception of experience and the nature of his idealism about appearances.¹⁶⁷ To begin with, note that if appearances were intentional objects of Experience, as characterized in section 2.6, the indicated indeterminacies would not obtain. The size of the empirical world and the cardinality of the multitude of parts of composite empirical substances would simply be whatever our final, completely determinate, comprehensive, unified, systematic account of the world says they are. So, by explaining why the size of the empirical world and the compositional structure of empirical substances are ¹⁶⁶ See B532–535/A504–507; B545–555/A518–527; B821/A793; MAN, 4:505 508; FM, 20:291. ¹⁶⁷ It is hard to see how these claims could be accommodated on two-aspect readings on which appearances are conceived of as things qua bearers of certain dispositions to appear to us in certain ways, or of properties of perceptually appearing to us in certain ways. On these readings, the spatial extension and temporal duration of the empirical world and the number of parts of empirical objects seem perfectly determinate.

88

      

indeterminate, we can hope to gain clues for how to improve our characterization of the one experience that ontologically specifies appearances. Kant offers the following explanation, which, because of its importance, is worth quoting in full: If one regards the two propositions ‘the world is infinite with respect to its size [Grösse]’ and ‘the world is finite with respect to its size’ as contradictorily opposed, one assumes that the world (the whole series of appearances) is a thing in itself. For it [the world as a thing in itself] remains, even if I remove the infinite or finite regress in the series of its appearances. But if I take away this assumption or this transcendental illusion and deny that it is a thing in itself, the contradictory opposition of both claims is transformed into a mere dialectical one; and since the world does not at all exist in itself (independently of the regressive series of my representations), it exists neither as an in itself infinite whole nor as an in itself finite whole. It is to be found only in the empirical regress of the series of appearances and not at all for itself. Therefore, if the latter [the series of appearances] is always conditioned, it is never given in its entirety, and the world is not an unconditioned whole and, accordingly, does not exist as such, neither with infinite nor with finite size. What has been said here concerning the first cosmological idea, namely, of the absolute totality of size in the appearance, applies also to all the others. The series of conditions is to be found only in the regressive synthesis itself, but not in itself in the appearance as a separate thing that is given before all regress. For this reason, I also will have to say: the multitude of parts in a given appearance is in itself neither finite nor infinite, since appearances are nothing that exists in itself, and the parts are first given through and in the regress of the decomposing synthesis, which regress is never given absolutely in its entirety, neither as finite nor as infinite. The same holds for the series of causes that are ordered above one another or the conditioned up to the unconditioned necessary existence, which can never be regarded in itself as either finite or infinite in its totality, since as series of subordinated representations it consists only in the dynamical regress but cannot at all exist in itself before this regress and as series that subsists for itself. (B532 533/A504 505)¹⁶⁸

The key claims in this passage are that, on Kant’s transcendental idealist view that allows us to see the thesis and antithesis as not contradictory, the empirical world does not exist in itself but only in the empirical regress of the series of appearances, that the parts of ¹⁶⁸ Also see MAN 4:506–507: “For what is actual only through being given in representation, of that also not more is given than as much as is found in the representation, i.e., as far as the progress of representations reaches. Thus, of appearances whose division goes to infinity one can only say that there are as many parts of the appearance as we may give, i.e., as far as we ever may divide. For the parts, as belonging to the existence of an appearance, exist only in thought, namely, in the division itself. Now it is true that the division goes to infinity but it is still never given as infinite. Thus, it does not follow from this that the divisible contains in itself an infinite multitude of parts in themselves and apart from our representation just because its division goes to infinity. For it is not the thing but only its representation, whose division, although it can be extended to infinity, and although there is a ground for it in the object (that is in itself unknown), can still never be completed and thus never be given in its entirety, and accordingly also does not demonstrate an actual infinite multitude in the object (which would be an explicit contradiction).” Also see B549/A521: “The regress in the series of appearances in the world, as a determination of the size of the world, goes in indefinitum, which says as much as that the sensible world has no absolute size, but the empirical regress (through which alone it can be given on the side of its conditions) has a rule, namely, to proceed from each member of the series, as a conditioned, always to a more remote one (be it through one’s own experience or the leading thread of history or the chain of effects and their causes) . . . ” Also see B551 B552/A523 524; Prol, 4:342.

   

89

appearances are only given in and through the regress of the empirical synthesis, and that neither regress is given in its entirety, be it finite or infinite. On my reading, by “empirical regress of the synthesis” Kant here means the kind of progressive construction of an account of the empirical world briefly described above, a construction that is fueled by the observation of new appearances and by the discovery of further special empirical laws that are grounded in the transcendental principles. Relying on transcendentally and empirically grounded laws of nature, be they newly discovered or long established, and aided by observation, exploration, experiments, and record-keeping, we extend our account of the empirical world by working our way from the middle-sized dry goods of our immediate spatiotemporal vicinity to their “conditions” and the conditions of these conditions, to use Kant’s terminology. These conditions include spatial conditions in that the boundaries of empirical objects are conditioned by the objects around them, temporal conditions in that later moments in the objects’ durations are conditioned by earlier moments, material conditions in that empirical objects are conditioned by their material parts, causal conditions in that changes in the states of empirical objects are conditioned by external causes, and modal conditions in that the existence of empirical objects, as contingent beings, is conditioned by other existent beings.¹⁶⁹ The construction of experience, accordingly, comprises regresses to appearances that are increasingly remote in space and increasingly earlier in time and in the chain of causes and effects as well as to increasingly smaller parts of matter. The claim that the empirical world, empirical objects, and their parts do not exist in themselves but only in these empirical regresses is a way of expressing the transcendental idealist core thesis that empirical objects and, with them, the empirical world are not things in themselves but appearances, which means that they are fully mind-dependent and their existence is not an existence in itself but an existence ‘in us’ or ‘in our representations.’ Since empirical objects are not things in themselves but appearances, the entire series of the indicated kinds of conditions for a given conditioned empirical object are not automatically given with the object. If the conditioned and its conditions of the relevant kind were things in themselves, reason’s previously mentioned ‘reification move’ from the rule to look for the conditions for any conditioned thing, to the assumption that for any conditioned thing the entire series of conditions also exists— would be legitimate, or at least defensible. As Kant puts it, “if the conditioned as well as its condition are things in themselves, then if the first is given not only the regress to the latter is set as a task, but it [the condition] is thereby really already given, and since this holds for all members of the series, the complete series of conditions, and thus the unconditioned, is given at the same time, or rather presupposed, through the fact that the conditioned is

¹⁶⁹ See B438–442/A411–415. It is worth noting that there is another regress to further and further conditions that is part of the construction of experience but is not explicitly thematized anywhere in the antinomy chapter. This further regress enters center stage in the Metaphysical Foundations; it is the regress to increasingly more encompassing material frames of reference in order to determine the actual or true (as opposed to merely apparent) rectilinear motions of appearances. Each rectilinear motion is conditioned by a material frame of reference that is assumed to be at rest but which, as material, is itself in motion with respect to other, more encompassing material frames of reference. In the limit, this regress would lead to the center of mass frame of the entire universe, a limit that, for reasons similar to the ones to be discussed here, cannot be reached by us. See MAN, 4:481–482, 490, 554–565. For a detailed discussion of this kind of regress and an interesting reading according to which Kant views Newton’s theory as specifying a constructive procedure for finding increasingly better approximations to such a privileged frame of reference, see Friedman 1992, esp. ch. 3, 136–65, and Friedman 2013, esp. ch. 4, §34, 474–508.

90

      

given, which was possible only through this series” (B526/A498).¹⁷⁰ But in the case of appearances, since they only exist in our representations, and since it does not follow that, if a conditioned thing is represented, its condition is thereby also represented, the regress from a conditioned thing to its spatial, temporal, material, causal, or modal conditions is not directly given with it but is only “demanded and set as a task” (527/A499). Accordingly, if an appearance is given, the entire series of its spatial, temporal, material, causal, or modal conditions are not thereby given at the same time and cannot be presupposed as existing; rather, each condition’s existence depends on us carrying out the empirical regress to its particular spot in the series of conditions. Finally, the formulation that none of the regresses are or could be “given to us in their entirety” expresses the crucial fact, to be discussed presently, that the regresses to all these conditions are not only not given at the same time with the conditioned but also such that we will never and could never complete them. This fact is crucial both for Kant’s solution of the mathematical antinomies and for our present question why we cannot have Experience. It is crucial for Kant’s solution because it explains why the size of the world and the compositional structure of empirical substances is indeterminate. Since the empirical world and empirical objects are appearances, the size of the former and compositional structure of the latter are determined by the corresponding empirical regresses. So, if these regresses are neither determinately finite nor determinately infinite, the size of the empirical world and the cardinality of the multitude of parts of empirical substances are also neither determinately finite nor determinately infinite.¹⁷¹ And the indicated fact is crucial for our current ¹⁷⁰ In section 6.6, we will take a closer look at the questions whether Kant is indeed committed to regarding this principle of reason as true of things in themselves and whether there may even be a sense in which he regards the principle as true of appearances after all. To anticipate, I lean toward the reading that Kant regards the principle as universally true if it is understood as a principle about transcendental conditions, that is, conditions from the point of view of fundamental ontology. The reason why the principle does not apply to the spatial, temporal, material, causal, or modal conditions of empirical objects understood as appearances is that these conditions are not transcendental conditions but merely empirical conditions. (Since, from the point of view of fundamental ontology, appearances essentially are intentional objects of experience, their transcendental conditions are the formal and material conditions of experience.) By contrast, if empirical objects were things in themselves, their spatial, temporal, material, causal, and modal conditions would be transcendental conditions and the principle would apply to them. ¹⁷¹ It is worth noting that (a) even though the so-called dynamical antinomies, that is, antinomies number three and four also involve empirical regresses, namely, from effects to their causes and from contingent beings to their existential conditions, and (b) the contradictory opposition of their thesis and antithesis, to which Kant’s transcendental realist opponent is committed, is also grounded in the opposition between reason’s demanded completion of a certain regress to ultimate conditions, on the one hand, and the infinite version of the same kind of regress that reason demands as given in its entirety, on the other hand, Kant’s solution differs in interesting ways from his solution of the mathematical antinomies. The conditions at issue in the mathematical antinomies are all sensible, just as the empirical objects and their properties that they condition. But, on the assumption of transcendental idealism, which allows us to introduce a distinction between sensible appearances and supersensible things in themselves that ground them, the conditions at issue in the dynamical antinomies may be understood as sensible or supersensible. That is, there may be not only sensible but also supersensible causes for sensible events, and there may be not only sensible but also supersensible existential conditions for sensible objects. So, while it is correct that the empirical regress in the series of causes is never given to us as finite and as terminating in uncaused causes, and the empirical regress in the series of existential conditions of contingent objects is never given to us as finite and as terminating in necessary beings, it still may be, assuming transcendental idealism, that there are uncaused supersensible causes of sensible events and a necessary supersensible being that ultimately conditions the existence of empirical objects. That is, given transcendental idealism, the theses of the dynamical antinomies—that the causality according to the laws of nature is not the only kind of causality that explains events in the empirical world, and there is causality through freedom, and that there is an absolutely necessary being in the world—are not only not contradictorily opposed to the corresponding antitheses, they may be true. Moreover, while it is also correct that the empirical regress in the series of causes and the empirical regress in the series of existential conditions of contingent objects are never given to us as infinite, it is still true that all events in the empirical world can be explained solely by appeal to the causal laws of nature, and that all empirical

   

91

question because it means that Experience, qua final comprehensive, completely determinate account of the empirical world that stands at the ideal end of all scientific and observational progress, is impossible for us. So, why does Kant think that none of the empirical regresses comprised in the construction of experience is given to us or could possibly be given to us in its entirety? Or, in other words, why does Kant think that the construction of experience cannot possibly be completed, at least not by us? It is especially difficult to see how he can rule out the possibility that a finite empirical regress could be given to us in its entirety, that is, that we could discover unconditioned conditions, as for instance simple parts as the ultimate constituents of material objects, if not by direct observations then through observationbased scientific theorizing. So, let us start with that option. On my reading, Kant’s most fundamental reason for thinking that we could not possibly discover any ultimate conditions in our empirical regresses is that the nature of our cognitive faculties and their representational capabilities make it impossible for us to represent appearances as unconditioned conditions. Given the way in which sensibility and the understanding represent appearances and their determinations, any given member of a series of conditions that we generate with our empirical regresses as part of the construction of experience automatically points to another condition without which it cannot be fully represented, that is, to a further member in the series, which, in turn, points us to its condition without which it cannot be fully represented, and so on ad indefinitum. Even though appearances are represented without a concomitant explicit representation of their various conditions, it is thus implicit in all of their representations that there are such conditions, which would have to be included in a completely determinate representation. If that is how we represent appearances, it directly follows that the representation of appearances as unconditioned conditions is impossible for us. More specifically, with respect to the regresses to the spatial, temporal, and material conditions of appearances that are relevant for the mathematical antinomies, that space and time are the forms of sensibility entails that the way in which we represent appearances reflects the way in which we represent space and time. According to Kant’s analysis, we represent space and time as infinitely divisible by representing any determinate finite space and any determinate finite duration as divisible into smaller parts, and space and time as infinite by representing any determinate finite space and any determinate finite duration as contained in a larger space and a larger duration, respectively.¹⁷² And since space and time are forms of sensibility, we accordingly represent appearances as infinitely divisible,¹⁷³ as existential conditions of empirical objects are themselves contingent. That is, given transcendental idealism, the antitheses of the dynamical antinomies—that there is no freedom in the empirical world and everything in the empirical world happens according to the laws of nature, and that there is no necessary being in the empirical world—are also both true. See B558 560/A530 A532; B563 5/A535 A557; B588 592/A560 564. Also see note 2, chapter 3. ¹⁷² See B39 B40/A25; B47–B48/A31–32. Kant’s account of space and time as forms of sensibility will be discussed in detail in sections 3.4 and 4.2. ¹⁷³ In the antinomy chapter, Kant directly moves from the infinite divisibility of space to the infinite divisibility of matter. See B553/A525: “Its [a body’s] divisibility is grounded in the divisibility of space which constitutes the possibility of the body as an extended whole. The body is thus infinitely divisible, without, however, therefore being composed of infinitely many parts.” But in the Metaphysical Foundations of Natural Science, he makes clear that the infinite divisibility of matter follows from the infinite divisibility of space only “if it has been demonstrated previously that in every part of space there is material substance,” and not just the activity of a material substance (MAN, 4:504). This important qualification is needed to rule out the kind of monadological theory of matter that Kant himself had advocated in his early Physical Monadology, according to which substances fill space, not

92

      

delineated by neighboring appearances, and as preceded by previous and succeeded by posterior appearances. Given this manner of representing appearances, it is not difficult to see that the empirical regress to increasingly smaller parts of appearances cannot possibly terminate in ultimate parts, and the empirical regresses to increasingly remote spatial and temporal conditions of appearances cannot possibly terminate in absolute spatial or temporal boundaries of the empirical world. Since any given appearance is represented as divisible into smaller parts, the regress by means of which we extend our experience to smaller and smaller material parts goes on ad infinitum and cannot possibly bottom out in the representation of appearances as simple. And the representation of an appearance that marks the ultimate spatial end or the ultimate temporal beginning of the empirical world would mean representing it as delineated by mere empty space or preceded by mere empty time, respectively. But empty space and empty time outside of the world can neither be perceived nor inferred on the basis of the perception of objects that are causally connected with them, which means that they are literally nothing, and thus cannot possibly function as the desired conditions. [I]n an empirical regress no experience of an absolute boundary can be happened upon, and [no experience] of a condition as one that is empirically absolutely unconditioned. The reason for this is that such an experience would have to contain a delineation of appearances through nothing, or the empty, which the continued regress would have to be able to discover by means of a perception, which is impossible. (B545/A517)¹⁷⁴

Kant pithily summarizes this line of reasoning by saying that if we assumed a first beginning in time or ultimate boundaries in space or simple parts of bodies, the empirical world would be “too small” for the understanding.¹⁷⁵ Reason, in its search for the through the plurality of their parts, but through their activity. See PM, 1:480 482. That is, while it is correct to say that the structural features of space are inherited by appearances, this happens not quite as directly as Kant’s presentation in the Critique and my presentation in the main text suggest. For further discussion, see Jauernig (forthcoming 2021b). Also remember that the claim that all bodies are infinitely divisible is different from the claim that all bodies are composed of infinitely many parts. Kant endorses the former but rejects the latter; and he can do so because he is a transcendental idealist. See B552/A524: “Nevertheless it is in no way permitted to say of such a whole that is divisible to infinity that it consists of infinitely many parts. For even though all parts are contained in the intuition of the whole, still the entire division is not contained in it, which consists only in the progressive decomposition, or the regress itself, which first makes the series actual. Now, since this regress is infinite, all of the members (parts) to which it progresses are contained in the given whole as an aggregate, but the entire series of the division is not contained in it, a series that is successively infinite and never whole and which cannot constitute an infinite multitude nor a taking together of this multitude in a whole.” Also see MAN, 4:506 507, quoted in note 168. ¹⁷⁴ Also see B455+457/A427+429; B459+461/A432+433; B548–549/A520–521. ¹⁷⁵ See B514–515/A486–487: “ . . . on whatever side of the unconditioned of the regressive synthesis of appearances a cosmological idea may come down, it would still be too large or too small for every concept of the understanding. . . . For assume first that the world has no beginning, then it is too large for your concept; for this concept, which consists in a successive regress, can never reach the entire elapsed eternity. Assume that it has a beginning, then again it is too small for your concept of the understanding in the necessary empirical regress. For, since the beginning still presupposes a time that precedes it, it is not yet unconditioned, and the law of the empirical use of the understanding still obliges you to ask for a higher temporal condition, and the world is thus obviously too small for this law. Exactly the same holds for the double answer to the question about the size of the world with respect to space. For if it is infinite and without limits, then it is too large for any possible empirical concept. If it is finite and limited, you rightly ask what determines this limit? . . . Thus a limited world is too small for your concept. Second, if every appearance in space (matter) consists of infinitely many parts, then the regress of the division is always too large for your concept; and if the division of space is supposed to stop with some part of it (the simple), then it is too small for the idea of the unconditioned. For this part still admits a regress to multiple parts that are contained in it.”

   

93

unconditioned, would be satisfied, but the understanding could not handle it, so to speak. Our intellectual cognitive faculty is constitutionally unable to represent ultimate empirical conditions; in its collaboration with sensibility, it always hankers after the empirical condition for any given appearance and for the condition of the condition and so on without end. Hence, we cannot hope for a completion of the empirical regresses that are part of the construction of experience in a finite number of steps. Turning to the second option, that is, that the relevant regresses could be given to us in their entirety as infinite, Kant’s main reason for ruling it out is that an infinite regress given to us in its entirety would amount to a completed or actual infinity, which, at least in the circumstances under consideration, is impossible. Why is a completed or actual infinity impossible in these circumstances? The more straightforward reason is that we are talking about empirical regresses, that is, regresses that happen in time and take time, which means that in order to complete them humanity would have to live, not only forever, but through a completed, infinitely long duration, which is impossible.¹⁷⁶ The deeper reason is that our understanding is discursive. On my reading, by characterizing human understanding as discursive Kant primarily means that it carries out all of its operations in piecemeal fashion, so to speak, that is, in successive steps each of which take time. This holds, in particular, for the comprehension of concepts, which consists in the successive, individual apprehension of the marks that compose them in a time consuming, step-wise procedure, that is, in ‘running through’ (discurrere) the marks one by one. (Kant conceives of concepts as sets of other concepts, the latter of which he calls the ‘marks’ or ‘parts’ of the former.) By contrast, an intuitive intellect can apprehend all of the components of its representations with consciousness at once, or at one glance, as one might say, without having to go through them one-by-one, even if there are infinitely many of them.¹⁷⁷ The indicated piecemeal mode of operation to which our discursive understanding is restricted also applies to any kind of regress in thought from something conditioned to its conditions. As a result, all of our concepts, of necessity, are only finitely complex, and any regress from a conditioned thing to its conditions that we could possibly complete with the understanding does not exceed a determinate finite number of steps, which is as much as to say that a completed infinite regress is impossible for us.¹⁷⁸ In the discussion of the antinomies, Kant

¹⁷⁶ See B548/A520: “Thus, I cannot say: the world is infinite with respect to past time or space. For such a concept of magnitude, as a given infinity, is empirically absolutely impossible and, accordingly, also with respect to the world as an object of the senses.” ¹⁷⁷ See VT, 8:389: “For the discursive understanding must expend much work by means of the former [cognition through concepts] for the analysis and again for the synthesis of its concepts according to principles and must ascend many steps in a tiresome way in order to make progress in cognition, whereas an intellectual intuition would grasp and represent the object immediately and at once.” See V-Th/Baumbach 28:1267: “Our understanding is discursive, i.e., we cognize things through general marks that we determine by and by in such a way that they signify an individual. But in this way I cognize things only successively and not at once according to all of their predicates. But these are obvious shortcomings; for this reason, God cannot have a human understanding. He must have an understanding that represents things at once, that intuits them, so to speak.” For further discussion of the sketched reading of the discursivity of the understanding, see Jauernig 2019 and Jauernig (in preparation). ¹⁷⁸ See R4079, 17:406: “The difficulty to represent a simultaneous quantum as infinite is grounded in the nature of human understanding, which can think a totum according to its possibility only synthetically, i.e., by successively adding one to one. But the synthesis that is supposed to go to infinity is never complete. On the other hand, one can indeed think a successive infinite, precisely because the synthesis has no end. But exactly for that reason it is impossible according to the nature of our understanding to think the successive series as whole and completed in the idea of a being.” Also see R5903, 18:379–380; V-Met/Volckman, 28:440; also recall B552/ A524, quoted in note 173.

94

      

sums this up by saying that if we assumed that the world is infinite in space and time, or that appearances are composed of infinitely many parts, the world and appearances would turn out to be “too large” for the understanding, which, due to its discursive nature, cannot complete any operations that require infinitely many steps.¹⁷⁹ Again, reason, in its pursuit of the unconditioned, would be all in favor of a completed infinite synthesis, but the understanding is constitutionally incapable of complying. In sum, the empirical regresses relevant for the mathematical antinomies are never given to us in their entirety and could not be given to us in their entirety, not as finite because the structural features of space and time that are reflected in the way in which the understanding thinks series of conditions in space and time make it incapable of representing ultimate empirical beginnings, limits, or parts, and not as infinite because the discursive nature of the understanding makes it incapable of completing any infinitely complex operations. With respect to Kant’s conception of experience, this means more generally that, due to the nature of our cognitive faculties, the construction of experience, including in particular the extension of our account of the empirical world to more remote regions of space and time and to smaller constituent parts of appearances is of necessity a persistently incomplete work in progress. This explains, not only why the size of the empirical world and the composition profile of appearances are indeterminate, but also why Experience is impossible for us to achieve. As time goes on, our account of the empirical world becomes more and more comprehensive, refined, and determinate but a final, completely determinate account that comprises all appearances and all determinations of appearances that are possible in principle, given the nature of our faculties and the prevailing ultimate material conditions for the possibility of experience, is necessarily beyond our reach. Given this understanding of Kant’s conception of experience, it is no surprise to read, in the Opus Postumum, that experience is an “asymptotic approximation to the empirical completeness of perceptions,”¹⁸⁰ and that we are working on an “asymptotic approximation” to experience (where the latter is to be understood as Experience).¹⁸¹ Experience (with a capital E) is the imaginary infinite limit that we asymptotically approach in the construction of experience in the long run, but it is not something that we could possibly reach.

2.8 Actual or Possible Experience? We are still looking for the one experience that ontologically specifies appearances. Here is what we know so far about Kant’s conception of the construction of experience. The experience of the community of human minds at a given time in history is a shared, or at least sharable, account of the empirical world that includes: a list of general laws of nature governing the behavior of all appearances as well as of special empirical laws of nature governing the behavior of appearances of a certain kind that have been specified by science so far, the former of which are grounded in the transcendental principles of the Critique and the latter of which are grounded in both the transcendental principles and the available empirical evidence; a list of general commonsensical descriptions of causal laws proposed so far that are inductively based on observations and govern various different kinds of

¹⁷⁹ See note 175.

¹⁸⁰ See OP, 21:53, 61.

¹⁸¹ See OP, 21:46, 90; OP, 22:99, 102.

   ?

95

phenomena;¹⁸² a record of all appearances that have been perceived to date as well as of all appearances that have been inferred so far on the basis of known laws and the available empirical evidence; and a collection of empirical judgments and judgments of experience made so far concerning the determinations of these appearances, including their relations to each other. For ease of reference, I will call such an account of the empirical world at a given time an ‘ew-account at t,’ where ‘ew’ stands for ‘empirical world.’ As science and exploration progresses and we further pursue the various ‘empirical regresses’ from given appearances to the different kinds of empirical conditions discussed in section 2.7, our ewaccounts become more detailed, determinate, refined, and comprehensive, both with respect to the inventory of appearances and their determinations contained in the world and with respect to the laws governing them. In order to safeguard against potentially ending up with an uncomfortably fickle or eternally provisional empirical world (in the eventuality that appearances are ontologically specified by actual experience), we must also assume that the evolution of our ew-accounts is conservative and cumulative so that there are no revisions or rejections of earlier ew-accounts. Every later ew-account in the series incorporates all of the earlier ones in some way. This does not mean that we have to deny the obvious fact that the progress of science frequently involves the correction and rejection of earlier scientific theories or of earlier commonsensical descriptions of certain causal laws or of the existence of certain previously inferred objects. We merely have to assume that (1) no presumed law that is properly grounded in the transcendental principles of the Critique is ever up for revision, and (2) no ew-account includes any descriptions of appearances or statements of presumed laws, be they scientific or commonsensical, that conflict with a transcendentally properly grounded law that has been, will be, or could be, specified by science, keeping fixed the transcendental level of reality.¹⁸³ Of course, there probably will be many such descriptions and statements on the record at any given time; they just will not be part of the ew-account at that time. The construction of experience thus corresponds to a series of ew-accounts whose complexity, comprehensiveness, systematicity, unification, and determinateness steadily increase as time goes on such that each ew-account in the series encompasses all ew-accounts prior to it. This series asymptotically approaches Experience in the ideal infinite limit but without possibly ever reaching it. Now, the impossibility for us of attaining Experience does not disqualify it in general from playing the role of the special kind of representation that ontologically specifies the ontological furniture of the empirical world that we are trying to identify. But, as I see it, it disqualifies it from playing this role in Kant’s version of idealism. Kant characterizes appearances both as objects of experience and as objects of possible experience; but if ¹⁸² Typically, descriptions of this kind will say not much more than that there is some law-like causal connection between events of type A and events of type B without yet specifying the precise causal mechanism or any quantitative relationship of relevant magnitudes associated with these event types. The latter further determination of the law is the job of science. ¹⁸³ In other words, not all of science and not all of our commonsensical inductive inferences are part of the construction of experience. Note that the only clear examples of scientific laws that are properly grounded in the transcendental principles that Kant himself can point to are his three laws of mechanics and, possibly, the universal law of gravitation. There is no room in his philosophy for the possibility that these laws could be superseded, let alone that they could be superseded in a non-conservative way. Of course, I do not mean to suggest that Kant is right about this—and it looks as if he may have been wrong about it. Like several neo-Kantians, I believe, though, that Kant’s account can be amended to allow for the conservative supercedence of Newtonian physics without damage to the ‘spirit’ of critical idealism. But this is a story for another day.

96

      

appearances were ontologically specified by Experience, they would be objects of impossible experience, which strikes me as sufficiently unKantian to dismiss this option. We can thus confirm, as our first result, that Experience is not the one experience that we are looking for, on which the ontological specification of appearances depends. An important question that has been hanging over our head for a while now is whether appearances are ontologically specified by actual or possible experience. We still have some more leg-work to do before we can tackle this question head-on, the first step of which is to get clear on what exactly actual and possible experience are supposed to be. Actual experience is simply the series of ew-accounts that the community of human minds develops throughout the course of history. What exactly possible experience is, on Kant’s theory, is unfortunately not as easily stated. In order to explicate the multifaceted meaning of this term in most expedient fashion, it is helpful to avail ourselves of the notion of a possible world and of a modest version of possible worlds semantics for modal expressions. From our discussion of Kant’s tiered conception of reality in section 2.2, we know that, according to critical idealism, every possible world in which human minds exist comprises (at least) two components, namely, a mind-independent, transcendental level of reality at which things in themselves exist and a mind-dependent, empirical level of reality at which appearances exist. Since the forms of our cognitive faculties are essential to us as human beings, the formal conditions of experience are the same at all possible worlds in which human minds exist. But, for all we know, not all possible worlds in which human minds exist agree with respect to their transcendental levels; indeed, if human freedom is supposed to be really possible they cannot all agree. That is, the ultimate material conditions for the possibility of experience are not the same at all possible worlds in which human minds exist. This means that what sensations and perceptions are possible for us, and thus which ew-accounts are possible for us, depends on which possible world we are talking about. Unicorns may not be objects of possible experience in the actual world, but they may be so in other possible worlds, due to a difference in what kind of things in themselves exist in these worlds. Accordingly, there are various different senses in which an ew-account can be said to be possible. The three most central of these senses can be characterized as follows. First, an ew-account ε is possible at possible world w if, and only if, a community of human minds could discover ε given the ultimate material conditions for the possibility of experience obtaining at w. That is, more explicitly, ε is possible at w if, and only if, there is a possible world w*, whose transcendental level of reality is the same as w’s, and ε is on the collective books of the community of human minds in w*. Second, we will say that an ew-account ε is generally possible if, and only if, there is a possible world at which ε is on the collective books of the community of human minds. Third, if we drop the implicit restriction of experience to human experience, we arrive at an even more general notion of possibility—which Kant countenances as well. We will define that an ew-account ε is abstractly possible if, and only if, there is a possible world at which ε is on the collective books of a community of finite minds, be they human or non-human.¹⁸⁴ In the service of conceptual clarity, corresponding to the world-indexed notion of a possible ew-account we ¹⁸⁴ For example, in his discussion of the question whether “the field of possibility is larger than the field that contains everything actual” in the section on the postulates of empirical thinking, one of the hypothetical situations that Kant briefly mentions in which the field of the possible would be larger than the field of the actual assumes that there are non-human finite minds with forms of sensibility and forms of the understanding that are different from ours; see B282–283/A230. Kant thinks that we cannot know if such minds are possible, but if they

   ?

97

may also define a world-indexed notion of an actual ew-account. An ew-account ε is actual at possible world w if, and only if, ε is on the collective books of the community of human minds in w. Finally, it is worth noting that Kant’s central notion of real (as opposed to merely logical) possibility that applies to appearances, which we may call ‘empirical possibility,’ is closely tied up with the notion of possible experience: an object is empirically possible if, and only if, it is an object of possible experience, that is, if and only if, it is represented in a possible ew-account.¹⁸⁵ Accordingly, we can specify that an object is empirically actual at w/empirically possible at possible world w/generally empirically possible/abstractly empirically possible if, and only if, it is represented in an ew-account that is actual at w/possible at w/generally possible/abstractly possible. With all of these clarifications in place, let us return to our question of how to think about the one experience that ontologically specifies the ontological furniture of the empirical world. In virtue of what kind of experience do appearances exist and have all of their determinations and other ontological ingredients? This question can be understood as asking, more generally, about the transcendental conditions of the existence of appearances in any possible world in which they exist, or, more specifically, about the transcendental conditions of their actual existence, that is, of their existence in the actual world. Since we have already established that appearances are ontologically specified by human experience, we can directly set aside the suggestion that appearances exist in possible world w/in the actual world in virtue of being represented in an ew-account that is abstractly possible. That there is a possible world in which some strange creatures are objects of possible experience for Neptunians does not have any implications for the existence of these creatures as appearances to us. Similarly, the suggestion that appearances exist in possible world w/in the actual world in virtue of being represented in an ewaccount that is generally possible can also be discarded right away. That there is some possible world in which, say, unicorns are objects of possible experience (now again understood as human experience) is surely irrelevant for whether or not unicorns exist in w/in the actual world (unless the world in question is w/the actual world). Whether appearances exist in w/in the actual world in virtue of possible experience or in virtue of actual experience, it seems incontrovertible that the experience in question must be possible or actual at w/at the actual world. It is a bit cumbersome to simultaneously address our question in its more general and its more specific form and to carry the explicit indices to a possible world or the actual world around with us. Indeed, this is not only cumbersome but obsolete, since the answer to the more general version of the question can easily be derived from the specific version, and vice versa, by simply exchanging the appropriate indices. Accordingly, in the following, I will explicitly discuss only the more specific version of our question that asks in virtue of what kind of experience appearances exist in the actual world, which also appears to be the case that Kant himself is primarily concerned with. Having made clear that I am talking about the actual world, for simplicity’s sake I will drop the explicit indices and the explicit qualifications from now were, the field of the possible could be identified with the realm of the abstractly empirically possible, as defined in the main text below, in which case it would be clearly larger than the field of the empirically actual. ¹⁸⁵ One can identify a veritable arsenal of various different kinds of possibility and necessity in Kant’s writings, both with respect to appearances and with respect to things in themselves. Fortunately, for our present concerns, there is no need to enter this arsenal any further than just delineated in the main text. Interested readers should take a look at Stang 2016, especially, chs 7 and 8.

98

      

on. That is, unless otherwise indicated, for the remainder of this chapter, ‘possible/actual experience’ and ‘possible/actual ew-account’ should be understood as short for ‘(human) experience that is possible/actual at the actual world’ and ‘ew-account that is possible/ actual (for us) at the actual world’ respectively, and ‘exists’ and ‘exists in the empirical world’ should be understood as short for ‘exists in the actual world’ and ‘exists in the empirical world at the empirical level of reality of the actual world’ respectively. The proposal that appearances exist and have all of their determinations and other ontological ingredients in virtue of being represented in Experience, that is, in a final, completely determinate, comprehensive, thoroughly unified, systematic account of the empirical world, is untenable because the attainment of Experience is impossible for us. Assuming that our approach to Experience is asymptotic, that is, that in the long run we are getting arbitrarily close to it but without ever reaching it, we also must conclude that it is untenable to say that appearances exist in virtue of being represented in the possible ewaccount that is most determinate and most comprehensive and, hence, closest to Experience. For if the series of ew-accounts approaches Experience in an asymptotic manner, talk about the most comprehensive and most determinate possible ew-account is as meaningless as talk about the negative real number that is largest and, hence, closest to 0. There simply is no such largest negative real number; and there simply is no possible ewaccount that is most comprehensive and most determinate. How about saying that appearances exist in virtue of being represented in the actual ew-account that is most determinate and most comprehensive? Whether the notion of a most comprehensive and most determinate actual ew-account is meaningful or not depends on what the transcendental level of reality is like. More specifically, it depends on whether the transcendental level is such that, at some point in time, all human minds will be effectively annihilated without hope of ever being re-created.¹⁸⁶ In this case, there will be an actual ew-account that is most comprehensive and most determinate and, hence, closest to Experience, namely, the one on the books at the moment of annihilation of the last human mind standing, or, perhaps more appropriately, hovering.¹⁸⁷ But if it so happens that human minds continue to grace the world with their presence in indefinitum, there is no most comprehensive and most determinate actual ew-account but merely a never ending series of accounts of increasing comprehensiveness and determinateness. Given that we appear to be well underway toward annihilating ourselves, the proposal that appearances exist in virtue of being represented in the most comprehensive and most determinate actual ewaccount might indeed seem promising for the case of the actual world. But since human ¹⁸⁶ I use the somewhat awkward formulation ‘the transcendental level is such that . . . ’ since the human minds on which the constitution of the empirical world depends do not themselves exist at the transcendental level of reality. The representations that constitute experience are representations of empirical human minds, which exist at the empirical level. ¹⁸⁷ Would this not conflict with Kant’s claim that the size of the empirical world and the composition profiles of appearances are indeterminate? It is true that the final ew-account left behind by the last human minds, just like any possible ew-account, would cover a determinate finite spatiotemporal region and would include descriptions of the decomposition of appearances into their constituents that feature a determinate (finite) number of parts. But it is also true that, for the reasons detailed in the previous section, it would not include a ‘that is all’ clause either or representations of ultimate spatial boundaries or of an ultimate beginning or end in time or of ultimate parts that cannot be further divided. And Kant seems to think—not implausibly, it seems to me—that the intentional object of a representation that neither includes a ‘that is all’ clause nor represents ultimate boundaries nor specifies that the object’s size is infinite is indeterminate with respect to size, and that the intentional object of a representation that neither includes a ‘that is all clause’ nor represents ultimate parts nor specifies that there are no simple parts is indeterminate with respect to its composition profile.

   ?

99

annihilation does not appear to be a necessary event, the proposal cannot serve as a general explication of the experience-dependence of appearances that covers all possible worlds in which appearances exist—which, however, is what we ultimately want. For this reason, we will set the proposal aside for the time being and, unless indicated otherwise, assume for the remainder of this chapter that the human species will never face extinction. Since the strategy of trying to single out a particular ew-account that ontologically specifies the empirical world—Experience or the most comprehensive and most determinate possible ew-account or the most comprehensive and most determinate actual ew-account—has proved to be problematic, we find ourselves with the following final two options for how to cash out the experience-dependence of appearances. Appearances exist and have all of their determinations and other ontological ingredients in virtue of being represented (1) in some actual ew-account, or, alternatively, (2) in some possible ewaccount.¹⁸⁸ Call option (1) ‘A’ and option (2) ‘P.’ I will wrap up this chapter by arguing that A is the more plausible reading of Kant’s conception of appearances. Questions of Kant exegesis aside, from an independent point of view, A seems preferable to P, if only because it renders appearances less ontologically suspect. For something to exist in virtue of being represented is already somewhat peculiar, but for something to exist in virtue of being possibly represented sounds like a situation in which a magician pulls an actual rabbit out of a possible hat. Since Kant is not at risk for winning any popularity contests for his idealism anyway, we should try not to saddle him with an especially eye-brow raising version if it can be avoided. Turning to Kant’s text, one has to admit that he does not only characterize appearances as objects of experience but also often speaks of them as objects of possible experience. But Kant’s use of this formulation by itself does not amount to conclusive evidence for P as long as there are other ways of accounting for this use, which, indeed, there are. Even if appearances are defined as intentional objects of actual experience, it may, on occasion, be useful to talk about objects of possible experience. Similarly, when Kant uses the expression ‘objects of possible experience’ he could simply be thinking of possible (not actual) appearances or even generally possible appearances, which, everybody agrees, depend on possible experience. For example, Kant repeatedly says that the categories and the synthetic a priori principles of the understanding apply to all objects of possible experience. But, clearly, he does not mean to say only that they apply to actual appearances; he means that they apply to anything that could possibly be an appearance for us, that is, they apply to all appearances in any possible world in which human minds exist. That these principles are necessarily true of all appearances is precisely what is remarkable about them. Furthermore, sometimes Kant can be read as using ‘possible’ and ‘actual,’ not in a proper modal sense but in a primarily temporal sense, in which the terms mean ‘future’ and ‘present’ relative to the present moment. Possible perceptions or experiences in this sense are perceptions or experiences that, with respect to the present moment, will actually be had later on.¹⁸⁹ ¹⁸⁸ Remember that, as stipulated above, all of these claims are still supposed to be understood as indexed to the actual world. That is, unless otherwise indicated, we are talking about the existence of appearances in the actual world and about ew-accounts that are actual/possible at the actual world. ¹⁸⁹ See B521/A493: “Nothing is actually given to us apart from perception and the empirical progress from this to other possible perceptions. For in themselves appearances, as mere representations, are actual only in perception, which, indeed, is nothing other than the actuality of an empirical representation, i.e., appearance. To call an appearance an actual thing before the perception either means that we must find such a perception in the progress of experience, or it has no meaning at all.” Also see B184/A144: “The schema of possibility is the

100

      

In the foregoing discussions in this chapter, I tried deliberately to maintain neutrality between A and P as best as I could in order not to pre-judge the issue, but it is now time to acknowledge that much of the textual evidence considered before coheres much better with A. A case in point is the textual evidence that we looked at in our discussion of how perception underwrites the existence of appearances. Throughout the section, Kant never characterizes the perceptions that establish the existence of appearances as possible but does explicitly characterize them as actual at several places. For example, in the previously quoted central passage, in which he first explicates the postulate of actuality, Kant says that the postulate requires the object’s connection “with some actual perception, in accordance with the analogies of experience” (B272/A225, my emphasis). And later on he adds that in this case “the existence of the thing still hangs together with our perceptions in a possible experience, and we can get to the thing in the series of possible perceptions from our actual perception according to the guidance of the analogies” (B273/A225–226, my emphasis).¹⁹⁰ Another case in point is the textual evidence that we looked at in our discussion of the mathematical antinomies in the previous section. Kant’s formulations that the empirical world exists only “in the empirical regress of the series of appearances,” that appearances exist only “in the progress of experience,” and the parts of appearances are given only “in the regress of the decomposing synthesis,” which recur throughout the antinomy chapter, strongly suggest that he takes appearances to be ontologically specified by actual experience. Kant uses these formulations to make the point that, due to the status of empirical objects as appearances, the series of their various conditions are not automatically given with them. But, more specifically, as Kant’s discussion makes clear, he also means to assert, not only that the series of the conditions for a given conditioned appearance is never given to us as a totality, but also that each condition is given to us and exists only in virtue of us actually extending the relevant empirical regress to where it is located in the series. For example, in the previously referenced passage at B527/A499, Kant puts the point as follows: Now it does not follow at all that if the conditioned (in the appearance) is given, then the synthesis that constitutes its empirical condition is thereby also given and presupposed, rather it [the synthesis] first takes place in the regress and never without it. But what one can say in such a case is that a regress to the conditions, i.e., a progressive empirical synthesis on this side, is demanded or set as a task, and that there cannot be a lack of conditions that are given through this regress. (B527/A498-99)¹⁹¹ harmony of the synthesis of different representations with the conditions of time in general . . . , and thus [it is] the determination of the representation of a thing at some time.” Also see B524/A496. ¹⁹⁰ But does Kant not also talk about possible perceptions and possible experience here? Yes, but these possible perceptions have nothing to do with establishing the existence of the thing in question. He is talking about in principle possible perceptions of this thing and of the objects that are involved in the causal chain that links our actual perception to it. Since, due to their causal connection to our actual perception, all of these objects have been established to exist, they must also be in principle perceivable (even though, as Kant goes on to explain, they may not be actually perceivable due to the roughness of our senses, that is, there may not be a possible world in which our senses are the same as in the actual world and in which these objects are perceived). The context makes it quite clear that, in the quoted passage, Kant is using ‘possible experience’ to refer to such a series of in principle possible perceptions of objects in a causal chain. ¹⁹¹ Also recall MAN, 4:506: “For what is actual only through being given in representation, of that also not more is given than as much as is found in the representation, i.e., as far as the progress of representations reaches. Thus, of appearances whose division goes to infinity one can only say that there are as many parts of the appearance as we may give, i.e., as far as we ever may divide.”

   ?

101

If appearances were ontologically specified by possible experience, the series of their conditions also would not and could not be given as a totality—for the reasons discussed in section 2.7, the cognition of such a series as a totality is impossible for us—but it would exist as a never ending series even before we actually embarked on the regress. For it is certainly possible for us to cognize the conditions of a given appearance and the conditions for the conditions and so on. The reading that the regress in which appearances and the empirical world exist must be an actual regress is also called for by the formulation that this regress “is set as a task.” Saying that I set a possible regress as a task for myself is hardly intelligible. The last lingering doubt about whether Kant is talking about an actual regress is quelled when he writes, a little later on, that in the case of appearances he could not presuppose “the absolute totality of the synthesis and the series that is represented by means of it, . . . because here all members of the series are possible only through the successive regress, which is given only through one’s actually carrying it out” (B528 529/ A500 501, my emphasis).¹⁹² In addition to the strong textual support just considered, a somewhat speculative but instructive systematic consideration in favor of A is that it is not obvious that the formal and ultimate material conditions for the possibility of experience, that is, the forms of our cognitive faculties and the things in themselves that happen to exist, together uniquely determine one possible experience. Perhaps these conditions are compatible with a multiplicity of different series of possible ew-accounts that asymptotically approach different comprehensive, completely determinate, systematic, unified accounts of the empirical world. For ease of reference, I will designate this eventuality as ‘underdetermination’ for short, since it amounts to an underdetermination of possible experience by the formal and ultimate material conditions for the possibility of experience. To be sure, the ontological inventories of the empirical world yielded by such different series of possible ew-accounts could not differ that wildly, simply because these accounts are all bound by

¹⁹² Indeed, one might even go so far as to read Kant as saying that, from the point of view of fundamental ontology, the parts of appearances and the remoter regions of the empirical world only first come into existence when our experience reaches them (where the ‘when’ refers to a certain point in time in the empirical world). See V-Met/Mron, 29:856: “The world of sense is not a whole of things but merely a sum of appearances (the world of sense is nothing but the concatenation of my representations of things that I receive from sense according to the relations of space and time; hence, it lies merely in my head and is not given in itself but merely in the progress of my experience of things, e.g., a rose is not red before I see it. For its redness rests on the nature of the eyes. The cause of its redness is in it, but that it is actually red depends on my eye . . . ).” This conception of an evolution of the ontological furniture of the empirical world that parallels the evolution of our ew-accounts is not easy to wrap one’s head around, since it has some time-bending implications. For instance, suppose for the sake of the example that modern-day paleontology is a transcendentally grounded science. On the conception under consideration, from the point of view of fundamental ontology, before the rise of paleontology it was indeterminate whether there are dinosaurs and the empirical world did not include any of them, but once paleontology had assumed its place in the canon of transcendentally grounded scientific theories, dinosaurs were included, namely, as existing about 200 million years ago. (What should we say about all this from the empirical point of view? I will leave this question as an exercise for the reader. Also note that, on the proposed reading, it is still the case that when and for how long dinosaurs exist in the empirical world is independent of when and for how long anybody actually represents them. Only their first coming into existence depends on when somebody actually first represents them.) Of course, choosing A does not force us to read Kant in the indicated way; all that A implies is that for an appearance to exist it has to be represented in actual experience at some point, which is compatible with holding that, from the point of view of fundamental ontology, dinosaurs were included in the empirical world already before the actual development of paleontology. But only A can accommodate the reading that appearances literally first come into existence when they are first represented by one of us (even though their existence might date back millions of years). However, because of the mentioned time-bending implications of this reading and because of a lack of compelling textual evidence that this was indeed Kant’s view, I will refrain from adopting it.

102

      

the same constraints, that is, the same formal and material conditions for the possibility of experience. But there still might be room for discrepancies due to variations in how we sort things into genera and species and subsume various regularities under special empirical laws, which could lead to differences in what qualitative determinations, in particular causal determinations, objects are represented to have, for instance. If such an underdetermination obtained, we would have another reason for adopting A. For, in this case, if we went with P, we would be committed to the view that all of the appearances that are represented in these different possible experiences exist, which would make the empirical world not only overcrowded but inconsistent. Admittedly, there is no decisive textual evidence that would force us to conclude that Kant thinks that this kind of underdetermination obtains or could obtain, but he does make some remarks that could be read in this way.¹⁹³ At any rate, other things being equal, it speaks for A that by adopting it we can avoid having to worry about a possible underdetermination problem altogether. Given the foregoing considerations, I conclude that Kant conceives of appearances as described by A. Removing the implicit index to the actual world that we have been carrying for much of this section, we can say more precisely and more generally that (outer) appearances exist and have all of their determinations and other ontological ingredients in possible world w in virtue of being represented in an ew-account that is actual in w, that is, an account of the empirical world of the kind described at the beginning of this section that, at some point in history, is on the collective books of the community of human minds in w. If we define actual experience at w as the conjunction of all ew-accounts that are actual at w, the same point can be put by saying that appearances

¹⁹³ As mentioned in note 184, in the section on the postulates of empirical thinking, Kant considers the question “whether the field of possibility is larger than the field that contains everything actual” (B282/A230). In the course of this discussion, he contemplates several hypothetical scenarios in which the field of the possible would be larger than the field of the actual. Kant glosses the more specific version of the question whether the field of the empirically possible is larger than the field of the empirically actual as asking “whether all things, as appearances, taken together belong to the sum total and the context of one single experience, of which every given perception is a part, which thus could not be connected with any other appearances, or whether my perceptions could belong to more than one possible experience (in its general connection) . . . ” (B282 283/A230). In the same vein, later on he describes the hypothetical scenario in which the field of the empirically possible would be larger than the field of the empirically actual by saying that “in the thoroughgoing connection with what is given to me in perception, another series of experience is possible, and thus more than one single experience is possible that comprehends everything” (B284/A231 232). There are different ways in which Kant’s suggestion can be understood that the field of the empirically possible would be larger than the field of the empirically actual if, and only if, there were more than one possible all-encompassing experience that includes a given actual perception. One way to understand it would be to read it as an articulation of the possibility of the kind of underdetermination just considered in the main text. If, and only if, the formal and ultimate material conditions for the possibility of experience underdetermine possible experience in the indicated way, then the field of the empirically possible is larger than the field of the empirically actual.—But, one may wonder, does this not suggest that if there is no underdetermination then the field of the empirically possible is coextensive with the field of the empirically actual, which, in turn, would amount to an argument for P? No. The indicated result would imply that the field of the empirically possible is coextensive with the field of the empirically actual, but it would not amount to an argument for P. P is not compatible with Kant’s question to begin with. On the assumption of P, it would be clear from the outset that the empirically possible is co-extensive with the empirically actual. (And, as noted in the main text, the main effect of a possible underdetermination in this case would not be to complicate Kant’s question but to render P untenable.) But Kant regards the question as open—which, incidentally, provides additional evidence for A; at least, this is what I take him to mean when he says that the “solution [of this and related questions] falls only under the jurisdiction of reason” (B282/A230). Moreover, it can be argued (see note 195) that A is compatible with the claim that in worlds in which human beings do not go extinct, the empirically possible is coextensive with the empirically actual; and the actual world may well be one of them.

    

103

exist and have all of their determinations and other ontological ingredients in possible world w in virtue of being represented in actual experience at w. From now on, when I say that appearances essentially are intentional objects of experience, I should be understood as having in mind that they essentially are intentional objects of actual experience as just explicated. Lest people are worried that this choice of A leaves us with an empirical world whose ontological furniture is deploringly sparse and disconcertingly growing over time, let me emphasize again that when and for how long appearances exist does not depend on when and for how long an ew-account that represents them is entertained by a human mind but rather on whether there is such an ew-account at some point and when and for how long this account represents them to exist. The number of appearances in the empirical world at this moment is probably much larger than the number of appearances that are represented in our latest ew-account and may even include Neptunians.¹⁹⁴ That there are many such as of yet unrepresented actually existing appearances holds, in particular, if human minds do not go out of existence and the evolution of our ever more comprehensive and ever more determinate ew-accounts continues in indefinitum.¹⁹⁵

2.9 Inner Appearances and Inner Experience So far, we have been talking mostly about outer appearances. What about inner appearances? The first thing to say is that the term ‘inner appearance’ is ambiguous. When Kant uses the term in the plural, as he often does, he employs it to refer to mental states or representations and feelings.¹⁹⁶ Inner appearances in this sense are modifications of inner ¹⁹⁴ See B524/A495 496: “Thus, if I imagine all existing objects of the senses in all time and all spaces taken together, I do not place such objects in both of them before experience, but this representation is nothing but the thought of a possible experience in its absolute completeness. In it alone those objects are given (which are nothing but mere representations). But that one says that they exist before all of my experience only means that they are to be found in that part of experience to which, starting from perception, I must first proceed.” ¹⁹⁵ Indeed, one could even suggest that, on the assumption that human minds will not be annihilated at any point, the field of the empirical actual at the actual world is coextensive with the field of the empirically possible at the actual world. The idea would be that if humans do not go extinct and the series of increasingly comprehensive and increasingly determinate ew-accounts goes on in indefinitum, it turns out that for any object that is empirically possible at the actual world, at some point in history, there is an ew-account on the collective books of the community of human minds in which the object is represented. (Note that this claim is different from saying that we will complete an infinite regress, which would violate the strictures discussed in section 2.7.2 in connection with the antinomies. Neither possible experience nor actual experience can be identified with Experience; Experience is impossible.) It may take a while, but if Neptunians are empirically possible, human minds will discover them at some point, either through direct observation or by means of an inference based on direct observations and established empirical causal laws. Admittedly, it is a somewhat speculative question whether Kant entertained the possibility that the empirically possible might be coextensive with the empirically actual at the actual world. But there is some indirect textual evidence for it. As mentioned in note 193, in his discussion of the question whether the field of the possible is larger than the field of the actual, Kant suggests that the field of the empirically possible would be larger than the field of the empirically actual if, and only if, there could be more than one all-encompassing possible experience that includes a given perception; see B282–284/ A230–232. One way to read this suggestion is as saying that the field of the empirically possible is larger than the field of the empirically actual if, and only if, the formal and ultimate material conditions for the possibility of experience underdetermine possible experience. This reading, in turn, means that if there is no underdetermination, the empirically possible and the empirically actual are coextensive. But how could that be if P is not Kant’s view? Answer: A is compatible with the assumption that the empirically possible and the empirically actual are coextensive in possible worlds in which human minds do not go extinct. Kant regards it as an open question whether, in the actual world, the field of the empirically possible is larger than the field of the empirically actual because it is not clear whether the actual world is one of these worlds. ¹⁹⁶ See Prol, 4:336: “ . . . by means of outer experience I am equally conscious of the reality of bodies as outer appearances in space as by means of inner experience I am conscious of the existence of my soul in time, which

104

      

sense―representations and feelings―which are the result of affections of inner sense by (what I will call) the transcendental self. But Kant also employs the term ‘appearance,’ or ‘inner appearance,’ to refer to the “object of inner sense” or “the soul as object of inner sense,” as he frequently puts it, or the empirical self, as I will call it.¹⁹⁷ In the following, when I talk about inner appearances, I should be understood to have empirical selves in mind unless otherwise indicated. A brief but important note about terminology: Kant himself does not explicitly use the expressions ‘transcendental self ’ and ‘empirical self.’ He often avails himself of variations of the formulation that “inner sense presents ourselves to consciousness only how we appear to ourselves, not how we are in ourselves, since we intuit ourselves only how we are inwardly affected” (B153–154), and employs a variety of related terms to refer to the thing in itself that affects inner sense and grounds this appearance, such as “transcendental subject,” “subject in itself,” or “self proper.”¹⁹⁸ In the service of clarity, it seems desirable to work with a more streamlined terminology here—hence, my introduction of the term ‘transcendental self.’ My transcendental self is the thing in itself (or things in themselves) that brings about my inner sensations through transcendental affections of sensibility, and, more precisely, of inner sense, and, thus, appears to myself as the object of inner sense in my empirical consciousness.¹⁹⁹ From there it is only a short step to adopting the matching term ‘empirical self ’ to designate the appearance that is constituted at the end of this I also can cognize only as an object of inner sense through appearances that constitute an inner state and of which the being in itself that grounds these appearances is unknown to me.” See Anth, 7:144: “The perceptions of the senses (empirical representations with consciousness) can only be called inner appearances.” See BJ, 8:154: “The very same thing can also be shown with respect to the empirical concept of our soul, that it contains mere appearances of inner sense and not yet the determined concept of the subject itself.” ¹⁹⁷ See B156: “ . . . we also must admit of inner sense that through it we intuit ourselves only how we are inwardly affected by ourselves, that is, that as far as inner intuition is concerned, we cognize our own subject only as appearance but not according to what it is in itself.” B334/A278: “ . . . we who know even ourselves only through inner sense and, thus, as appearance . . . ” See FM, 20:269: “In the theory of all objects of the senses, as mere appearances, there is nothing that is more strangely noticeable than that I the object of inner sense, i.e., considered as soul, can become known to myself only as appearance, not according to that which I am as thing in itself . . . ” For the characterization of the soul, or the self, as the “object of inner sense,” also see A368: “ . . . by contrast, the object of inner sense (I myself with all of my representations) is immediately perceived, and its existence does not suffer any doubt.” See B427: “ . . . the presupposed dissimilarity of the object of inner sense (the soul) with the objects of outer senses.” See B471/A443: “ . . . the object of inner sense, the I, which thinks . . . ” See Prol, 4:336: “ . . . as much as the connection of the appearances of inner sense proves the reality of my soul (as object of inner sense) . . . ” See FM, 20:309: “But he [the human being] is also conscious of himself as an object of his outer sense, i.e., he has a body, with which the object of inner sense is connected that is called ‘soul’ of the human being.” ¹⁹⁸ See B520/A492: “ . . . and even the inner and sensible intuition of our mind (as object of consciousness) is . . . also not the self proper, such as it exists in itself, or the transcendental subject, but only an appearance that has been given to sensibility of this being that is unknown to us.” See A355; B427; B506/A478note; B573/A545. It is also useful to mention in this context that, in Kant’s use, the term ‘transcendental subject’ exhibits the same kind of ambiguity as the term ‘transcendental object.’ Sometimes the intended referent is a thing in itself, sometimes the intended referent is a projected bare particular = X that is presented as (a) distinct from our representations of it and (b) a bearer of determinations, a particular that functions as an ontological ingredient of appearances. And sometimes it is not clear which meaning is intended, see B404/A345–346: “ . . . the simple representation I that, for itself, is completely empty of any content, of which one cannot even say that it is a concept but a mere consciousness that accompanies all concepts. Through this I or he or it (the thing), which thinks nothing more is represented than a transcendental subject = x . . . ” The function of the concept of the transcendental object in the construction of outer experience will be addressed further in section 3.5; for a few more comments about the function of the concept of the transcendental subject, see note 212 and section 5.7. ¹⁹⁹ See B68 69: “If the faculty to become conscious of oneself is to collect (apprehend) that which lies in the mind, it must affect the latter and can bring about an intuition of itself only in that a manner, an intuition whose form, however, which previously lies ready in the mind as a ground, determines the manner in which the manifold is together in the mind in the representation of time; in that way it intuits itself, not as it would immediately selfactively represent itself, but according to the manner in which it is affected from the inside, and thus as it appears to itself and not as it is.” Also see KpV, 5:6: “ . . . since it [the Critique] emphasized to regard the objects of

    

105

process instead of employing Kant’s generic “object of inner sense.” In order to avoid confusion, it is also worth noting that my empirical self thus understood is to be distinguished from my empirical self-consciousness, just as my transcendental self is to be distinguished from my transcendental self-consciousness or the original synthetic unity of apperception. As I see it, the empirical self is also best conceived of as an intentional object, which exists in virtue of the fact that there is a finite mind that represents it in experience.²⁰⁰ But the way in which empirical selves and their determinations are represented by us in experience differs in some respects from how outer appearances and their determinations are represented by us in experience, which deserves a few brief comments. Again, some of what I am about to say will be fleshed out more fully only later on, in particular in section 3.6. In contrast to the constitution of the empirical world, which, as we have just seen, is a job for the entire community of human minds, the constitution of our empirical selves is a special job for each one of us. Moreover, since much of the presentational content that feeds into the constitution of an empirical self is, in principle, accessible only to the person whose empirical self it is, empirical selves remain private to a large extent, in contrast to empirical objects, which are public.²⁰¹ Another, somewhat more elusive difference concerns the manner in which the determinations of outer appearances and empirical selves are mind-dependent. The determinations of outer appearances are mind-dependent in the sense that these appearances have them in virtue of the fact that a finite mind presents them to have them, for example, Aachen Cathedral has the determination of being made of stone (among other materials) in virtue of the fact that this is how it is presented by us in outer experience. But one might be skeptical about the claim that I have the determination of currently perceiving a big orange cat in virtue of the fact that I represent myself as perceiving a big orange cat. To be sure, perceptions and, more generally, all of our representations and feelings in inner sense, which make up the determinations of our empirical selves, are mind-dependent in the sense that they cannot exist without being had, or entertained, by a finite mind, but, one might hold, they need not themselves be represented by a finite mind in order to exist. Indeed, one might go as far as to claim that, despite their mind-dependence in the sense just indicated, our representations and feelings in inner sense ought to be regarded as existing at the transcendental level of reality, precisely because their mind-dependence is different from the mind-dependence of the determinations of outer appearances. If these considerations were on the mark, our empirical selves could not be said to be minddependent in the sense specified in section 2.3.

experience as such only as appearances, and among them even our own subject, but still to give them things in themselves as their ground . . . ” ²⁰⁰ See Prol, 4:337: “ . . . the question whether bodies (as appearances of outer sense) exist outside of my thoughts as bodies, can be answered in the negative without any qualms; in this there is no difference to the question whether I myself as appearance of inner sense (soul according to empirical psychology) exist outside of my capacity of representation in time, for this question must also be answered in the negative.” Also see the quotations in note 139, chapter 3. ²⁰¹ Also, recall that by “empirical world” I mean the public physical world at the empirical level of reality that is shared by all of us. Empirical selves are not only private to some extent, they are also immaterial, as we will discuss in more detail in section 3.6. Thus, they are not part of the empirical world. There are also grounds, to be considered in section 3.6, for thinking that, on Kant’s view, empirical selves do not act on empirical objects, which is yet another reason not to count the former as belonging to the empirical world. Empirical selves are part of the empirical level of reality, though, of course.

106

      

These claims touch on issues that, to my mind, are among the trickiest in Kant’s theoretical philosophy, including his conception of inner sense and of what it means for a representation to be conscious as well as his understanding of the relation between our empirical and transcendental self and his conception of the referent of ‘I’ in the famous ‘I think.’ An exhaustive discussion of these issues would be out of place in the present context, but here are a few relevant considerations. (We will return to some of these questions in sections 3.6, 5.7, and 5.10.) To start with the question at which level of reality our representations and feelings in inner sense exist, I think it is safe to say that they are transcendentally ideal and thus do not exist at the transcendental level. How so? Since time is the form of inner sense, all of our representations and feelings in inner sense are essentially in time.²⁰² And since time is transcendentally ideal and does not exist at the transcendental level, anything that is in time is transcendentally ideal as well and does not exist at the transcendental level either.²⁰³ The reading that our representations and feelings in inner sense are transcendentally ideal also fits well with Kant’s frequent characterization of them as inner appearances. To say of a representation that it is an appearance (from the point of view of fundamental ontology) entails that it is transcendentally ideal.²⁰⁴ Note that the transcendental ideality of our representations and feelings that are determinations of our empirical self does not imply that they do not exist from the point of view of fundamental ontology. Just as outer appearances and their determinations, despite being transcendentally ideal, exist from the point of view of fundamental ontology, empirical selves and their determinations exist from the point of view of fundamental ontology, despite their transcendental ideality. Ultimately not that much hangs on whether the determinations of empirical selves are mind-dependent in exactly the same sense as the determinations of outer appearances, and so I will not dwell on this question for long, except to make two observations. First, if being conscious of a representation or feeling, at least in inner sense, counts as a special way of representing it, as Kant suggests at several places,²⁰⁵ and if the degree of consciousness of a ²⁰² See B50/A34: “Time is the formal condition a priori of all appearances in general. Space, as the pure form of all outer intuition as condition a priori is restricted merely to outer appearances. By contrast, since all representations, as determinations of the mind, belong to the inner state, but this inner state belongs, under the formal condition of inner intuition, and thus to time, therefore time is a condition a priori of all appearance in general, namely, the immediate condition of the inner appearances (our souls) and through that mediately also of outer appearances.” Also see note 17. ²⁰³ We will discuss this inference in section 4.1.1. Also note that Kant even repeatedly suggests that all of our representations are in inner sense, which would mean that all of our representations are transcendentally ideal and thus do not exist at the transcendental level. See B50/A34 quoted in note 202, as well as MS, 6:214: “In space there are only the objects of outer senses, but in time all of them, the objects of outer senses as well as of inner sense: since the representations of both are representations after all and insofar altogether belong to inner sense.” See R5636, 18:267: “All representations, they may originate wherever they want, ultimately are, as representations, modifications of inner sense, and their unity must be viewed from this perspective.” There are several complications in this vicinity having to do with Kant’s views about pure consciousness; but, for the present, we are only interested in representations and feelings that are properties of our empirical selves, all of which, undoubtedly, are in inner sense and, thus, transcendentally ideal. ²⁰⁴ Another argument for the claim that representations and even thoughts must be transcendentally ideal is that otherwise we could straightforwardly cognize our transcendental self or, at least, some supersensible aspect of ourselves, which Kant, however, rejects. This point is also made by Pistorius; see Pistorius, 1786, 94. ²⁰⁵ See Log, 9:33: “Properly speaking, consciousness is a representation that another representation is in me.” See V-Met-L1/Pölitz, 28:227: “Consciousness is a knowledge of what pertains to me. It is a representation of my representations, it is self-perception.” See V-Lo/Dohna, 24:701: “Cognition is relation of the representation to an object—connected with an action in the mind—consciousness (representation of our representation), which obscure representations lack.” See V-Anth/Dohna, Kowalewski, 98: “Inner sense is one’s own representation of our thoughts. The soul itself is object of its own representations.”

    

107

representation or feeling is essential to it, my conscious representations and feelings in inner sense can be described as existing at least partly in virtue of the fact that I represent them. Second, since, as we will see, my empirical self is nothing but a projected bearer of my conscious representations and feelings in inner sense, it is legitimate to say that my empirical self has those representations and feelings in virtue of being represented to have them. (This holds even if we were to decide upon further consideration that the description of my conscious representations and feelings as existing partly in virtue of being represented by me is untenable after all.) So, as I see it, our definition from section 2.3 of what it means for a property or ontological ingredient of a thing to be strongly mind-dependent does, in fact, apply to the determinations of empirical selves. But, to be on the safe side, I will add a further clause to the definition to ensure that this kind of determination is covered even if it turns out that the above reasoning went astray at some point. A property P/an ontological ingredient OI of an entity E is strongly mind-dependent if, and only if, either, necessarily, if E has P/comprises OI, then there is a finite mind M, and E has P/ comprises OI in virtue of the fact that M has, or could have, a possibly complex representation whose presentational content includes E having P/comprising OI, or P/OI is a representation or feeling in inner sense. So, with this little safety adjustment in place, our empirical selves can be characterized as fully mind-dependent entities, that is, as entities whose existence and all of whose determinations and other ontological ingredients are strongly mind-dependent. The main take-home message from these considerations for now is that despite the fact that outer appearances are public while inner appearances are to a large extent private, and despite some possible subtle differences with respect to the specific kind of minddependence of their determinations, inner appearances and outer appearances have a lot in common as far as their basic ontological status is concerned. They are all transcendentally ideal and do not exist at the transcendental level of reality, but they all exist at the empirical level of reality and are existents even from the point of view of fundamental ontology.²⁰⁶ Outer appearances are ontologically specified by outer experience; inner appearances are ontologically specified by inner experience, where the latter is also to be understood as actual experience, just like the former.²⁰⁷ Kant indicates that, like the construction of outer experience, the construction of inner experience also depends on the application of concepts.²⁰⁸ But it is questionable whether he understands the transition from perception, or empirical intuition, to experience to work in quite the same way in the two cases. This is questionable, not least because Kant famously holds that there can be no empirical science of psychology that would discover quantitative empirical causal laws, expressible in mathematical terms, that govern psychological phenomena.²⁰⁹ Since the identification of ²⁰⁶ Like outer appearances, inner appearances are also empirically real, that is, they exist at the empirical level and are empirically mind-independent. But since the sense in which inner appearances are empirically mindindependent differs slightly from the sense in which outer appearances are empirically mind-independent, the sense in which they are empirically real is also not exactly the same as the sense in which outer appearances are empirically real. We will get back to these differences in section 3.6. ²⁰⁷ Kant himself uses the terms “inner experience” and “outer experience”; for example, see B275; Prol, 4:265; Anth, 7:141. ²⁰⁸ See Anth, 7:142: “For this reason I cognize myself through inner experience only in the way I appear to myself. . . . Appearance is not a judgment at all but merely empirical intuition that becomes inner experience and thereby truth through reflection and the concept of the understanding resulting from it.” ²⁰⁹ See MAN, 4:471.

108

      

special empirical causal laws that flesh out, or ‘materialize,’ the analogies of experience and govern the interactions between appearances is a crucial element in the transition from judgments of perception to judgments of experience in the case of outer experience, the assertion that a science of psychology is impossible makes it plausible to suppose that Kant must conceive of this transition in a different way in the case of inner experience.²¹⁰ On my reading, the transition from inner empirical intuition to inner experience, and thus the constitution of inner appearances, that is, empirical selves, requires the following three main steps. First, I must recognize my conscious representations and feelings in inner sense as mine.²¹¹ Second, I must conceive of these representations and feelings as my, or more precisely my mind’s, accidents or determinations and, accordingly, must think of myself, or rather my mind, as a subject and substance relative to them.²¹² Third, I must be “conscious of my existence as determined in time,” as Kant puts it, that is, I must be aware of a determinate temporal order and determinate temporal durations of my mental states by way of relating them to the objective temporal order that governs the states of outer appearances.²¹³ If this is the right way to think about inner experience, it turns out that there is considerably less to say about it than about outer experience. This may be why, in developing his theory of experience, Kant himself mainly focuses on the case of outer experience and the constitution of outer appearances and only in passing applies his ²¹⁰ Kant’s repeated observation that we can make the application of the categories properly intelligible to ourselves only with respect to outer appearances may be interpreted as another hint that he conceives of the construction of inner and outer experience and the role of the categories in these constructions in different ways. See B291: “It is even more curious that in order to understand the possibility of things according to the categories, and thus in order to demonstrate the objective reality of the latter, we need not only intuitions but even always outer intuitions.” See MAN, 4:478: “It is also indeed rather curious (but cannot be explained in detail here) that general metaphysics in all cases where it requires examples (intuitions) in order to give meaning to its pure concepts of the understanding must take them always from the general doctrine of body, and thus from the form and the principles of outer intuition, and that, if those are not given in a complete way, it gropes around unsteadily and staggeringly among only senseless concepts.” ²¹¹ Kant clearly counts our feelings as belonging to inner sense and as being among the properties of our empirical selves. [For example, in the Critique of Practical Reason he describes the feeling of pleasure and pain as “a receptivity belonging to inner sense” (KpV, 5:58), and in the Critique of Judgment he explains that a judgment of taste “is called aesthetic because its determining ground is no concept but the feeling (of inner sense) of the harmony in the play of the faculties of the mind insofar as it can be sensed” (KU, 5:228).] This raises the following somewhat awkward question: how can I recognize my conscious feelings as mine? This question is a bit awkward because Kant’s famous answer to the corresponding question of how I can recognize my conscious representations as mine does not seem to be straightforwardly transferrable to the present case. The famous answer, which Kant develops in §16 of the transcendental deduction of the categories, is that I can be conscious of my representation as mine in virtue of having synthetically unified the manifold contained in them in my consciousness according to the transcendental unity of apperception; see B132–135. This, in turn, means that the manifold that is given in each one of my intuitions “stands necessarily under categories,” as Kant spells out in the following sections that culminate in §20; see B143. Does this mean that each one of my feelings also contains a manifold that is synthetically unified in accordance with the transcendental unity of apperception? ²¹² See A379: “ . . . in the connection of experience, matter as substance in the appearance is really given to outer sense, just as the thinking I, equally as substance in the appearance, is given before inner sense; and in both cases appearances must be connected among each other according to the rules that this category introduces into the connection of our outer as well as inner perceptions to an experience.” Also see Anth, 7:161. I take it that this second step in the construction of inner experience also involves the application of the concept of the transcendental subject, which is the ‘inner’ counterpart of the concept of the transcendental object. Just as the concept of the transcendental object, that is, of a bare particular something = X that is (a) distinct from both our representations of it and the cognizer entertaining these representations and (b) a bearer of determinations, is required for us to (directly) represent of outer appearances, so the concept of the transcendental subject, that is, of a bare particular something = X that is distinct from the representations of it and a bearer of determinations, and, more specifically, a subject of representations and feelings, is required for us to (directly) represent empirical selves. ²¹³ See B275–276. This third step, which is the core of Kant’s refutation of idealism, will be discussed in more detail in section 4.1.2.

    

109

results, mutatis mutandis, to the case of inner experience and the constitution of empirical selves.²¹⁴ In order not to complicate our discussion too much, we will follow Kant’s model and continue to keep our attention focused on outer experience and outer appearances. (Thus, as before, ‘appearances’ and ‘experience’ without any qualifications is to be read as ‘outer appearances’ and ‘outer experience’ unless noted otherwise.) We will return to the question of how our empirical selves fit into the picture in section 3.6.

²¹⁴ To be sure, as indicated in note 211, Kant’s theory of experience centrally includes an account of the conditions under which representations can be mine and belong to my self–consciousness, which, at the same time, explains the possibility of my empirical self as a unified entity that persists through time. But Kant’s ultimate purpose in developing this account, in the course of the transcendental deduction of the categories, is to show that and how the categories make experience of outer objects possible and thus a priori relate to them. In the subsequent chapter on the principles of the pure understanding, in which Kant spells out the a priori synthetic laws that reflect the formal conditions of experience and thus apply to all objects of experience, outer experience remains the main concern.

3 Kant’s Transcendental Idealism and Empirical Realism 3.1 Empirical Objects are not Things in Themselves but Appearances In our efforts to explicate Kant’s conception of appearances as intentional objects of experience, we have already had several passing encounters with Kant’s conception of empirical objects. It is time to make this conception explicit. On Kant’s account, empirical objects—or objects in space and time, or sensible objects, or objects of the senses, as he also calls them—are not things in themselves but (outer) appearances. Indeed, if we regard the objects of the senses as mere appearances, as is proper . . . (Prol, 4:314) But from my inquiries it transpired that the objects with which we are dealing in experience are not things in themselves but merely appearances. (KpV, 5:53)¹ The claim that empirical objects are appearances would not be very exciting if ‘empirical object’ were already understood as ‘intentional object of experience.’ The claim has bite, however, if we assume our common sense understanding of empirical objects as the objects that make up the ontological furniture of our lives, so to speak, including, trees, sticks, roses, opossums, and literally items of furniture such as tables and chairs but also less familiar objects investigated by science such as distant galaxies or elementary particles. This is how ‘empirical object’ is meant to be understood in Kant’s assertion that empirical objects are not things in themselves but appearances. Kant acknowledges that it is a “very natural assumption to take the objects of the senses for things in themselves” (KU, 5:344), a “common prejudice” (B768/A740), an assumption that “commonly happens” (Prol, 4:339). Nevertheless, his investigations force us to relinquish this very natural assumption and accept instead that empirical objects are appearances, that is, that they are fully minddependent intentional objects, which exist and have all of their determinations and other ontological ingredients in virtue of being represented in experience. The thesis that empirical objects are appearances and not things in themselves is one of the core tenets that define Kant’s transcendental idealism. Transcendental realists, by

¹ Also see B299/A240: “ . . . they would not mean anything if we could not always explain their significance with respect to appearances (empirical objects) . . . ” See B51/A34: “ . . . all appearances in general, i.e., all objects of the senses, are in time . . . ” B521/A492: “Thus the objects of experience are never given in themselves but only in experience and do not exist apart from/outside it [experience] at all.” Also see FM/LB, 20:338–39: “ . . . the human theoretical faculty of cognition could not reach beyond the objects of the senses and the limits of possible experience and these objects are not things in themselves but merely their appearances.” Further textual evidence will be provided throughout this chapter. The World According to Kant: Appearances and Things in Themselves in Critical Idealism. Anja Jauernig, Oxford University Press (2021). © Anja Jauernig. DOI: 10.1093/oso/9780199695386.003.0003

        

111

contrast, endorse the natural and common misclassification that empirical objects are things in themselves, that is, that they are mind-independent. In the transcendental Aesthetic we have proved sufficiently that everything that is intuited in space or time, and thus all objects of an experience that is possible for us, are nothing but appearances, i.e., mere representations, which, as they are represented, . . . do not have an existence outside our thoughts that is grounded in itself. This doctrine I call transcendental idealism. The realist in transcendental significance turns these modifications of our sensibility into things subsisting in themselves and thus mere representations into things in themselves. (B518–519/A490–91)

The mistaken view that empirical objects are things in themselves is the main reason why transcendental realists are stuck with the mathematical antinomies, as already noted in section 2.7.7.² Since we can escape the antinomies by recognizing empirical objects for what they are, namely, appearances, that is, by becoming transcendental idealists, the antinomies provide additional, indirect support for transcendental idealism.³ We will take a closer look at this indirect argument for transcendental idealism in section 6.6. The realization that empirical objects are not things in themselves but appearances also lies at the heart of Kant’s famous so-called Copernican Turn or Copernican Revolution in philosophy, which, in a well-known passage in the B-preface, he presents as the key to his critical philosophy. So far one assumed that all our cognition had to conform itself to the objects; but all attempts to make something out about them a priori through concepts, whereby our cognition would be enlarged, failed under this assumption. Let us thus try if we cannot succeed better in the tasks of metaphysics by assuming that the objects have to conform themselves to our cognition, which already coheres better with the demanded possibility of an a priori cognition of them, which is supposed to determine something about objects before they are given to us. It is here just as with the first thoughts of Copernicus who, since there was not much progress in the explanation of the motions in the heavens if he

² That the transcendental realist is stuck with the dynamical antinomies, that is, the third and fourth antinomy, has a slightly different ultimate reason although it also centrally involves a failure on his part to realize that appearances and things in themselves are ontologically distinct. In Kant’s assessment, the mathematical antinomies are resolved by realizing that the theses and antitheses are all false if empirical objects are understood to be appearances, while the dynamical antinomies are resolved by realizing that the theses and antitheses can all be true if we understand the theses to be about supersensible things in themselves and the antitheses to be about appearances. So, instead of saying that empirical objects are mistakenly regarded as things in themselves, the relevant error in the case of the dynamical antinomies is better described by saying that empirical objects are mistakenly taken to exhaust everything there is in the world, or that certain things in themselves (our intelligible self and necessary beings) are mistakenly regarded as empirical objects. See Letter to Garve, August 7, 1783, 10:341, note: “If one regards appearances as things in themselves and demands from them, as such, the absolutely unconditioned in the series of conditions, one ends up with nothing but contradictions, which, however, fall away by means of showing that the entirely unconditioned does not take place among appearances but only with things in themselves. If, on the other hand, one regards, the other way around, that which as thing in itself can contain the condition of something in the world as an appearance, one creates contradictions for oneself where none would be necessary, e.g., in the case of freedom, and this contradiction falls away as soon as this different meaning of objects is taken into account.” See FM, 20:291–292; B819/A791. Also see note 171, chapter 2. ³ See Prol, 4:347: “ . . . it is entirely impossible to escape the contradiction of reason with itself as long as one regards the objects of the world of sense as things in themselves and not as what they in fact are, namely, mere appearances . . . ” See B534–535/A506–507; Prol, 4:339 340; KU, 5:344; FM, 20:287–290; RCLXIX, 23:40.

112

 ’      

assumed that the entire army of stars revolved around the spectator, tried whether it might not work better if he let the observer revolve and left the stars in peace. (Bxvi)

This revolution in our way of thinking is undertaken, at least in part, in response to a problem in epistemology and the philosophy of mind, namely, to explain how we can cognize objects, and, in particular, how we can cognize objects a priori. The ‘turn’ also has profound epistemological consequences, including, eventually, the result that all of our substantive theoretical cognition (of a certain kind) is restricted to objects of possible experience. And it crucially involves a change in our conception of how concepts intentionally relate to, or apply to objects. But the most important ingredient in the revolution is the move to a new ontological position, namely, from transcendental realism to transcendental idealism. The new ontological position underwrites the new view of how concepts intentionally relate to objects and first allows us to account for the possibility of a priori cognition. This ontological dimension of the Copernican Turn is illustrated nicely in the following note: If the objects that are given to us were things in themselves and not mere appearances we would not have any cognition of them a priori at all. For if we take it from the objects the cognition would be empirical and not a priori; but if we want to make concepts of those objects for ourselves independently of them, they would have no relation at all to any object; thus, they would be concepts without content. From this one sees that they [the objects that are given to us] must be appearances. (LB B12, 1:116, 23:20)⁴

Before the revolution, empirical objects were assumed to be things in themselves, that is, they were assumed to be mind-independent, and concepts were understood to intentionally relate to objects only if they are (partly) caused by them. On this view, it is unintelligible how we could cognize objects a priori. After the revolution, empirical objects are recognized to be appearances, that is, they are recognized to be fully minddependent and to be conditioned by the forms of our cognitive faculties. On this view, since our a priori concepts are partly constitutive of objects, it is intelligible how they can a priori relate to objects and how we can cognize objects a priori. As explicated in section 2.5.2, we can a priori cognize those features of objects that are reflections of the formal conditions of experience to which any representation must conform that represents a proper object for us. That Kant’s idealism about empirical objects allows him to solve the vexing problem of how synthetic a priori cognitions are possible, that is, substantive

⁴ See A128–129: “If the objects, with which our cognition has to do, were things in themselves we could not have any concepts of them a priori at all. For from where should we take them? If we take them from the object (without examining for now how this object could become known to us) our concepts would be merely empirical and not concepts a priori. If we take them from within ourselves, in that case that which is merely in us cannot determine the quality of an object that is distinct from our representations, i.e., it cannot be a reason why there should be a thing to which something like what we have in our thoughts pertained, rather than for all this representation to be empty. On the other hand, if we are dealing everywhere only with appearances, it is not only possible but also necessary that certain concepts a priori precede the empirical cognition of objects. For as appearances they make up an object that is merely in us, since a mere modification of our sensibility is not encountered outside us at all.” See Prol, 4:375, note: “Proper idealism always has a delusional purpose and also cannot have any other purpose; but my idealism is only there to make intelligible the possibility of our cognition a priori of objects of experience, which is a problem that, so far, has not been solved, in fact, has not even been raised.” Also see B124 127/A92 94; Letter to Herz, February 21, 1772, 10:130–132.

        

113

cognitions that, despite being a priori, intentionally relate to, or are about, objects, is another important argument in its favor, in addition to the indirect argument based on the antinomies.⁵ The central and prominent place in Kant’s ontological theory of the thesis that empirical objects are appearances is the reason why, of all the names he could have invented for it, he chose to call it ‘transcendental/critical idealism.’ If there were nothing idealist about Kant’s philosophy, as some commentators have argued,⁶ this choice would not only be misleading but quite bizarre, given that the label ‘idealist’ was as little a badge of honor then as it is now. This is why Kant is so concerned to distinguish his own version of idealism from the more familiar version exemplified by the likes of Berkeley. It is true that Kant had second thoughts about the name ‘transcendental idealism,’ but he only worried about the appropriateness of the qualification ‘transcendental,’ which, upon further reflection, he wanted to replace with ‘critical’ or ‘formal.’⁷ He never wavered on calling himself an idealist. It is crucial to appreciate that by claiming that empirical objects are appearances and not things in themselves, Kant is endorsing a genuine kind of idealism about empirical objects, that is, a position according to which empirical objects are fully mind-dependent, that is, mind-dependent with respect to their existence and all of their determinations and other ontological ingredients. It is not just our concept of an object or certain aspects of empirical objects or things considered in certain ways or things qua bearers of certain kinds of properties, of which Kant claims that they are mind-dependent. His claim is that empirical objects—as they are in themselves, if you will—are appearances and not things in themselves. In addition to formulations such as that “the objects of the senses exist only in experience” (Prol, 4:342), or that “the objects of our senses are not things in themselves but merely appearances, i.e., representations . . . ” (R5636, 18:268)—of which opponents of the reading that Kant is a genuine idealist might claim that they must be understood as referring to objects as they appear to us in our sensible representations—Kant also frequently makes the same point by simply talking about bodies or matter without adding any further qualifications. Since, according to what has been shown by now, the senses never and in no single part reveal things in themselves but merely their appearances, but the latter are mere representations of sensibility, ‘thus all bodies together with the space in which they are located must be regarded as nothing but mere representations in us and exist nowhere else than merely in our thoughts’. (Prol, 4:288) In the transcendental Aesthetic we have proved indisputably that bodies are mere appearances of our outer sense and not things in themselves. (A357) Accordingly, there then remains no other way out than to admit that bodies are not at all things in themselves and that their sensible representation to which we affix the name ⁵ See KU, 5:197: “The understanding, through the possibility of its a priori laws for nature, gives proof that nature is cognized by us only as appearance, and thus at the same time gives indication of a supersensible substrate of nature but leaves it completely undetermined.” ⁶ See Collins 1999, ix: “These words [transcendental idealism] naturally suggest that Kant is presenting some form of idealism, but in fact he is not.” In this context, also see Baum 1986, 21f.; Abela 2002; Baum 2011. ⁷ See Prol, 4:293–294, 337, 375. A possible reason for Kant’s later preference of the label ‘critical idealism’ will be offered in section 5.1.

114

 ’      

body is nothing but the appearance of something, which as thing in itself alone can contain the simple but remains completely uncognizable for us . . . (ÜE, 8:209)⁸

These passages clearly show that, on Kant’s view, it is not just some properties or aspects of bodies that are mind-dependent; rather, bodies are appearances and, thus, fully minddependent, without any mind-independent ontological ingredients whatsoever. And empirical objects obviously are bodies.

3.2 How to Simultaneously be an Idealist and a Realist Who Is Opposed to Berkeley It may be objected and has been objected that reading Kant as a genuine idealist about empirical objects is incompatible with both his commitment to empirical realism and his repeated protestations that his idealism is very different from Berkeley’s.⁹ That Kant is a transcendental idealist, so the objection goes, must mean, among other things, that he is not a genuine idealist about empirical objects, in contrast to Berkeley. Apart from the complaint that the classic two-world view is preposterous and the misguided worry that it shipwrecks over ascribing the view to Kant that things in themselves exist and affect us, which we will address in section 5.10.1, all other arguments against it in the literature (known to me) that are at all to be taken seriously boil down to variations of the objection just described. In response, I concede that, from a contemporary point of view, it is natural to think that being an empirical realist is incompatible with being a genuine idealist about empirical objects. For empirical realism nowadays is often understood as the view that there are (in the sense of exist) empirical objects, and empirical objects are mind-independent.¹⁰ Similarly, I grant that, since the full mind-dependence of empirical objects is the most salient feature of Berkeley’s idealism, absent any further information it is natural think that Kant’s rejection of Berkeley’s idealism must mean that Kant himself does not take empirical objects to be fully mind-dependent. However, if we want to understand what Kant means by his declaration for empirical realism and against Berkeley, we will do well to resist the temptation to impute our own preconceptions to him and take a closer look at how he himself conceives of empirical realism and its relation to transcendental idealism and how he himself characterizes the difference between his idealism and the idealism of Berkeley. This is the project for the remainder of the present chapter—although our inquiry into the differences between Kant and Berkeley will continue in section 5.6. I will argue that being an empirical realist about empirical objects is perfectly compatible with being a genuine idealist about them at the same time, and that there is plenty of room for disagreement between Kant and Berkeley, even though both hold that empirical objects are fully mind-dependent. Moreover, once we have sufficiently fleshed out Kant’s empirical realist position, we will also find that it is actually the two-aspect view that cannot properly accommodate it. I will sketch the basic proposal for how to read Kant as both a genuine

⁸ Also see B150; A358 359; A371–372; A383; A387. ⁹ See Robinson 1994, 416; Abela 2002; Allais 2004, 660–5; Allais 2007, esp. 476–7; Allais 2015, esp. 43 56. ¹⁰ See, for example, Bill Brewer’s empirical realism, as articulated in Brewer 2011, according to which the direct objects of our perceptual experience are mind-independent physical objects.

        

115

idealist and a genuine realist and opponent of Berkeley in the present section and flesh out the proposal in the sections to follow. Before proceeding to these tasks, another brief terminological observation is in order. The label ‘transcendental idealism’ can be used and is used in the literature in several different senses of varying degrees of generality. The most narrow sense is the sense in which I have mainly used the term so far, in which it refers to a couple of specific ontological theses within Kant’s theoretical philosophy, one of which is the thesis that empirical objects are appearances. Transcendental idealism in this sense is opposed to transcendental realism and complemented by empirical realism in a way to be explicated presently. The label ‘transcendental idealism’ can also be used more generally to stand for the conjunction of Kant’s transcendental idealism in the narrow sense and his empirical realism plus a few other theses, both ontological and transcendental-psychological, so to speak, that Kant himself does not mention in his explicit characterizations of transcendental idealism in the narrow sense and of empirical realism but that still play an important role in defining his ontological position overall. As anticipated in section 1.1, these are the two main uses of the term by Kant himself. In the literature, ‘transcendental idealism’ is also sometimes employed, unhelpfully, as a cover term for Kant’s entire theoretical philosophy as presented in the Critique, which does not only comprise transcendental idealism in the sense just explicated but also several important epistemological results, as for instance, the claim that all of our substantive theoretical cognition (of a certain kind) is restricted to the objects of possible experience or even as a cover term for Kant’s critical philosophy as a whole, including his theoretical and practical philosophy as well as his aesthetics. Right now, we are mainly interested in transcendental idealism in the first two senses. In the service of conceptual clarity, I will continue to use ‘transcendental idealism’ exclusively in the first, narrow sense and will follow Kant’s example in employing ‘critical idealism’ to refer to his ontological position overall, which comprises transcendental idealism, empirical realism, and a cluster of several other claims that further characterize the particular flavor of his idealism. We will focus on transcendental idealism and empirical realism for now and take a closer look at the other claims that are comprised in critical idealism in chapter 5. The key to understanding how Kant can be both a genuine idealist and a genuine realist about empirical objects is to read him as committed to a tiered conception of reality, just like Leibniz. As already outlined in section 2.2, on my view, Kant conceives of reality as comprising at least two different ontological levels that are home to genuine existents, the transcendental level and the empirical level.¹¹ From (what I have dubbed) the point of view of fundamental ontology, that is, a meta-point of view from which we can investigate the nature and structure of reality in general, the transcendental level can be described as a level of mind-independent reality, or “absolute reality,” as Kant also often says,¹² while the empirical level can be described as a level of mind-dependent reality. The qualification that

¹¹ As also already indicated in section 2.2, I find it useful to think of Kant’s ontological scheme as three-tiered, comprising yet another level at which other fully mind-dependent objects such as fictional characters and the objects of dreams and hallucinations exist, which, although not genuine existents, may be judged to have some moderate degree of reality (in the sense of realitas, or Sachgehalt) and thus should have a level of reality at which they exist. But since there is no need to appeal to this putative third level in order to characterize transcendental idealism and empirical realism, we will set it aside for the following discussion. ¹² See B52/A35; B53/A36; B54/A37; B55/A38; B56/A39; B564/A536; LB D9, 1:207–208, 23:81.

116

 ’      

Kant is a transcendental idealist is not meant to indicate that he is not a genuine idealist; rather, it is meant to indicate with respect to which level of reality he is an idealist about the objects in question. More specifically, ‘ideal’ and ‘real’ are terms that are relative to a level of reality, and the modifiers ‘transcendentally’ and ‘empirically’ indicate which level is relevant. ‘Ideal’ and ‘real’ thus have both a transcendental and an empirical sense; being transcendentally ideal/real and being empirically ideal/real are two different kinds of being ideal/real. That Kant is a transcendental idealist about empirical objects, or that he regards empirical objects as transcendentally ideal, means that, for Kant, empirical objects are ideal with respect to the transcendental level of reality, which, in turn, means, that, from the point of view of fundamental ontology, they are fully mind-dependent and do not exist at the transcendental level. Kant’s transcendental idealism thus fully deserves its name; it is a genuine kind of idealism. That Kant is an empirical realist about empirical objects, or that he regards empirical objects as empirically real, means that, for Kant, empirical objects are real with respect to the empirical level of reality, which, in turn means, that, from the point of view of fundamental ontology, they are mind-independent with respect to the empirical level and exist (namely, at the empirical level). In contrast to other fully mind-dependent objects such as the intentional objects of fictions, dreams, illusions, and hallucinations, empirical objects conform to Kant’s formal conditions of proper objecthood (as discussed in section 2.5.2) and are properly directly grounded in things in themselves (as briefly described in section 2.5.2 and as will be discussed in more detail in chapter 5), which gives them enough reality to count as existents from the point of view of fundamental ontology, although from this fundamental perspective their existence is mind-dependent. Kant’s empirical realism thus fully deserves its name; it is a genuine kind of realism. And since his idealism and realism about empirical objects are relative to different levels of reality, Kant can be both a genuine idealist and a genuine realist at the same time. Moreover, the problem with Berkeley’s theory is, not that it classifies empirical objects as fully minddependent, but that it, in effect, classifies them as having the same ontological status as the intentional objects of fictions, dreams, illusions, and hallucinations in Kant’s scheme. That is, the problem is that Berkeley is not only a transcendental but also an empirical idealist about empirical objects. It is useful to note that ‘ideal’ and ‘real’ and their cognates such as ‘idealism/realism’ and ‘ideality/reality’ are not the only concepts for which Kant distinguishes a transcendental and an empirical sense. On the reading proposed here, this is not surprising. Given the explicated stratification of reality into different ontological levels, which are distinguished by the manner of being of the entities that exist at them, it is to be expected that the proper description of the properties of an object that characterize its ontological status and ‘location’ in Kant’s ontological scheme—call them ‘ontological properties’ for short from now on—must vary depending on the point of view from which the description is formulated. More specifically, if we adopt the point of view of fundamental ontology, the proper description of an ontological property must include a specification of the level of reality with respect to which the description is to be evaluated. Another important example of concepts that refer to such ontological properties is the pair ‘appearance’ and ‘thing in itself.’ Empirical objects are things in themselves in the empirical sense, that is, things in themselves with respect to the empirical level of reality, while the intentional objects of perceptions, illusions, dreams, fictions, and hallucinations, are appearances in the empirical sense, that is, appearances with respect to the empirical level of reality. But empirical

        

117

objects are not things in themselves in the transcendental sense, that is, they are not things in themselves with respect to the transcendental level of reality; empirical objects are appearances in the transcendental sense, that is, appearances with respect to the transcendental level of reality. For example, a rainbow, understood as a band of colors, is an empirical appearance of a collection of water droplets; the collection of water droplets is an empirical thing in itself but a transcendental appearance of some mind-independent thing (or things) X that grounds it; and X is a transcendental thing in itself.¹³ The terms ‘in/inside us’ and ‘outside us,’ or ‘inner’ and ‘outer,’ are yet other crucial examples of terms that are ambiguous between an empirical and a transcendental meaning. Kant explicitly flags this case at A373, saying that “the expression outside us carries with it an inevitable ambiguity in that it sometimes signifies something that exists distinct from us as thing in itself, and sometimes something that belongs merely to outer appearance” (A373), which he proposes to remedy by “distinguishing empirically outer objects from those that may be called outer in the transcendental sense by directly referring to the former as objects that are encountered in space.”¹⁴ (Now that we know that there is a transcendental and an empirical version of the distinction between things in themselves and appearances, there is potential for confusion. I will continue to employ ‘thing in itself ’ and ‘appearance’ without further qualifications in their transcendental sense; if I intend the empirical sense, I will use ‘empirical thing in itself ’ and ‘empirical appearance.’) In order to prevent confusion, I want to emphasize that, despite certain surface similarities, my interpretation of the distinction between transcendental idealism and empirical realism as grounded in the distinction between a transcendental and empirical level of reality is a very different animal from the two-aspect interpretation of Kant’s transcendental distinction between things in themselves and appearances, according to which it amounts to a distinction between two ways of considering the same things, namely, as they are in themselves and as they appear to us. To acknowledge the similarities, my reading allows us to distinguish between different perspectives from which we can consider reality: the empirical point of view from which we look at the empirical level of reality, the transcendental point of view from which we look at the transcendental level, and the point of view of fundamental ontology from which we look at reality in general. It is fair to say on my reading that if we adopt the transcendental point of view we are concerned with reality as it is independently from how we represent it, while if we adopt the empirical point of view we are, in effect, concerned with reality as it is represented by us. It is also fair to say on my reading that we can consider the very same things, namely, ¹³ See B62 63/A45 46; FM, 20:269. The empirical version of the distinction between things in themselves and appearances will be discussed in detail in chapter 5, see esp. sections 5.3 and 5.4. In this context, it is also worth mentioning that I take the expression ‘thing in the appearance,’ which Kant also uses, often in contrast with ‘thing in itself,’ to be an amalgamation of ‘thing (in itself) in the empirical sense’ and ‘appearance in the transcendental sense’ and as meant to signify entities that can be described in both of these ways, that is, empirical objects. See B332–333/A276–277: “What the things in themselves may be I do not know but also do not need to know, since I can never encounter a thing other than in the appearance.” See Prol, 4:341: “When I talk about objects in time and space, I do not talk about things in themselves because I do not know anything about them but only about things in the appearance, i.e., of experience as a special manner of cognizing objects, with which humans alone are favored.” ¹⁴ Also see A385: “Thus there may well be something outside us to which this appearance corresponds that we call matter; but in this very quality as appearance it is not outside us but merely in us as a thought, even though this thought represents it as located outside us through the indicated sense [outer sense].” Also see A370; Prol, 4:336 337. The term ‘dualism’ is another example of a term that has an empirical and a transcendental sense on account of involving characterizations of the ontological status of objects; see A379.

118

 ’      

empirical objects, from different points of view. From the point of view of fundamental ontology, they are fully mind-dependent; from the point of view of the empirical level of reality, they are mind-independent. But this is where the similarities end. On my interpretation, the transcendental distinction between appearances and things in themselves is a distinction within a perspective, namely, the transcendental perspective, or, what comes down to the same thing in this case, the perspective of fundamental ontology. But according to the two-aspect view, the transcendental distinction maps onto the distinction between the two perspectives from which we can consider empirical objects. Furthermore, on my view, empirical objects considered from the transcendental point of view are appearances, while empirical objects considered from the empirical point of view are things in themselves. By contrast, on the two aspect-view, empirical objects considered as they are independently from how we represent them are things in themselves, while empirical objects considered as they are (sensibly) represented by us are appearances. With respect to the methodological two-aspect view, there is also the further difference that my distinction between the transcendental and the empirical perspectives is grounded in an ontological distinction between two different levels of reality. As I see it, in order to arrive at a proper understanding of Kant’s critical idealism it is important to appreciate not only his transcendental distinction between things in themselves and appearances but also the distinction between the transcendental and the empirical level of reality—which is orthogonal to the transcendental distinction—as well as the corresponding distinction between different points of view from which we can investigate reality. In section 1.3, I announced that most passages in which Kant talks about different ways of considering things or otherwise uses some kind of two-aspect language, and thus may be taken to articulate a two-aspect view of the transcendental distinction between appearances and things in themselves, are perfectly compatible with a two-world understanding of this distinction as long as the passages are properly understood. The considerations in this section put us in the position to explicate three batches of passages of this kind, with more to follow as we go along. First, in many passages in which Kant uses the formulation that objects can be considered as appearances or as things in themselves he is talking about different ways in which empirical objects are considered from the point of view of competing philosophical theories. From Kant’s own transcendental idealist point of view, empirical objects are considered as appearances, that is, as fully mind-dependent, while from the transcendental realist point of view, they are considered as things in themselves, that is, as mind-independent.¹⁵ Second, we sometimes find Kant distinguishing between the consideration of an empirical object from the empirical point of view and the consideration of the ¹⁵ See B251–252/A206: “ . . . if I consider all things not as phenomena but as things in themselves, and as objects of the mere understanding, they can be regarded as if dependent for their existence on foreign causes; but this would bring with it entirely different meanings of the words and would not apply to appearances, as possible objects of experience.” See Prol, 4:344: “If the objects of the senses were regarded as things in themselves and the above stated laws of nature as laws of things in themselves, the contradiction would be inevitable. Similarly, if the subject of freedom were imagined as mere appearance, like the other objects, the contradiction could not be avoided. . . . But if natural necessity is applied merely to appearances and freedom merely to things in themselves, no contradiction arises . . . ” See Letter to Garve, August 7, 1783, 10:341note: “It [the key to the solution of the antinomies] consists in that one can take all objects that are given to us according to two concepts, as appearances, on the one hand, and as things in themselves, on the other hand.” Also see in this sense Bxxvii; Bxviii–xix note; Bxx; B56/A39. Also note that, far from supporting the two-aspect reading, the formulation that objects can be considered as appearances or as things in themselves turns out to be a rather awkward affair for it, especially for its methodological variant. On this reading, considering an object as an appearance would amount to considering the object as a thing considered as it appears to us, and considering an object as a thing in itself would amount to considering the object as a thing considered as it is in itself. Prauss acknowledges and discusses this difficulty, see Prauss 1974, 33–7.

     

119

same object from the transcendental point of view or the point of view of fundamental reality. From the point of view of the empirical level of reality, empirical objects are things, or things in themselves, while from the point of view of the transcendental level, or from the point of view of fundamental ontology, empirical objects are appearances.¹⁶ Relatedly, third, Kant sometimes contrasts different senses in which the term ‘object’ or ‘thing’ can be used, namely, in an empirical and a transcendental sense.¹⁷ Things in the empirical sense, or with respect to the empirical level of reality, are appearances (in the transcendental sense of ‘appearance’), or empirical objects; things in the transcendental sense, or things with respect to the transcendental level of reality, are things in themselves (in the transcendental sense of ‘things in themselves’).¹⁸ In these passages, Kant is not saying that the same thing can be considered in two different ways, as it appears to us and as it is in itself. Rather, he is saying that the terms ‘thing’ or ‘object’ can be understood in two different ways, namely, as referring to appearances or as referring to things in themselves, depending on which level of reality we are talking about. The foregoing remarks provide an initial sketch of how the relation between Kant’s transcendental idealism and his empirical realism and his critique of Berkeley are supposed to be understood. In the following sections, this sketch will be fleshed out by way of an explication of the various specific theses that, on my reading, define transcendental idealism, transcendental realism, empirical realism, and empirical idealism, respectively.¹⁹ We will begin by considering the core theses of these positions that concern the ontological status of empirical objects as well as space and time. The corresponding theses about the self will be briefly addressed in the final section of this chapter. Additional theses of critical idealism will be identified in chapter 5, where we will also further flesh out the differences between Kant’s critical idealism and the ‘ordinary’ idealism of Berkeley.

3.3 Core Theses of Transcendental Idealism and Empirical Realism I: the Ontological Status of Empirical Objects 3.3.1 Empirical Objects in Transcendental/Empirical Idealism/Realism Please note that, unless stated otherwise, all of the following theses and related claims and definitions that characterize transcendental/empirical idealism/realism on my reading are

¹⁶ Recall B235 236/A190 191: “But now as soon as I bring up my concept of an object to transcendental significance, the house is not a thing in itself but merely an appearance, i.e., representation whose transcendental object is unknown; thus, what do I mean by the question how the manifold may be connected in the appearance itself (which after all is nothing in itself)? Here that which lies in the successive apprehension is regarded as representation but the appearance, even though it is nothing but the sum total of these representations, is regarded as their object with which my concept is supposed to cohere that I draw from the representations of apprehension.” ¹⁷ Obviously, in this context ‘object’ does not mean ‘proper object’ but must be understood a bit more loosely. ¹⁸ See Bxxvii: “ . . . it [the Critique] teaches to take the object in two senses, namely, as appearance or as thing in itself.” See A104: “Here it is necessary to clarify to oneself what one means by the expression of an object of a representation. We said above: that appearances themselves are nothing but sensible representations, which in themselves, in this very manner, must not be regarded as objects (outside of the faculty of representations).” See B507/A479: “Since we are here talking merely about a thing as object of a possible experience and not as a thing in itself . . . ” See Letter to Beck, 11:314: “ . . . I wanted to show that these objects of possible experience as objects of the senses reveal objects, not as things in themselves but only as appearances . . . ” Also see B43/A27; Bxxv–xxvi; B307. ¹⁹ For a very different, ‘non-ontological’ account of Kant’s conception of transcendental idealism, transcendental realism, empirical realism, and the relation between them, see Allison 2004, 20–49; and Allison 2006.

120

 ’      

to be understood as articulated from the point of view of fundamental ontology, even though, for expediency’s sake, I will no longer add an explicit clause to that effect every time; this also holds for the footnotes. The first thesis of transcendental idealism is the thesis that we just discussed in section 3.1, namely, that empirical objects are not things in themselves but appearances. For clarity’s sake, it is useful to distinguish a weaker, general version and a stronger, specific version of this thesis, which differ with respect to how much is built into the notion of an appearance. As noted before, ‘appearance’ in the specific sense refers to Kantian appearances, that is, appearances as conceived in Kant’s critical idealism, which essentially are intentional objects of experience as described in Kant’s theory of experience. But the term can also be taken in a general sense, where it means ‘something that is fully mind-dependent.’ As I read Kant, when he asserts that empirical objects are appearances in the context of distinguishing his position from transcendental realism he usually has the general sense in mind. But when he presents the details of his own view, he is typically relying on the specific sense. Accordingly, there are two versions of the first thesis of transcendental idealism, depending on how ‘appearance’ is understood. These theses can be formulated as follows.²⁰ (TI1) (a) Empirical objects are not things in themselves in the transcendental sense, or with respect to the transcendental level of reality, that is, they are not mind-independent; rather, (b), empirical objects are appearances in the transcendental sense, that is, they are fully mind-dependent.²¹ (TI1-strong) (a) Empirical objects are not things in themselves in the transcendental sense, that is, they are not mind-independent; rather, (b), empirical objects are Kantian outer appearances, i.e., they essentially are intentional objects of outer experience as characterized in Kant’s theory of experience, and, thus, fully mind-dependent. The relevant senses of mind-independence and full mind-dependence in these theses (and the theses to follow) are the senses defined in section 2.3, with the further specification that in (b) we are talking about a dependence on not just any finite minds but on human minds, as before, and that the relevant representations that ontologically specify appearances are actual (as opposed to merely possible) representations. Also note that, since the transcendental level of reality is a level of mind-independent reality, ‘x is mind-dependent’ when asserted from the point of view of fundamental ontology is equivalent to ‘x is minddependent in the transcendental sense, or with respect to the transcendental level.’²² That empirical objects are fully mind-dependent can also be expressed, metaphorically, by ²⁰ As also already noted, the term ‘thing in itself ’ can also be understood in a general and a specific sense, as referring to mind-independent things or as referring to Kantian things in themselves, that is, mind-independent things as conceived in Kant’s critical idealism. But since the claim that empirical objects are not things in themselves in the general sense entails that they are not things in themselves in the specific Kantian sense, there is no need to add a special thesis to this effect to our list of theses of transcendental idealism. By contrast, the claim that empirical objects are appearances in the general sense does not entail that they are appearances in the specific Kantian sense, which is why it makes sense to include the latter claim in our list. ²¹ I am using the formulation that empirical objects are not things in themselves but appearances to mirror Kant’s own preferred formulation. But, of course, since being an appearance entails not being a thing in itself, the thesis could also be expressed more briefly by saying merely that empirical objects are appearances. ²² The latter, in turn, is equivalent to ‘x is mind-dependent from the transcendental point of view,’ that is, from a point of view located at the transcendental level from which one can ‘see’ only the transcendental level.

     

121

means of Kant’s often employed formulation that they and all of their determinations and other ontological ingredients exist only ‘in us,’ or ‘in our representations,’ in the transcendental sense,²³ or, more concisely, by saying that empirical objects are transcendentally ideal.²⁴ So, an alternative, crisper formulation of TI1 is this: (TI1-concise) Empirical objects are transcendentally ideal, that is, they are ideal in the transcendental sense, or with respect to the transcendental level of reality. As anticipated in section 2.2, it is a largely terminological question whether we want to say that empirical objects exist at the transcendental level of reality, albeit only in a minddependent way. For the reasons outlined there, I prefer to say that they do not exist at the transcendental level, not even in a mind-dependent way—only mind-independent entities can exist at a mind-independent level of reality. Also note that although TI1-strong-b is ostensibly only about empirical objects, it implicitly also tells us something about their relation to entities at the transcendental level. For, as will be discussed in more detail in chapter 5, Kantian outer appearances are essentially grounded in things in themselves, and so by characterizing empirical objects as Kantian outer appearances TI1-strong-b implies that empirical objects are grounded in transcendentally real things. Opposed to TI1 is the first thesis of transcendental realism, TR1. (TR1) (i) There exist empirical objects; and (ii) empirical objects are (what a Kantian would call) things in themselves in the transcendental sense, that is, they are mindindependent.²⁵ My reason for explicitly including (i) in the first thesis of transcendental realism is that a genuine realist position arguably must incorporate some kind of existence claim, and (ii) by itself does not amount to one. And my reason for the choice of the slightly awkward formulation ‘there exist’ instead of ‘there are’ is to make explicit that we are talking about a genuine existence claim―and not just a ‘being claim,’ as one might put it―that is, a claim that ascribes existence to certain entities from the point of view of fundamental ontology. The claim that empirical objects are mind-independent can also be expressed by saying that, if they exist, empirical objects and all of their determinations and other ontological ingredients exist ‘outside our representations’ or ‘outside us’ in the transcendental sense. Following Kant’s previously cited characterization of the transcendental sense of ‘outside us’ as signifying “something that exists distinct from us as thing in itself” (A373), we might also define an additional, stronger sense of ‘transcendentally outside us.’ Taking ‘distinct from us’ at face value as meaning ‘distinct from all of us,’ we will say that a thing is strongly ‘outside us’ in the transcendental sense if, and only if, it is mind-independent and distinct from all human minds.²⁶ Using this terminology, we can describe the view of a

²³ See the quotations collected in section 2.1. ²⁴ Recall B535/A507: “It follows from this that appearances in general are nothing outside of our representations, which is precisely what we wanted to signify with their transcendental ideality.” Also see B518–519/ A490–491; A370. ²⁵ Remember that since the transcendental level is a level of mind-independent reality, from the point of view of fundamental ontology, ‘x is transcendentally mind-independent’ is equivalent to ‘x is mind-independent.’ ²⁶ We will countenance yet another, slightly weaker reading of ‘outside us’ in section 5.9.1.

122

 ’      

transcendental realist who is also a dualist—that is, who believes that there are two fundamental kinds of existents in the world, namely, minds and bodies—by saying that he regards empirical objects to be (what a Kantian who is aware of our terminology would call) strongly ‘outside us’ in the transcendental sense. Taken together, TR1-i and TR1-ii entail that the ‘place’ in reality where empirical objects exist is the transcendental level. Spelling out ‘is real’ as ‘exists and is mind-independent’ and, accordingly, ‘is transcendentally real’ as ‘exists at the transcendental level and is transcendentally mind-independent,’ TR1 can also be formulated more economically as follows: (TR1-concise) Empirical objects are (what a Kantian would call) transcendentally real, that is, they are real in the transcendental sense, or with respect to the transcendental level of reality, and, as such, genuinely exist.²⁷ The qualifications in brackets that according to the transcendental realist position empirical objects are what a Kantian would call things in themselves in the transcendental sense, or transcendentally real, are added to acknowledge that a transcendental realist himself would not describe his position in this way, even though he is, in fact, committed to the view thus described. Since a typical transcendental realist does not allow different levels of reality in his ontology but considers what, on Kant’s scheme, is the transcendental level as exhausting the realm of all genuinely existing entities, there is no need for him to relativize the ontological terms in question to a specific level of reality. He would describe his view by saying simply that there are (in the sense of exist) empirical objects, and empirical objects are mind-independent. But in order to make it easier to discern the logical relations between TR1 and TI1, and TR1 and ER1 (to be discussed presently), it is useful to have a Kantian-style formulation of TR1 on the books, since TI1 and ER1 can only be stated in Kantian terms. And lest there be any doubt about it, let me also note once more that Kant endorses TI1 in all of its variants, and rejects TR1 in all of its variants. TI1 is complemented by ER1, the first thesis of empirical realism. (ER1) (i) There exist empirical objects; and (ii) empirical objects are things in themselves in the empirical sense, or with respect to the empirical level of reality, that is, they are mind-independent in the empirical sense, or with respect to the empirical level of reality. Kant’s theory of experience provides him with the resources to draw a meaningful distinction between empirical objects, which essentially are intentional objects of experience, on the one hand, and the intentional objects of all of our other representations, including fictions, dreams, hallucinations, and illusions, on the other hand. The intentional objects of experience are the only intentional objects that both conform to Kant’s formal conditions of proper objecthood that correspond to the formal conditions of experience and objectivity and are grounded in things in themselves in a properly direct way so that they can be said to exist from the point of view of fundamental ontology. Another way to

²⁷ Note that, as I understand Kant’s terminology, ‘x is transcendentally ideal’ is not equivalent to ‘x is not transcendentally real.’ If x is transcendentally ideal it follows that x is not transcendentally real; but being transcendentally ideal is not the only way in which x can fail to be transcendentally real; simply not existing at the transcendental level suffices as well.

     

123

formulate the claim that empirical objects are empirically mind-independent is to say that, if they exist, empirical objects and all of their determinations and other ontological ingredients exist ‘outside our representations’ or ‘outside us’ in the empirical sense. In fact, Kant arguably is also committed to the stronger view that empirical objects are a strongly ‘outside us’ in the empirical sense, that is, that they are empirically mindindependent and distinct from all human minds.²⁸ The transcendental realist’s basic error, that is, his confusion of empirical objects with things in themselves in the transcendental sense, can thus also be described by saying that he mistakenly regards empirical objects as ‘outside us’ or as strongly ‘outside us’ in the transcendental sense, while in truth they are only strongly ‘outside us’ in the empirical sense. The transcendental realist thus regards outer appearances (if one admits their reality) as things in themselves, which exist independently of us and sensibility, and thus also would be outside us according to pure concepts of the understanding. (A369)

Taken together, ER1-i and ER1-ii entail that the ‘place’ in reality where empirical objects exist is the empirical level. Again, taking ‘is real’ to mean ‘exists and is mind-independent’ and, accordingly, ‘is empirically real’ to mean ‘exists at the empirical level and is empirically mind-independent,’ ER1 can be expressed more briefly as follows: (ER1-concise) Empirical objects are empirically real, that is, they are real in the empirical sense, or with respect to the empirical level, and, as such, genuinely exist.²⁹ Given that the empirical level is a mind-dependent level of reality, the claim that empirical objects are empirically mind-independent might look suspicious at first glance. What exactly could it mean for an object to be mind-independent with respect to the empirical level? One strategy to answer this question would be to make use of a switch in perspective and say that an entity E is empirically mind-independent from the point of view of fundamental ontology if, and only if, E is mind-independent from the empirical point of view, that is, from a point of view at the empirical level from which we can ‘see’ only the empirical level. ‘Is mind-independent’ on the right-hand side of this definition is then to be understood along the usual lines, that is, roughly in the sense of ‘neither exists in virtue of being represented by a finite mind nor has any of its determinations or other ontological ingredients in virtue of being represented by a finite mind.’ Another strategy, which is my preferred strategy, would be to stick to our customary meta-perspective, that is, to the point of view of fundamental ontology, and make use of the fact that experience determines what is the case at the empirical level. On this basis, we can define that an entity E is empirically mind-independent if, and only if, experience presents E as mind-independent. And, as noted in section 2.5.2, the way in which outer experience presents empirical objects as

²⁸ As will be discussed in section 3.6, Kant is a dualist with respect to the empirical level of reality and holds that empirical objects, which are material, are distinct from empirical selves, which are immaterial. ²⁹ As in the transcendental case, on my understanding of Kant’s terminology, ‘x is empirically ideal’ is not equivalent to ‘x is not empirically real.’ If x is empirically ideal it follows that x is not empirically real; but being empirically ideal is not the only way in which x can fail to be empirically real; simply not existing at the empirical level suffices as well. For example, things in themselves (in the transcendental sense) are neither empirically real nor empirically ideal.

124

 ’      

mind-independent is by presenting them as being in space.³⁰ This fits well with Kant’s remark at A373 that in order not to get confused by the ambiguity of ‘outside us’ we could refer to objects that are empirically ‘outside us,’ that is, objects that are empirically mindindependent and distinct from us, as objects that can be encountered in space. Having defined what it means to be empirically mind-independent, it does not take much work to define what it means to be empirically mind-dependent. There is only one small complication to take note of, namely, that, since we are dealing with a level-relative notion of mind-independence, we cannot simply define ‘is empirically mind-dependent,’ or ‘is empirically ideal,’ as ‘is not empirically mind-independent.’ One way of not being empirically mind-independent, in the sense of either one of our two definitions, is to be mind-independent (which is equivalent to being transcendentally mind-independent). But, of course, we do not want to say that things in themselves are empirically minddependent. Accordingly, we will say that an entity E is empirically mind-dependent, or empirically ideal, if, and only if, E is fully mind-dependent and not empirically mindindependent, that is, it is fully mind-dependent and experience does not present it as mindindependent.³¹ Being empirically mind-dependent thus entails being transcendentally fully mind-dependent. This is as it should be. Empirical mind-dependence is supposed to be a stronger kind of mind-dependence than transcendental mind-dependence.³² ³⁰ We will briefly talk about how inner experience presents the empirical self as mind-independent in section 3.6. ³¹ Is it a fact at the empirical level that, say, the objects of hallucinations are (empirically) mind-dependent? This question is less straightforward than it might seem on first glance, but my answer is no. As I see it, the objects of hallucinations, illusions, dreams, and fictions are empirically mind-dependent because they are fully minddependent and because experience does not represent them at all, which also means that it does not present them as mind-independent. On that assumption, since experience determines the facts at the empirical level, and since experience does not present the objects of hallucinations as mind-dependent, it is not a fact at the empirical level that the objects of hallucinations are mind-dependent. That the objects of hallucinations are empirically minddependent is a fact from the point of view of fundamental ontology and at the transcendental level, a fact that we ascertain as transcendental philosophers by, among other things, examining experience and what it represents. ³² One complication: Berkeley, who is Kant’s ‘favorite’ empirical idealist, holds that at least some empirical objects, at least some of the time, are ontologically specified by representations of God. (This view is a central element of his explanation for the continued existence of empirical objects when no finite mind is looking at them. See Berkeley 1713, Works 2:230–231; also see Berkeley 1710, §6, Works 2:43, and §48, Works 2:62.) Since our definition of full mind-dependence from section 2.3 is formulated in terms of representations of finite minds, Berkeleyan empirical objects that are temporarily ontologically specified by representations of God are thus not fully mind-dependent, according to our definition, and hence not empirically mind-dependent either during their God-dependent periods of existence The reason why we went with a definition of full mind-dependence on which the relevant minds are finite minds is that things in themselves may well depend on God, and, more specifically, God’s representations, both for their initial creation and for their continued existence. More specifically, it might be that the way in which God creates and concurs with things in themselves is by representing them, in which case it would be true that things in themselves exist and have all of their determinations and other ontological ingredients in virtue of being represented by God. So, we cannot simply lift the restriction to finite minds in our definition of full mind-dependence in order to be able to accommodate Berkeleyan empirical objects during the periods of their existence when no finite mind is representing them because that would potentially let things in themselves in the door as well, which, however, are supposed to be mind-independent. The way out of the difficulty is to recognize that except for the different ‘representers’ there does not seem to be any difference in how Berkeley conceives of the dependence of empirical objects on God, on the one hand, and on human minds, on the other hand. Any of the empirical objects that exist and have their determinations at a certain time in virtue of being represented by God could also exist with the very same determinations at the same time in virtue of being represented by a human mind, at least in principle. But this is not true for things in themselves. If anybody has the special ability to create and concur with things in themselves by representing them then it is God, and God alone. So, we can modify our definition of full mind-dependence by saying that an entity E is fully mind-dependent* if, and only if, all of E’s determinations and other ontological ingredients are strongly mind-dependent*, where a property P/an ontological ingredient OI of an entity E is strongly mind-dependent* if, and only if, necessarily, if E has P/comprises OI, then there is a mind M, and E has P/comprises OI in virtue of the fact that M has a possibly complex representation R whose presentational content includes E having P/comprising OI, and there is a finite

     

125

Empirical realism is to be contrasted with empirical idealism, or, more precisely, with dogmatic empirical idealism.³³ Following our established scheme, the first thesis of empirical idealism can be formulated as follows. (EI1) (a) Empirical objects are not (what a Kantian would call) things in themselves in the empirical sense; rather (b) empirical objects are (what a Kantian would call) appearances in the empirical sense, that is, they are (what a Kantian would call) mind-dependent in the empirical sense. In other words, the empirical idealist is committed to the view that empirical objects exist only ‘in us’ or ‘in our representations’ in the empirical sense. Accordingly, they do not genuinely exist at all. More concisely: (EI1-concise) Empirical objects are (what a Kantian would call) empirically ideal, or ideal with respect to the empirical level of reality. As in the case of our first two formulations of TR1, these formulations of EI1 use Kantian terminology that the proponents of this thesis would not employ themselves. For just like the transcendental realist, the empirical idealist (as understood by Kant) does not believe in multiple levels of reality. On his view, there is only one realm of genuinely existing entities―which corresponds to the Kantian transcendental level of reality―and empirical objects are not part of it. The empirical idealist himself (as understood by Kant) would simply say that empirical objects are fully mind-dependent and have the same ontological status as the intentional objects of dreams, illusions, fictions, and hallucinations. But, again, in order to facilitate the comparison between EI1 and TI1, and EI1 and ER1, it is useful to present EI1 also by using Kantian terminology. And lest there be any doubt about it, Kant endorses ER1 and rejects EI1, in all of their variants.

3.3.2 Kant versus Berkeley, Part I In Kant’s eyes, Berkeley is an empirical idealist in the sense just described. To be sure, Berkeley himself does not share Kant’s assessment that, according to his philosophical theory, empirical objects do not genuinely exist and cannot be meaningfully distinguished mind who could have a representation whose presentational content includes E having P/comprising OI. This definition classifies Berkeleyan empirical objects as fully mind-dependent*, including empirical objects that are temporarily ontologically specified by God’s representations, but the final conjunct prevents things in themselves that depend on God’s representations from counting as fully mind-dependent*. So, a more careful characterization of empirical mind-dependence (than the one just offered in the main text) would be to say that an entity E is empirically mind-dependent, or empirically ideal, if, and only if, E is fully mind-dependent* and not empirically mind-independent. (Note that any entity that is fully mind-dependent* is also fully mind-dependent.) In the following, when I talk about empirical objects as conceived by Berkeley being fully mind-dependent, I should be understood as having full mind-dependence* in mind, and when I talk about these objects being empirically mind-dependent, I should be understood as having the indicated more careful notion of empirical minddependence in mind. ³³ Kant also recognizes a problematic kind of empirical idealism. While the dogmatic empirical idealist denies the existence of empirical objects (according to Kant’s reading), the problematic idealist merely doubts their existence. In the following, it should be understood that by ‘empirical idealism’ I mean ‘dogmatic empirical idealism.’

126

 ’      

from the intentional objects of dreams, fictions, or hallucinations. On the contrary, he presents it as one of the main advantages of his conception of empirical objects as complex sensible ideas, or as intentional objects of complex sensible ideas, especially in comparison with Locke’s conception, that it makes his view invulnerable to skeptical worries and allows him to assert, with certainty, that empirical objects exist.³⁴ He also explicitly addresses the objection that, on his view, there is no proper distinction between “real things, and chimeras formed by the imagination, or the visions of a dream, since they are all equally in the mind” (Berkeley 1713, Works 2:235). His reply is that there is such a distinction after all, namely, one that is grounded in the greater clarity and vividness of the ideas of the senses on which real objects depend compared to the ideas of the imagination or the visions of a dream, as well as on the involuntariness of the ideas of the senses and their coherence with our memories and subsequent sensible ideas, which are lacking in the case of imaginings and dreams.³⁵ Be it that he was not aware of the finer details of Berkeley’s theory, or that he did not care about the details, or that he judged the theory to be inadequate, Kant holds that, within the framework of Berkeley’s account, ultimately no good sense can be made of the claim that empirical objects are real.³⁶ According to Kant’s understanding of Berkeley, due to his impoverished account of experience, and in particular, due to the absence of any formal conditions of experience and objectivity, Berkeley does not have the wherewithal to ³⁴ See Berkeley 1713, Works 2:248–249: “HYLAS: But after all, Philonous, when I consider the substance of what you advance against scepticism, it amounts to no more than this. We are sure that we really see, hear, feel; in a word, that we are affected with sensible impressions. PHILONOUS: And how are we concerned any farther? I see this cherry, I feel it, I taste it: and I am sure nothing cannot be seen, or felt, or tasted: it is therefore real. . . . Hence, when I see, and feel, and taste, in sundry certain manners, I am sure the cherry exists, or is real; its reality being in my opinion nothing abstracted from those sensations. But if by the word cherry you mean an unknown nature distinct from all those sensible qualities, and by its existence something distinct from its being perceived; then indeed I own, neither you nor I, nor anyone else can be sure it exists.” Also see Berkeley 1713, Works 2:206, 211–212, 229–230, 258–259, 262. ³⁵ See Berkeley 1713, Works 2:235: “PHILONOUS: The ideas formed by the imagination are faint and indistinct; they have besides an entire dependence on the will. But the ideas perceived by sense, that is, real things, are more vivid and clear, and being imprinted on the mind by a spirit distinct from us, have not a like dependence on our will. There is therefore no danger of confounding these with the foregoing: and there is as little of confounding them with the visions of a dream, which are dim, irregular, and confused. And though they should happen to be never so lively and natural, yet by their not being connected, and of a piece with the preceding and subsequent transactions of our lives, they might easily be distinguished from realities. In short, by whatever method you distinguish things from chimeras on your own scheme, the same, it is evident, will hold also upon mine. For it must be, I presume, by some perceived difference, and I am not for depriving you of any one thing that you perceive.” ³⁶ How well did Kant know the ‘real’ Berkeley? As evidenced by Herder’s lecture notes, Kant appears to have had some exposure to the Siris in the early 1760s in some form or another, unless Herder added the relevant comment on his own. See V-Met/Herder, 28:42: “In that way, Bishop Berkeley doubted in the treatise on the benefit of tar-water for our body whether there are bodies at all.” It is noteworthy that Baumgarten’s Metaphysica, which Kant used as his textbook for the lecture in question, does not contain a reference to the Siris in the corresponding sections. See Baumgarten 1757, §§392–393, 213. We also know from Warda’s inventory of Kant’s books that he owned a copy of a German translation of the Three Dialogues. See Warda 1922, 46, entry #12. This edition was first published in 1781, that is, too late to have been studied by Kant before writing the A-edition of the Critique (which was published in 1781). But he may have read it before or while preparing the Prolegomena or the B-edition. On the other hand, from 1781 to 1787 Kant was extraordinarily busy. In addition to fulfilling his demanding teaching obligations (he lectured between nine and twelve hours and gave several repetitoria per week) and performing his administrative duties at the university (he was dean during the winter semester 1783–84, for example), he was working out the other parts of his critical philosophy, publishing the Groundwork to the Metaphysics of Morals in 1785, the Metaphysical Foundations of Natural Science in 1786, and the Critique of Practical Reason in 1788, to name only the major book-length publications; not to mention that he bought a house in 1783, which came with the usual aggravations about unreliable contractors, and kept a rather busy social calendar. For a fascinating account of Kant’s life around that time, see Kuehn 2001, chs 6 and 7. Given this packed schedule, it would not be a huge surprise if he had not found the time to carefully read the Three Dialogues.

     

127

support a meaningful distinction between proper objects and mere appearances and between experience, on the one hand, and dreams, illusions, hallucinations, and fictions, on the other hand. Also, as we will discuss in section 5.6, empirical objects on Berkeley’s account―as Kant understands it―are not grounded in transcendentally real things, or, at least, they are not grounded in the right kind of transcendentally real things. As a result, as far as their ontological status is concerned, empirical objects end up on a par with merely pseudo-existing entities such as perception-immanent red roses, Frodo, Jane’s illusory raccoon, Joe’s hallucinated oasis, and my dreamed dragon. As we may put the point with respect to our level scheme, Berkeley’s theory lacks the resources to support the introduction of a second level of reality that, although a level of mind-dependent reality, is populated by genuine existents. Against the background of our discussion of Kant’s formal conditions of experience and objectivity in section 2.5.2, it should come as no surprise that Kant pinpoints as the ultimate source of these shortcomings Berkeley’s failure to appreciate that our ability to represent proper objects and thereby introduce a distinction between truth and illusion in the realm of fully mind-dependent objects crucially depends on the forms of our cognitive faculties, that is, the pure (or a priori) intuitions of space and time, and the categories, which first give rise to the required formal conditions of objectivity and, thus, truth. Space and time together with everything that they contain in them are not things in themselves or their properties but belong merely to appearances of them; up to this point I am of one mind with these idealists [all real idealists from the Eleatic School up to the bishop Berkeley]. But they, and among them especially Berkeley, regarded space as a merely empirical representation which, just like the appearances in it, would become known to us together with all its determinations only by means of experience or perception; by contrast, I show first that space (and similarly time to which Berkeley paid no heed) together with all its determinations can be cognized by us a priori, since it, as well as time, inheres in us as before all perception or experience as a pure form of sensibility and makes all of their intuition and thus also all appearances possible. From this it follows that, since truth is grounded on universal and necessary laws as its criteria, experience for Berkeley could not have any criteria of truth because the appearances of it [experience] were not grounded a priori in anything (by him), from which it then followed that it [experience] was nothing but mere illusion, whereas for us space and time (in connection with the pure concepts of the understanding) prescribe a priori the law to all possible experience, which, at the same time, provides the secure criterion to distinguish in it truth from illusion. (Prol, 4:374 375)³⁷ ³⁷ Also see B296/A237: “ . . . these rules of the understanding are not only a priori true but even the source of all truth, i.e., of the correspondence of our cognition with the objects, through containing in them the ground for the possibility of experience, as the sum total of all cognition, in which objects may be given to us . . . ” See Prol, 4:336: “ . . . and since space, together with all appearances that it contains, belongs to the representations whose connection according to laws of experience proves their objective truth . . . ” See VA-Prol, 23:54: “My so-called idealism is of its very own kind; through it alone all our cognitions a priori, even geometry, get objective reality, i.e., they can be related to real objects, while otherwise they would be mere representations of imagined idealist things . . . ” See B520–521/A492 493: “But in space and time the empirical truth of appearances is satisfactorily secured and sufficiently distinguished from the kinship with a dream if both properly and thoroughly hang together in an experience according to empirical laws.” See R5221, 18:123: “Everything that happens, happens according to a rule, is determined in general, can be cognized a priori. Thereby we distinguish what is objective from what is subjective play (fiction), truth from illusion.” See R5636, 18:268: “Since the objects of our senses are

128

 ’      

So, by declaring himself for empirical realism and against Berkeley. Kant does not mean to deny the thesis that, from the point of view of fundamental ontology, empirical objects are fully mind-dependent. This is a thesis with respect to which both philosophers “are of one mind”—which is why they both call themselves idealists. Rather, his point is to draw attention to crucial differences in their theories of experience and empirical objects and to highlight the ways in which Berkeley’s theory falls short. These shortcomings concern the lack of any formal conditions of experience and objectivity and, furthermore, the lack of any proper material conditions for the genuine existence of empirical objects, as we will discuss in more detail in section 5.6.³⁸ When it comes to Kant’s distancing maneuvers with respect to Berkeley, it is important to keep the context in mind in which these maneuvers take place. Most of Kant’s references to Berkeley occur in the Prolegomena and the B-edition of the Critique.³⁹ This clustering is hardly surprising. Kant wrote the Prolegomena not only to provide a shorter, more popular presentation of the material contained in the Critique but also to respond to the infamous, unfavorable review of the A-edition by Garve and Feder (Garve–Feder hereafter), which presents as the most noteworthy feature of the work that cost him so much time and effort that it contains a system of idealism á la Berkeley.⁴⁰ It is important not to forget this context when evaluating Kant’s public remarks about Berkeley, both because it means that the vehemence of his protestations that his idealism is very different from Berkeley’s may very well be, above all, an expression of his pique at the abominable review and, more importantly, because his main concern may not be to distance himself from Berkeley’s actual position but from the position that his critics associate with Berkeley.⁴¹ In general, there is something to be said for the assessment that none of the parties in this early debate about critical idealism are particularly interested in what Berkeley actually says. The critics are primarily interested in conveying that Kant’s idealism is the worst possible kind of idealism, namely, the ‘it’s all in your head just like a dream’ kind, while Kant is primarily interested in making clear that this is a severe mischaracterization of his view. Berkeley not things in themselves but only appearances, i.e., representations, whose objective reality consists solely in the endurance and unity of the connection of their manifold, the objects do not give the concepts, but the concepts bring it about that we have objects of cognition in them [these objects].” See R5642, 18:280: “Dialectic. Up to this point we have only been dealing with representations, in whose exposition, their principles and use is only truth, and there was no idealism. For truth consists merely in the connection of representations in a thorough-going way according to laws of the understanding. In that consists all difference to a dream. Not in that the pictures exist separately from the mind for themselves in this way.” Also see B519/A491; B522/A494; VAProl, 23:57–58; R5400, 18:172. ³⁸ The thesis that, in Kant’s eyes, one of the crucial differences between his idealism and Berkeley’s lies in their different accounts of objectivity—or, rather, Berkeley’s lack thereof—is also defended by Dina Emundts in Emundts 2008. ³⁹ There are not very many references to Berkeley in Kant’s works and lectures. As indicated in note 36, there is a very early reference in Herder’s transcript of a metaphysics lecture in 1762–64. All other references to Berkeley (14 in total) fall in the period from 1782 to 1793, with half a dozen references in the Prolegomena (1783), one in a lecture from 82/83, and two in the B-edition of the Critique (1787). (The later references, all between 1790–93, are to be found in lecture transcripts, except for one, which is in a Reflexion.) ⁴⁰ See Garve and Feder 1782. Christian Garve wrote the initial review, which, however, was substantially revised, for the worse, by Johann Georg Heinrich Feder, the chief editor of the journal in which the review appeared, the Göttinger Gelehrte Anzeigen. ⁴¹ Regarding the first point, some of Kant’s comments on Berkley in his lectures are much more conciliatory. See V-Met-K2/Heinze, 28:770: “Berkeley wanted to say that bodies as such are not things in themselves, but he expressed himself wrongly, which is why he appears to be an idealist.” (‘Idealist’ here must be understood as ‘empirical idealist.’) Also see V-Anth/Dohna, Kowalewski, 83. Regarding the second point, in his notes for the Prolegomena, instead of writing that he will examine if the Critique contains a system of Berkleyan idealism, Kant writes “let us see if there really is an idealism to be found in it [the Critique] in the sense in which the author understands it,” (VProl, 23:53) where “the author” is the (then) anonymous reviewer.

      

129

appears to function mainly as a convenient poster-boy for this bad kind of idealism but the details of his actual position never really come up.⁴²

3.4 Core Theses of Transcendental Idealism and Empirical Realism II: the Ontological Status of Space and Time 3.4.1 Space and Time in Transcendental/Empirical Idealism/Realism TI1 and ER1 are essential elements of transcendental idealism and empirical realism, respectively, but Kant’s own descriptions of these views make clear that they do not only concern the ontological status of empirical objects but also and even more fundamentally the ontological status of space and time. . . . one necessarily must distinguish a twofold idealism, the transcendental and the empirical one. But by transcendental idealism of all appearances I understand the doctrine according to which we regard them altogether as mere representations and not as things in themselves, and according to which time and space are only sensible forms of our intuition but not determinations or conditions of objects as things in themselves that are given for themselves. To this idealism is opposed a transcendental realism, which regards time and space as something that is given in itself (independently of our sensibility). The transcendental realist imagines outer appearances (if one grants their reality) as things in themselves, which exist independently from us and our sensibility, and thus would also be outside us according to pure concepts of the understanding. (A369) Our investigations thus establish the reality (i.e., the objective validity) of space with respect to everything that we can encounter outside us as an object, but at the same time the ideality of space with respect to things if they are considered by reason in themselves, i.e., without taking into account the quality of our sensibility. We thus assert the empirical reality of space (with respect to all possible outer experience) although [we also assert] its transcendental ideality, i.e., that it is nothing as soon as we leave out the condition for the possibility of all experience and assume it to be something that grounds things in themselves. (B43–44/A27–28)⁴³

So, in order to fully capture the essence of transcendental idealism and empirical realism, we need to add a few more theses that characterize the ontological status of space and time. I will first present all of these theses, together with their transcendental realist and

⁴² Garve–Feder’s only specific comment about the Kant–Berkeley relation is that Kant’s system rests on a certain conception “of sensations as mere modifications of ourselves (on which Berkeley also mainly builds his idealism)” (Garve–Feder 1782, 41). ⁴³ Also see B52/A36: “In this thus consists the transcendental ideality of time, according to which, if one abstracts from the subjective conditions of sensible intuition, it is nothing at all, and can be attributed to the things in themselves (without their relation to our intuition) neither as subsisting nor as inhering.” See R4637, 17:640: “The ideality of space does not annul its reality with respect to bodies, i.e., with respect to all outer objects of sensibility, and space does really pertain to them, but [the ideality of space] distinguishes only objects of the senses as such from things in themselves. No space pertains to a thing in itself (as condition or determination), but every object of the outer senses is thought through the condition of space.” See B43/A27; B51–52/A34 35; B55/A38, KU, 5:189; FM, 20:268–269.

130

 ’      

empirical idealist companions, and then move to a more detailed explanation of their meaning and implications. As before, all of these theses are to be understood as articulated from the point of view of fundamental ontology. Also, as before, ‘appearance’ and ‘thing in itself ’ without any further qualifications should be understood in the transcendental sense. The second main thesis of transcendental idealism, in explicit and in concise form, can be stated as follows: (TI2) (a) Space and time are neither things in themselves in the transcendental sense nor constituted by determinations of things in themselves in the transcendental sense; rather, (b) space and time are fully mind-dependent.⁴⁴ (TI2-concise)

Space and time are transcendentally ideal.

TI2-a is materially equivalent to the claim that things in themselves do not have any spatial or temporal determinations, that is, they have no extension and thus lack shape, dimensionality, and orientation, do not instantiate any spatial location or distance relations, have no duration, and their states do not instantiate any temporal location or distance relations. For there are only two possible ways for space/time and mind-independent things to be related such that the latter have spatial/temporal determinations: (i) space/time is mindindependent and functions as a quasi-container for the things, in which case the things have their spatial/temporal determinations in virtue of being contained in the quasicontainer, or (ii) space/time is constituted by the spatial/temporal determinations of the things.⁴⁵ So, if neither one of these options holds, as stated in TI2-a, it follows that things in themselves do not have any spatial or temporal determinations. Conversely, if things in themselves do not have any spatial or temporal determinations, it directly follows that space and time are not constituted by determinations of things in themselves, since the only determinations that would be candidates for this constitution are spatial and temporal determinations. And since space and time themselves have spatial and temporal determinations, the claim that things in themselves do not have any spatial or temporal determinations also implies that space and time are not things in themselves. Accordingly, TI2-a can also be expressed by saying that things in themselves do not have any spatial or temporal determinations. Moreover, since it is reasonable to say that a thing is in space/ time only if it has some spatial/temporal determinations, TI2-a also implies that things in themselves are not in space or time. That space and time are fully mind-dependent also means that they do not exist at the transcendental level of reality. The second main thesis of transcendental realism is this: (TR2) (i) Space and time exist; and (ii) () space and time are mind-independent, and, () (what a Kantian would call) things in themselves in the transcendental sense, namely, ⁴⁴ Recall that TI1-a, the claim that empirical objects are not things in themselves, is entailed by TI1-b, the claim that empirical objects are fully mind-dependent. Similarly, TI2-a, the claim that space and time are neither things in themselves in the transcendental sense nor constituted by determinations of things in themselves in the transcendental sense, is also entailed by TI2-b, the claim that space and time are fully mind-dependent. So, TI2-b by itself would already be sufficient to capture TI2. As before, I am using the indicated more elaborate, partly redundant, formulation to mirror the way in which Kant himself tends to present the second core thesis of transcendental idealism, especially the strong version of this thesis that we will discuss in section 3.4.2. See note 56. ⁴⁵ Remember that being mind-independent means existing mind-independently and having only mindindependent ontological ingredients, including, only mind-independent determinations.

      

131

empirical objects, are in space and time and, thus, have spatial and temporal determinations.⁴⁶ (TR2-concise)

Space and time are (what a Kantian would call) transcendentally real.

As before, the transcendental realist himself would not formulate his second thesis in these Kantian terms. In his own words, he would say that space and time exist and are mindindependent, and that empirical objects, which are mind-independent, are in space and time and, thus, have spatial and temporal determinations. There are different variants of transcendental realism about space and time. One could take space and time to be selfsubsisting mind-independent things that function as quasi-containers for other mindindependent things, as on Newtonian substantivalism.⁴⁷ Or one could regard them as mind-independent properties of God, as on a Spinozist or a Morean view. Or one could see them as constituted by the mind-independent spatial and temporal relations of mindindependent finite things, as on Leibnizian relationalism.⁴⁸ Also note that as a transcendental realist one could also hold that there are some things in themselves, namely, souls, that are only in time but not in space, and thus have only temporal but no spatial determinations. But since not all transcendental realists endorse this claim, there is no ⁴⁶ Since things in themselves have spatial and temporal determinations only if space and time are things in themselves or are constituted by determinations of things in themselves, the claim that things in themselves, which are mind-independent and thus have no mind-dependent determinations, have spatial and temporal determinations implies that space and time are mind-independent. So, TR2-ii could be expressed by TR2-ii- alone. I have included TR2-ii- for improved clarity. ⁴⁷ Newton himself is reluctant to classify space and time as substances (or accidents) and prefers to treat them as ontologically sui generis, since he understands them as emanent effects of God. See Newton 1668, 131 2 (translation by Hall and Hall): “Perhaps it may be expected that I should define extension as substance or accident or else nothing at all. But by no means, for it has its own manner of existence which fits neither substances nor accidents. It is not substance; on the one hand, because it is not absolute in itself but is as it were an emanent effect of God, or a disposition of all being. . . . Moreover, since we can clearly conceive extension existing without any subject, as when we may imagine spaces outside the world or places empty of body . . . , it follows that [extension] does not exist as an accident inherent in some subject. And hence it is not an accident.” So, if substantivalism is understood as the view that space and time are self-subsisting substances, Newton cannot be classified as a substantivalist, strictly speaking; hence, the label ‘Newtonian.’ On the other hand, he also holds that space and time are ontologically independent from and ontologically prior to empirical objects, for which they function as quasi-substantial containers. For this reason, the customary practice of counting him as a substantivalist seems defensible. Also note that substantivalism about space as just explicated must be carefully distinguished from a position that one might call absolutism about the structure of space. An absolutist about the structure of space holds that the structure of space includes a privileged global rest frame with respect to which true (as opposed to merely apparent) rectilinear motions are determined. Being a substantivalist about space is compatible with being a non-absolutist about the structure of space (and being a relationalist about space is compatible with being an absolutist about the structure of space). In order to forestall confusion, I will exclusively use the label ‘substantivalism’ (and not ‘absolutism’) to refer to the Newtonian ontological position about space just characterized in the main text. ⁴⁸ It is questionable whether Leibniz himself actually endorses this view―as I read him, he clearly does not―but there certainly are philosophers in the Leibnizian tradition who conceive of space and time in this way; hence the label ‘Leibnizian.’ Kant’s understanding of Leibniz’s own view seems to have changed over time. In the Critique, he presents him as holding that space and time are constituted by relations between things in themselves and their states, respectively, which are confusedly represented; see B331–332/A275–276. But in the Metaphysical Foundations, he says of a “a great man who perhaps contributes more than anyone else to the reputation of mathematics in Germany”―by which, I think, he quite clearly has Leibniz in mind―that he tried to remind us “that space belongs only to the appearance of outer things; alas, he was not understood. One took this claim as if he wanted to say that space appears to us, and otherwise it is a thing or a relation of things in themselves, and the mathematician considers it only how it appears. But instead they should have understood by it that space is not at all a property that pertains to a thing apart from our senses in itself but only the subjective form of our sensibility under which objects of outer senses appear to us, objects that we do not know how they are in themselves; this appearance we call matter” (MAN, 4:507). Also see in this sense ÜE, 8:248 249. Also see Jauernig (forthcoming 2021b).

132

 ’      

need to take account of it for the purpose of characterizing transcendental realism about space and time. Taken together, TR2-i and TR2-ii entail that the ‘place’ in reality where space and time exist is the transcendental level of reality. Kant endorses TI2 and rejects TR2. Moving on to empirical realism, its second main thesis can be stated as follows. (ER2) (i) Empirical space and empirical time exist; and (ii) () empirical space and empirical time are mind-independent in the empirical sense, and () things in themselves in the empirical sense, namely, empirical objects, are in space and time and, thus, have spatial and temporal determinations. (ER2-concise)

Empirical space and empirical time are empirically real.

The reason for the specification that the space and time whose empirical reality is proclaimed by ER2 are empirical space and empirical time is that one can conceptually distinguish two types of space and time in Kant’s account. Kant himself does not offer a corresponding terminological distinction, but, for ease of communication, I will call the first type ‘pure space’ and ‘pure time’ and the second type ‘empirical space’ and ‘empirical time,’ respectively. We will examine the relation between these two types of space and time in detail in section 3.4.3. For now, the following three observations will be sufficient. First, pure space and pure time are the objects represented by our original, most fundamental representations of space and time, namely, our pure intuitions of space and time, and thus correspond to what we ultimately mean by ‘space’ and ‘time,’ respectively. Second, empirical space and time are the space and time in which empirical objects are in the empirical world. Third, while the general formulation of TI2, as stated above, is true for both empirical space and time and pure space and time on Kant’s view―both empirical space and time and pure space and time are fully mind-dependent―the claim that space and time exist is true only for empirical space and time.⁴⁹ With respect to ER2-ii, it is worth noting that Kant is also committed to the view that there are some things in themselves in the empirical sense, namely, empirical selves, that are only in time and, thus, have only temporal but no spatial determinations. Furthermore, in addition to ER2-ii- and the claim about empirical selves just stated, Kant also endorses the corresponding stronger claims that empirical objects necessarily are in space and time and necessarily have spatial and temporal determinations, and that empirical selves necessarily are in time and necessarily have temporal determinations.⁵⁰ But there is no need to include any of these further claims in our basic characterization of empirical realism about space and time. Taken together, ER2-i and ER2-ii entail that the ‘place’ in reality where empirical space and time exist is the empirical level of reality. ⁴⁹ The claims that space and time are empirically mind-independent and that empirical objects have spatial and temporal determinations and, thus, are in space and time are also true for both empirical space and time and pure space and time, as we will see shortly. ⁵⁰ For example, Kant takes the a priori synthetic theorems of Euclidean geometry to be true of empirical objects. This claim forms an important premise in his central argument for the thesis that space is a form of sensibility, to be discussed in section 4.2.3. Kant’s explanation of how it is possible that the a priori synthetic theorems of Euclidean geometry are true of empirical objects reveals that empirical objects necessarily have spatial determinations. Furthermore, Kant regards it as analytic and thus necessary that bodies are extended; see B11–12. If it is necessary that bodies are extended, it is also necessary that empirical objects are in space. For nothing can be extended without being in space, and empirical objects are bodies.

      

133

The empirical mind-independence of empirical space and time calls for a brief comment. The general characterization of what it means for an entity to be empirically mindindependent, namely, that experience presents it as mind-independent, applies in this case as well. But how exactly does experience present empirical space and time as mindindependent? Answering this question is a bit tricky, in part because we have not spelled out yet what exactly empirical space and time are supposed to be. We will examine this question in detail in section 3.4.3, but to anticipate the bottom line, on my reading, empirical space is constituted by the spatial determinations of empirical objects, and empirical time is constituted by the temporal determinations of empirical objects. We already know that experience presents empirical objects as mind-independent by presenting them as being in space. More precisely, we should say that experience presents empirical objects as mind-independent by presenting them as being in pure space. As we will discuss in detail in section 4.2, the a priori intuition of space is our original, most fundamental representation of space, which first defines what we mean by ‘space’; moreover, as we will see in section 3.4.2, pure space essentially is an intentional object of this intuition. The key to answering the question of how presenting an object in pure space could amount to presenting it as mind-independent, I suggest, is the realization that our original, most fundamental representation of space is also our original representation of an ‘outside.’ That is, it is an essential and primitive feature of this representation that it presents its object as ‘outside us.’ Pure space is the ‘outside’ where we present empirical objects as being when we present them as ‘outside us.’ And, obviously, in order for that to work, we must present pure space as ‘outside us’ too.⁵¹ So, any representation that is a representation of pure space necessarily presents it as ‘outside us’ and thereby as mindindependent. On the assumption that empirical space is the totality of spatial determinations of empirical objects, and empirical time is the totality of temporal determinations of empirical objects, the empirical mind-independence of empirical space and time can then be understood as parasitic on the empirical mind-independence of empirical objects. Experience presents empirical space and empirical time as mind-independent by way of presenting empirical objects as both mind-independent, namely, by presenting them as being in pure space and as having spatial and temporal determinations.⁵² Note that the claim that experience presents empirical objects as mind-independent by presenting them as being in pure space also means that pure space is represented in experience as well, at least in part, albeit not as existing.⁵³ And since pure space is necessarily presented as mindindependent, if it is represented at all, it also follows that experience presents pure space as mind-independent, which, in turn, means that pure space is empirically mindindependent, even though it does not exist at the empirical level of reality.⁵⁴ ⁵¹ See R5400, 18:172: “The question whether anything is outside of me is as much as if I asked whether I imagine a real space. For this space is something outside of me. But this does not mean that something exists in itself . . . ” ⁵² Recall that all of the determinations of mind-independent objects are mind-independent. ⁵³ As discussed in section 2.5.3, the way in which experience presents objects as existing is by presenting them as causally connected to actual perceptions and sensations. Pure space and pure time are not presented as causally connected to actual perceptions and sensations. They cannot be perceived and do not causally interact with perceivable objects. ⁵⁴ I stress this point because if pure space were empirically mind-dependent while empirical objects are empirically mind-independent we would be left with a counterexample to the claim that if a space is minddependent in a certain sense, all objects in this space are mind-dependent in the same sense. This would be problematic because Kant relies on the claim in question in his argument for the transcendental ideality of

134

 ’      

Corresponding to the second main thesis of transcendental idealism, we can formulate the second main thesis of empirical idealism thus: (EI2) (a) (What a Kantian would call) empirical space and empirical time are neither (what a Kantian would call) things in themselves in the empirical sense nor constituted by (what a Kantian would call) determinations of things in themselves in the empirical sense; rather (b) (what a Kantian would call) empirical space and empirical time are (what a Kantian would call) mind-dependent in the empirical sense. (EI2-concise) (What a Kantian would call) empirical space and empirical time are (what a Kantian would call) empirically ideal. In his own words, the empirical idealist would describe his position by saying that space and time, or, more precisely, the space and time in which empirical objects are in the empirical world, are fully mind-dependent and have the same ontological status as the intentional objects of dreams, illusions, fictions, and hallucinations.⁵⁵ Given their empirically ideal status, space and time do not genuinely exist at all, according to empirical idealism. Note that empirical idealists typically agree with empirical realists that empirical objects are in space and time; their disagreement concerns the ontological status of these entities, not their relation. Kant endorses ER2 and rejects EI2. Interestingly, he also endorses a thesis about pure space and time that is an intriguing combination of claims that are similar to (1) the negation of ER2-i, (2) ER2-ii-, and (3) EI2-a, respectively, namely, the thesis that (i) pure space and time do not exist, (ii) pure space and time are mind-independent in the empirical sense, and (iii), pure space and time are neither things in themselves in the empirical sense nor constituted by determinations of things in themselves in the empirical sense.

3.4.2 Pure Space and Time are (Nothing but) Forms of Sensibility In our discussion of the first thesis of transcendental idealism, we noted that Kant is not only committed to the general version, TI1, but also to a stronger, more specific version. TI1-strong not only informs us that empirical objects are transcendentally ideal, as TI1 does, but also tells us more concretely, in its more specific (b)-component, what Kant takes empirical objects to be, namely, intentional objects of outer experience as conceived in his

empirical objects, to be discussed in section 4.1.1. How about pure time? Does experience also represent pure time, at least in part, and, if it does, does it also present it as mind-independent? Spelling out the details of the story is a bit more complicated, but my answer is yes to both questions. Just to sketch the basic story line: Since, as we will discuss in section 4.2, we can represent specific temporal determinations of objects only by presenting them as comprised in one all-encompassing time, namely, pure time, experience also presents empirical objects as being in pure time, and thus also represents pure time, at least in part. The complication in the temporal case is that, in contrast to our pure intuition of space, our pure intuition of time, which is the representation that ontologically specifies pure time, does not have it built into it that it necessarily presents its object as mind-independent. But experience still presents pure time as mind-independent, namely, by way of presenting empirical objects both as being in pure time and as mind-independent (namely, by presenting them as being in space). ⁵⁵ Also note that, since there are no such things as (what a Kantian would call) things in themselves in the empirical sense on the empirical idealist view, EI2-a is not materially equivalent to the claim that (what a Kantian would call) things in themselves in the empirical sense do not have any spatial or temporal determinations.

      

135

theory of experience. In a similar way, we can also identify a stronger version of TI2 with a more specific (b)-component, which informs us not only that space and time are transcendentally ideal, as TI2 does, but also tells us more concretely, in its more specific (b)component, what Kant takes space and time to be. Indeed, on my reading, we can identify two stronger versions of TI2, TI2-pure-strong and TI2-empirical-strong because, as just noted in section 3.4.1, Kant distinguishes two types of space and time, namely, pure space and time, on the one hand, and empirical space and time, on the other hand. In this section, we will take a closer look at Kant’s conception of pure space and time and the strong version of TI2-pure, which states, in its (b)-component, that pure space and pure time are forms of sensibility, or forms of intuition, or forms of appearances, as Kant also puts it (where, as before, ‘sensibility’ is to be understood as ‘human sensibility’). We will examine the TI2-empirical-strong in the following section. (Terminological note: from now on, in all following sections and chapters, in characterizations of Kant’s views, ‘space’ and ‘time’ without any further qualifications should be understood as ‘pure space’ and ‘pure time,’ respectively. If the context calls for it, I will add the qualification ‘pure’ for emphasis and to safeguard against possible confusion. Whenever empirical space or empirical time is the intended referent, I will explicitly say so.) (TI2-pure) (a) Pure space and pure time are neither things in themselves in the transcendental sense nor constituted by determinations of things in themselves in the transcendental sense; rather (b) pure space and pure time are fully mind-dependent. (TI2-pure-strong) (a) Pure space and pure time are neither things in themselves in the transcendental sense nor constituted by determinations of things in themselves in the transcendental sense; rather (b) pure space and pure time are (nothing but) forms of sensibility and thus fully mind-dependent.⁵⁶ In Kant’s own words: (a) Space does not at all represent a property of some things in themselves or them in their relation to each other, i.e., [it does not represent] a determination of them that adhered to objects themselves and that remained even if one were to abstract from all subjective conditions of intuition . . . (b) Space is nothing but the form of all appearances of outer senses, i.e., the subjective condition of sensibility, under which alone outer intuition is possible for us. (B42/A26)⁵⁷ (a) Time is not something that subsists for itself or adheres to things as objective determination, and thus remained if one abstracted from all subjective conditions of their intuition. . . . [T]ime is nothing but the subjective condition under which all

⁵⁶ I have included both (a) and (b) in this formulation of TI2-pure-strong to match Kant’s own presentation in the much-discussed ‘Conclusions from the Above Concepts’ in the Transcendental Aesthetic, which is the passage quoted next in the main text, which we will discuss in detail in section 4.2. Of course, since (b) entails (a), TI2-pure-strong can also be presented by just stating (b). ⁵⁷ See B457/A430, note: “Space is merely the form of outer intuition (formal intuition) but not a real object that can be intuited outwardly.” See B120/A87: “ . . . the outer sensible world of which space is the pure form of its intuition.” Also see B460/A432.

136

 ’      

intuitions can take place in us. For in this case this form of inner intuition can be represented prior to the objects, and hence a priori. (B49/A32 33)⁵⁸

In order to understand the precise meaning of the claim that pure space and time are (nothing but) forms of sensibility and to see how this entails that they are fully minddependent, we need to get clearer on what exactly a ‘form of sensibility’ is supposed to be. It is one of the main tenets of Kant’s theory of the mind that sensations themselves are formless. As we learn in the first section of the Transcendental Aesthetic, the general scheme of order in conformity with which sensations are intuited in particular relations is not given with our sensations but is due to us, namely, due to the specific constitution of sensibility, and “lies ready in the mind a priori.”⁵⁹ Empirical intuitions depend both on things in themselves that affect us and on sensibility, which is constituted such that, if affected, it produces, with help from the imagination, representations that necessarily present their objects as having particular spatial or temporal determinations that are necessarily governed by certain general schemes of order. Similarly, experience depends both on things in themselves and on sensibility and the understanding, which are constituted such that, in the construction of experience, as detailed in chapter 2, they produce representations that necessarily present their objects as having particular spatial and temporal determinations that are necessarily governed by certain general schemes of order. There are several candidates for being called ‘the forms of sensibility’ in this account: (1) certain features or properties of the faculty of sensibility that are responsible for all intuitions and experience necessarily presenting their objects as having spatial and temporal determinations that are necessarily governed by certain general schemes of order; (2) certain general schemes of order that are responsible for all intuitions and experience necessarily presenting their objects as having spatial and temporal determinations that are necessarily governed by them; or (3) the sum total of the represented spatial and temporal determinations. The third option can be set aside quickly. The forms of sensibility are supposed to be partly responsible for the spatial and temporal determinations represented in intuition and experience, and it makes not much sense to say that these determinations are partly responsible for themselves. Moreover, these determinations depend not only on sensibility but also on how matters stand in the realm of things in themselves. That experience presents tomatoes as having a shape that conforms to the postulates of Euclidean geometry is (exclusively) due to sensibility, but that it presents them as having a particular tomato-shape also depends on the realm of things in themselves. The first option is a more serious contender, and it may well be that at certain places where Kant employs the term ‘forms of sensibility’ he indeed is thinking of certain structural features ⁵⁸ See B47/A31: “Time is not a discursive or, as one calls it, general concept but a pure form of sensible intuition.” See B51–52/A35: “ . . . if we take the objects as they might be in themselves, then time is nothing. It is of objective validity only with respect to appearances. . . . Time is merely a subjective condition of our (human) intuition . . . and in itself, outside of the subject, nothing.” See B56/A39: “For they [space and time] both taken together are pure forms of all sensible intuition . . . ” ⁵⁹ The full passage reads as follows, B34/A20: “Since that in which sensations alone can order themselves and can be placed in a certain form cannot again itself be sensation, it follows that while the matter of all appearance is given to us only a posteriori, their form must altogether lie ready for them in the mind a priori, and thus must be able to be considered in abstraction from all sensation. I call all representations pure (in the transcendental understanding), in which nothing is found that belongs to sensation. Thus, the pure form of sensible intuitions in general will be found in the mind a priori, [a form] in which all manifold of appearances is intuited in certain relations.”

      

137

or properties of the faculty of sensibility. But, I submit, when he identifies space and time with the forms of sensibility, he must be using the term in a different sense. Since, arguably, the faculty of sensibility is not mind-dependent (in the relevant sense⁶⁰), if space and time were forms of sensibility understood as properties or features that characterize its particular nature, they would not be mind-dependent either. Moreover, as noted above, Kant calls space and time not only “forms of sensibility” but also “forms of intuition,” and says of them that they are “forms of appearances.” I take these to be clear indications that what he has in mind when he classifies space and time as forms of sensibility is not forms of our receptive cognitive faculty itself but forms of the representations, or, more precisely, of the presentational contents of the representations, that are produced by sensibility or in collaboration with sensibility. That Kant must be thinking of space as a form of presentational contents, rather than as a form of representations, becomes clear upon a moment’s reflection. His view is obviously not that my mental state that is an empirical intuition of, say, a tomato is extended and round, but that it presents a tomato as extended and round. The case of time is a bit more complicated; for at first glance it might seem as if time is both a form of presentational contents, such as the intentional objects of (outer) experience, which are presented as being in time, and a form of representations, namely, of all our representations in inner sense. But if we take seriously the conception that for us to be conscious of a representation in inner sense is to represent that representation to ourselves, as briefly discussed in section 2.9, time as the form of inner sense can also be understood as a form of presentational contents. The forms of sensibility determine part of the ‘form’ of the intentional objects that are presented in intuition and experience, which structures their ‘matter.’ (The other part of their ‘form’ is determined by the forms of the understanding, or the categories.) This leaves us with the second option mentioned above as the only viable choice: the forms of sensibility are certain general schemes of order. More specifically, the forms of sensibility are general schemes of order that are responsible for all intuitions of a certain type (namely, the inner or outer type) of finite objects necessarily presenting these objects as having determinations of a corresponding type (namely, the temporal or spatial type) that are necessarily governed by these schemes of order.⁶¹ I take it that this ordering function of the forms of sensibility is what Kant is getting at when he explicates the term “form of appearances” as referring to “that which makes that the manifold of the appearance can be ordered in certain relations” (B34).⁶² The proposed characterization of what a form of sensibility is supposed to be is a good start to explicate the meaning of TI2-pure-strong but it is not the whole story yet. Kant not only describes the forms of sensibility as forms of intuition, he goes so far as to claim that ⁶⁰ Of course, as a cognitive faculty of a mind, sensibility is mind-dependent in some sense. But it is not true that sensibility exists, or has any of its determinations, in virtue of being represented by a finite mind, that is, sensibility is not mind-dependent in the sense defined in section 2.3, which is the only relevant sense of ‘mind-dependent’ for our present concerns. ⁶¹ Note that the reason why I say that the forms of sensibility are responsible for all intuitions of finite objects of a certain type necessarily presenting these objects as having determinations of a certain corresponding type is that, as we will see in a second, the forms of sensibility essentially are infinite intentional objects of the pure intuitions of space and time. So, if we did not restrict the claim in question to finite objects, we would be saddled with the assertion that the forms of sensibility are responsible for the way in which the pure intuitions of space and time represent those very forms, which would be problematically circular. ⁶² Also see V-Met/Heinze, 28:178: “In our sense there is a form or a ground how the sensible impressions are coordinated with one another, this is the subjective form in which all manifold things appear. It is thus a form of sensibility, in which the things appear next to each other or after each other.”

138

 ’      

“this pure form of sensibility will also itself be called pure intuition” (B34–35/A20). Similarly, he frequently refers to space and time—which, as we now know, are forms of sensibility—as pure intuitions.⁶³ Our characterization does not shed any light on why the forms of sensibility are themselves pure intuitions. Relatedly, we also have not explained yet how it follows from the fact that space and time are (nothing but) forms of sensibility that they are fully mind-dependent. To make some headway in these respects, it is useful to take note of the results of the metaphysical expositions of the concepts of space and time, to be examined in detail in section 4.2.2.2, according to which our original and most fundamental representations of space and time are a priori intuitions, and, similarly, of what Kant tells us about the content of a priori intuitions, namely, that “pure intuition contains merely the form under which something is intuited” (B74–75/A50–51), that is, that pure intuitions represent the forms of sensibility.⁶⁴ Indeed, since the forms of sensibility govern the spatial and temporal determinations of appearances represented in perception and experience, they are reflected in these determinations, and thus also at least partly represented in perception and experience. Kant repeatedly illustrates this by way of a thought experiment. In this thought experiment he asks his readers to isolate certain features of the spatial form of sensibility in their empirical intuitions by a process of abstraction. If we remove all of the contributions of the understanding as well as all a posteriori elements from empirical intuitions we are left with elements of pure intuition, which are aspects of one of the forms of sensibility, space. Thus, if I separate from the representation of a body that which the understanding thinks of it, such as substance, divisibility etc., as well as that which belongs to sensation, such as impenetrability, hardness, color etc., something remains to me from this empirical intuition, namely, extension and shape. These belong to pure intuition, which takes place in the mind a priori as a mere form of sensibility, also without any real object of the senses or sensation. (B35/A20–21)⁶⁵

⁶³ See B39/A24–25: “Space is not a discursive or, as one says, general concept of relations of things in general but a pure intuition.” See B43/A27: “The persistent form of this receptivity, which we call sensibility, is a necessary condition of all relations in which objects are intuited as outside us and, if one abstracts from these objects, a pure intuition, which carries the name space.” Also see MAN, 4:475, note: “Granted: that these pure intuitions can never be anything other than mere forms of appearances of outer or of the inner sense (space and time) . . . ” See R4673, 17:638: “Space is nothing but the intuition of the mere form also without any given matter, and thus is pure intuition.” ⁶⁴ See FM, 20:266: “An intuition that is supposed to be possible a priori can only concern the form under which the object is intuited, for that is what it means to represent something a priori. . . . Now a representation that merely concerns the form is called pure intuition, which, if it is supposed to be possible, must be independent of experience.” See FM, 20:267: “The form of the object, as it alone can be represented in an intuition a priori . . . ” See B748/A720: “Now of all intuition none is given a priori except the mere form of appearances, space and time . . . ” Also see Prol, 4:284; V-Met/Schön, 28:482 484. ⁶⁵ See B36/A22: “In the Transcendental Aesthetic we thus first will isolate sensibility by removing everything that the understanding thinks in this through its concepts, so that nothing remains but empirical intuition. Second, we will remove from this still everything that belongs to sensation, so that nothing remains but pure intuition and the mere form of appearances, which is the only thing that sensibility can provide a priori. In this examination we will find that there are two pure forms of sensible intuition as principles of cognition a priori, namely, space and time . . . ” See Prol, 4:283: “But both representations [of space and time] are mere intuitions; for if from the empirical intuitions of bodies and their changes (motion) one leaves out everything that belongs to sensation, space and time remain, which are thus pure intuitions that a priori ground the former, and thus themselves can never be left out but which demonstrate precisely by being pure intuitions a priori that they are mere forms of our sensibility, which must precede all empirical intuition, i.e., the perception of real objects . . . ” Also see B5 6; ÜE, 8:240.

      

139

All of this can be taken to suggest that the forms of sensibility essentially are intentional objects of pure intuitions and, as such, essentially forms of presentational contents. That is, a form of sensibility is not a general form that happens to apply, among other things, to presentational contents; rather, a form of sensibility is a form that could never apply to anything other than presentational contents because it itself is a presentational content. Since all intentional objects are fully mind-dependent, this conception of what it means to be a form of sensibility directly makes clear why space and time, as forms of sensibility, are fully mind-dependent. It also explains why Kant often describes the forms of sensibility, as well as space and time, as pure intuitions. Just as ‘representation’ is ambiguous between the mental state that does the representing and the presentational content that is represented, ‘intuition’ is ambiguous between the mental state and the intuited content. So, just as Kant’s often repeated claim that appearances are ‘mere representations’ is to be understood as saying that they are intentional objects of representations—namely, of the representations constituting experience—his claims that the forms of sensibility and space and time are pure intuitions are to be understood as saying that they all are intentional objects of pure intuitions. So, in sum, we can record the following definition of a form of sensibility, or a form of intuition. S is a form of sensibility/intuition if, and only if (1) S essentially is an intentional object of a pure intuition of a certain type T; (2) S is responsible for all intuitions of type T of finite objects necessarily presenting these objects as having determinations of a corresponding type T*; and (3) S is a general scheme of order that necessarily governs all determinations of type T*.⁶⁶ That the forms so characterized are forms of sensibility, that is, forms that are determined by the nature of sensibility and not some other cognitive faculty, is captured in the definition by the characterization that the representations whose intentional objects they are, are pure intuitions, and that the presentational contents whose determinations they govern are presentational contents of intuitions. For, as we learn in the first section of the Transcendental Aesthetic, “sensibility alone provides us with intuitions” (B33/A19). With an eye toward our later discussion, we will say that conditions (2) and (3) define what it means to be a form of intuition in the broad sense, while conditions (1), (2), and (3) define what it means to be a form of sensibility, or a form of intuition simpliciter.⁶⁷ ⁶⁶ Note that there is a difference between saying that an object is necessarily represented to have spatial determinations and all of these spatial determinations are necessarily governed by the Euclidean postulates, on the one hand, and saying that an object is represented to necessarily have Euclidean spatial determinations. On my reading, intuitions are incapable of representing the latter, as discussed in section 2.4. ⁶⁷ This distinction between a form of intuition in the broad sense and a form of sensibility/intuition, mirrors Kant’s distinction in a famous footnote between a form of intuition and a formal intuition. (What I am calling ‘form of intuition in the broad sense’ corresponds to what Kant, in the footnote, is calling ‘form of intuition’; and what I am calling ‘form of sensibility/form of intuition’ corresponds to what Kant is calling ‘formal intuition.’) See B160 161, note: “Space represented as an object (as one really requires for geometry), contains more than the mere form of intuition, namely, comprehension of the manifold that is given according to the form of sensibility in an intuitive representation, so that the form of intuition gives merely a manifold, the formal intuition gives unity of the representation. In the Transcendental Aesthetic, I counted this unity merely to sensibility only to observe that it precedes all concepts, although it presupposes a synthesis that does not belong to the senses but through which all concepts of space and time first become possible. For since through it (in that the understanding determines sensibility) space or time as intuitions are first given, the unity of this intuition a priori belongs to space and time and not to the concept of the understanding. (§24.)” Kant introduced this distinction only in the B-edition of the Transcendental Deduction and did not revise the rest of the text in light of it, so that “form of intuition” in his usage remains ambiguous between signifying either a form of intuition or a formal intuition, as distinguished in the footnote. That is, when he uses ‘form of intuition’ anywhere else we should not assume that he means ‘form of intuition’ in the sense of the footnote; indeed, he usually means ‘formal intuition’ in the sense of the footnote (which is why I am using ‘form of intuition’ simpliciter to refer to what, in the footnote, he calls ‘formal intuition’).

140

 ’      

That the forms of sensibility do not only apply to the objects represented in intuition but also the objects represented in experience follows from the fact that experience is, by necessity, constructed on the basis of empirical intuitions, which furnish us with the ‘matter’ for the objects of experience and underwrite their existence.⁶⁸ Finally, the forms of sensibility determine part of the ‘form’ of the objects of intuition and experience; this is why, I take it, Kant sometimes uses the label ‘forms of appearances’ when he talks about the forms of sensibility. Kant often uses the formulation that space and time are “nothing but,” or “mere,” forms of sensibility. In the secondary literature, these qualifications are usually taken to signify that space and time are only forms of appearances, as opposed to being both forms of appearances and forms of something else, in particular, as opposed to being both forms of appearances and forms of things in themselves. (A form of things in themselves would be a general scheme of order that [necessarily] governs or is responsible for and governs all determinations of a certain type of things in themselves.) The latter option is (a version of) the famous third possibility that Kant supposedly overlooked and whose actuality he supposedly failed to rule out in his argument for transcendental idealism in the Transcendental Aesthetic.⁶⁹ But if we adopt the understanding of a form of sensibility just proposed, the alleged third possibility vanishes into thin air. If something is a form of sensibility in the sense just explicated, it could not possibly be a form of things in themselves because forms of sensibility essentially are forms of presentational contents, and things in themselves are not presentational contents. So, if Kant’s argument in the Transcendental Aesthetic succeeds in establishing that space and time are forms of sensibility thus understood, it is impervious to the notorious objection of the neglected third possibility, at least in the version just formulated. (Whether it does thus succeed will be discussed in section 4.2.3.) Whether the forms of sensibility could at the same time be, if not forms of things in themselves, then forms of appearances that depend on non-human minds, for example, forms of the sensibility of angles or forms of the way in which God represents the world when He had too much to drink, does not matter for Kant’s concerns in the Transcendental Aesthetic. What Kant needs for the purpose of his own philosophy, and what he wants for the purpose of the debate about the nature of space and time with his contemporaries— where, as Kant sees it, all serious proposals in contention are versions of the view that space and time are forms of things in themselves⁷⁰—is only that space and time are forms of appearances to us, and that they are not forms of things in themselves. And in order to state that claim, it is enough to say that space and time are forms of sensibility. But, then, one naturally wonders, why does Kant so frequently add the “nothing but” and “mere” qualifications anyway when he characterizes space and time as forms of sensibility? There are several possible reasons. It could be that, apart from the explicated

⁶⁸ This point will be discussed in more detail in section 4.2.3. ⁶⁹ We will take a closer look at this objection in section 4.2.5. ⁷⁰ That this is indeed the dialectical context in which Kant views his theory of space (and time), comes out clearly at the beginning of the Metaphysical Exposition where he canvasses the different possibilities on offer of what space and time could be and only lists that they might be things in themselves or determinations of things in themselves or forms of sensibility. See B37–38/A23: “What, then, are space and time? Are they real beings? Are they merely determinations or also relations of things but still such [determinations] that belonged to them also in themselves, even if they were not intuited, or are they such [determinations] that only adhere to the form of intuition alone and thus to the subjective quality of our mind without which these predicates could not be applied to any thing?”

      

141

meaning of ‘form of sensibility,’ Kant also uses this term in a more general sense, that is, in the sense of a form of intuition in the broad sense as captured by conditions (2) and (3) alone or perhaps even only condition (3) alone, as meaning a general scheme of order that necessarily governs all determinations of a certain type of all objects that are represented in a certain type of intuition. A form of sensibility thus understood could also be a form of things in themselves, which would call for adding a ‘nothing but’ or ‘mere’ to forestall misunderstandings when using the term ‘form of sensibility’ in this more general sense to characterize space and time. More likely, though, Kant could have chosen to include these qualifications for primarily rhetorical purposes. Even though redundant, strictly speaking, he probably added ‘nothing but’ or ‘mere’ for emphasis and to underline that space and time are ontologically less fundamental than they are commonly taken to be. Kant also often uses the formulation that appearances are nothing but, or mere, representations, by which he obviously does not intend to say that appearances are only intentional objects of representations as opposed to being both intentional objects and something else. Rather, he means to convey the ontological lightweight character of appearances. Instead of being mind-independent objects, they are nothing but, or mere, intentional objects. Similarly, the formulation that space and time are nothing but forms of sensibility can be understood as highlighting their ontological lightweight character. Instead of being forms of things in themselves, or mind-independent forms of mind-independent objects, space and time are merely forms of sensibility, or fully mind-dependent forms of fully mind-dependent objects. Since I take it that this is what is actually going on in the relevant passages, I put ‘nothing but’ in brackets in our formulation of T2-strong-pure above; it is not needed, strictly speaking, but it is useful for purposes of emphasis. Incidentally, the proposed account of what a form of sensibility is also explains the supersensible nature of things in themselves and why we cannot intuit them. The forms of sensibility, as defined, must be applicable to anything sensible or intuitable. Given that these forms essentially apply to presentational contents, it follows that anything that is sensible or intuitable is a presentational content. But this means that no transcendentally real thing is sensible or intuitable, or, put positively, that all transcendentally real things are non-sensible or supersensible, as Kant prefers to say, and that we cannot intuit them.

3.4.3 Empirical Space and Time The main question that arises with respect to ER2-i, the first half of the second thesis of empirical realism, which states that empirical space and empirical time exist, is how exactly Kant conceives of empirical space and time. More specifically, an important question in this context is how exactly empirical space and time are related to pure space and pure time, which, as we just learned, essentially are objects of pure intuitions. Bearing in mind that substantivalism and relationalism are the two main variants of transcendental realism about space and time that Kant considers, another specific question is whether, as an empirical realist about space and time, he endorses an analogous version of one of these views with respect to the empirical level. That is, does he hold that empirical space and time are self-subsisting things at the empirical level, or that they are constituted by the spatial and temporal determinations of empirical objects, or does he conceive of them in yet another way? The present section is devoted to the examination of these questions.

142

 ’      

A good place to start is with the distinction between pure space and time and empirical space and time. In response to his wannabe nemesis Eberhard, who takes himself to have landed a serious punch against Kant by “completely apodictically” proving “the truth that space and time have both subjective and objective grounds,” and that “their ultimate objective grounds are things in themselves,” Kant makes the following striking reply: Now every reader of the Critique will admit that these are exactly my own assertions, and thus that Mr. Eberhard, with his apodictic proofs . . . , has asserted nothing contrary to the Critique. But that these objective grounds, namely, the things in themselves, are not to be looked for in space and time but in that which the Critique calls their outer- or supersensible substrate (noumenon), this was my assertion with respect to which Mr. Eberhard wanted to prove the opposite without ever coming quite out with it, including here in his final result. (ÜE, 8:207)

Kant is saying here that his real disagreement with Eberhard concerns the claim that things in themselves are in space and time, which Eberhard affirms and Kant denies, but not the claim that space and time have things in themselves as their objective grounds, which both of them affirm. I do not share Kant’s confidence that every reader of the Critique will readily admit that he regards things in themselves as the objective grounds of space and time. Indeed, this claim will probably be puzzling to many readers because it apparently conflicts with his central doctrine that space and time are (nothing but) forms of sensibility—which, quite clearly, is the doctrine that Eberhard is targeting as well. But Kant’s assertion makes perfect sense if we recognize that he distinguishes between pure space and time and empirical space and time and intends his remark as a comment about empirical space and time. While pure space and time, which are ontologically specified by pure intuitions, only have subjective grounds, namely, the nature of sensibility, empirical space and time, which are what Eberhard is talking about and which, on Kant’s account, are ontologically specified by experience, are grounded in both the nature of our cognitive faculties and in things in themselves. In order to get a better grip on the distinction between pure space and time and empirical space and time, it is helpful to compare their properties. The properties of pure space and time are what pure intuitions present them to be; they depend exclusively on the nature of sensibility. The properties of pure time are exhausted by what we might call ‘structural’ properties; the properties of pure space also include geometric properties in addition to structural properties. One of the main structural features of pure space and pure time is that each one of them is an “infinite given magnitude” (B39). As Kant’s remarks in the Transcendental Aesthetic make clear, he means by this both that pure space and pure time are infinite as far as their extension/duration is concerned and that they are infinitely divisible.⁷¹ Other structural features of time include that it is one-dimensional and that its parts are ordered by the ‘is before’ relation, which is irreflexive, antisymmetric, transitive, and connex; other structural features of space include that it is three-

⁷¹ See B39–40/A25, B47–48/A31 32. He also classifies pure space and time as continuous, which he characterizes as being such that none of their parts is the smallest, or being such that all of their parts are themselves spaces, or times, respectively. That is, Kant basically equates being continuous and being infinitely divisible. See B211/A169.

      

143

dimensional, orientable, and has the standard topology of a three-dimensional real space, R³.⁷² Among the geometric properties of pure space, we find the properties of being such that the Pythagorean Theorem holds and that the internal angle sum of any triangle in it is 180 degrees, properties that mark the space as Euclidean. The properties of empirical space and time are what experience presents them to be. As just noted, since the particular spatial and temporal determinations of appearances that are represented in experience are governed by and are in part due to the forms of sensibility, these forms are reflected in the represented determinations and also, at least partly, represented in experience. Accordingly, empirical space and time and pure space and time share a lot of properties, properties that depend exclusively on the nature of sensibility. Empirical space has all of the structural and geometric properties that pure space has with one possible exception; empirical time has all of the structural properties that pure time has with one possible exception. The one possible exception in both cases is the structural property of being infinite with respect to extension/duration. Whether empirical space and time are infinite in this sense depends on how we conceive of their ontological status at the empirical level of reality and, for the case of space, on whether experience comprises a representation of an ether, or of a special kind of subtle matter, that fills the space between the heavenly bodies and is distributed throughout space to infinity. We will return to the question whether empirical space and time are infinite below after having completed our survey of their other properties. Since the forms of sensibility are among the conditions for the possibility of experience, the properties that empirical space and time share with pure space and pure time are essential to them. Indeed, since the pure intuitions of space and time are our original and most fundamental representations of space and time, and thus define what we mean by ‘space’ and ‘time,’ the properties that empirical space and time share with pure space and time are essential to anything that can count as a kind of space or time for us.⁷³ But empirical space and time also have some properties that pure space and time do not have, and they could be said to have some properties essentially that are not essential to pure space and time. First, one might hold that it is essential to empirical space and time that empirical objects are ‘in’ them, while this is not essential to pure space and time.⁷⁴ More specifically, one might say that it is essential to empirical space and time that empirical objects take up particular determinate volumes of space and endure throughout particular determinate periods of time. Since these relational properties do not only depend on the forms of sensibility but also on things in themselves, and since different possible worlds in which human minds exist can differ with respect to their transcendental level of reality, the particular relational properties that empirical space and time have with respect to empirical objects in a given possible world are not essential to them. But one ⁷² A relation R on set S is connex if, and only if, for every two members a, b of S, either aRb or bRa. An orientable space allows for incongruent counterparts, that is, objects whose internal spatial relations among their parts are exactly the same but that cannot be brought to coincidence through rigid motions (in the space in question), for example, a left and a right hand. ⁷³ Of course, in thinking that being Euclidean is a property of pure space Kant turns out to have made a mistake—a mistake that, I should add, is fairly easy to correct, though, without having to change much else in Kant’s theory. ⁷⁴ As we will discuss below, since we necessarily represent every determinate space and time and, hence, every object with a determinate shape and duration as ‘cut-out’ parts of pure space and time, respectively, empirical objects are also in pure space and time, not just empirical space and time. But, and this is the important point for now, it is not essential to pure space and time that empirical objects are in them.

144

 ’      

could still insist that it is essential to empirical space and time to have some relational properties with respect to empirical objects, while this is not essential to pure space and time. Second, empirical space and empirical time also have certain properties that depend on some of the concepts that are involved in the construction of experience, while pure space and time are determined by the nature of sensibility alone. For reasons that will become apparent in a minute, I will call these properties ‘dynamical properties.’ The dynamical properties of empirical space and time are reflections of certain geometric properties of what we now call spacetime, including, in particular, its inertial structure. This structure is required for us to be able to represent determinate durations by allowing us to specify a measure of time; it is also required for us to be able to introduce a meaningful distinction between actual or true and merely apparent motions, which is an important step in the construction of experience.⁷⁵ The dynamical properties of empirical time include, most importantly, a specific temporal distance function that allows us to represent determinate durations; the dynamical properties of empirical space and time taken together, that is, taken as empirical spacetime, include, most notably, the property of being such that bodies that rotate in space are subject to certain inertial forces, as for example, the centrifugal force that causes the water in a rotating bucket to rise up along the walls. In brief, Kantian empirical space and time taken together have all the properties of what is now known as Galilean spacetime.⁷⁶ In contrast to the structural properties, the dynamical properties of empirical spacetime depend not only on the forms of sensibility but also on the forms of the understanding, including, in particular, the category of causality. It is not entirely clear, though, if Kant thinks that the dynamical properties exclusively depend on the forms of our cognitive faculties or also on things in themselves. Kant’s derivations of his mechanical laws in the Metaphysical Foundations, which encode the indicated Galilean spacetime structure, rely on the concept of matter as the movable in space, which Kant explicitly classifies as empirical.⁷⁷ This could be taken to suggest that Kant’s mechanical laws and, with them, the Galilean structure of empirical spacetime only obtain in all possible worlds in which matter is the movable in space (and the movable in space is matter), which, in turn, depends, at least partly, on what kind of things in themselves exist. Similarly, one could argue that in his insistence against Eberhard that empirical space and time have things in themselves as their objective grounds, he is not only thinking of their particular relations to empirical objects but also of their dynamical properties, as dependent on things in themselves. Given that possible worlds in which human minds exist can differ with

⁷⁵ See MAN, 4:554–558. Also note that in order for us to be able to represent true rectilinear motions and true motions with specific determinate velocities even more structure would be required. Briefly put, we would need to represent a privileged frame of reference at rest, that is, we would need to represent ‘absolute space.’ In Kant’s words, MAN, 4:560: “ . . . all motion and rest must be reduced to absolute space if their appearance is supposed to be transformed into a determinate concept of experience (that unifies all appearances).” Unfortunately, as Kant explains, the determination or representation of such a privileged frame of reference in experience, is beyond us. See MAN, 4:559: “It [absolute space] cannot be an object of experience; for space without matter is not an object of experience, but still it is a necessary concept of reason, and thus nothing more than a mere idea.” For an illuminating and detailed discussion of Kant’s construction of the spatiotemporal framework of Newtonian theory in the Metaphysical Foundations, see Friedman 1992, ch. 3, 136–64, and Friedman 2013, esp. ch. 4, 413–562. Also see note 169, chapter 2. ⁷⁶ For a helpful explanation of how to distinguish the various spacetime structures associated with different physical theories of the past with the help of modern mathematics, see Friedman 1983. ⁷⁷ See B58/A41, MAN, 4:470.

      

145

respect to their transcendental levels of reality, if the dynamical properties of empirical spacetime depend, at least in part, on things in themselves, then the particular dynamical properties that empirical spacetime has in the actual world may not be essential to it. To be sure, despite their reliance on the empirical concept of matter as the movable in space, Kant regards his laws of mechanics as a priori and hence necessary and as belonging among the constitutive principles that form the backbone for the construction of experience and the constitution of appearances. But this necessity could be understood as hypothetical or conditional. In all possible worlds in which human minds exist and matter is the movable in space, the derived particular mechanical laws hold without exception and function as constitutive principles, and empirical spacetime has the structure of Galilean spacetime.⁷⁸ I am inclined to read Kant in this way. At any event, whether or not the specific dynamical properties that empirical spacetime has in the actual world are essential to it, one could still hold that it is essential to empirical spacetime to have some dynamical properties such that empirical time is measurable in some way, and that empirical spacetime has some kind of geometric structure. Now, against this background, the following possible substantivalist and relationalist readings of Kant’s views about the ontological status of space and time at the empirical level of reality suggest themselves. According to the substantivalist reading, empirical space and time are self-subsisting things in themselves (in the empirical sense) in their own right, which exist at the empirical level of reality independently of and in addition to empirical objects, empirical selves, and their determinations (as well as independently of and in addition to the ether, assuming it exists). The most straightforward move for the substantivalist would be to simply numerically identify empirical space and time with pure space and time, respectively. On this reading, that empirical objects are in them is not essential to empirical space and time; neither is it essential to empirical time that empirical selves are it. Similarly, the dynamical properties of empirical space and time are also not essential to them. Empirical space and time are pure space and time with some additional contingent dynamical properties. On this version of substantivalism, empirical space and time would be infinite with respect to their extension and duration, respectively.⁷⁹ According to the relationalist reading, empirical space consists in the sum total of (a) all “filled” determinate spaces, that is, all spatial volumes dynamically occupied by matter so that it resists everything moveable that strives to enter the space in question (see MAN, 4:496), spaces that are characterized by the determinations of extension, threedimensionality, shape, and spatial size of the empirical objects occupying them, where the size is measured according to the Euclidean distance function (that is, the Pythagorean theorem) and specified with respect to a particular unit of measurement, and (b) all instances of spatial location relations and ‘thick’ spatial distance relations of ⁷⁸ And, who knows, maybe it is a metaphysical fact that in all possible worlds in which human minds exist the transcendental level of reality is such that matter is the movable in space. In that case, the necessity of Kant’s laws of mechanics would be unrestricted (apart from the restriction to worlds in which human minds exist): in all possible worlds in which human minds exist, Kant’s laws hold and empirical spacetime has the structure of Galilean spacetime. ⁷⁹ How could experience present empirical space and time as infinite? Answering this question is one of the challenges that proponents of the substantivalist reading must face. The most straightforward option is to say that, due to being based on the pure intuitions of space and time, the representations of space and time in experience present them such that for any determinate finite space and any determinate finite duration there is a larger space and a larger duration, in which they are enclosed, respectively, which, one might argue, amounts to presenting an infinite space and time in which all of these finite spaces and times are contained.

146

 ’      

empirical objects, where the former comprise instances of the relations of being in between and being next to, and the latter are of the form of being a particular Euclidean distance away from X specified with respect to a particular unit of measurement. Empirical time consists in the sum total of (a) all ‘filled’ determinate times, that is, all successions of states of matter, successions that are characterized by the determinations of duration and temporal size of the empirical object whose states make up the succession in question, where the size is determined according to a measure derived from the distance function of Galilean spacetime and specified with respect to a particular unit of measurement, and (b) all instances of temporal location relations and ‘thick’ temporal distance relations of empirical objects, where the former comprise instances of the relations of being before, being after, and being simultaneous with, and the latter are of the form of being a particular Galilean temporal distance before/after X specified with respect to a particular unit of measurement. More briefly, empirical space and empirical time are constituted by, in the sense of being the sum total of, the spatial and temporal determinations of empirical objects (and of the ether, if it exists), respectively.⁸⁰ Given their richness, the listed spatial and temporal determinations of empirical objects easily encode all of the structural and dynamical properties of empirical space and time just described.⁸¹ On the relationalist reading, it is essential to empirical space and time that empirical objects are in them and that they have some kind of dynamical properties, while neither feature is essential to pure space and time. Empirical space and time are understood as ‘materialized,’ more determinate versions of pure space and time, as it were, which embody the properties of pure space and time (except, possibly, infinity) and, thereby, actualize or realize them at the empirical level of reality. Since the totality of spatial and temporal determinations of empirical objects (and the ether, if it exists) is distinct from the general schemes of order that necessarily govern these determinations, it would be incorrect to say that empirical space and time thus understood, as materialized, more determinate, versions of pure space and time, are numerically identical to pure space and time. A better way of characterizing their relation is to say that empirical space and time conceived in the indicated fashion are more determinate instances of pure space and time. According to Kant’s analysis in the Transcendental Aesthetic, in contrast to concepts, which contain their instances under them, our pure intuition of space and time literally contain their instances in them, or,

⁸⁰ What about the temporal determinations of empirical selves? Are they not part of the constitutive elements of empirical time? One could certainly have a relationalist reading on which these determinations are included among the determinations that constitute empirical time, but on my reading they are not included. Arguably, it is a necessary condition on the objects whose temporal determinations constitute the same time that they all mutually causally interact with each other. After all, as we learn in Kant’s discussion in the chapter on the analogies of experience, the causal relations between empirical objects, including their mutual causal interactions, are the grounds for their objective temporal determinations. But, as we will see in section 3.6, there are reasons to assume that empirical selves do not mutually causally interact with empirical objects, on Kant’s view. To be sure, we present our existence as determined in time, that is, we present our empirical self as having temporal determinations. But Kant seems to think of this kind of representation as depending on us relating our inner determinations to the objective temporal order of empirical objects; see B275 276. And so it seems plausible to say that empirical time is constituted by the temporal determinations of empirical objects, while the temporal determinations of empirical selves are parasitic on the objective temporal order thus constituted. ⁸¹ Talk about various properties of empirical space and time has its natural home in the framework of substantivalism. On the relationalist reading, this talk is a bit misleading and should be understood as a convenient representational ploy. According to relationalism, all there is to space and time are the spatial and temporal determinations of empirical objects, but the nature of these determinations can efficiently be represented by talking about space and time as if they were self-subsisting things with various kinds of properties.

      

147

more precisely, in their presentational contents.⁸² All determinate spaces and times, including empirical space and time, are necessarily contained in pure space and time, and thus also fully mind-dependent.⁸³ Empirical space and time instantiate pure space and time at the empirical level of reality; they are transcendentally ideal and empirically real. Pure space and time as such, that is, as general ordering schemes in accordance with which all particular spatial and temporal determinations are necessarily represented in experience, are transcendentally ideal but not empirically real; they do not exist at the empirical level of reality.⁸⁴ Regarding the question of whether empirical space conceived of as materialized instance of pure space is infinite, the answer depends on whether experience includes the representation of an ether, understood as some kind of subtle matter, that is distributed throughout space to infinity. Kant’s views about such an ether seem to have evolved over time. There are some remarks in his pre-critical writings that show him to be sympathetic to the assumption of the existence of an ether.⁸⁵ There is no mention of the ether in the Critique, but in the Metaphysical Foundations Kant considers it as part of what he seems to regard as a not entirely implausible account of the gravitational attraction of heavenly bodies.⁸⁶ In the Opus Postumum, the ether enters center stage in Kant’s developing views about cosmology and the generation and constitution of matter, where he describes it (or the “heat substance” [Wärmestoff], as he also calls it) as “the hypostasized space itself, so to speak, in which everything moves” (OP, 21:224), and as “that which makes space into an object of the senses” (OP, 22:115) and attempts a deduction of the ether’s existence based on a priori principles by showing it to be a material condition for the possibility of experience.⁸⁷ If experience includes the representation of an infinitely distributed ether, empirical space is infinite. If experience does not include such an ether representation, empirical space, just like empirical time, is indeterminate as far as its extension is concerned because, according to Kant’s solution to the first antinomy discussed in chapter 2, the size of the empirical world, understood as the sum total of all empirical objects, is neither finite nor infinite in space or time but rather indeterminate.⁸⁸

⁸² See B39/A25: “For, first, one can only represent a unified space, and if one talks about many spaces, one understands by that only part of one and the same all-unified space. These parts also cannot precede the allencompassing space as its constituents, as it were (through which a composition were possible), but can only be thought in it. It is essentially unified, the manifold in it, and thus the general concept of it, rests merely on limitations. From this it follows that with respect to it [space] an intuition a priori (that is not empirical) grounds all concepts of it.” Also see B39 40: “Space is represented as an infinite given magnitude. Now, one must conceive of every concept as a representation that is contained in an infinite multitude of different possible representations (as their common mark), and thus contains them under itself; but no concept, as such, can be conceived in such a way as if it contained an infinite multitude of representations in itself. Nevertheless, space is thought in this way (for the parts of space to infinity are simultaneous). Therefore, the original representation of space is an intuition a priori and not a concept.” To prevent confusion, also note that, despite their generality as ordering schemes and despite having instances, pure space and time are still particulars, on Kant’s view, not abstract universals. A proper way of characterizing the difference between universals and particulars on Kant’s view would be to say that universals have instances that contain them, while particulars either have no instances at all or have instances that are contained in them. ⁸³ This claim and this inference will be considered in detail in chapter 4. ⁸⁴ Remember, though, that pure space and time are not empirically ideal either, given that experience represents them, at least in part, and presents them as mind-independent. See notes 29 and 54. ⁸⁵ See NTH, 1:307; MP, 1:487; BDG, 2:113. ⁸⁶ See MAN, 4:563 564. ⁸⁷ See OP, 21:218–224; 378–383; 424–428; 545; OP, 22:211–215; 593–596; 612–613. ⁸⁸ If there were an ether, the empirical world, understood as the sum total of all empirical objects, would still be indeterminate with respect to its extension, but empirical space would be infinite—at least, this would be so on the assumption that the ether is not an empirical object. (Why would we assume this? One reason just is that

148

 ’      

As already anticipated in section 3.4.1, in my assessment, Kant is a relationalist about empirical space and time of the kind just described. In his comments on the antithesis of the first antinomy, Kant expresses his agreement with the Leibnizians over the rejection of the existence of substantival time and space by way of what is a fairly clear reference to the Leibniz–Clark correspondence, which contains Leibniz’s most celebrated attack on Newtonian substantivalism and an exposition of his own relationalist theory of space and time.⁸⁹ I am quite happy with the latter part of this opinion of the philosophers of the Leibnizian school [i.e., that it is impossible for there to be an absolute time before the beginning of the world or an absolute space outside of the actual world]. Space is merely the form of outer intuition but not an actual object that can be outwardly intuited and not a correlate of appearances but the form of appearances themselves. Thus, space [taken] absolutely (for itself alone) cannot occur as something determining in the existence of things, since it is not an object but merely the form of possible objects. Thus, things, as appearances, indeed determine space, i.e., among all possible predicates of it (size and relation) they bring it about that these or those belong to actuality; but, the other way around, space, as something that exists for itself, cannot determine the actuality of things with respect to size or figure, since in itself it is nothing actual. (B459+461/A431+433)⁹⁰

This is not only a rejection of substantivalism about space and time understood as things in themselves. This is a rejection of substantivalism at any level of reality and a rather clear endorsement of relationalism about empirical space and time. Another reason that speaks for the relationalist and against the substantivalist reading has to do with the question of how to think about the grounds for the existence of empirical space and time. As indicated in section 2.5.3, one of the crucial factors that ensures that appearances exist from the point of view of fundamental ontology is that they are grounded in things in themselves in a properly direct way, in contrast to the intentional objects of dreams, hallucinations, and fictions, which are not so grounded. Kant’s emphasis against Eberhard, that on his view empirical space and time have things in themselves as their objective grounds, suggests that the grounding of empirical space and time in things in themselves plays an equally important role with respect to their genuine existence. On the relationalist reading, empirical space and time are indeed grounded in things in themselves in that the spatial and temporal determinations of empirical objects depend not only on the forms of sensibility but also on what things in themselves are around and affect us. On the substantivalist reading, by contrast, since empirical space and time’s dynamical properties and relations to empirical objects are taken to be merely contingent and their remaining otherwise we would have trouble accommodating the central claim of Kant’s solution to the first antinomy that the extension of the empirical world is indeterminate.) ⁸⁹ The Leibniz–Clarke correspondence was published in German translation by Heinrich Köhler in 1720 and re-issued in 1740 by Caspar Jacob Huth. Kant was quite certainly familiar with it. ⁹⁰ Also see B457/A429, note: “Empirical intuition is thus not composed from appearances and space (perception and empty intuition). One is not the synthesis correlate of the other but merely connected in one and the same empirical intuition as its matter and form. If one wants to posit one of these two elements apart from the other (space apart from all appearances), all sorts of empty determinations of outer intuition result from this that are still not possible perceptions, e.g., motion or rest of the world in infinite empty space, a determination of the relation of both with respect to one another that can never be perceived and thus is also the predicate of a mere object of thought.”

      

149

properties depend on the forms of sensibility alone, empirical space and time turn out not to be grounded in things in themselves in a sufficiently robust sense. The substantivalist thus cannot make sense of Kant’s remark to Eberhard nor can he justify why empirical space and time as conceived by him should be included in the august company of fully mind-dependent, yet genuinely existing entities. The suggestion to move to a modified version of substantivalism and take the dynamical properties, understood as dependent on things in themselves, to be essential to empirical space and time after all, does not represent a satisfactory solution either. For the distinctive feature of the substantivalist reading is precisely that space and time are conceived of as independent of empirical objects (as well as of empirical selves) and as capable of existing even if there are no empirical objects (or empirical selves). But this means that a proponent of this modified version of substantivalism would be committed to the view that the grounding of empirical space and time in things in themselves obtains even in possible worlds in which there are no appearances and, hence, no affections of sensibility by things in themselves or at least, no affections of the usual kind that produce sensations. (Substantival space and time themselves are not sensible.) This would be a grounding of an entirely different kind compared to the grounding of appearances in things in themselves, which essentially depends on the affection of sensibility by things in themselves. What this other kind of grounding could possibly consist in remains a mystery. In a similar vein, with respect to the related task of identifying conditions for the existence of empirical space and time at the empirical level, the relationalist has the significant advantage over the substantivalist that, since, according to his view, empirical space and time simply are the sum total of all spatial and temporal determinations of empirical objects, respectively, and since empirical objects are appearances, he can just point to our previously (in section 2.5.3) identified conditions for the existence of appearances at the empirical level. Once the existence of empirical objects is accounted for, the existence of all of their determinations and, with them, the existence of empirical space and time is accounted for as well.⁹¹ Since the substantivalist takes empirical space and time to be capable of existing even if there are no empirical objects (and no empirical selves), he cannot adopt the same ploy and must come up with a special story, within Kant’s framework, of how experience presents empirical space and time as actually existing. The prospects for success in this project are dim. As we saw in our earlier discussion, the way in which empirical thinking presents objects as actually existing is by presenting them as causally connected to actual perceptions and sensations, and, more specifically, by presenting them either as the proximate causes of actual sensations or as causally connected to objects that it presents as the proximate causes of actual sensations. But, in contrast to empirical objects, empirical space and time as conceived by the substantivalist are neither perceived nor perceivable, not even in principle, and, accordingly, are also not ever presented as the proximate causes of actual sensations.⁹² It is true that, by ascribing to empirical spacetime an inertial structure, experience in effect presents it as causally connected to certain perceived empirical objects, such as the rotating water in Newton’s

⁹¹ Recall that, on Kant’s view, empirical objects essentially are in space and time. ⁹² As before, by ‘in principle perceivable’ I mean that the appearance would be perceived if an observer were properly located with respect to it in space and time or if our sense organs had greater acuity, range, or scope.

150

 ’      

bucket whose surface exhibits a concave shape. But this circumstance is of no help to the substantivalist in the present situation because the indicated representation does not force a substantivalist conception of space and time on us. Empirical spacetime as conceived by the relationalist can perform the role of being the cause for the observed inertial phenomena just as well. That is, by presenting empirical spacetime as causally connected to various observed inertial phenomena experience does not present empirical spacetime as conceived by the substantivalist as existing; it merely presents empirical spacetime as existing. For all these reasons, I thus conclude that Kant is a relationalist about empirical space and time. Accordingly, we can add the following general and strong versions of TI2-empirical to our list of theses of transcendental idealism, the latter of which, in its (b)-component, tells us more concretely what Kant takes empirical space and time to be. (TI2-empirical) (a) Empirical space and empirical time are neither things in themselves in the transcendental sense nor constituted by determinations of things in themselves in the transcendental sense; rather (b) empirical space and empirical time are fully minddependent. (TI2-empirical-strong) (a) Empirical space and empirical time are neither things in themselves in the transcendental sense nor constituted by determinations of things in themselves in the transcendental sense; rather (b) empirical space and empirical time are constituted by the spatial and temporal determinations of appearances in the transcendental sense, respectively, and thus are fully mind-dependent.

3.4.4 Kant’s Theory of Space and Time Combines Substantivalist and Relationalist Elements It is important to highlight that the special kind of relationalism about empirical space and time that I am attributing to Kant differs significantly from ordinary forms of relationalism. The rich spatial and temporal determinations of empirical objects to which empirical space and time can be reduced on Kant’s view go far beyond the spatial and temporal relations that usually figure in relationalist accounts of space and time. The ordinary relationalist aims to construct all of the properties of space and time by appealing only to such meager spatial and temporal relations of empirical objects as being next to, being in between, being equidistant from, being before/after, and being simultaneous with. By contrast, Kant can help himself to the indicated rich spatial and temporal determinations that already embody all of the structural and dynamical properties of empirical space and time—including such ‘global’ properties as space’s three-dimensionality and orientability—because, according to his transcendental idealist conception of the relation between objects and our representations, some of our representations, including the pure intuitions of space and time as well as our a priori representations of the laws of mechanics, are both prior to empirical objects and crucially involved in their constitution. This is a reflection of a more general feature of Kant’s theory of space and time overall, namely, that it has some important elements in common with substantivalism. It is a central tenet of substantivalism that space and time are ontologically prior to empirical objects in the sense of being among their existence conditions. The ordinary relationalist categorically denies this claim, but Kant can agree with it since, on his view, pure space and

      

151

time are ontologically prior to empirical objects in the indicated sense.⁹³ Given that, on Kant’s view, pure space and time do not genuinely exist, the way in which their ontological priority manifests itself on his view differs slightly from the way it manifests itself on the substantivalist view. But there is still enough substantive agreement to align Kant and substantivalism in opposition to ordinary relationalism on the question of whether and in what way space and time are ontologically prior to empirical objects. For ease of communication, we will say that a property P is space-infected (time-infected) if, and only if, it is essentially bound up with space (time) in that objects can have P only if they are in space (time); a determination is space-infected (time-infected) if, and only if, it is an instance of a space-infected (time-infected) property. This terminology is useful because the determinations of empirical objects that I have been counting among their spatial and temporal determinations—their extension, shape, three-dimensionality, spatial size, spatial locations with respect to and distances from other empirical objects, their duration, and the temporal locations of their states with respect to and the temporal distances of their states from each other and the states of other empirical objects—are not alone in having a spatial or temporal character, as one might put it. For example, being moveable is not a spatial or temporal property, according to our list, but it is both space-infected and time-infected in sense just specified. No object is moveable unless it is in space and time. Using this terminology, for the substantivalist, the ontological priority of space and time compared to empirical objects manifests itself in, among other things, that (a) space and time can exist even if there are no empirical objects in them, while empirical objects cannot exist unless they are in space and time, and (b) empirical objects have all of their space-infected and time-infected determinations, including their spatial and temporal determinations, some of which are essential to them, partly in virtue of being in space and time, while space and time do not have any of their essential determinations in virtue of or partly in virtue of the empirical objects in them. For Kant, the ontological priority of pure space and time compared to empirical objects manifests itself in, among other things, that (i) we can represent pure space and time without any empirical objects in them and prior to ever having represented empirical objects in them, while we cannot represent empirical objects unless we represent them in pure space and time and rely for their representation on logically prior representations of pure space and time, and (ii) empirical objects have all of their space-infected and time-infected determinations, including their spatial and temporal determinations, some of which are essential to them, partly in virtue of being represented in, and thus being in, pure space and time, while pure space and time do not have any of their essential determinations in virtue of or partly in virtue of the empirical objects represented in them or in them. Note that the logical priority of our representations of pure space and time compared to the representations of empirical objects, on the transcendental idealist account, directly translates into an ontological priority of pure space and time compared to empirical objects because, according to transcendental idealism, pure space and time as well as empirical objects are intentional objects that exist ‘in our representations.’ Accordingly, Kant’s claim (i) can also be formulated by saying that pure space and time can have being even if there are no empirical objects in them, while ⁹³ See Bxxvi–xxvii: “That space and time are only forms of sensible intuition, thus also only conditions of the existence of things as appearances . . . is proved in the analytic part of the Critique.” See A24/B38–39; A31/B46; B323/A267; B469/A441. Also see V-Met-L1/Pölitz, 28:179.

152

 ’      

empirical objects can have being only if they are in space and time, a formulation that underlines the similarity with claim (a) of the substantivalist. The ordinary relationalist, by contrast, disagrees with all of the indicated claims (a, b, i, and ii).⁹⁴ In this context, it is also useful to highlight again that, although pure space and time as such do not exist at the empirical level of reality, it is still legitimate to say that empirical objects are not only in empirical space and time but also in pure space and time—which is what I just said when I described the way in which the ontological priority of pure space and time compared to empirical objects is to be cashed out on the Kantian account. As Kant emphasizes in the so-called space arguments and time arguments in the Transcendental Aesthetic (to be examined in section 4.2.2.2), in which he shows that our original, most fundamental representations of space and time are pure intuitions, we can represent determinate spaces and times and, hence, objects with determinate shapes and determinate durations in no other way than by representing them as ‘cut-out’ parts of the one space and the one time that are the intentional objects of our a priori intuitions, respectively. This is an expression of the crucial difference between concepts and intuitions, already noted in section 3.4.3, that the instances of concepts are contained under them, while the instances of intuitions are, literally, contained in them, or, more precisely, in their presentational contents. Accordingly, due to the fact that we necessarily represent all determinate spaces and times by presenting them as parts of pure space and time, pure space and time are, at least partly, represented in experience, and empirical objects are presented as being in them. And this holds despite the fact that experience does not present pure space and time as existing, which it fails to do because it does not present pure space and time as causally connected to any actual sensations.⁹⁵ If empirical objects and space and time were things in themselves it would be highly implausible to say both that (1) empirical objects genuinely exist and are in space and time, and (2) space and time are ontologically prior to empirical objects but do not genuinely exist. But since empirical objects, as well as space and time, are not things in themselves but intentional objects, there is nothing implausible about maintaining both (1) and (2)—provided ‘space’ and ‘time’ in this context are understood as ‘pure space’ and ‘pure time.’ Since pure space and time are forms of sensibility and, thus, part of the formal conditions for the possibility of experience, they are ontologically prior to empirical objects, the intentional objects of experience. And since, more specifically, representing empirical objects requires presenting them as being in pure space and time, empirical objects are in pure space and time. But since in order for an intentional object to genuinely exist it also must be grounded in things in ⁹⁴ Ordinary relationalists need not deny that we can represent space and time without any objects in them. Even if we first acquire our concepts of space and time through the observation of empirical objects and their spatial and temporal determinations, once we have acquired these concepts through comparison, abstraction, and reflection, we can use them to represent space and time without any objects in them. But ordinary relationalists (as Kant understands them) deny that we can represent space and time prior to ever having represented empirical objects in space and time. Relatedly, ordinary relationalists need not deny that we cannot represent empirical objects without representing them in space and time. It is generally conceded all around, for example, that empirical objects, or bodies, are necessarily extended, which means, in turn, that in order to represent empirical objects we must represent extension, which amounts to representing them in space. (This also holds even if there is only one object in existence, for example, a tomato. In that case, space would be exhausted by the volume taken up by the tomato.) But ordinary relationalists deny that we cannot represent empirical objects unless we rely on logically prior representations of space and time. For example, for an ordinary relationalist (as understood by Kant), our perceptions of empirical objects do not rely on any logically prior representations of space and time. ⁹⁵ As we learned in section 2.5.3, the only way in which empirical thinking can present objects as actually existing, according to Kant’s analysis, is by presenting them as causally connected to actual sensations.

      

153

themselves, pure space and time do not genuinely exist, in contrast to empirical objects, even though the former are included among the existence conditions for the latter.⁹⁶ That Kant’s critical theory of space and time comprises both substantivalist and relationalist elements reflects the path that he traveled to arrive at this theory. In his younger days, Kant was a relationalist of the ordinary variety and believed that empirical objects were ontologically prior to space and time.⁹⁷ But from 1768 onwards he consistently subscribes to the view that space and time are ontologically prior to empirical objects although the sense in which this ontological priority is spelled out changes slightly over time as his theory of space and time evolves. He first came to see their ontological priority, or, more precisely, the ontological priority of space, through his notorious argument from incongruent counterparts as presented in the essay “On the first ground of the distinction of regions in space.” Incongruent counterparts are objects that are similar in shape and equal in size but cannot be put into each other’s places by rigid motions (understood as motions through the three-dimensional space in which the relevant objects exist), as for instance a left and a right hand. There is no need for us to examine this argument in detail here.⁹⁸ Suffice it to say that its main upshot in Kant’s eyes is that the orientation of a spatial object, for example, the rightness or leftness of a hand, cannot be explained in ordinary relationalist terms but only by taking into account its relation to space understood as ontologically prior to all objects in it.⁹⁹ In the mentioned essay, Kant takes this argument to establish the existence of substantival space conceived of as a thing in itself that exists independently of the objects in it. But soon after and before the publication of his Inaugural Dissertation in 1770, he convinces himself that the conception of space and time as selfsubsisting things in themselves or as properties of God is untenable. As evidenced by his remarks in the Critique, Kant seems to have come to regard the Newtonian view that space and time are infinite quasi-substances that can exist on their own without any objects in them as simply absurd.¹⁰⁰ He also points out that if we insist that space and time are ⁹⁶ So, if A is ontologically prior to B, A is not necessarily ontologically more fundamental than B. (Arguably, genuinely existing intentional objects are ontologically more fundamental than intentional objects that do not genuinely exist.) For the claim that the critical Kant regards space and time, understood as forms of sensibility, as conditions for the existence of the objects in them, see B69: “Thus I do not say that that bodies merely illusorily appear to be outside of me, or my soul merely illusorily appears to be given in my self-consciousness, when I claim that the quality of space and time, according to which, as condition of their existence, I posit both, lies in my manner of intuiting and not in these objects in themselves.” See V-Met/Volckmann, 28:437: “ . . . these two [space and time] appear to be the sovereign conditions of the existence of things and are so in fact, for this is the form of our sensibility . . . ” Also see the references in note 93. ⁹⁷ See WSLK, 1:23: “It is easy to demonstrate that there would be no space and no extension if substances did not possess a force to act outside of themselves. For without this force there is no connection, and without a connection there is no order, and without an order, finally, there is no space.” See NTH, 1:308: “Attraction is without doubt a property of matter that is equally far extended as [the property of] coexistence, which brings about space in that it connects substances through mutual dependencies . . . ” See PM, 1:481: “ . . . however, space is not a substance but merely the special appearance of the outer relations of substances . . . ” See BG, 2:71: “I doubt that anybody has ever correctly explained what space is. Without getting into this, I am certain that where it is there must be outer relations . . . ” ⁹⁸ A number of useful papers about Kant’s argument are collected in Van Cleve and Frederick 1991. ⁹⁹ See GUGR, 2:378, 381 383, esp. 383: “From this it is clear that the determinations of space are not consequences of the situations of the parts of matter with respect to each other, but that the latter are consequences of the former, and thus that there can be differences in the quality of bodies, and real differences at that, that relate solely to absolute and original space, since only through it is the relation of corporeal objects possible . . . ” ¹⁰⁰ See B56/A39: “For if they opt for the first choice [to regard space and time as subsisting things in themselves] (which is usually the party of the mathematical researchers of nature) they must assume two eternal and infinite self-subsisting non-things (space and time), which are there (without there being something real),

154

 ’      

conditions for the existence of all things that are in space and time and conceive of space and time as pertaining to things in themselves, we would have to hold that space and time are conditions for the existence of God, which conflicts with the view that God’s existence is unconditioned.¹⁰¹ With respect to the assumption that space and time are properties of God, Kant appears to hold that it necessarily leads to Spinozism, which must be avoided at all costs.¹⁰² But this means that if space and time were to exist in the realm of things in themselves they would have to be constituted by the relations between things in themselves and their states, as the Leibnizians claim, which, however, conflicts with the main upshot of the argument from incongruent counterparts that space and time are ontologically prior to all objects in them. The way out of this difficulty—as well as several other difficulties that Kant was also struggling with at the time—is to let go of the presupposition that space and time are things in themselves or forms of things in themselves and recognize them as forms of sensibility instead, which are necessary conditions for the possibility of experience and, thus, for the existence of the objects of experience.¹⁰³ This move allows Kant not only to dispose of the Newtonian, the Spinozist, and the Leibnizian transcendental realist views in one fell swoop but also to hold on to the claim that space and time are ontologically prior to all objects in them, while at the same time resurrecting the relationalist insight from his ‘youth’ that, as far as genuine existents are concerned, all there is in the empirical world are empirical objects (and maybe an ether) and their spatial and temporal determinations.¹⁰⁴ This account of Kant’s path to the doctrine that space and time are forms of sensibility can instructively be contrasted with Kant’s reconstruction of Berkeley’s path to idealism, which reveals another facet of his criticism of the “good bishop.” The two paths have important similarities, but in Kant’s story about Berkeley the protagonist takes a wrong turn at the final critical juncture. According to Kant, Berkeley correctly conceives of space as a condition for the existence of all objects in it and correctly regards the view that space is a self-subsisting thing in itself or a property of a thing in itself as untenable. But since he failed to recognize the third option of understanding space to be a form of sensibility, he was inevitably led to the conclusion that space and all objects in it are mere imaginings.

only in order to comprise everything real.” See B70–71: “For if one regards space and time as qualities that, according to their possibility, would have to be encountered in things in themselves and reflects on the incoherencies in which one then entangles oneself, in that two infinite things that are not substances nor something that really inheres in substances but still something existing, indeed the necessary condition for the existence of all things, still remain even though all existing things are removed, one cannot begrudge the good Berkeley that he degraded bodies to mere illusion.” ¹⁰¹ See B71–72: “ . . . for as conditions of all existence in general they [space and time] also would have to be conditions of the existence of God. Unless one wants to turn them into objective forms of all things, nothing remains but to turn them into subjective forms of our outer as well as inner intuition . . . ” See KpV, 5:101–102. ¹⁰² See V-Met-L2/Pölitz, 28:567: “If I regard space as a being in itself, Spinozism is irrefutable, i.e., the parts of the world are parts of the deity.” See V-Met-K3/Arnoldt, 29:977: “Yes, if space and time were properties of things in themselves, they would have to be properties of God . and the hypothesis of Spinozism indeed consisted in that one regarded space and time as things in themselves and as properties of God.” See V-Met/Dohna, 28:666; V-MetK2/Heinze, 28:803; V-Met-K3/Arnoldt, 28:829. ¹⁰³ Another such difficulty that played an important role in Kant’s journey to critical idealism for whose solution the classification of space and time as forms of sensibility is the key is the problem of how to account for the applicability of geometry and arithmetic to empirical objects, given that they are a priori sciences. We will talk about how this problem is solved by regarding space and time as forms of sensibility in section 4.2. ¹⁰⁴ This path to transcendental idealism about space and time is reflected in the version of Kant’s main ideality argument in the Critique that he presents for the case of time. See B49/A32 33; also see note 36, chapter 4. The version of the argument that he presents for the case of space is slightly different. We will look at this argument in detail in section 4.2.

      

155

“ . . . the second kind of idealism is the dogmatic idealism of Berkeley who declares space, with all things to which it attaches as an inseparable condition, for something that is in itself impossible, and thus also all things in space for mere imaginings. Dogmatic idealism is unavoidable if one regards space as a property that is supposed to pertain to things in themselves; for in that case it, together with everything for which it serves as a condition, is a non-thing. (B274)¹⁰⁵

Kant’s classification of space as a form of sensibility saves him from this error and allows him to regard empirical space as well as the objects in space as genuine existents, despite the ontological priority of pure space compared to the objects in space and the nonexistence of pure space as such.

3.4.5 Empirical Objects Genuinely Have Spatial and Temporal Determinations ER2-ii, the second half of the second core thesis of empirical realism, according to which space and time are empirically mind-independent, and empirical objects are in space and time and, thus, have spatial and temporal determinations, also prompts an important cluster of questions that deserve brief consideration. As all theses of idealism and realism collected in this chapter, this thesis is to be understood as formulated from the point of view of fundamental ontology. But a skeptical reader might wonder whether, in the framework of the interpretation presented here, on which Kant is read as a genuine idealist about empirical objects, this claim should not rather be understood as stated from the empirical point of view. While conceding that it is correct to say that, from the empirical point of view, empirical objects have spatial and temporal determinations, the skeptical reader might argue that, from the point of view of fundamental ontology, all that can be legitimately asserted on the proposed interpretation is that empirical objects are represented to have spatial and temporal determinations but not that they genuinely have them. And the skeptical reader might continue, this constitutes a problem for my interpretation because it conflicts with Kant’s empirical realism.¹⁰⁶ Based on the textual evidence cited in our discussion in the foregoing sections and in the corresponding footnotes, I concur that Kant’s empirical realism centrally includes the claim that empirical objects genuinely have spatial and temporal determinations, understood as articulated from the point of view of fundamental ontology, and so I agree with the imagined skeptical reader that it is an important desideratum for a satisfactory interpretation of his critical idealism that this claim can be properly accommodated. The interpretation presented here satisfies this demand, though, despite the doubts of our imagined skeptical reader. To be sure, it is an integral part of my reading that, from the point of view of fundamental ontology, empirical objects have all of their spatial and temporal determinations in virtue of being represented by finite minds to have them, just as it is an integral ¹⁰⁵ Also see B70 71; B519/A491. ¹⁰⁶ In a similar vein, one finds the objection in the literature that, on the classic two-world view, statements about appearances must be understood as claims about how things merely seem to us to be, which founders over Kant’s empirical realism, or as claims about our representations, including, for example, the claim that our representations are spatially extended, which is absurd. See Prichard 1909, 71–100.

156

 ’      

part of my reading that, from the point of view of fundamental ontology, empirical objects exist and have all of their other determinations in virtue of being represented by finite minds in this way. This is just what it means to say that they are fully mind-dependent. But all of this is perfectly compatible with—and, indeed, requires—also ascribing the claim to Kant that, from the point of view of fundamental ontology, empirical objects have spatial and temporal determinations, as opposed to merely being represented to have them, or merely appearing to have them. To see this, first, recall that, according to my interpretation, despite being fully minddependent, empirical objects exist from the point of view of fundamental ontology because they not only conform to Kant’s formal conditions of proper objecthood but also are grounded in things in themselves in a properly direct way. This is in contrast to other intentional objects, such as Frodo or the dragon from my dream, which do not satisfy all of Kant’s formal conditions of proper objecthood and are not grounded in things in themselves in the right way, and thus, from the point of view of fundamental ontology, can be described only as pseudo-existing. Now, since the objects that exist at the empirical level also exist from the point of view of fundamental ontology, it stands to reason to count the facts that obtain at the empirical level of reality with respect to these genuine existents also as facts from the point of view of fundamental ontology. And since it is a fact at the empirical level that empirical objects have spatial and temporal determinations, this is also a fact from the point of view of fundamental ontology. Of course, from this fundamental point of view, the fact in question is a special kind of fact, namely, a mind-dependent fact, just as the existence of empirical objects is a special kind of existence, namely, a minddependent existence. But it is still a kind of genuine fact, just as it is a kind of genuine existence.¹⁰⁷ I take it that this is part of what Kant has in mind in the following passage: By contrast, our transcendental idealism allows that the objects of outer intuition really are exactly as they are intuited in space, and in time all changes [really are] as inner sense represents them. But this space itself, together with time, and, together with both, all appearances still are in themselves not things but nothing but representations and cannot at all exist outside of our mind. (B520/A492)¹⁰⁸

Second, even if we set aside for the moment that empirical objects genuinely exist, the worry of our imagined critic can be put to rest. For, given the kind of conception of intentional objects that we have been working with, it is plausible to hold that, from the point of view of fundamental ontology, all intentional objects genuinely have the determinations that they are presented to have in the representations that ontologically specify them. This includes empirical objects, as intentional objects of experience, but also fictional

¹⁰⁷ We have not defined explicitly what it means for a fact to be mind-dependent. Roughly put, a minddependent fact is a fact that obtains in virtue of being represented by a finite mind. I leave it as an exercise for the reader to spell this out more precisely, using our discussion in section 2.3 as a guide. ¹⁰⁸ See as well B43/A27: “The proposition ‘all things are next to one another in space’ is valid under the restriction that these things are taken as objects of our sensible intuition. If I add the condition to the concept and say ‘all things, as outer appearances, are next to one another in space,’ then the rule is valid universally and without any restriction.” Also see B55/A38. In the passage quoted in the main text, when Kant says that all appearances are not things but merely representations, he obviously means that they are not things in the transcendental sense, or with respect to the transcendental level of reality, precisely because, with respect to that level, they are fully minddependent. See the discussion in section 3.2, and, in particular, note 18.

      

157

characters and the intentional objects of perceptions, illusions, hallucinations, and dreams. Take Frodo and his determination of having hairy feet or the dragon from my dream and his determination of having a scaly back or the rose that is the intentional object of my current perception and its determination of being red. Our imagined skeptical reader would judge that, from the point of view of fundamental ontology, Frodo does not have hairy feet but is merely represented to have them by Tolkien’s novels, the dragon does not have a scaly back but is merely dreamed by me to have one, and the perception-immanent rose is not red but merely appears to me to be so in my perception. As I see it, this assessment betrays a misunderstanding of the nature of intentional objects. The proper way to express the ontologically light-weight character of Frodo, my dragon, and the perception-immanent rose from the point of view of fundamental ontology is to say, not that they do not genuinely have the determinations that they are represented to have, but that they are fully mind-dependent and that they do not genuinely exist.¹⁰⁹ Given that they are intentional objects, and given how The Lord of the Rings, my dream, and my perception present them, it is correct to say, even from the point of view of fundamental ontology, that Frodo has hairy feet, the dragon has a scaly back, and the perception-immanent rose is red.¹¹⁰ To be sure, all of these determinations are mind-dependent, but that does not take away from the fact that they are genuinely had by the intentional objects that are presented to have them in the representations that ontologically specify them.¹¹¹ Similarly, given that empirical objects are intentional objects, and given how experience presents them, it is correct to say, even from the point of view of fundamental ontology, that empirical objects genuinely have spatial and temporal determinations. As I read Kant, this understanding of intentional objects and their determinations is presupposed in the secondary quality analogy that he uses to illustrate certain aspects of his transcendental idealism. As I will argue in detail in section 5.4, it is part of the intended meaning of the analogy that, just as secondary qualities can be attributed, without any qualifications, to empirical appearances, for example, redness to the intentional object of a perception that is caused by a rose, spatial and temporal determinations can be attributed, without any qualifications, to transcendental appearances, for example, extension to an opossum as which some unknown thing in itself (or things in themselves) X appears. What

¹⁰⁹ If you are thinking ‘But they are presented as existing!,’ hold your horses. We will get to this complication presently. ¹¹⁰ We find a parallel situation when we turn to non-actual possibilia. Suppose we are talking about a certain merely possible raccoon. The proper thing to say is, not that the raccoon does not have a pointy snout but only possibly has a pointy snout, but that the raccoon, together with his pointy snout, does not actually exist but only possibly exists. ¹¹¹ Similarly, the same properties can be instantiated by objects at different levels of reality (where the instantiation relation is also the same in each case). These instances differ only with respect to their degree of reality as determined by their location in the hierarchy. (See section 2.2.) For example, the sole difference between Frodo’s determination of having feet and my determination of having feet, according to critical idealism, is that my determination is mind-dependent and genuinely exists while Frodo’s determination is mind-dependent and only pseudo-exists. But they are instances of the very same property of having feet. Or the determination of being a ground of a thing in itself that affects us instantiates the same property of being a ground as the determination of being a ground of an empirical object that causes a change in another empirical object’s state of motion. These determinations, both of which genuinely exist, differ only in that the former is mind-independent, while the latter is mind-dependent. However, it is important to emphasize that not all properties can be instantiated at all levels of reality. Since space and time are (nothing but) forms of sensibility, all spatial determinations and temporal determinations and, more generally, all determinations that are space-infected or time-infected necessarily are mind-dependent. Accordingly, no space-infected and no time-infected property can be instantiated at the transcendental level of reality.

158

 ’      

would be a mistake is to say that empirical objects have secondary qualities, or that things in themselves have spatial and temporal determinations. In such a situation it would be appropriate to point out that the public rose is not red but is merely perceived to be red, and that X is not extended but merely appears as an extended opossum. But as long as we restrict our attributions of secondary qualities to empirical appearances and of spatial and temporal determinations to transcendental appearances, no qualification and correction is called for, not even from the point of view of fundamental ontology.¹¹² But wait. Experience presents empirical objects as mind-independent. So, if it is indeed the case, as just argued, that even from the point of view of fundamental ontology all intentional objects genuinely have all of the properties that they are presented to have in the representations that ontologically specify them, it seems as if we are now committed to the claim that empirical objects are mind-independent from the point of view of fundamental ontology. But this claim is flat-out inconsistent with the first core thesis of transcendental idealism that, from the point of view of fundamental ontology, empirical objects are fully mind-dependent. Even worse, it also seems to follow that Frodo exists and is mind-independent from the point of view of fundamental ontology, since that is how The Lord of the Rings presents him. This worry can be put to rest as follows. First, in the foregoing discussion, I did not claim that, from the point of view of fundamental ontology, all intentional objects genuinely have all properties that they are presented to have in the representations that ontologically specify them. I merely claimed that they genuinely have all determinations that they are presented to have in the relevant representations. But existence, be it mind-dependent or mind-independent existence, is not a determination. As Kant famously puts the point in his critique of the ontological argument for the existence of God: Existence is obviously not a real predicate, i.e., a concept of something that could be added to the concept of a thing. It is merely the positing of a thing or of certain determinations in themselves. . . . Thus if I think a thing through whatever and however many predicates I want (even in thoroughgoing determination), by still adding that the thing exists not the least is added to the thing. For otherwise not the very same but more would exist than what I had thought in the concept, and I could not say that the very object of my concept exists. (B626–628/A598–600)

¹¹² With respect to the related objections to the classic two-world view mentioned in note 106, the alleged problem that claims about appearances must be understood as claims about our representations could arise only on those versions of the classic two-world view on which appearances are identified with representations, understood as mental states that do the representing. Setting aside that it is not obviously absurd to say that representations are extended—Hume, for one, does not find it absurd, for example—if appearances are conceived of as intentional objects of representations, as on my view, the problem evaporates. Representations need not be extended in order to be able to represent extended objects. The objection that, on the classic two-world view, statements about appearances must be understood as claims about how things merely seem to be rests on a confusion that fuels the two-aspect view in general, as we will discuss below. Statements about appearances are not claims about the things that ground them; they are claims about the things that are those appearances. Suppose that a thing in itself, X, appears to us as an opossum. In this situation, if I make a statement about the appearance, for example, that it is extended, I make a claim about the opossum, not X, even though the opossum is what X appears to us as. And the opossum does not merely appear to be extended; the opossum is extended. Of course, it is also legitimate to say that X appears as an extended opossum. This is indeed a statement about how things appear to us. But in this case, I am not talking about appearances or empirical objects but about things in themselves, leaving Kant’s empirical realism uncompromised.

 , ,   - 

159

Second, the properties of existing and of being mind-independent are ontological properties of objects, which characterize their ontological status and determine their ‘location’ in Kant’s scheme of reality and whose proper description, accordingly, varies depending on one’s point of view. The determinations of an object, for example, Frodo’s properties of being extended and of having hairy feet, pertain to it irrespective of its ontological status and thus can be ascribed to it without any qualifications from any possible point of view in Kant’s ontological scheme, including the point of view of fundamental ontology. By contrast, the ascription of ontological properties to an object from the point of view of fundamental ontology requires the specification of the level of reality with respect to which the ascription is to be evaluated. This is why we must distinguish between a transcendental and an empirical sense for all concepts that refer to ontological properties, such as the concepts ‘exists’ and ‘is mind-independent,’ while no such distinction is required for concepts that refer to determinations of objects, such as the concepts ‘is extended’ and ‘has hairy feet.’ And it is also why we cannot drop the qualification ‘empirically’ or ‘with respect to the empirical level’ in our description of empirical objects as mind-independent from the point of view of fundamental ontology, despite the fact that, from the empirical point of view, empirical objects can be described as mind-independent without any qualifications. Similarly, when speaking from the point of view of fundamental ontology, if, in talking about Frodo’s ontological status, we want to express something that is true, we have no choice but to specify that Frodo exists and is mind-independent in the world of The Lord of the Rings, even though, from the point of view of the world of The Lord of the Rings, Frodo exists and is mind-independent without any qualifications. So, even if we grant—which I am happy to do—that, from the point of view of fundamental ontology, intentional objects genuinely have, not only all determinations, but also all ontological properties that they are presented to have in the representations that ontologically specify them, we will still not be committed to the view that, from the point of view of fundamental ontology, empirical objects are mind-independent, and Frodo exists and is mindindependent. All that follows is that, from the point of view of fundamental ontology, empirical objects are empirically mind-independent, and Frodo exists and is mindindependent in the world of The Lord of the Rings.

3.5 Empirical Objects, Empirical Realism, and the Two-aspect View Among the lessons from our discussion in the previous sections is that reading Kant as a genuine idealist about empirical objects, space, and time, is not only demanded by the text but also perfectly compatible with simultaneously reading him as a genuine realist about them provided that one properly understands the qualifications that he is a transcendental idealist and an empirical realist. The lessons to be spelled out in this section are that the conception of empirical objects that two-aspect commentators ascribe to Kant is incompatible both with his empirical idealism and his empirical realism, in particular, his empirical realism about space and time. It is supposed to be one of the main advantages of the two-aspect reading, at least in its ontological variant, that it can make straightforward sense of Kant’s empirical realism about empirical objects in that, in contrast to the two-world reading, it does not present Kant as committed to the view that the ordinary objects of experience are ontologically

160

 ’      

separate from anything that is ‘truly real.’ In this context, realism is usually understood as being tied to (transcendental) mind-independence: x is ‘truly real’ only if it is (transcendentally) mind-independent or at least has a (transcendentally) mind-independent aspect. In contrast to the two-world view, which denies that empirical objects have any (transcendentally) mind-independent features or ontological ingredients, the two-aspect view holds that they do. And so, the argument goes, the two-aspect view is the only reading on which Kant turns out to be a genuine empirical realist. Sometimes two-aspect commentators claim that, on their reading, while certain properties of empirical objects are minddependent―we will call them ‘appearance properties’ from now on―the existence of empirical objects is mind-independent.¹¹³ This is a problematic way of putting the point, though, because (a) Kant clearly holds that at least some appearance properties of empirical objects are essential to them, for example, their extension, and (b) as discussed in section 2.3, if an essential property of a thing is mind-dependent, then the existence of the thing is mind-dependent as well, given that the thing cannot exist without its essential properties. So, the two-aspect commentators should stick with the more general formulation that, on their reading, empirical objects are genuine existents that have mindindependent aspects. The first problem with the two-aspect conception of empirical objects as genuine existents with mind-independent aspects is that it is incompatible with Kant’s genuine idealism about empirical objects. As we learned in section 2.1, appearances are “nothing apart from representations” and “exist only in us,” that is, they are fully mind-dependent and include no mind-independent ontological ingredients whatsoever. And since the text also clearly establishes that empirical objects are appearances, as discussed in section 3.1, it directly follows that, on Kant’s view, empirical objects are fully mind-dependent and have no mind-independent aspects of any kind. That is, the strategy for how to accommodate Kant’s empirical realism typically embraced by proponents of the two-aspect view, in both the one-object and the same-things version (as characterized in section 1.2), comes at the price of leaving them incapable of properly accommodating his genuine idealism. The second problem with the two-aspect conception of empirical objects as genuine existents with mind-independent aspects is that it also leaves them incapable of properly accommodating Kant’s empirical realism. As we just learned in section 3.4, it is part and parcel of Kant’s empirical realism that empirical objects have spatial and temporal determinations. For emphasis, we might put this by saying that it is a central tenet of Kant’s empirical realism that there are genuine existents that genuinely have genuine spatial and temporal determinations, and these existents are empirical objects. But, according to the two-aspect interpretation, there are no objects like that, and so empirical objects are not like that either. For efficiency’s sake, in the following explication of this claim I will mostly focus on the case of space and spatial determinations; but exactly similar remarks apply to the case of time and temporal determinations. Since it is uncontroversial that space is (nothing but) a form of sensibility on Kant’s view, the proponents of the two-aspect interpretation must admit, and typically do admit, that spatial determinations belong to the appearance aspect of things. But this means that they cannot attribute these determinations to things that genuinely exist and have some mind-independent aspects without

¹¹³ See, for example, Allais 2004, esp. 665.

 , ,   - 

161

adding some accompanying qualifications. They have several options for how to do so. They could qualify, (1) the things, (2), the way in which spatial determinations belong to them, or (3) the spatial determinations. That is, they could say either (1) that it is not genuine existents but genuine existents (considered) as they appear to us, that have spatial determinations or (2) that genuine existents do not genuinely have spatial determinations but merely appear to have them or (3) that genuine existents do not have genuine spatial determinations but merely determinations of being spatial-in-relation-to-us, or of being spatial-when-considered-under-the-conditions-of-human-sensibility, or of appearing spatial, or merely dispositions to cause intuitions in us of themselves as spatial, or something similar along these lines. But none of these options are viable if the goal is to give up on neither the two-aspect interpretation nor empirical realism. Option (1) is a non-starter. Either it amounts to a violation of Kant’s empirical realism by including the denial of the claim that there are genuine existents that have spatial determinations. Or, if a proponent of this option insists that a genuine existent (considered) as it appears to us is itself a genuine existent, it amounts to the adoption of a two-world view. For given that spatial determinations are mind-dependent, the only objects to which they can be ascribed without the addition of any qualifications along the lines of options (2) and (3) are objects that are fully mind-dependent.¹¹⁴ Option (2) is blatantly incompatible with Kant’s empirical realism. An object that merely appears to have spatial determinations is not genuinely spatial. That leaves option (3), which, unsurprisingly, is the most popular chosen option in the literature by far.¹¹⁵ But option (3) is also incompatible with Kant’s empirical realism, although in the present case this is not quite as obvious as it was for the previous two options. To reiterate, on Kant’s view, empirical objects are genuine existents that genuinely have genuine spatial determinations. The proper way to attribute, say, extension to a rose, on Kant’s view, and the way in which Kant in fact attributes extension to empirical objects such as roses, is to say, simply, that the rose is extended. Moreover, there is no indication anywhere that Kant holds that ‘is extended’ does not mean ‘is extended’ but rather ‘has the determination of being-extended-in-relation-to-us’ or ‘has the determination of appearing extended.’ ‘Is extended’ in Kant’s usage just means ‘is extended.’ The claims that the rose has the determination of being extended-in-relation-to-us, or of appearing extended, etc. are at best misleading and do not go far enough and, at worst, unintelligible. Only the straightforward formulation that the rose is extended expresses Kant’s realist view about space with respect to the empirical level of reality. Appearance talk becomes an issue only when we ask further questions about the ontological status of space and time and the objects in them from the point of view of fundamental ontology. From this fundamental point of view, space and time are transcendentally ideal, and, accordingly, all objects in them are appearances in the transcendental sense, including the rose—although, even from the point of view of fundamental ontology, the proper way of attributing extension to the rose is to say, simply, that it is extended. In sum, the two-aspect view, be it the one-object or the same-things version, is incompatible with Kant’s empirical realism about space and ¹¹⁴ That option (1), if it is cashed out in such a way as to not conflict with Kant’s empirical realism, amounts to the adoption of a two-world view will become even clearer below in our discussion of possible strategies of proponents of the one-object two-aspect view for how to escape the charge that they are committed to a bunch of contradictions. Option (1) is not a frequent choice in the literature but it is represented; it is adopted in Roche 2011, for example. ¹¹⁵ See for example Allison 2004, 16, 42–45; Rosefeldt 2007; Allais 2007; Allais 2015, 117 22.

162

 ’      

time since it cannot accommodate its central thesis that empirical objects, which are genuine existents, genuinely have genuine spatial and temporal determinations. Both versions of the two-aspect view are incompatible with Kant’s empirical realism about space and time. But the one-object version is in especially hot water. In addition to the pressure to come up with a way of ascribing spatial and temporal determinations to empirical objects that respects both their mind-dependence and Kant’s empirical realism about space and time, one-object commentators must contend with the objection that their reading implies that Kant is committed to a host of contradictory statements, which affirm and deny the same spatial or temporal predicates of the very same thing. The objection that the one-object view saddles Kant with many contradictions is frequently raised in the literature.¹¹⁶ I agree that this is a serious problem for the view but I would like to emphasize (what is not usually emphasized, namely,) that the best version of the objection gets part of its bite from the fact that one of the two principal escape routes for the one-object commentators leads them straight into the conflict with Kant’s empirical realism about space and time just described. Kant is clearly committed to the claims that (a) things in themselves do not have any spatial or temporal determinations, and (b) appearances have spatial and temporal determinations. (a) and (b) pose a problem for the one-object twoaspect interpretation because these claims cannot both be accommodated unless one either rejects the defining core tenet of the one-object view, according to which appearances and things in themselves are numerically identical, or gives up on Kant’s empirical realism. How so? Given the commitment of the one-object commentators to the indicated core tenet, (a) and (b) raise the question of how they can avoid having to ascribe numerous flat-out contradictions to Kant of the form that the very same thing both has and does not have certain spatial or temporal determinations, for example, that the very same rose both is and is not extended. In general, if one is faced with the threat of a contradiction of the form ‘S is F, and S is not F’ there are two obvious basic strategies that one could pursue in order to diffuse the threat. Sloppily put, one could play around either with the subject terms or with the predicate terms that are involved in the seemingly contradictory statement. Let us look at these two strategies in turn. First, one could argue that the apparently contradictory predicates ‘is F’ and ‘is not F’ are not really contradictory after all, because of an equivocation or because they are implicitly relativized in some way, for example, to different times, places, respects, standards of comparison, or points of view. We have already seen how the one-object commentators could reformulate the relevant predicates for the case at hand if they adopted this strategy, namely, by choosing option (2) or (3) described above. For example, with respect to the statement ‘the rose is extended, and the rose is not extended,’ they could argue that, if properly spelled out, the predicates in question are ‘appears to us as extended,’ or, more promisingly, ‘has the property of appearing extended to human subjects,’ or ‘has the property of being extended-inrelation-to-us’ or something similar along these lines, on the one hand, and ‘is not extended,’ on the other hand. And, indeed, there is nothing contradictory about a rose that appears extended, or has the property of appearing extended, but is not extended. But, while the suggested strategy is successful in dispelling the looming contradiction, it is not ¹¹⁶ A particularly insightful presentation of the objection can be found in Van Cleve 1999, 146–8. Also see Stang 2014.

 , ,   - 

163

available to anybody who is committed to empirical realism about space and time, like Kant. As we just reminded ourselves, Kant holds that there are genuine existents that genuinely have genuine spatial determinations, namely, empirical objects, and empirical objects are appearances. For Kant, roses genuinely are genuinely extended. On the oneobject proposal under consideration, empirical objects do not genuinely have genuine spatial determinations, nor are there any genuine existents that genuinely have genuine spatial determinations. So, the proposed first strategy of the one-object commentators for how to avoid having to ascribe to Kant a multitude of contradictory statements comes at the cost of not being able to accommodate his empirical realism about space and time. The second obvious general strategy for diffusing a threatening contradiction of the form ‘S is F, and S is not F’ is to try to show that the subject term ‘S’ to which the contradictory predicates are attributed is ambiguous and does not have the same referent in both cases. For example, the threatening contradiction in the statement ‘my real estate property is in good shape, and my real estate property is in bad shape’ could be dispelled by reformulating it more precisely as ‘my house is in good shape, and my yard is in bad shape.’ Something like this strategy seems to be at work when one-object commentators pull the familiar ‘thing as,’ ‘thing considered as,’ and ‘thing qua’ constructions out of their toolbox—except that they have the additional constraint to contend with that on their view the things to which the reformulated subject terms refer are supposed to be numerically identical after all. Applying this strategy to the case at hand, the one-object commentators would argue that the statement ‘the rose is extended, and the rose is not extended’ is to be reformulated more precisely as ‘the rose (considered) as it appears to us, or the rose qua bearer of appearance properties, is extended, and the same rose (considered) as it is in itself, or the same rose qua bearer of (what one might call) in-itself properties, is not extended.’ Does this proposal live up to the task of resolving the looming contradiction, while neither engendering a rejection of empirical realism about space and time nor engendering a rejection of the core tenet of the one-object view that appearances and things in themselves are numerically identical? As hard as I try, I find it impossible to make sense of the reformulated statement while both keeping the numerical identity claim in place and resisting the temptation to read the first part of the statement as a way of saying that the rose merely appears to be extended, or merely has the property of appearing extended—which would amount to sliding back to the previous proposal and giving up on Kant’s empirical realism. One reason why it is so difficult to make sense of the statement in question is that it is unclear how exactly the ‘thing qua’ or ‘thing considered as’ constructions and the corresponding property attributions are supposed to be understood. The methodological one-object view is in the toughest spot here. How could a mere change in the way in which we consider a thing possibly go hand in hand with a change in its genuine (as opposed to merely apparent) spatial or temporal determinations?¹¹⁷ But even setting this problem aside, it still remains a mystery how the ‘considered as’ and ‘qua’ qualifications could accomplish what they are intended to accomplish, namely, to make the looming contradiction go away, without at the same time introducing a distinction between different objects or leading us straight back to the previously considered unsatisfactory proposal that brings with it the rejection of empirical realism about space and ¹¹⁷ See Van Cleve 1999, 8: “As I sit typing these words, I have shoes on my feet. But consider me apart from my shoes: so considered, am I barefoot? I am inclined to say no; consider me how you will, I am not now barefoot.”

164

 ’      

time. If we assume that ‘is F’ and ‘is not F’ are in fact contradictory and that the relevant context is extensional, it seems undeniable that reformulating a seemingly contradictory statement such as ‘S is F, and S is not F’ by changing the subject terms can successfully dissolve the threatening contradiction only if the reformulated subject terms refer to numerically distinct things. In terms of our real estate example above, the suggested reformulation successfully resolves the contradiction only because my house is in fact numerically distinct from my yard. So, as I see it, the only way in which the ‘considered as’ and ‘qua’ qualifications of the two-aspect commentators could make the looming contradiction go away without running afoul of Kant’s empirical realism is if they were understood to introduce a distinction between different objects. The thing-(considered)-asit-appears-to-us, or the thing-qua-bearer-of-appearance-properties, is one object; this object is extended. The thing-(considered)-as-it-is-in-itself, or the thing qua-bearer-ofin-itself-properties, is another, second object; this object is not extended. But, of course, as soon as this concession is made we have abandoned the one-object view. The considerations in the previous paragraph might prompt the following objection on behalf of the one-object view. The claim that the reformulation of the subject terms in the statement ‘S is F, and S is not F’ removes the threatening contradiction only if the reformulated subject terms have numerically distinct referents has limited scope. As acknowledged in the formulation in the previous paragraph, the claim only holds for extensional contexts, that is, contexts in which co-referential terms can be substituted for each other without changing the truth-value of the relevant sentence. If we are dealing with an intensional (that is, not extensional) context, ‘S1 is F’ and ‘S2 is not-F’ can both be true, even though ‘S1’ and ‘S2’ have the same referent.¹¹⁸ For example, take the case of Jocasta, the mother and tragic wife of Oedipus, and the property of being known-by-Oedipusto-be-his-wife, and consider the contradictory statement ‘Jocasta is known-by-Oedipusto-be-his-wife, and Jocasta is not known-by-Oedipus-to-be-his-wife,’ which conforms to our schema ‘S is F, and S is not F.’ The sentence ‘Oedipus knows that S is his wife’ involves or creates an intensional context. Accordingly, the predicate ‘is known-by-Oedipus-to-behis-wife’ is an intensional predicate, as one might say, that is, a predicate that can at the same time be truthfully affirmed of and truthfully be denied of subject terms that have the same referent provided the subject terms present the referent in two appropriately different ways. Due to the intensional nature of this predicate, the threatening contradiction can thus be dissolved by appropriately reformulating the subject terms in the problematic statement to yield ‘Jocasta is known-by-Oedipus-to-be-his-wife, and Oedipus’s mother is not known-by-Oedipus-to-be-his-wife.’ This sentence is not contradictory, even though the reformulated subject-terms, ‘Jocasta’ and ‘Oedipus’s mother,’ have the same referent. And, the objection continues, this is exactly how attributions of spatial and temporal determinations are to be understood on the one-object view, that is, as attributions in intensional contexts. Just as the claims that Jocasta is known-by-Oedipus-to-be-his-wife and that Oedipus’s mother is not known-by-Oedipus-to-be-his-wife are both true, even though Jocasta and Oedipus’s mother are numerically identical, the claims that things in themselves do not have any spatial or temporal determinations, as stated in thesis (a) above, and that appearances have spatial and temporal determinations, as stated in thesis ¹¹⁸ Another way of making the same point would be to remind us that what is known as ‘Leibniz’s law,’ that is, the law that if X and Y are identical, then all of their properties are the same, is valid only in extensional contexts.

 , ,   - 

165

(b) above, are both true, even though appearances and things in themselves are numerically identical.¹¹⁹ In response, ingenious as this improved subject-reformulation strategy of the one-object commentators is, if one takes the trouble of carefully thinking the proposal through it becomes clear that the new strategy, in effect, amounts to a combination of the original subject-reformulation strategy and a close cousin of the predicate-reformulation strategy considered earlier and fails for the same reason as the latter, namely, for the reason that it brings with it a rejection of empirical realism about space and time. So, let us take the trouble of carefully thinking the proposal through. Merely saying that spatial predicates are intensional predicates does not make it so. And, indeed, we commonly do not think of spatial predicates as intensional. The truth-value of ‘S is extended’ is commonly understood not to depend on how S is presented. Sentences that involve intensional contexts typically contain or can be transformed into sentences that contain some kind of epistemic or modal operators such as ‘S knows that . . . ,’ ‘S believes that . . . ,’ ‘It appears to S that . . . .,’ ‘It is necessary that . . . ,’ etc. Similarly, what is special about intensional predicates such as the predicate ‘is known-by-Oedipus-to-be-his-wife’ is their complex internal structure and their incorporation of one of the indicated kind of operators. So, in order to make the view plausible that, despite appearances to the contrary, spatial predicates are intensional, one would have to maintain that they are structurally more complex than they appear to be on the surface and implicitly incorporate some kind of epistemic operator. Given that we are dealing with Kant’s transcendental distinction between appearances and things in themselves, the only reasonable candidate for the relevant operator is ‘It appears to human subjects that . . . ’ or something similar. Accordingly, the one-object commentator must maintain that, although spatial predicates on the surface seem to be of the simple form ‘is F,’ they actually are of the form ‘appears to human subjects as F’ or ‘has the property of appearing to human subjects as F.’ Assuming this account of how and why spatial predicates are to be regarded as intensional, we are now ready to spell out how the new and improved subject-reformulation strategy plays out with respect to our test sentence ‘the rose is extended, and the rose is not extended.’ According to the original subjectreformulation strategy, this sentence is to be reformulated more precisely as ‘the rose as it appears to us is extended, and the rose as it is in itself is not extended.’ This proposal floundered over the objection that the reformulated sentence can be seen as noncontradictory only if the terms ‘the rose as it appears to us’ and ‘the rose as it is in itself ’ refer to numerically distinct things. Proponents of the improved subject-reformulation strategy hold on to the suggested reformulation of the subject terms but at the same time insist that the predicate ‘is extended’ is intensional since it can be unpacked as the predicate ‘appears extended to human subjects,’ or ‘has the property of appearing extended to human subjects.’ So, according to the improved subject-reformulation strategy, our test sentence is to be reformulated more precisely as ‘the rose as it appears to us has the

¹¹⁹ A reading along these lines has been proposed by Andrew Roche, who argues that the key to avoiding the threatening contradiction is the recognition that when we compare things in themselves and appearances we are dealing with a context in which spatiotemporal properties become intensional properties, that is, “properties whose true attribution to objects depends on how those objects are described, or conceived, or perhaps more generally, represented” (Roche 2013, 597). Since “we can establish that A and B are numerically distinct only by way of extensional properties,” “that A and B have different intensional properties . . . shows nothing.” (Roche 2013, 597)

166

 ’      

property of appearing extended to human subjects, and the rose as it is in itself does not have the property of appearing extended to human subjects.’ The good news for the oneobject commentators is that this reformulated sentence is indeed not contradictory even if the reformulated subject terms, ‘the rose as it appears to us’ and ‘the rose as it is in itself,’ are understood to refer to the same thing. So, in this respect, the present strategy is indeed an improvement over the initial subject-reformulation strategy. There is some bad news as well, however. As is made patently obvious by the foregoing analysis, the improved subjectreformulation strategy in effect amounts to running the original subject-reformulation strategy and a close cousin of the predicate-reformulation strategy in tandem and, just like the original predicate-reformulation strategy, prohibits the one-object commentators from being able to accommodate Kant’s empirical realism. The key move in the original predicate-reformulation strategy consisted in reformulating the seemingly contradictory predicates in such a way that they are no longer contradictory and thus can both be affirmed of the same original subject term without saddling us with a contradiction. The present key move consists in reformulating the seemingly contradictory predicates in such a way that it can be reasonably claimed that we are dealing with an intensional context, which allows us, after also appropriately reformulating the subject terms, to affirm one of the reformulated predicates of one of the reformulated subject terms and the other reformulated predicate of the other reformulated subject term without ending up with a contradiction, even though the reformulated predicates are contradictory and the reformulated subject terms have the same referent. Just like the predicate-reformulation proposal, the present proposal enables the one-object commentators to avoid having to ascribe contradictions to Kant without forcing them to relinquish their key tenet that appearances and things in themselves are numerically identical. But, also just like the predicate-reformulation proposal, the present proposal accomplishes this feat only at the cost of leaving the one-object commentators unable to do justice to Kant’s empirical realism about space and time. Presumably, Kant would agree that the rose as it appears to us has the property of appearing extended to us and, more generally, that empirical objects as they appear to us have properties of appearing as spatial and temporal to us. But these trivial claims fall far short of what Kant’s empirical realism amounts to in this context. Kant’s empirical realist view of the matter is that the rose is extended and, more generally, that empirical objects have genuine spatial and temporal determinations, which are claims that the one-object commentator cannot accommodate. Hence, although an improvement over the original subject-reformulation strategy, the present strategy of the one-object commentators for how to avoid having to ascribe to Kant numerous glaring contradictions turns out to be unsatisfactory for ultimately the same reason as the first strategy considered above, namely, that the resulting position is incompatible with Kant’s empirical realism about space and time. The moral of the story is that if one wants to be able to read Kant as an empirical realist about space and time—which he clearly is—one cannot understand his transcendental distinction between things in themselves and appearances in the way proposed by the two-aspect interpretation, and especially not in the way proposed by the one-object version of this interpretation. The troubles of proponents of the one-object view with respect to their conception of empirical objects do not end there, though. Regardless of whether they regard empirical objects as the appearance aspects of genuine existents or as genuine existents that have two aspects, a necessary condition for the viability of their interpretation is that it makes sense

 , ,   - 

167

to talk about or to consider or at least to talk about considering empirical objects as they are apart from their appearance aspect.¹²⁰ (Considering empirical objects as they are apart from their appearance aspect is what considering things as they are in themselves amounts to, on the two-aspect view.) The problem is that if we take into account how Kant conceives of the nature of empirical objects it turns out that considering an empirical object as it is apart from its appearance aspect is impossible—and not just for us but for anybody, including God or any other putative pure intelligence. Similarly, talk about an empirical object as it is apart from its appearance aspect is, strictly speaking, nonsense. How so? Kant holds that empirical objects essentially are in space and time,¹²¹ and many of their essential determinations, namely, most of their essential determinations that are knowable to us, are space-infected or time-infected, in the terminology introduced in section 3.4.4, that is, they are instances of properties that are essentially bound up with space or time in the sense that objects can have them only if they are in space or time, respectively.¹²² The space-infected and time-infected determinations of a thing comprise its spatial and temporal determinations but also such determinations as being moveable or having a certain velocity. Now, since on Kant’s view space and time are (nothing but) forms of sensibility, all spaceinfected and time-infected determinations of a thing belong to its appearance aspect, as we will discuss in more detail in section 4.1.1. So, no matter how the one-object commentators want to cash out in detail what it means to consider an empirical object as it is apart from its appearance aspect, what this consideration must involve at minimum is considering it apart from its space-infected and time-infected determinations. But this means that considering an empirical object as it is apart from its appearance aspect requires considering it apart from a good deal of what is essential to it. This is plainly impossible.¹²³ Abstracting from several essential determinations of a thing, even if they do not exhaust its essence, leaves us with something that can no longer be meaningfully identified with the thing that we started out with. After all, for a determination to be essential to a thing means precisely that nothing that lacks the determination could possibly be the thing. What sense could be made of the assertion that if we or God, for that matter, abstract from my bicycle’s size, shape, weight, material and mechanical constitution, spatiotemporal position, causal powers, causal history, and origin, God or we are still considering the same, numerically

¹²⁰ For the case of the first conception, we should say, strictly speaking, that it must make sense to talk about or consider the things that as they appear to us are empirical objects as they are apart from their appearance aspect. Since this is a bit cumbersome, I will stick to the simpler formulation just used in the main text. This simplification does not affect the argument. ¹²¹ See note 50. ¹²² For Kant, as for pretty much every other modern philosopher, apart from being extended, the other main essential property of each empirical object is that it is material, that is, that it is a chunk of matter. In the Metaphysical Foundations, Kant provides an account of all the central essential properties of matter known to us, including that it is (a) “the movable in space” (MAN, 4:480); (b) “the movable insofar it fills a space” (MAN, 4:496), which it accomplishes “not through its mere existence but through a special moving force” (MAN, 4:497), namely, a repulsive force, which is counteracted by a corresponding attractive force (see MAN, 4:498–499, 508–510); (c) impenetrable (MAN, 4:501); (d) “infinitely divisible, namely, into parts each of which is again matter” (MAN, 4:503); and (e) “the movable insofar as it as such has moving force” (MAN, 4:536). Also see B321/A265; B341/ A285; BJPMM, 8:153. All of these properties are space-infected and time-infected. ¹²³ This point is also made by Thomas Pogge and Ralph Walker; see Pogge 1991, 495 496; Walker 2010, 825 6. Note that even if we thought that the appearance properties of empirical objects were exhausted by their spatial and temporal determinations, considering them as they are apart from how they appear to us would amount to considering them apart from a good deal of what is essential to them.

168

 ’      

identical thing, that is, my bicycle? To my mind, the answer is ‘none whatsoever.’¹²⁴ For the same reason, the expression ‘empirical object as it is apart from its appearance aspect’ simply does not make any sense. Needless to add, it makes even less sense to talk about empirical objects as they are apart from their appearance aspect, if all of their determinations belong to their appearance aspect—which I indeed take to be Kant’s view. This is an opportune moment to acknowledge that Kant himself occasionally uses formulations that suggest that he is talking about objects considered as they are apart from their relation to sensibility. But this kind of formulation must be understood as referring, not to individual things, let alone individual things in themselves, but to the transcendental object in its function in the constitution of appearances. The concept of the transcendental object is a central notion in Kant’s theory of the human mind’s conceptual arsenal, so to speak. This concept does not only play an important role in Kant’s theory of experience; it also is part of our concept of a thing in itself and an integral element in the fictions that are the ideas of reason, to be discussed in section 6.3, for example, the fiction of God as the creator of the world.¹²⁵ As already noted, the concept of the transcendental object is the concept of a bare particular something = X that is, (a) distinct both from our representations of it and from the cognizer entertaining these representations and (b) a bearer of determinations. In Kant’s theory of experience, the concept of the transcendental object is involved in the constitution of both the intentional objects of perceptions and the intentional objects of experience. As we learned in sections 2.4–2.8, Kant’s theory of experience incorporates an account of the various stages in which we, thanks to the specific make-up of our cognitive faculties, gradually order and unify the sensible material that is given to us through sensations and thereby bring about representations with increasing degrees of objectivity, a process that begins with and is fueled by perception and aims at the construction of experience, which alone represents genuinely existing, public, proper objects. One of the first main challenges in trying to provide such an account is to explicate how the intentional objects of perceptions are constituted. The task here is to explain how we transform a manifold of sensible qualia that are presented as ordered in certain spatial relations into a (direct) representation of a ‘lightweight’ object.¹²⁶ As already noted in sections 2.2. and 2.4, this transformation centrally involves an application of the concept of the transcendental object. In Kant’s theory of perception, the transcendental object is the projected bare bearer of the determinations that correspond to the sensations in empirical intuition. It is added in thought to the sensations in such a way as to yield a perception whose presentational content is no longer merely sensible qualia in certain spatial relations but a particular thing with a certain shape that has the sensible qualia as its

¹²⁴ Could we not secure the reference of our considerations to the bicycle simply by pointing at it while performing all of these abstractions? No. One can only point at things that stand in a determinate spatial relation to whatever one is pointing with, but spatial relations are among the determinations from which we are abstracting. ¹²⁵ See B344–346/A288–289; A393; B568/A540; B593 594/A565 A566; B707/A679. This multi-purpose character of the concept of a transcendental object also means that we have to pay careful attention to the context in which Kant uses the term ‘transcendental object.’ Another reason to be sensitive to the context is that this term can also be used to refer to objects in the transcendental sense, that is, objects that exist at the transcendental level of reality, as discussed in section 3.1. ¹²⁶ The qualification ‘lightweight’ is intended to signify that we are not talking about (direct) representations of proper objects yet. Only representations that conform to all of Kant’s formal conditions of experience and objectivity have proper objects as their intentional objects. As discussed in section 2.5.2, perceptions do not conform to all of these conditions.

 , ,   - 

169

determinations and is distinct from the perception.¹²⁷ As we saw in section 2.5.2, the concept of the transcendental object plays a similar, vital role in the later stages of the constitution of the objects of experience by, again, providing the conception of a bare skeleton of a particular thing = X that we ‘dress up’ by presenting it in all the ways necessary for it to count as a proper object for us, for example, by presenting it as outside us, as the proximate cause of certain perceptions, and as causally connected to other objects. This explains why and in what sense the transcendental object can be identified with the empirical object or the perceptual object considered as it is in itself, or as it is apart from its relation to sensibility.¹²⁸ Objects thus considered are neither individuated, nor can they be identified with things in themselves in the transcendental sense. As Kant himself emphasizes, we have no concept of the transcendental object that is involved in the construction of experience “other than merely of the object of a sensible intuition in general, which is thus the same for all appearances” (A253).¹²⁹ That is, Kant agrees that once we consider an object as it is apart from sensibility, we are no longer considering the same individual but are merely contemplating a particular something = X that is distinct from us and our sensible representations, a something that is exactly the same in all cases no matter with which object we started out.¹³⁰ Moreover, from the point of view of fundamental ontology, the transcendental object in both perception and experience is just as mind-dependent as the appearances constituted with its help. It is an ontological ingredient of appearances that is contributed by the understanding in the course of the

¹²⁷ Recall A250: “All our representations are in fact related to some object by the understanding, and since appearances are nothing but representations, the understanding relates them to something, as the object of sensible intuition: but this something is to that extent only the transcendental object. But this signifies a something = x, of which we do not know anything at all nor in general can know (according to the current constitution of our understanding) but which only serves as a correlate of the unity of apperception for the unity of the manifold in the sensible intuition, by means of which the understanding unifies the manifold in the concept of an object. This transcendental object cannot at all be separated from the sensible data, since then nothing remains through which it would be thought. It is thus not an object of cognition in itself but merely the representation of appearances, under the concept of an object in general, which is determinable through the manifold of the latter.” Also see A104 105; A109; B304/A247. ¹²⁸ For example, I understand B55/A38 along these lines where Kant says of the appearance that “it always has two sides, the one where the object is considered in itself (irrespective of the manner of intuiting it but whose quality for this reason always remains unknown), the other where we look at the form of intuition of this object, which must not be looked for in the object in itself but in the subject to whom this object appears but which still really and necessarily pertains to the appearance of this object.” Also see A366: “What thing in itself (transcendental object) matter is, is completely unknown to us . . . ” Also see B306. The reading of Kant’s conception of the transcendental object presented here is similar to the reading presented in Adickes 1924, 97–108; and Prauss 1974, 98–114. I do not follow Prauss, however, in regarding the concept of a thing in itself as a concept on a second level of transcendental reflection whose function is to explicitly express the result of the first level of reflection that our experience has an objective side, as it were, namely, the side that is captured in the concept of the transcendental object. ¹²⁹ Here is the passage in full, A253: “The object to which I relate the appearance in general is the transcendental object, i.e., the completely undetermined thought of something in general. This cannot be called the noumenon; for I do not know of it what it is in itself and have no concept of it at all other than merely of the object of a sensible intuition in general, which is thus the same for all appearances.” ¹³⁰ A particular thing is not the same as an individual or an individuated thing. On my reading, the transcendental object is a particularizing element in appearances but not an individuating element. That the transcendental object is included in an appearance as one of its ontological ingredients makes it so that the appearance is a particular thing, as opposed to a universal, and that the determinations of the appearance are particular property instances. But appearances are individuated by means of the specific determinations with which the bare particular that corresponds to the transcendental object is ‘dressed up’ in the course of the construction of experience, in particular, its spatial and temporal determinations. On my reading, appearances, although individuated, are not individuals, however, because they are not completely determinate. See the discussion in section 4.2.2.2, especially note 53, chapter 4.

170

 ’      

construction of experience through the application of the concept of the transcendental object.¹³¹ Having started with the explication of Kantian formulations that touch on the ontological status of empirical objects whose meaning is not immediately transparent, I will conclude this section with a few explanatory comments about the formulations that (a) ‘appearances are not things in themselves but mere representations,’ and that (b) certain philosophers mistakenly ‘regard appearances or mere representations as things in themselves,’ which Kant uses repeatedly and, which, on first hearing, sound a bit odd. On my reading, Kant employs these formulations in two different ways.¹³² First, he sometimes uses (a) to express the first thesis of transcendental idealism that empirical objects are not things in themselves but fully mind-dependent entities and (b) to describe the basic error of the transcendental realist of mistakenly regarding empirical objects as things in themselves. Since on Kant’s view empirical objects are appearances, he often uses ‘appearance’ and ‘empirical object’ interchangeably, which explains why we find these alternative formulations. Kant is a bit sloppy when he substitutes ‘appearance’ for ‘empirical object’ in describing the transcendental realist position because many a transcendental realist would probably not recognize himself in the description that he regards appearances or mere representations as things in themselves.¹³³ What Kant should have said more carefully in this case—and what, undoubtedly, he means, as the context of the passages ¹³¹ V-Met-L2/Pölitz, 28:550: “Through the senses we can only cognize the properties or predicates of the object, the object itself lies in the understanding.” See R5554, 18:230: “Noumenon signifies (basically) always the same, namely, the transcendental object of sensible intuition (but this is not a real object or given thing but a concept in relation to which appearances have unity), for something must correspond to this [the intuition], even though we know nothing but its appearance. But we cannot say that the (pure) categories have objects; they merely determine the transcendental object in relation to our sensibility through the synthesis of the manifold of intuition.” See R5654, 18:312: “We have to note here that every object signifies something that is distinguished from the representation but which is only in the understanding . . . ” In the Opus Postumum, one can find a number of passages that appear to support a two-aspect fictionalist reading of the transcendental distinction; see note 56, chapter 1. In my assessment, the best way to understand what is going on there is that one of the views that Kant is trying on for size in his final years is a view that dispenses with things in themselves, understood as actually existing mind-independent grounds of appearances, and that settles on the transcendental object, understood as the projected bearer of sensible determinations, as the legitimate heir to the title ‘thing in itself.’ ¹³² I will arrange the relevant quotations in two groups; the quotations in group I illustrate the first way in which Kant uses the formulations in question; the quotations in group II illustrate the second way. Group I. See A369: “But by transcendental idealism of all appearances I understand the doctrine according to which we regard them altogether as mere representations and not as things in themselves . . . ” See B235/A190: “If appearances were things in themselves, no human being could figure out from the succession of the representations of their manifold how the latter is connected in the object.” See B563/A536: “If appearances were things in themselves, and thus space and time forms of the existence of things in themselves . . . ” See B564–565/A536–537: “If appearances are things in themselves, freedom cannot be saved. . . . But if, on the other hand, appearances are regarded as nothing more than what they in fact are, namely, not as things in themselves but mere representations . . . ” See B768/A740: “ . . . but it turned out that it [the antithetic of pure reason] rests on an error, namely, that, according to the common prejudice, one regarded appearances as things in themselves . . . ” Group II. See B206/A165: “Appearances are not things in themselves.” See A108–109: “Appearances are the only objects that can be given to us immediately. . . . But these appearances are not things in themselves but themselves only representations . . . ” See A101: “ . . . appearances are not things in themselves but the mere play of our representations that in the end amount to determinations of inner sense.” See A114: “But if one remembers that this nature in itself is nothing but a sum total of appearances and thus not a thing in itself but merely a multitude of representations of the mind . . . ” See B549/A521: “Since it, as appearance, can be none of the two, for appearance is not a thing in itself . . . ” See V-Met-L2/Pölitz, 28:569: “The size of appearances cannot be given; for appearance is not a thing in itself and does not have a size.” It is also worthwhile to note that trying to make sense of these formulations within the framework of the two-aspect view is an extremely daunting task, which I leave as an exercise for the reader who fancies the two-aspect view. ¹³³ Kant to Newton: “You regard empirical objects as things in themselves, that is, as mind-independent entities?” Newton: “That’s right.” Kant to Newton: “You regard appearances and thus mere representations as things in themselves?” Newton: “Huh?”

 , ,   - 

171

makes clear—is that the transcendental realist regards what, unbeknownst to him, actually are appearances and thus ‘mere representations,’ namely, empirical objects, as things in themselves. Second, Kant also uses formulation (a) to express his transcendental distinction, according to which appearances and things in themselves are numerically distinct and fundamentally different kinds of existents, namely, fully mind-dependent and mindindependent ones, respectively, and formulation (b) to describe the error of the LeibnizWolffians of confusing appearances with things in themselves. As already noted in section 2.4, Kant reads the Leibniz-Wolffians as holding that sensible and intellectual representations do not differ in kind but merely with respect to their degree of distinctness or confusedness. Sensible representations are confused concepts, that is, concepts whose components are not individually apprehended with consciousness.¹³⁴ Kant finds this conception problematic for several reasons, but one of the main reasons is that it leads them to a “transcendental amphiboly,” that is, “a confusion of the object of the pure understanding with the appearance” (B326/A270).¹³⁵ Since the Leibniz-Wolffians, as committed rationalists, believe that objects can be cognized in terms of pure concepts alone if, and only if, they are mind-independent, and, thus, that the concept ‘object of the pure understanding,’ or ‘noumenon,’ and the concept ‘thing in itself ’ are extensionally equivalent, their transcendental amphiboly at the same time amounts to a confusion of things in themselves with appearances.¹³⁶ The confusion consists in that, on account of their conception of sensible representations as confused concepts, they understand the distinction between things in themselves and appearances in two-aspect fashion as merely amounting to a distinction between different ways in which we represent the very same mind-independent entities, namely, distinctly in terms of pure concepts, on the one hand, and confusedly in terms of sensible representations, on the other hand.¹³⁷ As a ¹³⁴ See FM, 20:278: “Regarding the Leibnizian principle of the logical difference of the indistinctness and distinctness of representations, when he claims that the former kind of representation, which we called mere intuition, is basically only the confused concept of its object, and, thus, intuition is distinguished from concepts of things merely with respect to the degree of consciousness, not specifically . . . ” Also see the other references in note 80, chapter 2. For a reminder of Kant’s conception of what it means for a representation to be confused (which he prefers to call ‘indistinct’), see note 65, chapter 2. ¹³⁵ The complaint that the Leibniz-Wolffians confuse noumena with phenomena can already be found MSI, 2:395: “Thus, I am afraid that Wolff, through this difference between sensible and intellectual cognition, which for him is only a logical difference, might have completely destroyed that valuable old institution to distinguish the nature of the phenomena from that which is understood by the intellect, causing great damage to philosophy . . . ” This Leibniz-Wolffian confusion is the focus of attention in the chapter on the so-called amphiboly of the concepts of reflection, which contains Kant’s most sustained critical engagement with the Leibniz-Wolffian philosophy in the Critique. We will take a closer look at the amphiboly chapter in sections 6.2 and 6.4; also see section 5.3. For a more general examination of Kant’s many worries regarding the Leibniz-Wolffian conception of the difference between sensible and intellectual representations, see Jauernig (forthcoming 2021a). ¹³⁶ See B320/A264: “Leibniz took appearances for things in themselves and thus for intelligibilia, i.e., objects of the pure understanding (although he gave them the name of phenomena because of the confusion of their representations) . . . ” Note that, on Kant’s analysis, there are different ways in which one can commit a transcendental amphiboly. Locke is guilty of such a confusion as well, but in contrast to the Leibniz-Wolffians who “intellectualized appearances,” that is, who took appearances to be confusedly represented things in themselves, Locke “sensibilized the concepts of the understanding” (B327/A271), that is, he took the concepts of the understanding to be worked-over, glorified ideas of sensation (and reflection), and, thus, conceived of noumena as ‘refined’ phenomena, as it were. ¹³⁷ See B326–27/A270–71: “He [Leibniz] did not regard the conditions of sensible intuition as original, which carry their own differences with them; for sensibility was for him merely a confused kind of representation and not a special source of representation; appearance was for him a representation of a thing in itself, although distinguished with respect to logical form from the cognition through the understanding, since the former, given its usual lack of analysis, draws a certain mix of coordinated representations into the concept of the thing,

172

 ’      

consequence of their amphiboly, the Leibniz-Wolffians take empirical objects to be things in themselves, albeit confusedly perceived ones, and thus commit the basic error of transcendental realism. But, in their case, the underlying confusion is more complex and subtle than in the case of a more typical, garden-variety transcendental realist. The LeibnizWolffians recognize that the distinction between appearances and things in themselves has an important role to play in a satisfactory epistemology and ontology—a fact that is not recognized by all transcendental realists—and they also recognize that empirical objects are appearances—which is hardly ever recognized by other transcendental realists—but they (like proponents of the methodological two-aspect view) conceive of the distinction between appearances and things in themselves in the wrong way, namely, as a mere distinction between different ways in which we cognitively relate to things and not as a distinction between two fundamentally different kinds of things.¹³⁸

3.6 Core Theses of Transcendental Idealism and Empirical Realism III: the Ontological Status of Empirical selves As indicated in section 2.9, Kant’s account of the construction of inner experience and the constitution of empirical selves appears to deviate in fairly significant ways from his account of the construction of outer experience and the constitution of empirical objects. Empirical objects and empirical selves also differ in that the former are public and contained in both time and space, while the latter are largely private and contained only in time. Still, Kant leaves no doubt that, as far as their basic ontological status is concerned, empirical objects and empirical selves are in exactly the same boat: they genuinely exist, but they are appearances, that is, fully mind-dependent intentional objects, and not things in themselves. They [certain opponents of Kant’s view] did not hope to be able to apodictically demonstrate the absolute reality of space, since they were confronted by idealism, according to which the reality of outer objects is not capable of a strict proof. By contrast, the reality of the object of our inner senses (of myself and my state) is immediately clear through consciousness. The former [outer objects] might be a mere illusion, but the latter [the

which the understanding knows to separate from it.” See B60 61/A43: “That our whole sensibility is nothing but the confused representation of things, which contains merely what belongs to them in themselves but only by way of a heaping together of marks and partial representations that we do not separate from one another with consciousness, is a falsification of the concept of sensibility and of appearance, which renders the entire doctrine of them useless and empty. The distinction between an indistinct and a distinct representation is merely logical and does not concern the content . . . ” Also see FM, 20:278; ÜE, 8:218. Also see Schultz 1790, 791. ¹³⁸ In his reply to Eberhard, who subscribes to a Leibniz-Wolffian style two-aspect conception of the distinction between things in themselves and appearances, Kant expresses his disagreement with this conception by saying that it amounts “not to any distinction in the things but merely in the degree of our faculty of perception” and concludes with the comment that, if this were all that his transcendental distinction came down to, it would be “mere child’s play” (ÜE, 8:208–209). For the purposes of illustrating how the Leibniz-Wolffian amphiboly goes beyond the basic error of transcendental realism, imagine the following two situations. In situation 1, you and Kant are facing an elephant. Kant asks: “Is this elephant a thing in itself or an appearance?” If you answer that it is a thing in itself, you are a transcendental realist, and Kant will disagree with you. In situation 2, you and Kant are facing a gnu. Kant asks: “We agree that what we have here is the appearance of a gnu; but is the gnu a mindindependent thing?” If you answer that it is, you could be a Leibniz-Wolffian (or a two-aspect commentator), and Kant will disagree with you.

     

173

object of inner sense] is, in their opinion, indisputably something real. But they did not take into account that both still belong merely to appearance, although one must not dispute their reality as representations . . . (B55/A38)¹³⁹

Given this similarity between empirical objects and empirical selves as far as their basic ontological status is concerned, it is easy to formulate theses of transcendental idealism and empirical realism about empirical selves that mirror the corresponding theses about empirical objects that we collected in section 3.3.1. (TI3) (a) Empirical selves are not things in themselves in the transcendental sense, or with respect to the transcendental level of reality, that is, they are not mind-independent; rather (b) empirical selves are appearances in the transcendental sense, that is, they are fully mind-dependent. (TI3-strong) (a) Empirical selves are not things in themselves in the transcendental sense, that is, they are not mind-independent; rather, (b), empirical selves are Kantian inner appearances, that is, they essentially are intentional objects of inner experience as characterized in Kant’s theory of experience, and, thus, fully mind-dependent. (TI3-concise) Empirical selves are transcendentally ideal, that is, they are ideal in the transcendental sense, or with respect to the transcendental level of reality. (ER3) (i) Empirical selves exist; and (ii) empirical selves are things in themselves in the empirical sense, or with respect to the empirical level of reality, that is, they are mindindependent in the empirical sense, or with respect to the empirical level of reality. The claim that our empirical selves are transcendentally ideal would not be very exciting if ‘empirical self ’ simply meant ‘intentional object of inner experience.’ The claim has bite, however, if we assume our common sense understanding of empirical selves as the selves whose determinations are revealed in our self-consciousness. This is how ‘empirical self ’ is meant to be understood in the claim that our empirical selves are appearances and not things in themselves. That empirical selves are fully mind-dependent can also be expressed metaphorically by saying that they only exist ‘in us,’ or ‘in our representations,’ and entails that they do not exist at the transcendental level of reality.¹⁴⁰ Empirical selves are empirically mind-independent in that inner experience presents them as being mindindependent. While outer experience presents empirical objects as mind-independent by presenting them as being in space, inner experience presents empirical selves as mindindependent by presenting them as being in empirical time, by way of relating their mental states to the objective temporal order of outer appearances. This amounts to a

¹³⁹ See B69: “If I say: in space and time both the intuition of outer objects and the self-intuition of the mind represents both how it affects [sic.] our sense, i.e., as it appears, this is not meant to say that these objects are a mere illusion.” See B156: “ . . . thus we must order the determinations of inner sense as appearances in time in exactly the same way as we must order the determinations of the outer senses in space; and, thus, if we admit of the latter that we thereby cognize objects only insofar as we are outwardly affected, we must also admit of inner sense that we thereby intuit ourselves only in the way in which we are inwardly affected by ourselves, i.e., that as far as inner intuition is concerned, we cognize our own subject merely as appearance but not according to what it is in itself.” Also see B68; B158; B334/A278; A372–373; KpV, 5:6; Anth, 7:142. ¹⁴⁰ As we will see in section 5.8, there is also a quite literal sense in which empirical selves exist ‘in us,’ namely, in that they are among the constitutive parts of ourselves as human beings.

174

 ’      

representation of empirical selves as mind-independent since, in this construction, empirical time is presented as being mind-independent, as explicated in section 3.4.1.¹⁴¹ Empirical selves are not proper objects; in contrast to outer experience, inner experience does not conform to Kant’s formal conditions of objectivity, and, in contrast to empirical objects, empirical selves do not conform to all of Kant’s formal conditions of proper objecthood, for example, they are not in space. Nevertheless, empirical selves are distinguished from the intentional objects of dreams, hallucinations, illusions, and fictions, not only in that experience presents them as mind-independent, but also by having states whose order is correlated in law-like fashion with the objective temporal order that obtains with respect to the states of empirical objects and by being grounded in a properly direct way in entities that exist at the transcendental level of reality, namely, in transcendental selves, on account of which they possess the required additional ontological weight to count as genuine existents. Note that TI3-strong-b implies that empirical selves are grounded in things in themselves by characterizing them as intentional objects of inner experience. For inner experience depends on inner sensations, just like outer experience depends on outer sensations, and inner sensations, just like outer sensations, are due to affections of sensibility by things in themselves, namely, in this case, transcendental selves. The thesis of the transcendental ideality of empirical selves indicates another difference between Kant’s position and more traditional forms of idealism such as Berkeley’s, which Kant also calls “ordinary idealism.”¹⁴² While the general gist of the differences between Kant’s idealism and Berkeley’s idealism discussed in section 3.3.2 can be expressed by saying that critical idealism is, in various respects, less idealistic than ordinary idealism, Kant’s endorsement of TI3 reflects a respect in which critical idealism is more idealistic, so to speak. Proponents of ordinary idealism typically restrict their idealism to empirical objects as well as space and time but are committed to the transcendental reality of the selves whose determinations are revealed in our self-consciousness. For example, both Berkeley and Leibniz hold that, whereas empirical objects are fully mind-dependent, the self that we cognize through inner experience is a transcendentally real thing in itself, namely, an immaterial thinking substance or soul. By contrast, for Kant, the self whose determinations are disclosed to us in inner sense is transcendentally ideal just like empirical objects.¹⁴³ As noted in section 2.9, the empirical self is the object that is ontologically specified when I reflect on my representations in inner sense, recognize them all as mine, think of myself as a substance in which those representations inhere as its determinations, and am ¹⁴¹ As explicated in note 80, that we present our empirical selves as being in empirical time does not have to be taken to mean that our inner temporal determinations are among the temporal determinations from which empirical time is constituted. As indicated, I do not take them to be among these constitutive elements. Also note that there is some sense in which our empirical self can be said to be empirically mind-dependent, namely, in that it itself is an empirical mind and, thus, cannot exist at the empirical level of reality unless there is a mind at the empirical level of reality. But this is a very different sense of empirical mind-dependence compared to the empirical mind-dependence of the intentional objects of fictions, illusions, hallucinations and dreams that is at issue in the characterization of empirical idealism. ¹⁴² See, for example, B519, note. As will be explicated in section 5.6, ordinary idealism includes empirical idealism, as characterized in sections 3.3.1 and 3.4.1, as a central element but also a few more ontological theses, to be discussed in the indicated later section. ¹⁴³ See B519 520/A491 492. In this respect, Kant’s position has more affinity with Hume’s. Note that, as we will see in section 5.7, Kant seems to agree that our self-consciousness in thinking gives us some kind of cognitive access to a transcendentally real part of ourselves; but the self whose determinations are revealed in our selfconsciousness is the empirical self.

     

175

aware of my existence as determined in time relative to the objective temporal order of empirical objects. The empirical self that is constituted in this way is, thus, a mind. More specifically, empirical selves are immaterial or non-corporeal minds. In the transcendental Aesthetic, we have proved indisputably that bodies are mere appearances of our outer sense and not things in themselves. In conformity with this we can rightfully say that our thinking subject is not corporeal, that is, that, since it is represented by us as an object of inner sense, it, insofar as it thinks, cannot be an object of outer senses, i.e., it cannot be an appearance in space. (A357)¹⁴⁴

So, Kant is a dualist with respect to the empirical level of reality: there are material bodies and immaterial minds.¹⁴⁵ Kant’s commitment to dualism naturally raises a question that all dualists must face, namely, whether he regards minds and bodies in the empirical realm to be capable of interacting, a question that proves to be somewhat tricky. To anticipate, on my view, Kant is best read as an epiphenomenalist with respect to the empirical level. Empirical objects empirically causally influence empirical minds, but empirical minds do not empirically causally influence empirical objects. Prima facie, it might appear as if Kant is committed to the view that the mental, understood as immaterial or non-physical, causally influences the material or physical at the empirical level. For he has no qualms about treating our desires and inclinations as part of the empirical causal chains that determine our actions, and it might seem reasonable to assume that he conceives of those desires and inclinations as states of the empirical self and, hence, as mental.¹⁴⁶ On second glance, however, reasons emerge for being cautious about ascribing the view to him that the desires and inclinations that empirically determine our actions are mental states. There is something to be said for the reading that, as someone who values physics and regards its laws as universally valid with respect to the empirical world, that is, the public physical world, the spatiotemporal realm of bodies, Kant holds not only that the empirical world is deterministic in the sense ¹⁴⁴ See V-Met/Mron, 29:905: “That the soul is not matter can be discerned clearly; but it still may be the case that the substratum of matter is the same as the substratum of the soul. But the phenomena are different.” Also see V-Met-L1/Pölitz, 28:271–272. ¹⁴⁵ See A370: “The transcendental idealist, by contrast, can be an empirical realist, and thus a dualist, as one calls him, i.e., he can admit the existence of matter . . . ” See A379: “If one asks whether, accordingly, dualism alone holds sway in psychology, the answer is: indeed! but only in the empirical understanding; that is, in the connection of experience, matter as substance in the appearance is really given to outer sense, just as the thinking I, equally as substance in the appearance, is given before inner sense.” See A367; A357–358. ‘Dualism’ in this context obviously does not mean that there are only two things, but that there are only two fundamental kinds of things. ¹⁴⁶ See GMS, 4:453; KpV, 5:96–97. More generally, in several lecture notes he explicitly asserts that the soul affects the nervous system, in particular, the nerves in the brain, and that the body affects the soul by means of the nervous system. See V-Met/Mron, 29:909; V-Met-L1/Pölitz, 28:259–260. And since he regularly uses ‘soul’ loosely to refer to the empirical self (see note 197, chapter 2), we might take this as evidence for his endorsement of mind– body interaction at the empirical level of reality. On the other hand, it is clear neither if Kant is indeed thinking about the empirical self in these passages—he could be thinking about a thinking thing in itself—nor if he is speaking entirely in his own voice. Also, the second passage comes from lecture notes that presumably date from the mid-1770s, which is relatively early, and thus need to be taken with a grain of salt as a putative source of information about Kant’s critical views. But then, yet again, in the late essay on progress in metaphysics he says of the human body that “the object of inner sense . . . that is called ‘soul’ of the human being” is “connected with it” (FM, 20:309). To be sure, the assertion that the empirical self and the body are connected does not necessarily imply that they interact, but it does suggest it. Of course, it is also not implausible to assume that Kant is talking ‘with the vulgar’ here, so to speak, for ease of communication. All of this is to say that there is no conclusive direct textual evidence with respect to the question of whether the mind and the body at the empirical level of reality interact such that the direction of causal influence goes both ways.

176

 ’      

that every physical event is determined by an empirical cause that necessitates it according to a law but also that no physical event is empirically caused by anything that is not itself physical.¹⁴⁷ In this context, it is interesting to note how Kant characterizes the determinism that reigns in the empirical realm when discussing the third antinomy, which expresses an alleged incompatibility between freedom (thesis) and determinism (antithesis). He describes the position of the antithesis by saying that everything in the world, including our actions, “happens merely according to laws of nature” (B473/A445), that “there is nothing but nature in which we have to look for the connection and order of the events in the world” (B475/A447), and that there is “a connection according to general laws of appearances that determine each other necessarily, which one calls nature” (B479/A451). It is plausible to think that ‘nature’ here means ‘corporeal nature’ or ‘physical nature,’ that is, the sum total of all empirical objects and their states, governed by the laws of physics. If this is what Kant means, the alleged incompatibility at the heart of the third antinomy is between a world that contains free actions, understood as actions that have no physical causes, on the one hand, and a world that contains only actions that are physically caused, on the other hand. But this means, in turn, that Kant must understand the desires and inclinations that empirically cause our actions as physical states. Otherwise, we would not have to ‘go transcendental idealist’ and understand free actions as actions that are transcendentally uncaused to solve the antinomy by showing that, with free actions thus understood, the thesis is compatible with the antithesis after all, and both can be true at the same time.¹⁴⁸ Admittedly, at several places, most explicitly in the preface to the Metaphysical Foundations, Kant distinguishes between “two main parts of nature, of which one contains the objects of outer sense, the other the object of inner sense,” which is why the doctrine of nature can be divided into “the doctrine of body and the doctrine of the soul, of which the former considers extended nature, the latter thinking nature” (MAN, 4:467).¹⁴⁹ So, why should we not assume that by ‘nature’ in the antinomy passages Kant also means nature in the broad sense, which includes both material and immaterial nature? Here is a reason: in an intriguing passage in the preface to the Anthropology, Kant says quite explicitly that examining the natural causes that move human beings basically amounts to studying (or speculating about) the human brain. Physiological Anthropology is concerned with the research of what nature makes of the human being, pragmatic anthropology is concerned with what he as freely acting being makes of himself or what he can and ought to make of himself.—He who wonders about natural causes, e.g., about what the faculty of memory may be based on, can speculate back and forth (according to Descartes) about the remaining traces in the brain of the impressions that the received sensations leave; but in this he must admit that in this play of his representations he is a mere spectator and must let nature run its course in that he does not know the nerves and fibers of the brain nor knows how to use them for his ¹⁴⁷ Of course, on Kant’s view, physical events can also be caused by supersensible, free agents, but, in this case, we are talking about transcendental causes. See the discussion at the end of section 2.5.3. ¹⁴⁸ For a further explication of Kant’s general strategy for solving the third antinomy, see note 171, chapter 2, and note 2 above. ¹⁴⁹ See B712/A684: “This nature is twofold, either thinking or corporeal nature.”

     

177

purpose, and thus [he must admit] that all theoretical speculation about this is a mere loss. (Anth, 7:119)

Against the background of this passage, it is noteworthy that in the antinomy chapter Kant characterizes the task of explaining the possibility of freedom as “not physiological but transcendental” (B563/A535). If ‘physiological’ here is used in the same way as in ‘physiological anthropology’ in the Anthropology passage and not merely as a more fancy variation on ‘natural,’ understood in the broad sense from the Metaphysical Foundations passage, Kant would be implicitly admitting that the two relevant options for how to account for human actions are (a) by appeal to their transcendental causes and (b) by appeal to their physical causes. That Kant indeed has the sense of ‘physiological’ from the Anthropology in mind is strongly suggested when, later on in the antinomy chapter, he says that “with respect to this empirical character there is no freedom, but we can consider the human being only according to this empirical character if we merely observe and want to examine physiologically the moving causes for his actions, as it happens in anthropology” (B578/A550). So, assuming that Kant holds that the natural explanation of human actions is an explanation in terms of its physiological and, hence, physical causes, we have to conclude that the desires and inclinations that empirically determine our actions must be understood as states of our bodies, presumably of our brain. Of course, all of this is compatible with holding that Kant is committed to the existence of mental states at the empirical level of reality, that is, mental states understood as states of immaterial minds, which he clearly is, and even mental states that correspond to (but are not reducible to) the physical desires and inclinations that determine our actions. These mental states, which are determinations of our empirical selves, are the subject matter of the second kind of doctrine of nature, the doctrine of the soul.¹⁵⁰ But mental desires and mental inclinations are not part of the empirical causal chains that bring about our actions. The previous considerations do not speak against the possibility that Kant allows that physical events empirically cause mental events. And, as I read him, he in fact allows it. For starters, he often talks in a way that strongly suggests that he regards the empirical affection of our sense organs by empirical objects to be part of a causal chain that ends with sensations, understood as mental states.¹⁵¹ Moreover, as discussed in section 2.5.3, an important step in the construction of experience and the constitution of empirical objects consists in the transition from perceptions to more objective representations, which, in turn, crucially involves us representing empirical objects as the proximate causes of the sensations contained in our perceptions. That is, the representation of empirical causal ¹⁵⁰ See A379, quoted in note 145. ¹⁵¹ See Anth, 7:153–158; MAN, 4:476. Interestingly, there is also some evidence that Kant holds that for every mental state, and thus every sensation, there is a brain-state that corresponds to it. It is not uncommon for philosophers in the modern period to acknowledge the existence of so-called material ideas, which are conceived of as impressions on or motions of the brain and function as physical correlates or substrates for ideas in the soul. These ideae materiales are prominent in the Cartesian tradition, but one also reliably finds them in the writings of philosophers closer to Kant’s intellectual milieu, including Wolff, Baumgarten, and Platner. See Wolff 1734, §§112–21, §118, 88–92; Baumgarten 1757, §560, 297; Platner 1772, §231–9, 65–9. And, indeed, Kant himself mentions them at several additional places other than in his metaphysics lectures. See TG, 2:325, note; VA-Söm, 13:400; R4230, 17:468. Granted, Kant seems skeptical about whether the hypothesis of these material ideas is the right way to think about the physiological correlates of our mental states. But he appears to agree that there must be some kind of physiological correlates; see Söm, 12:32.

178

 ’      

influences from empirical objects on empirical minds is an integral part of experience from the start. And since experience ontologically specifies what is the case at the empirical level, we can thus conclude that there is physical-to-mental causation at the empirical level. Admittedly, given his view that a proper science of psychology is impossible, Kant probably does not regard the prospects for a science of physico-psychology as any rosier, and so presumably does not think that there could ever be a scientific account of how empirical objects cause sensations in us. But as long as experience presents empirical objects as causally efficacious in this way, they are thus causally efficacious, even if, due to the present and future absence of a scientific account, the causal mechanism and the laws governing this kind of causation are indeterminate. In section 3.5, we examined several reasons that speak against the two-aspect reading of the transcendental distinction between outer appearances and things in themselves. But one might think that, whereas the two-world reading has clear advantages over the twoaspect reading in the ‘outer’ case, the two-aspect reading is clearly preferable in the ‘inner’ case, simply because each one of us is only one, not two. This is a fair challenge to the twoworld view. I will respond to it in section 5.8, at which point we will have sufficiently familiarized ourselves with certain peculiarities of the grounding relation that obtains between transcendental selves and empirical selves, peculiarities that will play a role in my response. This concludes my description and explanation of the core theses of Kant’s transcendental idealism and empirical realism. Keep in mind, though, that these theses do not fully capture his ontological position overall yet. We will examine further core tenets of Kant’s critical idealism, about things in themselves and their grounding of appearances, in chapter 5, after having taken a look at the foundational structure of and Kant’s main arguments for his transcendental idealism and empirical realism in chapter 4.

4 Kant’s Case for Transcendental Idealism and Empirical Realism 4.1 The Foundational Structure of Transcendental Idealism and Empirical Realism 4.1.1 Arguments for the Transcendental Ideality of Empirical Objects and Empirical Selves Based on the Transcendental Ideality of Space and Time In our project of trying to explicate the nature and meaning of Kant’s transcendental idealism and empirical realism, we have so far mainly focused on identifying the content of these doctrines by looking at what Kant more or less explicitly tells us about them. It is time to consider some arguments and to examine the logical relations between the theses of transcendental idealism and empirical realism identified in chapter 3, thereby revealing the foundational structure of the part of Kant’s critical idealism that is defined by these theses. Although we will be working through various arguments, the main purpose of this chapter remains primarily exegetical. That is, our immediate project is to identify how Kant conceives of his case for transcendental idealism and empirical realism, since this will reinforce and further illuminate the interpretation developed so far. Whether Kant’s case is ultimately convincing is a question for another day, but at the very least the following reconstruction should be a useful prolegomenon for any future all-out defense attempt.¹ One can find several different arguments in the Critique for the first core thesis of transcendental idealism, TI1, the thesis of the transcendental ideality of empirical objects, according to which empirical objects are not things in themselves but appearances. We have already encountered two of these arguments: the argument based on the antinomies, which Kant sketches in the antinomy chapter, and the argument based on the general question of how a priori synthetic judgments are possible, which Kant sketches in the B-preface. Adopting TI1 allows us to solve the antinomies and to answer the famous possibility question. Kant describes the argument in the antinomy chapter as a kind of ‘fallback’ demonstration, “in case somebody is not satisfied by the direct proof in the Transcendental Aesthetic” (B534/A506). This suggests that he takes his main argument for transcendental idealism to be contained in the Transcendental Aesthetic. As he puts it,

¹ As we will see, whether Kant’s case is deemed ultimately convincing depends primarily on how convincing his argument for the thesis of the transcendental ideality of space and time is taken to be, which we will examine in section 4.2. The World According to Kant: Appearances and Things in Themselves in Critical Idealism. Anja Jauernig, Oxford University Press (2021). © Anja Jauernig. DOI: 10.1093/oso/9780199695386.003.0004

180

 ’        

“[i]n the transcendental Aesthetic we have proved sufficiently that everything that is intuited in space or time, and thus all objects of an experience that is possible for us, are nothing but appearances, i.e., mere representations . . . (B518–519/A490–491)²

This demonstration in the Transcendental Aesthetic of the theses of the transcendental ideality of empirical objects (TI1), comprises two sub-arguments. The first sub-argument consists in the derivation of TI1 from TI2, the thesis of the transcendental ideality of space and time. The second sub-argument consists in a demonstration of TI2. An exactly analogous demonstration can be provided for the thesis of the transcendental ideality of empirical selves (TI3). So, while the theses of the transcendental ideality of empirical objects and empirical selves may be the most striking claims of Kant’s critical idealism, the most fundamental claim is the thesis of the transcendental ideality of space and time. Indeed, TI2 is arguably one of the most important doctrines, if not the most important doctrine, of Kant’s theoretical philosophy overall. This assessment is confirmed by Kant’s s declaration that “the doctrine of the ideality of space and time” is one of the “two hinges” around which metaphysics turns (FM, 20:311)—the other hinge being “the doctrine of the reality of the concept of freedom,” which is one of the most important doctrines, if not the most important doctrine, of Kant’s practical philosophy—and by his pronouncement that an alternative, better name for his critical idealism would be “the principle of the ideality of space and time” (Letter to Beck, December 2, 1792, 11:395). Against the foil of statements like these, it must be regarded as a serious flaw of any interpretation of Kant’s critical idealism if it does not allocate a prominent position to the thesis of the transcendental ideality of space and time.³ Given the fundamental place of TI2 in Kant’s critical idealism, the argument for it must bear a lot of weight and deserves a separate discussion, which we will take up in section 4.2. For present purposes, we will assume TI2 and all more specific versions of TI2 as given, and examine how TI1 and TI3 can be justified on their basis. We will also assume that empirical objects are in space and time and thus have spatial and temporal determinations, and that empirical selves are in time and thus have temporal determinations, claims that I take to be uncontroversial. There is room for disagreement about the nature of space and time and their precise relation to empirical objects—one could be an idealist or realist, or a relationalist or substantivalist, for example. But the basic claims that empirical objects are in space and time, and empirical selves are in time, seem to be universally agreed upon by everybody involved in the debate and confirmed by incontrovertible empirical evidence. As before, unless noted otherwise, all Kantian theses and related claims that we will discuss in this chapter are to be understood as articulated from the point of view of fundamental ontology. Also, ‘thing in itself,’ ‘appearance,’ and ‘mind-(in)dependent’ without any further qualifications are to be understood in the transcendental sense. If we grant that empirical objects are in space and time, there is a straightforward argument that leads to the conclusion that empirical objects are not things in themselves from part (a) of TI2-empirical, the claim that empirical space and time, that is, the space

² Also see A357: “In the transcendental Aesthetic we have proved indisputably that bodies are mere appearances of our outer sense and not things in themselves.” ³ Rae Langton’s reading, which has many virtues otherwise, is an interpretation of this kind, for example; on her account, the transcendental ideality of space and time plays virtually no role at all; see Langton 1998.

 

181

and time in which empirical objects are in the empirical world, are neither things in themselves nor constituted by determinations of things in themselves. For this claim entails that things in themselves are not in empirical space or empirical time. 1. Empirical objects are in empirical space and empirical time. (Uncontroversial premise.) 2. Things in themselves are not in empirical space or in empirical time. (From TI2-empirical-a.) 3. Empirical objects are not things in themselves. (TI1-a, from 1 and 2, and so-called Leibniz’s law, according to which things that have different properties are numerically distinct.) An exactly similar argument can be formulated for the TI3-a, the thesis that empirical selves are not things in themselves, given that empirical selves uncontroversially are in empirical time. For efficiency’s sake, I will only explicitly discuss the argument for TI1-a. This argument is fine as far as it goes, but it calls for a few comments.⁴ First, in order for the argument to be valid within the framework of Kant’s own theory—as opposed to being merely useful as a dialectical weapon in the debate with the transcendental realist—it must be legitimate for Kant to assert (1) from the point of view of fundamental ontology. For, obviously, (2) and (3) are claims from this fundamental perspective. On the interpretation presented here, this assertion is indeed legitimate for Kant, as explicated in section 3.4.5. Empirical objects genuinely are in space and time and genuinely have spatial and temporal determinations, as opposed to merely appearing or being represented in this way. Second, the argument only gets us part of the way to the ultimate goal—where, in this case, the ultimate goal is the thesis of the transcendental ideality of empirical objects. The conclusion that empirical objects are not things in themselves (TI1-a) means that they are not mind-independent, and, thus, that at least some of their determinations or other ontological ingredients, or even their existence, are mind-dependent. But it does not allow us to infer that they are fully mind-dependent (TI1-b), as defined in section 2.3, that is, that their existence and all of their ontological ingredients, including all of their determinations, are mind-dependent. In order to arrive at this stronger conclusion, we have to appeal to a different premise, namely, TI2-pure-b, the thesis that pure space and time are fully mind-dependent. Ditto for TI3-b, the claim that empirical selves are fully mind-dependent. Third, as we will see, within Kant’s scheme, the justification for TI2-empirical-a, on which the sketched argument depends, also ultimately rests on TI2-pure-b, the claim that pure space and time are fully mind-dependent. So, since in order to fully justify the sketched argument we thus have to appeal to TI2-pure-b, and since, as just noted, TI2-pure-b is strong enough to underwrite a stronger conclusion, namely, TI1-b, which, moreover, is the conclusion that we are ultimately interested in, we may as well directly go

⁴ In addition to the three comments to be considered presently, proponents of the one-object view may also want to object that the argument is invalid since Leibniz’s law is valid only in extensional contexts, but the attribution of spatial or temporal predicates creates an intensional context. We already disposed of this kind of objection in section 3.5.

182

 ’        

after this stronger result, based on TI2-pure-b, and set the sketched argument aside. So, then, let us proceed to this task. Since TI1-b (the claim that empirical objects are fully mind-dependent) entails TI1-a (the claim that empirical objects are not things in themselves), in order to justify TI1 it is enough for us to justify TI1-b. The same holds for TI3, the corresponding thesis about empirical selves, which can be justified by justifying TI3-b (the claim that empirical selves are fully mind-dependent). Again, for efficiency’s sake, I will only explicitly discuss the justification for the thesis of the transcendental ideality of empirical objects; the justification for the thesis of the transcendental ideality of empirical selves is exactly parallel. Also note that, since TI2-pure-strong-b (the claim that pure space and pure time are [nothing but] forms of sensibility) entails TI2-pure-b (the claim that pure space and pure time are fully mind-dependent), TI2-pure-strong-b could also serve in place of TI2-pure-b as one of the main premises in the argument for TI1 that we are about to examine. TI2-pure-b is the main load-bearing beam of the argument for TI1, but it needs to be supplemented with two additional premises in order for us to be able to conclude that empirical objects are fully mind-dependent. The first of these extra premises is needed to underwrite the intermediary conclusion that empirical objects are not only in empirical space and empirical time but also in pure space and pure time. For if this were not the case, we could not draw any conclusions about the mode of being of empirical objects from the mode of being of pure space and pure time. The required premise is the claim that any determinate space, that is, any determinate figure and any determinate spatial volume, and, more generally, any spatial determination is necessarily contained in pure space, and any determinate time, that is, any determinate temporal duration, and, more generally, any temporal determination is necessarily contained in pure time.⁵ For future reference, I will call this claim ‘S-.’ There are two closely related possible derivations of the conclusion that empirical objects are in pure space and pure time based on S- and the premise that empirical objects are in empirical space and empirical time. First, given that empirical objects are in empirical space and thus fill determinate spatial volumes and stand in various spatial relations to one another, and given that empirical objects are in empirical time and thus fill determinate durations and stand in various temporal relations to one another, S- allows us to conclude that empirical objects, via their spatial and temporal determinations, are as much in pure space and pure time as they are in empirical space and empirical time. Second, since empirical space is a determinate space or at least constituted by spatial determinations, and since empirical time is a determinate time or at least constituted by temporal determinations, S- implies that empirical space and empirical time are necessarily contained in pure space and in pure time, respectively.⁶ On the plausible assumption that if A is in B and B is contained in C, then A is in C, we can then infer from the fact that empirical objects are in empirical space and empirical time that they are also in pure space and pure time. In sections 3.4.3 and 3.4.4, we already got ⁵ For Kant’s usage of the terms ‘determinate space’ and ‘determinate time’ in this sense, see B137–138; B202–203. ⁶ More strongly, since every space can contain determinate spaces and other spatial determinations, and since every time can contain determinate times and other temporal determinations, S- implies that every space is either identical to or necessarily contained in pure space, and every time is either identical to or necessarily contained in pure time.

 

183

glimpses of Kant’s reasons for endorsing S-; we will spell out an explicit justification for it in section 4.2.3.4. For now, we just register that it is needed to derive TI1 from TI2. A second additional premise is needed for our argument because the inference to the claim that empirical objects are fully mind-dependent from the full mind-dependence of the space and time in which they are is not justified on every possible conception of the relation between space and time and the objects in them. Views on which empirical objects are ontologically prior to space and time quite obviously leave room for the possibility that space and time are fully mind-dependent but empirical objects have mind-independent determinations or other mind-independent ontological ingredients, and thus are not fully mind-dependent according to our definition. But even views on which space and time are ontologically prior to empirical objects do not automatically rule out the unwanted possibility although this is less obvious in their case. As indicated in section 3.4.4, substantivalists and Kant agree that space and time are ontologically prior to empirical objects in the sense that the former are among the existence conditions for the latter, and they also agree that this ontological priority manifests itself in, among other things, that (1) space and time can have being even if there are no empirical objects in them, while empirical objects cannot have being unless they are in space and time, and (2) empirical objects have all of their space-infected and time-infected determinations, including all of their spatial and temporal determinations, partly in virtue of being in space and time, while space and time do not have any of their determinations in virtue of, or partly in virtue of, the empirical objects in them.⁷ Claim (2) justifies the inference from the full minddependence of space and time to the claim that all space-infected and all time-infected determinations of empirical objects are mind-dependent. But it leaves open the possibility that empirical objects have other determinations or other ontological ingredients that are mind-independent.⁸ One possible strategy for closing this loophole would be to insist that, if an object is in space/time, it not only has some spatial/temporal determinations but is literally contained in space/time in the sense of being fully immersed in, or enveloped by, it such that all of its determinations and ontological ingredients are space-infected/time-infected. For a while, I had myself convinced that this was indeed Kant’s argumentative strategy but, upon further consideration, I came to the conclusion that the claim that no object in space/time can have any determinations or ontological ingredients that are not space-infected/timeinfected is simply not very plausible.⁹ So, we will modify the proposed strategy a bit. We ⁷ As you will recall from section 3.4.4, a determination is space-infected/time-infected if, and only if, it is an instance of a property that is space-infected/time-infected, and a property is space-infected/time-infected if, and only if, it is essentially bound up with space/time in that objects can have the property only if they are in space/ time. ⁸ TI2-pure-strong-b (the claim that space and time are [nothing but] forms of sensibility) is stronger than TI2-pure-b but, by itself, is also not sufficient for a justification of TI1-b. Since the forms of sensibility essentially are forms of presentational contents, TI2-strong-b allows us to conclude that anything in pure space and pure time, anything of which pure space and pure time are forms, is a presentational content. So, on the assumption of S-, it can be inferred that those aspects of empirical objects that are in space and time, that is, those aspects that are constituted by their space- and time-infected determinations, are presentational contents and, hence mind-dependent. But, again, we are not yet in a position to rule out that empirical objects have some other determinations or other ontological ingredients that are mind-independent. ⁹ For example, one might hold that every empirical object, on Kant’s view, includes a bare particularizing ontological ingredient in which all of its determinations inhere. But it is hard to make sense of the claim that bare particulars are somehow space-infected or time-infected, precisely because they are bare. Similarly, many

184

 ’        

will retain the idea that, if an object is in space/time, it is fully immersed in it, but we will cash out what this full immersion entails in a slightly weaker, more plausible way. We will say that if x is in space S/time T, x is fully immersed in S/T such that x has some spatial/ temporal determinations and, possibly, some space-infected/time-infected determinations, and the mode of being of all of x’s determinations and other ontological ingredients is the same as the mode of being of S/T. For future reference, I will call this claim ‘S.’ People who are committed to the ontological priority of space and time compared to empirical objects are not forced to accept S-, nor are people who are committed to the ontological priority of empirical objects compared to space and time forced to reject it. But S- fits much better into the portfolio of the former group. The view that the mode of being of objects in space and time is inherited from, or determined by, the mode of being of space and time provides a plausible justification for S- and is another natural way in which one could see the ontological priority of space and time compared to empirical objects manifest itself (in addition to the ways described by (1) and (2) above at the beginning of this paragraph). And, indeed, substantivalists typically do accept S-—as does Kant, on my reading. So, here, then, is a reconstruction of an argument for the thesis of the transcendental ideality of empirical objects based on the thesis of the transcendental ideality of space and time. 1. Pure space and pure time are fully mind-dependent. (TI2-pure-b.) 2. Any determinate space and, more generally, any spatial determination is necessarily contained in pure space, and any determinate time and, more generally, any temporal determination is necessarily contained in pure time. (S.) 3. Empirical objects are in empirical space and empirical time. (Uncontroversial premise.) 4. Empirical objects are in pure space and pure time. (From 2 and 3, as explicated above.) 5. If x is in space S/time T, x is fully immersed in S/T, which means, among other things, that the mode of being of all of x’s determinations and other ontological ingredients is the same as the mode of being of S/T. (From S-.) 6. All of the determinations and ontological ingredients of empirical objects are minddependent. (From 1, 4, and 5.) 7. Empirical objects are fully mind-dependent. (From 6 and the definition of ‘fully mind-dependent.’) In short and more informally, empirical objects are fully mind-dependent because (a) they are in pure space and time, (b) any object in space and time is fully immersed in them and thus shares their mode of being, and (c) pure space and time are fully mind-dependent. I take it that this line of reasoning is what Kant has in mind when he argues, as he often does, that since space and time are ‘nothing but representations,’ that is, fully mindempirical objects have the property of being a ground of some kind. But while the determination of being an empirical cause is certainly time-infected, the determination of being a ground arguably is not.

 

185

dependent intentional objects, anything that is in space and time must also be a ‘mere representation,’ that is, a fully mind-dependent intentional object. One must well note this paradoxical but true proposition: that nothing is in space other than what is represented in it. For space itself is nothing but representation; consequently, what is in it must be contained in the representation, and nothing at all is in space except insofar as it is really represented in it. This is a proposition that must indeed sound odd, that a thing can exist only in its representation, but which loses here what is objectionable because the things with which we are dealing are not things in themselves but merely appearances, i.e., representations. (A374, note)¹⁰

Note that TI2-pure-b, plus S-, and the result, derived from S-, that empirical space and empirical time are necessarily contained in pure space and pure time, respectively, also enable us to conclude that empirical space and empirical time are fully mind-dependent (TI2-empirical-b),¹¹ which, in turn, implies that empirical space and empirical time are neither things in themselves nor constituted by determinations of things in themselves (TI2-empirical-a). This is the justification of TI2-empirical-a based on TI2-pure-b that I mentioned above in my third observation about the argument for TI1-a based on TI2-empirical-a with which we began our discussion in the present section of how TI1 can be deduced from TI2. So much for the derivation of the transcendental ideality of empirical objects from the transcendental ideality of space and time. How about TI1-strong, and, in particular, TI1-strong-b, the thesis that empirical objects are Kantian outer appearances, that is, intentional objects of outer experience as characterized in Kant’s theory of experience? We can move from the claim that empirical objects are appearances in the general sense of being fully mind-dependent (TI1-b) to TI1-strong-b, if we grant, or want to be able to hold on to, the common-sense view that empirical objects are proper objects that genuinely exist. Given the nature of our cognitive faculties, the only kind of representation that can ontologically specify proper objects that genuinely exist is outer experience as characterized in Kant’s theory, as discussed in section 2.5.3. So, if empirical objects are fully minddependent, the only way that we can assert that they are proper objects that genuinely exist is if we identify them with intentional objects of outer experience as characterized in Kant’s theory. Less drily put, given that genuine idealism about empirical objects is true (TI1-b), if we do not want to rule out the existence of the empirical world—that is, if we do not want to be like Berkeley (as Kant reads him)—we must adopt Kant’s conception of empirical objects (TI1-strong-b). If we grant that empirical selves genuinely exist, a similar train of thought gets us from the thesis that empirical selves are appearances in the general sense ¹⁰ Also see A375: “ . . . that is, something real in space corresponds to our outer intuitions. Of course, space itself, with all its appearances, as representations, is only in me, but in this space still the real, or the material of all objects of outer intuition, is given really and independently from all fantasy, and it is also impossible that in this space something outside us (in the transcendental sense) should be given, since space itself outside our sensibility is nothing.” ¹¹ Also note that we could derive TI2-empirical-a from TI2-pure-b and S- by relying on a different assumption instead of S-, namely, that (a) pure space/time is ontologically prior to empirical space/time, where this is understood as implying minimally that empirical space/time has all of its space-infected/time-infected determinations partly in virtue of being in pure space/time, and (b) any kind of space/time is necessarily fully spatial/temporal in character such that it is impossible for there to be a space/time that has any ontological ingredients other than space-infected/time-infected determinations.

186

 ’        

(TI3-b) to the thesis that empirical selves are Kantian inner appearances, that is, intentional objects of inner experience as characterized in Kant’s theory of experience (TI3-strong-b). Kant himself does not spell this out but we could say that, according to his analysis of our cognitive capacities, inner experience, as characterized in his theory, is the only kind of representation that can ontologically specify immaterial objects that genuinely exist. So, given that empirical selves are fully mind-dependent, if we want to accommodate the view that empirical selves genuinely exist, then we must conceive of them as intentional objects of inner experience as characterized in Kant’s theory.

4.1.2 Arguments for Empirical Realism Once it is settled that empirical objects are intentional objects of outer experience (TI1-strong-b) and empirical selves are intentional objects of inner experience (TI3-strong-b), the theses that empirical objects and empirical selves are empirically mind-independent (ER1-ii and ER3-ii), empirical space and empirical time are empirically mind-independent (ER2-ii-), and things in themselves in the empirical sense are in space and time and, thus, have spatial and temporal determinations (ER2-ii-), directly follow as well. For, since experience ontologically specifies empirical objects and empirical selves, all it takes for them to be empirically mind-independent is that experience presents them in these ways, which, in fact, it does, as discussed in sections 3.3.1 and 3.6. And since it is uncontroversial that empirical objects are in space in space and time, the result that empirical objects are empirically mind-independent and, thus, things in themselves in the empirical sense directly entails that things in themselves in the empirical sense are in space and time and, thus, have spatial and temporal determinations. Moreover, on the assumption that empirical space and empirical time are the sum total of all spatial and temporal determinations of empirical objects—a view that, as we have seen, Kant endorses for a variety of reasons, including that substantivalism, even restricted to the empirical realm, is inherently implausible—the empirical mind-independence of empirical objects directly entails the empirical mind-independence of empirical space and time too. Incidentally, Kant’s relationalist conception of empirical space and time and TI1-b (the thesis that empirical objects are appearances) also directly imply that empirical space and empirical time are constituted by determinations of appearances, as stated in TI2-empirical-strong-b. This almost completes our project of justifying and unearthing the logical relations between the various theses that define Kant’s transcendental idealism and empirical realism. All that is left to do—apart from identifying Kant’s arguments for TI2-pure-b and TI2-pure-strong-b, which we will attempt in section 4.2—is to reconstruct possible justifications for ER1-i, ER2-i, and ER3-i, the claims that empirical objects, empirical space and time, and empirical selves exist. As already noted in section 3.4.3, on the assumption of Kantian relationalism about empirical space and time, the thesis that there exist empirical objects (ER1-i) provides a sufficient basis for concluding that empirical space and empirical time exist (ER2-i). So, our sole remaining task in this section is to examine how Kant can justify ER1-i and ER3-i—which proves to be a little tricky. More precisely, the justification for ER1-i proves to be tricky. The justification for ER3-i is comparatively straightforward. We are justified in asserting that empirical selves exist because each one of us is assured of the existence of our own empirical self on the basis of

 

187

our (empirical) self-consciousness.¹² On first glance, it may seem that providing a justification for the thesis of the existence of empirical objects is not a difficult assignment either. For once we have established that empirical objects are intentional objects of outer experience (TI1-strong-b), their existence directly follows from us having outer experience, which, one might hold, we obviously do have. That was easy. But the problem is that, on closer inspection, it turns out to be far from clear that we have outer experience of the kind described in Kant’s theory. This is not at all obvious because we do not know if the alleged outer perceptions that feed into the construction of what we take to be outer experience are, indeed, genuine perceptions, rather than what I will call ‘pseudo-perceptions.’ As briefly discussed in section 2.4, genuine perceptions are the representations that our mind generates by means of the application of a certain arsenal of cognitive unification operations encapsulating the categories of quantity and quality to genuine sensations, that is, qualia-representations that are the direct result of transcendental affections of sensibility by things in themselves. Pseudo-perceptions are representations that we mistakenly regard as genuine perceptions but that our mind generated in other ways, for example, by means of the application of cognitive operations to genuine sensations that differ from the operations employed in genuine perception, as in the case of illusions, or by means of the application of various cognitive operations to pseudo-sensations, that is, sensations that are just like genuine sensations except for not being the direct result of affections of sensibility by things in themselves, as in the case of hallucinations or dreams.¹³ Genuine experience is based on genuine perceptions; what I will call ‘pseudo-experience’ is based on pseudo-perceptions. I take it that a growing recognition of the epistemic problem of how we can be certain that we have genuine outer experience rather than outer pseudoexperience is the reason for Kant’s prolonged efforts to come up with a satisfactory refutation of what he calls ‘problematic idealism,’ that is, the kind of idealism exemplified by Descartes, or, more precisely, Descartes in the first two Meditations. The problematic idealist holds that the existence of empirical objects ‘outside us’ is dubitable and uncertain, while we can be certain of our own existence based on the consciousness of our representations or based on inner experience. Kant proposes (at least) three, increasingly sophisticated versions of an argument against this kind of problematic idealism that are meant to show that the existence of empirical objects is as certain as the existence of our self. The first version can be found in his discussion of the fourth paralogism in the A-edition of the Critique; the second, improved version is contained in §49 of the Prolegomena; the third, even further improved version constitutes the so-called refutation of idealism in the B-edition of the Critique. All versions of the argument crucially rely on the classification of empirical objects as appearances (TI1-b), and, more specifically, as intentional objects of outer experience (TI1-strong-b), and centrally include the inference to the existence of empirical objects from the claim that we have outer experience of the kind described in Kant’s theory. So, let us take a closer look. ¹² Textual evidence for the reading that Kant is committed to the indicated justification of ER3-i will be presented shortly, in connection with our discussion of his various justification attempts for ER1-i. Note that the justification for ER3-i is not quite as straightforward as it looks, though; see the comment in the following note. ¹³ But is there not a similar problem in the case of inner experience? How can we be sure that we indeed have inner experience as characterized by Kant? It would lead us too far afield to provide an exhaustive discussion of this question here, but the short answer, I take it, is that, given how our mind, in particular, inner sense, works, inner pseudo-perceptions are impossible. Providing a satisfactory defense of this answer is a tall order that goes beyond the scope of the present investigation.

188

 ’        

In the section on the fourth paralogism in the A-edition, Kant argues that, since empirical objects ‘outside us’ (in the empirical sense) are nothing but intentional objects of our representations that are presented as being in space, their existence is as indubitable as the existence of these representations, of which we are certain based on the testimony of our self-consciousness. But this means that the existence of objects ‘outside us’ is as indubitable as our own existence, of which we are also certain based on the testimony of our self-consciousness. For I am conscious of my representations; hence, they and I myself exist who has those representations. But outer objects (bodies) are mere appearances and thus also nothing but a kind of my representations, whose objects are something only through these representations but nothing separated from these representations. Therefore, outer objects exist just as well as I exist myself, and both do so on the testimony of my selfconsciousness, only with the difference that the representation of my self, as the thinking subject, is merely related to the inner sense, but the representations that signify extended beings are also related to the outer sense. (A370)

This argument is not an unqualified success—which is presumably why Kant removed it from the B-edition. Surely, it is not the case that the intentional objects of any kind of actually existing representations that present their objects as being in space exist. Joe’s hallucinated oasis and the purple dragon from my dream are both presented as being in space in actually existing representations but none of them exists. (They exist in their respective intentional ‘worlds’ but only pseudo-exist from the point of view of fundamental ontology.) The version of the argument from §49 of the Prolegomena improves on the A-edition version precisely with respect to this point. Kant emphasizes there that the representations on which outer empirical objects depend are special in that they present their objects as part of a law-governed system, which “proves their objective truth, just as the connection of the appearances of inner sense proves the reality of my soul (as an object of inner sense)” (Prol, 4:336). In other words, the doubt about whether, not only our soul as object of inner experience but also the objects of outer experience exist can be easily removed . . . by examining the connection of appearances in both according to general laws of experience, and, if the representation of outer objects thoroughly coheres with it, we cannot doubt that they should not [sic.] constitute veridical experience. (Prol, 4:337)

Kant seems to say here that the fact that the intentional objects of outer experience form a unified, coherent system that is governed by various laws allows us to distinguish between them and the intentional objects of dreams and hallucination and provides us with a justification for the claim that outer experience attests to the existence of empirical objects ‘outside us’ (in the empirical sense) just as inner experience, which also presents its object as governed by certain laws, attests to the existence of our empirical self. In brief, Kant’s argument appears to be that the conformity of outer experience to his formal conditions of experience and objectivity, as discussed in section 2.5.2, ensures that the existence of

 

189

empirical objects is directly given through outer experience, just as the existence of our empirical self is directly given through inner experience. Although this version of the argument for ER1-i is certainly better than the version of the argument in the paralogism chapter of the A-edition, it is not perfect either. Kant is right to claim that experience is distinguished from all typical illusions, hallucinations, and dreams through its conformity to his formal conditions of experience. These kinds of representations tend not to present their intentional objects as governed by universal causal laws, for example. But this is not enough to clinch the case. First, there are other kinds of representations that are not experience but still conform to these formal conditions. At this juncture, it is a natural thought that we might be able to fix up Kant’s argument on his behalf by appealing to not only the formal but also the material conditions of experience. After all, as discussed in section 2.5.3, the formal and material conditions taken together appear to be sufficient to distinguish between experience and all other kinds of representations. The second problem is, that, unfortunately, for our present purposes, this ploy turns out to be a dead end as well. In our previous discussion, we were interested in the conditions that a representation must satisfy in order to be experience. But in the present context, we are interested in conditions that would allow us to tell whether a given representation is experience. And in contrast to the formal conditions of experience, not all of the material conditions are usable as such epistemic criteria. More specifically, neither the formal conditions of experience nor the material conditions would enable us to distinguish genuine experience, which is based on genuine perceptions, from pseudo-experience, which is based on mere pseudoperceptions, such as might be supplied through illusions, hallucinations, or dreams. Experience and pseudo-experience are both constructed in such a way as to conform to Kant’s formal conditions of experience, and they both appear to us to conform to the material conditions as well. In order to tell them apart, we would have to be able to discriminate between genuine perceptions and pseudo-perceptions, which, however, appears to be beyond us. As I see it, with this problem we are touching on one of the shiftiest elements of Kant’s philosophy, an element with which Kant himself continued to struggle from the publication of the A-edition, through the Prolegomena, to the B-edition, and beyond.¹⁴ The crucial question, in the context of justifying the claim that some empirical objects exist, is how we could ever be sure that what we take to be experience indeed is genuine experience, rather than a unified, systematic account of a merely imaginary world that satisfies Kant’s formal conditions of experience but is based on pseudo-perceptions. One strategy to respond to this question in the context of Kant’s discussion in the Prolegomena would be to backpedal a little and admit that we cannot conclusively prove that we have genuine experience but insist, based on an inference to the best explanation, that the conformity of what we take to be experience to Kant’s formal conditions of experience at least justifies us in assuming that, for the most part, it really is genuine. If all or most of the representations that we take to be perceptions were mere dreams, illusions, or hallucinations, it would be close to a miracle that we manage to construct on their basis a unified, coherent account of

¹⁴ There are a number of reflections, dated later than the publication of the B-edition, which show Kant still hard at work at finding the proper formulation for his argument for the existence of outer empirical objects. See R5653–5655, 18:306–316; R6311–6316, 18:607–623; R6319, 18:633; R6323, 18:642–643.

190

 ’        

the world that conforms to Kant’s formal conditions of experience, an account whose possibility depends on the existence of strong, persistent regularities in the content of these representations. By contrast, if the representations that we take to be perceptions indeed are genuine, and thus, in part, due to the affections of sensibility by things in themselves, we can explain the regularities exhibited by their content as reflecting certain (unknown) regularities in the realm of things in themselves.¹⁵ Another strategy for how to respond to our question is implemented in Kant’s version of his argument for ER1-i in the B-edition of the Critique, the refutation of idealism. So, again, we find that in his later attempt to demonstrate the existence of empirical objects, Kant sets out to improve the argument precisely at the place where, according to our analysis, the earlier version runs into trouble. The version of Kant’s argument for ER1-i in the Refutation of Idealism is significantly more sophisticated than the earlier versions, but it also crucially turns on the ontological status of empirical objects as intentional objects.¹⁶ In a nutshell, Kant argues that inner experience depends on outer experience, and since the Cartesian grants that it is indubitable that we have inner experience, we can conclude that we also have outer experience and not just outer pseudo-experience. The required proof thus must establish that we also have experience of outer objects and not merely fantasy, which cannot very well happen unless one can demonstrate that even our inner experience, which is indubitable to Descartes, is only possible under the presupposition of outer experience. (B275)

More specifically, according to Kant’s argument, the representation of a determinate temporal order of our inner states, including, in particular, the representation of determinate durations, is a core element of our inner experience. But the representation of such a temporal order is possible only if we can represent something that persists in time, with respect to which this order can be determined. But, for reasons to be addressed below, we can represent something that persists in time only in outer experience. So, if we assume that we have inner experience, which Descartes admits, we are justified in concluding that we have outer experience as well.¹⁷ But this means that we can be as certain of the existence of empirical objects as of the existence of our self because empirical objects are nothing but intentional objects of outer experience.

¹⁵ See B723–24/A695–96: “If one asks . . . whether there is something distinct from the world that contains the ground of the order of the world and its connection according to general laws, the answer is: without a doubt. For the world is a sum of appearance, so there must be some transcendental ground of it, i.e., a ground that is thinkable only for the pure understanding.” ¹⁶ The literature on Kant’s refutation of idealism is vast; a great place to start is Paul Guyer’s detailed discussion in Guyer 1987, 279–329. ¹⁷ Note that, as Kant repeatedly points out himself, this argument does not show and is not intended to show that we cannot ever be mistaken about particular representations that we take to be genuine perceptions. See B278–279: “From the fact that the existence of outer object is required for the possibility of a determinate consciousness of ourselves, it does not follow that every intuitive representation of outer things at the same time includes their existence, for this representation can very well be the mere result of the imagination (in dreams as well as in insanity); but it is it only through the reproduction of former outer perceptions, which, as has been shown, are possible only through the reality of outer objects. Here we only wanted to prove that inner experience in general is possible only through outer experience in general. Whether this or that presumed experience may be mere fantasy must be figured out according to its special determinations and through comparison with the criteria of all real experience.” Also see Bxli.

 

191

I am conscious of my existence as determined in time. All determination of time presupposes something persistent in perception. This persistent thing, however, cannot be an intuition in me. For all grounds of the determination of my existence that can be encountered in me are representations and require, as such, themselves some persistent thing that is distinct from them, with respect to which their change, and thus my existence in time, in which they change, can be determined. Hence, the perception of this persistent thing is possible only through a thing outside of me and not through the mere representation of a thing outside of me. Therefore, the determination of my existence in time is possible only through the existence of real things, which I perceive outside me. (B275–276, with Kant’s corrections, as stated in the preface at Bxxxix)¹⁸

The most obvious question about this argument is on what grounds Kant can maintain the crucial claim that we can represent something persistent in time only on the basis of genuine perceptions, in outer experience but not on the basis of mere “fantasies,” hallucinations, or dreams. In my assessment, the ultimate reason for this claim is not very far from the inference to the best explanation that I sketched above on Kant’s behalf as a supplement to the Prolegomena-version of the argument. Presenting something as persistent in time requires genuine perception and outer experience, not only because our representations themselves are constantly changing and do not persist in time but ultimately also because the desired representation of something persistent requires that the perceptions on which it is based exhibit strong, persistent regularities as far as their content is concerned, regularities that would be nothing short of miraculous if these perceptions in truth were mere dreams, hallucinations, illusions, or fantasies. By contrast, if they are genuine perceptions, that is, if they are due, in part, to affections of sensibility by things in themselves, the persistent, strong regularities in their contents can be seen as reflecting certain (unknown) regularities in the realm of things in themselves.¹⁹ It is worth noting that the sketched reading of the refutation of idealism allows us to give a very satisfying answer to a question about this argument that is haunting the literature, namely, the question of whether it is really only meant to establish the existence of fully mind-dependent entities that are presented as being ‘outside us’ or whether it is aimed at the more ambitious conclusion that there are mind-independent entities, that is, entities that are ‘outside us’ in the transcendental sense. More briefly, the question is whether Kant intends to show that there are empirically real but transcendentally ideal entities or whether he intends to show that there are transcendentally real entities. On my reading, Kant can have it all. The refutation, in effect, accomplishes both of the indicated tasks, even though its primarily intended, explicitly advertised conclusion is the former, more modest result. Since empirical objects are grounded in things in themselves―on account of the fact that experience is tightly based on perceptions, which are, in part, due to affections of sensibility by things in themselves―by proving the existence of empirical objects Kant has, implicitly, also demonstrated the existence of things in themselves, that is, of things that are ‘outside us’ in the transcendental sense.²⁰ ¹⁸ Also see Kant’s sketch of the argument in the B-preface, Bxxxix–Bxli note. ¹⁹ There is more to say about Kant’s various refutations of idealism, but, fortunately, there is no need to pursue the tricky issues surrounding them any further in the present context. ²⁰ That the existence of empirical objects entails the existence of things in themselves is the reason, I take it, why at some places where his official topic is the refutation of problematic idealism Kant sounds as if he is talking

192

 ’        

It is also important to appreciate that Kant’s arguments in support of ER1-i (the claim that there exist empirical objects) just discussed do not amount to—nor are meant to be—a general refutation of external-world skepticism from an independent point of view. None of these arguments will sway anybody who rejects the thesis of the transcendental ideality of empirical objects (TI1) or does not buy into Kant’s conception of empirical objects as intentional objects of outer experience. Indeed, I think it is fair to say that external-world skepticism ordinarily is understood as challenging our ability to know about the existence of empirical objects conceived of as things in themselves, that is, conceived of according to transcendental realism. Kant’s arguments, by contrast, respond to a challenge that is directed at our ability to know about the existence of empirical objects conceived of as appearances, that is, conceived of according to transcendental idealism. But the core thesis of transcendental idealism that empirical objects are appearances and not things in themselves is just another way of saying that empirical objects understood as things in themselves do not exist. So, especially people who want to refute skepticism as ordinarily understood will not welcome Kant as an ally. To them, endorsing TI1 basically amounts to giving the game away. Telling these people that Kant’s refutation implicitly proves the existence of things in themselves might make them feel a little bit better but not much. For their question was not if some I-do-notknow-what exists ‘outside us’ in the transcendental sense but if tables and chairs exist ‘outside us’ in the transcendental sense. And Kant’s answer in response to the latter question remains a resounding no. Having emphasized that Kant’s refutation of problematic idealism only works within the framework of transcendental idealism, I should add, though, that Kant also thinks that trying to answer the skeptic within in the framework of transcendental realism is a hopeless undertaking. We already noted in section 3.4.4 above that, in Kant’s assessment, there is a natural path that leads from being a transcendental realist about space and time to being an empirical idealist—at least if space and time are understood as ontologically prior to all objects in them, which is how they ought to be understood on Kant’s view. This is the path taken by Berkeley, according to Kant’s analysis. But Kant also sees a natural path to empirical idealism that starts from being a transcendental realist about empirical objects. For, as the skeptic rightly says, it is impossible to demonstrate the existence of empirical objects ‘outside us’ if they are conceived of as things in themselves, that is, as things that are ‘outside us’ in the transcendental sense. The transcendental realist thus represents outer appearances (if one grants their reality) as things in themselves, which exist independently from us and sensibility, and thus would also be outside us according to pure concepts of the understanding. It is really this transcendental realist who later plays the role of the empirical idealist and who, after falsely presupposing of the objects of the senses that in order to be outer objects they must have their existence in themselves also without the senses, finds all of our representations of the senses to be insufficient in this respect to make their reality certain. (A369)

about a demonstration of the existence of things in themselves, for example, in some of the reflections cited in note 14. We will examine the indicated Kantian argument for the existence of things in themselves in more detail in section 5.10.2.

 

193

So, Kant’s general response to external empirical world skepticism may be summarized by saying that anybody who wants to demonstrate the existence of external empirical objects must first pledge allegiance to transcendental idealism á la Kant. The objects whose existence can be secured in this way are ‘outside us’ only in the empirical sense, but this is all that we can ever hope for anyway. That transcendental idealism enables us to counter skepticism about the external empirical world and demonstrate the existence of empirical objects can, thus, be recorded as another, supplementary argument in its favor.²¹ With an eye toward things to come, it is also worth highlighting that it is not the project of proving the existence of things in themselves in general, that is, the existence of entities that are mind-independent, that Kant takes to be hopeless but the more specific transcendental realist project of proving the existence of empirical objects understood as mindindependent things. Kant himself notes this shortly after the passage just quoted, when he says that one can, indeed, admit that something that may be outside us in the transcendental understanding is the cause of our outer intuitions, but this is not the object that we understand by the representations of matter and of corporeal things; for those are merely appearances, i.e., mere kinds of representations, that are always only in us and whose reality is grounded in immediate consciousness, just as the consciousness of my own thoughts. (A372)

In other words, it is legitimate to admit and, indeed, possible to show, as we will see in section 5.10.2, that there are things ‘outside us’ in the transcendental sense. What would be a mistake, though, is to identify these things with empirical objects, since empirical objects are only ‘in us’ in the transcendental sense. As a final comment with respect to ER1-i, the thesis that there exist empirical objects, I would like to explicitly point out that neither the foundational structure of Kant’s critical idealism nor any of the arguments for the other theses defining this position would be much affected if instead of providing an argument for ER1-i he treated it as a basic assumption of his philosophy. This is worth saying both because the refutation of idealism is not only one of Kant’s more controversial arguments but also one of the arguments with which he himself appears not to have been fully satisfied—as evidenced by his frequent attempts at revising it—and because ER1-i is among Kant’s less controversial theses, commonsensical as it is.²²

²¹ See A378–379: “If we let outer objects pass for things in themselves it is absolutely impossible to understand how we should get to the cognition of their reality outside us by relying merely on the representation that is in us. . . . Thus skeptical idealism forces us to take the only rescue that remains to us, namely, the ideality of all appearances, which we have shown in the transcendental Aesthetic independently of these consequences, which we could not foresee back then.” Also see Prol, 4:337. ²² If we assume ER1-i, the claim that we have outer experience as described in Kant’s theory can be justified by appeal to TI1-strong-b, the thesis that empirical objects essentially are intentional objects of outer experience as characterized in Kant’s theory.

194

 ’        

4.2 The Argument for Transcendental Idealism in the Transcendental Aesthetic 4.2.1 Context and Set-up As reported in section 4.1, Kant takes his main proof for transcendental idealism to be contained in the Transcendental Aesthetic. In section 4.1.1, we already examined the second step of this proof, which consists in the derivation of the theses of the transcendental ideality of empirical objects and empirical selves (TI1 and TI3) from the thesis of the transcendental ideality of space and time (TI2), and, more specifically, from the thesis that pure space and pure time are fully mind-dependent (TI2-pure-b), a derivation that is supplemented by two additional premises, namely, S- and S-. The first step of the proof is an argument for TI2, and, more specifically, TI2-pure-b. In his discussion in the Transcendental Aesthetic, Kant pays no more than scant attention to spelling out the second step of the proof for transcendental idealism.²³ His primary focus lies on demonstrating TI2-pure-b, as well as TI2-purestrong-b, the thesis that pure space and pure time are (nothing but) forms of sensibility. Our project for the remainder of this chapter is to reconstruct these demonstrations, which will complete our account of Kant’s case for transcendental idealism and empirical realism. There is much debate in the literature about how to understand Kant’s argument for transcendental idealism in the Transcendental Aesthetic. The so-called argument from geometry, which commentators locate in the Transcendental Exposition of the Concept of Space (B40 41, also see A24) as well as in the General Remarks to the Transcendental Aesthetic (B64 67/A46 49),²⁴ has received an especially voluminous amount of attention, in particular in Anglo-American circles. (Following Kant, by ‘geometry’ I will mean ‘Euclidean geometry’ unless otherwise noted.) Many commentators believe that this is the argument on which Kant’s transcendental idealist claims about space and the objects in space ultimately rest, at least as far as the Transcendental Aesthetic is concerned.²⁵ Similarly, on this kind of interpretation, Kant’s transcendental idealist claims about time and the objects in time are understood to stand or fall with an analogous argument with respect to the general theory of motion, an argument that is hinted at in the third argument in the Metaphysical Exposition of the Concept of Time (B47/A31) and spelled out further in the Transcendental Exposition of the Concept of Time (B48–49). (Most commentators tend to restrict their discussion to the argument from geometry, though.) There are many different proposals for how to spell out the argument from geometry in detail, but most of them converge on reconstructing it as a ‘bottom-up,’ regressive or analytic inference from ²³ The clearest articulation of the second step of the proof for transcendental idealism in the Transcendental Aesthetic is Kant’s previously quoted summary in the General Remarks with respect to the Transcendental Aesthetic in B59/A42: “Thus we wanted to say: that all our intuition is nothing but the representation of appearance . . . and that if we take away the subject or even merely the subjective quality of the senses in general, all quality, all relations of objects in space and time, indeed even space and time themselves would disappear and can as appearances not exist in themselves but only in us.” ²⁴ Also see sections 6 to 12, and the first remark, in the Prolegomena in the chapter “How is pure mathematics possible?”; Prol, 4:280–287. ²⁵ A prominent proponent of this kind of interpretation is Strawson; see Strawson 1966, 277: “We saw that, as far as the Transcendental Aesthetic is concerned, the doctrine of the transcendental subjectivity of space rests on no other discernible support than that provided by the argument from geometry.”

       

195

our knowledge of Euclidean geometry, conceived of as a body of a priori and, hence, necessary, synthetic or substantive propositions that are true of empirical objects, to the necessary (and sufficient) conditions for the possibility of such knowledge. The analogous argument for the case of time, accordingly, is to be understood as a regressive or analytic inference from our knowledge of certain a priori and, hence, necessary, synthetic or substantive propositions about motion that are true of empirical objects to the necessary (and sufficient) conditions for the possibility of such knowledge. Knowledge of the indicated necessary synthetic truths about empirical objects, so the reconstruction goes, is possible—or at least can be explained without having recourse to a pre-established harmony—(if and) only if our representations of space and time are a priori intuitions, the objects in space and time are not things in themselves but appearances, and space and time are nothing but forms of sensibility; and, of course, opinions diverge about how exactly these conclusions are to be understood.²⁶ (In the following, when I talk about the argument from geometry and the argument from the general theory of motion, I should be understood as having the arguments just sketched in mind.) On this kind of reading, Kant’s argument for transcendental idealism in the Transcendental Aesthetic turns out to be a version of the argument based on the general question of how a priori synthetic judgments are possible, as outlined in the B-preface. In contrast to the reading that Kant’s case for transcendental idealism in the Transcendental Aesthetic hangs on the argument from geometry and an analogous argument from the general theory of motion, other commentators insist that considerations concerning the possibility of geometry or the general theory of motion are not essential to the argument for transcendental idealism in the Transcendental Aesthetic. Rather, the weight of the argument squarely rests on the fact that we have a priori intuitions of space and time, which Kant establishes in the Metaphysical Exposition of the Concept of Space (B37–40/A22–25) and the Metaphysical Exposition of the Concept of Time (B46–48/A30–32), respectively, without appealing to geometry or the general theory of motion in any way. On this kind of reading, the Transcendental Expositions merely provide additional support for the crucial results of the Metaphysical Expositions that we have a priori intuitions of space and time, which serve as the main premises in Kant’s argument for transcendental idealism.²⁷ My reading falls in between these two kinds of interpretations but is closer to the second. I agree with the commentators in the second group that the argument from geometry and the argument from the general theory of motion do not constitute the main proof for transcendental idealism in the Transcendental Aesthetic to which Kant refers at B518 519/A490 491 and B534/A506, quoted at the beginning of section 4.1.1. In this context, it is useful to draw attention to Kant’s comment in the Prolegomena that in the Critique he proceeded according to the synthetic method, while in the Prolegomena he

²⁶ Reconstructions of the argument from geometry along these and similar lines can be found, for example, in Kemp Smith 1918, 111–12, 115–16; Strawson 1966, 277–8; Guyer 1987, 354–69; Van Cleve 1999, 34–43. Note that not all commentators mentioned here are committed to the view that the argument from geometry is Kant’s main or only argument for the thesis that space is transcendentally ideal and nothing but a form of sensibility. It is possible to hold both that Kant endorses the argument from geometry as reconstructed in the main text and that he is in possession of other, possibly more fundamental arguments for the ideality thesis. ²⁷ A prominent proponent of this kind of reading is Allison, for example, see Allison 2004, esp. 117–18. Also see Horstmann 1976; Willaschek 1997; Allais 2010 and Allais 2015, ch. 8.

196

 ’        

proceeds according to the analytic method.²⁸ When proceeding according to the analytic or regressive method, one reasons ‘bottom up,’ as it were. One assumes certain conditioned facts (broadly conceived) as given and moves from them to their principles, or the necessary conditions for their possibility. When proceeding according to the synthetic or progressive method, by contrast, one reasons ‘top down’ from principles to certain consequences.²⁹ Granted, this cannot be taken to mean that the Critique does not contain any analytic or regressive arguments at all. There are fairly clear examples of regressive lines of reasoning at various prominent places throughout the book.³⁰ Nevertheless, I regard Kant’s comments in the Prolegomena about the synthetic procedure in the Critique versus the analytic procedure in the Prolegomena as counting against the reading that the argument from geometry and the argument from the general theory of motion amount to Kant’s main proof for transcendental idealism in the Transcendental Aesthetic. For in these comments, he explicitly pinpoints these very arguments as the regressive line of reasoning in the Prolegomena that qualifies this work as proceeding according to the analytic method.³¹ Furthermore, I also agree with the commentators in the second group that, in his main argument in the Transcendental Aesthetic, Kant establishes the transcendental ideality of space and time and, with it, the transcendental ideality of empirical objects and of empirical selves, by relying on the premise that we have a priori intuitions of space and time and without appealing to geometry or the general theory of motion in any way. So, if we understand by ‘transcendental idealism’ the conjunction of the theses of the transcendental ideality of empirical objects, of empirical selves, and of space and time, Kant’s main argument for transcendental idealism in the Transcendental Aesthetic indeed turns out to be independent of any considerations involving our a priori synthetic cognitions in geometry and the general theory of motion.

²⁸ See Prol, 4:274–275: “In the Critique of Pure Reason, I went to work synthetically with respect to this question [whether metaphysics is at all possible], namely, in such a way that I researched in pure reason itself and tried to determine according to principles in this source itself the elements as well as the laws of its use. This work is difficult and requires a determined reader, to think himself into a system that does not presuppose anything as given except reason itself, and thus tries to develop cognition from its original seeds without relying on any kind of factum. Prolegomena, on the other hand, are supposed to be preliminary exercises. . . . They thus must rely on something that one already knows as reliable, from where one can start with faith and ascend to the sources that one does not know yet, and whose discovery will not only explain what one knew but at the same time present a sphere of many cognitions that all arise from these sources. The methodical procedure of the Prolegomena, in particular, of those that are supposed to prepare us for a future metaphysics, is thus going to be analytic.” Also see Prol, 4:263, 276, 279. ²⁹ See Logic, 9:149; Prol, 4:276, note. ³⁰ The inference to the transcendental unity of apperception as a condition for the possibility of (empirical) self-consciousness in section 16 of the B-deduction is case in point; see B131–135. With reference to Kant’s remark in the Prolegomena that the Critique “does not presuppose anything as given except reason itself ” (Prol, 4:274), one could argue that the employment of regressive arguments that rely on ‘facts of reason,’ or facts of transcendental psychology, is compatible with the synthetic method of the Critique. The indicated argument based on the fact of our empirical self-consciousness could then be seen as such a ‘sanctioned’ analytic argument. ³¹ See Prol, 4:275: “Fortunately, it so happens that, although we cannot assume that metaphysics is actual as a science, we can say with confidence that certain pure synthetic cognitions a priori are actual and given, namely, pure mathematics and pure natural science. . . . We thus have at least some undisputed synthetic cognition a priori and must not ask whether it is possible (for it is actual) but only how it is possible, in order to be able to derive from the principle of the possibility of the given cognitions all others as well.” That the analytic format of the argument from geometry seems to be in tension with the synthetic method of the Critique has not gone unnoticed in literature; for example, see Kemp Smith 1918, 112. An interesting example of an alternative, synthetic reconstruction of Kant’s geometry-based argument in the Transcendental Exposition in the Critique is suggested in Shabel 2004.

       

197

On the other hand, I also agree with the commentators in the first group that, if confronted with the argument from geometry and the argument from the general theory of motion, Kant would give his blessing to them. More specifically, I concur that when he revisits the nature of geometric cognition in the General Remarks (B59 72/A41 49), the final section of the Transcendental Aesthetic, his observations are most naturally read as expressing the kind of regressive argument based on the fact of our having a priori synthetic cognitions in geometry of the kind sketched above.³² That is, even though the argument from geometry and the argument from the general theory of motion do not represent Kant’s main argument for transcendental idealism in the Transcendental Aesthetic, I concur that they are among his arguments for it and can be found, as supplementary arguments, in this section of the book. Moreover and more importantly, I also think that considerations concerning the possibility of our a priori synthetic cognitions in geometry and the general theory of motion, as presented in the Transcendental Expositions, do play a significant, if not essential, role in Kant’s main overall argument of the Transcendental Aesthetic. This main overall argument is supposed to establish not only the theses of the transcendental ideality of empirical objects, of empirical selves, and of space and time but also the thesis that pure space and pure time are (nothing but) forms of sensibility (TI2-pure-b), a thesis that is considerably stronger than the claim that pure space and pure time are fully minddependent and, thus, transcendentally ideal. Given that being fully mind-dependent is only one of several conditions for being a form of sensibility, as discussed in section 3.4.2, asserting that space and time are transcendentally ideal is compatible with denying that they are such forms. Berkeley’s view, for example, includes the former claim—albeit in a formulation that is free of the indicated Kantian terminology—but not the latter. So, once we have established that pure space and pure time and, with them, empirical objects, empirical selves, and empirical space and time are transcendentally ideal, there still remains quite a bit of work to do in order to show that pure space and pure time are (nothing but) forms of sensibility. Specifically, what remains to be done is to show that pure space and pure time are not only fully mind-dependent but also satisfy the other necessary conditions for being a form of sensibility, namely, that they are partly responsible for and necessarily govern the spatial/temporal determinations of all objects of intuition and, hence, of all objects of experience, that is, that they are not only fully mind-dependent but also (what I call) forms of intuition in the broad sense. And, on my reading, the demonstration within his main overall argument in the Transcendental Aesthetic that Kant offers for the claim that pure space and pure time are forms of intuition in the broad sense centrally includes considerations concerning the possibility of geometry and the general theory of motion in the justification for one of its main premises. At the same time, it is important to appreciate that these considerations are not essential to this demonstration. As we will see, the premise in question—the claim that the a priori

³² See B65/A48: “If there were not a faculty in you to intuit a priori; if this subjective condition, with respect to its form, were not at the same time the general condition a priori under which alone the object of this (outer) intuition is itself possible; if the object (the triangle) were something in itself without a relation to your subject: how could you say that what lies necessarily in your subjective conditions to construct a triangle also must necessarily pertain to the triangle in itself? . . . ” Similarly, there are several reflections in which Kant comes close to expressing a version of the argument from geometry or the general theory of motion; see R6346, 18:671; R6355, 18:680 681.

198

 ’        

intuitions of space and time represent general schemes of order that govern (or, more strongly, necessarily govern) the spatial and temporal determinations of all empirical objects—can also be justified in another way that does not involve any appeal to these mathematical sciences. Furthermore, as we will see as well, the thesis that pure space and pure time are forms of intuition in the broad sense can also be demonstrated in an entirely different way that does not depend on considerations concerning the possibility of geometry and the general theory of motion either. This alternative demonstration, which is also hinted at in the text of the Transcendental Aesthetic, is primarily based on considerations concerning the possibility of us having a priori intuitions of space and time. Nevertheless, it remains a fact that the demonstration of the thesis that pure space and pure time are forms of intuition in the broad sense to which Kant gives pride of place in the Transcendental Aesthetic is an argument whose main premise can be justified by thinking about the conditions for the possibility of geometry and the general theory of motion, and that the justification for this premise that Kant foregrounds in his discussion is indeed the one that involves these kinds of considerations.³³ So, if we take transcendental idealism to comprise not only the theses of the transcendental ideality of empirical objects, of empirical selves, and of space and time but also the thesis that pure space and pure time are (nothing but) forms of sensibility (TI2-pure-strong-b), Kant’s main argument for transcendental idealism in the Transcendental Aesthetic, at least in the form highlighted by Kant himself, turns out to rely on considerations concerning the possibility of geometry and the general theory of motion. And even if we take TI2-pure-strong-b to go beyond what Kant means by transcendental idealism, strictly speaking, it still is a core thesis of his critical idealism, that is, of his ontological position overall, and represents a pillar of his theory of experience and the human mind. So, even if Kant’s main argument for transcendental idealism in the Transcendental Aesthetic, strictly speaking, does not have anything to do with geometry or the general theory of motion, an important part of his main overall argument for critical idealism and for his theory of experience that is offered there does have quite a bit to do with these sciences, at least in the form that, for better or worse, he chose to put in the spotlight. Having said that the main argument for TI2-pure-strong-b in the Transcendental Aesthetic foregrounded by Kant involves considerations concerning the conditions under which a priori synthetic cognitions in geometry and the general theory of motion are possible, it is worth highlighting up front that the main line of reasoning in this argument concerns the conditions under which a certain kind of a priori intuition is possible, just like the argument for the thesis of the transcendental ideality of space and time. This is obscured, or at least not sufficiently made clear, by the reconstructions of the arguments from geometry and the general theory of motion sketched earlier, which focus on the conditions for the possibility of a certain kind of a priori knowledge. That is, even though, on my reading, considerations concerning the possibility of geometry and of the ³³ Also note that, arguably, this justification is to be classified as an analytic argument or, at least, involves a regressive line of reasoning, in contrast to the indicated alternative justification that does not rely on considerations concerning the possibility of geometry and the general theory of motion. So, if providing a demonstration with a premise whose justification consists in an analytic argument or involves an analytic line of reasoning, were incompatible with the synthetic method (which is far from clear), then we would have to conclude that (1) in the Transcendental Aesthetic, Kant did not quite manage to live up to his aspirations of proceeding purely synthetically, and (2) instead of highlighting the analytic-style justification for the relevant premise, he should have highlighted the indicated alternative synthetic-style one.

       

199

general theory of motion play a role in Kant’s main overall argument in the Transcendental Aesthetic, the argument from geometry and from the general theory of motion, as reconstructed above, are not the proper way to capture these considerations and their function in this argument. Another point worth emphasizing before we take a closer look at Kant’s main overall argument in the Transcendental Aesthetic is that the intuitive nature of our original representations of space and time is as crucial for this argument as their a priori nature. No analogous argument can be constructed with respect to the a priori concepts of the understanding, a.k.a. the categories. The result that the schematized categories, that is, the versions of the categories that include their own sensible application conditions, apply only to appearances and not to things in themselves—that they are conceptual forms of appearances but not of things in themselves, as one might say—follows from the fact that they apply only to sensible objects, and that all sensible objects, qua being objects in space and/or time, are appearances and not things in themselves. But whether the unschematized categories apply to things in themselves is a further question that cannot be decided by merely meditating on their apriority. Appropriately primed by these preliminary considerations, we are now ready for a closer inspection of Kant’s arguments for the theses that pure space and pure time are transcendentally ideal (TI2-pure-b), and that they are (nothing but) forms of sensibility (TI2-purestrong-b). Kant presents the two claims that constitute TI2-pure-strong—that (a) pure space and pure time are neither things in themselves nor constituted by determinations of things in themselves, and (b) pure space and pure time are (nothing but) forms of sensibility—together with brief summary statements of his reasons for endorsing them as conclusions (a) and (b) under the heading ‘Conclusions from the above Concepts’ for the case of space (B42/A26) and as conclusion (a) under the heading ‘Conclusion from these Concepts’ for the case of time (B49/A32–33).³⁴ Given its central role in Kant’s case for his critical idealism, I will refer to the argument for TI2-pure-strong that is condensed in these paragraphs as ‘the master argument.’ The master argument is Kant’s main overall argument in the Transcendental Aesthetic. Kant’s indication in the section headings that the conclusions in question follow “from the above concepts” makes it plausible to assume that he takes the premises of the master argument to be justified by the results from the foregoing sections of the Aesthetic. These sections comprise an introductory section in which Kant explicates the nature and function of intuition in general, including the difference between pure/a priori and empirical intuition (§1, B33–36/A19–22), and two sections that deal specifically with our representation of space, the previously mentioned Metaphysical Exposition of the Concept of Space (§2) and the Transcendental Exposition of the Concept of Space (§3), the latter of which is directly followed by the Conclusions from the Above Concepts, and two sections that deal specifically with our representation of time, the previously mentioned Metaphysical Exposition of the Concept of Time (§4) and the Transcendental Exposition of the Concept of Time (§5), the latter of which is directly followed by the Conclusions from these Concepts (§6).³⁵ Although there are some non³⁴ Also see Prol, §6–11, 4:280–284; FM, 20:267–268. ³⁵ Kant gives the Conclusions from these Concepts, in which he spells out the implications of the previous results for the nature of time, a special section number (§6), but he does not give a special section number to the Conclusions from the Above Concepts, in which he spells out the implications of the previous results for the nature of space. This accounts for the fact that the Metaphysical Exposition of the Concept of Time bears the

200

 ’        

negligible differences in Kant’s presentation of the master argument for the case of space and for the case of time, I believe that he would agree that the version of the argument that he presents for the case of space, mutatis mutandis, also works for the case of time, and vice versa.³⁶ In order to streamline our discussion, I will thus focus on the case of space and the version of the master argument that Kant presents in the Conclusions from the Above Concepts, with the understanding that an analogous argument can be put together for the case of time. The master argument breaks down into two parts. Part (A) is captured in paragraph (a) of the Conclusions from the Above Concept and establishes TI2-pure-strong-a restricted to space (which is the same as TI2-pure-a restricted to space), that is, the thesis that pure space is neither a thing in itself nor constituted by determinations of things in themselves. In addition, the premises of part (A) are also strong enough to support an argument for TI2-pure-b restricted to space, the thesis that pure space is transcendentally ideal. Part (B) of the master argument is captured in paragraph (b) of the Conclusions from the Above Concepts and establishes TI2-pure-strong-b restricted to space, the thesis that pure space is (nothing but) a form of sensibility.³⁷ For ease of communication, from now on I will refer to this section simply as ‘Conclusions,’ to the Metaphysical Exposition of the Concept of Space as ‘Metaphysical Exposition,’ and to the Transcendental Exposition of the Concept of Space as ‘Transcendental Exposition.’ Also, I will no longer add the explicit qualification that the space with whose transcendental ideality and status as a form of sensibility we are concerned is pure space, but this is how I should be understood unless indicated otherwise. We will examine the two parts of the master argument in turn, starting with part (A). As before, our primary focus will be on trying to figure out how Kant intends his argument to work, as opposed to evaluating its soundness from an independent point of view. At the

section number 4, even though it is the fifth section in the Transcendental Aesthetic. Another potential source of confusion is that Kant’s organization of this material in the B-edition differs notably from his organization of it in the A-edition. There are no separate sections for the transcendental expositions in the A-edition. The third space argument in A contains part of the considerations that make up the transcendental exposition of the concept of space in B and is, accordingly, not included among the space arguments in the metaphysical exposition of the concept of space in B. Similar remarks apply to the case of time, except that Kant decided to retain the third time argument in the metaphysical exposition of the concept of time in B. ³⁶ The master argument for the case of space is presented in two parts under (a) and (b) in the Conclusions from the Above Concepts, while the master argument for the case of time is presented under (a) in the Conclusions from these Concepts. On my reading, the second part of the version of the master argument that Kant presents for the case of time is closely analogous to the second part of the version of the master argument that he presents for the case of space, to be discussed in section 4.2.3. But the first part of the version of the master argument that Kant presents for the case of time differs significantly from the first part of the version of the master argument that he presents for the case of space, and maps fairly closely onto (part of) the argument that, on my reading, was one of the avenues that led him to transcendental idealism about space and time in the first place, as explicated in section 3.4.4. Since substantivalism about time is plainly absurd, and since time is ontologically prior to the things in time, time cannot be a self-subsisting thing or be constituted by determinations of things in time, in particular, not if these things are things in themselves. See B49/A32–33: “(a) Time is not something that subsists for itself or adheres to things as objective determination and thus remained if one abstracted from all subjective conditions of their intuition. For in the first case, it would be something that is real without a real object. But as far as the second is concerned, as a determination or order that inheres in things themselves it could not precede the objects as their condition and be a priori cognized through synthetic propositions and intuited. But the latter indeed happens very well if time is nothing but the subjective condition under which all intuitions can take place in us. For in this case this form of inner intuition can be represented prior to the objects and, hence, a priori.” ³⁷ As we will see, one of the main premises of part (A) serves as the main premise in the first step of part (B), which establishes TI2-pure-b restricted to space, that is, the thesis that pure space is transcendental ideal. That is, Kant also could have divided the master argument into a part that demonstrates that pure space is transcendentally ideal and a part that demonstrates that it is a form of intuition in the broad sense.

       

201

same time, my hope is that by carefully reviewing Kant’s justification for his premises, I will convince you that the argument is philosophically interesting and merits our attention.

4.2.2 Part (A) of the Master Argument: Space is Neither a Thing in Itself nor Constituted by the Determinations of Things in Themselves 4.2.2.1 Reconstruction of the argument Here is the relevant passage from the Conclusions in which Kant summarizes part (A) of the master argument: (a) Space does not at all represent a property of some things in themselves, or them in their relation to each other, i.e., [it represents] no determination of them that adhered to objects themselves and that remained even if one were to abstract from all subjective conditions of intuition. For neither absolute nor relative determinations can be intuited prior to the presence [Dasein] of the things to which they pertain, and thus they cannot be intuited a priori. (B42/A26)

This argument depends on two main premises, one of which is stated explicitly and the other one of which remains implicit. The implicit premise is not hard to guess; it is that we have an a priori intuition of space. The explicit premise is that no determination of a thing can be intuited a priori, on the grounds that no determination can be intuited prior to the presence of the thing to which it pertains. It is tempting to read Kant here as meaning that no property of a thing in itself, or no mind-independent property, can be intuited a priori—and many commentators who take Kant’s argument for the transcendental ideality of space to depend on the nature of intuition do read him in this way.³⁸ I concur that the features of objects of which it is claimed here that they cannot be intuited a priori must be restricted in some way. For there are properties that we can intuit a priori, including, being spherical or square or having some other particular shape that conforms to the Euclidean postulates. Nevertheless, I think the limitation to mind-independent properties, or properties of things in themselves, does not capture what Kant has in mind. On my reading, when Kant says that no determinations of a thing can be intuited a priori he is talking about particular property instances and particular relation instances of genuine existents, where the latter, as you will recall, are entities that exist, not merely in some intentional ‘world,’ but from the point of view of fundamental ontology. We may be able to a priori intuit the general property of being spherical but we cannot a priori intuit, say, the sphericalness of my balancing ball, or that my balancing ball is spherical. In my reconstruction of the argument, to be presented below, I will thus render the second premise in the indicated way, as talking about the determinations of genuine existents. Accordingly, my reconstruction contains an intermediary conclusion, which asserts that space is not constituted by the determinations of genuine existents, that is stronger than the conclusion that we are after, which asserts that space is not constituted by the determinations of things

³⁸ See Willaschek 1997, esp. 554–6; Allison 2004, 122–5; Setiya 2004; Allais 2010, esp. 56–66; Allais 2015, esp. 194–201.

202

 ’        

in themselves. Readers who are reluctant to go along with my rendering of the second premise are invited to substitute for it a more restrictive, weaker version, according to which no determination of a thing in itself can be intuited a priori, and to directly derive the desired conclusion from it, without the added step of the indicated intermediary conclusion. With respect to Kant’s formulation of the conclusion of the argument, that is, that space is not constituted by determinations of things in themselves, it is to be observed that this is not all that we want yet. The stated result rules out the Spinozist view that space is a determination of God and the Leibnizian view that space is constituted by the spatial determinations of things in themselves, but it leaves the Newtonian substantivalist view unscathed, according to which space is a self-subsisting thing in itself that functions as a quasi-container for other things in themselves. There is no question, however, that, by the Conclusions section, Kant takes himself to be justified in rejecting the Newtonian view as well, despite failing to mention it explicitly in his presentation under (a). Among other things, this is made clear by the fact that, in his presentation of the corresponding conclusion in the case of time, he does include an explicit denial of the Newtonian view by adding the claim that “time is not something that subsists for itself” (B49/A32). Similarly, in §7 of the Transcendental Aesthetic, which follows right on the heels of the Conclusions of these Concepts, Kant characterizes his opponents as “those who claim the absolute reality of space and time, whether they assume them to be subsisting or merely inhering” (B56/A39), where the ‘subsisting’ option clearly is substantivalism. And since his previous arguments are supposed to have defeated all of his opponents, it is safe to assume that this also applies to the proponents of substantivalism. Having settled that Kant’s conclusions include the rejection of the view that space and time are subsisting things in themselves, the question arises whether he takes this rejection to be justified by the argument sketched under (a) that we are presently considering or on separate grounds. As I see it, the answer to this question is ‘both.’ As already briefly mentioned earlier, Kant seems to regard the view that space and time are self-subsisting things in themselves as simply absurd, and thus as not in particular need of a special refutation that goes beyond drawing attention to its absurdity.³⁹ At the same time, the argument presently under consideration quite clearly also refutes the Newtonian view. Since it is impossible to intuit an entity without intuiting at least some of its determinations—indeed, intuiting an entity seems to consist in intuiting at least some of its determinations—the premise that no determination of a genuine existent can be intuited a priori entails that no genuine existent can be intuited a priori, which, in turn, allows us to draw the desired stronger conclusion that space is neither constituted by the determinations of things in themselves nor a thing in itself. Since the rejection of substantivalism on grounds of absurdity is bound to have only limited appeal, I will formulate my reconstruction of part (A) of the master argument in such a way as to make clear that it is strong enough to support the stronger conclusion just stated.

³⁹ As indicated in note 36, this is also what we find by way of explanation of the rejection of Newtonian substantivalism under (a) in the conclusions section on time. See B49/A32: “ . . . for in the former case [if one assumes that time subsists for itself] it would be something that, without any real object, is still real.” Kant seems to regard this consequence as obviously absurd. Also recall B56/A39 and B70–71, quoted in note 100, chapter 3.

       

203

Master argument, part (A): 1. We have an a priori intuition of space. (First premise, in need of support.) 2. No genuine existent and no determination of a genuine existent can be intuited a priori. (Second premise, in need of support.)⁴⁰ 3. Space is not a genuine existent nor is it constituted by determinations of genuine existents. (Intermediary conclusion, from 1 and 2.)⁴¹ 4. If space were a thing in itself or constituted by determinations of things in themselves, it would be a genuine existent or constituted by determinations of genuine existents, respectively. (Third premise, uncontroversial.) 5. Space is not a thing in itself nor is it constituted by determinations of things in themselves. (TI2-pure-a=TI2-pure-strong-a, restricted to space; from 3 and 4.) With respect to the intermediary conclusion in line 3, it is important to keep in mind that the space that we are talking about here is (what later on in the overall argument turns out to be) pure space, not empirical space. On my reading, as you will recall, empirical space is constituted by determinations of genuine existents, namely, empirical objects, and, thus, is itself a genuine existent. The argument tells us that, by contrast, pure space, that is, the space that is represented by the a priori intuition of space, is not so constituted and does not exist from the point of view of fundamental ontology. Another important feature of the argument that deserves to be highlighted is that its two premises are also sufficient to deduce the conclusion that space is fully mind-dependent (TI2-pure-b, restricted to space). On the plausible assumption, which we already implicitly adopted all the way back in section 2.2, that any entity, that is, anything that is real to some degree, is either a genuine existent or a fully mind-dependent pseudo-existent, and the further plausible assumption that space is not nothing, the intermediary conclusion that space is not a genuine existent directly entails that space is fully mind-dependent. That is, part (A) of the master argument, in effect, is an argument for the transcendental ideality of space. In his presentation in part (a) of the Conclusions, Kant himself does not explicitly draw attention to this fact but he quite clearly agrees with this assessment.⁴² As presented, part (A) of the master argument is valid. Whether it is convincing depends on how well its two premises can be supported. So, let us take a look at that.

4.2.2.2 Premise 1: We have an a priori intuition of space There is consensus in the literature that the main result of the metaphysical exposition of the concept of space is supposed to be that our representation of space—or, as one should ⁴⁰ As just noted, since the claim that no determination of a genuine existent can be intuited a priori entails that no genuine existent can be intuited a priori, the former claim by itself could be used as second premise. I opted to include the latter claim to maximize transparency. ⁴¹ This inference also rests on the assumption that if x is constituted by certain determinations and x can be intuited a priori, then the determinations by which x is constituted can be intuited as well. I take this assumption to be uncontroversial, at least for the kind of constitution relation in question, where A being constituted by X, Y, and Z means as much as that A consists in X, Y, and Z, taken together. ⁴² Indeed, in the relevant sections in the Transcendental Aesthetic Kant seems to equate the claim that space and time are neither things in themselves nor constituted by determinations of things in themselves with the claim that they are transcendental ideal; see B44/A28, B52/A36. This is conceptually a bit sloppy but excusable from a justificatory point of view, given that the argument by means of which he establishes the former claim also establishes the latter.

204

 ’        

say more precisely, our original representation of space—is an a priori intuition.⁴³ (The meaning of the qualification ‘original’ will be explicated presently.) That is, commentators tend to agree that the main intended outcome of the metaphysical exposition is, or at least includes, the first premise of part (A) of the master argument. But while there is widespread agreement about the intended conclusions of Kant’s arguments in this section of the Transcendental Aesthetic, there is considerable disagreement about how exactly the arguments are to be understood. For reasons of space, we cannot wade very deeply into this controversy here. I will confine myself to a brief sketch of my reading, with a focus on bringing out what the arguments reveal about Kant’s views about our representations of space and his conception of intuition.⁴⁴ The metaphysical exposition of the concept of space proceeds by analyzing what “belongs” to this concept, and, more specifically, what constitutes it as “a priori given” (B38/A23), where ‘concept’ here should be understood in a general sense, I believe, as including any of our ordinary representations of space, be they conceptual or intuitive. The indicated analysis uncovers that all of these representations, as well as all representations of spatial properties and relations, rely on and thus presuppose a more basic representation of space—which, I take it, is not conscious and never explicitly entertained as such—and that this more basic representation is an a priori intuition. The qualification ‘original’ in the conclusion that our original representation of space is an a priori intuition, I submit, is thus to be understood as indicating that this representation is our most fundamental representation of space, which grounds all others, including, in particular, all conceptual representations, strictly speaking, and first defines what we mean by ‘space.’ The a priori nature of the original representation of space is established in the first two space arguments (B38–39/A23–24); its intuitive nature is demonstrated in the third and fourth space argument (in the numbering of the B-edition, B39–40/A24–25).⁴⁵ The most important feature of our ordinary representations of space that plays a vital role in all four arguments is that, according to Kant’s analysis, we necessarily conceive of and present any spatial determination as delineated in, and any determinate space, that is, any determinate figure and any determinate spatial volume, as a ‘cut-out’ part of the same all-encompassing space. For example, when we represent a sphere, we necessarily present it as taking up a certain spherical volume in space, and when we represent two spheres that are 3 meters apart from each other, we necessarily present their distance as a separation in space that one could measure by, for example, placing a measuring rod between them. For ease of reference, I will call this aspect of our ordinary representations of space their ‘fundamental feature.’⁴⁶ The fundamental feature of our ordinary representations of space allows us to ⁴³ See B40: “Therefore, the original representation of space is an intuition a priori and not a concept.” Also see B40–41; for the case of time, see B48/A32. ⁴⁴ Helpful discussions of the metaphysical expositions can be found in Vaihinger, 1892, 151–263; Allison 2004, 99–116; Falkenstein 1995, 159–252. An excellent analysis of the first space argument—and convincing critique of Allison’s (1983) earlier reading—is given in Warren 1998. ⁴⁵ The corresponding results for the case of time are established in the Metaphysical Exposition of the Concept of Time. The a priori nature of the original representation of time is shown in the first three arguments (B46–47/ A30–31), and its intuitive nature is shown in the fourth and fifth arguments (B47–48/A31–32). ⁴⁶ While it seems difficult to deny that we necessarily present every particular spatial determination as delineated in, and every determinate finite space as a ‘cut-out’ part of some larger space, one might be skeptical of the stronger claim that we necessarily present every particular spatial determination as delineated in, and every determinate finite space as a ‘cut-out’ part of the same space. The first, second, and fourth space argument depend only on the weaker claim; but the third argument, as we will see, depends on the stronger one.

       

205

infer that these representations rely on and implicitly presuppose a more basic representation of space that, as further analysis reveals, is a priori and exhibits telltale features that mark it as an intuition. Note that the indicated fundamental feature is indeed special to our ordinary representations of space (and time). We do not represent colors or color relations of objects in this way, for example. When I represent a tomato, I do not present its red color as delineated in an all-encompassing color space. Neither the representations of particular colors nor the representations of particular color relations implicitly presuppose an underlying more basic representation of color space in which these colors and color relations are presented. But I do present the shape of the tomato as delineated in space, whose representation is thus implicitly presupposed in my tomato representation. More specifically, the first space argument turns on the following claim: [I]n order for certain sensations to be related to something outside of me (i.e., to something that is in another place of space than the one in which I am), and similarly in order for me to be able to represent them as outside and next to one another, and thus not merely as different but as being in different places, the representation of space must be already presupposed. (B38/A23)

This claim is a direct consequence of the fundamental feature of our ordinary representations of space just described. The representation of any particular spatial relation, be it a spatial relation between me and an object that empirically causes certain sensations in me or between different objects that cause different sensations in me, presupposes a more basic representation of space in which this spatial relation is presented. But this means that empiricism about the original representation of space cannot be right. For according to the empiricist account, we first acquire this representation through the perception of particular spatial relations and other spatial determinations, and the performance of various basic cognitive operations on these perceptions such as comparing and abstracting. But, as we just learned, any representation of particular spatial relations or other spatial determinations already presupposes a representation of space, which means that we could not possibly have acquired the original representation in the way proposed by the empiricist.⁴⁷ So, since the original representation of space cannot be empirical, it must be a priori. The second space argument relies on a previous result from the second section of the Introduction to the Critique, a result that could be expressed by the slogan that ‘necessity is a sign of apriority.’ Experience can only ever teach us how things are but not how they have to be. The slogan applies to both propositions and features of objects. That is, on Kant’s view, the necessity of a proposition indicates that our cognition of it (if we have it) must be a priori in the sense of consisting in a priori representations and being justified on a priori grounds. And the necessity of a feature of an object indicates that our cognition of it (if we have it) must be a priori in the sense of relying on a priori representations. For instance, using Kant’s examples, that the theorems of geometry are necessary indicates that our cognition of them must be a priori; and that objects necessarily are either substances or accidents indicates that the concept of substance and the concept of accidents that inhere ⁴⁷ See B38/A23: “Consequently, the representation of space cannot have been borrowed from the relations of outer appearances through experience, but this outer experience is itself first possible through the imagined representation.”

206

 ’        

in a substance must be a priori.⁴⁸ The second space argument applies this result to the feature of bodies of necessarily being in space, which, Kant assumes, we can cognize. Given that ‘necessity is a sign of apriority,’ we can conclude from this feature that the representation of space is a priori.⁴⁹ This reading of the second space argument is supported by the fact that Kant’s second example in the Introduction that he offers to illustrate his necessity criterion for apriority is precisely the feature of bodies to necessarily be in space. Not only in judgments but also in concepts an a priori origin of some of them shows itself. If you, by and by, leave out from your empirical concept of a body everything that is empirical in it: the color, the hardness or softness, the heaviness, even the impenetrability, there still remains the space that it (which is now completely gone) occupied. . . . Thus, you must admit, persuaded by the necessity through which this concept pushes itself on you, that it has its seat a priori in your faculty of cognition. (B6)

That we cognize bodies as necessarily being in space is also, in part, a consequence of the fundamental feature of our ordinary representations of space. It is analytic and, hence, necessary that bodies are extended, which means that bodies necessarily occupy a determinate volume of space. And, due to the fundamental feature of our ordinary representations of space, we necessarily represent any determinate volume of space by presenting it as a ‘cut-out’ part of, and thus as contained in space. This means that, while we can represent space without bodies in it, we cannot represent bodies without presenting them as being in space, which, in turn, serves as the basis for our cognition that bodies necessarily are in space.⁵⁰ The third and fourth space arguments are meant to establish that the original representation of space is an intuition, as opposed to a concept. In both arguments, Kant proceeds by showing that the original representation of space lacks various features that are characteristic of concepts but satisfies features that, as we gather from the arguments, are characteristic of intuitions. The basic difference between intuitions and concepts on which these arguments depend is that intuitions are singular representations, while concepts are essentially general, to use terminology that Kant himself explicitly introduces only later on.⁵¹

⁴⁸ See B2/A2; B4 6. ⁴⁹ See B38 39/A24: “Space is a necessary representation, a priori, that grounds all outer intuitions. One can never represent that there is no space although one can very well think that no objects are encountered in it. It thus is regarded as the condition of the possibility of appearances and not a determination that depends on them and is a representation a priori that necessarily grounds outer appearances.” ⁵⁰ The claim that the representation of space is logically prior to and a condition for the possibility of the representation of bodies is also explicitly featured in Kant’s own presentation of the second space argument (see the previous note). Many commentators take Kant to deduce the apriority of the original representation of space directly from one or both of these claims. See, for example, Allison 2004, 104 8. On my reading, the function of the relevant claims in the argument is to establish that we cognize bodies as necessarily being in space. And it is the latter claim from which Kant then infers the apriority of the original representation of space, based on his necessity criterion for apriority, as just explicated in the main text. ⁵¹ See B136, note: “Space and time and all of their parts are intuitions, and thus singular representation with the manifold that they contain in themselves (see the transcendental Aesthetic), and thus not mere concepts through which the very same consciousness is encountered as contained in many representations but many representations as in one and in its consciousness . . . ” Also recall B376–377/A320: “ . . . an objective perception is cognition (cognitio). This is either intuition or concept (intuitus vel conceptus). The former relates immediately to the object and is singular; the latter relates mediately to the object, by means of a mark that can be common to several things.” Also see Log, 9:9: “An intuition is a singular representation . . . , a concept is a general or reflected representation.” Also see FM, 20:325; Anth, 7:196.

       

207

In the literature, the singularity of intuitions is commonly cashed out as meaning no more than that they refer to exactly one particular thing (if they refer), in contrast to concepts, which refer to classes of things (if they refer). I concur that this characterization is a correct description of how the singularity–generality difference typically manifests itself with respect to the reference of intuitions and concepts. But, in my assessment, it falls short of capturing the difference that Kant is after. For one thing, there are some concepts that also refer to exactly one particular thing, for example, the concept ‘the oldest head of state in office right now.’ More importantly, the proposed characterization focuses on a difference between intuitions and concepts that is a consequence of their singularity and generality, respectively, but fails to properly describe what singularity/generality consists in to begin with. On my reading, by calling intuitions ‘singular,’ Kant means to highlight a special feature of their representational capabilities, namely, that they reveal a special kind of presentational content to the mind of the cognizer who entertains them. Intuitions directly represent individuals, that is, they have individuals as their intentional objects— they represent individuals in their individuality, as one might say—which also explains why they refer to exactly one particular thing (if they refer).⁵² By contrast, the representational capabilities of concepts are limited in such a way that they are incapable of directly representing individuals. Concepts typically directly represent properties, which also explains why they typically refer to classes of things (if they refer). And even in those rare cases where they directly represent or refer to particular objects, concepts represent only certain aspects of these objects but not them in their individuality. For example, the concept ‘the oldest head of state in office right now,’ despite directly representing and referring to one particular object, reveals only a limited number of determinations of its object, namely, that it is a person, a head of state, quite old, and in office right now. According to Kant, this difference in the representational capabilities of intuitions and concepts is grounded in that (a) individuals are what he calls ‘completely determinate,’ (b) in order to directly represent an individual, a representation must be analogously completely determinate, and (c) intuitions are the only completely determinate representations that we are capable of. “Since only singular things or individuals are completely determinate, there can be completely determinate cognitions only in form of intuitions but not in form of concepts; in the case of the latter, the logical determination can never be regarded as completed.” (Log, 9:99)⁵³ “But a lowest concept . . . or a lowest species, under which no other would be contained anymore, does not exist in the series of species and genera, since such a concept cannot possibly be determined. For even if we have a concept that we apply immediately to individuals, there still can be specific differences with respect to it, which we either do not notice or do not take into consideration.” (Log, 9:97)⁵⁴ ⁵² Recall that I use ‘R directly represents O’ as short for ‘O is R’s intentional object’ and that I reserve the expression ‘R refers to O’ to characterize intentional relations of representations to objects that genuinely exist. ⁵³ See V-Met-L2/ Pölitz, 28:560: “An individual or ens singulare is that [sic.] insofar it is completely determinate.” ⁵⁴ See V-Lo/Pölitz, 24:569: “ . . . the lowest concept cannot be determined. But I could say ‘lowest cognition,’ because cognitions comprise both concepts and intuitions. A lowest cognition thus would be immediate intuition because it contains only a single thing.” See B683/A655: “ . . . reason demands in its entire expansion that no

208

 ’        

Kant does not provide a crisp, explicit definition of what he means by describing individuals and representations of individuals as completely determinate. A comprehensive discussion of this question—which turns out to be quite complicated—would lead us to far afield. But, without wading into any messy details, on my reading, the basic idea can be captured by saying that an entity E is completely determinate if, and only if, (1) for every possible determinable, it is determinate whether E has it or not, and (2) for every determinable of E, it is determinate which corresponding determination E has.⁵⁵ And a representation R of type T is completely determinate if, and only if, (i) R represents a completely determinate entity E, (ii) R captures every determination of E that can be represented in representations of type T, and (iii) R is sufficient to individuate E in the sense of distinguishing it from all other entities.⁵⁶ That intuitions are the only completely determinate representations that we are capable of is a direct consequence of the nature of completely determinate representations and the nature of our cognitive faculties, sensibility and the understanding. On the assumption that “individuality involves infinity,” to use a memorable formulation from Leibniz,⁵⁷ that is, on the assumption that all individuals have infinitely many determinations and that, in order to represent an object in its individuality, infinitely determinations must be specified—which is an assumption that Kant quite clearly endorses—it follows that every completely determinate representation is infinitely complex in the sense of comprising infinitely many representations in it.⁵⁸ As described in section 2.7.2, due to the discursive nature of our intellect, we comprehend concepts by apprehending the marks that compose them in

species is regarded as the lowest in itself because, since it is still always a concept, which only contains that which is common to several things, it cannot be completely determined and, thus, cannot right away be related to an individual, and, therefore, must always contain different concepts, i.e., sub-species, under itself.” See V-Th/ Baumbach, 28:1244: “All our cognitions of things are representations through concepts that I determine more and more until they are completely determinate, and then I have the concept of an individual.” ⁵⁵ For a refresher on how to understand the notion of a determinable, revisit note 158, chapter 2. Note that, due to their status as intentional objects of experience (which is a conceptual representation), appearances are not completely determinate. For example, the number of parts of any appearance is necessarily indeterminate, as we saw in section 2.7.2 in our discussion of the mathematical antinomies. Appearances are still individuated particular objects; they are just not individuals, strictly speaking. Also see note 130, chapter 3. ⁵⁶ See B602/A573: “The proposition: everything that exists is completely determinate does not mean only that of every pair of opposite predicates one belongs to it but also that of all possible predicates always one belongs to it; through this proposition not only predicates are compared logically with one another, but the thing itself is transcendentally compared with the sum total [Inbegriff] of all possible predicates. It wants to say as much as that in order to completely cognize a thing, one must cognize everything possible and determine it through it, be it affirmatively or negatively.” See Log, 9:99: “The highest completed determination would give a completely determinate concept . . . , that is, a concept in addition to which one could not think any further determination.” See V-Met/Volckmann, 28:410: “ . . . for different objects are entirely undetermined by the concepts that we have of them, for example, the concept of a human being is entirely undetermined, it can be learned or not, it can be man or woman etc., I say one of the two must pertain to it, and if I posit one, it is determined; something is called undetermined only with respect to that which we think of it through our concept, e.g., the concept of Δ is undetermined, for it can be obtuse or acute etc. So, determinable is what we call a general concept and an object that is not yet completely determined . . . .” A detailed discussion of Kant’s views about complete determination, both with respect to objects and with respect to representations, is provided in Jauernig (in preparation). ⁵⁷ See Leibniz 1704, A VI.6:289–290: “The most important thing in this is that individuality involves infinity, and only someone who is capable of grasping the infinite could have knowledge of the principle of individuation of a given thing.” ⁵⁸ See V-Lo/Wiener, 24:931: “But this complete determination of a thing is impossible because an infinite cognition would be required to visit all of the predicates that pertain to a thing, and I can thus proceed to infinity while still not completely determining the thing.” When Kant says in this passage that a complete determination of a thing is impossible, he should be understood to mean that a complete conceptual determination of a thing is impossible, for reasons that will become clear presently. The important point for now is that the passage supports the reading that complete determination requires infinitely many determinations.

       

209

piecemeal fashion, one by one, in a process that takes time. Consequently, all of our concepts are, of necessity, only finitely complex and thus never completely determinate. This explains why our concepts are incapable of directly representing individuals. By contrast, intuitions are completely determinate, according to Kant, and, thus, directly represent individuals, which is reflected in their infinite complexity. In this context, the Leibnizian strand in Kant’s account of intuition, as discussed in section 2.4, is relevant again. Even though, due to the limitation of our mind, we can individually apprehend only finitely many parts of any intuition with consciousness, all intuitions are still infinitely complex in that their unconscious basic components include infinitely many ‘little’ intuitions on which their emergent components supervene through con-fusion. Thanks to their infinitely complex basic components, intuitions are capable of directly representing individuals in their individuality, albeit only unconsciously.⁵⁹ With this clarification of Kant’s conception of the singularity of intuitions in hand, let us return to the remaining space arguments in the Metaphysical Exposition. In the third argument, Kant shows that the underlying representation of space that is implicitly presupposed in all our ordinary representations of space is an intuition by establishing that it is singular in the sense just explicated, that is, that it is a direct representation of an individual, a completely determinate representation of exactly one particular thing. He does this by appealing to the fundamental feature of our ordinary representations of space, that we necessarily present all spatial determinations as delineated in, and all determinate finite spaces as ‘cut-out’ parts of the same one, all-encompassing space. . . . one can represent only a unified space, and if one talks about multiple spaces one understands by that only parts of one and the same all-unified space. These parts also cannot precede the unified all-encompassing space as its constituents, as it were (through which a composition was possible), but can only be thought in it. It is essentially unified, the manifold in it, and thus the general concept of spaces in general, rests merely on limitations. (B39/A25)

That we conceive of any spatial determination as delineated in, and any determinate space as a ‘cut-out’ part of the same one, all-encompassing space that is represented by our original representation of space directly allows us to conclude that this representation is intentionally related to exactly one particular thing.⁶⁰ But, I take it, we can also conclude that this representation is completely determinate. This latter part of the argument remains

⁵⁹ That our conscious intuitions are only finitely complex is also what Kant is getting at in his response to Beck’s proposal to define intuitions as completely determinate representations, see Letter to Beck, July 3, 1792, 11:347: “What you say of your definition of intuition, that it is a completely determinate representation with respect to a given manifold, I do not have anything to add against that except that the complete determination must be understood objectively and not as taking place in the subject (since we cannot possibly know all determinations of an object of empirical intuition), in which case the definition would say no more than that it is a representation of a singular thing that is given.” On my reading, by ‘subject’ Kant here means as much as the conscious subject. For further discussion of Kant’s account of the difference between intuitions and concepts along the indicated lines, see Jauernig 2019. ⁶⁰ One might question the main premise of this argument, as already indicated in note 46. Do we indeed necessarily present every particular spatial determination as delineated in, and every determinate finite space as a ‘cut-out’ part of the same space, or do we merely necessarily represent each one of them in some larger space, where it remains undetermined whether this larger space is the same for all of these spatial relations and determinate spaces?

210

 ’        

largely implicit in Kant’s own presentation, but the main idea can be spelled out as follows. That we conceive of spatial determinations and determinate spaces in the indicated way can be taken to establish that the underlying original representation of space, on which the representations of these spatial determinations and spaces are based, at least implicitly represents all possible spatial determinations comprised in, and all possible determinate spaces that are parts of the space represented by it. And, so the argument goes, a representation of a space that represents all possible spatial determinations comprised in it and all possible determinate spaces that are parts of it is completely determinate. Informally put, once you have told me which possible spatial determinations and which possible determinate spaces are contained in space S, you have told me all there is to know about S; you have completely determined S. So, our original representation of space is a completely determinate representation that intentionally relates to exactly one particular thing, that is, it is a singular representation. Given what we just learned about the nature of concepts and intuitions, this result directly implies that our original representation of space is not a concept. And on the assumption that intuitions are the only singular representations that we are capable of (which Kant appears to endorse), it also directly follows that this representation is an intuition. The fourth space argument provides further support for the result that the original representation of space is an intuition by establishing that it is infinitely complex. Kant appears to infer its infinite complexity from the fact that we present space as infinitely divisible. It is not directly clear how this inference is supposed to work, and Kant does not offer much help, except to say that “space is thought in this way,” that is, as containing an infinite multitude of representations in itself, “for the parts of space to infinity are simultaneous” (B40). I suspect that Kant is reasoning along the following lines. According to our ordinary representations of space, for any determinate space, there are multiple smaller determinate spaces contained in it. That is what it means to think of space as infinitely divisible. Given the fundamental feature of our ordinary representations of space, each determinate space is necessarily presented as contained in the same one allencompassing space as one of its parts. Since all parts of space are simultaneous, the underlying, original representation of space must thus comprise representations of all these determinate spaces and hence be infinitely complex.⁶¹ Given what we just learned about the nature of concepts and intuitions, in particular, the inevitably finite complexity of all our concepts, this result, in turn, directly implies that the original representation of space is not a concept, and, on the assumption that intuitions are the only infinitely complex representations that we are capable of (which Kant appears to endorse), it also directly follows that this representation is an intuition.⁶² This completes our review of the ⁶¹ Also see Kä, 20:419: “In the former [metaphysics] it [space] is original and only one (unified) space, in the latter [geometry] it is derived, and there are (many) spaces, of which, however, the geometer, in conformity with the metaphysician, must admit, on account of the foundational representation of space, that they can be thought only as parts of the unified original space. Now one can call a magnitude in comparison with which every specifiable similar magnitude is equal only to a part of it no other than infinite. Accordingly, the geometer, just as the metaphysician, represents original space as infinite and, more specifically, as infinitely given.” ⁶² See B39 40: “Space is represented as an infinite given magnitude. Now, one must conceive of every concept as a representation that is contained in an infinite multitude of different possible representations (as their common mark), and thus contains these under itself; but no concept, as such, can be conceived in such a way as if it contained an infinite multitude of representations in itself. Nevertheless, space is thought in this way (for all parts of space to infinity are simultaneous). Therefore, the original representation of space is an intuition a priori and not a concept.”

       

211

arguments in the Metaphysical Exposition for the claim that our original representation of space is an a priori intuition, a claim that directly implies the first premise of part (A) of the master argument, according to which we have an a priori intuition of space. Commentators disagree about the overall purpose of the Transcendental Exposition, and, in particular, about the function of the argument from geometry that is presented there. They also disagree about the structure of, and the exact content of the different steps in Kant’s reasoning. But most commentators agree that, whatever else Kant takes his argument in this section to show, one of the conclusions that he draws is that the representation of space on which our geometric cognitions are based is an a priori intuition. That is, there is consensus in the literature that the first premise of part (A) of the master argument receives additional support from Kant’s considerations concerning the possibility of geometry as an a priori synthetic science in the Transcendental Exposition (provided these considerations are defensible). A transcendental exposition is “an explanation of a concept, as a principle, from which the possibility of other synthetic cognitions a priori can be understood,” where it is required “1) that such cognitions really flow from the given concept, [and] 2) that these cognitions are possible only by presupposing a [sic.; should be ‘the’] given way of explaining this concept” (B40). I take this to mean that, in a transcendental exposition of a concept, the possibility of certain synthetic a priori cognitions is made intelligible by showing that we actually have these cognitions if, and only if, a certain understanding of this concept, or rather of the object of this concept, is assumed. If we indeed have cognitions of the relevant kind, our transcendental exposition allows us then to conclude that the assumed understanding of the concept is correct. The a priori synthetic cognitions whose possibility is supposed to be made intelligible in the transcendental exposition of the concept of space are our a priori cognitions of the theorems of geometry, and the understanding in question is Kant’s understanding of space as a form of sensibility. For now, we are only interested in the first step of the desired explanation, which is covered in the second paragraph of the Transcendental Exposition (“Geometry is a science . . . ”). This first step consists in identifying what the representation of space on which our geometric cognitions are based must be so that these cognitions are (i) cognitions of synthetic truths, that is, truths that are not logical truths nor truths that are true in virtue of the meaning of the involved concepts alone but substantive truth about an object, namely, pure space, and (ii) cognitions of necessary truths. This step yields the conclusions that the representation of space on which our geometric cognitions are based must be (i) an intuition, on the grounds that it is impossible to demonstrate that the geometric axioms are true of pure space by mere conceptual analysis, and (ii) a priori, on the grounds that the geometric theorems are necessary, and experience can only teach us how things are but not how they have to be, or, more briefly, because ‘necessity is a sign of apriority,’ as explicated above.⁶³ There is quite a bit more to say about part (i) of this argument, but since we do not ⁶³ See B40–41: “Geometry is a science that determines the properties of space synthetically and still a priori. What then must the representation of space be so that such a cognition of it is possible? It must originally be intuition; for from a mere concept no propositions can be drawn that go beyond the concept, which, however, happens in geometry (Introduction V.). But this intuition must be encountered in us a priori, i.e., before all perception of an object, and thus be pure, not empirical intuition. For the geometric propositions are altogether apodictic, i.e., connected with the consciousness of their necessity, e.g., space has only three dimensions; but such propositions cannot be empirical or judgments of experience nor can they be inferred from them (Introduction II.).” Also see B14 (=Introduction V.): “For since one found that the inferences of the mathematicians all proceed

212

 ’        

want to get side-tracked I will merely note that, on my reading, the crucial idea here is that the a priori intuition of space, which comprises intuitions of all possible geometric figures in it, provides a model for Euclidean geometry, thanks to its singularity. We will return to the second step of the desired explanation below in our discussion of part (B) of the master argument. For our present purposes, we only need the result that the possibility of our a priori synthetic cognitions of geometry, understood as substantive a priori cognitions about pure space, can be explained only if the representation of space on which these cognitions are based is an a priori intuition. For this result, together with the assumption that we do have a priori synthetic geometric cognitions, provides additional confirmation for premise one of part (A).

4.2.2.3 Premise 2: No genuine existent and no determination of a genuine existent can be intuited a priori As I see it, the justification for the second premise of part (A) of the master argument rests on a certain necessary condition for us to have an intuition of an object, a condition that is underwritten by the sensible nature of intuitions, which, in turn, is an expression of our finitude and the ontologically uncreative nature of our mind.⁶⁴ The sensible nature of intuitions is one of two other features of intuitions—in addition to their singularity—that are introduced in Kant’s discussion in the Transcendental Aesthetic before his presentation of the master argument. The second other feature is the immediacy of intuitions—that is, that they intentionally relate to objects without mediation by any other representation⁶⁵— which, however, has no direct bearing on the master argument, according to my reading.⁶⁶ according to the principle of contradiction . . . , one convinced oneself that the foundational principles were cognized from the principle of contradiction as well; in which they erred; for a synthetic proposition can indeed be understood from the principle of contradiction but only in such a way that another synthetic proposition is presupposed from which it can be inferred but never in itself.” ⁶⁴ We will take a closer look at the ontologically uncreative nature of our mind, due to our mind’s finitude, and its relation to the passive nature of sensibility and the sensible nature of our intuitions, in section 5.10.2. ⁶⁵ See note 155, chapter 2. ⁶⁶ In this assessment, I am disagreeing with Markus Willaschek und Lucy Allais who, like me, also understand Kant’s argument for the transcendental ideality of space and time to rest squarely on his conception of the nature of intuition but who, unlike me, regard the immediacy of the relation of intuitions to objects as the crucial feature of intuitions on which the argument depends. Willaschek suggests that Kant is working with an externalist conception of how intuitions acquire their intentional relation to objects: intuitions are not only caused by affections of sensibility through objects but their reference is determined in this way as well. Intuitions are about and represent the objects to whose affections they are due, and their immediate relation to objects consists in that they are causally and, hence, directly related to these objects. Assuming this understanding of intuitions, it follows that the object of an a priori intuition cannot be a real, subject-independent object, as Willaschek puts it, for otherwise the intuition would not be a priori. See Willaschek 1997, 547–8, and 554–6. My reading of Kant’s support for premise 2 of part (A) of the master argument is similar to Willaschek’s, mutatis mutandis (he understands the premise slightly differently), but, in my assessment, Kant does not subscribe to an externalist, causal account of the reference of intuitions, as will be explicated presently. Also see note 72. Allais argues that the immediacy of intuitions is to be understood according to a relational view of the referential relation of intuitions to their objects, according to which intuitions involve the presence to consciousness of the particular things they represent, where “if an object is present to consciousness, it is not merely something that causally affects a subject’s mental states. Rather, the object is in the subject’s consciousness” (Allais 2015, 159). Since, according to Allais’s reading, mind-independent things can be present to consciousness only through affecting us, it follows that a priori intuitions do not present us with mind-independent features of reality. See Allais 2010, 59–65; Allais 2015, 153–63; 194–201. As can be gleaned from my brief account of Kant’s conception of the singularity of intuitions above, I am happy to concur that, on Kant’s view, intuitions involve the presence of the particular things they represent—or, as the case may be, of the intuitive counterparts of these things (understood as explicated in section 2.5.4)—to the mind of the cognizer entertaining them. But I find it difficult to make sense of the claim that intuitions involve the presence to consciousness of the particular things they represent in the context of Allais’s two-aspect interpretation. This difficulty is closely related to the problem, discussed in note 12, chapter 2, that, in

       

213

In whatever way and through whatever means a cognition may be related to objects, the one through which it relates to them immediately and toward which all thinking aims as a means is intuition. But this takes place only insofar as the object is given to us; but this, in turn, is possible, for us human beings at least, only through it affecting the mind in a certain way. The capacity (receptivity) to receive representations through the manner in which we are affected by objects is called sensibility. Thus, by means of sensibility, objects are given to us and it alone furnishes us intuitions; through the understanding, however, they are thought, and from it arise concepts. But all thinking eventually must be related to intuitions, and thus in our case to sensibility, be it directly (directe), or indirectly (indirecte), by means of certain marks, since no object can be given to us in another way. (B33/A19)

Kant’s precise understanding of the sensible nature of intuitions emerges only throughout the later discussion, in the Transcendental Aesthetic and beyond. For example, it only emerges later on what kind of affections by what kind of things our intuitions depend and what sort of relation between intuitions and these things is thereby established. At this early juncture in the dialectic, the quoted general description is all that Kant offers his readers. This deliberate vagueness is to be expected. Part of the point of the argument of the Transcendental Aesthetic is to first introduce the transcendental distinction between things in themselves and appearances and to clarify their different relations to intuitions. Moreover, Kant naturally wants his argument to have the broadest possible appeal, which

order to make room for her anti-phenomenalist reading of Kant’s conception of appearances, Allais claims that, on Kant’s view, representations can have mind-independent or extra-mental constituents or elements, which, however, is highly implausible. (On the last point, also see note 9, chapter 2, and the discussion of the implications of Kant’s critique of the Leibniz-Wolffian view that we confusedly perceive mind-independent things for an Allais-style two-aspect reading in section 5.3.) On Allais’s reading, there are particular things that have mindindependent and mind-dependent properties. These particular things as bearers of mind-independent properties are things in themselves; these particular things as bearers of mind-dependent properties are appearances. Accordingly, by ascribing the view to Kant that intuitions involve the presence to consciousness of the particular things they represent, she is at the same time ascribing the view to him that, when we have empirical intuitions, things with mind-independent properties, or things with mind-independent aspects, are in, or present to, our consciousness. To my mind, this view is complete anathema to Kant. Just as representations are ‘in us’ and fully mind-dependent, anything that is in, or present to, our consciousness is ‘in us’ and fully mind-dependent. In an important passage in the Prolegomena, to be discussed in the main text below, Kant explains why we cannot have intuitions of things in themselves, not even empirical ones, by pointing out that, even if we assume that things in themselves affect us, “it is unintelligible how the intuition of a present thing could reveal it to me how it is in itself because its properties cannot migrate into my capacity of representing” (Prol, 4:282, my emphasis). I believe that Kant would give a similar response to Allais’s reading as well. No particular thing with mind-independent properties, or a mind-independent aspect, can be in, or present to, our consciousness because these mindindependent properties or aspects cannot simply migrate into it. At best, only the mind-dependent properties or aspects of the particular things in question can be in, or present to, our consciousness; but, of course, in that case it can no longer be said that particular things are in, or present to, our consciousness. We can get at the same kind of worry from a slightly different direction by reflecting on how Kant’s claim that empirical intuitions represent, and can represent, only appearances but not things in themselves is to be cashed out on an Allais-style two-aspect view. This claim must be read as saying that empirical intuitions represent only how things appear to us but not how they are in themselves. This means that Allais is committed to the reading that empirical intuitions present particular things to consciousness even though they represent only one aspect of them, namely, their appearance aspect. But this is like saying that Nelly, the elephant, is present to my consciousness, even though I only see Nelly’s belly. In order for a particular thing to be in, or present to, my consciousness, all of it must be present to my consciousness; otherwise, what is present to my consciousness is not the particular thing but merely an aspect of the thing. Accordingly, it seems to me that Allais ought to give up either her two-aspect interpretation or her relational reading of the reference and immediacy of intuitions. On my reading, particular empirical objects can be in, and present to, consciousness because I take them to be fully mind-dependent and, thus, capable of ‘migrating,’ in their entirety, into my capacity of representing, to use Kant’s flowery language.

214

 ’        

makes it desirable not to build any potentially controversial, specific assumptions into the terms that figure in it. A straightforward, general characterization of what it means for intuitions to be sensible, based on Kant’s remarks in section 1 of the Transcendental Aesthetic, is that they are produced by sensibility, an essentially passive cognitive faculty, and thus depend on affections of our mind.⁶⁷ Note that this dependence on affection extends to every intuition-token and not just to the first generation of the corresponding intuition-type. As Kant puts it in the passage just quoted, intuitions take place only insofar as the object is given to us and, thus only insofar as we are affected.⁶⁸ Given that intuitions are sensible in the sense just described, and thus depend on affections of our mind by things that are present, we can start our examination of Kant’s support of premise 2 with the general thesis that if a subject has an intuition of an object or of a determination of an object, then the subject’s mind is affected by something present.⁶⁹ Moreover, given Kant’s visual resemblance theory of the intentional relation of intuitions to objects, as explicated in section 2.5.4, according to which intuitions intentionally relate to objects in virtue of the resemblance, in terms of looks, between the presentational content of the intuition and the intuitive determinations of the object, it is plausible to precisify this thesis to say that if a subject has an intuition of a genuine existent or of a determination of a genuine existent, then that existent is present and materially affects the subject’s mind. (The qualification ‘materially’ will be explicated in a minute.) This precisification is plausible because without assuming some kind of causal connection or, more generally, some kind of determination relation that links the existent and the intuition, it would appear to be a mere

⁶⁷ See B72: “ . . . our outer as well as inner manner of intuiting, which is called sensible because it is not original, i.e., one through which the existence itself of the object of the intuition is given (and that, as far as we can see, can only pertain to the ur-being) but which depends on the existence of the object and thus is possible only through the affection of the capacity of representation of the subject by it.” See B75/A51: “If we want to call the receptivity of our mind to receive representations insofar it is somehow affected sensibility, the capacity to produce representations on its own, or the spontaneity of cognition, by contrast, is the understanding. Our nature brings with it that intuition can never be other than sensible, i.e., contain only the manner how we are affected by objects.” Also see B93/A68; B129. For the passive character of sensibility, also see B33–34/A19–20: “The capacity (receptivity) to receive representations through the manner in which we are affected by objects is called sensibility . . . ” Also see Anth, 7:140–141. ⁶⁸ This is what distinguishes the sensible character of intuitions from the sensible character of empirical concepts. Empirical concepts are sensible in that the process by means of which we first acquire them involves affections of our mind by external objects. But, once we have acquired them, these concepts can be entertained without any external affection taking place. For example, once I have formed the empirical concept ‘dog’ by comparing various dog perceptions, abstracting from what is different between them, and reflecting on what is common among them, I can entertain the concept ‘dog’ again at a later time without being affected by anything apart from my own mind. By contrast, even though I might be able to remember certain aspects of a given empirical intuition of a dog without being affected by anything apart from my own mind, I cannot have a qualitatively identical empirical intuition at a later time without also undergoing the same kind of external affection again as when I had the original dog-intuition in the first place. ⁶⁹ Another commentator who highlights the dependence of intuitions on affections as the feature that is of crucial relevance for Kant’s argument for the transcendental ideality of space is Kieran Setiya, see Setiya 2004, esp. 71–4. But Setiya understands Kant’s question of how a priori intuitions are possible as a challenge to come up with an account of how we could stand in cognitive relations to a particular thing without being affected, a challenge that he answers by suggesting that space is a sense-datum. “Suppose that the sense-datum of space stands for space itself by way of an isomorphism of identity. Then our cognitive relation to the sense-datum of space—a relation that does not depend on affection—will be a relation to space itself. In other words, this will be a view on which we represent space in intuition a priori” (Setiya 2004, 75). But a representation that does not depend on affection is not an intuition, on Kant’s account; intuitions are essentially sensible. That is, Setiya’s answer to Kant’s question of how a priori intuitions are possible amounts to an explanation of how representations are possible that, like intuitions, put us in cognitive relations to particular things but, unlike intuitions, do not depend on affections. But that is not what Kant is asking about. I will return to the question of how we can have representations that are both a priori and depend on affections in section 4.2.4.

       

215

coincidence, and thus quite mysterious how it could be, that the presentational content of the intuition—or, more precisely, the presentational content of the ‘little’ intuitions on which the intuition supervenes—resembles the intuitive determinations of the existent. Moreover, since our mind is finite and, thus, ontologically uncreative, we do not bring objects into existence by intuiting them. So, given that we are talking about intuitions of genuine existents and of determinations of genuine existents, the direction of the causal connection or determination relation that links the existent and the intuition must be existent-to-intuition, not intuition-to-existent.⁷⁰ To be sure, there is a potential alternative explanation of the visual resemblance between the presentational content of the intuition and its referent available, namely, an explanation in form of a pre-established harmony that God maintains out of the goodness of his heart to ensure that human beings can successfully cognitively relate to the world. But, for Kant, appealing to such a pre-established harmony not only yields a less satisfactory explanation than the indicated causal account, it amounts to intellectually selling out. So, taking our lead from Kant, we will simply ignore the possibility of a pre-established harmony in what follows. The qualification in the stated thesis that the relevant kind of affection is ‘material’ has been added to take account of the fact that a priori intuitions also require affections, simply because, like all intuitions, they are essentially sensible. I will call this other kind of affection that is involved in the generation of pure intuitions ‘formal affection.’ As we will discuss in section 4.2.4, formal affections are transcendental affections of sensibility through the mind’s own activity. Material affections, be they empirical or transcendental, are affections of the mind by means of which sensations are generated. The proposed necessary condition for the reference of intuitions to genuine existents is fairly weak; it can be expected to be palatable to a wide and diverse audience, and it is also part of Kant’s considered critical idealist view, as fully developed in later parts of the Critique.⁷¹ And it is all that we need to support our second premise. 1. If a subject has an intuition of a genuine existent or of a determination of a genuine existent, then the existent is present and materially affects the subject’s mind. (Thesis supported by the sensible nature of intuitions, the finitude and ontologically uncreative nature of the human mind, and considerations concerning the reference of intuitions, including the rejection of a pre-established harmony.) 2. A priori intuitions are independent from any material affections of our mind by genuine existents. (Uncontroversial premise.)

⁷⁰ This line of reasoning has a long history in Kant’s thought; see Letter to Herz, February 21, 1772, 10:130: “For I asked myself: on what ground does the relation of that which in us one calls representation to the object rest? If the representation contains only the manner in which the subject is affected by the object it is easily intelligible how it is in conformity with the latter as an effect to its cause, and how this determination of our mind could represent something, i.e., have an object. . . . Similarly, if that which in us is called representation were active with respect to the object, i.e., if through it itself the object were created, as one imagines divine cognitions as the archetypes of things, the conformity of it with the objects could also be understood.” Also note that the indicated line of reasoning does not require that the intuited existent is a thing in itself. Given that we do not literally create appearances by intuiting them—which is impossible for us, given the finitude of our mind and its ontologically uncreative nature—the intentional relation between our intuitions and appearances is intelligible only if there is some kind of determining link between them on account of which the presentational content of the intuitions resembles the appearances. ⁷¹ The thesis holds for both appearances and things in themselves, which, according to Kant’s critical idealist ontology, are the only genuine existents in the world.

216

 ’        

3. No genuine existent and no determination of a genuine existent can be intuited a priori. (Second premise of part (A) of the master argument, from 1 and 2.) Note that this argument does not turn on some kind of causal theory of reference, as one might be tempted to think. Granted, if intuitions referred to genuine existents in virtue of being partly caused by them, the desired conclusion would follow as well.⁷² But, on my view, Kant does not subscribe to such a causal theory of reference but to a kind of visual resemblance theory, as just reiterated. The appeal to causal relations between intuitions and their referents enters the account because of the sensible nature of intuitions and in the service of making the resemblance in terms of looks between the referent and the presentational content of the intuition (or, more precisely, of the presentational content of the ‘little’ intuitions on which the intuition supervenes) intelligible, on which the reference of the intuition depends. That the proposed reconstruction of the justification for the second premise of part A of the master argument adequately captures what Kant has in mind is confirmed by his explanation in the Prolegomena of why it is impossible for us to a priori intuit things in themselves. If our intuition had to be such that it represented things as they are in themselves, no intuition at all would happen a priori but it would always be empirical. For what is contained in the object itself, I can only know if it is present and given to me. To be sure, even in this case it is unintelligible how the intuition of a present thing should reveal it to me how it is in itself because its properties cannot migrate into my capacity of representing; but even if we grant this possibility, such an intuition would not take place a priori,

⁷² As indicated before, Willaschek reads part (A) of the master argument in this kind of way. See note 66. My main worry about ascribing to Kant an externalist, causal theory of reference is that this does not comport with his preoccupation with the question of how to understand the cognitive, intentional relations of our representations to objects. A mere causal relation is not a cognitive, intentional relation. Willaschek himself actually acknowledges as much in the final section of his essay, where he admits that we arrive at a representation that has an intentional relation to an object only by applying a concept whose marks express the sensations contained in the intuition. For example, an intuition that contains sensations of redness and roundness is related to a tomato by our application of the concept ‘tomato’ which contains the marks ‘red’ and ‘round.’ Willaschek does not spell out how the reference of intuitions to objects is to be understood once this complication is taken into account, but I suppose that he takes it to comprise both the immediate causal relation and the conceptually mediated relation to objects as integral elements. See Willaschek 1997, 560. Setting aside that one could question whether, after this modification, Willaschek can still legitimately advertise his account of the referential relation of intuitions to objects as a form of externalism, the main problem with the new account is that the indicated relations do not seem to ‘target’ the same object: the conceptually mediated relation connects the intuition to a round, red thing, while the immediate, causal relation connects it to an unknowable, non-spatial thing that affects us (or to several unknowable, non-spatial things that affect us). Willaschek’s response consists in an appeal to the one-object view and the claim that the round, red thing is in fact not distinct from the affecting thing but the same thing considered in a different way. Not surprisingly, then, the ultimate shortcoming of Willaschek’s account of the reference of intuitions is a specific version of one of the main shortcomings of the one-object interpretation. We will discuss this shortcoming in section 5.5, but to anticipate briefly, the main worry is that attributing the claim to Kant that the tomato and the affecting thing are numerically identical amounts to ascribing the belief in a pre-established harmony to him. According to Kant’s account of the operations of outer sense, it could be that the sensations of redness and roundness are due to affections by multiple things in themselves, and the thing in itself that is responsible for the sensation of redness could also be responsible, say, for the sensation of greenness that is part of our intuition of a zucchini that is right next to the tomato. The assumption that none of these possibilities actually obtains is an assumption of a pre-established harmony, that is, a harmony between how we represent the empirical world to be carved up into different things and how the transcendental level is actually carved up.

       

217

i.e., before the object is put in front of me; for without that no ground of the relation of my representation to it can be imagined, unless it were to rest on inspiration. (Prol, 4:282)⁷³

According to Kant’s visual resemblance account of the reference of intuitions, in order for intuitions to refer to things in themselves, their presentational content would have to resemble, in terms of looks, the intuitive determinations of the things in themselves. As Kant comments in the passage, it is unintelligible how this could ever happen, given that the properties of things in themselves “cannot migrate into my capacity of representing.” It is not immediately obvious what he is getting at here, but my best guess is that he is anticipating the later result that, since space and time are (nothing but) forms of sensibility, to which anything that is intuited must conform, all properties of objects that we can intuit, and, thus, that can make it into “our capacity of representing,” must be mind-dependent. Another thought that might also be part of what Kant has in mind here is that it is hard to see how something that is mind-independent could ever resemble the presentational content of an intuition, given that presentational contents are essentially mind-dependent, or, to speak with Berkeley, given that “an idea can be like nothing but an idea.”⁷⁴ At any rate, despite the unintelligibility of how the presentational content of intuitions could possibly resemble the determinations of things in themselves, for the sake of the argument Kant is willing to assume that this is possible after all and to grant that our intuitions can refer to things in themselves. This reference would be utterly mysterious, however, Kant goes on, if things in themselves were not partly causally responsible for the intuitions. If there were no such causal connection, the resemblance between the presentational content of the intuitions and the determinations of the things in themselves would amount to a kind of pre-established harmony and rest on “inspiration.” So, even if we allow, for the sake of the argument, that intuitions of things in themselves are possible, a priori intuitions of things in themselves still turn out to be impossible because they would constitute a violation of the condition that intuitions refer to genuine existents only if those existents materially affect our mind. It is also worth highlighting that, according to Kant’s ultimate account of the sensible nature of intuitions and their relation to objects, even though appearances or empirical objects are the only genuine existents that we can intuit, things in themselves are still required to affect sensibility for us to have intuitions of genuine existents. In his ultimate account, Kant is not only committed to the thesis that if a subject has an intuition of a genuine existent or of a determination of a genuine existent, then that existent is present

⁷³ Also see Prol, 4:284: “If one wanted to doubt in the least that both [space and time] are not determinations that inhere in things in themselves but only determinations that inhere in their relation to sensibility, I would like to know how one can find it possible to know a priori, and thus prior to all acquaintance with the things, namely, before they are given to us, what their intuition must be like, which, however, is the case here with respect to space and time.” ⁷⁴ See Berkeley 1710, Works 2:45: “§8. But say you, though the ideas themselves do not exist without the mind, yet there may be things like them whereof they are copies or resemblances, which things exist without the mind, in an unthinking substance. I answer, an idea can be like nothing but an idea; a colour or figure can be like nothing but another colour or figure. If we look but ever so little into our thoughts, we shall find it impossible for us to conceive a likeness except only between our ideas. Again, I ask whether those supposed originals or external things, of which our ideas are the pictures or representations, be themselves perceivable or no? If they are, then they are ideas, and we have gained our point; but if you say they are not, I appeal to any one whether it be sense, to assert a colour is like something which is invisible; hard or soft, like something which is intangible; and so of the rest.”

218

 ’        

and materially affects the subject’s mind but also to the stronger claim that if a subject has an intuition of an empirical object or of a determination of an empirical object, the empirical object is present and materially empirically affects the subject’s empirical mind, via the subject’s sense organs, and a thing in itself is present and materially transcendentally affects the subject’s sensibility. To complete our discussion of part (A) of the master argument, we need to address one important objection that might be raised to it as presented so far. The feature of the original representation of space, on the strength of which Kant identifies it as an intuition and justifies the first premise, is its singularity. But the feature of intuitions on which the justification for the second premise rests is their sensible nature. This invites the objection that as long as it has not been shown that no representation of ours can be singular without also being sensible the argument does not go through. In other words, the objection pushes on the implicit assumption of the metaphysical and transcendental expositions that we are capable of only one kind of singular representations.⁷⁵ Without going into too much detail, although Kant himself does not provide an explicit argument for the claim that all of our singular representations are necessarily sensible, I believe that such an argument is available to him. As we will discuss in more detail in section 5.10.2, arguably the most fundamental presupposition of Kant’s philosophical system is that the human mind is essentially finite. Given this essential finitude, it is quite striking that we are capable of entertaining infinitely complex representations at all. This capability is nevertheless compatible with our finitude because it is subject to several significant limitations, in contrast to the corresponding capacity of an infinite intellect, which is not limited in any way. We have encountered two of these limitations already, namely, that (1) we cannot individually apprehend all of the infinitely many component parts of our intuitions with consciousness, with the result that all our conscious intuitions are confused to some degree, and (2) we do not create objects by intuiting them. Another, related limitation is that our intellect cannot spontaneously produce any infinitely complex representations on its own, be they distinct or confused. The only way in which an infinitely complex representation can be generated in our mind is passively, by way of an affection of our mind. That is, due to our finitude, all of our infinitely complex representations must be sensible. And since all singular representations necessarily are infinitely complex, it follows that all of our singular representations necessarily are sensible too.

4.2.3 Part (B) of the Master Argument: Space is (Nothing but) a Form of Sensibility 4.2.3.1 Reconstruction of the argument Part (B) of the master argument is contained in passage (b) in the Conclusions, which reads as follows:

⁷⁵ An analogous objection can also be raised to reconstructions of the argument that identify the immediacy of intuitions as the ground for the impossibility of a priori intuitions of things in themselves, as on Willaschek’s and Allais’s readings. As long as it has not been shown that any representation of ours that exhibits the features of the original representation of space on the strength of which the metaphysical and transcendental expositions identify it as an intuition is also immediate in the relevant sense, the argument does not go through.

       

219

“(b) Space is nothing but the form of all appearances of our senses, i.e., the subjective condition of sensibility, under which alone our outer intuition is possible. Now since the receptivity of the subject to be affected by objects necessarily precedes all intuitions of these objects, it is intelligible how the form of all appearances could be given before all real perceptions, and [how it could be given] a priori in the mind, and how it as a pure intuition in which all objects must be determined could contain principles of their relations before all experience. (B42/A26)

As in his presentation of part (A), Kant begins his presentation of part (B) of the master argument by stating its conclusion. The conclusion of part (B) is TI2-pure-strong-b, restricted to space, that is, the thesis that space is a form of sensibility/intuition, or, equivalently but more explicitly, that pure space is nothing but a form of sensibility/ intuition, or, more precisely, that space is nothing but a form of outer intuition.⁷⁶ The reason for this conclusion offered by Kant is that by assuming it we can explain two seemingly mysterious circumstances, namely, first, how we could have an a priori intuition of space, and, second, how this a priori intuition of space “could contain principles of the relations of all objects before all experience.” By the latter expression, I understand Kant to mean how this a priori intuition could encode certain principles, namely, the Euclidean postulates, that govern all determinations of a certain type that all empirical objects have, namely, their spatial determinations, or, equivalently, how this a priori intuition could represent a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects.⁷⁷ As I see it, the relevant model of explanation that Kant has in mind is the kind of explanation that is featured in the transcendental exposition of space, which was just completed in the section immediately preceding the Conclusions. As noted already, a transcendental exposition is “an explanation of a concept, as a principle, from which the possibility of other synthetic cognitions a priori can be understood,” where it is required “1) that such cognitions really flow from the given concept, [and] 2) that these cognitions are possible only by presupposing a [sic.; should be ‘the’] given way of explaining this concept” (B40). Accordingly, I take it that the explanation of the possibility of us having an a priori intuition that represents space as well as a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects, of which Kant claims that it can be given on the assumption that space is a form of outer intuition, includes showing that we have such an a priori intuition if, and only if, space is a form of outer intuition. For our present purposes of deriving the desired conclusion—that space is a form of outer intuition and, thus, a form of sensibility—we primarily need to worry only about the only-if-part of this explanation, which, unfortunately, turns out to be the more challenging part, although in the course of explicating it we will have occasion to examine an important element of the if-part as well.

⁷⁶ As explicated in section 3.4.2, since a form of sensibility, by definition, is an intentional object, the addition that space is nothing but a form of sensibility—as opposed to being both a form of sensibility and a form of things in themselves—is, strictly speaking, redundant. Also, as before, I will continue to use ‘space’ without further qualification as short for ‘pure space’; whenever I have a different space in mind, I will explicitly say so. ⁷⁷ Here and in the following, I will assume it to be uncontroversial that it is essential for empirical objects to have some spatial determinations. It is essential for empirical objects to be extended; and being extended implies having spatial determinations. So, the claim that a certain scheme of order governs all spatial determinations of empirical objects implies that this scheme governs determinations of all empirical objects.

220

 ’        

We can thus start our examination of part (B) of the master argument with the following general reconstruction. Part (B) of the master argument, general version: 1. We have an a priori intuition that represents space as well as a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects. 2. We have an a priori intuition that represents space as well as a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects only if space is a form of outer intuition. 3. Space is a form of outer intuition and, thus, a form of sensibility. (TI2-pure-strong-b, restricted to space; from 1 and 2.) As explicated in section 3.4.2, there are several conditions for being a form of sensibility/ intuition, on my reading. To remind us of the definition established there, S is a form of sensibility/intuition if, and only if, (1) S essentially is an intentional object of a pure intuition of a certain type T, (2) S is responsible for all intuitions of type T of finite objects necessarily presenting these objects as having determinations of a corresponding type T*, and (3) S is a general scheme of order that necessarily governs all determinations of type T*.⁷⁸ (2) and (3) define what it means to be (what I am calling) a form of intuition in the broad sense, while (1), (2), and (3) define what it means to be a form of intuition simpliciter, a.k.a. a form of sensibility.⁷⁹ On my reading, Kant’s demonstration in part (B) of the master argument to the effect that space satisfies these conditions proceeds in two steps. The first step establishes that space satisfies condition (1) by relying on the claim that we have an a priori intuition of space as one of its main premises. The second step establishes that space satisfies not only condition (1) but also conditions (2) and (3) by relying on the claim that the a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects as one of its main premises. For clarity’s sake, it pays to explicitly distinguish these different steps in the argument. Part (B) of the master argument, detailed version: 1. We have an a priori intuition of space. (First premise, in need of support.) 2. We have an a priori intuition of space only if space essentially is an intentional object of an a priori intuition. (Second premise, in need of support.) 3. Space essentially is an intentional object of an a priori intuition. (From 1 and 2.) 4. The a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects. (Third premise, in need of support.)

⁷⁸ Also recall that the reason why I say that the forms of sensibility are responsible for all intuitions of a certain type of finite objects necessarily presenting these objects as having determinations of a certain corresponding type is that it would be circular to hold that the forms of sensibility are responsible for the way in which the pure intuitions of space and time represent them, given that these intuitions ontologically specify them in the first place. (Pure space and time are infinite.) ⁷⁹ For the distinction between a form of intuition in the broad sense and a form of sensibility, see note 67, chapter 3.

       

221

5. The a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects only if space is a form of outer intuition in the broad sense, that is, a general scheme of order, defined by the Euclidean postulates, that is responsible for all outer intuitions of finite objects necessarily presenting these objects as having spatial determinations and necessarily governs all spatial determinations. (Fourth premise, in need of support.) 6. Space is a form of outer intuition in the broad sense. (From 4 and 5.) 7. Space is a form of sensibility. (TI2-pure-strong-b, restricted to space; from 3 and 6, and the definition of a form of sensibility.) The first step of the argument is captured in lines 1 to 3, the second step is captured in lines 4 to 6, and line 7 brings them both together in the desired conclusion. It is worth flagging that the conclusion of the second step can also be derived in a different way, by relying on the claim that we have an a priori intuition of space and a different necessary condition for this to be the case, which can be ferreted out by considering the question of how we can have representations that are both sensible and a priori. We will return to this alternative derivation in section 4.2.4. For now, we will focus on part (B) of the master argument as Kant presents it in the Conclusions, which is the argument just reconstructed. Another point worth flagging is that Kant is committed to and takes himself to be justified in asserting, not only that our a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects, which I have identified as the third premise, but also the modally stronger claim that our a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that necessarily governs all spatial determinations of empirical objects. And, according to Kant’s account, the necessary conformity of all spatial determinations of empirical objects to the Euclidean theorems ultimately is also due to space’s status as a form of intuition. That is, we could also use the claim that our a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that necessarily governs all spatial determinations of empirical objects as our third premise. But since, as I see it, we can get away with relying on the modally weaker claim, and since it is generally a prudent strategy to construct one’s arguments with the weakest and thus least controversial premises that support the desired conclusion, we will stick with the reconstruction just provided that relies on the weaker claim.⁸⁰ I should add, though, that sufficiently little hangs on this choice so that readers who, after reading my account of Kant’s justification for the premises of part (B), remain skeptical about whether the desired conclusion can indeed be shown to express a necessary condition for the truth of the ⁸⁰ Incidentally, this marks another difference of the argument reconstructed here compared to the argument from geometry in its usual reconstruction, as sketched in section 4.2.1, which relies on the alleged necessity with which the Euclidean postulates and theorems apply to all empirical objects. On the other hand, it is worth noting that Kant himself may not even recognize a difference in modal strength between the claim that all spatial determinations of empirical objects are governed by a certain scheme of order, on the one hand, and the claim that all spatial determinations of empirical objects are necessarily governed by a certain scheme of order, on the other hand. For, on Kant’s view, strict universality and necessity invariably go together; see B4: “Necessity and strict universality are thus reliable criteria of a cognition a priori and also belong inseparably together.” That is, Kant seems to hold that if the scheme of order in question governs absolutely all spatial determinations of empirical objects, then this scheme necessarily governs these determinations. (One way to read this would be to say that ‘absolutely all Xs’ is to be understood as ‘all actual and all possible Xs.’)

222

 ’        

modally weaker claim and worry that it rises to the level of a necessary condition only with respect to the modally stronger claim, are welcome to substitute the latter for the former in premises three and four. The justifications for the original versions of these premises can fairly easily be ‘enhanced’ so as to support the thus modified versions as well, which, for the sake of a clean presentation, however, will be left as an exercise for the skeptical reader. Now, as stated, the argument is valid. But is it sound?

4.2.3.2 First step: space essentially is an intentional object of an priori intuition This step, in effect, amounts to a recapitulation of part (A) of the master argument, and thus can be dealt with rather quickly. As noted in section 4.2.2.1, the two premises of part (A) of the master argument—that we have an a priori intuition of space, and that no genuine existent and no determination of a genuine existent can be intuited a priori— together with the assumption that any entity, that is, anything that is real to some degree, is either a genuine existent or a fully mind-dependent pseudo-existent, support not only the conclusion that space is neither a genuine existent nor constituted by determinations of genuine existents but also the result that space is fully mind-dependent and thus essentially an intentional object. And since the a priori intuition of space is our most fundamental representation of space, which grounds all others and first defines what we mean by ‘space,’ as demonstrated in the metaphysical exposition, it is furthermore safe to conclude that this intuition is the representation that ontologically specifies space, that is, that it is the representation in virtue of being represented in which space has being and all of its determinations and other ontological ingredients. So, space essentially is an intentional object of the a priori intuition of space. This line of reasoning can be summarized by saying that we have an a priori intuition of space only if space essentially is an intentional object of an a priori intuition, as claimed in the second premise of part (B). And since we do have an a priori intuition of space, as shown previously and as stated in the first premise of part (B), it follows that space essentially is an intentional object of an a priori intuition, as stated in line 3 of part (B). 4.2.3.3 Second step: space is a form of intuition in the broad sense Before we turn to a detailed consideration of Kant’s justifications for premises three and four on which the second step of part (B) depends, it will be helpful to take a brief look at the context in which these justifications are proposed. On my reading, the second step of part (B) of the master argument is the place at which considerations concerning the possibility of geometry as an a priori synthetic science enter Kant’s deliberations. More specifically, the second step of part (B) largely corresponds to the second step of the transcendental exposition of the concept of space, which is contained in the third paragraph of the Transcendental Exposition. As already noted in section 4.2.2.2, in the first step of the transcendental exposition (contained in the second paragraph of the section), Kant shows that the representation of space on which our geometric cognitions are based must be an a priori intuition, on the grounds that otherwise these cognitions would be neither synthetic nor of necessary truths. As just rehearsed, this result allows us to conclude furthermore that space is transcendentally ideal and essentially an intentional object of an a priori intuition. Kant presents the second step of the transcendental exposition as follows (in the third and final substantive paragraph of the section).

       

223

Now, how can an outer intuition be in the mind that precedes the objects themselves and in which the concept of the latter can be determined a priori? Obviously not in any other way than insofar as it is located merely in the subject, as the formal quality of it to be affected by objects and thereby receive an immediate representation of them, i.e., intuition, that is, only as form of outer sense in general. (B41)

The question with which Kant begins this passage—how can outer intuition be in the mind that precedes the objects themselves and in which the concept of the latter can be determined a priori?—amounts to the same question that, in his presentation of part (B) of the master argument, he expresses by asking how the a priori intuition of space “in which all objects must be determined could contain principles of their relations before all experience,” that is, how the a priori intuition of space could represent a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects. This how-question would not arise if we did not actually have such an intuition. And, in the context of the transcendental exposition, it is clear that the reason on the basis of which Kant regards it as established that we do have such an intuition is that all empirical objects conform to the a priori substantive theorems of Euclidean geometry. In order for this conformity to obtain, the a priori intuition of space, on which our a priori cognitions of the theorems of Euclidean geometry are based, must represent a general scheme of order that governs all spatial determinations of empirical objects. The claim that all empirical objects conform to the a priori substantive theorems of Euclidean geometry can, of course, be called into question, but this is not the place to do so; we will grant it for the purpose of the argument. So, in the second step of the transcendental exposition, Kant completes the promised explanation of the possibility of geometry as an a priori synthetic science by showing that all empirical objects conform to the substantive theorems that we a priori cognize in geometry if, and only if, (1) the a priori intuition of space, on which our geometric theorizing is based, represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects, which, in turn, is possible if, and only if, (2) space is a form of outer intuition in the broad sense— at least if we set aside the possibility of a pre-established harmony.⁸¹ We find the same argumentative structure in the corresponding sections in the Prolegomena (Prol, 4:280–282). In §6, Kant wonders how the a priori synthetic judgments of mathematics are possible, which he explains in §7 by pointing out that the first and highest condition for the possibility of mathematics is that it is grounded in pure intuition, and, more specifically, in the case of geometry, the a priori intuition of space. In §8, he turns to the second step of the argument, which centers on the question of how it is possible that we can cognize something about objects based on the a priori intuition of space. The answer is revealed in §9, where we learn that this is possible only if the intuition represents a form of sensibility.

⁸¹ Also see B196/A157: “Although we thus a priori cognize much in synthetic judgements about space in general or the figures that the productive imagination describes in it, so that we indeed have no need of experience for it; this cognition would still be nothing but play with a mere figment of the brain if space were not to be regarded as the condition of the appearances that constitute the material for outer experience.”

224

 ’        

Thus it is possible only in one unique way that my intuition precedes the actuality of the object and takes place as a cognition a priori, namely, if it contains nothing other than the form of sensibility, which in my subject precedes all actual impressions through which I am affected by objects. (Prol, 4:282)⁸²

Having thus identified the proper context in which the second step of part (B) of the master argument must be read, namely, the transcendental exposition of space to whose second step it largely corresponds, and having outlined the main line of reasoning of the second step, we can now zoom in on Kant’s justifications for the third and fourth premise.

4.2.3.4 Premise 3: The a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects Given the foregoing considerations, the justification for this premise can be presented in the form of the following brief argument. 1. Our a priori cognitions of the substantive theorems of Euclidean geometry are based on the a priori intuition of space. (Result from step 1 of the transcendental exposition of space.) 2. All empirical objects conform to the substantive theorems that we a priori cognize in Euclidean geometry. (Alleged fact, granted for the sake of the argument.) 3. Our a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects. (Third premise of part (B) of the master argument; from 1 and 2.) As already noted in section 4.2.1, while I take it that considerations concerning the possibility of geometry as an a priori synthetic science are central to the main argument for the claim that space is a form of sensibility that Kant foregrounds in the Transcendental Aesthetic, I agree with the view that these kinds of considerations are not essential to establishing this result. One of my reasons for agreeing with this view is that the third premise of part (B) of the master argument, which, according to Kant’s presentation of this argument in the Transcendental Aesthetic, is to be justified in the geometry-involving way just presented, can also be supported in another way. On first glance, it might seem as if the third premise is a fairly direct consequence of the comparatively uncontroversial assumption that what we might call the kind of spatiality of an object in space, for example, that it is three-dimensional and has a Euclidean nature, corresponds to the kind of spatiality of the space in which it happens to be, for example, that this space is a three-dimensional Euclidean space.⁸³ Given this assumption and given that, as our original representation of space, the a priori intuition of space represents a general scheme of order that characterizes

⁸² See Prol, 4:284: “But this [how one can know a priori what the intuition of things must be like] is very well intelligible as soon as both [space and time] are regarded as nothing more than formal conditions of our sensibility and the objects merely as appearances; for in that case the form of appearance, i.e., pure intuition, can indeed be represented out of ourselves, i.e., a priori.” ⁸³ This assumption is uncontroversial, at least, if we also grant that “in every part of space there is material substance, i.e., parts that are in themselves moveable,” which Kant establishes in the Metaphysical Foundations, see MAN, 4:504.

       

225

the kind of spatiality of space, including its Euclidean nature, it follows that this a priori intuition also represents a general scheme of order that characterizes the kind of spatiality of all objects in space, and, thus, governs their spatial determinations in accordance with the Euclidean postulates. And since it is uncontroversial that all empirical objects are in space, it seems that we have thus easily justified our third premise. On second glance, however, one realizes that in order for this justification to be cogent, another assumption must be made, namely, that the space in which empirical objects are in the empirical world, that is, empirical space, necessarily has the same kind of spatiality as the space that is represented in our a priori intuition of space, that is, pure space. One way in which this assumption, in turn, could be supported is by appeal to the alleged fact that all empirical objects necessarily conform to the theorems of geometry. Since we cognize these theorems on the basis of our a priori intuition of pure space, if the kind of spatiality of empirical space were or could be different from the kind of spatiality of pure space, the geometric theorems would not necessarily apply to all empirical objects. If this were the only possible line of support for the assumption in question, we would thus have discovered that the indicated ‘alternative’ justification for the third premise of part (B) of the master argument depends on considerations concerning the possibility of geometry as an a priori synthetic science after all, and thus is not really an alternative. There is another line of support, however, which does not depend on any such considerations. Recall S-, according to which any determinate space, that is, any determinate figure or any determinate spatial volume, and, more generally, any spatial determination is necessarily contained in pure space, and any determinate time, that is, any determinate temporal duration, and, more generally, any temporal determination is necessarily contained in pure time. As we saw in section 4.1.1, this claim turns out to be required as one of two additional premises for deriving TI1, the thesis of the transcendental ideality of empirical objects, from TI2, the thesis of the transcendental ideality of space and time. In our earlier discussion, we postponed an examination of how S might itself be justified, but this is a good place to take up this examination. For this claim can also serve as a premise in a derivation of the assumption currently at issue, that is, that empirical space necessarily has the same kind of spatiality as pure space. As already noted in our earlier discussion in section 4.1.1, given that empirical space is a determinate space or at least constituted by spatial determinations, and empirical time is a determinate time or at least constituted by temporal determinations, S directly implies that empirical space and empirical time are necessarily contained in pure space and in pure time, respectively. Indeed, given that every space can contain determinate spaces and other spatial determinations, and every time can contain determinate times and other temporal determinations, S- implies that every space is either identical to or necessarily contained in pure space, and every time is either identical to or necessarily contained in pure time. So, if we grant that anything in space has the same kind of spatiality as the space in which it is contained, these results allow us to infer that empirical space necessarily has the same kind of spatiality as pure space.⁸⁴ We already encountered the main justifying ground for S⁸⁴ Kant’s belief that the features of pure space include features that determine a particular metric, namely, the Euclidean one, must be regarded as a mistake. Not every possible space is Euclidean. But I also think that this mistake is not fatal. The features of pure space that explicitly figure in the metaphysical exposition are more

226

 ’        

 in the context of our discussion of the Metaphysical Exposition in section 4.2.2.2. By way of an analysis of our ordinary representations of space, the metaphysical exposition shows that these representations rely on and presuppose a more fundamental representation of space—our original representation of space—which turns out to be an a priori intuition. One of the central features that this analysis uncovers is what I have called the fundamental feature of our ordinary representations of space, namely, that we necessarily conceive of and present any spatial determination as being delineated in, and any determinate space as a ‘cut-out’ part of the same all-encompassing space. So, the analysis reveals that for something to count as a determinate space or other spatial determination for us it must be contained in the one all-encompassing space that is represented by the a priori intuition of space. As Kant puts it in a previously quoted passage in the context of the third space argument, “one can represent only a unified space, and if one talks about multiple spaces, one understands by that only parts of one and the same all-unified space” (B39/A25). Similarly, the metaphysical exposition of the concept of time shows, by way of an analysis of our ordinary representations of time, that these representations rely on and presuppose a more fundamental representation of time—our original representation of time—which turns out to be an a priori intuition. This analysis uncovers that we necessarily conceive of and present any temporal determination as being delineated in, and any determinate time as a ‘cut-out’ part of the same all-encompassing time. So, the analysis reveals that anything that counts as a determinate time or other temporal determination for us must be contained in the one all-encompassing time that is represented by the a priori intuition of time. As Kant puts it, “[d]ifferent times are only parts of the very same time” (B47/A31–32). In contrast to all concepts of space and time, which contain their instances under themselves, the a priori intuitions of space and time, which first define what we mean by ‘space’ and ‘time,’ contain all of their instances in themselves, or, more precisely, in their presentational content. And that is all that we need in order to justify S-. Against this background, the promised alternative justification for the third premise of part (B) of the master argument that does not depend on any considerations concerning the possibility of geometry as an a priori synthetic science can be spelled out as follows: 1. Any determinate space and, more generally, any spatial determination is necessarily contained in pure space. (S- restricted to space, from the metaphysical exposition of the concept of space.) 2. Empirical space is necessarily contained in pure space. (From 1, as explicated above.) 3. Anything that is contained in a space has the same kind of spatiality as that space. (Supplemental assumption granted for the argument.) 4. Empirical space necessarily has the same kind of spatiality as pure space. (From 2 and 3.) 5. All empirical objects are in empirical space. (Uncontroversial premise.) general than specific metric features and independent of Euclideanism. Kant could retreat to a more general conception of pure space, that is, more general than conceiving of it as Euclidean space, without any substantial loss to the overall argument. Obviously, on this modified account, the explanandum of the transcendental exposition would also have to be modified accordingly. Instead of explaining the possibility of Euclidean geometry as an a priori synthetic science, we would have to explain the possibility of, say, Riemannian geometry as an a priori synthetic science.

       

227

6. All empirical objects have the same kind of spatiality as pure space. (From 3, 4, and 5.) 7. The a priori intuition of space represents the kind of spatiality of pure space, including its Euclidean nature. (From the result of the metaphysical exposition that the a priori intuition of space is the original representation of space.) 8. The a priori intuition of space represents the kind of spatiality of all empirical objects, including their Euclidean nature. (From 6 and 7.) 9. The a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects. (Third premise of part (B) of the master argument; from 8, by spelling out what ‘representing a certain kind of spatiality’ means.)

4.2.3.5 Premise 4: The a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects only if space is a form of outer intuition in the broad sense Regardless of whether we arrive at the result that the a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects by thinking about the conditions for the applicability of the a priori theorems of geometry to these objects or by thinking about how all objects in space inevitably have the same kind of spatiality as pure space, the question arises how we can possibly have such an intuition. This is the question that Kant explicitly puts at the center of all of his presentations of part (B) of the master argument, in the Transcendental Exposition, in paragraph (b) of the Conclusions, and in §8 of the Prolegomena.⁸⁵ How can we have an a priori intuition on the basis of which we can learn something about empirical objects? To bring into sharper relief why our possession of such an intuition is remarkable, it is useful to distinguish two different kinds of puzzle that arise with respect it. The first puzzle consists in the general question of how there could be a representation that is both sensible and a priori. Given that all our intuitions are sensible and thus generated through affections of sensibility, it might seem as if any kind of a priori intuition is impossible for us, not just a priori intuitions of genuine existents. For, on the prima facie plausible assumption that a representation’s apriority is equivalent to it not depending on any affection of sensibility, an a priori intuition appears to be a contradiction in terms. So, in order to explain how we could have an a priori intuition of space, we need to make intelligible how it is possible for a representation of ours to be both sensible and a priori. The second puzzle consists in the more specific question of how an intuition that is a priori and, thus, not caused by empirical objects, could still encode substantive information about them. What is the connection between the a priori intuition and empirical objects that accounts for this mysterious correspondence? Although not an intuition of genuine existents or their determinations, the a priori intuition of space, in effect, amounts to a representation of a general property, or of a determinable, that is instantiated in all genuine existents of a certain kind, namely, the determinable of all empirical objects of having some ⁸⁵ Recall B41: “How can an outer intuition be in the mind that precedes the objects themselves and in which the concept of the latter can be determined a priori?” And B42/A26: “ . . . how the form of all appearances could be given . . . a priori in the mind, and how it as a pure intuition in which all objects must be determined could contain principles of their relations before all experience.” And Prol, 4:281: “For now the question is: how is it possible to intuit anything a priori?” Also see Prol, 4:282: “But how can intuition of an object precede the object itself?”

228

 ’        

spatial determinations all of which are governed by a general scheme of order that is defined by the Euclidean postulates. For ease of reference, in the following I will refer to this determinable as ‘being Euclidean.’ One could thus say that the a priori intuition of space, in effect, at the same time is an a priori intuition of the determinable of being Euclidean that all empirical objects possess. So, the second puzzle can also be captured by the question of how to make sense of the fact that the a priori intuition of space is not only intentionally related to space, its intentional object, but also refers to a determinable of each empirical object. This is a puzzling question, indeed, at least if we continue to disregard the possibility of a pre-established harmony. Given that the a priori intuition of space is not causally dependent on empirical objects, what could possibly account for the resemblance or match between the intuition’s presentational content and the determinable of being Euclidean of these objects on which the reference of the former to the latter depends? In §8 of the Prolegomena, right after having concluded that our mathematical cognitions must be based on an a priori intuition, Kant gives voice to this two-pronged kind of puzzlement, the second prong of which arises as soon as we reflect on the fact that the mathematical theorems derived in this way are true of empirical objects. The difficulty appears to increase rather than decrease with this step. For now the question is: how is it possible to intuit something a priori? Intuition is a representation as it would immediately depend on the presence of the object. For this reason, it appears to be impossible to a priori originally intuit, since in that case the intuition would have to take place without an object that was present earlier or is present now to which it is related, and thus could not be intuition. (Prol 4:281–282)

I take it that the reason for the seeming mysteriousness of our ability to a priori intuit general properties or determinables of genuine existents that Kant means to express here is that both the sensible nature of intuitions and the requirement of a causal or otherwise determining connection between an intuition and its referent appear to be incompatible with us having intuitions of the indicated kind. Given that intuitions depend on affections of sensibility and refer in virtue of the resemblance between their presentational content and their referent, it seems that unless a genuine existent is present and affects us we cannot possibly intuit any of its properties, not even its determinables. While a full explanation of the possibility of an priori intuition of space that represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects also requires addressing the question of how we can have representations that are both sensible and a priori, as far as the justification for premise four of part (B) of the master argument is concerned, only the question of the necessary conditions for the possibility of the reference of this intuition to the determinable of being Euclidean of each empirical object is directly relevant. So, we will first think about the latter question and return to the former question in section 4.2.4 after our examination of part (B) has been concluded. So, then how is it possible for the a priori intuition of space to encode substantive information about empirical objects? Or, more precisely, how is it possible for the a priori intuition of space to refer to the determinable of being Euclidean of each empirical object? As noted, Kant holds that, unless we assume that there is some kind of causal connection or

       

229

other kind of determination relation between an intuition and its referent, the resemblance between the presentational content of the intuition and the intuitive features of the referent, on which the reference depends, would be a miracle. If there is no causal connection or other kind of determination relation between the a priori intuition of space and empirical objects, it would be mysterious how the former could encode any substantive information about the latter. In the case of empirical intuitions, as we have seen, the relevant determination relation consists in that the empirical object to which or to whose determinations the intuition refers empirically causes the intuition. But this is not the only possible determination relation that can obtain between intuitions and objects. The intuition can also be an at least partial cause or determining ground of the object. We are already familiar with this kind of reversal of perspective from Kant’s celebrated Copernican Turn. But while the perspective reversal in the ‘classic’ Copernican Turn, as described in the B-preface, is performed in order to directly make the possibility of a priori synthetic judgments intelligible, in the present case it is performed, in the first place, in order to make the possibility of the reference of a priori intuitions intelligible.⁸⁶ So, unless we want to embrace a pre-established harmony, the a priori intuition of space at the same time refers to the determinable of being Euclidean of each empirical object only if it determines these objects in an appropriate way. How could this determination work? How can the a priori intuition of space determine empirical objects in such a way that they each have the determinable of being Euclidean? We already know the general shape of the answer, which can be conveyed by saying that the required determination works in virtue of the fact that the determinable of being Euclidean turns out to be part of the ‘form’ of empirical objects, and this ‘form’ is due to us. So, let us try to pin this down more precisely. Given that we are not talking about acts of willing, where objects are created according to a pre-conceived plan, the only way in which the determination in question can happen is if the determinable of being Euclidean is (strongly) mind-dependent,⁸⁷ and, for every empirical object, the a priori intuition of space either ontologically specifies that the object is Euclidean or determines the representation that ontologically specifies that the object is Euclidean in such a way that this representation necessarily presents the object as being Euclidean.⁸⁸ Call the representation that ⁸⁶ See B124–125/A92–93: “There are only two possible cases in which synthetic representation and its objects can come together, necessarily relate to one another, and meet each other, as it were: either if the object makes the representation or the representation makes the object possible. In the former case, the relation is only empirical and the representation is never possible a priori. And this holds for appearance with respect to what in them belongs to sensation. But in the second case, since representation in itself does not bring about its object as far as its existence is concerned (for we are not talking about the causality by means of the will), the representation is still a priori determining with respect to the object if it is possible through it alone to cognize something as an object.” Of course, over the course of the investigation in the Transcendental Analytic it emerges that explicating the possibility of the reference of a priori intuitions, as well as of a priori concepts, is an integral part of the larger project of explicating the possibility of a priori synthetic cognitions. In the Critique of Practical Reason, Kant is very explicit about this connection when he says of the main task of the first Critique, which as he puts it is to explain “how pure reason can cognize objects a priori,” that “it requires that one explicates beforehand how intuitions are possible a priori, without which no object is given to us anywhere and [without which] no object can be cognized synthetically either” (KpV, 5:45). Also see B73 and Letter to Herz, February 12, 1772, 10:130–132. ⁸⁷ The second step of part (B) of the master argument could thus be re-formulated as an argument for the claim that all empirical objects are mind-dependent. See KU, 5:351: “Just as the ideality of the objects of the senses as appearances is the only way in which the possibility can be explained that their forms can be determined a priori . . . ” (Of course, demonstrating the conclusion that all empirical objects are fully mind-dependent requires additional work, as we know.) ⁸⁸ As a reminder, a representation R ontologically specifies a property P of object O if, and only if, O has P in virtue of R presenting O as having P.

230

 ’        

ontologically specifies that empirical objects are Euclidean ‘experience*.’⁸⁹ Since the a priori intuition of space does not represent any empirical objects, it does not ontologically specify that they have any properties either, at least not by itself. This means that the a priori intuition of space refers to the determinable of being Euclidean of each empirical object only if this determinable is (strongly) mind-dependent, and the intuition determines experience* in such a way that experience* necessarily presents its objects as being Euclidean. What remains to be shown to complete the justification for premise four is that the a priori intuition of space determines experience* in such a way that it necessarily presents its objects as being Euclidean only if space is a form of outer intuition in the broad sense, that is, a general scheme of order, defined by the Euclidean postulates, that is responsible for all outer intuitions of finite objects necessarily presenting their objects as having spatial determinations and necessarily governs all spatial determinations. This last bit is the trickiest part of the argument. Reconstructing it requires chasing down several lines of reasoning that are only hinted at in Kant’s text. But this is a fair price to pay for a clear grasp of part (B) of the master argument. It is instructive to start by briefly revisiting how Kant himself explains the key move in the second main step of the transcendental exposition’s explanation of the possibility of geometry as an a priori synthetic science, which, as noted, largely corresponds to the second step of part (B) of the master argument that we are presently examining. As we already know, this second step of the transcendental exposition is aimed at making intelligible, without having recourse to a pre-established harmony, of how it could be that the a priori substantive theorems of geometry apply to all empirical objects. The key move is to claim that the a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects if, and only if, space is a form of outer intuition in the broad sense. The question is, why exactly should this be so? In the Transcendental Exposition in the Critique, Kant does not volunteer any additional justification for the indicated key move and seems to take it as self-evident. But in the corresponding sections in the Prolegomena, he offers a few brief explanatory comments. Right after the previously quoted passage from §9, which states that cognitions of empirical objects based on the a priori intuition of space are possible if, and only if, the latter represents nothing but a form of sensibility, Kant explains that this is so because we can know a priori “that objects of the senses can be intuited only in conformity with this form of sensibility,” which means that “propositions that merely concern this form of sensible intuition are possible and valid for objects of the senses, and, similarly, conversely, that intuitions that are a priori possible can never concern other things than objects of our senses” (Prol, 4:282). And later on in the first remark, after highlighting that “our sensible representation is in no way a representation of things in themselves but only of the manner in which they appear to us,” he pinpoints as the reason for the conformity of all objects to the theorems of Euclidean geometry that “sensibility whose form lies at the foundation of geometry is that on which the possibility of outer appearances rests” (Prol, 4:287). We find a sketch of roughly the same explanation in the opening section of the transcendental deduction of the categories in the Critique, where ⁸⁹ Of course, we later learn that experience* is or is included in experience as conceived in Kant’s theory and as explicated in sections 2.5–2.8. But, at the stage of the argument that we are currently considering, this has not yet been established.

       

231

Kant contrasts the ease with which it can be explicated that and how the forms of sensibility refer to determinables of empirical objects with the enormous difficulty of explicating that and how the forms of the understanding, that is, the categories, refer to determinables of empirical objects. Above [in the transcendental Aesthetic], we were able to make intelligible with little effort with respect to the concepts of space and time how they as cognitions a priori still must be necessarily related to objects and made [sic.] possible a synthetic cognition of them independently of all experience. For since only by means of such pure forms of sensibility an object can appear to us, i.e., be an object of empirical intuition, time and space are thus pure intuitions, which a priori contain the condition for the possibility of the objects as appearances, and the synthesis in them has objective validity. (B121–122/A89)⁹⁰

Based on these texts, it appears as if Kant takes the claim that the a priori intuition of space refers to the determinable of being Euclidean of each empirical object if, and only if, space is a form of outer intuition in the broad sense, to be justified for the simple reasons that (1) the a priori intuition of space determines all other outer intuitions in such a way that they necessarily present their objects as being Euclidean if, and only if, space is a form of outer intuition in the broad sense, and (2) all empirical objects essentially are intentional objects of outer empirical intuitions, or, more cautiously, all empirical objects essentially are sensible objects, that is, objects that can be represented by empirical intuitions. Unfortunately, this straightforward justification is not without problems. The problem with (1) is a comparatively minor quibble, but, for the sake of full disclosure, it is worth putting it down for the record. Clearly, it is true that if space, which we know to be the presentational content of the a priori intuition of space (from the first step of the argument), is a general scheme of order, defined by the Euclidean postulates, that is responsible for all outer intuitions of finite objects necessarily presenting their objects as having spatial determinations and necessarily governs all spatial determinations, then the a priori intuition of space determines all outer intuitions of finite objects in such a way that they necessarily present their objects as being Euclidean. That is, it is safe to say that, if space is a form of outer intuition in the broad sense, then the a priori intuition of space is responsible for all outer intuitions of finite objects necessarily presenting their objects as being Euclidean. It is not quite as clear, however, even if we continue to set aside the possibility of a pre-established harmony, whether and why the a priori intuition of space should be responsible for all outer intuitions of finite objects necessarily presenting their objects as being Euclidean only if space is a form of outer intuition in the broad sense. But, with some good will and in the absence of any alternative explanation of how the a priori intuition of space could be responsible for all outer intuitions of finite objects necessarily presenting their objects as being Euclidean, we will grant Kant that it is responsible for it if, and only if, space is a form of outer intuition in the broad sense. The problems with (2)—the claim that (i) all empirical objects essentially are intentional objects of outer empirical intuitions, or, more cautiously, (ii), all empirical objects ⁹⁰ Also see B122–123/A90: “For that the objects of sensible intuition must conform to the formal conditions of sensibility that lie a priori in the mind is clear from the fact that otherwise they would not be objects for us.” Also see FM 20:267.

232

 ’        

essentially are sensible objects, that is, objects that can be represented by empirical intuitions—are more serious. In the context of part (B) of the master argument, Kant is not entitled to assert (2i); and (2ii), although unobjectionable on its own, turns out to be insufficient to yield the desired justification for the key move of the second step of the transcendental exposition presently under discussion. Regarding (2i), while the (strong) mind-dependence of the determinable of being Euclidean of each empirical object can be identified as a necessary condition for the reference of the a priori intuition of space to this determinable, none of our foregoing considerations in the context of part (B) allow us to conclude that empirical objects are fully mind-dependent, let alone that they are ontologically specified by outer empirical intuitions (which, later on, they turn out not to be). Regarding (2ii), the claim that all empirical objects essentially are sensible objects is fairly uncontroversial. Furthermore, the representability of all empirical objects by outer empirical intuitions indeed implies that, if space is a form of outer intuition in the broad sense, all actual and possible empirical appearances of empirical objects in perception are necessarily presented as being Euclidean due to a determination by the a priori intuition of space, and may be granted to imply that all said appearances are necessarily presented as Euclidean due to a determination by the a priori intuition of space only if space is a form of outer intuition in the broad sense. But none of this is sufficient to show that all empirical objects are Euclidean due to a determination of the a priori intuition of space if, and only if, space is a form of outer intuition in the broad sense. In order for Kant’s strategy to work of justifying the key move in the second step of the transcendental exposition of space by relying on claim (1) that we just granted to him—the claim that the a priori intuition of space determines all outer intuitions of finite objects in such a way that they necessarily present their objects as being Euclidean if, and only if, space is a form of outer intuition in the broad sense—two additional claims must be established. First, it must be shown that outer intuitions of finite objects, in virtue of necessarily presenting their objects as being Euclidean, determine experience* in such a way that it necessarily presents its objects as being Euclidean. In that case, (1) allows us to conclude that, if space is a form of outer intuition in the broad sense, then experience* is determined by the a priori intuition of space in such a way that it necessarily presents its objects as being Euclidean. Second, in order to secure the only-if direction, it must be demonstrated that experience* necessarily presents its objects as being Euclidean only if it is determined to do so in the indicated way by outer intuitions of finite objects. Or, at least, it must be shown that any potential other pathway for experience* to be determined by the a priori intuition of space in the relevant way is also somehow necessarily bound up with space being a form of outer intuition in the broad sense. None of these tasks are easy, but the second one is an especially tall order—which is the main reason why I said above, in section 4.2.3.1, that the only-if-part of Kant’s explanation of how we can have an a priori intuition that represents space as well as a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects is more challenging than its if-part. Starting with task number one, as I see it, the key to establishing the relevant determination-connection between outer intuitions of finite objects and experience* is Kant’s claim that only through sensibility can objects be given to us, which makes its first appearances in the Critique in the first section of the Transcendental Aesthetic (B33/A19). As noted, a central component of what Kant has in mind with this dictum is that only

       

233

perception can furnish us with the ‘matter’ for the objects that are represented in our representations. As a result, all of our representations that are not purely ‘formal,’ including all representations of proper objects, are necessarily grounded in perception in some way. Now, if we assume that sensibility has forms, understood according to our definition of a form of sensibility, and that space is a form of outer intuition in the broad sense, it is essential to all intentional objects of empirical outer intuitions to have a ‘form’ in addition to their ‘matter,’ and this ‘form’ essentially includes that these objects are Euclidean. Accordingly, the ‘matter’ of the intentional objects of empirical outer intuitions is not and could not be given to us without being informed by the ‘form’ of being Euclidean. As a result, the ‘form’ of being Euclidean of the intentional objects of perception is necessarily carried over to the intentional objects of all representations that are grounded in perception for the sake of the ‘matter’ of their objects, including all intentional objects that are proper objects. So, if space is a form of outer intuition in the broad sense, the ‘form’ of being Euclidean is, in effect, a condition for the possibility of us representing proper objects, in the sense that all proper objects must be presented as having this ‘form,’ including all empirical objects.⁹¹ So, if space is a form of outer intuition in the broad sense, all representations of empirical objects, including experience*, necessarily present them as being Euclidean due to a determination by empirical outer intuitions, thanks to the necessary grounding of all representations that are not purely ‘formal’ in perception. Importantly, that empirical outer intuitions necessarily present their objects as being Euclidean plays a crucial role in this determination. This circumstance makes it plausible to conclude that, if space is a form of outer intuition in the broad sense, the a priori intuition of space ultimately determines experience* in such a way that it necessarily presents its objects as being Euclidean. For the a priori intuition of space is what determines all empirical outer intuitions in such a way that they necessarily present their objects as being Euclidean. Turning to the second task, in order to be able to infer from the result just derived the conclusion that the a priori intuition of space determines experience* in such a way that it necessarily presents its objects as being Euclidean only if space is a form of outer intuition in the broad sense, we need to establish that there is no other pathway for experience* to be determined by the a priori intuition of space in such a way that it necessarily presents its objects as being Euclidean, apart from the one just described that relies on the mediation by empirical outer intuitions. Or, at least, we need to show that there could not be such an alternative pathway unless space were a form of outer intuition in the broad sense. Let us start by thinking about whether there are any representations that could reasonably be said to be determined by the a priori intuition of space such that they necessarily present their objects as being Euclidean without any mediation by empirical outer intuitions or, more strongly, without any mediation by outer intuitions of finite objects.⁹² All representations that do not even present their objects as being Euclidean can be set aside right away. This group of representations includes inner intuitions, the categories, the ideas of reason, all representations of things in themselves, and all concepts of ‘non-tangible’ objects such as ⁹¹ See B29–30/A15: “Insofar as sensibility contained a priori representations that make up the condition under which objects can be given to us, it would belong to transcendental philosophy.” ⁹² If the determination of these representations by the a priori intuition of space were mediated by outer intuitions of finite objects, then it could not take place unless space were a form of outer intuition in the broad sense.

234

 ’        

justice, freedom, or happiness. Many empirical concepts present their objects as being Euclidean but since these concepts are generated by comparing, reflecting on, and abstracting from, empirical outer intuitions, their determination by the a priori intuition of space to necessarily present their objects as being Euclidean is mediated by empirical outer intuitions. But there are certain concepts that do seem to be determined by the a priori intuition of space in the relevant way, without the mediation of outer intuitions of finite objects, namely, the a priori sensible concepts of geometry, understood as embodying the Euclidean postulates, as for example the concept of a Euclidean triangle, that is, a triangle that has an internal angle sum of 180 degrees. Call these concepts ‘Euclidean concepts.’ Since the a priori intuition of space is our most fundamental representation of space that grounds all others, it is certainly correct to say that all Euclidean concepts are determined by the a priori intuition of space in such a way that they necessarily present their objects as being Euclidean. So, unless we can make plausible that experience* could neither include these Euclidean concepts nor be determined by them in such a way that it necessarily presents its objects as being Euclidean unless space were a form of outer intuition, the only-if-part of Kant’s explanation of the conformity of all empirical objects to the a priori synthetic theorems of Euclidean geometry, as well as premise four of part (B) of the master argument would be in jeopardy.⁹³ A full discussion of this question would require getting deep into Kant’s theory of schemata, which lies beyond the scope of the present investigation. So, I will limit myself to simply reporting that, on my reading, Kant takes schemata to be more fundamental than the corresponding concepts in that the latter

⁹³ Note that, on my reading, experience, as conceived in Kant’s final theory, does turn out to include Euclidean concepts as integral elements; indeed, these concepts are the representations in experience that ontologically specify that empirical objects are Euclidean. There is a lot to be said about the role of Euclidean concepts in the construction of experience, a role that, to my mind, has gone deplorably underappreciated by Kant himself. On my reading, these concepts are best understood as ‘spatial pendants’ to both the schematized categories and the concepts that figure in the formulation of the laws of the general theory of motion. A detailed discussion of these issues would lead us too far afield, but it is worthwhile to note that, although Kant’s accounts of space and time in the Transcendental Aesthetic are exactly parallel, there is a massive asymmetry in his treatment of them in the Analytic. A major topic of the Analytic is the nature and function of the schemata of the categories, which turn out to be transcendental time-determinations, that is, roughly put, representations of rules for the construction of time in experience in accordance with the structure represented in the a priori intuition of time. (See in particular the chapter on the schematism, B176–187/A137–147, and the section on the analogies of experience in the chapter on the transcendental principles in the Critique, B218–265/A176–218.) But what about transcendental spacedeterminations? Space also must be constructed in experience in accordance with the structure represented in the a priori intuition of space. As I see it, the Euclidean concepts are transcendental space-determinations and, as such, should have received some explicit discussion in the Analytic. (Revealingly, in paragraph 24 of the transcendental deduction, where we are first introduced to the idea that the categories apply to all empirical objects because they are implicitly involved in the construction of experience in the form of their schemata, Kant calls the synthesis of the imagination by means of which the transcendental unity of apperception is first brought to bear on the sensibly given material ‘figurative’ synthesis, that is, a synthesis directed at the generation of representations of figures in space; see B151.) If Kant were planning a third, revised edition, I would urge him to try to ameliorate this imbalance in his treatment of space and time. With respect to our concerns in the main text, Kant’s account of the schemata of the categories as transcendental time-determinations arguably also means that the schematized categories, which are the representations within experience that ontologically specify that empirical objects have temporal determinations that are necessarily governed by the general scheme of order that is pure time, are determined by the a priori intuition of time in such a way that they necessarily present their objects as having temporal determinations that are necessarily governed by the general scheme of order that is pure time. So, unless we can show that this determination of the schematized categories by the pure intuition of time is mediated by inner intuitions of finite objects or, at least, that there could not be any schematized categories that are determined in this way unless time were a form of inner intuition, the only-if-part of Kant’s explanation of the conformity of all empirical objects to the theorems of the pure theory of motion, and, with it, premise four of part (B) of the corresponding master argument for the case of time would be in jeopardy.

       

235

are derived from the former by means of a special kind of abstraction.⁹⁴ Moreover, due to their hybrid nature of being both intellectual and sensible, schemata are necessarily connected to the intuitions whose construction-rules they represent, in the sense that we cannot have a schema of a given concept unless we also have the intuition that is constructed by following the rule represented in the schema and that represents an instance of the concept.⁹⁵ Accordingly, the schemata of the Euclidean concepts are logically prior to them, and there cannot be any Euclidean concepts, understood as concepts that are determined by the a priori intuition of space in such a way that they necessarily present their objects as being Euclidean, without there also being outer intuitions of finite objects that are determined by the a priori intuition of space in such a way that they present their objects as being Euclidean. Since this result eliminates the most obvious and, as best I can tell, only threat to the claim that the a priori intuition of space determines experience* in such way that it necessarily presents its objects as being Euclidean only if space is a form of outer intuition in the broad sense, we might treat this claim as thereby justified. This completes the challenging last step in the justification for the fourth premise of part (B) of the master argument, which can be summarized in its entirety in the form of the following argument. 1. The a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects only if it refers to the determinable of being Euclidean of each empirical object. (First premise; from an analysis of what representing the indicated general scheme of order, in effect, amounts to.) 2. The a priori intuition of space refers to the determinable of being Euclidean of each empirical object only if it determines all empirical objects in such a way that they have the determinable of being Euclidean. (Second premise; from Kant’s account of the reference of intuitions, and the rejection of the possibility of a pre-established harmony.) 3. The a priori intuition of space determines all empirical objects in such a way that they have the determinable of being Euclidean only if this determinable is (strongly) mind-dependent, and the intuition determines experience*, that is, the representation that ontologically specifies that empirical objects are Euclidean, in such a way that it necessarily presents its objects as being Euclidean. (Third premise; from an analysis of how a representation could possibly determine a determinable of an empirical object.) 4. The a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects only if the determinable of being Euclidean of each empirical object is (strongly) minddependent, and the intuition determines experience* in such a way that it necessarily presents its objects as being Euclidean. (Intermediary conclusion; from 1, 2, and 3.) 5. The a priori intuition of space determines experience* in such a way that it necessarily presents its objects as being Euclidean if space is a form of outer intuition in the

⁹⁴ In this reading, I am in agreement with Heidegger’s assessment that the imagination, as the faculty in charge of the schematism, is the most fundamental cognitive faculty in Kant’s account of the human mind. See Heidegger 1991 (*1929), esp., §29, 146–55. ⁹⁵ See B177–188/A138–142.

236

 ’        

broad sense. (Fourth premise; from the fairly uncontroversial thesis that the a priori intuition of space determines all outer intuitions of finite objects in such a way that they necessarily present their objects as being Euclidean if, and only if, space is a form of outer intuition in the broad sense, and the assumption of a necessary grounding of all representations of proper objects in empirical outer intuitions.) 6. The a priori intuition of space determines experience* in such a way that it necessarily presents its objects as being Euclidean only if space is a form of outer intuition in the broad sense. (From 5, a plausibility argument to the effect that the Euclidean concepts are the only representations that are determined by the a priori intuition of space, without the mediation of outer intuitions of finite objects, in such a way that they necessarily present their objects as being Euclidean, and the result, based on considerations concerning the relation between schemata and their corresponding concepts, that we can have Euclidean concepts only if space is a form of intuition in the broad sense.) 7. The a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects only if space is a form of outer intuition in the broad sense. (Premise four of part (B) of the master argument; from 4 and 6.) Admittedly, the very last bit of this justification for premise four, captured in brackets in line 6, is a bit precarious—like most arguments that aim to establish a non-existence claim. It is thus useful to put a fallback strategy on the table for saving part (B) of the master argument, in case the indicated precariousness might tempt some readers to reject the entire justification attempt. This fallback strategy exploits the circumstance that it is not entirely obvious what kind of objects Kant has in mind when, in presenting part (B), he wonders how the a priori intuition of space could “contain principles of the relations of all objects before all experience” (B42/A26). The reading presented above, according to which he is thinking about empirical objects, is certainly the most straightforward and prima facie most plausible option. But taking him to be concerned about both empirical objects and the intentional objects of perception is an option too. On the latter reading, the third premise of part (B) states that the a priori intuition of space represents a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects and objects of perceptions, and the fourth premise says that this is the case only if space is a form of outer intuition in the broad sense. If we grant that the a priori substantive theorems of Euclidean geometry apply not only to empirical objects but also to the objects of perception, the first justification for premise three discussed above also works for the modified version of this premise. And since the objects of perception are spatial and, thus, in space, the second justification discussed above works as well. The justification for the modified version of premise four, mutatis mutandis, runs exactly parallel to the justification for the original version just presented up to and including line 4. But in this case, there is no need to take the detour, as it were, through the if-claim stated in line 5 to get to the only-if claim stated in line 6. In the modified version of the argument, the representations that the a priori intuition of space is supposed to determine in such a way that they necessarily present their objects as being Euclidean are not only all representations that ontologically specify that empirical objects are Euclidean but also all representations that ontologically specify that the objects of perception are Euclidean. Since the

       

237

representations that ontologically specify the objects of perception are obviously perceptions, that is, empirical intuitions, the thesis that the a priori intuition of space determines all outer intuitions of finite objects in such a way that they necessarily present their objects as being Euclidean if, and only if, space is a form of outer intuition in the broad sense, allows us to conclude directly that the a priori intuition of space determines all representations that ontologically specify that the objects of perception are Euclidean in such a way that they present their objects as being Euclidean only if space is a form of outer intuition in the broad sense. And since a necessary condition for p is also a necessary condition for p and q, this result directly implies the desired conclusion that the a priori intuition of space determines all representations that ontologically specify that the objects of perception are Euclidean and all representations that ontologically specify that empirical objects are Euclidean in such a way that they necessarily present their objects as being Euclidean only if space is a form of outer intuition in the broad sense.⁹⁶ More generally, in concluding I want to emphasize that even if the step from line 5 to 6 in the above justification for premise four of the original reconstruction of part (B) of the master argument were judged a failure and the fallback strategy just suggested were rejected, the damage would be far from devastating. We still would have shown that space is not only an intentional object of an a priori intuition but also a general scheme of order that is responsible for experience* necessarily presenting its objects as having spatial determinations that are necessarily governed by this scheme of order. That is, although we would not have shown that space is a form of outer intuition, we would still have shown that space is a form of experience—which is nothing to scoff at. This completes our examination of part (B) of the master argument and, with it, of the master argument itself, by means of which Kant demonstrates TI2-pure-strong, according to which space and, analogously, time are neither things in themselves nor constituted by the determinations of things in themselves but (nothing but) forms of sensibility.

4.2.4 How Can an Intuition be both a priori and Sensible? All that remains is to tie up the loose ends left dangling earlier concerning the question of how the apriority of the intuition of space can be reconciled with its sensible nature, a question whose discussion, while not integral to the master argument, is integral to a complete explanation of the possibility of us having an a priori intuition of space. How could there be a representation that both is a priori and depends on affections of sensibility? Thinking about this question will also reveal a further reason for my view that considerations concerning the possibility of geometry as an a priori synthetic science are not essential to establishing that space is a form of sensibility, despite being central to the main considerations that Kant highlights in the Transcendental Aesthetic in support of

⁹⁶ But is that not a rather lame argument? It does not even establish any kind of connection between experience* and the a priori intuition of space. Maybe this argument is a bit disappointing, but bear in mind that the full explanation of how our a priori intuition of space can be a representation of both space and a general scheme of order, defined by the Euclidean postulates, that governs all spatial determinations of empirical objects and of objects of perception includes not only the only-if-part on which we have focused here but also an if-part. And establishing a connection, via empirical outer intuitions, between experience* and the a priori intuition of space is a central element of this if-part, as we just saw above.

238

 ’        

this result. For it turns out that Kant’s account of how intuitions can be a priori, despite being sensible, readily translates into an alternative argument for the claim that space is a form of outer intuition in the broad sense. A representation is empirical if, and only if, it contains, or is in some way derived from representations that contain, sensations.⁹⁷ Sensations are produced by sensibility in response to it being affected by things in themselves. Accordingly, a representation is empirical if, and only if, it depends on affections of sensibility by things in themselves that produce sensations. These are the kind of affections to which I earlier referred as ‘material transcendental affections.’ (All of the affections featured in this section are transcendental affections. From now on, I will no longer explicitly say that, but this is how I should be understood.) This label reflects the fact that material affections provide us with the ‘matter’ of the objects that are represented in our representations. Given that material affections lead to empirical representations, the only way for there to be a priori intuitions, that is, representations that are non-empirical but still sensible, is if, in addition to material affections, there are also affections of a different kind, namely, affections that do not produce sensations. How should we think about these other affections? More specifically, what sort of contribution to our cognitions of objects should we ascribe to them, and what kind of entities are we to suppose are the affecting parties in this case? Empirical intuitions do not only contain sensations. They contain sensations that are ordered in certain ways; they represent objects that have ‘matter’ and ‘form.’ Given that material affections provide us with the ‘matter’ of objects, it is not unreasonable to assume that the indicated other kind of affections to which sensibility is subject provide us with the ‘form’ of objects or, at least, part of it. As Kant puts it in the Prolegomena, “this capacity to intuit a priori does not pertain to the matter of appearance, i.e., that which in it is sensation, since that makes up the empirical, but only to their form, i.e., space and time.” (Prol, 4:284) For this reason, I will call the kind of affections that generate a priori intuitions ‘formal affections.’ Turning to the question of the ‘perpetrator’ of these formal affections, Kant does not say very much about it, but the following passage contains some valuable hints. Now that which as representation can precede any action of thinking something is intuition and, if it contains nothing but relations, the form of intuition, which, since it represents nothing unless something is posited in the mind, can be nothing other but the manner in which the mind is affected through its own activity, namely, this positing of its representation, and thus is affected through itself, i.e., an inner sense according to its form. (B67–68)⁹⁸

These remarks suggest that formal affections are a special kind of self-affections of the mind.⁹⁹ More specifically, they are affections of sensibility through the activity of the mind

⁹⁷ See B34–35. ⁹⁸ Also see V-Met-K2/Heinze, 28:713: “All apprehension of the objects of the senses is an action of the mind whereby the human being affects itself. The effect belongs to receptivity, but the action itself to spontaneity. This building of the object into my power of representing is an affection of the subject.” Also see B151–152; B153–154. ⁹⁹ I say that formal affections are a special kind of self-affections of the mind because the material affections by means of which inner sensations are produced, which feature in inner empirical intuitions, are also a kind of selfaffections of the mind.

       

239

of “positing,” or taking up, representations in inner sense, in particular and in the first instance, sensations. The main idea seems to be that the affection of sensibility through this positing activity has the result that the intuitions that are generated in this way present the contents of the original representations as standing in particular relations of a certain type in accordance with certain general schemes of order, schemes that are determined by the nature of sensibility. Only those sensation-producing affections of sensibility by things in themselves that are attended by corresponding formal affections of sensibility by the positing activity of the mind lead to empirical intuitions that represent objects with both ‘matter’ and ‘form.’ Material affections yield sensations, which correspond to the ‘matter’ of the objects that are represented in our empirical intuitions. Formal affections, supplemented by certain unifying operations, yield a priori intuitions whose presentational contents are general schemes of order that are responsible for and reflect the ‘form’ of the objects that are represented in our empirical intuitions.¹⁰⁰ Or, in other words, formal affections, on their own, yield a priori intuitions whose presentational contents are general schemes of order that are responsible for all intuitions of a certain type necessarily presenting their objects as having determinations of a certain type that are necessarily governed by these schemes, that is, they yield a priori intuitions that represent forms of intuition in the broad sense. So, the receptive faculty of our mind that is sensibility includes both a capacity to produce sensations and a capacity to produce a priori intuitions that represent forms of intuition in the broad sense. While the latter capacity is usually exercised in connection with the former, it can also be exercised on its own since it depends only on a special kind of self-affection of the mind. I take it that this is what Kant is getting at in passage (b) of the Conclusions when he says that “since the receptivity of the subject to be affected by objects necessarily precedes all intuitions of these objects, it is intelligible how the form of all appearances could be given before all real perceptions, and [how it could be given] a priori in the mind” (B42/A26). Given the distinction between material and formal affections of sensibility just discussed, it is easy to see how the sensible nature of the intuition of space is compatible with its apriority. While all representations that depend on some affections of sensibility are sensible, only representations that depend on material affections are empirical. So, as long as the intuition of space depends only on formal affections, there is no problem about it being both sensible and a priori. The sketched account of the genesis and content of a priori intuitions establishes that the a priori intuitions of space and time directly represent forms of intuition in the broad sense, as just noted. Together with the result that space and time essentially are intentional objects of a priori intuitions (as established in part (A) of the master argument), this yields the conclusion that space and time are forms of sensibility. So, if we take on board the sketched account, we thus have an alternative demonstration for the thesis that space and time are forms of sensibility up our sleeves. This alternative demonstration proceeds by

¹⁰⁰ I say that formal affections, supplemented by some unifying operations, yield a priori intuitions because, like all intuitions, a priori intuitions also must be governed by the transcendental unity of apperception to be something for us. Relatedly, also note that formal affections are involved in both the generation of the a priori intuition of space and the generation of a priori intuitions of finite determinate spaces, for example, the a priori intuition of a triangle. But the generation of the latter depends on yet further unifying operations, encoded by various concepts, for example, the concept of a Euclidean triangle. See B160 161, note, quoted in note 67, chapter 3.

240

 ’        

simply reflecting on what it takes for there to be non-empirical but still sensible representations and does not involve any considerations about the possibility of the reference of a priori intuitions to determinables of empirical objects or about the possibility of geometry as an a priori synthetic science. While this alternative demonstration is not part of Kant’s official main argument for TI2-pure-strong in the Transcendental Aesthetic or in the corresponding sections in the Prolegomena, it is adumbrated in both.¹⁰¹ One possible reason why Kant chose part (B) of the master argument as its ‘star’ argument in the Transcendental Aesthetic instead might be that the above story of the genesis and content of a priori intuitions involves a few steps that rely on assumptions that, although not implausible, have less warrant than the assumptions involved in part (B). I am thinking here in particular of the central assumption that the affections that are involved in the generation of a priori intuitions are responsible for the ‘form’ of the objects of our representations, given that the affections that produce sensations and are involved in the generation of empirical intuitions are responsible for the ‘matter’ of the objects of our representations. This assumption is underwritten by Kant’s claim in the first section of the Transcendental Aesthetic that [s]ince that in which sensations alone can order themselves and can be placed in a certain form cannot again itself be sensation, it follows that while the matter of all appearance is given to us only a posteriori, their form must altogether lie ready for them in the mind a priori, and thus must be able to be considered in abstraction from all sensation. (B34/A20)

Unfortunately, this claim is far from obvious. Without further argument, it is not clear why sensations cannot furnish us with both the ‘matter’ and the ‘form’ of objects. To my mind, the most charitable reading of the intended function of the outlined story about the generation and content of a priori intuitions is to understand it as meant, not to provide an alternative demonstration for the claim that space and time are forms of intuition in the broad sense, but to supplement Kant’s explanation of how we can have an a priori intuition of space by fleshing out how there can be representations that are both sensible and a priori.

4.2.5 Coda: the Allegedly Neglected Third Possibility Before concluding our discussion of Kant’s arguments for TI2-pure-strong in the Transcendental Aesthetic, I feel obliged to add a few comments about one of the oldest and most notorious objections to his theoretical philosophy, already briefly mentioned in section 3.4.2, the so-called neglected third possibility objection (NTPO for short from now on). This objection is directed at exactly the part in Kant’s argument in the Transcendental Aesthetic that we just examined, that is, the part where he concludes from “the above ¹⁰¹ In the Critique, see in particular §1 of the Transcendental Aesthetic. In the Prolegomena, see, for example, Prol, 4:283 (my emphasis): “ . . . space and time remain, . . . which, through the fact that they are pure intuitions a priori, demonstrate that they are mere forms of our sensibility, which must precede all empirical intuition, i.e., perception of actual objects, and according to which objects can be cognized a priori but of course only how they appear to us.”

       

241

concepts” that space and time are transcendentally ideal and (nothing but) forms of sensibility. The objection was first articulated by Kant’s contemporary Hermann Pistorius but became notorious through a famous dispute between Adolf Trendelenburg and Kuno Fischer about whether there is a “gap” in Kant’s argument in the Transcendental Aesthetic. (Trendelenburg thinks there is a gap, Fischer thinks there is not.)¹⁰² NTPO is presented in various formulations, but the common gist is that the earlier sections before the Conclusions at best show that our original representations of space and time are a priori intuitions, and space and time are forms of sensibility (in some sense¹⁰³), and that these results are not sufficient for the conclusion that space and time are transcendentally ideal and nothing but forms of sensibility. The earlier results, so the objection goes, may allow Kant to rule out the possibility that space and time are nothing but forms of things in themselves, that is, that they are only ‘objective,’ as the participants in the historical debate like to put it.¹⁰⁴ But they are compatible with two further possibilities. The second possibility still in the running is the option that Kant himself favors, namely, that space and time are nothing but forms of sensibility, that is, that they are only ‘subjective.’ The third possibility is that space and time are both forms of sensibility and forms of things in themselves, that is, that they are objective and subjective. Or, alternatively and more weakly, the third possibility is that there are two distinct but very similar kinds of space and time, one of which is subjective and one of which is objective. According to the objection, Kant overlooked the third possibility and consequently failed to show that it does not actually obtain. As a result, he is not justified in asserting that space and time are transcendentally ideal and nothing but forms of sensibility, which also means that his further inference to the transcendental ideality of empirical objects is not warranted either. How damaging is NTPO to Kant’s case for transcendental idealism in the Transcendental Aesthetic? The first thing to say to an objector who formulates NTPO by claiming that Kant overlooked and failed to rule out the actuality of, the possibility that space and time are both forms of sensibility and forms of things in themselves is that he (the objector) is confused about what a form of sensibility is, or, more charitably, that he is working with a conception of a form of sensibility that deviates from Kant’s. As already noted in section 3.4.2, on Kant’s conception, it is plainly impossible for a form of sensibility to be a form of things in themselves as well. This kind of response to NTPO was made famous by Reinhold. He argues that a form of sensibility is a form of representation or

¹⁰² See Pistorius 1786, 99–107; Pistorius 1788, 432–8. (Interestingly, Pistorius himself sees the third possibility in that space and time might be grounded in things in themselves. See Pistorius 1786, esp. 100; 1788, 437 8. As we saw in section 3.4.3, Kant, in fact, happily admits that empirical space and time are grounded in things in themselves.) Also see Jakob 1786, 26; Maaß 1788, esp. 120–2. See Trendelenburg 1862, 163 4; Fischer 1865, 173 80; Trendelenburg 1867; Fischer 1869, iv–xvi, 263 5, 315 16, 322 5, 328 30, 335 6, 338 40, 547 50; Trendelenburg 1869; Fischer 1870. My discussion of NTPO is indebted to Vaihinger’s excellent account in Vaihinger 1892, 134–51, 290–326. Vaihinger’s verdict, however, is that NTPO does present a genuine problem for Kant and that there really is a gap in the argument in the Transcendental Aesthetic, while my verdict is that there is no such problem and no such gap. Also see Allison 2004, 129–32, whose discussion of NTPO owes much to Vaihinger as well. In his ultimate response to NTPO, Allison avails himself of a proposal by Falkenstein whose work on the topic is well worth consulting too. See Falkenstein 1989, and Falkentein 1995, ch. 9, 289–309. ¹⁰³ Obviously, proponents of this objection understand what it means to be a form of sensibility in a weaker sense than it is understood here. We will address this point shortly. ¹⁰⁴ Kant himself occasionally uses the formulation that space and time are “objective forms” to characterize the option that space and time are forms of things in themselves, for example, see B71–72.

242

 ’        

representing, and since “the form of representation is that through which representation distinguishes itself from everything that is not representation,” anybody who admits “that the thing in itself is not a representation” also has “to admit that the form of representation does not pertain to it” (Reinhold 1789, 246).¹⁰⁵ I would not put the response in quite these terms, but I agree with Reinhold that the key to seeing why the strong version of the alleged third possibility, as formulated above, is in fact not a possibility is that the forms of sensibility are essentially tied to representations, or, more precisely, to presentational contents. As discussed, on my reading, in order for a general scheme of order to be a form of sensibility it is not enough for it to be a form of intuition in the broad sense, that is, to be responsible for all intuitions of finite objects necessarily presenting their objects as having determinations of a certain type and to necessarily govern all determinations of that type. In addition, the general scheme of order in question must itself be an intentional object of an a priori intuition, which means that a form of sensibility essentially is a form of presentational contents. And since things in themselves are not presentational contents, no form of sensibility could possibly pertain to them. So, if we can conclude from the “above concepts” that space and time are forms of sensibility, we can also conclude that they are nothing but forms of sensibility and could not possibly be forms of things in themselves as well. Accordingly, on a charitable reading of the strong version of NTPO, we should assume that the objector is working with a conception of a form of sensibility that differs from Kant’s and leaves room for a form of sensibility that is a form of things in themselves as well. The most obvious candidate for such an alternative conception equates the forms of sensibility with forms of intuition in the broad sense. A general scheme of order that is a form of intuition in the broad sense could also be a form of things in themselves in the sense that it is mind-independent and governs a certain type of determinations of things in themselves or even in the sense that it is mind-independent and both governs and is partly responsible for a certain type of determinations of things in themselves. NTPO would then be the complaint that the “above concepts” allow Kant to infer only that space and time are forms of intuition in the broad sense, which, however, is not enough to rule out the actuality of the neglected third possibility that they are both forms of intuition in the broad sense and forms of things in themselves. In response, Kant neither neglects the indicated possibility nor does he fail to show that it is not actual. As we just saw, the structure of his master argument is not that he concludes from the “above concepts” that space and time are forms of intuition in the broad sense and then blithely infers from this result that space and time are not forms of things in themselves. Rather, the claim that space and time are not forms of things in themselves is established first, in part (A) of the argument, as summarized in paragraph (a) of the Conclusions. The thesis that space and time are forms of intuition in the broad sense is only shown in part (B) of the argument, as summarized in paragraph (b) of the Conclusions, and, more precisely, in part (B)’s second step. The master argument thus starts out by explicitly refuting the actuality of the allegedly neglected third possibility under consideration. Space and time are not both forms of intuition in the broad sense and

¹⁰⁵ For a discussion of Reinhold’s response and the reactions triggered by it, see Vaihinger 1892, 312 15.

       

243

forms of things in themselves, since they are demonstrably not forms of things in themselves. This is the point in the dialectic at which proponents of NTPO commonly retreat to a weaker version of the third possibility, which, they hope, translates into a more challenging objection. Maybe Kant has shown that there is a kind of space and a kind of time that are forms of sensibility (in my sense) and, hence, not forms of things in themselves. But what he has not addressed is the possibility that there is another kind of space and another kind of time that are forms of things in themselves. This other kind of space and this other kind of time would be general schemes of order that are isomorphic to, or perhaps even exactly similar to, the forms of sensibility, except for being mind-independent and pertaining to things in themselves. In this case, there would be two kinds of space and time: subjective space and subjective time, which are forms of sensibility, and objective space and objective time, which are forms of things in themselves. As long as this possibility has not been ruled out or at least been shown not to be actual, the general thesis that space and time are transcendentally ideal and nothing but forms of sensibility and, with it, the thesis that empirical objects are transcendentally ideal, remain unjustified. Our previous discussion in this chapter puts us in the position to dispose of this version of NTPO in expedient fashion as well. All we need is (1) S-, established in the metaphysical expositions, which, you will recall, states that any determinate space and, more generally, any spatial determination is necessarily contained in the space represented by our a priori intuition of space, that is, pure space, and any determinate time and, more generally, any temporal determination is necessarily contained in the time represented by our a priori intuition of time, that is, pure time, (2) the result, entailed by S-, that any space is either identical to or necessarily contained in pure space, and any time is either identical to or necessarily contained in pure time,¹⁰⁶ and (3) the claim, entailed by S-, that if x is in space S/time T, the mode of being of all of x’s determinations and other ontological ingredients is the same as the mode of being of S/T. (1), (2), (3), and the full mind-dependence of pure space and pure time allow us thus to conclude that there cannot possibly be any kind of space or any kind of time that is mind-independent.¹⁰⁷ Whether there are or could be mindindependent general schemes of order that pertain to things in themselves and that are isomorphic to space and time or in every way like space and time except for being mindindependent, is completely irrelevant for Kant’s master argument and his argument for the transcendental ideality of empirical objects.¹⁰⁸ Whatever those mind-independent general

¹⁰⁶ Recall B39/A25: “ . . . one can represent only a unified space, and if one talks about multiple spaces one understands by that only parts of one and the same all-unified space . . . ” Also recall B47/A31–32: “Different times are only parts of the very same time . . . ” ¹⁰⁷ We can derive the same result, without appealing to S-, by relying on (a) the ontological priority of pure space and pure time compared to all other kinds of space and time, understood as implying minimally that any space/time has all of its space-infected/time-infected determinations partly in virtue of being in pure space/time, (b) the plausible assumption that any kind of space/time is necessarily fully spatial/ temporal in character, respectively, such that it is impossible for there to be a space/time that has any ontological ingredients that are not space-infected/time-infected, and (c) the full mind-dependence of pure space and pure time. ¹⁰⁸ Indeed, I take it that Kant would be happy to admit that there could be such ordering schemes, at least as an epistemic possibility.

244

 ’        

schemes of order might be, neither one of them would count as a kind of space or a kind of time for us. Given that our a priori intuitions of space and time are our original representations of space and time, which first define what we mean by ‘space’ and ‘time,’ respectively, anything that counts as a kind of space or a kind of time for us is identical to or necessarily contained in pure space or pure time, respectively, and hence fully minddependent.¹⁰⁹

¹⁰⁹ This response to NTPO is similar in spirit to the one suggested in Willaschek 1997, esp. 554–6. Willaschek claims that the result of the Metaphysical Exposition is that our representation of space is an a priori intuition, which means that “if we understand by ‘space’ that to which our representation of space refers, then space cannot be a ‘real’ object.” That is, “even if there should be something independently from our ‘mind’ to which all general marks of space apply (three dimensionality, homogeneity, infinity, etc.), it would not be space, insofar as we understand by this the object of our representation of space” (Willaschek 1997, 555, my translation). In my assessment, this response to NTPO is on the right track, but as presented by Willaschek it is a bit too quick. A proponent of NTPO could come back and point out—plausibly in my view—that, surely, Kant does not want to deny the obvious fact that we also have conceptual representations of space. In the Metaphysical Exposition, we learn, among other things, that we have an a priori intuition of space; but we do not learn there or anywhere else that we do not have any concepts of space. So, even if it is true that the space that is the object of our a priori intuition must be mind-dependent, the possibility has not been ruled out yet that there is a different kind of space that is mind-independent and pertains to things in themselves, a space to which we can refer with our concepts of space. The most obvious way in which Willaschek could pre-empt this kind of rebuttal is by showing that our concepts of space depend on our a priori intuition of space in such a way that the former must refer to whatever the latter refers to or to parts of whatever the latter refers to. And, to my mind, the best way to do that, in turn, would be to utilize the result of the Metaphysical Exposition that our original representation of space, on which all others depend, is the a priori intuition of space. The resulting argument, when spelled out, would be a variant of the argument just sketched in the main text.

5 Things in Themselves and their Relation to Appearances 5.1 Things in Themselves Ground Appearances, Appearances are Appearances of Things in Themselves Kant is committed to the transcendental ideality of several existents—empirical objects, empirical selves, and space and time—but he is also committed to the transcendental reality of some existents and has views about how these different kinds of existents are related to each other. Kant often explicitly expresses his conviction that there are things in themselves, as for example in the Prolegomena where he explains that, while it would be nonsense to hope that we can know anything about things in themselves, “it would be an even greater nonsense if we did not admit any things in themselves at all” (Prol, 4:350).¹ His commitment to the existence of things at the transcendental level of reality may be the reason why, upon further reflection, he preferred the label ‘critical idealism’ to ‘transcendental idealism’ as a name for his overall ontological position. ‘Transcendental idealism’ might be taken to suggest that he holds that all there is, is transcendentally ideal―which, however, is not his view. ‘Critical idealism’ is thus preferable in that it indicates that his position includes a genuine idealist element but without misleadingly suggesting that for him there are no other but transcendentally ideal entities. Even though things in themselves and appearances are distinct and exist at different levels of reality, some things in themselves stand in a special, close relationship to appearances. They are the supersensible grounds of appearances. . . . [R]eason relentlessly demands the unconditioned for the given conditioned, which however is never to be found if one regards the sensible as belonging to things in themselves and does not rather posit for it, as mere appearance, something supersensible (the intelligible substrate of nature outside us and in us) as thing in itself as its ground. (KU, 5:345)²

Appearances and their states thus have empirical causes and empirical existence conditions at the empirical level of reality, namely, other empirical objects and events, as well as transcendental grounds at the transcendental level of reality, namely, things in themselves, ¹ Also see KpV, 5:6: “ . . . since it [the Critique] emphasized to regard the objects of experience as such only as appearances, and among them even our own subject, but still to give them things in themselves as their ground, and, thus, not to take everything supersensible as fantasy and its concept as empty of content . . . ” Also see Prol, 4:361 361 and note 56, chapter 6. ² Also see KU, 5:196: “The understanding, through the possibility of its a priori laws for nature, gives proof that nature is cognized by us only as appearance and thus at the same time gives indication of a supersensible substrate of nature, but leaves it completely undetermined.” See B428; B641 642/A613 614; Prol, 4:354; Prol, 4:35; ÜE, 8:203; KU, 5:344; KU, 5:409; KU, 5:422. The World According to Kant: Appearances and Things in Themselves in Critical Idealism. Anja Jauernig, Oxford University Press (2021). © Anja Jauernig. DOI: 10.1093/oso/9780199695386.003.0005

246

       

which, as ultimate material conditions for the possibility of experience, are among the transcendental existence conditions of appearances. The grounding of appearances in things in themselves obtains thanks to the important role that perception, and, more specifically, sensation, plays in the construction of experience. Things in themselves ground appearances in that they transcendentally affect sensibility and thereby bring about sensations, and that sensations and perceptions are the material conditions for the possibility of experience, which contribute the ‘matter’ of appearances and underwrite their existence, as explicated in section 2.5.3. After . . . asking ‘Who (what) gives sensibility its matter, i.e., the sensations?’ he [Eberhard] believes himself to have spoken against the Critique in saying: ‘We can choose what we want—we end up with things in themselves.’ Now, that is exactly the constant assertion of the Critique; except that it posits this ground of the matter of sensible representations not itself in things, as objects of the senses, but in something supersensible, which is the underlying ground of the former and of which we cannot have any cognition. (ÜE, 8:215) The sensible faculty of intuition, basically, is only a receptivity to be affected with representations in a certain manner. . . . The non-sensible cause of these representations is completely unknown to us, and we thus cannot intuit it as an object . . . . (B522/A494)³

Another way of putting the point that things in themselves ground appearances is to say that appearances are appearances of things in themselves; they are how things in themselves appear to us, or, more precisely, they are what things in themselves appear to us as. Indeed, if we regard the objects of the senses appropriately as mere appearances, as is proper, we still admit thereby at the same time that a thing in itself grounds them although we are not acquainted with it as it is in itself but merely with its appearance, i.e., the manner in which our senses are affected by this unknown something. (Prol, 4:314–315) The representation of a body in intuition contains nothing at all that could pertain to an object in itself but merely the appearance of something and the manner in which we are affected by it . . . . (B61/A44)⁴

³ Also see B334/A278: “Its [sensibility’s] relation to an object . . . without doubt lies too deeply hidden for us . . . to be able to use such an inappropriate tool of our investigation in order to find something other than always appearances again whose non-sensible cause, however, we would like to investigate.” See V-Met/Mron, 29:861: “If we look at appearances, they all hang together according to the laws of nature. But all appearances still also have a transcendental cause, which we do not know.” Also see B235/A190; A358; B574/A546. ⁴ See Prol, 4:290: “ . . . it was proved by us that sensibility consists, not in this logical difference of clarity or obscureness, but in the genetic difference of the origin of the cognition itself, since sensible cognition represents things not at all as they are but merely the manner in which they affect our senses, and thus that through it merely appearances, not the things themselves are given to the understanding for reflection.” See GMS, 4:451: “ . . . all representations that come to us involuntarily (such as those of the senses) do not let us cognize the objects other than how they affect us, where what they may be in themselves remains unknown to us, and thus . . . , as far as this kind of representations is concerned, . . . we only get to the cognition of appearances, and never of the things in themselves. Once this distinction is made . . . it follows automatically that one has to admit and assume behind the appearances still something else that is not appearance, namely, things in themselves, although we humbly acknowledge on our own accord that we cannot get closer to them and never know what they are in themselves, since they can never be known to us but always only how they affect us.” See A387: “But we should take note that bodies are not objects in themselves that are present to us but a mere appearance of who knows what unknown object, that motion is not the effect of this unknown cause, but merely the appearance of its influence on our senses . . . ” See B55/A38; Prol, 4:289.

    

247

That appearances are appearances of things in themselves that ground them underwrites their ontological status as existents from the point of view of fundamental ontology and sets them apart from other intentional objects that are not so grounded, or, at least, not in a comparable, properly direct way. I take it that this is what Kant has in mind when he says of the ‘matter’ of appearances that it is the “real” or the “reality” in them “through which something existing is given.”⁵ The ‘matter’ of appearances is the real in them in that, since it depends on sensation, which, in turn, depends on affections of sensibility by things in themselves, it ‘anchors’ appearances in the most fundamental level of reality and thereby gives them the extra ontological weight that is required for them to count as genuine existents. In this way, the grounding of appearances in things in themselves also adds another degree of robustness to Kant’s empirical realism, so to speak. The empirical reality of empirical objects consists not only in the conformity of experience to Kant’s formal conditions of objectivity but also in the grounding of empirical objects in transcendentally real things.⁶ Mere coherence and lawfulness is not enough for experience; experience also depends on ‘friction’ supplied by entities that exist at the transcendental level of reality, as it were. That appearances are grounded in things in themselves is the reason why TI1-strong-b and TI3-strong-b, according to which empirical objects are Kantian outer appearances, and empirical selves are Kantian inner appearances, although ostensibly only about empirical objects and empirical selves, amount to claims about the relation between them and entities at the transcendental level of reality. So, appearances and some things in themselves are intimately connected; and yet, despite their close relationship, they are numerically distinct and do not share any ontological ingredients. Based on the foregoing remarks and our discussion of the role of perception in the construction of experience in section 2.5.3—providing the ‘matter’ for appearances and underwriting their existence—we already have a pretty good grip on what it means for appearances to be grounded in things in themselves. Nevertheless, there is still plenty of potential for confusion about how exactly the grounding relation and the corresponding appearance relation should be understood and about what these relations do and do not entail with respect to the relata that instantiate them. The potential for confusion is great, both because appearance-talk is inherently confusion prone and because the appearance relation is familiar to us from our ordinary thought and talk as well as from our prior ⁵ See KU, 5:189: “Sensation (in this case outer sensation) expresses just as well the merely subjective of our representations of things outside us, but it properly expresses the material (real) of them (through which something existing is given) . . . ” See B609/A581: “ . . . but that which makes up matter, the reality in the appearance, (what conforms to sensation) must be given . . . ” See B182/A143: “Reality in the pure concept of the understanding is that which corresponds to a sensation in general; thus that whose concept in itself indicates a being (in time). . . . that which in appearances corresponds to sensation is the transcendental matter of all objects, as things in themselves (thingness, reality).” See B210/A168: “What in empirical intuition corresponds to sensation is reality (realitas phaenomenon) . . . ” See KU, 5:203: “All relation of representations can be objective, even that of the sensations (and in this case it signifies the real of an empirical representation) . . . ” See V-Met/ Mron, 29:829: “The matter of all appearances is the sensation, and what corresponds to it is real.” See R5814, 18:361. ⁶ The special grounding of empirical objects in things in themselves on Kant’s view is also emphasized by Karl Ameriks, in Ameriks 2011, esp. 38–40. But Ameriks takes the “characteristic of requiring a particular kind of more fundamental real ground” as “distinctive of Kant’s notion of ideality” (Ameriks 2011, 39), by which he means transcendental ideality. While I agree that this characteristic is distinctive of Kant’s critical idealism, I do not agree that it is part of Kant’s understanding of transcendentally ideality. To say of something that it is transcendentally ideal just means that it is ideal with respect to the transcendental level of reality, that is, that it is fully minddependent (from the point of view of fundamental ontology). On my reading, the characteristic of requiring a particular kind of more fundamental real ground is distinctive of Kant’s notion of empirical reality.

248

       

exposure to philosophical reflections about it. The problem is that, in the context of Kant’s philosophy, the expressions ‘appearance’ and ‘thing in itself ’ function as technical terms, whose meaning and use differ in important ways from how these terms are used in ordinary language or in philosophical discussions that take their cue from ordinary language. If we do not pay proper attention to these differences, confusion is inevitable.

5.2 Confusion Prevention I: the Ambiguity of the Term ‘Appearance’ To begin with, one has to be careful not to get thrown off track by the ambiguities of the expression ‘appearance of x.’ A first ambiguity is specific to Kant, due to his recognition of both a transcendental and an empirical version of the distinction between things in themselves and appearances. The empirical version, to be discussed in more detail in the following section, is also recognized in our ordinary thought and talk and distinguishes between empirical objects and their appearances to us in perceptions or illusions. The latter are empirical appearances, or appearances in the empirical sense. So, the first thing that one needs to get clear on when encountering the expression ‘appearance of x’ in Kant is whether he is talking about empirical or transcendental appearances. A second ambiguity, which can be detected, not only in Kant’s use, but also in ordinary English and even more so in ordinary German, at least in some contexts, concerns the ‘x’ in the expressions ‘appearance of x’ and ‘x appears.’ Assume that A appears to a subject S as B, that is, A does the appearing, as it were, and B is the appearance. In response to the question ‘what is the appearance an appearance of?’, in Kant’s use, as well as in ordinary English and German, at least in some contexts, it is equally acceptable to say that it is an appearance of A or that it is an appearance of B. Similarly, in response to the question ‘what appears to S?’ it is equally acceptable to say that A appears to S as it is to say that B appears to S. That is, these questions could be understood as asking about the transcendental ground or proximate cause of S’s appearance representation, on the one hand, or about its presentational content, on the other hand. As an example of this ambiguity in ordinary English, take the case where a large spider in the dimly lit bicycle room in my apartment building appears to me as a small cockroach. It is equally permissible to describe this situation by saying either that the appearance confronting me is an appearance of a spider, and that a spider appears to me, or that the appearance confronting me is an appearance of a cockroach, and that a cockroach appears to me. When Kant says that appearances are appearances of things in themselves or that things in themselves appear to us, as in some of the quotations above, he is using ‘appearance of x’ and ‘x appears’ in the first sense, that is, in the sense ‘appearance of A’/‘A appears’ from our template above, and intends to signify the transcendental ground of the appearance. But when he talks about appearances of empirical objects and means ‘appearance’ in the transcendental sense, for example, the appearance of a house, he is using the term ‘appearance of x’ in the second sense, that is, in the sense of ‘appearance of B,’ and intends to signify the content of the appearance representation. It is crucial to be clear about these ambiguities so as to avoid being tricked into drawing any false conclusions. One false conclusion, based on Kant’s talk about appearances of empirical objects, would be to assume that Kant’s transcendental distinction between things in themselves and appearances just is the distinction between empirical objects, such as a house, and their

     ‘  ’

249

appearances, while, in truth, it is a distinction between supersensible entities and their appearances, such as a house, which are empirical objects. Another false conclusion would be to infer that Kant must regard appearances and things in themselves as numerically identical, on the basis of claims such as that we can consider the same thing as it appears to us and as it is in itself, or that we cannot know the things that appear to us as they are in themselves, or that we can cognize or know certain things only as they appear to us, or as appearances, or only in the appearance, but not as they are in themselves. The expression ‘the thing that appears to us’ in this context must be understood in the sense of ‘A appears to us’ in our template. To consider a transcendentally real thing, call it ‘X,’ as it appears to us does not mean to consider it as an appearance; it means to consider it as appearing as a certain appearance, for example, to consider X as appearing as an opossum, but not to consider it as an opossum. Saying that we cannot know the things that appear to us as they are in themselves is not to say that we cannot know the things that are appearances as they are in themselves—although our current knowledge of this kind is fairly limited, given that science is still progressing—but that we do not know the things that do the appearing apart from the fact that they appear to us. Similarly, the claim that we can cognize or know certain things only as they appear to us, or as appearances, or only in the appearance, but not as they are in themselves, is not to say that the things that are appearances are the same things as the things that are unknowable in themselves and do the appearing. Rather, it is to say that we can cognize or know the things that do the appearing only in the manner in which they appear to us, or more precisely, that we cannot cognize or know anything about them apart from the fact that they appear to us in this way, or as these appearances.⁷ Also note that it makes sense to say that it is a feature or aspect of certain transcendentally real things that they appear to us as certain appearances. But despite the use of the term ‘aspect,’ this description does not amount to any kind of concession to the two-aspect interpretation. Asserting that it is an aspect of a thing in itself that it appears as a certain appearance is perfectly compatible with holding that the thing in itself and the appearance are numerically distinct and do not share any ontological ingredients. Similar remarks apply to the expression ‘(empirical) intuition of x.’ The ‘x’ could stand for the presentational content, the referent or empirical proximate cause, or the transcendental ground of the intuition. Usually, when Kant employs this kind of expression, for example, when he talks about the intuition of a house, he is using it in the first or second sense. But since empirical intuitions are partly due to affections of sensibility by things in themselves, he also sometimes uses it in the third sense, for example, when he says that “the receptivity of the subject to be affected by objects necessarily precedes all ⁷ Additional passages that ought to be read in this way (that is, passages in addition to some of the passages cited in section 5.1) include the following: ÜE, 8:208: “ . . . the distinction between the cognition of objects as appearances and the concept of them according to what they are as things in themselves . . . ” See Bxx: “ . . . that consequently the unconditioned must be found not in things insofar as we know them (as they are given to us), but in them, insofar as we do not know them, as things in themselves . . . ” Bxxvi: “That . . . we cannot have cognition of an object as a thing in itself but only insofar as it is an object of sensible intuition, i.e., as appearance, is proved in the analytical part of the Critique.” See Prol, 4:289: “ . . . things are given to us as objects of the senses that are outside us, but of what they may be in themselves we do not know anything, but only know their appearances, i.e., the representations that they effect in us by affecting our sensibility.” See R 6317, 18:627: “ . . . it [intuition], as pure as it may be, is always only sensible and represents the object not as thing in itself, but merely in the appearance . . . ” See V-Met-K3/Arnoldt, 29:1020: “But we cannot cognize things in themselves, what and how they are in themselves, but only in the appearance.” Also see B67/A49.

250

       

intuitions of these objects” (B42/A26).⁸ Again, it is vital to beware of this ambiguity in order not to mistakenly assume that, on Kant’s view, the things that are represented in, or referred to by our intuitions and the things that transcendentally affect sensibility to produce (or contribute to the production of) those intuitions are the same things, or that empirical objects transcendentally affect sensibility and thereby bring about intuitions of themselves. Kant’s view is that things in themselves transcendentally affect sensibility and thereby bring about (or contribute to the bringing about of) intuitions, which are (a) of themselves in the sense that they are the intuitions’ transcendental grounds and (b) of empirical objects in the sense that these objects are the intuitions’ referents and empirical proximate causes, causes that are distinct from the transcendental grounds.

5.3 Confusion Prevention II: the Empirical Distinction between Things in Themselves and Appearances Turning to potential confusions due to our prior understanding of the appearance relation, we often say of empirical objects that they appear to us as something that they are not, or that they appear in a way that obscures, or at least does not transparently represent, their true nature, for example, when we say that a cat appears as a raccoon, or that a straight stick in water appears bent, or, to use Kant’s examples, that a collection of rain drops appears as a rainbow, a rose appears red, or Saturn appears to have handlebars.⁹ In these kinds of cases, the cat, the stick, the collection of rain drops, the rose, and Saturn are regarded as things in themselves, while the raccoon, the property of being bent, the rainbow (understood as a band of colors), the property of being red, and the property of having handlebars are regarded as appearances, or appearance properties (which, as before, are to be understood as mind-dependent properties). This is a distinction between things in themselves and appearances with respect to the empirical level of reality, or in an empirical sense.¹⁰ We usually distinguish among appearances that which essentially adheres to their intuition and which holds for every human sense in general from that which pertains to it only contingently in that it is not valid with respect to the relation of sensibility in general but merely with respect to the special position or organization of this or that sense. And in this case one calls the former cognition one that represents the object in itself but the second one merely its appearance. But this distinction is merely empirical . . . (B62/A45)¹¹

⁸ Also see B59/A42: “ . . . the things that we intuit are not in themselves as what we intuit them nor are their relations constituted in themselves such as they appear to us . . . ” ⁹ See B63/A45; B45/A29–30; B70; V-Met/Mron, 29:857. ¹⁰ Note that the term ‘appearance’ in the empirical sense can even be used in situations in which there is no thing that appears, in the sense of ‘doing the appearing.’ For example, the objects represented in hallucinations or dreams, such as Joe’s oasis and my purple dragon from section 2.2, may also be called ‘appearances’ in the empirical sense. ¹¹ See FM, 20:269: “Furthermore, it is also to be noted that appearance taken in the transcendental sense, where one says of things that they are appearances (phenomena), is a concept of an entirely different meaning than if I say that this thing appears to me such or such, which is supposed to indicate the physical appearances and can be called apparentia [Apparenz] or seeming. For in the language of experience, these objects of the senses are thought of as things in themselves, since I can only compare them with other objects of the senses, e.g., the sky with all its stars, even though it is a mere appearance, and if we say of it that it appears to be a vaulted ceiling, the seeming

  

251

The empirical distinction between things in themselves and appearances is a distinction between empirical objects and their appearances in the subjective representations of particular human subjects, such as perceptions and illusions.¹² In such an empirical context, where something appears to us as something other than it is (a rose appears red, a straight stick appears bent, a cat appears as a raccoon), the common verdict is that there are not two existents (a non-colored rose with certain surface reflectance properties and a red rose, a straight stick and a bent stick, a cat and a raccoon) but only one (a noncolored rose with certain surface reflectance properties, a straight stick, a cat).¹³ In these cases, the empirical proximate cause of the sensations contained in the perception or illusion is the existent that the perception or illusion refers to, even though the presentational content of the perception or illusion does not capture the empirical object in itself but only how it appears to us. (Probably needless to say, in the case of dreams or hallucinations, there is no referent at all.) Accordingly, insofar as one wanted to say that there is an existent that has the properties that are represented in perceptions and illusions, the only choice would be to hold that these properties belong to the empirical object as it appears to a particular kind of observer in a particular kind of circumstance, or that they are appearance properties of the empirical object, that is, properties of appearing in a certain way to a particular kind of observer in a particular kind of circumstance. On this understanding, the empirical object as it appears to a particular kind of observer in a particular kind of circumstance, or the empirical object as bearer of certain appearance properties, could be regarded as empirical appearance, while the empirical object as it is in itself could be regarded as an empirical thing in itself. In other words, the empirical distinction between things in themselves and appearances could naturally and reasonably be understood in two-aspect fashion, and often is understood in this way. I take it that Kant would agree with this assessment—that the empirical distinction could reasonably be understood in two-aspect fashion—even though his own preferred understanding is slightly different and amounts to a two-world reading but also with only one existent. Instead of identifying the appearance with the empirical object as it appears to a particular kind of observer in a particular kind of circumstance, or as a bearer of certain appearance properties, as on the two-aspect reading just sketched, one could also identify it with the intentional object of the relevant perception or illusion. This is how Kant usually tends to proceed, as we will see in section 5.4. On a view like this, the thing in itself in the empirical sense, that is, the empirical object, is numerically distinct from and does not

here signifies the subjective in the representation of a thing, which can be a cause to mistakenly regard it as objective in a judgment.” Also see B313–314/A258; A393; Anth, 7:146. ¹² Illusions, strictly speaking, I take it, are essentially conscious, just like perceptions. ¹³ The claim that empirical objects as they are in themselves do not have color properties might strike some readers as surprising. But it is important to remember that the properties of empirical objects, as transcendental appearances, are determined by our ew-accounts, which are accounts of the empirical world based on science. And colors, understood as sensations or qualia, which is how Kant understands them, are simply not part of the scientific description of empirical objects, neither in the eighteenth century nor today. (Of course, if colors were understood as powers to cause color sensations in us, then empirical objects could also be said to be colored.) The colors that empirical objects appear to have depend on the specific circumstances in which these objects are observed and can depend on specific features of individual observers (as Kant himself points out in B45/A29–30 and FM, 20:268 269), but the properties that they have in themselves do not suffer from any of these dependencies. We will talk more about Kant’s understanding of secondary qualities, including colors, in section 5.4.

252

       

share any ontological ingredients with the appearance in the empirical sense. But only empirical objects are genuine entities; the intentional objects of perceptions and illusions do not genuinely exist; their way of being is mere pseudo-existence. This is why this view is a two-world one-existent view. Although Kant himself understands even the empirical distinction between appearances and things in themselves in two-world fashion, it still is fair to say that it could naturally and reasonably be understood in two-aspect fashion and, on any reasonable understanding, does not amount to a distinction between distinct existents. This changes radically, however, once we turn to the transcendental case, which is why the two cases must be carefully distinguished. In saying that the two cases must be carefully distinguished, my main point is not that the distinction between a rose and its red appearance or a straight stick and its bent appearance is not an instance of Kant’s transcendental distinction. This is true and bears repeating but should also be rather obvious. My point is to highlight that our familiarity with the empirical distinction makes it natural and tempting to project the way in which we tend to think about the appearance relation in the empirical case onto the transcendental case and use it as a model for understanding the transcendental distinction, and that doing so would be a grave mistake.¹⁴ Kant shares this assessment. Here is how he continues the passage quoted above: But this distinction is merely empirical. If one stops there (as it ordinarily happens) and does not regard this empirical intuition, in turn, as mere appearance (as it should happen), so that in it nothing can be encountered that has anything to do with a thing in itself, our transcendental distinction is lost . . . (B62/A45)

Similarly, a little earlier Kant expresses the same kind of warning as follows: The purpose of this remark is merely to prevent one from thinking to explicate the asserted ideality of space through hugely inadequate examples, since namely, say, colors, taste etc. are rightly not regarded as properties of things but merely as modifications of our subject, which even can be different in different people. For in this case that which is originally merely appearance, e.g., a rose, in the empirical understanding passes for a thing in itself, which, however, can appear differently to each eye with respect to color. By contrast, the transcendental concept of appearances in space is a critical reminder that nothing at all that is intuited in space is a thing in itself, and that space is not a form of things that belonged to them in themselves, but that the objects in themselves are not known to us at all, and that what we call outer objects are nothing but mere representations of our sensibility, whose form is space but whose true correlate, i.e., the thing itself, neither is nor can be cognized through this but about which we also never inquire in experience. (B45/A29–30)¹⁵ ¹⁴ Allais has exactly the opposite view: “My suggestion is in fact that starting by thinking about this ordinary empirical distinction is the best way to understand Kant’s position” (Allais 2004, 673). See Allais 2007, 478. Another commentator who uses the bent-stick case as a model for transcendental idealism, with results quite similar to Allais’s, is Andrew Roche in Roche 2013, see esp. 594–9. ¹⁵ Also see B52 53/A36: “But this ideality [of time] is as little to be compared with the subreptions of sensation as the ideality of space, since in that case one presupposes of the appearance itself in which these predicates inhere that it has objective reality, which completely falls away here, except insofar as it [the reality] is merely empirical, i.e., regards the object itself merely as appearance.” Also see FM, 20:268 269.

  

253

By insisting that his transcendental distinction must be carefully distinguished from the empirical distinction, part of what Kant means to emphasize is that appearances in the empirical sense are more subjective and typically more private than appearances in the transcendental sense. Perceptions and illusions do not conform to all of Kant’s conditions of objectivity and experience—as spelled out in sections 2.5.2 and 2.5.3—and their intentional objects depend not only on the cognitive make-up of the human mind in general and on what things in themselves exist but also on special contingent empirical circumstances, such as the particular vantage point of the observer and the prevalent lighting conditions, and, possibly, on special features of the observer’s cognitive capacities or sense organs. But that is not all that Kant wants to convey here.¹⁶ He also means to highlight that there are several crucial differences with respect to how our representations, appearances, and things in themselves are related in the two cases, despite some undeniable similarities (on which more in section 5.4). In the transcendental case, where a thing in itself appears as an empirical object, for example, where X appears as an opossum, it is also legitimate to say that the presentational content of experience does not capture X but only how X appears to us, namely, as an opossum. But it is not legitimate to say that X is what experience refers to, let alone what our opossum perceptions refer to, nor is it legitimate to say that there is only one existent, namely X, and that X is the opossum in itself. There are two existents, without any ontological overlap between them, X and the opossum, one mind-independent and one fully minddependent. Given that X is the transcendental ground of the opossum, we may say that experience, our opossum perceptions, and the opossum, all indicate, or signify, the existence of X, in a way comparable to how smoke may be said to indicate the presence of fire or a fever may be said to signify the presence of an infection. But the existent that experience and our perceptions refer to is the fully mind-dependent opossum.¹⁷ Kant urges caution in using the empirical distinction between things in themselves and appearances as a model for how to think about the transcendental distinction precisely because, on the very natural two-aspect understanding of the empirical distinction, it is only certain properties or aspects of empirical objects that are regarded as minddependent, while the crucial point in the transcendental case is that empirical objects, together with all of their determinations and other ontological ingredients, are to be ¹⁶ On the interpretations by Allais and Rosefeldt, for example, the difference between the empirical and the transcendental case does not amount to much more than this kind of difference with respect to the respective degree of subjectivity of the relevant appearances. See Allais 2007, 480–1; Rosefeldt 2007, 182–3. Roche provides a more sophisticated analysis of the difference between the two cases, but in arguing that “transcendental idealism uses only the entirely familiar notion of appearing: appearinge,” that is, appearing in the empirical sense, and that appearances in the transcendental sense are merely “objects as they ought to appeare to us” (Roche 2013, 604), he also defends a view on which appearances in the transcendental sense are not different in kind but merely more objective versions of appearances in the empirical sense. ¹⁷ See A372–373: “Now one can admit that something that is outside us in the transcendental understanding is the cause of our outer intuitions, but this is not the object that we understand by the representations of matter and corporeal things; for those are merely appearances, i.e., mere kinds of representations, which are always only in us. . . . The transcendental object is equally unknown with respect to inner as with respect to outer intuition. But we are also not talking about it but about the empirical object, which is then called an outer object if it is represented in space, and an inner object if it is represented only in relations of time; space and time, however, are only to be found in us.” See A378: “They [the idealistic objections] drive us with force . . . to regard all perceptions . . . merely as a consciousness of that which belongs to our sensibility, and [to regard] the outer objects of them not as things in themselves but only as representations . . . , which are called outer because they adhere to the sense that we call outer sense, whose intuition is space, but which is itself nothing other than an inner kind of representations . . . ”

254

       

regarded as mind-dependent. This is exactly what Kant highlights in the passage from B45/A29 30 just quoted. In the case of secondary qualities and the empirical distinction between things in themselves and appearances, the empirical object to which the secondary qualities are ascribed “passes for a thing in itself.” But in the case of spatiotemporal properties and the transcendental distinction, the empirical objects to which the spatiotemporal properties are ascribed are themselves appearances, or “nothing but mere representations of our sensibility.”¹⁸ This is also precisely what I take Kant to be getting at in the last sentence of the passage at B62/A45: if one stops with the distinction between the properties that an empirical object has in itself and its appearance properties, as it usually happens, and does not regard this empirical object, in turn, as mere appearance, as one should, our transcendental distinction is lost.¹⁹ As the foregoing remarks make clear, a central question of contention in the dispute with proponents of the two-aspect view who regard the empirical distinction between things in themselves and appearances as a suitable model for the transcendental distinction is what kind of things Kant takes the referents of perceptions to be. All parties agree that these referents are empirical objects, but this is where the agreement ends. Since the twoaspect commentators in question hold that empirical objects have an in-itself aspect that is mind-independent, they are committed to the view that we perceive mind-independent existents, albeit not in a way that discloses how they are in themselves but only how they appear to us. Lucy Allais puts the point by saying that perceptions do not “transparently reveal the mind-independent natures of the objects that appear” (Allais 2007, 474; also see Allais 2015, ch. 5.iii). On my reading, by contrast, Kant holds that the referents of perceptions are appearances, that is, fully mind-dependent existents. Mind-independent existents or mind-independent aspects of existents are not perceived at all, not even nontransparently. With respect to our example of X that appears as an opossum, on my reading, what I perceive is the opossum, not X; on the relevant two-aspect view, X is what I perceive, albeit non-transparently and only as an opossum. In section 3.1, we saw that Kant classifies empirical objects as appearances without adding any qualifications. On the assumption that the referents of perceptions are empirical objects, this result also directly implies that these referents are appearances, not mind-independent things that have an appearance aspect. This reading receives additional textual support from Kant’s critique of the Leibniz-Wolffian conception of the difference between sensible and intellectual representations. As already reported in section 3.5, one of the main reasons why Kant finds the Leibniz-Wolffian conception of sensible representations as confused concepts problematic is that it leads them to a ‘transcendental amphiboly,’ that is, in the first place, a confusion of phenomena with noumena and, in the second place, on account of their view that the objects of the pure understanding are things in themselves and vice versa, a confusion of appearances with things in themselves. The confusion consists in that they misunderstand ¹⁸ Similar remarks apply to the corresponding passage regarding the transcendental ideality of time at B52 53/ A36, cited in note 15. In the empirical case, the empirical object to which the secondary qualities are ascribed is regarded as ‘objectively real,’ that is, as existing mind-independently (from the empirical point of view). But in the transcendental case, the empirical object to which the spatiotemporal properties are ascribed is itself an appearance (from the point of view of fundamental ontology). ¹⁹ James Van Cleve arrives at a similar assessment in his discussion of the question of whether the distinction between primary and secondary qualities could be used as a model for the transcendental distinction: “It is because the secondary quality model involves a residual object that Kant thinks it is insufficient to illustrate his own view” (Van Cleve 1999, 168).

  

255

the distinction between appearances and things in themselves in two-aspect fashion as merely amounting to a distinction between different ways in which we represent the very same mind-independent existents, namely, distinctly in terms of pure concepts, on the one hand, and confusedly in terms of sensible representations, on the other hand, rather than understanding it as a distinction between different existents. For the same reasons, the Leibniz-Wolffians are also committed to the view that we perceive mindindependent existents, albeit only confusedly, that is, albeit only as they appear to us.²⁰ Kant has no love for either one of these consequences of the Leibniz-Wolffian misconception of the difference between sensible and intellectual representation. His rejection of the first consequence—that the distinction between things in themselves and appearances is not a distinction between different things but merely a difference in how we cognitively relate to things—illustrates his disagreement with the methodological twoaspect view, as noted in section 3.5. His rejection of the second consequence—that the referents of our perceptions are mind-independent things, which we perceive only as they appear to us but not as they are in themselves—illustrates his disagreement with two-aspect views that model the transcendental distinction between things in themselves and appearances on the empirical distinction. Our present focus is on the latter disagreement. By contrast, the representation of a body in intuition does not contain anything at all that could pertain to a thing in itself but merely the appearance of something and the manner in which we are affected through it. And this receptivity of our ability to cognize is called sensibility and remains utterly distinct from the cognition of the object in itself, even if one were to grasp it (the appearance) down to its ground. The Leibniz-Wolffian philosophy gave a completely illegitimate perspective to all investigations concerning the nature and origin of our cognitions, in that it considered the difference of sensibility from the intellectual as merely logical, even though it is obviously transcendental and does not merely concern their form of distinctness or indistinctness but their origin and content, so that by means of the former we cognize the nature of things in themselves not merely indistinctly but not at all, and, as soon as we take away our subjective quality, the represented object with the properties that sensible intuition gave to it, is nowhere to be found nor can be found, insofar as this very subjective quality determines its form, as appearance. (B61–62/A44)

Sensible and intellectual representations not only differ with respect to how they represent, they also differ with respect to what they represent and, more specifically, what kind of things they refer to.²¹ Kant’s critical correction to the Leibniz-Wolffian view is to say, not ²⁰ For textual evidence, revisit notes 136 and 137, chapter 3. ²¹ See ÜE, 8:219–220: “Thus, sensibility, as Mr. Eberhard ascribes this concept to Leibniz, either is distinguished from the cognition of the understanding merely through the logical form (confusion), while still containing only intellectual representations of things in themselves as far as the content is concerned, or it is distinguished from it also transcendentally, i.e., with respect to origin and content, in that it contains nothing at all of the quality of objects in themselves but merely the manner in which the subject is affected, may it otherwise be as distinct as it wants. In the latter case, that is the claim of the Critique, to which one cannot oppose the former opinion without positing sensibility [to consist] merely in the confusion of representations that the given intuition contains.” See Anth, 7:140–141, note: “To posit sensibility [to consist] merely in the indistinctness of representations, but intellectuality in distinctness, and, thereby, to posit a merely formal (logical) difference of consciousness instead of a real (psychological) [difference] that concerns not only the form but also the content of thinking, was a

256

       

that instead of confusedly perceiving mind-independent things we perceive them nontransparently, but that we perceive them not at all.²² The referents of perceptions are appearances, that is, fully mind-dependent things, not mind-independent things with appearance properties. As discussed in section 2.4, Kant in fact agrees with the LeibnizWolffians that perceptions involve a considerable amount of confusion, or indistinctness, in the sense that they comprise many ‘little’ intuitions that are not individually apprehended with consciousness. But he insists that even if we could resolve all of this indistinctness, the result would still not be a representation of a thing in itself but a distinct (sensible) representation of an appearance.²³ And if, per impossibile, we could divest ourselves of the forms of sensibility, the result, again, would not be a representation of a mind-independent thing but the complete disappearance of the referent of our perception, which, as Kant puts it in the quoted passage, would “nowhere to be found,” precisely because it is nothing but an appearance, that is, because it is a fully mind-dependent thing, whose existence and all of whose determinations depend, at least in part, on the forms of sensibility. Having clarified that the referents of perceptions are appearances for Kant, it is also important to underline that, on account of their grounding in things in themselves, appearances can rightly be said to indicate, or signify, things in themselves. [T]hey [appearances] really relate to something that is distinct from them (and thus completely dissimilar), in that appearances always presuppose a thing in itself and, accordingly, indicate it, one may know it more closely or not. (Prol, 4:355)

Kant’s often repeated claim that objects can be given to us only through sensibility undoubtedly is primarily concerned with appearances. Appearances are given to us only through sensibility in that only perception furnishes us with their ‘matter’ and underwrites their existence. But I believe it is legitimate to read Kant’s dictum as having the important

big mistake of the Leibniz-Wolffian school . . . ” I take it that, in these passages, by ‘content’ Kant means not only the presentational content but also the referent of the representations in question. ²² See Prol, 4:287: “Pure mathematics and in particular pure geometry can have objective reality only under the condition that it applies merely to objects of the senses, with respect to which, however, the principle holds that our sensible representation is in no way a representation of things in themselves, but merely of the manner in which they appear to us.” See Prol, 4:290: “After one had first spoiled all philosophical insight into the nature of sensible cognition by positing sensibility merely in a confused manner of representing, through which we cognize the things still as they are, just without having the capacity to bring everything in this representation of ours to clear consciousness; contrary to that, it has been proved by us that sensibility does not consist in this logical difference of clearness or obscurity but in the genetic one of the origin of the cognition itself, since sensible cognition does not represent things in themselves at all but merely the manner in which they affect our senses, and thus that through it merely appearances, not things themselves are given to the understanding for reflection, after this necessary correction an objection arises . . . ” ²³ See ÜE, 8:219: “One or the other: either the intuition is completely intellectual as far as the object is concerned, i.e., we intuit the things as they are in themselves, and in that case sensibility consists merely in the confusion that is inseparable from such a complex intuition; or it is not intellectual, we understand by it only the manner in which we are affected by an object that is in itself completely unknown to us, and in that case sensibility does not at all consist in confusion so that, on the contrary, its intuition may have the highest degree of distinctness, as the case may be . . . , but would still not in the least contain anything more than mere appearance.” See B60/A43: “Even if we could bring this intuition of ours to the highest degree of distinctness, we would not get any closer to the quality of things in themselves . . . ” See V-Met-L2/Pölitz, 28:584.

   

257

subtext that, to the extent to which they can be given to us, things in themselves are also given to us only through sensibility in that their existence is disclosed to us by way of our perceptions of appearances.²⁴ But the claim that appearances—or, more generally, appearances, perceptions, and sensations—indicate things in themselves and the claim that we confusedly or non-transparently perceive things in themselves are quite different. A raised flag atop Buckingham palace indicates or signifies the presence of the Queen, but by seeing the raised flag I do not see the Queen, not even non-transparently. Similarly, the opossum in my yard indicates or signifies that there is at least one thing in itself that grounds the opossum, but by seeing the opossum I do not see the thing itself that grounds it, not even non-transparently or confusedly. In sum, the empirical distinction between things in themselves and appearances is a distinction between empirical objects, which are empirically mind-independent and genuinely exist, on the one hand, and the intentional objects of perceptions or illusions, which are empirically mind-dependent and do not genuinely exist, on the other hand. Empirical objects are the referents and (partial) proximate empirical causes of perceptions and illusions. It is thus possible and natural to understand the empirical distinction as a distinction between empirical objects as they are in themselves and empirical objects as they appear to us, or as bearers of certain empirically minddependent properties. (Again, this is not how Kant himself understands it; but it is a reasonable conception.) By contrast, the transcendental distinction is a distinction between things in themselves (in the transcendental sense), which are transcendentally mind-independent and genuinely exist, on the one hand, and empirical objects, the intentional objects of experience, which also genuinely exist but are transcendentally fully mind-dependent, on the other hand. Things in themselves are (partial) proximate transcendental causes of perceptions, which feed into the construction of experience; perceptions and experience can be said to indicate the existence of things in themselves, but they do not refer to them. The transcendental distinction is thus a distinction between distinct existents that belong to different ontological kinds and share no ontological ingredients but are connected in that things in themselves (in the transcendental sense) ground empirical objects (which are appearances in the transcendental sense).

5.4 Confusion Prevention III: the Secondary Qualities Analogy Two-aspect commentators who propose to use the distinction between primary and secondary qualities as a model for the transcendental distinction are likely to protest at this point that there are other passages in which Kant himself embraces the secondary quality analogy, in marked contrast to the passages just discussed in section 5.3, and thus seems to endorse the strategy of modeling the transcendental distinction on the empirical

²⁴ In an extended sense, this also holds for human transcendental minds, even if we assume that they can be given to us through our self-consciousness in thinking, to be discussed in section 5.7. As we will see, selfconsciously thinking centrally involves reflecting on my representations and feelings as mine and, thus, depends on inner perception. This is why Kant classifies the ‘I think’ as an empirical proposition and says of it that it expresses “an indeterminate empirical intuition, i.e., perception” (B422, note).

258

       

distinction between things in themselves and appearances. The first of these (focused on by Allais) is a long passage from the Prolegomena, the second one (focused on by Rosefeldt) is to be found in a footnote at B69–70 in the Critique. That one could say, without detracting from the real existence of outer things, of a number of their predicates that they do not belong to these things in themselves but merely to their appearances and have no existence of their own outside of our representation is something that was already generally accepted and admitted long before the times of Locke but mostly after him. To these belong warmth, color, taste etc. But that I in addition to these, due to important reasons, also count the remaining qualities of bodies among the mere appearances, qualities which are called primary, [such as] extension, place and in general space with everything that adheres to it (impenetrability or materiality, shape etc.), against that one cannot present the least ground for inadmissibility; and as little as somebody who wants to admit colors, not as properties that adhere to the object in itself, but only to the sense of sight as modifications can for this reason be called an idealist, as little can my doctrine be called idealistic, only because I hold that still more, indeed, all properties that make up the intuition of a body, merely belong to its appearance; for the existence of the thing that appears is thereby not negated as in genuine idealism, but it is only shown that we cannot at all cognize it as it is in itself through the senses. (Prol, 4:289–290) The predicates of the appearance can be attributed to the object itself, in relation to our sense, e.g., the red color or the smell to the rose; but illusion [Schein] can never be attributed as predicate to the object precisely because it attributes to the object for itself what pertains to it only in relation to the senses or to the subject in general, e.g., the two handlebars that one initially attributed to Saturn. What is not to be encountered at all in the object in itself but always in its relation to the subject and is inseparable from the representation of the former [“latter” in the Academy edition] is appearance, and so the predicates of space and time are rightly attributed to the objects of the senses as such and in this is no illusion. By contrast, if I attribute redness to the rose in itself, to Saturn the handlebars, or to all outer objects extension in themselves, without looking to a certain relation of these objects to the subject and restricting my judgment to it, then first illusion arises. (B69–70, note)

I agree that in these passages Kant draws an analogy between secondary qualities and spatiotemporal properties and between their respective proper attributions to objects. But holding that there are interesting and instructive parallels between these kinds of properties and attributions is perfectly compatible with the view, expressed at B45/A29–30, B53/ A36, and B62/A45 just discussed, that the transcendental distinction between things in themselves and appearances must be carefully distinguished from the empirical distinction and does not boil down to a distinction between the properties of an empirical object that it has in itself and its appearance properties. So, let us take a closer look at the passages in question. Allais and Rosefeldt want to argue, on the strength of the secondary quality analogy expressed in the quoted passages, that for Kant appearances (in the transcendental sense) are extra-mental things as bearers of mind-dependent (Allais) or subject-relative

   

259

(Rosefeldt) properties, while things in themselves (in the transcendental sense) are extra-mental things as bearers of mind-independent or not subject-relative properties. In order to be able to make that argument, they must ascribe to Kant a conception of secondary qualities on which they are real properties of extra-mental things but also mind-dependent or subject-relative in some sense. Clearly, the conception that Kant articulates—and endorses—at B45/A29, according to which “colors, tastes etc. are rightly not regarded as properties of things but merely as modifications of our subject” does not fit this bill (my emphasis). We may call this conception of secondary qualities ‘Berkeley’s conception.’ Berkeley identifies secondary qualities with sensations, or, more carefully and more charitably put, with the qualia disclosed in sensation, for example, red qualia or smells-like-wet-dog qualia, and declares that they exist only in our mind.²⁵ Allais concedes that Kant appears to be working with Berkeley’s conception at B45/A29–30, where he explicitly disallows the secondary quality analogy, but claims that he has a different conception in mind in the Prolegomena passage, where he endorses the analogy.²⁶ Rosefeldt adopts a different strategy and argues that Kant’s underlying account of secondary qualities is the same in the Prolegomena and the Critique but that there is an ambiguity in his use of such terms as ‘red’ or ‘smelly,’ which he employs to refer both to real properties in the objects and, at other times, to sensations.²⁷ Allais and Rosefeldt develop their proposals for how to understand Kant’s conception of secondary qualities that is at work in the analogy by way of (ostensibly) rejecting Locke’s conception as a suitable model, after already having rejected Berkeley’s conception. Locke identifies secondary qualities with powers of the object to produce sensations in our mind that do not resemble any properties in the object, for example, the power of a rose to produce red sensations in us or the power in a wet dog to produce smells-like-wet-dog sensations in us. On his view, secondary qualities are ‘in the object,’ that is, they are real properties of the object, which, in turn, is assumed to exist outside of our mind.²⁸ Allais claims that if we adopt Locke’s conception and “identify the appearances of things with what the secondary qualities ‘in ²⁵ See Berkeley 1713, Works 2:180: “PHILONOUS: Tell me then once more, do you acknowledge that heat and cold, sweetness and bitterness (meaning those qualities which are perceived by the senses) do not exist without the mind? HYLAS: I see it is to no purpose to hold out, so I give up the cause as to those mentioned qualities. Though I profess it sounds oddly, to say that sugar is not sweet.” See Berkeley 1713, Works 2:187: “PHILONOUS: Well then, since you give up the point as to those sensible qualities, which are alone thought colours by all mankind beside, you may hold what you please with regard to those invisible ones of the philosophers. It is not my business to dispute about them; only I would advise you to bethink your self, whether considering the inquiry we are upon, it be prudent for you to affirm, the red and blue which we see are not real colours, but certain unknown motions and figures which no man ever did or can see, are truly so. Are not these shocking notions, and are not they subject to as many ridiculous inferences, as those you were obliged to renounce before in the case of sounds? HYLAS: I frankly own, Philonous, that it is in vain to stand out any longer. Colours, sounds, tastes, in a word, all those termed secondary qualities, have certainly no existence without the mind.” ²⁶ See Allais 2007, 466–8; Allais 2015, 127. ²⁷ See Rosefeldt 2007, 187–8. ²⁸ See Locke 1690, II.viii.15, 137: “From whence I think it easy to draw this observation, that the ideas of primary qualities of bodies are resemblances of them, and their patterns do really exist in the bodies themselves; but the ideas, produced in us by these secondary qualities, have no resemblance of them at all. There is nothing like our ideas, existing in the bodies themselves. They are in the bodies, we denominate from them, only a power to produce those sensations in us: And what is sweet, blue, or warm in idea, is but the certain bulk, figure, and motion of the insensible parts in the bodies themselves, which we call so.” See Locke 1690, II.viii.22, 140: “I hope, I shall be pardoned this little excursion into natural philosophy, it being necessary in our present enquiry to distinguish the primary and real qualities of bodies, which are always in them (viz. solidity, extension, figure, number, and motion, or rest, and are sometimes perceived by us, viz. when the bodies they are in are big enough singly to be discerned) from those secondary and imputed qualities, which are but the powers of several combinations of those primary ones, when they operate, without being distinctly discerned; whereby we may also come to know what ideas are, and what are not, resemblances of something really existing in the bodies, we denominate from them.”

260

       

truth’ are—causal powers of objects,” then “our picture is in danger of losing any kind of idealism” (Allais 2007, 467).²⁹ Her alternative proposal is to identify secondary qualities with properties of perceptually appearing to human perceivers in a certain way, properties that she calls “essentially manifest qualities” (see Allais 2015, chs 5 and 6), for example, the property of a rose of appearing to human perceivers as red. Secondary qualities thus understood are real properties of extra-mental things but they are mind-dependent because “they exist only in relation to possible . . . experience” (Allais 2007, 476), or because things have them “only in their perceptually appearing in the visual experience of subjects like us” (Allais 2015, 122).³⁰ Rosefeldt rejects Locke’s conception as a model on the grounds that secondary qualities thus understood are not subject-relative. “The motion of the molecules of an object, for example, is the cause for the . . . warmth-sensation in touching the object and, thus, would be a secondary quality in Locke’s sense. But molecular motion is not a subject-relative state” (Rosefeldt 2007, 188 9). His ‘alternative’ proposal is to identify secondary qualities with dispositions to cause certain sensations in subjects like us, for example, the disposition of the rose to cause red sensations in human subjects in appropriate circumstances.³¹ I have put ‘alternative’ in scare quotes because this just is Locke’s account. For Locke, secondary qualities are not the causal bases of dispositions, which is how Rosefeldt apparently reads him, but the dispositions. Molecular motion is the cause of the warmth-sensation but it is not a causal power; it is the primary quality in virtue of which an object has the power to cause the sensation of warmth in a human perceiver. Clearly, it is this power that Locke identifies with the secondary quality, not the causal basis. At any rate, according to Rosefeldt, if Kant holds the conception of secondary qualities that he ascribes to him, “it becomes intelligible why Kant so often makes the mistake of not distinguishing in a terminologically clear fashion talk about colors as properties of objects from talk about color sensations as mental properties” (Rosefeldt 2007, 193, my translation). Allais and Rosefeldt claim that the cited passages from Kant should be read as expressing the view that the appearance properties of empirical objects in the transcendental case are to be understood in exactly the same way as secondary qualities according to the conceptions just sketched. For example, that the rose is extended means ²⁹ Also see Allais 2015, 126, note. ³⁰ I have to confess that I find it difficult to get a precise grip on what an essentially manifest property is supposed to be without falling back on Locke’s conception of secondary qualities. Allais is concerned to emphasize that her account of secondary qualities as ways of perceptually appearing must be distinguished from a dispositional account such as Locke’s. But as long as we do not want to say, implausibly, that objects are colored only when somebody looks at them, it seems to me that the only reasonable way to cash out what it means for an object to have the property of visually appearing red to us, if this is supposed to be understood as a real property in the object, is to say, with Locke, that the object has the disposition to cause red sensations in us. Moreover, even if we grant that the property of appearing red to us is distinct from the disposition to cause red sensations in us, it is also not clear to me in what way Allais’s account introduces any more or any stronger kind of mind-dependence into the picture than Locke’s. The property of appearing red to human minds and the disposition to cause red sensations in human minds are both mind-dependent in the weak sense that they cannot be described or conceived of without reference to human minds. One could also say that both are mind-dependent in the sense that their existence depends on the possible existence of human minds and on the ability of these minds to have sensations and perceptions. To be sure, roses may not have the property of appearing red to Martians. But in that case roses will not have the disposition to cause red sensations in Martians either. Furthermore, no matter what the Martians perceive, a rose, for the Martians, has both the property of appearing red to us and the disposition to cause red sensations in us. So, I do not see how Kant turns out to be more of an idealist on Allais’s essentially manifest account than on a Lockean account. ³¹ See Rosefeldt 2007, 188–95. Rosefeldt spells this out by saying that the rose has “the property to have some property such that through the presence of this property in appropriate circumstances red-sensations are caused in subjects with normal human sensibilities” (Rosefeldt 2007, 190).

   

261

that it has the property of appearing as extended to beings like us, or that it has the disposition to cause intuitions of itself as extended in subjects whose forms of sensibility are space and time.³² Before evaluating these readings of our two passages, I briefly want to make one general observation about the Allais–Rosefeldt interpretation. Granting for the sake of the argument the basic two-aspect picture according to which appearances are things qua bearers of appearance properties and things in themselves are things qua bearers of in-itself properties, the proposed account of understanding the appearance properties as secondary qualities as just described cannot be the whole story. The account only works for those appearance properties of empirical objects that fall under the general rubric ‘ways of being sensed.’ As we just saw, Allais explicitly describes the relevant properties in this way, that is, as ‘ways of perceptually appearing.’ That is, the account only works for the intuitive determinations of empirical objects (using the terminology introduced in section 2.5.4), such as the property of being extended or of having a certain shape. The problem is that there are other properties of empirical objects that do not fall under the general rubric ‘ways of being sensed’ but which clearly count as appearance properties on Kant’s view. These properties primarily depend on the pure concepts of the understanding and do not ever perceptually appear to us, as for example, the property of changing only in response to being acted upon by an external cause or the property of having a certain force of motion. Any object has these properties only qua appearance. But it does not make much sense to say that the rose has a disposition to cause in us representations of it as changing only in response to being acted upon by an external cause, or that it has the property of perceptually appearing to us as having a certain force of motion. (The object’s motion perceptually appears to us, but not the moving force.)³³ Now, the four main questions with respect to the two passages, from the Prolegomena (4:289–290) and B69–70, on which Allais and Rosefeldt principally rest their interpretations are: (1) what conception of secondary qualities does Kant employ?; (2) what does he say about secondary qualities?; (3) how exactly is the analogy supposed to work that Kant presents?; and (4) what does he intend to show by means of the analogy? Starting with the Prolegomena passage, Kant states in no uncertain terms that secondary qualities “have no existence of their own outside of our representation” and that colors are to be seen “not as properties that adhere to the object in itself but only to the sense of sight as modifications.” These formulations make it quite clear that, pace Allais, the conception of secondary qualities in play there is exactly the same as in B45/A29–30 (and B53/A36 and B62/ A45), namely, Berkeley’s. And even if there were an ambiguity in Kant’s use of terms such as ‘red’ or ‘smelly,’ as suggested by Rosefeldt, in that sometimes he uses them to refer to ³² See Allais 2007, 476–7; Allais 2015, ch. 6; Rosefeldt 2007, 192. ³³ A related problem that primarily affects Rosefeldt’s account is that Kant, being the scientifically minded modern philosopher in the tradition of Descartes and Leibniz that he is, agrees with the analysis of the mechanical philosophy that the causal bases of the powers of empirical objects to cause certain sensations in us are to be cashed out in terms of their (traditionally so-called) primary properties. For example, the power of, say, a stone to cause the sensation of warmth in us is to be understood as based in the motion of the molecules composing it. But, on Kant’s view, that the stone is composed of molecules that are in motion is itself an appearance property of the stone, that is, a property that, if we apply Rosefeldt’s general proposal for how to understand appearance properties, must also be understood as a disposition of some kind. So, that the stone is warm means . . . what exactly? That it has a disposition to cause a representation of it in us as composed of molecules in a certain state of motion such that this disposition is the causal basis of the disposition to cause a sensation of warmth in us? How could a disposition to cause certain representations in us be the causal basis for another disposition?

262

       

sensations and sometimes he uses them to refer to causal powers or dispositions, there is no ambiguity here. Kant explicitly tells us how color terms are supposed to be understood for the purpose of his analogy, namely, as referring to modifications of the sense of sight, that is, as referring to color sensations or qualia. So, what does Kant say about secondary qualities in our passage? On my reading, when Kant claims that one could say of a number of the predicates of outer things that “they do not belong to these things in themselves but merely to their appearances,” he simply means that secondary qualities such as colors and smells, understood as sensations or qualia, can rightly be attributed, without any qualifications, to the appearances of empirical objects, that is, that they can rightly be attributed to the intentional objects of our perceptions that refer to empirical objects. (Obviously, ‘appearance’ here is to be understood in the empirical sense.) The intentional object of my perception that is (in part) empirically caused by a rose is red, as opposed to merely appearing to be so, just as the intentional object of my perception that is (in part) empirically caused by a straight stick in water is bent, as opposed to merely appearing to be so. How, then, should we understand the analogy to the transcendental case? Just as secondary qualities can rightly be attributed, without any qualifications, to the appearances of empirical objects (where ‘appearance’ is understood in the empirical sense), spatiotemporal properties can rightly be attributed, without any qualifications, to the appearances of things in themselves (where ‘appearances’ and ‘things in themselves’ are understood in the transcendental sense), that is, to empirical objects.³⁴ The opossum that is grounded in X is extended, as opposed to merely appearing to be so. Turning to the analogy’s purpose, it is clear that, in the passage under consideration, Kant does not use the secondary quality analogy in order to provide a general illustration of his transcendental distinction between appearances and things in themselves. He wants to make a specific point, namely, that his doctrine of the transcendental ideality of space and time does not commit him to the view that things in themselves (in the transcendental sense) do not exist. In the empirical case, by conceding that redness, understood as a red quale, does not belong to the rose in itself but only to its appearance (in the empirical sense), we are not calling into question that the rose exists. Similarly, in the transcendental case, by conceding that no spatiotemporal property whatsoever belongs to things in themselves and that all kinds of spatiotemporal properties only belong to their appearances (in the transcendental sense), we are not calling into question that things in themselves exist. In terms of our previous example, by conceding that only the opossum is in space and time but not X, I am not denying that X exists. That this is the proper way of understanding Kant’s secondary quality analogy in the Prolegomena is confirmed by how he leads up to it. The question at issue in the relevant section is whether his claim from the Critique that “all bodies together with the space in which they dwell must be regarded as nothing but mere representations and exist nowhere else than merely in our thoughts” is not “obvious

³⁴ Also see in this sense B55/A38: “ . . . since we are looking at the form of the intuition of this object, which must not be searched for in the object in itself but in the subject to which this object appears, but which still actually and necessarily pertains to the appearance of this object.”

   

263

idealism” (Prol, 4:288). In the paragraph right before the secondary quality analogy, Kant answers this question by saying that . . . I certainly admit that there are bodies outside us, i.e., things that, although completely unknown to us with respect to what they may be in themselves, we know through the representations that their influence on our sensibility provides for us, and which we call body; a word that thus signifies only the appearance of this unknown but nevertheless real object. Can one call this idealism? It is exactly the opposite of it. (Prol, 4:289)

That is, while it is true that bodies, on Kant’s view, are appearances, that is, ‘representations,’ or, more precisely, intentional objects of representations, he still holds that there are genuinely existing mind-independent things that ground these appearances, which goes against idealism as traditionally understood.³⁵ Also note that the quoted sentence is a clear echo of the concluding sentence of the passage from B45/A29–30—where Kant warns us not to use the distinction between primary and secondary qualities as a model for the transcendental distinction—in which he asserts that “what we call outer objects are nothing but mere representations of sensibility, whose form is space but whose true correlate, i.e., the thing itself, neither is nor can be cognized through this but about which we also never inquire in experience.” On my reading, this echoing is no surprise. The two discussions, in the relevant passages from the Prolegomena and the Critique, are cut from the same cloth, so to speak; there is no change with respect to how Kant conceives of secondary qualities, and both underline that, on Kant’s view, empirical objects, and not merely some of their properties, are mind-dependent (from the point of view of fundamental ontology). Moving on to the passage at B69–70, it must be admitted that the formulation that “the predicates of the appearance can be attributed to the object itself in relation to our sense, e.g., the red color or smell to the rose” taken in isolation does invite the reading that secondary qualities, such as colors, can be attributed to empirical objects if we understand them as dispositions to cause certain sensations in us. But, against the background of all relevant textual evidence, I do not believe that this is how the formulation was intended to be understood. To begin with, to my mind, the textual evidence overall establishes quite clearly that Kant’s go-to conception of secondary qualities is Berkeley’s and not Locke’s. That is, for Kant secondary qualities, by default, are sensations, or, more precisely, qualia that are disclosed in sensation, and not dispositions of empirical objects.³⁶ For example, in addition to the passages from the Critique (B45/A29–30, B53/A36, B62/A45) and the Prolegomena (4:289–90) already cited, where he describes secondary qualities as “modifications of the subject” and as having “no existence of their own outside of our representation,” he states earlier in the Critique that “colors are not properties of bodies to whose intuition they adhere but also only modifications of the sense of sight, which is affected by light in a certain way” (A28). Or in the Anthropology, we find him asserting that

³⁵ There is quite a bit more to say about this rather tricky passage, and I will do so in section 5.6. Impatient readers who are wondering about Kant’s somewhat confusing use of ‘body’ in this passage should consult note 57. ³⁶ But note that, in contrast to Berkeley, Kant neither takes the primary–secondary quality distinction to be spurious nor does he regard spatiotemporal properties as mere sensations or qualia.

264

       

“to someone who of the seven colors has never seen the color red, one can never make this sensation intelligible” (Anth, 7:168), where the context leaves no doubt that “this sensation” refers to “the color red.”³⁷ Rosefeldt collects a number of passages, in which, he claims, Kant “talks as a matter of course about colors as properties of extra-mental objects” (Rosefeldt 2007, 186). As I see it, none of these passages show that Kant takes colors to be properties of empirical objects rather than sensations, or qualia. (And they show even less that he regards colors as properties of extra-mental objects, if “extra-mental” is understood in the transcendental sense.) These passages are: (a) “The green color of meadows belongs to objective sensation, as perception of an object of the sense; but the pleasantness of it belongs to subjective sensation, through which no object is represented, i.e., to feeling . . . ” (KU, 5:206) (b) “ . . . if I separate from the representation of a body that which the understanding thinks of it . . . , as well as that which belongs to sensation of it, such as impenetrability, hardness, color etc., something remains for me from this empirical intuition, namely, extension and shape. These belong to pure intuition . . . ” (B35/A20 21) (c) “The analytic unity of consciousness adheres to all common concepts as such, e.g., if I think red in general, I imagine thereby a quality that (as mark) can be encountered in something, or can be combined with other representations . . . ” (B133, note) (d) “If cinnabar were sometimes red, sometimes black . . . ” (A100) (e) “By and by leave out from your empirical concept of a body everything that is empirical: the color, the hardness or softness, the heaviness, even impenetrability itself, there still remains the space that it occupied . . . ” (B5) In response, in (a) and (b) Kant explicitly characterizes secondary qualities as that which “belongs to sensation.” It is possible that he means by these expressions dispositions to cause sensations in us but the most straightforward understanding is that he simply means sensations. In (c), Kant does not say what the “something” is in which red can be encountered as a mark but one plausible reading is that he means the intentional objects of our perceptions, which are not empirical objects but appearances of empirical objects. This reading is supported by the fact that he goes on to say that red “can be combined with other representations” (my emphasis), which suggests that he regards red as a representation too, that is, as a modification of our mind, or a sensation. Regarding (d) and (e), Kant’s endorsement of Berkeley’s conception of secondary qualities surely does not commit him to relinquishing the right to participate in our normal way of talking about colors or smells. The mere fact that he uses the formulation that cinnabar is red (or that meadows are green, as in (a)) or, more generally, that empirical objects have colors, does not entail anything about his philosophical views on the matter. Everybody talks “as a matter of course” about secondary qualities as properties of empirical objects; but this does not mean that everybody also conceives of them as properties of empirical objects. Now, even though Kant himself regards secondary qualities as sensations or qualia, he certainly acknowledges that

³⁷ Also see Anth, 7:155: “But as regards the vital sense, it is not only moved . . . but also strengthened by music, as a regular play of sensations of hearing, which is thus, as it were, a language of mere sensations (without all concepts). The noises are here tones and that for hearing what colors are for sight.”

   

265

empirical objects have powers or dispositions to produce sensations in us,³⁸ and he is certainly also aware, among other things from his study of Locke, that one could identify secondary qualities with these causal powers or dispositions, instead of identifying them with the resulting sensations or qualia. So, it could be, I suppose, that, as a one-time special exception, he employed this alternative conception of secondary qualities in our footnote at B69–70, deviating from his standard practice. But it is not very likely. This unlikelihood is underscored by Kant’s final sentence of the passage in question, which says that “if I attribute redness to the rose in itself . . . without looking to a certain relation of these objects to the subject and restricting my judgment to it, then first illusion arises.” Clearly, in this instance, by “redness” he cannot mean the disposition to cause red sensations in us because this attribution would involve “looking to a certain relation of the object to the subject.” What Kant has in mind by “redness” here, obviously, is a red sensation, or a red quale. If I attribute to the rose redness understood as a red quale, then illusion arises. Given that Kant generally conceives of secondary qualities as sensations or qualia, and that he conceives of the color red in this way in his second mention of it in our footnote, I conclude that this is also how he conceives of it in his first mention of it in our footnote. Assuming, then, that Kant is relying on his standard Berkeleyan conception of secondary qualities in the relevant passage, what is he saying about secondary qualities here? Specifically, how should we understand the formulation in the first sentence of the note that “the predicates of the appearance can be attributed to the object itself, in relation to our sense, e.g., the red color or the smell to the rose,” and the formulation in the last sentence that illusion can be avoided if I attribute redness to the rose itself while “looking to the relation of this object to the subject and restricting my judgment to it”? On my reading, what Kant wants to convey is simply that the proper way to attribute a secondary quality to an empirical object such as redness to a rose is, not by saying ‘the rose is red,’ but by saying ‘the rose appears red to me’ or more generally, ‘the rose appears red to standard human observers in standard circumstances,’ or, more briefly, ‘the perceptual appearance of the rose is red.’ In this way, I am making a judgment about the rose itself but I am also “restricting my judgment” to the appearance relation in which the rose stands to me or to standard human observers in standard circumstances.³⁹ That is, on my reading, in the footnote at B69–70 Kant is basically making the same point about secondary qualities as in the Prolegomena passage that we discussed a minute ago. Accordingly, I also understand the analogy with spatiotemporal properties in basically the same way as before. Just as the right way to attribute secondary qualities to empirical objects is to say that the appearances of the empirical objects have the quality (where ‘appearance’ is understood in the empirical sense), or that empirical objects appear to have the quality, namely, by appearing as objects that have the quality, so the right way to attribute spatiotemporal properties to things in themselves (in the transcendental sense) is to say that the appearances of things in themselves have spatiotemporal properties (where ‘appearance’ is understood in the transcendental sense), or that things in themselves ³⁸ For example, in the very next paragraph after the passage from the Prolegomena just discussed, he says that he cannot connect any meaning with the claim that “the sensation of redness has similarity with the property of cinnabar that causes this sensation in me” (Prol, 4:290). ³⁹ Similarly, the proper way to attribute handlebars to Saturn is to say, not that Saturn has handlebars, but that Saturn appears to me, or, more generally, to people with a subprime telescope, to have handlebars, or that Saturn’s telescope-mediated perceptual appearance has handlebars.

266

       

appear to have spatiotemporal properties, namely, by appearing as objects that have spatiotemporal properties. This is precisely what Kant says in our footnote when he turns to discussing spatiotemporal properties and asserts that “what is not to be encountered at all in the object in itself but always in its relation to the subject and is inseparable from the representation of the former is appearance, and so the predicates of space and time are rightly attributed to the objects of the senses as such and in this is no illusion.” The predicates of space and time—which, like secondary qualities, have an irreducible dependency on the subject but, unlike secondary qualities, are essential to our representation of any proper object—are rightly attributed to the objects of the senses as such because these objects are appearances. Finally, the purpose of the analogy with secondary qualities in the passage under consideration, again, is not to provide a general illustration of the transcendental distinction between things in themselves and appearances but to make a more specific point. Kant uses the analogy to illustrate that there is an important difference between appearance and illusion, and that with his claim that all objects in space and time are appearances, on account of the fact that space and time are fully mind-dependent and conditions for the existence of these objects, he does not mean to say that we are subject to some kind of illusion about these objects.⁴⁰ In the case of secondary qualities, we would be subject to an illusion only if we were to attribute redness to a rose in itself (in the empirical sense); but saying that the redness belongs to the appearance of the rose (in the empirical sense) does not involve any kind of illusion. Similarly, in the transcendental case, we would be subject to an illusion only if we were to attribute spatiotemporal properties to things in themselves (in the transcendental sense); but saying that spatiotemporal properties belong to their appearances (in the transcendental sense), that is, to empirical objects such as the opossum, does not involve any kind of illusion. None of the similarities between secondary qualities and spatiotemporal properties that Kant points out in the two examined passages imply that he regards spatiotemporal properties as a special kind of secondary qualities, as suggested by Allais and Rosefeldt, or detract from the fundamental differences between the empirical and the transcendental distinction between things in themselves and appearances examined in section 5.3. This is why we are dealing with an analogy. In particular, none of the highlighted similarities between the two cases imply that the objects to which we attribute secondary qualities, that is, empirical objects, considered in themselves are to be identified with things in themselves (in the transcendental sense), or things that are mind-independent from the point of view of fundamental ontology. As discussed in sections 2.2 and 3.3.1, there is a sense in which empirical objects can be characterized as mind-independent, namely, as empirically mindindependent or mind-independent with respect to the empirical level of reality—in contrast to the intentional objects of perceptions, illusions, dreams, and hallucinations, which are empirically mind-dependent. But once we ‘go transcendental,’ so to speak, or

⁴⁰ This assessment is supported by how the paragraph begins to which our footnote is appended: “If I say that in space and time intuition, of outer objects as well as the self-intuition of the mind, represents both only how it [sic.] affects our senses, i.e., as it appears, this is not supposed to mean that these objects are a mere illusion” (B69). Similarly, just before the footnote, Kant reiterates that his claim is not “that bodies merely seem [‘scheinen’ as opposed to ‘erscheinen’] to be outside of me, or my soul merely seems to be given in my self-consciousness” when he asserts that “the quality of space and time, according to which I posit both, as condition of their existence, lies in my manner of intuiting and not in these objects in themselves” (B69).

 -- 

267

take up the point of view of fundamental ontology, we see that empirical objects are fully mind-dependent, like all other intentional objects. So, my overall assessment of the secondary quality analogy can be summarized as follows. The analogy is potentially misleading, namely, if we understand it as saying merely that spatiotemporal properties are similar to, or ‘just like,’ secondary qualities. For in the familiar empirical context, the empirical objects to which the secondary qualities are ascribed in our common talk are understood to be mind-independent, such as the rose which is said to be red. The danger is to mistake this empirical mind-independence for transcendental mind-independence or mind-independence from the point of view of fundamental ontology and to regard empirical objects that are things in themselves in the empirical sense as things in themselves in the transcendental sense, or as having aspects that are mind-independent in the transcendental sense. This is what Kant warns us about at B45/A29–30, B53/A36, and B62/A45. But the analogy can be helpful, namely, if it is understood as saying that the relation of secondary qualities to empirical objects is similar to the relation of spatiotemporal properties to things in themselves (in the transcendental sense). Understood that way, the analogy illustrates that the claim that spatiotemporal properties belong only to the appearances of things in themselves (in the transcendental sense) means neither that things in themselves do not exist nor that we are subject to some kind of illusion, just as the claim that secondary qualities belong only to the appearances (in the empirical sense) of empirical objects means neither that empirical objects do not exist nor that we are subject to some kind of illusion. This is what Kant tells us in the Prolegomena and in the footnote at B69. (Terminological comment: Going forward, ‘thing in itself ’ and ‘appearance’ without any qualifications should again be understood in the transcendental sense, as before; if the empirical sense is intended, I will explicitly say so.)

5.5 Confusion Prevention IV: the Grounding of Appearances in Things in Themselves Need not be a One-to-one Mapping Another potential general confusion concerning the (transcendental) appearance relation, which is also aided by our ordinary understanding of the term ‘appearance,’ that must be forestalled is thinking that this relation, as well as the corresponding grounding relation, must be one-to-one. There is no reason to suppose that every thing in itself in existence affects sensibility. But even if we restrict our attention to only those things in themselves that do affect sensibility, we have no reason to assume that for each appearance there is exactly one thing in itself that appears as this appearance, and that for each thing in itself that affects sensibility there is exactly one appearance that is grounded in it. At least, we have no reason for this assumption in the case of outer appearances, on which we will focus for now; we will return to the question of how to think about the relation between inner appearances and the things in themselves that ground them in section 5.7. Why might the relation between appearances and things in themselves that appear as these appearances not be a one-to-one mapping? There are mainly two reasons, namely, that, according to Kant’s theory of the human mind, (1) outer sense, which provides us with outer sensations on the basis of which appearances are constituted, can be affected by multiple different things, and (2) which sensations feed into the constitution of which appearances is determined, not by the things in themselves that affect outer sense and

268

       

thereby bring about these sensations, but by the particular way in which our cognitive faculties operate. According to Kant’s account of sensibility, as discussed in sections 3.4.2 and 4.2, the general scheme of order in conformity with which we intuit outer sensations in particular spatial relations is not given with the sensations but is due to the nature of sensibility. Similarly, as we saw in sections 2.4–2.8, the combination of outer sensations in unified perceptions of particular objects, as well as the combination of the representations of particular objects in a unified experience of the empirical world, is the work of the synthesis of the understanding under the guidance of its a priori forms, that is, the categories, which express the unity of consciousness.⁴¹ Accordingly, a given outer sensation could be due to the affection of outer sense by multiple different things in themselves, and different outer sensations that are due to the affection of outer sense by multiple different things in themselves could feed into the constitution of the same appearance. So, a given appearance could be grounded in multiple different things in themselves. Similarly, multiple outer sensations that are due to the affection of outer sense by the same thing in itself could feed into the constitution of different appearances. So, the same thing in itself could ground multiple different appearances. Consequently, the proper attitude within the framework of Kant’s critical philosophy with respect to the issue of how exactly the inhabitants of the realm of things in themselves and the inhabitants of the realm of outer appearances line up is agnosticism. Conversely, reading Kant as committed to a one-to-one mapping between outer appearances and things in themselves that ground outer appearances, or appear as them, would amount to ascribing to him a belief in a rather astonishing pre-established harmony. He would have to believe that providence arranged it such that, for each appearance, all outer sensations that, due to the make-up of our cognitive faculties, feed into its constitution are due to affections of sensibility by one unique thing in itself. Given the indicated mandatory agnosticism about how the inhabitants of the realm of things in themselves and the inhabitants of the realm of outer appearances line up, instead of saying things such as ‘a certain thing in itself, X, appears as an opossum,’ we should say more carefully that a certain thing in itself, X, or certain things in themselves, XYZ, appear as an opossum. Or, if we want to be extra careful, we could say in an even less committal fashion that the realm of the supersensible, the transcendental level of reality, grounds the realm of the sensible, the empirical level of reality, such that the sensible realm includes an opossum.⁴² Since the latter kinds of formulations are a bit tedious, for efficiency’s sake I will continue to use the more manageable ‘X appears as (or grounds) an opossum,’ now that we have clarified how the appearance relation and the corresponding grounding relation should be understood. The core thesis of the one-object view that every appearance is numerically identical to a thing in itself is the basis for several objections that can be raised against this version of the two-aspect interpretation. Some of them were already considered in section 3.5, including that, upon analysis, it turns out that this core thesis cannot be meaningfully stated, and that it commits the one-object commentators to a bunch of contradictions from which they can escape only on pain of not being able to account for Kant’s empirical realism or of having ⁴¹ See B103/A77; A99 100, A101 102; A105 110; A111 112; A113 114; A116 119; A123; A125; B129 130; B137–138; B159–165. ⁴² See R6358, 18:683: “The sensible as such considered in general points toward something supersensible.”

  ,  

269

to abandon the one-object view. Another objection is that the numerical identity claim entails that the relation between appearances and the things in themselves that appear as them necessarily is a one-to-one mapping. As we just learned, this implication is incompatible with Kant’s account of the constitution of appearances and the workings of our cognitive faculties, which allows that the relation in question may be one-to-many, manyto-one, or many-to-many. According to the one-object interpretation, Kant is committed to the assumption of an astounding pre-established harmony right at the center of his critical idealism.⁴³ To be sure, given that we are quite successful in representing in our experience unified objects that persist through time and obey various empirical laws, it is plausible to assume that the realm of things in themselves exhibits some kind of regularities that are reflected in our experience. As I put the point in section 2.7.1, things in themselves have to show a certain amount of cooperation in order for experience to be possible for us by providing ‘matter’ that is appropriate for our cognitive scheme and our limited cognitive firepower, so to speak.⁴⁴ But to acknowledge that our experience reflects some kind of unknown regularities in the realm of things in themselves is a long way from boldly stating that things in themselves cooperate with our cognitive machinery to such an extent that for every appearance there is a thing in itself that appears as this appearance and is numerically identical with it.

5.6 Kant versus Berkeley, Part II Given the importance of the thesis that appearances are grounded in things in themselves with respect to the particular brand of idealism that Kant is advocating, it is not surprising to find that he regards this thesis—or, more precisely, a particular version of this thesis, as ⁴³ It is important to note that this pre-established harmony objection to the one-object view, as we may call it, goes beyond the other objections just mentioned, which also target the one-object view’s core claim of the numerical identity between appearances and the things in themselves that appear as those appearances. The fact that things in themselves and appearances have different kinds of essential properties and different identity conditions by itself does not show that, according to Kant’s theory, a given appearance may be grounded in multiple different things in themselves, and a given thing in itself may appear as multiple different appearances. (For example, Ralph Walker seems to move directly from the former to the latter without any additional argumentation; see Walker 2010, 826.) The reasons why Kant’s theory allows for the possibility of one-tomany, many-to-one, and many-to-many relations between outer appearances and the things in themselves that ground them are his account of the workings of outer sense and of how sensations are operated upon in the constitution of appearances, as just explicated. As we will see in section 5.7, Kant’s theory, arguably, includes a commitment to a one-to-one correspondence between empirical selves and the transcendental selves that ground them, even though their essential properties and identity conditions are also different from one another. It might also be helpful to highlight explicitly that the pre-established harmony objection is not an epistemological worry. It is sometimes objected to the one-object view that its numerical identity thesis entails all sorts of propositions about things in themselves that we should not be able to know, according to Kant’s own epistemology, such as that there are at least as many things in themselves as there are appearances, or that things in themselves are individuated in ways that match how appearances are individuated. Personally, I am not as moved by these epistemological worries—to my mind, Kant’s claim that we cannot theoretically cognize things in themselves must be taken with a sizable grain of salt—but for now I just wish to stress that the pre-established harmony objection is of a different nature. It gets its ammunition, not from Kant’s epistemology and its implications about what we can know, but from his theory of the workings of our cognitive faculties in the constitution of appearances and its implications about what is and could be the case. ⁴⁴ And, indeed, in the Critique of Judgement it turns out that the assumption of the purposiveness of objects for our cognitive faculties is a transcendental principle of judgment without which experience would not be possible for us. See KU, 5:179 186. Also see our discussion in section 4.1.2 of the inference to the best explanation involved in Kant’s various attempts at refuting idealism by showing that what we take to be outer experience indeed is outer experience rather than mere pseudo-experience.

270

       

we will see—as also playing an important role in distinguishing critical idealism from ordinary idealism. On Kant’s analysis, for Berkeley and other ordinary idealists, experience is on a par with dreams, illusions, hallucinations, and mere imaginings for two reasons. First, as discussed in section 3.3.2, due to their failure to recognize the essential role of the forms of our cognitive faculties in the construction of experience, their theory lacks formal conditions of experience and objectivity that would allow them to distinguish experience from these other kinds of representations and thereby introduce a distinction between proper objects and mere appearances in the realm of mind-dependent objects. Second, their theory also lacks proper material conditions for the genuine existence of empirical objects understood as fully mind-dependent intentional objects. It is a bit tricky to spell out this second main difference between ordinary idealism and Kant’s critical idealism precisely. Minimally, it consists in that, in contrast to Kant, ordinary idealists are committed to the view that there are no mind-independent things that ground and in that sense appear to us as empirical objects—or, more precisely, that there are no mind-independent things apart from God that ground empirical objects—just as there are no mindindependent things that ground and appear to us as the objects of a dream. Consequently, empirical objects as conceived by ordinary idealists lack the necessary extra ontological weight, so to speak, to count as genuine existents. Idealism. In a dream the objects that could appear to us are not there; thus, the representations are not appearances but imaginings. But in our theory that which appears to us as body is real and the cause of our representations, only that these representations are mere appearances, but in a dream mere illusion. (R5652, 18:305) Bishop Berkeley in Ireland went still further, for he claimed that bodies are even impossible, since one would always contradict oneself if one assumed them. This is the dogmatic or brute Idealism that no bodies exist outside us but that appearances are nothing and lie only in our senses and our imagination. But there is also a critical or transcendental idealism if one assumes that appearances in themselves are nothing but that there still really is something unknown that grounds them. This is correct. (V-Met/Mron, 29:928)⁴⁵

To be sure, since the ordinary idealist and Kant agree that empirical objects are fully mind-dependent, there is a kind of grounding of empirical objects in finite things in themselves that they both assert, namely, their grounding in finite mind-independent thinking beings on which the objects’ determinations and being depend.⁴⁶ For the ordinary idealist, empirical objects are grounded in finite mind-independent thinking beings in that

⁴⁵ See R6316, 18:621–622: “[I say] that the ideality of space and time, which is merely formal, does not contain the real idealism that alleges that to the perception of things in space no object outside of the representation is given; rather [I say] that to this object or to these outer objects (which remains undetermined) only the same form of space does not pertain under which we intuit it or them, since it belongs merely to the subjective manner of our faculty of representation in perception . . . ” See R6311, 18:610: “One divides idealism into problematic (the one of Descartes) and dogmatic (the one of Berkeley). The latter denies the existence of all things except for the existence of whoever is making the claim, the former, by contrast, only says that one cannot demonstrate this.” Also see OP, 21:440–441. ⁴⁶ Following Kant, I will assume that a thing in itself is infinite if, and only if, it is God. Readers who are uncomfortable with this assumption are welcome to substitute ‘non-divine’ for every occurrence of ‘finite’ in the following discussion.

  ,  

271

the beings’ representations ontologically specify the objects;⁴⁷ for Kant, empirical objects are grounded in finite mind-independent thinking beings in that the beings’ cognitive faculties are among the ultimate conditions for the possibility of experience, which ontologically specifies the objects.⁴⁸ But this kind of representation-mediated grounding is quite different from the affection-mediated grounding that is at issue in Kant’s thesis that appearances are grounded in things in themselves. Only things in themselves that ground empirical objects by way of affecting sensibility can be said to appear to us as, and correspond to empirical objects. It is this second difference to ordinary idealism―that, in contrast to critical idealism, ordinary idealism lacks proper material conditions for the genuine existence of empirical objects―that Kant is getting at with his characterization of his own idealism as “formal” and of ordinary idealism as “material,” and his claim that, in contrast to ordinary idealism, his own idealism pertains only to the ‘form’ of appearances but not to their ‘matter’ and does not include any doubt or denial of the existence of things.⁴⁹ Kant obviously does not mean by this that, on his account, empirical objects exist mind-independently. Kant agrees with the ordinary idealist that empirical objects are fully mind-dependent; that is why they are both idealists. Rather, he means that, far from denying the existence of finite things in themselves that ground and in that sense correspond to empirical objects, his critical idealism requires their existence on account of his conception of the constitution of empirical objects, according to which their ‘form’ is due to our cognitive faculties but their ‘matter’ must be given to us, namely, through affections of sensibility by things in themselves.⁵⁰ And since empirical objects are thus grounded in transcendentally real things—and since they conform to Kant’s formal conditions of proper objecthood—they also genuinely exist on his account, despite being fully mind-dependent. According to ordinary idealism, by contrast, neither the ‘form’ nor the ‘matter’ of empirical objects is due to affections of our minds by finite things in themselves. This is why ordinary idealism can be called ‘material’ idealism and can be said to incorporate a denial of the existence of ⁴⁷ As briefly discussed in note 32, chapter 3, on Berkeley’s view, at least some empirical objects, at least some of the time, are ontologically specified by God’s representations. But this additional bit of complexity of Berkeley’s account does not make any difference for our present concerns, and so we will ignore it. The role of God in the grounding of empirical objects on Berkeley’s view does merit a closer look, though, which we will undertake later on in this section. ⁴⁸ Experience, which is the representation that ontologically specifies empirical objects, according to Kant, is clearly a representation of human empirical minds. Empirical objects still also depend on human transcendental minds, however, since the forms of the transcendental cognitive faculties of the latter are the formal conditions for the possibility of experience. ⁴⁹ See B519, note: “I have also sometimes called it [transcendental idealism] formal idealism in order to distinguish it from the material one, i.e., the ordinary one, which even doubts or denies the existence of outer things.” See Prol, 4:293: “For that I myself gave my theory the name of a transcendental idealism does not justify anybody to confuse it with the empirical idealism of Descartes . . . or with the mystical and fantastical one of Berkeley. . . . For this idealism, so called by me, did not concern the existence of things (but doubting it properly constitutes idealism in the received sense), for to doubt it never occurred to me, but [it concerns] merely the sensible representation of the things, to which space and time belong foremost; and of them, and thus all appearances in general, I have merely shown that they are not things (but mere kinds of representations), and also not determinations that pertain to things in themselves.” See Letter to Beck, Dec. 4, 1792, 11:395: “Mr. Eberhard’s and Garve’s opinion of the identity of Berkley’s idealism with the critical one, which I could better call the principle of ideality of space and time, does not deserve the least attention. For I speak of ideality with respect to the form of the representation, but they turn this into ideality of it [the representation] with respect to the matter, i.e., the object and its existence itself.” ⁵⁰ The important link between the merely formal nature of Kant’s idealism and his commitment to the existence of a mind-independent reality of things in themselves is also emphasized by Riehl 1876, 432–5; Riehl 1908, 570.

272

       

things, namely, in the first place, a denial of the existence of finite things in themselves that ground and appear to us as empirical objects and, accordingly, in the second place, a denial of the genuine existence of empirical objects. The reason for the addition that critical idealism and ordinary idealism differ in that the former asserts the existence of finite things in themselves that ground and in that sense correspond to and appear as empirical objects, while the latter denies it, is that many ordinary idealists can agree with the claim that there is an infinite mind-independent thing that grounds empirical objects. Take Berkeley, Kant’s own ‘favorite’ ordinary idealist. Like any typical ordinary idealist, Berkeley is committed to the view that, while there are mindindependent minds or souls, all empirical objects are fully mind-dependent. But, in contrast to ‘absolute’ idealists, he concurs with Kant that our mind does not produce its representations of empirical objects completely on its own. Berkeley holds that all our sensations and, more generally, all our ideas of empirical objects are caused in us by God right when we have them. And so, although he probably would not endorse the formulation that empirical objects are appearances of God, he can and explicitly does endorse the claim that our sensations and perceptions are due to affections of our mind by a thing in itself, namely, an infinite, omniscient, omnipotent, omnibenevolent spirit, who can thus be said to ground empirical objects in a sense very close to the sense in which empirical objects are grounded in things in themselves on Kant’s account.⁵¹ Another idealist who subscribes to a view that can be described by saying that empirical objects are grounded in God in a sense very close to Kant’s sense is (the mature) Leibniz. Leibniz believes that God works in more efficient ways than Berkeley’s God, but he also holds that all of our representations of empirical objects are ultimately caused in us by God. Once in existence, monads, that is, simple souls, require no further affection by anything external to them in order to produce the representations that ontologically specify the objects in the empirical realm. The perceptions of each monad arise solely from its own depth according to its lawof-the-series that governs and fuels the evolution of its perceptual states. But in order for this to be possible, at the moment of creation God has to ‘wind up’ every monad and supply it with its first perceptual state and with its special law-of-the-series on which all of its future mental life depends.⁵² There are some obvious ways for Kant to distinguish his overall theory from the views just described. It is an important feature of Kant’s account that the human mind is not

⁵¹ See Berkeley 1713, Works 2:215: “PHILONOUS: From all which I conclude, there is a mind which affects me every moment with all the sensible impressions I perceive. And from the variety, order, and manner of these, I conclude the Author of them to be wise, powerful, and good, beyond comprehension.” See Berkeley 1713, Works 2:220: “PHILONOUS: You acknowledge then God alone to be the cause of our ideas, and that he causes them at the presence of those occasions.” See Berkeley 1713, Works 2:250: “PHILONOUS: Look you, Hylas, when I speak of objects as existing in the mind or imprinted on the senses; I would not be understood in the gross literal sense, as when bodies are said to exist in a place, or a seal to make an impression upon wax. My meaning is only that the mind comprehends or perceives them; and that it is affected from without, or by some being distinct from itself.” ⁵² See note 76, chapter 2; Leibniz 1698, G 4:520; and Leibniz 1714b, §7, G 6:607, §§56–59, G 6:616. Note that Leibniz also recognizes a sense in which each empirical object is grounded in an aggregate of monads. On his view, every monad has a body in the sense of being associated with a body, namely, the empirical object that it perceives most distinctly. See Leibniz 1714b, §62, G 6:617. An empirical object O is grounded in the aggregate of monads M1, M2, M3 etc. in that O is or includes as one of its parts the body of M1, M2, M3 etc.. But, obviously, this kind of grounding is also a species of mere representation-based grounding and is quite different from the kind of grounding that is at issue in Kant’s thesis that appearances are grounded in things in themselves. The latter grounding depends on the affection of human minds by finite things in themselves, an affection whose possibility Leibniz categorically denies.

  ,  

273

completely passive in the constitution of the empirical world but plays an active role by contributing the ‘form’ of empirical objects. By contrast, for both Berkeley and Leibniz the ‘matter’ as well as the ‘form’ of empirical objects is given to us by God. On Berkeley’s account, the human mind functions as a mere passive receptacle for God-given ideas; on Leibniz’s account, the human mind is as active as a turnspit, to recycle one of Kant’s more memorable metaphors.⁵³ This difference would obtain even if Kant held that the ‘matter’ of empirical objects is given to us by God through affections of sensibility. But this is a difference in their accounts of how the human mind works, a difference that does not make any extra difference for the particular kind of idealism to which they subscribe―that is, any difference in addition to the first main difference identified by Kant to the effect that the ordinary idealists lack appropriate formal conditions of objectivity, precisely because of their failure to recognize that the ‘form’ of all proper objects depends on the forms of our cognitive faculties. So, in order for Kant’s characterization to come out correctly that his assertion of the existence of things in themselves that ground, correspond to, and provide the ‘matter’ for empirical objects distinguishes his critical, formal idealism from ordinary, material idealism, we must assume that the things in themselves in question are distinct from God, or, at the very least, could be distinct from God, for all we know. And since Kant knew that for many ordinary idealists the representations of the souls on which empirical objects depend have their source in God,⁵⁴ it is plausible to assume furthermore that distinctness from God, or at least possible distinctness from God, and hence finitude, is indeed built into his conception of the things ‘outside us’ that ground outer appearances on his view.⁵⁵ Indeed, there are passages that quite clearly imply that finitude is not the only additional characteristic that is built into Kant’s conception of the things ‘outside us’ that ground outer appearances on his view. The context of these passages supports the reading that, ⁵³ See KpV, 5:97: “Here one only looks to the necessity of the connection of events in a temporal series, as it develops according to the law of nature, one may call the subject in which this process occurs automaton materiale, since a machine-being is fueled by matter, or one may call it with Leibniz automaton spirituale, since it is fueled by representations; and if the freedom of our will were no other than the latter . . . , it would basically not be any better than the freedom of a turnspit, which also carries out its movements on its own once it has been wound up.” Also see V-Met/Herder, 28:103: “God according to his omniscience of the effects erected, as it were, an automaton spirituale and made such a connection of the soul that just at the moment of change of the body a conforming change arose in the soul from an inner principle. But Leibniz did not achieve his goal, since he does not decrease the supernatural effect because God must have accomplished all of this in the moment of creation; indeed, he increases it, since much more supernatural arrangement pertains to this setting up of a multitude of machines. For since the body is affected by many things, one must also make arrangements on account of that. Marionettes explain everything, either one secretly winds them up every moment or one has set them up previously (what is probably more difficult). 2) idealism is, then, the most reasonable thought, for since all representations arise from the soul’s own power, the body, merely a plaything before God, is superfluous.” ⁵⁴ See, for example, the quotation from the metaphysics lecture transcript by Herder cited in the previous note. ⁵⁵ If some ordinary idealists hold that empirical objects are grounded in God, the question arises why this grounding should not be sufficient for empirical objects to count as genuine existents. To be sure, these ordinary idealists lack the kind of formal conditions of objectivity and proper objecthood to distinguish truth from illusion and proper objects from mere appearances that Kant can appeal to thanks to his more sophisticated theory of experience. But they could insist that the grounding of empirical objects in God both gives them extra ontological weight compared to other intentional objects and distinguishes them from mere appearances in the realm of fully mind-dependent objects. Since Kant does not explicitly address this question, I can do no more than speculate about what he might say in response. Here is a suggestion. It seems plausible to say that in order for a finite thing to genuinely exist, it must either be or be grounded in an entity that exists at the transcendental level of reality of the world. God himself, however, is not part of the world but outside it. See V-Met-L1/P€olitz, 28:343; V-Phil-Th/ P€ olitz, 28:1092–1093; V-Th/Volckmann, 28: 1193–1194. Accordingly, while being grounded in God might confer a certain amount of extra reality on an intentional object, it is not enough for the object to be part of the world and count as genuinely existing.

274

       

when Kant emphasizes that, contrary to the ordinary idealists, he holds that there are finite things ‘outside us’ that ground outer appearances, ‘outside us’ is to be understood as meaning not only ‘mind-independent’ but also ‘distinct from us.’ That is, Kant seems to be committing himself to the view that the things in themselves that ground outer appearances are both finite and not human minds. Here is a prominent, noteworthy passage from the Prolegomena in which Kant describes the difference between his critical idealism and ordinary idealism in this way. Idealism consists in the claim that there are no other beings than thinking beings, the other things that we believe to be perceiving in intuition were only representations in the thinking beings to which, in fact, no object corresponds that is located outside of them. Contrary to this I say: things are given to us as objects of the senses that are located outside us, alas, of that which they may be in themselves we do not know anything but merely know their appearances, i.e., the representations that they effect in us by affecting our senses. Accordingly, I certainly admit that there are bodies outside us, i.e., things that, although completely unknown to us with respect to what they may be in themselves, we know through the representations that their influence on our sensibility provides for us, and which we call body; a word that thus signifies only the appearance of this unknown but nevertheless real object. Can one call this idealism? It is exactly the opposite of it. (Prol, 4:288–289)⁵⁶

Because of its importance for understanding the nuances of critical idealism, this passage warrants some careful unpacking. Kant begins by describing the ordinary kind of idealism to which he is opposed. According to ordinary idealism, there are only thinking beings, or minds, and the intentional objects of their representations but no things “located outside” of the thinking beings that correspond to these intentional objects. The context and, especially, Kant’s use of ‘located outside them’ instead of merely ‘outside them’ strongly suggest the reading that the ordinary idealists deny the existence of finite things corresponding to empirical objects that are both mind-independent and distinct from all thinking beings and, hence, not thinking beings. The same reading is also implied by the way in which Kant opens the passage (“idealism consists in the claim that there are no other beings than thinking beings”), which suggests that the things corresponding to empirical objects whose existence the ordinary idealists deny are not thinking beings. Having sketched the view of the ordinary idealists, Kant then proceeds to describe his own view, which is supposed to be opposed to the idealist position just presented. Here is a paraphrase of what I take Kant to be saying there. On his view, certain things that affect sensibility are given to us as, or more precisely through, empirical objects, which are ‘outside us’ in the empirical sense, that is, in the sense of being presented in experience as being in space. For the things that affect sensibility thereby bring about sensations that feed into the construction of experience and the constitution of appearances, the intentional objects of experience, which is what empirical objects are. We do not know the things that affect us as they are in themselves; we only know them through their appearances, or, more precisely, we only know that they appear to us as empirical objects. These appearances are ⁵⁶ Also see See V-Met/Mron, 29:928: “Idealism is if one imagines that there are thinking beings outside oneself but no bodies.”

  ,  

275

what we call ‘bodies.’⁵⁷ And, according to critical idealism, bodies thus understood genuinely exist. But because bodies are grounded in things that affect us, by admitting the existence of bodies that are empirically outside us, Kant explains, he is also admitting the existence of unknown things ‘outside us’ that correspond to these intentional objects of our representations. And, this, he claims, is “exactly the opposite” of ordinary idealism. So much for what Kant (more or less) explicitly says. In order for Kant’s position sketched here to be exactly opposed to ordinary idealism, as he claims, the things ‘outside us’ that ground empirical objects on his view must be the same kind of things as the things whose existence the ordinary idealist denies according to Kant’s description of ordinary idealism. Given how Kant uses ‘outside x’ in his description of ordinary idealism, in the cited passage he thus appears to be committing himself to the claim that there are finite things grounding and thus corresponding to empirical objects, and that these things are ‘outside us’ in the sense of being both mind-independent and distinct from all thinking beings. That is, he seems to be saying that the transcendental grounds of empirical objects on his view are not minds. While the interpretation of the Prolegomena passage just explicated is the most literal interpretation that is closely based on the text, there is a slightly less literal interpretation, also closely based on the text, that has the advantage of leaving us with a slightly less bold overall upshot. Typically, when Kant talks about ‘us’ he is not talking about thinking beings in general but, more specifically, about human beings. So, an alternative reading of ‘outside us’ would be to understand it as meaning ‘mind-independent and distinct from all human minds.’ As you will recall, this is also how, in section 3.3.1, we cashed out Kant’s official characterization of the transcendental sense of ‘outside us’ at A373 (which I dubbed ‘strongly ‘outside us’ in the transcendental sense’), where he says that it applies to “something that exists distinct from us as thing in itself.” If we understand ‘outside us’ in this sense, Kant’s view that there are finite things ‘outside us’ that ground and correspond to empirical objects, is not exactly the opposite of ordinary idealism understood as centrally including the claim that there are no finite mind-independent beings corresponding to empirical objects that are distinct from all thinking beings—unless human minds are the only minds there are, which, however, we have no reason to assume. But there is another way of formulating the contentious claim of ordinary idealism that both captures what I take Kant to be after and, for Kant, comes down to “exactly the opposite” of the claim that there are things ‘outside us’ in the sense just indicated that ground and correspond to empirical objects. According to this alternative formulation, ordinary idealists hold that there are no finite mind-independent beings corresponding to empirical objects that are distinct from all those thinking beings whose representations ontologically specify empirical objects, that is, whose representations are the representations in virtue of being represented in which empirical objects exist and have all of their determinations and other ontological ingredients. For most ordinary idealists, the class of

⁵⁷ Kant’s use of ‘body’ in the quoted passage is a bit confusing. When the term first occurs, it refers to the things that affect us. But in both later occurrences, it refers to the appearances that are brought about through these affections. My sense of what is going on here is that when Kant first says that he admits that there are bodies ‘outside us’ he is addressing an imagined ordinary idealist in the idealist’s terms who holds that there are no bodies outside us, by which he (the idealist) means that there are no mind-independent things that are not thinking beings. After this opening salvo, Kant then clarifies that, on his own view and in his own usage, what we call ‘bodies’ are, in fact, not mind-independent things, but appearances of mind-independent things.

276

       

finite thinking beings whose representations ontologically specify empirical objects is coextensive with the class of all finite thinking beings anyway, be it that this is because they believe that the only thinking beings are human minds and God or because they believe that empirical objects depend on all kinds of finite minds, not only human minds.⁵⁸ And, on Kant’s view, the minds on which empirical objects depend are human minds, which means that, for him, a finite mind-independent thing that is distinct from all human minds is the same as a finite mind-independent thing that is distinct from all those thinking beings whose representations ontologically specify empirical objects. So, on this slightly less literal interpretation, the overall upshot of the Prolegomena passage is that the transcendental grounds of empirical objects on Kant’s view are not human minds, which is less bold than the upshot of the more literal interpretation that they are not minds. This less bold reading also has the further advantage that it coheres better with Kant’s professed official agnosticism about whether dualism or monism is the correct view with respect to the transcendental level of reality and his apparent private predilection for monism, to be discussed briefly in section 5.7. The view that the transcendental grounds of empirical objects are not minds would commit him to asserting a form of dualism at the transcendental level of reality. But the view that the transcendental grounds of empirical objects are not human minds is compatible both with being agnostic about whether dualism is the correct account of the transcendental level of reality and with favoring monism. For the finite things in themselves that ground empirical objects could still be minds, albeit not human minds, in which case they would be “not that dissimilar” from the things in themselves that ground empirical selves, as Kant says at B428, where he describes the advantages of critical idealism when it comes to the challenge of solving the mind–body problem.

5.7 Transcendental Selves and the Grounding of Empirical Selves The textual evidence leaves no room for doubt that, on Kant’s view, just as empirical objects are grounded in things in themselves, empirical selves are grounded in things in themselves.⁵⁹ As indicated before, I will call things in themselves that ground, or appear as, empirical selves ‘transcendental selves.’ The project for this section is to learn more about Kant’s conception of transcendental selves. We already know that, like all things in themselves, they are mind-independent and, hence, non-spatial and non-temporal as well as supersensible, and that, as grounds of empirical selves, they affect inner sense and thereby bring about inner sensations. But there is quite a bit more that we can glean about them from Kant’s discussion. To begin with, it seems pretty clear that Kant is committed to a one-to-one mapping between empirical selves and the transcendental selves that ground them. Even though the general scheme of order in conformity with which inner sensations are intuited in particular temporal relations is also not given with them but is due to us, and the

⁵⁸ Unless we understand ‘mind’ in a narrow sense as equivalent to ‘rational mind,’ Leibniz is an example of the latter type of idealist. For him, empirical objects are ontologically specified by the perceptions of all monads, which include human monads but also several kinds of non-human monads. ⁵⁹ See notes 197 and 198, chapter 2.

       

277

combination of our representations and feelings in the construction of inner experience is also the work of the synthesis of the understanding, there are several important differences between the ‘inner’ and the ‘outer’ case that account for the fact that there is a one-to-one mapping between empirical selves and transcendental selves that ground empirical selves, according to Kant’s theory, while there may not be a one-to-one mapping between empirical objects and things in themselves that ground empirical objects. Kant frequently characterizes the affections of inner sense as self-affections and says of the empirical self that it is how, or as what, I appear to myself.⁶⁰ That is, he identifies the ‘owner’ of the inner sense that is being affected and the thing in itself that is doing the affecting.⁶¹ Supplemented with an additional assumption about inner sense, to which Kant appears to be committed, this identification provides an explanation of the one-to-one mapping that obtains in the case of empirical selves and their transcendental grounds. The additional assumption is that no particular inner sense can belong to more than one thing in itself, and no thing in itself can have more than one particular inner sense. Call this thesis ‘U---.’ Given that, according to Kant’s theory of inner experience, each particular inner sense supplies the ‘matter’ for exactly one empirical self, on the assumption of U---, the identification of the ‘owner’ of the inner sense and the thing in itself that affects it directly implies that, for each empirical self, there is exactly one transcendental self that grounds it, and, for each transcendental self, there is exactly one empirical self that is grounded by it. This kind of explanation is not available in the ‘outer’ case. For not only does each particular outer sense produce sensations that correspond to the ‘matter’ of multiple different empirical objects, each particular outer sense can also be affected by multiple different things in themselves, and each thing in itself that is able to affect outer sense, can affect multiple different particular outer senses, as noted in section 5.6. Ascribing to Kant the belief in a one-to-one mapping between appearances and their transcendental grounds in the ‘outer’ case amounts to saddling him with a belief in a pre-established harmony, but doing so in the ‘inner’ case does not. Kant asserts in several passages that, for all we know, the things in themselves that ground empirical objects and the things in themselves that ground empirical selves might be the same kinds of things. That is, as briefly anticipated at the end of section 5.6, Kant expresses official agnosticism with respect to the question of whether dualism or a form of monism is the correct ontological position with respect to the transcendental level of reality.⁶²

⁶⁰ See B152 153: “Here is the place to make intelligible the paradox that everybody must have noticed in the exposition of the form of inner sense (§6.), namely, how it represents even ourselves to consciousness also only how we appear to us, and not how we are in ourselves, since we intuit ourselves only how we are inwardly affected . . . ” Also recall B156: “ . . . we must also admit of inner sense that we thereby intuit ourselves only in the way in which we are inwardly affected by ourselves, i.e., that as far as inner intuition is concerned, we cognize our own subject merely as appearance but not according to what it is in itself.” See R6311, 18:611: “If there are impressions on my inner sense, this presupposes that I affect myself (although it is inexplicable how this happens) . . . ” Also see notes 197, 198, chapter 2, and note 139, chapter 3. ⁶¹ The ‘owner’ of inner sense must be a thing in itself, and thus cannot be the empirical self, because we are here talking about transcendental affections by things in themselves. Obviously, it makes no sense to say that a mindindependent thing transcendentally affects a mere appearance. We will come back to this line of reasoning below. ⁶² Again, ‘dualism’ is here to be understood as the thesis that there are two fundamental kinds of things, not that there are two things; ditto for ‘monism,’ mutatis mutandis.

278

       

But if one wanted to expand the concept of dualism, as it usually happens, and take it in the transcendental understanding, neither it nor pneumatism, on the one hand, nor materialism, on the other hand, which are opposed to it, would have the least ground, in that in this case one would fail in the determination of one’s concepts and regard the difference in the way of representing objects that remain unknown to us according to what they are in themselves as a difference of these things themselves. I, represented through inner sense in time, and objects in space outside of me are but specifically completely distinct appearances but are thereby not thought as different things. The transcendental object that grounds outer appearances and similarly the one that grounds inner intuition is neither matter nor a thinking being in itself but a ground of appearances that is unknown to us, appearances that supply the empirical concepts of the first as well as of the second kind. (A379–380)⁶³

At least in some moments, Kant seems to be leaning toward monism, though, namely, when he presents the fact that his theory allows for the possibility of monism at the transcendental level as giving him an advantage over dualists of the transcendental realist persuasion with respect to the vexed problem of explaining the possibility of interactions between the mind and the body. There is general consensus among many philosophers that if the mind and the body are understood as different kinds of things in themselves—as on Descartes’ view, for example—it seems unintelligible how they could possibly interact. But, Kant points out, if we understand the empirical mind (empirical self) and the body as appearances of unknown things in themselves, which, for all we know, may be of the same fundamental kind, the problem is, if not completely solved, at least considerably reduced in that “no other difficulty remains than the one how a community of substances is possible in general . . . ” (B427), where ‘community’ in this context is a synonym for ‘interaction.’⁶⁴ The passage from A379–380 just quoted, in particular, the formulation that the things in themselves that ground appearances are “neither matter nor a thinking being in itself,” might be taken to suggest that Kant denies that there are any thinking beings or minds at the transcendental level of reality. But a more careful reading makes clear that what he means to deny in passages like these is only that there are things at the transcendental level of reality that fall under our empirical concept of a thinking being, that is, the concept of an empirical self, which is a substance in which representations and feelings inhere as its determinations and whose existence is determined in time. So, maybe Kant does not explicitly deny the existence of transcendentally real thinking beings but, one might argue, the chapter on the paralogisms of pure reason in the Critique shows that, in conformity with his professed epistemic humility about supersensible matters, he at least wants to remain agnostic about their existence. For it is the main upshot of this chapter that all attempts of previous philosophers to demonstrate the thesis that our self is a soul are fallacious, and that asserting this thesis constitutes a transgression of the bounds of theoretical cognition and theoretically meaningful discourse.⁶⁵ But, strictly speaking, all that the paralogisms chapter shows is that Kant regards the assertion of the existence of souls at the transcendental level of reality as problematic, that is, the assertion that there are ⁶³ See A358–360. ⁶⁴ Also see A384 396; FM, 20:308. For the equation of ‘community’ with ‘interaction,’ see B256/A211. ⁶⁵ See B399–431/A341–366.

       

279

immaterial, simple, thinking substances that persist identically through time. This does not imply that he is committed to suspending judgment with respect to the existence of any kind of mind-independent minds. So, are there any reasons to think that, despite his epistemic humility, Kant holds that there are transcendentally real beings that are minds? As I see, it, the answer is yes. There are some fairly strong reasons for the conclusion that, on Kant’s view, each human being has a transcendentally real part that is a kind of mind. This transcendentally real part does not fit our empirical conception of a mind nor is it a soul, but it is a kind of mind or thinking being. And so, given that there are human beings, there also are transcendentally real beings that are minds. The first cluster of reasons for thinking that Kant is committed to the view that each human being has a transcendentally real part that is a kind of mind has to do with (what we may call) Kant’s transcendental psychology. The ontological status of the transcendental cognitive faculties that are among the protagonists of Kant’s transcendental psychology— including sensibility, which comprises inner and outer sense, the understanding, the imagination, the original synthetic unity of apperception, and reason—and the epistemological status of Kant’s account of the operation of these faculties in the Critique are among the most hairy issues in his philosophy. What seems rather clear, though, is that he understands his assertions about these faculties and their operations as justified claims about real faculties and not as mere fictions,⁶⁶ and regards these faculties, not as appearances, but as mind-independent capacities.⁶⁷ After all, these transcendental faculties are among the ultimate conditions that make experience and thus appearances possible for us in the first place.⁶⁸ That the cognitive faculties of Kant’s transcendental psychology must be mind-independent is most obvious in the case of sensibility, the faculty that is transcendentally affected by things in themselves. For, as already noted previously, surely, whatever can be transcendentally affected by transcendentally real things must be transcendentally real itself. And if sensibility is mind-independent, it stands to reason that the other transcendental faculties, which closely work together with sensibility to bring about cognitions and other kinds of representations, are mind-independent as well.⁶⁹ What is less clear is how all of this can be reconciled with Kant’s epistemic humility regarding the supersensible—which, however, is a problem for another day. Since it is plausible to assume that faculties cannot exist on their own but must be faculties of some kind of being, we can thus conclude that all of the cognitive faculties of Kant’s transcendental psychology, as mind-independent, must be faculties of things in themselves. Moreover, ⁶⁶ On a fictionalist construal, Kant’s assertions about our various cognitive faculties are to be understood as part of a mere story about the nature of the human mind that we tell ourselves to make sense of our mental lives, but whose real possibility, let alone truth, we cannot possibly ascertain in any way. Although Kant’s philosophy embodies certain fictionalist elements, there is no indication in any of his writings that he conceives of his own transcendental psychology as a mere fiction, and his tone and manner of presentation strongly speak against this assumption. ⁶⁷ For purposes of illustration, see in particular A99–100; A101–102; A106–108. ⁶⁸ In an attenuated sense, this even holds for reason, which provides regulative principles without which science and, thus, experience, understood in the technical sense explicated in chapter 2, would not be possible for us. ⁶⁹ Some readers might object that it is a mistake to ‘ontologize’ the faculties of Kant’s transcendental psychology. They are mere necessary conditions for the possibility of cognition and experience, and no ontological import is intended. In response, I have to say that, unless this objection is meant as a plug for an instrumentalist or fictionalist reading of Kant’s theory of the mind—a reading that I regard as highly implausible, see note 66—I do not understand it. Surely, if x is a necessary condition for the possibility of y, and y exists, or is actually the case, or actually takes place, then x must exist as well, or actually be the case, or actually take place.

280

       

since, according to Kant’s account, all of these faculties are in various kinds of ‘communication’ and cooperate in the generation of various kinds of representations and cognitions, it is plausible to think that particular faculties that work together in this way belong to the same mind-independent being.⁷⁰ Finally, given that all of these faculties are human faculties, we can conclude that each human being must include a transcendentally real part that is the seat of these transcendental faculties, or the subject that exercises these faculties.⁷¹ This transcendentally real part is a mind in at least the sense that it is equipped with various cognitive faculties, which, in various ways, produce representations. The second cluster of reasons for thinking that Kant is committed to the view that each human being has a transcendentally real part that is a kind of mind has to do with his theory of self-consciousness. These reasons are not as conclusive as the ones just considered, but they are worth being put on the table. Kant’s theory of self-consciousness is a central element of his account of cognition, and there is a lot to say about it.⁷² In the present context, however, we are only interested in one, very limited aspect of this theory. Kant seems to hold that by being aware of ourselves as subjects of consciousness in thinking we have cognitive access to a part of ourselves that is transcendentally real. Indeed, he sometimes even sounds as if he believes that in self-consciously thinking we are aware that we have a supersensible part that belongs to the transcendental level of reality. Only the human being, who knows all of nature merely through the senses, cognizes himself also through mere apperception, namely, in actions and inner determinations that he cannot at all count among the impression of the senses, and is for himself, to be sure, in one part phenomenon but in another part a merely intelligible object, namely, with respect to certain faculties, since its actions cannot at all be counted among the receptivity of sensibility. (B574–575/A546–547)

⁷⁰ My sensibility and your understanding cannot produce any cognitions together. ⁷¹ Note that the requirement that all of the cognitive faculties featured in Kant’s transcendental psychology are faculties of the same mind-independent being is compatible with the possibility that this being is a complex entity that includes multiple constitutive parts, and different faculties are associated with different such parts. There are certain passages that might be taken to suggest that Kant distinguishes in the human being a mind-independent part that has understanding and reason and a mind-independent part that has sensibility. For example, in the Groundwork, he says that “the human being, in fact, finds a faculty in himself through which he is distinguished from all other things, indeed from himself insofar as he is affected by objects, and this faculty is reason” (GMS, 4:452). (Also see B574–575/A546–547, which is quoted in the following paragraph in the main text.) To be sure, as a subject with understanding and reason, the human being is also distinguished from his body, which is affected by objects and which belongs to the empirical realm. And in the kind of context in which remarks like the one just quoted tend to occur, namely, in contexts where Kant is arguing for a distinction between a supersensible and a sensible part of the human being, he might well primarily have this latter contrast in mind. But, then again, as an empirical object, the body is affected only by other empirical objects, but human beings are also affected by things in themselves. So, there must be a mind-independent part of the human being that is equipped with sensibility. And if the part that is equipped with understanding and reason is distinct from all parts by means of which the human being is affected by objects, it follows that there are two mind-independent parts, one that is equipped with understanding and reason, and one that is equipped with sensibility. The existence of two distinct constitutive parts of the human being at the transcendental level of reality, one spontaneous and active, the other one receptive and passive, would nicely mirror the existence of two distinct constitutive parts of the human being at the empirical level of reality, the empirical self and the body, the first of which is an active mind, and the second one of which is a passive body. Probably needless to add, I am not suggesting that this story is part of Kant’s official account of the transcendentally real side of the human being. The story is merely an illustration of how Kant may have thought about these matters. ⁷² Helpful extended discussions can be found in Kitcher 2011 and Longuenesse 2017, part II.

       

281

The human being who considers himself in such a manner as intelligence places himself thereby in a different order of things and in a relation to determining causes of a completely different kind if he thinks of himself as intelligence gifted with a will and, thus, causality, than if he perceives himself as a phenomenon in the world of sense (which he really is as well) and subjects his causality to outer determination according to laws of nature. . . . [T]hat he must imagine and think himself in this double fashion is grounded, as far as the first is concerned, in the consciousness of himself as an object that is affected through senses, as far as the second is concerned, in the consciousness of himself as an intelligence, i.e., as independent from sensible impressions in the use of reason (thus as belonging to the world of the understanding). (GMS, 4:457)⁷³

The question of how exactly to think about the cognitive access to ourselves that is afforded to us by our self-consciousness in thinking is tricky and touches on areas of Kant’s theory that, I suspect, were not completely clear in his own mind when he set out to write it down and remained somewhat fluid throughout the course of the critical period. More specifically, Kant seems to have changed his mind with respect to the question of whether, on the basis of our self-consciousness in thinking, we can indeed cognize ourselves as transcendentally real thinking beings, that is, be aware that we have a transcendentally real part and justifiedly ascribe to it the determination of thinking, as suggested in the first of the passages just quoted. A more modest, if also slightly more complicated, picture emerges in several newly added passages in the second edition of the Critique, including, in particular, the paralogisms chapter and Kant’s analysis there of the ontological import of the thought ‘I think.’ In these passages, we find a distinction between being conscious of ourselves as subjects and determining ourselves as objects through the ascription of certain determinations to ourselves, and the reminder that thoughts, as representations in inner sense, which occur in time, can be determinations only of our empirical self. According to Kant, ‘I think’ is an empirical proposition in the sense that my recognition that I think depends on the presence of some representations in inner sense that I am reflectively aware of as mine. Since time is the form of inner sense, all representations in inner sense are transcendentally ideal. So, if I ascribe those representations to myself as my determinations or, more generally, if, on the basis of reflecting on these representations as mine, I attribute to myself the determination of thinking, the object that I am determining in this way is the empirical self.⁷⁴ Nevertheless, ‘I think’ does not only disclose the existence of the empirical self, the object that is determined through the determination of thinking, but also of the

⁷³ Also see V-Met/Mron, 29:832: “One could say: we could indeed cognize ourselves? Yes, where we are selfactive [selbsttätig], one can say this does not belong to sensibility, but if we observe ourselves this is nothing but a series of inner appearances.” See V-Met/Mron, 29:882: “If the soul is conscious of itself without being conscious of its state, it is apperception. If it is also conscious of its state, it is sensation or perception.” Also see GMS, 4:451–452; KpV, 5:97–98; B430–431. ⁷⁴ See B429–430: “But the proposition, I think, insofar it says as much as: I exist as thinking, is not merely [a] logical function but determines the subject (which then is at the same time object) with respect to existence and cannot take place without inner sense whose intuition always provides the object not as thing in itself but merely as appearance. Thus, in it is no longer mere spontaneity of thinking but also receptivity of intuition, i.e., the thinking of myself is applied to the empirical intuition of the very same subject.” See Anth, 7:142: “I, as thinking being, am with me, as sensible being, one and the same subject; but as object of inner empirical intuition, i.e., insofar I am inwardly affected by sensations in time, . . . I cognize myself only how I appear to myself, not as thing in itself.” Also see B133–136; B428; FM, 20:270–271.

282

       

subject who acts in thinking ‘I think.’⁷⁵ But, in this context, Kant insists that my consciousness of myself as the subject of ‘I think’ does not amount to any kind of cognition of myself. Cognition requires determination, that is, the ascription of a determination, to the thing that is to be cognized. But, according to Kant, merely being conscious of myself in thinking does not reveal any determinations that I could ascribe to myself. In his words, “in the consciousness of myself in mere thinking I am the being itself, of which, however, nothing is given to me for thinking, of course” (B429). Or, a bit earlier in the text, he puts the same point by saying that the “indeterminate perception” that is expressed in ‘I think’ “signifies here only something real that has been given, namely, for thinking in general, thus not as appearance, also not as thing in itself (noumenon) but as something that indeed exists and is signified as such in the proposition: I think” (B422–423, note).⁷⁶ So, the thought ‘I think’ attests to the existence of the subject of consciousness in self-consciously thinking but this subject remains undetermined and thus uncognized. On first glance, it is not entirely clear how Kant conceives of this subject of consciousness. Basically, there are two options. First, he could be thinking of the transcendental subject understood as an ontological ingredient of the empirical self that is contributed by the understanding in the construction of inner experience and functions as the bare skeleton of a particular = X that we ‘dress’ up by attributing representations and feelings to it as its determinations.⁷⁷ Second, he could be thinking of the transcendental subject understood as a supersensible, mind-independent subject = X, that is, an “intelligible” subject that “belongs to the world of the understanding” as he puts it in the passages cited above.⁷⁸ To my mind, the second reading is much more plausible than the first. Given that my self-consciousness in thinking attests to the existence of my empirical self, it trivially

⁷⁵ See B423, note: “For it is to be noted that if I have called the proposition I think an empirical proposition, I do not want to say by that that the I in this proposition is an empirical representation; rather, it is purely intellectual since it belongs to thinking in general. However, without any empirical representation that provides the material for thinking, the actus, I think, still would not take place, and the empirical is only the condition of the application or of the use of the purely intellectual faculty.” ⁷⁶ Also see B157–158: “On the other hand, I am conscious of myself in the transcendental synthesis of the manifold of representations in general, and, thus, in the synthetic original unity of apperception, not how I appear to myself, nor how I am in myself, but only that I am. This representation is a thinking, not an intuition. Since in order to cognize ourselves, apart from the action of thinking that brings the manifold of any possible intuition to the unity of apperception, a determinate kind of intuition is required through which this manifold is given, it follows that my own existence is not appearance (even less so mere illusion) but the determination of my existence can only happen in conformity with the form of inner sense according to the special manner in which the manifold that I combine is given in inner intuition, and I thus have no cognition of me as I am but merely as I appear to myself.” See FM, 20:270: “I am conscious of myself is a thought that already contains a twofold I, the I as subject and the I as object. How it is possible that I, who thinks, can be an object (of intuition) for myself and in this way distinguish me from myself is absolutely impossible to explain, although it is an undoubted fact. . . . Of the I in the first sense (the subject of apperception), the logical I, as representation a priori, it is impossible to cognize anything further at all, what kind of being it is, or of what kind of nature it is; it is, as it were, like the substantial that remains if I have left out all accidents that inhere in it, but which cannot be cognized further at all, since the accidents were exactly that based on which I could cognize its nature. But the I in the second sense (as subject of perception), the psychological I, as empirical consciousness, is capable of manifold cognition . . . ” Also see 406 407; A402; B157, note. ⁷⁷ The transcendental subject thus understood is the ‘inner’ counterpart to the transcendental object understood as an ontological ingredient of outer appearances. See note 212, chapter 2. ⁷⁸ This kind of ambiguity between the conception of the subject of consciousness in thinking as a transcendental subject understood as an ontological ingredient of an empirical self and a transcendental subject understood as supersensible subject is illustrated by the following passage at B404/A345–346: “ . . . the simple representation I that, for itself, is completely empty of any content, of which one cannot even say that it is a concept, but a mere consciousness that accompanies all concepts. Through this I or he or it (the thing), which thinks nothing more is represented than a transcendental subject =x . . . ” See note 198, chapter 2.

       

283

follows that the existence of the transcendental subject understood as an ontological ingredient of the empirical self is thereby established as well. But if that were all, one would legitimately wonder what all the fuss is about. Also, in the course of the construction of inner experience the transcendental subject thus understood is determined through the ascription of representations and feelings to it as its determinations. Moreover, Kant’s formulations that in the consciousness of myself in thinking I am “the being itself [das Wesen selbst]” and that the indeterminate perception that is expressed in the ‘I think’ “signifies something real that has been given” make it very plausible to conclude that the subject to whose existence our self-consciousness in thinking attests is a mindindependent, supersensible thing. (Perceptions depend on supersensible, transcendental grounds.)⁷⁹ Last but not least, in his discussion in the paralogisms chapter, Kant anticipates that, in his practical philosophy, he will be able to determine the subject whose existence is revealed in the ‘I think’ and which could not be determined in the context of the theoretical philosophy, namely, by ascribing to it the determination of being autonomous or free.⁸⁰ And the subject that the practical philosophy determines to be free clearly is supersensible, as Kant explicitly states at many places.⁸¹ Accordingly, I conclude that, on Kant’s view, in self-consciously thinking I cognitively relate to myself as both an empirical self and a supersensible, mind-independent subject. Thus, ‘I’ in ‘I think’ has both an empirical and a transcendental sense and is a referring expression in both senses. ‘I’ in ‘I think’ in its empirical sense is intentionally related to a genuine existent that exists at the empirical level of reality, namely, the empirical self, and ‘I’ in ‘I think’ in its transcendental sense is intentionally related to a genuine existent that exists at the transcendental level of reality, namely, the supersensible subject of consciousness in thinking. This account of how to think about our cognitive access to ourselves in our selfconsciousness in thinking is more cautious and more subtle than the account suggested in the passages from B574–575/A546–547 and GMS, 4:457, quoted above, but it still allows us to conclude that Kant is committed to the view that each one of us has a transcendentally real part that is a mind of some kind. To be sure, the more cautious account no longer suggests that our self-consciousness in thinking discloses that we have such a transcendentally real part. Moreover, given what we just learned, the claim that we have a transcendentally real part that is a mind of some kind cannot be understood to mean that this part has the determination of thinking, or that thoughts inhere in it as its determinations. The only object to which we can justifiedly ascribe these kinds of determinations is the empirical self. Nevertheless, Kant’s account implies that each one of us has ⁷⁹ Kant’s formulation in the note at B423 that the indeterminate perception that is expressed in ‘I think’ does not signify anything that has been given “as a thing in itself (noumenon)” need not to be understood as saying that the subject of consciousness whose existence is disclosed in the thought ‘I think’ is not a thing in itself. What Kant means to say, I take it, is that this subject of consciousness is not a thing in itself understood as a noumenon, that is, an object that can be cognized by means of the pure understanding alone. The distinction between things in themselves and noumena will be explained in more detail in section 6.4. ⁸⁰ See B430 431. Also see FM, 20:292: “Freedom of the will is this supersensible which, through the moral law, is not only given as real in the subject, but, in practical respect, is also determining with respect to the object, which in theoretical respect would not be cognizable at all, which then is the proper final end of metaphysics.” See KpV, 5:105; B430–431. ⁸¹ See Kpv, 5:47: “The moral law is indeed a law of causality through freedom and thus of the possibility of a supersensible nature, as the metaphysical law of the events in the world of sense was a law of causality in sensible nature . . . ” That transcendental idealism allows freedom to be attributed to supersensible entities that are not part of the empirical realm, in which everything is governed by universal deterministic laws, is Kant’s solution to the third antinomy; see B560–586/A532–A558. Also see the discussion at the end of section 5.8.

284

       

a transcendentally real part to which we are cognitively related in self-consciously thinking, and that this part is a subject of consciousness in thinking, or a subject of spontaneity, as we might also say, and in that sense a mind.⁸² So, on the basis of Kant’s transcendental psychology, we can conclude that he is committed to the view that each one of us has a transcendentally real part that is equipped with various transcendental cognitive faculties, and, on the basis of his theory of selfconsciousness, we can conclude that he is committed to the view that each one of us has a transcendentally real part that is a subject of consciousness in self-consciously thinking. It is eminently plausible to identify these parts. The supersensible subject of consciousness in self-consciously thinking to which the ‘I’ in ‘I think’ refers just is the supersensible subject of our cognitive faculties. Going forward, I will call this transcendentally real part of the human being the ‘(human) transcendental mind.’⁸³ Given that Kant’s account of the transcendental affection of inner sense as a kind of self-affection implies that the thing that is doing the affecting, that is, the transcendental self, is identical with the thing that is being affected, that is, the thing that is equipped with sensibility, we can also conclude that transcendental selves are transcendental minds.⁸⁴ With respect to Kant’s professed agnosticism about whether a form of monism or a form of dualism is the proper account of the transcendental level of reality, this means that, while the possibility that, at the transcendental level, there are minds and non-minds, and the possibility that there are only minds, are both still in the running, the possibility that the only transcendentally real beings are non-minds is ruled out. We know that there are human transcendental minds.⁸⁵ It should also be added that the claim that we cannot ascribe thinking as a determination to transcendental minds does not imply that they do not have any intellectual representations. We cannot ascribe thoughts to them because, as Kant puts it, “thoughts, as factual determinations of the faculty of representation, also belong to the empirical representation of our state” (FM, 20:270), and, thus, can only ever be determinations of the empirical self, ⁸² One might object that, just as it is a mistake to ‘ontologize’ cognitive faculties, so it is a mistake to ‘ontologize’ subjects of consciousness in thinking. In response, it is Kant who insists that the ‘I think’ attests to the existence of the subject and refers to the being itself. The ‘I’ in ‘I think’ is a genuinely referring expression for Kant, even if it is understood in the transcendental sense. One might also worry that being a subject of consciousness is a determination, which means that Kant’s theory of self-consciousness (if justified) includes a modest cognition of a supersensible thing. This worry is closely related to the worry raised above that Kant’s transcendental psychology seems to be in tension with his epistemic humility about supersensible matters. Since we are not presently concerned with Kant’s epistemology, we will not worry about these worries here. I profusely worry about them in Jauernig (in preparation). ⁸³ It is worth acknowledging that (a) the sketched reading of Kant’s account of the ontological import of the thought ‘I think’ can be challenged, (b) Kant himself may not have achieved a completely settled view on this matter, and (c) one can imagine various modifications that Kant may have wanted to make with respect to this account in order to alleviate tensions with his theory of cognition, including, completely giving up on the idea that in our self-consciousness in thinking we are cognitively related to something transcendentally real. But Kant’s transcendental psychology is a non-negotiable element of his critical theoretical philosophy. So, even if we were to decide that Kant’s theory of self-consciousness does not warrant ascribing the view to him that there are human transcendental minds, we still have conclusive reasons to read him as firmly committed to this view. ⁸⁴ That Kant indeed identifies the supersensible, mind-independent being that ‘I’ in ‘I think’ refers to with the transcendental self, that is, the thing in itself that grounds the empirical self, is illustrated by the following passage, see Prol, 4:337: “ . . . the I in the proposition: I am, does not only signify the object of inner intuition (in time), but the subject of consciousness, just as body does not only signify the outer intuition (in space) but also the thing in itself that grounds this appearance . . . ” This passage implies that the subject of consciousness to which ‘I’ refers is the thing in itself that grounds the object of inner intuition. (In the passage, Kant talks about ‘I am’ and not ‘I think,’ but he often switches between them in these kinds of context; the same point holds for ‘I exist’ as well.) ⁸⁵ In section 5.10.2, we will take a closer at look at various ways in which the claim that there are human transcendental minds can be justified within Kant’s theoretical philosophy.

       

285

or the empirical mind. But precisely because thoughts belong to the empirical self, it is not implausible to think that the transcendental self—which is a transcendental mind—as the ground of the empirical self, has unconscious intellectual representations that, through self-affections of its own inner sense, appear to us as these thoughts. That the conscious thoughts of the empirical self in inner sense are appearances of unconscious intellectual representations of the transcendental self that grounds it seems to be what Kant has in mind when, in the context of contemplating the possibility that matter might be grounded in transcendental selves, he says that he can well assume that “in the substance in itself, to which extension belongs with respect to our outer sense, thoughts inhere that can be represented with consciousness through its own inner sense” (A359).⁸⁶ This reading also fits well with the fact that, as noted in section 2.9, Kant often refers to the representations in inner sense themselves as “inner appearances.”⁸⁷

5.8 The Human Being and its Multiple Distinct Parts, or Why the Empirical and the Transcendental Self are not Numerically Identical As just discussed, Kant is committed to the view that there is a one-to-one mapping between empirical selves and transcendental selves that ground empirical selves. Each empirical self is necessarily grounded in exactly one transcendental self, and each transcendental self necessarily grounds exactly one empirical self. This is a noteworthy modal difference between the relation of empirical selves and their transcendental grounds, on the one hand, and the relation of empirical objects and their transcendental grounds, on the other hand. Each empirical object could be grounded in more than one thing in itself, and each thing in itself could ground more than one empirical object. Despite this difference, there is also the important similarity between the two cases that, just as empirical objects are numerically distinct from the things in themselves that ground them, empirical selves are numerically distinct from the transcendental selves that ground them, with no ontological overlap between them. The project for this section is to argue for this two-world reading of the distinction between the empirical and the transcendental self by defending it against two central objections from proponents of the two-aspect view and by offering some additional considerations in its favor, that is, considerations in addition to the ones already surveyed in section 3.6, where we learned that the empirical self is fully minddependent, while the transcendental self is mind-independent, from which it directly follows that they are numerically distinct and do not share any ontological ingredients. To begin, it is worth emphasizing that Kant nowhere indicates in any way that the transcendental distinction between inner appearances and things in themselves is to be understood differently from the transcendental distinction between outer appearances and things in themselves. On the contrary, as we saw in section 3.6, what he says about inner appearances suggests that he takes the two cases to be exactly parallel. As we just reminded ourselves, there is the modal difference with respect to the relevant grounding relation that each empirical self is necessarily grounded in one unique transcendental self, while empirical objects could each be grounded in more than one thing in itself, and things in

⁸⁶ Also see FM, 270–271.

⁸⁷ See note 196, chapter 2.

286

       

themselves could each ground more than one empirical object. But this difference neither affects the nature of the grounding relation nor does it imply that, in contrast to outer appearances, inner appearances are numerically identical with their transcendental grounds. Consequently, all else being equal, readings on which the transcendental distinction between inner appearances and the things that ground them is understood in the same way as the transcendental distinction between outer appearances and the things that ground them are preferable to any kind of ‘mixed’ readings, on which these transcendental distinctions are understood differently. Moreover, given that there is much more systematic and textual evidence that speaks to the question of how to understand the transcendental distinction in the ‘outer’ case than there is for the ‘inner’ case, all else being equal, readings that are designed primarily with the ‘inner’ case in mind are less reliable than readings that primarily focus on the ‘outer’ case or take all of the available evidence into account. All of this means that, given the strong case for the two-world reading of the transcendental distinction between outer appearances and things in themselves developed in the preceding chapters, there is a strong initial presumption that the two-world reading is also the correct reading of the transcendental distinction between inner appearances and things in themselves. The main systematic objection against regarding the empirical self and the transcendental self as numerically distinct is that it seems obvious that each one of us is only one and not two, which might be taken to suggest that, in the case of the self, the two-aspect reading clearly wins the day.⁸⁸ Another central objection, based on the text, is that Kant is especially prone to using two-aspect language when he talks about the relation between the empirical and the transcendental self, which he does most frequently in discussions of the question of how there could be room for freedom and moral responsibility, given that all events in the empirical realm, including our actions, are causally determined by other empirical events. Here are some representative passages of this kind: [T]hus, it [a rational being] has two points of view from which it can consider itself and cognize laws for the use of its powers and thus for all its actions, first, insofar as it belongs to the world of sense, under natural laws (heteronomy), second, as belonging to the intelligible world, under laws that are independent of nature, not empirical, but grounded merely in reason. (GMS, 4:452) If one can think such an influence of the beings of understanding on appearances without contradiction, while natural necessity pertains to all connection of cause and effect in the world of sense, freedom can be conceded to the cause that is itself not an appearance (although it grounds it), and thus nature and freedom can be attributed without contradiction to the same thing but in a different respect, on the one hand as appearance, on the other hand as thing in itself. (Prol, 4:344)⁸⁹ ⁸⁸ Robert Adams, although not a two-aspect theorist in general, claims that “when practical reason is allowed into the discussion it is clear that Kant postulates both noumena that definitely are identical with some phenomena and noumena that definitely are not. We ourselves are the prime example of the former” (Adams 1997, 822). Erich Adickes makes a similar observation: “In the case of the I it basically goes without saying that one and the same being is at the same time thing in itself and appearance” (Adickes 1924, 27). ⁸⁹ See KpV, 5:95; “ . . . no path remains than to attribute the existence of a thing insofar it is determinable in time, accordingly also the causality according to the law of natural necessity, only to the appearance, but freedom to the very same being as a thing in itself.” See GMS, 4:457: “For that a thing in the appearance (which belonging to the world of sense) is subject to certain laws, of which the very same as thing or being in itself is independent, does not contain the slightest contradiction . . . ” See B566–569/A538–541; Prol, 4:345; KpV, 5:114; V-Met/Mron, 29:924.

       

287

Prima facie, these passages might strike the reader as expressing a two-aspect conception of the self. I agree that each one of us is only one and not two, and that two-aspect language naturally lends itself to talking about the transcendental self and the empirical self on Kant’s account. But I do not agree that this is sufficient to establish that Kant is committed to a two-aspect conception of the relation between them. The claim that each one of us is only one and not two is perfectly compatible with conceiving of this relation in two-world fashion, as is the use of two-aspect formulations to characterize these entities, including Kant’s specific formulations in the passages just quoted. How so? With respect to the oneness of each one of us, I would like to start with a comment about terminology. One might hold that it is built into how we normally use the term ‘self ’ that it is nonsense to talk about one’s multiple selves, just as it is built into how we use the term ‘I’ that none of us can refer to more than one thing with it. On this view, the victory of the two-aspect view in the present dispute might seem like a forgone conclusion. In response, first, I am not convinced that we ordinarily use these terms exclusively in the indicated way. For example, I take it that it is not hard to understand what I mean when I talk about my professional and my private self or my charming and my grumpy self. Similarly, it would be intelligible (and true) if I were to respond to both the question ‘Who is taller than 5 feet?’ and the question ‘Who, according to Kant’s practical philosophy, is immortal?’ with the answer ‘I am,’ even though in the first case I would be talking about my body and in second case about my soul. But even if we normally did use ‘self ’ and ‘I’ in the more restricted ways suggested above, this hardly would decide the issue under dispute. The terms ‘transcendental self ’ and ‘empirical self ’ are technical terms, introduced by me no less (although inspired by Kant’s text), which, according to my stipulation, refer to (a) the thing in itself that affects inner sense and appears as an inner appearance and (b) the inner appearance that is constituted on the basis of inner sensations, which are brought about by these affections. The substantive question at issue is how exactly the relation between the thing described in (a) and the thing described in (b) is to be understood, which, obviously, cannot be decided by meditating on the meaning of the word ‘self ’ in ordinary English. And, just to add, since I agree that each one of us is only one and not two, I am perfectly happy to reserve the special term ‘Self,’ with a capital ‘S,’ for the one entity that each one of us is. So, then, how can one hold that the empirical self and the transcendental self that grounds it are numerically distinct while at the same time agreeing that each one of us is only one and not two? I propose to answer this question by appealing to Kant’s conception of human beings, understood as an ontological, as opposed to biological, kind. It is uncontroversial in general that, although each human being is one, all human beings are composed of distinct parts, at least during their embodied time on earth. Even a transcendental realist who is also a physicalist would agree with this claim. For example, my hand and my ear are distinct parts of my body. A transcendental realist who is a dualist, such as Descartes, recognizes not only a multiplicity of parts in each human being but also ontologically different kinds of parts, namely, a body and its parts, all of which belong to the ontological kind ‘extended thing,’ and a soul, which belongs to the ontological kind ‘thinking thing.’ For ease of communication, we will say that a part P of an entity E is ontologically basic with respect to ontological kind K if, and only if, (1) P is of kind K, (2) either P does not contain any parts or P contains only parts that are of kind K, and (3)

288

       

P does not compose any larger part of E with other parts of E that are of kind K. More intuitively and sloppily put, an ontologically basic part of an entity is the ‘largest’ part of the entity of a particular ontological kind. Using this terminology, on Descartes’ view, each human being has two ontologically basic parts, a soul and a body. The former is ontological basic with respect to the kind ‘thinking thing,’ and the latter is ontologically basic with respect to the kind ‘extended thing.’ On Kant’s view, each human being has two primary ontologically basic parts, namely, a sensible part and a supersensible part,⁹⁰ the former of which contains two further, more specific ontologically basic parts,⁹¹ namely, an empirical self and a body,⁹² and the latter of which either is a transcendental self, that is, the transcendental ground of its empirical self, or is composed of two further, more specific ontologically basic parts, namely, a transcendental self and the transcendental ground of its body, depending on whether the latter ground is indeed part of the human being or not.⁹³ Call this Kantian conception of human beings the ‘multiple-parts view.’⁹⁴ Given the multiple-parts view, we can easily answer the question raised at the beginning of this ⁹⁰ Recall B574–575/A546–547: “Only the human being, who knows all of nature merely through the senses, cognizes himself also through mere apperception, namely, in actions and inner determinations that he cannot at all count among the impression of the senses, and is for himself, to be sure, in one part phenomenon, but in another part a merely intelligible object, namely, with respect to certain faculties, since its actions cannot at all be counted among the receptivity of sensibility” (my emphasis). In a metaphysics lecture, after pointing out the “noteworthy phenomenon in the case of the human being that in it freedom is unified with natural necessity,” in that “as natural being (phenomenon) every new action can be explained as determined by laws of natural necessity,” while “as intelligence (rational being)” the human being, “determining himself, independently of all laws of nature, takes from himself the ground to refrain from performing his action that he can or ought to do, which he would not be capable of according to the laws of nature,” Kant says that “we are forced to assume a union of both seemingly completely contradictory beings in him in his selfhood in virtue of which he is the very person who is affected by the lower forces, but determines himself through independence of reason” (V-Met-K3/Arnoldt, 29:1019–1020). ⁹¹ The reason that an ontological basic part can contain parts that are themselves ontologically basic is due to the fact that there are different, more or less general ontological kinds. Of course, given how we just defined what it means to be an ontologically basic part, no part that is ontologically basic with respect to kind K1 can contain a (proper) part that is ontologically basic with respect to kind K1. But it is possible for a part that is ontologically basic with respect to kind K1 to contain a part that is ontologically basic with respect to kind K2. For example, a part that is ontologically basic with respect to the kind ‘sensible thing’ can contain a (proper) part that is ontologically basic with respect to the kind ‘extended thing.’ ⁹² That Kant considers human beings to have bodies, and, more specifically, that he considers his body, like his empirical self, as a sensible ontologically basic part of himself (or of his Self), qua human being, almost goes without saying. If illustration is needed, it can be found in following passages. See Anth, 7:286: “ . . . the temperaments that we attribute to merely the soul probably also secretly have the corporeal in the human being for a contributing cause.” See MS, 6:423: “To deprive oneself of an integrating part as organ (mutilate), e.g., to donate or sell a tooth to plant it in the jaw of somebody else, or to allow castration to be done to oneself in order to live more comfortably as a singer, etc. belong to partial suicide.” See V-Met/Mron, 29:876–877: “The soul is merely our I, not the body, but body and soul together as human consciousness are also called I. In empirical psychology, we consider our I as soul and as human being.” See V-Mo/Collins, 27:369: “Now the body is the complete condition of life such that we have no other concept of our life than by means of our body, and since the use of our freedom is possible only through the body, we see that the body is a part of our self.” See V-Mo/Collins, 27:387: “The moral reason [why prostitution is wrong] thus is that the human being is not his property and cannot do with his body what he wants; for since the body belongs to his self, it constitutes with it a person; however, he cannot turn his person into a thing . . . ” ⁹³ The question of whether the transcendental grounds of the bodies of human beings are also among their parts goes beyond the limits of our cognition. Note that if they were among their parts, then, for all we know, there may also be partially overlapping human beings who share some of the things in themselves that ground their bodies. ⁹⁴ It is useful to emphasize that Kant is not the first philosopher to conceive of human beings as complex entities that span different levels of reality. As mentioned before (see note 30, chapter 2), Leibniz also holds that reality comprises various ontological levels. On Leibniz’s view, human beings and, more generally, all animals are composed of at least one monad, the soul of the animal, and a phenomenal body—and, possibly, other monads, namely, those monads that perceive parts of the body in question most distinctly and thus compose animals with these parts. Monads exist at the most fundamental, substantial level of created reality, while phenomenal bodies

       

289

paragraph. Each one of us is one and not two, despite the fact that our empirical self and our transcendental self are numerically distinct, in that each one of us is one particular human being of which the empirical self and the transcendental self are distinct parts.⁹⁵ My Self, the one entity that I am, is a particular human being. Since a part is a kind of aspect, on the sketched conception of human beings, it is correct to say that a given empirical self and the corresponding transcendental self are both aspects of the same entity, namely, a particular human being. But, despite the two-aspect language, this does not mean that, with respect to the transcendental distinction between the empirical self and the transcendental self, Kant holds a two-aspect view after all. On Kant’s multiple-parts view of the human being, the empirical self and the transcendental self are numerically distinct and do not share any ontological ingredients. It would be wrong to say that the empirical self is the transcendental self as it appears to us (although it is correct to say that the empirical self is what the transcendental self appears to us as), or that a certain thing qua bearer of appearance properties is the empirical self while the same thing qua bearer of in itself properties is the transcendental self. Granted, the difference between the two-world reading and the two-aspect reading with respect to the relation of the empirical self and the transcendental self is less striking and more subtle than the difference between these readings with respect to the relation of empirical objects and their transcendental grounds. This asymmetry is ultimately due to the fact that inner sense and outer sense operate in significantly different ways, on Kant’s account, as discussed above. But a subtle difference is still a difference. For x and y to be numerically distinct parts of a complex entity z, on the one hand, and for x to be z as it appears to us and for y to be z as it is in itself, or for x to be z qua bearer of certain properties and for y to be z qua bearer of certain other properties, on the other hand, indisputably are different relations between x, y, and z.⁹⁶ If we read Kant as committed to the indicated multiple-parts view of human beings, we can also explain his specific two-aspect formulations in the passages quoted above. I, the human being, belong to both the sensible and the intelligible world, can consider myself from two different points of view and as subject to two different sets of laws, and can attribute both nature and freedom to myself because I am composed of multiple distinct ontologically basic parts, some of which belong to the sensible world and are subject to the laws of nature, namely, my empirical self and my body, and at least one of which belongs to the intelligible world and is free, namely, my transcendental self. By way of illustration, consider, as an analogy, the question of how I can be in two different countries at once that have different laws. Straightforward answer: by sitting down right on the border, with my left side in one country and the right side in the other. Similarly, how can I be both cold and not cold at the same time? Straightforward answer: by having cold feet, say, but warm hands.

exist at the less real, mind-dependent phenomenal level; animals thus exist partly at the substantial level and partly at the phenomenal level. ⁹⁵ Also note that it is not my purpose here to provide a reconstruction of Kant’s account of persons, let alone of personal identity. I am merely talking about his account of human beings. ⁹⁶ Ralph Walker also suggests that Kant holds that (what Walker calls) the noumenal subject, on the one hand, and the phenomenal person or human being, on the other hand, are intimately related, but distinct parts of a composite whole; see Walker 2010, 836 7. (My reading of Kant on the relation of the transcendental and empirical self was developed independently.)

290

       

Many readers, especially, after having read the feet example, will probably have the following worry. Kant’s exact formulation in the passage quoted above and in other similar passages is that the thing as appearance is subject to the laws of nature, and that the thing as thing in itself is free. If the appearance and the thing in itself are understood as parts of me, this translates into the statement that, as one part of me, I am subject to the laws of nature, and, as another part of me, I am free, a statement that, one might say, sounds rather odd. And, one might add, ‘as a pair of feet, I am cold’ sounds more than odd. More generally, one might hold that while it is correct to say that I am my parts collectively, it is simply false that I am my parts individually, which also means that I cannot meaningfully attribute properties or determinations to myself as one of my parts. There are two possible responses to this worry. First, one could argue that even though an entity is not identical to its parts taken individually, strictly speaking, the special relation that obtains between it and each one of its parts is a kind of identity relation after all, namely, what we may call ‘partial identity,’ which can legitimately be expressed by saying that the entity is the part. The part is not all the entity is, but it is part of what it is. In this context, it is noteworthy that once we shift from talking about ordinary, spatial parts, such as feet, to talking about ontologically basic parts, such as the soul or the body, the relevant identity statements lose much of their odd sounding quality, at least to my ears. ‘I am a soul, and I am a body’ sounds acceptable to me.⁹⁷ And although not particularly elegant, it sounds similarly acceptable to me to say that, as a soul I am indivisible, and as a body I am divisible.⁹⁸ Statements of partial identity are less familiar than ordinary identity statements, in particular, if they concern such humble parts of ourselves as our feet, which is why they may sound a bit odd, but they are perfectly intelligible. This also holds for the statement that, as one of my ontologically basic parts, namely, as an appearance, I am subject to the laws of nature, and, as another one of my ontologically basic parts, namely, as a thing in itself, I am free. A second possible response to the indicated worry would be to maintain that, given Kant’s demonstrated tendency to occasional sloppiness in expression, the formulation that the thing as appearance is subject to the laws of nature should be understood as an imprecise, abbreviated way of saying that the part of the thing that is an appearance is subject to the laws of nature, or that with respect to its part that is an appearance the thing is subject to the laws of nature. In this context, it is also important to note that two-aspect commentators do not have an easy time accommodating Kant’s exact formulations either. On the two-aspect view, an appearance is a thing (considered) as it appears to us, or a thing qua bearer of appearance properties, and a thing in itself is a thing (considered) as it is in itself, or a thing qua bearer of in-itself properties. If we plug this into Kant’s formulations above we end up with the statements that nature can be attributed to the thing as thing (considered) as it appears to us, and freedom can be attributed to the thing as thing (considered) as it is itself, or that as a thing qua bearer of appearance properties I am subject to the laws of nature, and as a thing qua bearer of in-itself properties I am free. These statements also sound rather odd. ⁹⁷ And I do not seem to be alone. A casual internet search yields the following title of a 2010 book by Sami Jarroush: I am the Body, the Mind, and the Soul. Another one, from 1995 by Elisabeth Moltmann-Wende, is called I Am My Body: A Theology of Embodiment. ⁹⁸ Also see in this spirit FM, 20:308: “But he [the human being] is also conscious of himself as an object of his outer senses, i.e., he has a body, with which the object of inner sense is connected, which is called the soul of the human body.”

       

291

I suspect that the two-aspect commentators will want to protest against this kind of paint by numbers approach as too close-minded—that is, the approach of taking their understanding of what an appearance and a thing in itself is and simply inserting it into Kant’s formulations “as an appearance, the thing is . . . ” and “as a thing in itself, the thing is . . . ”— and insist that it is perfectly obvious what Kant must be taken to say on their reading, namely, that nature can be attributed to the thing (considered) as it appears to us, and freedom can be attributed to the thing (considered) as it is in itself, or that, considered as I appear to myself, I am subject to the laws of nature, and, considered as I am in myself, I am free. I am prepared to grant this kind of leeway to the two-aspect commentators in dealing with Kant’s exact formulations for the particular case at hand, but only if I am granted the same privilege and spared the paint by numbers approach as well. That is, even if the two-aspect commentators reject my first response to the formulation worry stated above and do not admit that the statement that, as one of my parts, namely, as appearance, I am subject to the laws of nature, is intelligible, they must at least grant me the second response, that, despite Kant’s slight sloppiness, it is perfectly obvious what he must be taken to say on my reading, namely, that nature can be attributed to the thing with respect to its appearance part, while freedom can be attributed to it with respect to its in-itself part, or that, as far as my appearance part is concerned, I am subject to the laws of nature, and, as far as my in-itself part is concerned, I am free. Kant often employs two-aspect language when talking about his solution to the apparent problem that freedom seems impossible, given that all events in the empirical realm, including our actions, are causally determined by other events in the empirical realm. As just explicated, on the assumption of the multiple-parts view of human beings, these kinds of passages are perfectly intelligible even if the transcendental self and the empirical self are understood as numerically distinct entities. But there are also many passages in which Kant explicitly describes his solution to the indicated problem in multiple-object terms. If the objects of the world of sense were regarded as things in themselves, and the laws of nature listed above were regarded as laws of things in themselves, the contradiction would be inevitable. Similarly, if the subject of freedom, like the other objects, were conceived of as a mere appearance, the contradiction could also not be prevented; for the very same would be simultaneously affirmed and denied of the same object in the same sense. But if natural necessity is related only to appearances and freedom only to things in themselves no contradiction ensues, even if one assumes or admits both kinds of causality, even though it might be difficult or impossible to make the latter kind intelligible. (Prol, 4:343, my emphasis)⁹⁹ ⁹⁹ Also see GMS, 4:459: “One can only show them [the people who declare freedom to be impossible] that the contradiction allegedly discovered in it by them lies in nothing else than that, since they necessarily had to regard the human being as appearance in order to apply the law of nature to human actions, and now that one demands of them that they should think him, as intelligence, also as thing in itself, they still regard him even then as appearance, where, thus, of course, the separation of his causality (i.e., of his will) from all laws of nature of the world of sense in one and the same subject would be a contradiction, which, however, falls away when they reflect and admit as they should that behind the appearances the things in themselves still must be there as their grounds (although hidden), of whose laws of effect one cannot demand that they are the same as the ones under which their appearances stand.” See V-Met/Mron, 29:924–925 (my emphasis): “Thus, the antinomy that freedom and necessity exists in the world is solved in that it is shown that these are not real opposites, in that they pertain to different objects, namely, the former to the mundus noumenon and the other to the mundus phenomenon, and thus both can take place at the same time.” Also see Bxxvii–xxvviii.

292

       

These kinds of multiple-object formulations cannot be chalked up to mere sloppiness. A multiple-object conception of the relation between the thing that is free and the thing that is subject to the laws of nature is required in order for the contradiction to be resolved in the way Kant indicates. How so? In section 3.5, I argued that the one-object commentators do not have a satisfactory solution to the problem that, on their reading, the very same thing turns out to be both extended and not extended. One of the points I made there is that, in an extensional context, there are only two ways in which an apparent contradiction of the form ‘S is F, and S is not F’ can be made to disappear, namely, (a) by showing that the apparently contradictory predicates are not really contradictory after all, for example, because of an ambiguity, or (b) by showing that the contradictory predicates are ascribed to distinct things after all. And since option (a) was ruled out for the case in question—empirical objects do not merely appear to be extended nor do they merely have the property of appearing as extended to perceives with our kind of sensibility, they really are extended—it emerged that the only way in which the one-object commentators could escape the contradiction is by admitting that the thing-(considered)-as-it-appears-to-us and the thing-(considered)-as-it-is-in-itself are two distinct things after all, that is, by joining the multiple-object camp. There are some helpful parallels between this case and the problem discussed by Kant in the passage just quoted. In this passage, Kant agrees that the way to get out of a looming contradiction in a case where the relevant predicates are truly contradictory, that is, where “the very same is simultaneously affirmed and denied in the same sense,” is by showing that they pertain to distinct things after all. Consider the apparently contradictory claim that Franz is both free and causally determined by the laws of nature, where we will assume, with Kant, that ‘is free’ and ‘is causally determined by the laws of nature’ are truly contradictory. Initially, one might try to resolve the apparent contradiction by arguing that Franz really is free but merely appears to be causally determined, or, conversely, that he really is causally determined but merely appears to be free. But, as in the extension case, this ploy is incompatible with Kant’s philosophical commitments. Just as he holds that empirical objects really are extended, he holds that human beings, like all sensible beings, really are causally determined by the laws of nature. Similarly, Kant makes it clear that the kind of freedom that he is talking about is not merely apparent freedom but real freedom. More generally, he is adamant that his solution to the third antinomy shows how the thesis, according to which there is freedom in the world, and the antithesis, according to which everything in the world is determined by the laws of nature, can both be true at the same time.¹⁰⁰ Since, by assumption, ‘is free’ and ‘is causally determined’ are contradictory predicates and, moreover, are not intensional predicates, that is, predicates that introduce an intensional context, the only way to resolve the apparent contradiction that Franz is both free and causally determined is thus to show that the thing that is free and the thing that is causally determined are not the same thing after all. Franz really is both free and causally determined at the same time because he is a complex entity that comprises a supersensible part and a sensible part, and his supersensible part is free, while his sensible part is causally determined. Similarly, the thesis and the antithesis of the third antinomy can both be true at the same time because they are

¹⁰⁰ See B560/A532; V-Met/Dohna, 28:658–659. See note 171, chapter 2, and note 2, chapter 3.

       

293

about distinct things, the world understood as the transcendental level of reality or as comprising all levels of reality, on the one hand, and the world understood as the empirical level of reality, on the other hand. There could be freedom in the world understood in the former sense, whereas everything is causally determined in the world understood in the latter sense. But this resolution is not available on the one-object view. On the one-object reading, there is only Franz, and there is only one world. Franz (considered) as he appears to us is determined by the laws of nature, while Franz (considered) as he is in himself is free. As in the extension case discussed earlier, I do not see how this maneuver could resolve the contradiction or make sense of the quoted passage and others like it, unless it is accompanied by the relinquishing of the central tenet of the one-object view that the thing in itself and the appearance are numerically identical. If Franz-(considered)-as-he-appearto-us and Franz-(considered)-as-he-is-in-himself are admitted to be distinct entities, the contradiction does disappear but only at the price of having thereby defected to the multiple-object reading. So, the two-world reading of the transcendental distinction between the empirical self and the transcendental self is not only compatible with Kant’s text, it is demanded by it.¹⁰¹ Much more would have to be said to explain how exactly human action, and, in particular, morally responsible human action, is to be understood on the proposed multiple-parts view of human beings. But this question—which is rather hairy on any reading of the relation between the empirical and transcendental self—encroaches upon the territory of Kant’s practical philosophy proper, which lies outside the scope of this book. The only question in this general neighborhood that Kant himself discusses in his theoretical philosophy is the limited question that we focused on as well, namely, whether the possibility of freedom is ruled out by the fact that everything in the empirical realm happens according to deterministic laws, a question that can be answered in the negative once we recognize that human beings have both sensible and supersensible parts.¹⁰² But there are two other questions raised by the multiple-parts view, which are of a theoretical (rather than practical) nature, that I would like to address briefly in concluding this section. After having been presented with this view, it is natural to wonder whether there are any other complex entities on Kant’s account that span different levels of reality. In particular, do empirical objects and the things in themselves that ground them also compose additional entities? Given our general cognitive limitations, and, more specifically, the limitation that we do not have any independent cognitive access to the transcendental grounds of empirical objects—that is, cognitive access apart from knowing that they ground empirical objects by way of affecting outer sense—it should come as no surprise that, according to Kant’s theory, we have no way of knowing the answers to these questions. One might try to argue that the possible lack of a one-to-one mapping between empirical objects and their transcendental grounds should be regarded as evidence against the view that each empirical object and its transcendental ground(s) compose an entity, since this would leave us with the possibility of partially overlapping entities, that is, ¹⁰¹ See B479/A451: “Moreover, even if in the meantime a transcendental faculty of freedom to begin changes in the world is conceded, this faculty at least would have to be outside of the world. . . . But to ascribe such faculty to the substances in the world can never be permitted, since in that case the connection of appearances that necessarily determine one another in accordance with general laws, which is called nature, would mostly disappear and with it the criterion of empirical truth, which distinguishes experience from a dream.” ¹⁰² See B585–586/A557–558.

294

       

distinct entities that share common parts, which, one might hold, are metaphysically impossible. But since the assumption that partially overlapping entities are metaphysically impossible itself seems hard to justify, this argument remains inconclusive.¹⁰³ There may be general metaphysical composition principles that specify necessary and sufficient conditions for multiple distinct entities to compose a new, additional entity, but it does not seem to be possible for us to find out what they are. For all we know, empirical objects and their transcendental grounds may be distinct parts of complex entities, but they also may not be. This directly leads to the second question raised by the multiple-parts view of the human being that deserves brief consideration, namely, how Kant could possibly justify it. In particular, on what grounds can we claim that each empirical self and the transcendental self that grounds it are distinct parts of a complex entity, while having to remain agnostic about whether empirical objects and their transcendental grounds compose complex entities? What is special about the case of the self? Just pointing to the by now familiar difference between the ‘inner’ and the ‘outer’ case, that, due to the different workings of inner and outer sense, each empirical self is necessarily grounded in only one transcendental self, while empirical objects could each be grounded in more than one thing in itself, and things in themselves could each ground more than one empirical object, is not sufficient for the justification that we are looking for. That we do not have to worry about the possibility of multiple distinct human beings who share the same transcendental self is reassuring, but it does not amount to any positive support for the claim that each empirical self and the transcendental self that grounds it are distinct parts of a complex entity. I believe that, on Kant’s view, this support rests on our self-consciousness. Briefly put, I take it that Kant’s answer to the question of what is special about the case of the self in this context would or at least could be that we have ‘inside’ information, in the literal sense of the word, based on our self-consciousness, and that this ‘inside’ information allows each one of us to justify the claim that our empirical self and our transcendental self are, indeed, ours. As reported, Kant holds that in being conscious of myself in selfconsciously thinking I am related to a supersensible subject of consciousness, or a transcendental mind. And he holds that in my ordinary empirical self-consciousness I am related to an empirical mind. Moreover, the way in which Kant talks about human action, especially in the context of his practical philosophy, suggest that he would accept the characterization that in self-consciously acting I am related to a body. The idea is that I could not be self-conscious in these various ways if the entities to which I am related in being self-conscious—a transcendental mind, an empirical mind, and a body—were not actually all parts of me, that is, if I were not actually each one of these different entities in the sense of being partially identical with them as explicated above. This is how my selfconsciousness, plus Kant’s account of self-consciousness, justifies the claim that my transcendental self—which is my transcendental mind—my empirical self, and my body are indeed all mine. Since we do not have any comparable ‘inside’ view in the case of empirical objects and their transcendental grounds, we have no reason to assume that they too compose additional, complex entities.

¹⁰³ Also, it might be that only some empirical objects compose additional entities with their transcendental grounds, namely, those empirical objects that have a unique transcendental ground. Such complex entities would not overlap with any other entities.

     

295

5.9 Additional Core Theses of Critical Idealism: Things in Themselves and the Grounding of Appearances 5.9.1 P, G, and E The theses of the transcendental ideality and empirical reality of empirical objects (or, more specifically, TI1-strong and ER1), the theses of the transcendental ideality and empirical reality of space and time (or, more specifically, TI2-pure-strong, TI2-empirical-strong, and ER2), and the theses of the transcendental ideality and empirical reality of empirical selves (or, more specifically, TI3-strong and ER3), constitute the core of Kant’s transcendental idealism and empirical realism, respectively. Since transcendental idealism and empirical realism are integral parts of Kant’s critical idealism, these theses are also among the core theses of critical idealism. But in order to characterize the particular flavor of Kant’s overall ontological position, several further theses must be taken into account that reflect his views about things in themselves and their relation to appearances. We will now turn to collecting these further core theses of critical idealism. Kant’s views about things in themselves and their relation to appearances are already sufficiently familiar to us from the foregoing discussion. So, our present job consists mainly in formulating suitable theses that adequately capture these views. We will collect these theses in two batches. The first batch contains theses to which Kant is minimally committed. The second batch contains alternative, stronger versions of the theses contained in the first batch, versions that some readers might be reluctant to ascribe to Kant, not least because of their strength, even though they are also reflected in the Kantian texts. This way of proceeding seems advisable both because Kant’s texts do not support a completely conclusive verdict on which versions of these theses he endorses and regards as part of his official position and because it gives us some room to maneuver when it comes to the question of how critical idealism could be defended. The stronger the position, the harder its justification. Having stronger and weaker versions of critical idealism on record makes it easier to respond to potential justification challenges. Each batch contains three clusters of theses: theses concerning the passive nature of sensibility, theses concerning the grounding of appearances in things in themselves, and theses concerning the existence of things in themselves. Each cluster comprises theses both about outer appearances and their grounds and about inner appearances and their grounds. As in our previous discussion, all of the following theses are to be understood as articulated from the point of view of fundamental ontology, and ‘thing in itself,’ ‘appearance,’ and ‘mind-(in)dependent’ without any qualifications are to be understood in the transcendental sense, and ‘experience’ and ‘appearance’ without any qualification are to be read as ‘outer experience’ and ‘outer appearance,’ respectively.¹⁰⁴ Additional theses of critical idealism, batch I: (P) Due to its essentially passive nature, sensibility can produce sensations and intuitions only in response to being transcendentally affected by things in themselves in the transcendental sense. ¹⁰⁴ For the sake of maximal clarity, I will explicitly include the qualifications ‘in the transcendental sense’ or ‘outer’–‘inner’ when applicable in the ‘official’ statements of the various further theses of critical idealism.

296

       

(P--) Due to its essentially passive nature, outer sense can produce (a) outer sensations only in response to being materially transcendentally affected by things in themselves in the transcendental sense and (b) a priori intuitions of space only in response to being formally transcendentally affected by the mind whose outer sense it is. (P--) Due to its essentially passive nature, inner sense can produce (a) inner sensations only in response to being materially transcendentally affected by things in themselves in the transcendental sense and (b) a priori intuitions of time only in response to being formally transcendentally affected by the mind whose inner sense it is. (G) Kantian appearances are essentially grounded in things in themselves in the transcendental sense. (G--) Kantian outer appearances, that is, intentional objects of outer experience as characterized in Kant’s theory, are essentially grounded in things in themselves in the transcendental sense. (G--) Kantian inner appearances, that is, intentional objects of inner experience as characterized in Kant’s theory, are essentially grounded in things in themselves in the transcendental sense. (E) There are things in themselves in the transcendental sense. (E--) There is at least one thing in itself in the transcendental sense that grounds empirical objects. (E--) There is at least one transcendental self, that is, a thing in itself in the transcendental sense that grounds an empirical self. The specification that the relevant affections are transcendental affections is a reminder that we are talking about affections at the transcendental level of reality between mindindependent things, while the sensations and intuitions that result from these affections, as mental states that occur in time, exist at the empirical level of reality. (This is why P, in all of its versions, counts as a thesis about how entities at the transcendental level and entities at the empirical level are related.) The distinction between material and formal affections is required to make sense of how there could be representations that are both sensible and a priori, as explicated in section 4.2.4. On the account of sensibility captured in the stated weak versions of P, inner and outer sense are primarily distinguished by their forms. Time is the form of inner sense; space is the form of outer sense. Outer sense is outer in that it is the ultimate formal condition for us presenting objects as being ‘outside us,’ which we accomplish by presenting them as being in space; inner sense is inner in that it is not outer in this sense.¹⁰⁵ ¹⁰⁵ It is also not implausible to suppose that the difference between inner and outer sense tracks a difference in the manner in which things in themselves transcendentally affect them for the production of sensations. We might capture this difference by saying that things in themselves bring about inner sensations by affecting inner sense inwardly, with their ‘inside,’ while they bring about outer sensations by affecting outer sense outwardly, with their ‘outside,’ so to speak. Outer sense thus can be said to be outer also in that it is receptive to material transcendental affections by the ‘outside’ of things in themselves, and inner sense thus can be said to be inner also in that it is receptive to material transcendental affections by the ‘inside’ of things in themselves. (Both inner and outer sense can be formally transcendentally affected from ‘inside,’ namely, by the mind’s own activity, whereby a priori intuitions are produced, which represent the forms of sensibility. This kind of formal affection from ‘inside’ is thus to be distinguished from inward material affections, or material affections by the ‘inside’ of things in themselves, understood in the way just explicated.)

     

297

There are reasons to think that the basic and weak versions of P, G, and E just recorded are not strong enough to capture Kant’s vision of the ontological landscape at the transcendental level of reality. As we saw in section 5.7, Kant’s transcendental psychology commits him to the view that there are human transcendental minds, that is, mind-independent ‘owners’ or subjects of various transcendental cognitive faculties. Moreover, his account of self-consciousness suggests that he believes that in self-consciously thinking we are related to a supersensible, mind-independent subject, a subject that can plausibly be identified with the subject of our transcendental cognitive faculties. In the context of his transcendental psychology, Kant also provides a more specific account of the operations of sensibility when he describes inner sense as producing inner sensations only in response to self-affections, that is, in response to affections by the thing in itself whose inner sense it is. Relatedly, he quite clearly holds, not only that he has a transcendental self that grounds his empirical self, but also that this transcendental self is his transcendental mind. And he seems to endorse the claim that he is a human being, that is, an entity whose ontologically basic parts comprise, minimally, a human transcendental self, an empirical self, and a body. Finally, as we saw in section 5.6, there are reasons to assume that Kant regards neither his own transcendental self as the only thing in itself in existence nor human transcendental minds as the only kind of things in themselves in existence. For he appears to endorse the view that empirical objects are grounded in things in themselves that are distinct from all human minds. Indeed, he even appears to endorse the view that empirical objects are grounded in things in themselves that are distinct from all human minds and God. These claims are quite a bit stronger than any implications of the versions of P, G, and E on record so far. So, against the background of our discussion in sections 5.6 and 5.7, we can thus formulate the following stronger versions of P, G, and E. Additional theses of critical idealism, batch II: (P--) Due to its essentially passive nature, outer sense can produce (a) outer sensations only in response to being materially transcendentally affected by things that are moderately strongly ‘outside us’ in the transcendental sense, that is, by things in themselves in the transcendental sense that are distinct from the thing in itself whose outer sense it is, and (b) a priori intuitions of space only in response to being formally transcendentally affected by the mind whose outer sense it is. (P--) Due to its essentially passive nature, inner sense can produce (a) inner sensations only in response to being materially transcendentally affected by the thing in itself whose inner sense it is and (b) a priori intuitions of time only in response to being formally transcendentally affected by the mind whose inner sense it is. (G--) Kantian outer appearances, that is, intentional objects of outer experience as characterized in Kant’s theory, are essentially grounded in things that are strongly ‘outside us’ in the transcendental sense, that is, things in themselves in the transcendental sense that are distinct from all human minds.¹⁰⁶ ¹⁰⁶ As things in themselves, the transcendental grounds of empirical objects are obviously distinct from all human empirical minds (which are appearances). But, according to P--, they are also distinct from all human transcendental minds. So, it is correct to say in general that the relevant things in themselves are distinct from all human minds.

298

       

(G--) Kantian outer appearances, that is, intentional objects of outer experience as characterized in Kant’s theory, are essentially grounded in finite things that are strongly ‘outside us’ in the transcendental sense, that is, things in themselves in the transcendental sense that are distinct from all human minds and God. (G--) Each Kantian inner appearance, that is, each intentional object of inner experience as characterized in Kant’s theory, is essentially grounded in one thing in itself in the transcendental sense that is (a) unique to the appearance and (b) a human transcendental mind. (E--) There is at least one thing in itself in the transcendental sense that grounds empirical objects and is not a human mind. (E--) There is at least one thing in itself in the transcendental sense that grounds empirical objects and is neither a human mind nor God. (E--) (a) There is at least one transcendental self, that is, a thing in itself in the transcendental sense that grounds an empirical self; and (b) there is at least one human transcendental mind. (E--) There is at least one transcendental self that is a human transcendental mind.¹⁰⁷ (E--) There is at least one human being, that is, an entity whose ontologically basic parts include, minimally, a body (which is an empirical object), an empirical self (which is an empirical mind), and a transcendental self (which is a transcendental mind). On the account of sensibility captured in the stated strong versions of P, inner and outer sense are distinguished not only by their forms. Inner sense can be materially affected only by the thing in itself whose inner sense it is. That is, the only material affections that it can undergo are affections from ‘inside.’ By contrast, outer sense can be materially affected only by things in themselves that are distinct from the thing in itself whose outer sense it is. That is, the only material affections that it can undergo are affections from ‘outside.’ Also note that even though P-- describes the things in themselves in response to whose affections outer sense can produce outer sensations only as distinct from the thing in itself whose outer sense it is, it leads to the same version of G, namely, G--, as a slightly stronger version of P would, according to which outer sense can produce outer sensations only in response to being affected by things in themselves that are distinct from all human minds. P--

¹⁰⁷ E--/ and E--/ together imply that there are at least two things in themselves in the transcendental sense. Also note that Kant is committed to the finitude of human minds and to their distinctness from God. So, --/ implies that there is a finite thing in itself that is distinct from God. Denying the distinctness of human minds from God would amount to a version of Spinozism, according to which human minds are parts or modes of God considered under the attribute of thought, a view that Kant clearly and firmly rejects. Kant points out that this view has the consequence that our thoughts either think themselves or are thought for us, each one of which is plainly absurd. The concept of a thought that thinks itself is utter nonsense; for such a thing would be “an accident that simultaneously exists as a subject for itself: a concept that cannot at all be found in human understanding and also cannot be brought into it” (WDO, 8:143, note). Moreover, our self-consciousness clearly establishes that it is each one of us who thinks his own thoughts, not somebody else. See V-Phil-Th/Pölitz, 28:1042; V-Met/Volckmann, 28:457–458; V-Th/ Baumbach, 28:1269.

     

299

, together with Kant’s theory of experience, entails that, for each human transcendental mind, every outer appearance is essentially grounded in things in themselves that are distinct from this mind. But since, in contrast to inner appearances, outer appearances are public, and the transcendental ground of an outer appearance is invariant with respect to changes of cognizer, it follows that every outer appearance is essentially grounded in things in themselves that are distinct from all human transcendental minds. It is also worthwhile to highlight that, given God’s traditionally alleged omnipotence, it is not plausible to assume that Kant would approve of a bold version of P, according to which outer sense can produce outer sensations only in response to being transcendentally affected by things in themselves that are distinct from all human minds and God. Surely, if God exists, then He can transcendentally affect outer sense. This is useful to highlight because it means that the claim that the transcendental grounds of empirical objects are distinct from God must be justified apart from and in addition to the claim that these grounds are distinct from all human minds. Only the latter claim can be derived from P-- and Kant’s theory of experience, as will be spelled out in section 5.10.2.

5.9.2 Bold and Timid Critical Idealism The different versions of P, G, and E listed in section 5.9.1 allow us to construct different versions of critical idealism that vary in strength. All versions that can plausibly be ascribed to Kant minimally incorporate the core tenets of transcendental idealism and empirical realism, discussed in chapters 3 and 4 and recapitulated at the beginning of section 5.9.1, plus P--, P-, G--, G--, E--, and E--.¹⁰⁸ I will call the version of critical idealism defined by these theses ‘timid critical idealism.’ Timid critical idealism is at the weak end of the spectrum of possible critical idealist views.¹⁰⁹ The version at the strong end of the spectrum in addition includes P--, P--, G-, G--, E--, E-, and E--. I will call it ‘bold critical idealism.’ Since there are many other combinations of the various versions of P, G, and E, as well as many other possible versions of these theses, whose strength falls in

¹⁰⁸ Given the centrality of Kant’s transcendental psychology for his critical philosophy, no version of critical idealism has E-- as its strongest ‘inner’ existence thesis. The commitment to the existence of at least one transcendental self and at least one human transcendental mind is a non-negotiable element of any version of critical idealism. ¹⁰⁹ P-- allows that outer sense can be affected by human transcendental minds and even that a particular outer sense can be affected by the thing in itself whose outer sense it is, in contrast to P-. More generally, P-- and G-- leave open the possibility that the entire outer empirical world is grounded in only one human transcendental mind, which grounds empirical objects by ‘outwardly’ affecting itself. (See note 105.) So, they cannot be used to justify the claim that there is more than one human mind, or that there are things in themselves that are not human minds, let alone that there are things in themselves corresponding to empirical objects that are not human minds. Still, weak as they are, these theses allow us to distinguish timid critical idealism from ordinary idealism in that the former allows that there might be things in themselves distinct from God, or even distinct from God and from all human minds, that ground empirical objects by affecting us, whereas the latter flat-out rejects their existence.

300

       

between the weak and strong versions just discussed, there are also many other possible versions of critical idealism whose strength falls in between the timid and bold versions just described. But in order not to overcomplicate our discussion, I will resist the temptation to explicitly spell out any more versions of critical idealism.¹¹⁰ Although there are fairly solid textual reasons to think that Kant personally believes bold critical idealism to be true, it can be questioned whether this is the version that he wants to or should want to present as his official position, given how bold it is.¹¹¹ There are a few passages in which he is indeed much more cautious about committing himself to a specific view about what kind of things in themselves affect us and ground empirical objects than in the Prolegomena and some of the related passages referenced above. For example, at A358 he entertains the possibility that “ . . . the something that grounds outer appearances, what affects our sense in such a way that it receives the representations of space, matter, and figure, this something, considered as noumenon (or better, as transcendental object), could still at the same time be the subject of thoughts, although, through the manner in which our outer sense is affected by it, we receive no intuition of representations, will etc. but merely of space and its determinations.” And at A385, he uses the guarded formulation that “there may well be something outside us to which this appearance corresponds that we call matter.” These passages suggest that the second main difference between critical idealism and ordinary idealism is not that the former asserts the existence of mindindependent grounds of empirical objects that are distinct from God and all human minds while the latter denies it, but merely that critical idealism allows that there might be such mind-independent grounds while ordinary idealism denies it. And, indeed, in a note in his private copy of the A-edition of the Critique, Kant characterizes the difference between critical idealism and ordinary idealism along these lines by saying that “pure idealism concerns the existence of things outside us. The critical one leaves it undecided and claims only that the form of their intuition is merely in us” (R XXVI E 18, 23:23).¹¹² One might also argue that his strongly worded remarks about the existence of things in themselves ‘outside us’ that ground empirical objects in the Prolegomena must be taken with a grain of salt. They are the fruit of Kant’s frustration over the Garve–Feder review of the first Critique and its characterization of critical idealism as a mere rehash of Berkeley.¹¹³ In light of the review, it would be understandable if Kant got carried away a little in his protestations that his idealism is not at all like Berkeley’s and overstated the differences between them.

¹¹⁰ In order to counteract theses-induced vertigo, the appendix at the end of this book provides a summary of bold and timid critical idealism, as well as of transcendental realism and ordinary idealism, by conveniently listing all of the theses that define them in one place. ¹¹¹ Note that proponents of the one-object view are committed to ascribing E-- to Kant (interpreted according to their reading), provided they agree that, for Kant, empirical objects are distinct from God and from human minds—which they obviously are. Similarly, since it is part and parcel of the one-object interpretation that each empirical self and the transcendental self that appears as that empirical self are numerically identical, they are also committed to a thesis that is analogous to G--. ¹¹² Relatedly, one might draw attention to the fact that, in his metaphysics lectures, Kant makes a point of explicitly rejecting dogmatic idealism and solipsism (or ‘egoism,’ as he calls it), which assert that there are no other beings than thinking beings, or other beings than our own mind, while not rejecting the problematic versions, that is, the versions that allow that there might be no other beings than thinking beings, or no other beings than our own mind. See V-Met-L1/Pölitz, 28:206–207. ¹¹³ See notes 40 and 42, chapter 3.

     

301

On the other hand, all of the more cautiously worded passages just quoted are from the A-edition, and thus need not be seen as expressing the position that Kant eventually settled on. One might hold that in 1781 Kant had not fully made up his mind yet about the strength of his idealism and that his eventual view coagulated in sharp relief only after the Garve–Feder review, which is why it is much more clearly in evidence in the Prolegomena and the B-edition.¹¹⁴ The infuriating review forced Kant to think more deeply and more clearly about the nature of his own idealism in contrast with ordinary idealism, which made him realize that the formulation that there “may well be” things in themselves distinct from us that ground empirical objects is much too timid to adequately capture his ontological commitments. So, which version of critical idealism should we crown as the official version? In order to answer this question, it would be very useful to know the answers to two related further questions. First, how well do the various versions of critical idealism cohere with Kant’s epistemology and theory of cognition, in particular, the so-called ‘unknowability doctrine,’ according to which all of our substantive theoretical cognition is restricted to the empirical realm, which means that things in themselves are theoretically unknowable or uncognizable for us? Second, how well can the various versions of critical idealism be justified? If it turns out that not all versions cohere equally well with Kant’s epistemology and theory of cognition or are not equally well justified, we will need to balance various desiderata in choosing the official position. In particular, we will need to decide on the relative importance and weight, in our eyes and in Kant’s eyes, of accomplishing the following aims: (1) minimizing the number and strength of basic commitments involved in justifying the official version of critical idealism that are not justifiable on more fundamental grounds, (2) minimizing the number and seriousness of adjustments to Kant’s epistemology and theory of cognition in order to make it coherent with the official version of critical idealism, and (3) maximizing overlap between the official version of critical idealism and bold critical idealism, which Kant quite clearly believes to be true. Many people will probably hold that, of course, the first two aims far outweigh the third. But it is not clear to me that they do, nor is it clear to me that Kant would agree with that assessment. To my mind, there is something to be said for the view that Kant’s official position is bold critical idealism, despite the fact that, due to its boldness, it is hardest to justify and least likely to turn out to be fully coherent with his epistemology and theory of cognition. To be sure, if it transpired that the basic commitments on which the justification for bold critical idealism rests are outrageously strong and outrageously implausible, and if adopting bold critical idealism would require throwing Kant’s entire epistemology and theory of cognition overboard, this reading would not be an option. As it turns out, however, the basic commitments that carry the weight of the justification for bold critical idealism are quite plausible and not all that much stronger than the basic commitments that carry the weight of the justification for timid critical idealism, and adopting bold critical idealism requires hardly any adjustments to Kant’s epistemology and theory of cognition at all. Since showing the latter presupposes giving a detailed account of Kant’s epistemology and theory of cognition, which goes ¹¹⁴ This is also how Erdmann assesses Kant’s note in his personal copy of the A-edition cited above, whose (not exactly flawless) edition of the notes in Kant’s copy is the only surviving record of them. See Erdmann 1881, 18: “This marginal note obviously belongs to a very early time; presumably it was written down before the appearance of the Göttinger review; for it stands in characteristic opposition both to the investigations in the Prolegomena and to the investigations of the new edition.”

302

       

beyond the scope of the present investigation, I will not be able to conclusively settle these issues here. But we can make substantive progress toward doing so. In the first part of section 5.10, I will sketch a general strategy for how to respond to the charge that critical idealism is in trouble because the various claims about things in themselves that are comprised in it are the kind of claims that Kant’s epistemology and theory of cognition declares to be out of reach for us. I contend that this strategy also works with respect to the theses of bold critical idealism, although showing as much will have to wait until a later occasion. In the second part of section 5.10, I will reconstruct justifications for all of the listed versions of P, G, and E, which will allow us to identify and compare the basic commitments on which the justification for the various versions of critical idealism rest. All of this will already go a long way toward establishing that bold critical idealism not only is Kant’s personal favorite version and believed by him to be true, but also can reasonably be regarded as his official position.

5.10 Kant’s Case for P, G, and E 5.10.1 Worries about P, G, and E P, G, and E were among the prime targets of criticism in the first phase of the reception of Kant’s philosophy. These theses are said to be problematic because (a) they are in tension with the unknowability doctrine, according to which all of our substantive theoretical cognition is restricted to the empirical realm and because (b) they conflict with the result of the Transcendental Analytic that the categories of causality and actuality meaningfully apply only to appearances and objects of possible experience. Call worries (a) and (b) with respect to P and G the ‘the problem of affection’ and with respect to  the ‘problem of existence.’ I concur that these problems are not to be taken lightly, and, although a detailed investigation of how Kant might respond to them goes beyond the scope of this book, I will briefly sketch a general response strategy below. But first and foremost, I would like to comment on the questionable dialectical merit of trying to turn the problems of affection and existence into objections to the classic two-world interpretation. Two-world interpretations typically ascribe versions of P, G, and E to Kant, and, thus, incur the burden of having to address the problems of affection and existence. And since, so the objection goes, these problems are very tough or even downright intractable, these interpretations must be rejected in favor of other readings that do not saddle Kant with these problems.¹¹⁵ Of course, I agree that if the choice is between two interpretations of an author that equally respect the textual evidence, it is sound hermeneutical practice to adopt the one that does not ascribe any theses to him that are in apparent conflict with other claims that he explicitly asserts. But this is not our present situation. There is incontrovertible textual evidence that Kant endorses P, G, and E. Indeed, since these theses were under fire right from the outset, and since several of his contemporaries

¹¹⁵ This kind of ploy is standard fare in the literature, for example, see Robinson 1994, 415; Rogerson 1999, 2.

 ’    , ,  

303

directly asked Kant about them, we even have his explicit assurance that he means them literally, as stated it in the Critique, and does not merely entertain them as useful fictions.¹¹⁶ It is certainly not sound hermeneutical practice to avoid perceived potential problems within the theory of one’s author by ignoring textual evidence or by assuming that he does not really mean what he says over and over again. So, far from being an advantage of an interpretation that it does not present Kant as committed to the literal truth of P, G, and E, it constitutes a major liability since it flies in the face of the textual evidence.¹¹⁷ But, one may wonder, what about those two-aspect interpretations that honor the textual evidence and do ascribe versions of P, G, and E to Kant but, due to their alternative conception of what Kant means by ‘thing in itself,’ give them a different reading? Are they not in a better position than two-world interpretations to integrate these theses with the rest of Kant’s theoretical philosophy? In a word, no. We can cognize as much or as little of empirical objects considered apart from our sensible epistemic conditions, or of empirical objects qua bearers of mind-independent properties, or of the mind-independent aspects of the things whose mind-dependent aspects empirical objects are, as we can cognize of supersensible mind-independent things that are numerically distinct from empirical objects and do not share any ontological ingredients with them, and the categories of causality and actuality apply as much or as little to the former as they apply to the latter.¹¹⁸ Insofar as the problems of affection and existence are genuine problems, they are problems, not just for proponents of the classic two-world view, but, in the first place, for Kant and, in the second place, for any interpreter who respects the relevant textual evidence. So, just briefly, how then could Kant respond to these problems? Let us first look at worry (a), that P, G, and E violate Kant’s doctrine that all of our substantive theoretical cognition is restricted to the empirical realm. This worry is a specific version of what is often called ‘the unknowability problem.’ This is the problem that Kant’s unknowability doctrine, that is, the doctrine that all of our substantive theoretical cognition is restricted to the empirical realm, and thus things in themselves are theoretically unknowable or uncognizable for us, appears to be in tension with the fact that he makes all sorts of claims about things in themselves in his theoretical philosophy, not just that they exist, affect us, and ground appearances, but also that they are ¹¹⁶ See again ÜE, 8:215, quoted at the beginning of this chapter, and EF, 12:371, quoted in section 1.4. ¹¹⁷ Interpretations of this kind include fictionalist readings and all readings on which the only kind of affection in Kant’s theory is empirical affection, that is, affection by appearances, examples of which are defended, in various ways, by Beck and many neo-Kantians such as Liebmann and Cohen. For discussion, see Vaihinger 1892, 49 51. Also see note 65, chapter 1. The latter kind of readings are not only textually problematic, they also incur the additional cost of landing us in a vicious circle or unacceptable infinite regress. For the sensations that are the result of the affections of sensibility, according to Kant, are supposed to provide the ‘matter’ of appearances in the first place. For more discussion of this additional problem, see note 133. Depending on the details of the particular interpretation, there may be additional costs for proponents of fictionalism too. For example, if the suggestion is, not that we construct a fiction according to which things in themselves affect us and cause sensations, but that fictional objects affect us and cause sensations, one can reasonably object that this makes matters even worse overall. Schulze, who reads Reinhold’s fictionalist re-interpretation in this way, puts his finger on the obvious problem. See Schulze 1792, 308, note: “If the thing in itself, in its distinction from the sensible representations, is merely an idea or a logical thing, in order to be able to show that the matter of sensible representations is something given by it one must presuppose that an idea and a merely logical thing can give something and affect receptivity, which, however, is obviously absurd.” ¹¹⁸ This point is also made by Langton in her critical discussion of Allison’s response to the problem of affection. See Langton 1998, 11.

304

       

supersensible, non-spatial, and non-temporal, for example. The unknowability problem rests on the presupposition that Kant’s claims about things in themselves in the Critique and related writings fall into the category of representations that the unknowability doctrine rules to be impossible for us with respect to things in themselves. The most obvious and most popular strategy for how to deal with the unknowability problem is to challenge this presupposition.¹¹⁹ My preferred way of implementing this strategy includes three steps. First, I argue that the unknowability doctrine declares only a special kind of substantive theoretical cognitions to be impossible for us with respect to things in themselves. For ease of communication, call the relevant kind of cognitions ‘cognitions in the strict sense.’ Second, I make the case that there are different kinds of justified, or at least permissible, claims about objects that we can make, according to Kant’s epistemology, only some of which are cognitions, according to his theory of cognition, and only some of which are cognitions in the strict sense. Third, I show that the indicated Kantian claims about things in themselves either are not cognitions at all, despite being justified or at least permissible, or are not cognitions in the strict sense. For example, to illustrate the first type of argument of step three, as we saw in section 5.7, one of the necessary conditions for a representation to count as a cognition is that it ascribes a determination to a thing, either by predicating it of the thing, as in a judgment, or by presenting the thing as having it, as in an intuition. But, as we know from Kant’s famous critique of the ontological argument for the existence of God, ‘existence’ is not real predicate, and the property of existing is not a real property or determination.¹²⁰ Accordingly, no existence claim is a cognition, not even justified existence claims. So, E is not in tension with the unknowability doctrine because it is not even a cognition, let alone a cognition in the strict sense.¹²¹ As noted in section 5.9.2, I submit that the outlined strategy for handling the unknowability problem largely works even if the seemingly problematic claims about things in themselves are the theses of bold critical idealism; spelling this out, however, is the topic for another day.¹²² This leaves worry (b), that P, G, and E conflict with the doctrine that the categories of causality and actuality meaningfully apply only to appearances and objects of possible experience. The most obvious and most popular response to this worry, which I endorse as well, is to argue that this doctrine only concerns the ‘schematized’ categories of causality and actuality but not their unschematized companions.¹²³ The schematized categories, that is, the versions of the pure concepts of the understanding that include their sensible application conditions, trivially apply only to

¹¹⁹ Different versions of this strategy are employed, for example, by Adams 1997, 809–13; Langton 1998, esp. 12–24, 41–3; Ameriks 2003a, 29–33; and Hogan 2009, esp. 57–60. ¹²⁰ See B625 629/A597 601. ¹²¹ See in this sense KU, 5:453: “ . . . there is no material for us for the determination of the ideas of the supersensible in that we would have to take it from the things in the world of sense, but such a material is not at all adequate to this object, but without all determination of it nothing more than the concept of a nonsensible something remains that contains the final ground of the sensible world, which does not constitute a cognition yet (as expansion of the concept) of its inner quality.” ¹²² See Jauernig (in preparation). ¹²³ See B166, note: “In order to avoid premature concern about the alarming disadvantageous consequences of this proposition [that no cognition a priori is possible except of objects of possible experience], I want to recall that in thinking the categories are not restricted by the conditions of our sensible intuition, but have an unlimited field . . . ”

 ’    , ,  

305

objects in the sensible realm, but the scope of the unschematized or pure categories is not so restricted.¹²⁴ There is much more to say about all of this, but since these matters concern Kant’s theory of cognition, their detailed examination must wait. But the previous considerations are enough to show that neither worry (a) nor worry (b) spells instant doom for Kant’s critical idealism, not even bold critical idealism and, more specifically, that neither worry constitutes a special problem for the classic two-world interpretation, which is all that we need for our present concerns.

5.10.2 Arguments for P, G, and E Even if one grants that the worries just articulated can be taken care of, one might still wonder whether and how Kant can justify P, G, and E, as well as the various more specific versions of these theses collected in section 5.9.1. This question is the topic for the present section. Among commentators who read Kant as committed to the existence of things in themselves one can find a variety of different views on the questions of whether a special justification for this commitment is required, whether Kant can and does provide such a justification, and what the justification consists in. There is no need to give an exhaustive summary of all of these views here, but I will mention some representative examples. First, there is the view that, while Kant should have argued for the existence of things in themselves, he fails to do so and merely asserts it in blameworthy dogmatic fashion.¹²⁵ Then there are various variants of the view that Kant neither provides nor needs to provide a special argument for the claim that things in themselves exist. For example, some twoaspect commentators point out that since things in themselves are numerically identical to ordinary empirical objects, the existence of the former is as certain as the existence of the latter and established on the very same grounds.¹²⁶ More generally and regardless of ¹²⁴ This kind of proposal is endorsed, in one form or another, for example, by Hartmann 1875, 55–7; Riehl 1876, 431–2; Riehl 1908, 569; Erdmann 1878, 73–4; Adickes 1924, 50–1; Adams 1997, 829–1; Van Cleve 1999, 137. One of the questions that I examine in Jauernig (in preparation) is how the pure categories of causality and actuality, despite their a priority, manage to apply, or intentionally relate, to things in themselves, according to Kant, which is not typically discussed in the relevant literature. ¹²⁵ See Ewing 1934, 197: “Kant gives no grounds for believing in things in themselves, but merely asserts their existence dogmatically.” ¹²⁶ See Allison 2004, 51: “[T]he temptation to worry about the existence of things in themselves disappears once it is recognized that Kant is not primarily concerned with a separate class of entities, which, unlike appearances, would supposedly ‘be there’ even if there were no finite cognizers. As we have seen, the concern is rather with the familiar objects of human experience, considered as they are in themselves.” For the record, I think that this kind of response, at best, works only for a methodological two-aspect view such as Allison’s. On an ontological version of the two-aspect view, according to which things in themselves are empirical objects qua bearers of in-itself properties, the question remains how Kant can justify the assumption that empirical objects in fact have any in-itself properties, or an in-itself aspect. This question does not appear any less pressing nor any easier to answer than the question of how Kant can justify the assumption that there are supersensible, mindindependent entities. (Indeed, Kant seems to regard the former question as particularly problematic; he goes as far as to say that there is as little justification for the assumption that sensible objects have in-itself properties as there is for the assumption that God or the soul exist; see B800/A772: “Merely intelligible beings [such as God or the soul, AJ], or merely intelligible properties of the things of the world of sense, can be assumed with no reasoned justification . . . ”) In fact, even proponents of the methodological two-aspect view face an uncomfortable question in this context, namely, how Kant can justify the assumption that we can consider the familiar objects of human experience as they are in themselves, or that there is something for us to consider once we abstract from the sensible conditions under which the familiar objects of human experience appear to us. This question is

306

       

whether one defends a two-aspect or a two-world interpretation, one could hold that Kant simply takes the existence of mind-independent things for granted, and that he is within his rights to do so. Erich Adickes has a view of this kind, for example. On his reading, Kant starts from a position where the existence of mind-independent things is assumed as a matter of course, and since in the course of his inquiry he does not encounter any reason to doubt their existence, he never revises this assumption. On this kind of interpretation, Kant’s endorsement of the existence of things in themselves reflects a basic, standing commitment on his part in which he is justified on the grounds that he has no reason to give it up.¹²⁷ There are also several commentators who read Kant as offering an argument for the existence of things in themselves, and many of them take this argument to be based on some version of G or P.¹²⁸ My own view falls somewhere in between the latter reading and the one favored by Adickes. I agree that Kant does not infer the existence of things in themselves in the framework of his transcendental philosophy in the way that, say, a contemporary physicist might infer the existence of unexpected elementary particles in the framework of the Standard Model in order to explain certain experimental results that would be unintelligible otherwise. In this respect, I concur with Adickes’s assessment that Kant believes in the existence of multiple things in themselves from the start and never seriously calls this belief into question. Nevertheless, I also think that Kant can and does justify E through an argument that centrally involves G and P. To be sure, there is no special chapter in the Critique entitled ‘justification for the claim that things in themselves exist,’ nor does Kant explicitly spell out a step-by-step demonstration for E anywhere. But there are quite a number of passages throughout his writings that contain fairly clear hints to the effect that he believes himself in possession of such a justificatory argument and about what this argument is supposed to be. These hints include, for example, his assertions in the Prolegomena that “appearances always presuppose a thing in itself and, accordingly indicate it” (Prol, 4:355), and that “the understanding, precisely by assuming appearances, admits the existence of things in themselves,” because by regarding empirical objects as appearances “we admit through this at the same time that

uncomfortable not least because, arguably, the relevant kind of consideration is impossible for us; see the discussion in section 3.5. ¹²⁷ See Adickes 1924, §2–4, 4–14; Adickes 1929, 2. In a slightly different vein, Adickes also claims that Kant’s belief in the existence of things in themselves is based on an immediate inner experience as if of something transcendent. See Adickes 1924, §5, 14–19. ¹²⁸ The most prominent commentator in this group is probably Schopenhauer; see Schopenhauer 1859, Werke, 1:595–597; Schopenhauer 1851, Werke, 4:104–108. Also see Riehl 1908, 366–70. Actually, Riehl seems to want to have it both ways. He stresses that Kant does not infer the existence of things in themselves but remains committed to it because he does not see any reason to doubt it; see Riehl 1908, 569–70. At the same time, he also insists that Kant proves the existence of things in themselves by way of analyzing the sources of our cognition, and, in particular, by exposing the receptive or passive character of sensibility, which can produce sensations only by being affected; see Riehl 1908, 571. A different kind of argument for the existence of things in themselves is offered by Allais. She understands things in themselves as bearers of in-itself properties, as noted before, and, more specifically, as things with intrinsic properties or inner natures. According to her account, since appearances are merely relational, and since all relations require intrinsic grounds, the existence of appearances entails that there are things with inner natures, or things in themselves. See Allais 2015, chs 9 and 10, 207–58. In my assessment, Kant himself would not endorse this argument, for various reasons, but primarily because he does not endorse the second main premise, that all relations require intrinsic grounds. See the discussion of what I call the ‘odd argument for transcendental idealism’ in sections 6.2 and 6.5 below. Probably needless to add, reading Kant as offering a justification for the claim that things in themselves exist and judging this justification to be successful are two different things. Schopenhauer regards Kant’s justification as a dismal failure, for instance.

 ’    , ,  

307

a thing itself is their ground” (Prol, 4:314–315).¹²⁹ As we will see, if one thinks this argument through to uncover its ultimate presuppositions one will eventually uncover a few basic commitments on Kant’s part that are not, and arguably need not be justified by way of yet another argument. So, I agree with Adickes that with E we are approaching rock-bottom in Kant’s theoretical philosophy, so to speak; but in my assessment the relevant basic, rock-bottom commitment is not expressed by E itself but concerns the foundation of P, or so I will argue. Kant’s justification for E--, the thesis that there is at least one thing in itself that grounds empirical objects, which is suggested by texts such as the passages from the Prolegomena just cited, can be stated in form of the following straightforward argument. 1. There exist empirical objects. (ER1-i.) 2. Empirical objects are Kantian outer appearances, that is, intentional objects of outer experience as described in Kant’s theory. (TI1-strong-b.) 3. Kantian outer appearances are essentially grounded in things in themselves. (G--.) 4. Empirical objects are essentially grounded in things in themselves. (From 2 and 3.) 5. There is at least one thing in itself that grounds empirical objects. (E-; from 1 and 4.) If we replace G-- in this argument with one of the stronger versions of G-, that is, G-- or G--, we can infer corresponding stronger conclusions, that is, that there is at least one thing in itself that grounds empirical objects and is not a human mind (E-), or that there is at least one thing in itself that grounds empirical objects and is neither a human mind nor God (E--). E is entailed by any of these results. An analogous argument for E-- can be put together by using the existence of my empirical self as a starting point, of which I am certain on the basis of my self-consciousness. 1. My empirical self exists. (ER3-i.) 2. Empirical selves are Kantian inner appearances, that is, intentional objects of inner experience as described in Kant’s theory. (TI3-strong-b.)

¹²⁹ Also see B564–565/A536–537: “But if, on the other hand, appearances are regarded as nothing more than what they in fact are, namely, not as things in themselves but mere representations that are connected according to empirical laws, they themselves must have grounds that are not appearances.” See V-Met/Mron, 29:857: “They [things in themselves] ground appearances, and I thus can well infer from the appearances to the reality of the things but not from the properties of the appearances to the properties of the things themselves.” See also Bxxvi– xxvii, where Kant says of the proposition “that appearance would be without anything that appears” that it is nonsense or implausible. For further relevant passages, see notes 2, 3, and 4. Many commentators dismiss the argument that is suggested by these kinds of passages as inadequate. See Hartmann 1875, 23–4; Allison 2004, 54–5; Adams 1997, 803, note. Adickes treats the alleged inadequacy of the argument as evidence that Kant did not intend these kinds of remarks as a demonstration for the existence of things in themselves; on his view, they simply express the basic, self-evident assumption that there are things in themselves because “as proofs they would lack any kind of demonstrative force” (Adickes 1924, 9).

308

       

3. Kantian inner appearances are essentially grounded in things in themselves. (G--.) 4. My empirical self is essentially grounded in a thing in itself. (From 2 and 3.) 5. There is at least one transcendental self, that is, a thing in itself that grounds an empirical self, namely, mine. (E--/E---a; from 1 and 4.) If we replace G-- in this argument with G-, we arrive at a corresponding stronger conclusion, according to which there is at least one transcendental self that is a transcendental mind (E--). In addition to the demonstration just sketched that relies on G-, there are two other ways in which one can establish the claim that there is at least one human transcendental mind (E---b). First, we can derive the existence of at least one human transcendental mind from the existence of empirical objects, utilizing the result that Kant’s transcendental psychology and theory of experience reveal the forms of the cognitive faculties of the human transcendental mind to be the formal conditions for the possibility of experience. 1. There exist empirical objects. (ER1-i.) 2. Empirical objects are Kantian outer appearances, that is, intentional objects of outer experience as described in Kant’s theory. (TI1-strong-b.) 3. There is outer experience. (From 1 and 2.) 4. The forms of the cognitive faculties of the human transcendental mind are the formal conditions for the possibility of outer experience. (From Kant’s transcendental psychology and theory of experience.) 5. There is at least one human transcendental mind. (E---b; from 3 and 4.) Second, as we saw in section 5.7, Kant seems to think that the existence of my transcendental mind can be demonstrated by a cogito-style argument. My self-consciousness in thinking as expressed in the thought ‘I think’ attests to the existence of both an empirical self and a supersensible, mind-independent subject of consciousness.¹³⁰ On the assumption that only a mind can be a subject of consciousness in self-consciously thinking, this allows me to conclude that there is at least one human transcendental mind, namely, mine. My self-consciousness is also the center piece of a possible justificatory argument for E--, the claim that there is at least one human being, that is, an entity whose ontologically basic parts include, minimally, a body, an empirical self (which is an empirical mind), and a transcendental self (which is a transcendental mind). Given that my body is an ordinary empirical object among ordinary objects and part of the lawgoverned system of empirical objects that constitutes the empirical world, (ER1-i) makes it reasonable for me to hold that my body exists. The existence of my empirical self, my transcendental self, and my transcendental mind have already been established through the previous arguments. P---a, the claim that, due to its essentially

¹³⁰ See B422–423, note.

 ’    , ,  

309

passive nature, inner sense can produce inner sensations only in response to being materially transcendentally affected by the thing in itself whose inner sense it is, allows us to conclude that all transcendental selves are transcendental minds. On the assumption that no entity can have more than one mind, this result allows me to deduce that the transcendental mind of which I am conscious in my self-consciousness in thinking is my transcendental self. Now, my being conscious of a body, an empirical self, and a transcendental mind in self-consciously acting, perceiving, and thinking, respectively, reveals that my body, my empirical self, and my transcendental mind/self are all parts of me. Accordingly, I can conclude that there is an entity that includes a body, an empirical self (which is an empirical mind), and a transcendental self (which is a transcendental mind) as its parts, namely, me. The argument for E-- outlined above centrally depends on G--, according to which Kantian outer appearances are essentially grounded in things in themselves. Strictly speaking, G-- does not need a justification because it merely spells out one aspect of how outer appearances are defined in Kant’s theory. But more can be said about Kant’s conception of experience in order to explain why and in what sense its intentional objects turn out to be essentially grounded in things in themselves. This explanation, already intimately familiar to us from our previous discussion, can be summarized in the form of the following argument. 1. The intentional objects of outer experience depend on outer perceptions, which include outer sensations as integral elements, in that outer perceptions provide the objects’ ‘matter’ and underwrite their existence. (Kant’s theory of outer experience.) 2. Due to its essentially passive nature, outer sense can produce outer sensations only in response to being materially transcendentally affected by things in themselves. (P---a.) 3. Kantian outer appearances, that is, the intentional objects of outer experience, are essentially grounded in things in themselves. (G--; from 1 and 2.) If we replace P-- in this argument with P--, according to which outer sense can produce outer sensations only in response to being materially transcendentally affected by things in themselves that are distinct from the thing in itself whose outer sense it is, we can derive corresponding stronger results, namely, in a first step, that, for each transcendental mind, each Kantian outer appearance is essentially grounded in things in themselves that are distinct from it, and based on this result and the public nature of outer appearances, in a second step, that all Kantian outer appearances are essentially grounded in things in themselves that are distinct from all human minds (G--). The same kind of argument can also be used to establish G-- by replacing P-- in line 2 of the above reconstruction with P-- and all occurrences of the term ‘outer’ with the term ‘inner’ in all other lines. The derivation of G-- proceeds in roughly the same way, except that we need to add a few more assumptions that are supported by Kant’s transcendental psychology.

310

       

1. The intentional objects of inner experience depend on inner perceptions, which include inner sensations as integral elements, in that inner perceptions provide the objects’ ‘matter’ and underwrite their existence. (Kant’s theory of inner experience.) 2. Due to its essentially passive nature, inner sense can produce inner sensations only in response to being materially transcendentally affected by the thing in itself whose inner sense it is. (P---a.) 3. Each particular inner sense necessarily supplies inner sensations for the constitution of only one intentional object of inner experience. (Additional assumption; from Kant’s transcendental psychology.) 4. No particular inner sense can belong to more than one thing in itself, and no thing in itself can have more than one particular inner sense. (U---; from Kant’s transcendental psychology.) 5. Each Kantian inner appearance, that is, each intentional object of inner experience, is essentially grounded in one unique thing in itself that is human and an ‘owner’ of a particular inner sense. (From 1, 2, 3, and 4 and the fact that the experience in question is human experience.) 6. A human thing in itself that is an ‘owner’ of a particular inner sense is a human transcendental mind. (Additional assumption; from Kant’s transcendental psychology.) 7. Each Kantian inner appearance, that is, each intentional object of inner experience, is essentially grounded in one thing in itself in that is (a) unique to the appearance and (b) a human transcendental mind. (G--; from 5 and 6.) E---b, the thesis that there is at least one human transcendental mind, thus turns out to be the only version of E whose demonstration does not rely on a version of P. P accordingly emerges as part of the backbone of Kant’s demonstration of almost all versions of E.¹³¹ Before tackling the question of how Kant might justify any or all of these versions of P, I want to make two observations about it. The first observation concerns an obvious worry that naturally arises with respect to it, namely, that prima facie it is unclear why the things that affect sensibility and thereby bring about sensations could not be appearances, that is, empirical objects. In response to this worry, recall that the kind of affections that are the topic of P are transcendental affections. So, the question whether, on Kant’s view, empirical objects causally interact with empirical selves at the empirical level of reality and empirically cause outer sensations has no bearing on the issue presently under discussion. The question in front of us, more explicitly put, is why we should assume that fully minddependent entities (empirical objects) are incapable of affecting a mind-independent

¹³¹ That P indeed plays a crucial role in establishing the existence of things in themselves in Kant’s mind is also illustrated by the following passage, GMS, 4:451: “Once this distinction [between appearances and things in themselves] has been made (if only through the noted difference between representations that are given to us from somewhere else and with respect to which we are passive, and those representations that we bring forth only out of ourselves and with respect to which we demonstrate our activity), it follows automatically that one has to admit and assume behind the appearances still something else that is not appearance, namely, things in themselves . . . ”

 ’    , ,  

311

cognitive faculty (sensibility).¹³² One straightforward reason for this view that Kant could put forward is that to assume otherwise would simply be implausible. I take it that this implausibility is part of what he has in mind when he says of the claim that “the representations of outer objects (appearances) cannot be external causes of the representations in our mind,” when presented as an objection to the view that the mind and the body can interact, that it would be “a completely senseless objection, because nobody would think of regarding what he has already acknowledged to be a mere representation as an external cause” (A390–391). In other words, the objection that the intentional objects of our representations cannot affect our mind would be an attack on a strawman because nobody in their right mind would ever think that they could affect it, since this thought is obviously absurd. One could go on and explain this absurdity by appeal to the fact that, since intentional objects only have those determinations that they are presented to have in the representations that ontologically specify them, their causal reach is necessarily limited to objects that are part of the same intentional ‘world.’ Frodo can causally influence other characters in the world of the Lord of the Rings, but he cannot causally influence Harry Potter, for example. And since sensibility is not an intentional object and thus not part of any intentional ‘world,’ no intentional object can affect it, including empirical objects. Another reason for rejecting the assumption that empirical objects can transcendentally affect sensibility comes to the fore when we take into account that the ‘matter’ of appearances is given to us and their existence is underwritten through the sensations that are produced by these affections. If empirical objects transcendentally affected sensibility, they would be involved in producing their very own ‘matter’ and in transcendentally conditioning their own existence, which would be viciously circular or at least amount to an unacceptable infinite regress.¹³³ We can avoid this circle and regress if we assume that the entities that transcendentally affect sensibility are mind-independent and thus do not depend on the outcome of these affections. This is what I understand Kant to be getting at when he says that

¹³² ‘Mind-independent’ is to be understood in the usual way, that is, as defined in section 2.3. As noted before, in another sense, sensibility is obviously mind-dependent namely, in the sense that it is a cognitive faculty of a mind. ¹³³ To see this, consider a given appearance and call it ‘A1.’ To assume that A1’s ‘matter’ is given to us through transcendental affections of sensibility by A1 itself is viciously circular. A1 cannot ontologically precede its very own ‘matter,’ which, after all, is essential to it. So, if the ‘matter’ of appearances is given to us through transcendental affections of sensibility by appearances, A1’s ‘matter’ must be given to us through transcendental affections of sensibility by another appearance, call it ‘A2.’ But since A2 is an appearance, its ‘matter’ must be given to us through transcendental affections of sensibility by appearances as well. To assume that A2’s ‘matter’ is given to us through transcendental affections of sensibility by A2 itself or by A1 is viciously circular. So, A2’s ‘matter’ must be given to us through transcendental affections of sensibility by another appearance, call it ‘A3,’ and so on ad infinitum. As is made clear by the passage from A251–252 that is quoted at the end of the present paragraph in the main text, Kant seems to regard both the circle and the infinite regress as unacceptable. (He there speaks of a “continual circle” that we want to avoid.) While it is obvious why the circle would be unacceptable, it is admittedly not as clear why the regress would be unacceptable. I take it that the unacceptability of the regress is due to the fact that our perceptions are supposed to not only provide the ‘matter’ of empirical objects but also underwrite their existence. In order for appearances to count as genuine existents, perceptions must ‘anchor’ them to the transcendental level of reality in some way. The indicated infinite regress, however, does not ever bottom out in any transcendentally real grounds, and thus must be rejected. Also recall, from our discussion in section 3.6, that the fact that empirical objects empirically affect our sense organs and thereby bring about sensations does not lead to a problematic circle. Neither our sensations nor the empirical causes of our sensations are among the empirical existence conditions of empirical objects.

312

       

. . . it follows naturally from the concept of an appearance in general: that something must correspond to it which in itself is not appearance because appearance can be nothing for itself and apart from our way of representing, and, thus, if we do not want to end up with a continual circle, the word appearance already indicates a relation to something whose immediate representation is but sensible but which also must be something in itself, also without this quality of our sensibility (on which the form of our intuition is based), i.e., an object that is independent from sensibility. (A251–252)

The second observation about P that I wish to make is that it can also be used as central premise in an alternative argument for the existence of things in themselves—that is, alternative to the argument sketched above—whose other main premise is considerably weaker than the claim that some empirical objects or some empirical selves exist, on which the earlier argument depends. On Kant’s view, our mind is essentially uncreative with respect to the ‘matter’ of any kind of intentional objects, not just the ‘matter’ of Kantian appearances. The faculty that provides the ‘matter’ for the constitution of other kinds of intentional objects—such as the objects of fictions, dreams, and hallucinations—is the imagination. Unlike sensibility, the imagination is not passive in that it does not have to be affected by something else to be set into operation, and it can even be productive and generate novel representations, for example, the representation of a unicorn wearing a fedora. But the imagination is essentially unoriginal in that in both its reproductions and its productions it is restricted to working with material that it has received from somewhere else, namely, from sensibility. The imagination (in other words) is either composing (productive) or merely recalling (reproductive). However, the productive imagination still is not creative on this account, namely, it is not able to produce a sensible representation that was never given to our sensible capacity, but one can always detect the material for it. . . . So, even though the imagination is a great artist, even a magician, it is still not creative but must take the material for its generations from the senses. (Anth, 7:167–168)¹³⁴

For example, in producing an intuitive representation of our stylish unicorn the imagination must rely on earlier perceptions of horses, horns, and fedoras. As a companion thesis to P about sensibility, we can thus record the following thesis about the imagination, which I will call ‘U.’

¹³⁴ See Anth, 7:178: “But the imagination is not as creative as one tends to claim.” See R341, 15:134: “The imagination is not productive with respect to sensations but merely with respect to intuitions.” Kant’s view that the representations of outer things by the imagination in dreams or hallucinations depend on earlier outer perceptions also finds expression in the following passages, see B278: “ . . . it does not follow that every intuitive representation of outer things at the same time includes their existence, for this representation might well be the mere effect of the imagination (in dreams as well as in insanity); but it is this only through the reproduction of former outer perceptions, which, as has been shown, is possible only through the reality of outer objects.” See R5653, 18:307: “Thus the possibility to represent things in space in intuition is grounded in the consciousness of a determination through other things, which signifies nothing more than my original passivity in which I am not active at all. That a dream brings about deception with respect to existents outside of me does not prove anything against that; for outer perceptions had to precede it regardless. It is impossible to originally get a representation of something as outside of me without being in fact passive.” See R314, 15:124: “The generation of representations is either passive or active. The former either through the objects (senses) or one representation through the other (phantasy and praesagium). The active one is never of matter but of form . . . ”

 ’    , ,  

313

(U) Due to its essentially unoriginal nature, the imagination can produce pseudo-sensations only by reproducing and recombining sensations. A pseudo-sensation is a representation that is exactly like a sensation except that it is not the direct result of an affection of sensibility by things in themselves. Sensations are produced by sensibility in response to transcendental affections; pseudo-sensations are produced by the imagination through the reproduction and recombination of previous sensations. The upshot of the unoriginality thesis is that we would not be able to have any representations of objects that have both ‘form’ and ‘matter’—not even dreams, hallucinations, or imaginings of any kind—if it were not for the affection of sensibility by things in themselves. So, then, here is a reconstruction of the promised alternative argument for E. 1. We have representations that include sensations or pseudo-sensations, be it perceptions, dreams, hallucinations, or imaginings. (Uncontroversial premise.) 2. Sensations are produced by sensibility; pseudo-sensations are produced by the imagination. (Kant’s theory of the human mind.) 3. Due to its essentially unoriginal nature, the imagination can produce pseudosensations only by reproducing and recombining sensations. (U.) 4. Due to its essentially passive nature, sensibility can produce sensations only in response to being transcendentally affected by things in themselves. (P.) 5. We can have representations that include sensations or pseudo-sensations, be it perceptions, dreams, hallucinations, or imaginings, only if there are things in themselves that transcendentally affect sensibility. (From 2, 3, and 4.) 6. There are things in themselves. (E, from 1 and 5.)¹³⁵ One might think that this argument for E does not present any real advantage over the argument sketched earlier, despite the fact that its first premise is weaker than the claim that some empirical objects or some empirical selves exist, on which the earlier argument depends. For the new argument includes an extra premise, namely, U, which incurs a new justificatory burden. But, as we will see presently, a closer examination of Kant’s ultimate ground for P reveals that this ground also supports U. This means that the dependence of the present argument on both of these theses does not offset the advantage compared to the earlier argument that is gained by the weaker first premise. For in depending on G, the earlier argument also depends on P and thus on its ground. P and, more generally, the view that our mind essentially incorporates a passive component on which all of our ‘matter’ providing representations depend, is one of the main load-bearing beams of Kant’s theoretical philosophy. Kant is often presented as the great synthesizer of rationalism and empiricism—with some qualifications defensibly so, in ¹³⁵ If we replace P in line 4 with P---a, every occurrence of ‘sensation’ by ‘outer sensation’ and every occurrence of ‘pseudo-sensation’ by ‘outer pseudo-sensation,’ we arrive at an argument for the claim that there is at least one thing in itself that is distinct from us. And if we replace P in line 4 with P---a, every occurrence of ‘sensation’ by ‘inner sensation’ and every occurrence of ‘pseudosensation’ by ‘inner pseudo-sensation,’ we arrive at an argument for the claim that there is at least one thing in itself that is a human transcendental mind.

314

       

my opinion—and I think it would be fair to say that the indicated view constitutes the central empiricist element of his theoretical philosophy. But with P we have not reached rock-bottom yet. On my reading, Kant conceives of the essentially passive nature of sensibility, as well as of the essentially unoriginal nature of the imagination, as grounded in an even more fundamental feature of the human mind. This feature is our mind’s finitude, more specifically, its finitude as expressed in an inability to be ontologically creative. We already had occasion (in sections 2.7.2 and 4.2.2.2) to look more closely at one other important way in which our mind’s finitude manifests itself on Kant’s view, namely, in the discursive nature of our intellect, which is to be contrasted with the intuitive nature of a divine, infinite intellect. We also already had occasion (in section 4.2.2.3) to indicate that the thesis of the ontologically uncreative nature of our mind plays an important role in Kant’s critical idealism. We are now ready to explicate this thesis and spell out in more detail what its important role consists in. The ontologically uncreative nature of our mind, which engenders the passive nature of our sensibility and the unoriginal nature of our imagination, is a second important way in which our mind’s finitude manifests itself.¹³⁶ This is in contrast to the ontologically creative nature of a divine infinite mind, which is purely active and original and does not stand in need of being given any material to bring forth representations or genuine existents. Our manner of intuiting, Kant says, is called sensible because it is not original, i.e., one through which the existence itself of the object of the intuition is given (and that, as far as we can see, can only pertain to the urbeing) but which depends on the existence of the object and thus is possible only through the affection of the capacity of representation of the subject by it. . . . . It is also not necessary that we restrict the manner of intuiting in space and time to the sensibility of human beings; it may be that all finite thinking beings must necessarily agree with human beings in this respect (although we cannot decide this), but it does not on this account cease to be sensibility, precisely because it is derived (intuitus derivativus), not original (intuitus originarius), and thus not intellectual intuition, which, for the reason just given, appears to pertain alone to the ur-being but never to a being that is dependent with respect to its existence as well as its intuition (which determines its existence in relation to given objects). (B72) If we had intellectual intuitions, our understanding would have to be creative and itself produce the things. Since this is not the case, the things must produce the representations in us, and [they must do] that through sensible intuition. The understanding then adds to experience nothing but the form. (V-Met/Mron, 29:880–881) Here we have to recall that we have before us the finite, not the infinite spirit. The finite spirit is one that does not get to work other than through passivity, that reaches the absolute only through limits, acts and forms only insofar as it receives material. (OP, 21:76)¹³⁷ ¹³⁶ Other commentators who stress the centrality in Kant’s philosophy of the claims that the human mind is finite, and that this finitude finds expression in, among other things, the passive nature of sensibility, include Hartmann 1875, 57; Heimsoeth 1924, 122–4; and Heidegger 1991 (*1929), 20–35, 243 6. ¹³⁷ See B145: “But from one element alone I could not abstract in the above proof after all, namely, from the fact that the manifold for the intuition must be given still prior to the synthesis of the understanding and independently from it; but how remains here undetermined. For if I wanted to conceive of an understanding that itself

 ’    , ,  

315

It is not surprising that Kant stresses these crucial defining features of his philosophy in the Opus Postumum, that is, that the minds on which the empirical realm depends on his view are finite minds, and that his idealism is an idealism of forms that are applied to ‘matter’ that must be given to these finite minds. For the Opus Postumum was written at a time when the idealist systems of his German Idealist successors were already on the rise, in which these restrictions, in one form or another, are softened or even thrown overboard. There is quite a bit more to say about the connection between the discursivity of the understanding and the passivity of sensibility, as well as about the connection between the purely active nature of a divine intellect and its ability to create by originally intuiting and thereby cognize things in themselves.¹³⁸ But for now we are mainly interested in the connection between the finitude of the human mind, on the one hand, and the passivity of (human) sensibility and the unoriginality of the (human) imagination, on the other hand, a connection that is not immediately intelligible on first glance. It is worth noting that there is a long tradition of philosophers, from Plato and Aristotle, over Augustine and Aquinas, to Leibniz, who conceive of finite beings as essentially having passive capacities and who associate pure activity with a divine infinite intellect. Of course, this observation does not amount to a justification for the assumption that being finite implies being passive in some way, much less for P and U, but it shows that, for somebody in this tradition, this assumption would have been familiar and natural. For Kant, the connecting link between the finitude of our mind and the particular kind of passivity of sensibility and the particular kind of unoriginality of our imagination that are featured in his theory is our mind’s inability to be ontologically creative. Creation ex nihilo is the prerogative of a divine, infinite intellect. Only a divine, infinite intellect has the power to generate objects that have some degree of reality from literally nothing, let alone objects that are genuine existents. The intentional objects of our dreams, hallucinations, and imaginings do not genuinely exist, but they are not nothing and thus have some degree of reality. Empirical objects, qua intentional objects of experience, do not only have some degree of reality, they even genuinely exist, namely, at the empirical level of reality. So, if our imagination could generate dreams, hallucinations, and imaginings without being dependent on the deliverances of suitable ‘matter’ from sensibility, that is, if it were original, or if sensibility could produce sensations on its own power without having to be affected by things in themselves, that is, if it were not passive, our mind’s abilities would rival some of the creative powers of the divine intellect—which, however, is impossible, given our finitude. This also explains why the existence of empirical objects, despite their full mind-dependence, must be ‘given’ to us, on Kant’s view. Their existence must be given to us because our mind, due to its finitude, lacks the power to create them ex nihilo, and thus depends for their generation on the ‘assistance’ of things in themselves that affect us. intuits (such as a divine understanding that does not represent given objects but through whose representation the objects themselves are at the same time given or created), the categories would not have any meaning at all with respect to such a cognition.” See R6041, 18:431: “The divine understanding is called the highest and pure understanding, which cognizes things absolutely, as they are in themselves. It is not sensibly conditioned. It is not receptivity but absolute spontaneity. It is intellectus originarius, not derivativus. Its cognitions are intuitions, not concepts, but not sensible intuitions but ideas that do not presuppose things but make them possible. Intellectus archetypus.” See OP, 21:90: “Transcendental Philosophy is not a science of objects that are a priori given to the subject by reason. For this would be self-creation of imagining. Rather, it is similar to a science of forms, under which objects would have to appear if they were given.” ¹³⁸ See Jauernig (in preparation) and Jauernig 2019.

316

       

The claim that, on account of its finitude, the human mind is incapable of being ontologically creative and, accordingly, cannot actively generate any ‘matter’ for the objects of our representations is Kant’s rock-bottom commitment that I promised we would uncover by thinking his reasoning in support of the claim that there are things in themselves through to its ultimate presuppositions. (U) Due to its essentially finite nature, the human mind is ontologically uncreative and thus incapable of actively or spontaneously generating any ‘matter’ for the objects of its representations. All weak versions of P, and thus of G, as well as E and U are underwritten by U. And as we saw in section 4.2.2.3, U is also one of the central assumptions on which Kant’s justification for the crucial second premise of part (A) of the master argument depends, according to which no genuine existent and no determination of a genuine existent can be intuited a priori. And since part (A) of the master argument is Kant’s main argument for the transcendental ideality of space and time, on which, in turn, all other theses of transcendental idealism and empirical realism depend in one form or another, U turns out to be one of the main foundations on which the entire edifice of Kant’s critical idealism rests. U itself does not need any justification in Kant’s eyes, I take it. From Kant’s point of view, it is not only obviously true that the human mind is finite but also that this finitude entails that the human mind is ontologically uncreative.¹³⁹ What about justificatory arguments for P--, P-, and G--? Although U does not entail P--, according to which outer sense can produce sensations only in response to being affected by things in themselves that are distinct from the thing in itself whose outer sense it is, it does lend some support to it. For there is not much difference with respect to creative power between a mind that actively brings about the ‘matter’ of empirical objects and a mind that brings about the ‘matter’ of empirical objects by affecting itself. If the former ability conflicts with the uncreative nature of our mind, it is plausible to think that the latter ability conflicts with it as well. A potential problem with this line of reasoning is that it might appear to prove too much. For the same kind of argument could be given for the claim that, due to the ontologically uncreative nature of our mind, the ‘matter’ of empirical selves cannot be the result of self-affections, a claim that is in direct conflict with P--. Or, put conversely, if my uncreativity is compatible with me constituting my empirical self without any assistance from things in themselves that are distinct from my transcendental mind, why would it not also be compatible with me constituting empirical objects without any assistance from things in themselves that are distinct from my transcendental mind? One possible response strategy

¹³⁹ But why assume that this uncreativity manifests itself in that we are incapable of actively generating any ‘matter’ for the objects of our representations? Why could it not manifest itself in that we cannot actively generate the ‘form’ of the objects of our representations, or their ‘form’ or their ‘matter’? The uncreativity of the human mind manifests itself in that, for all our representations of objects, some ingredient must be given to us; call this ingredient ‘matter.’ But it so happens that we have some a priori representations that refer to certain general features of all objects (as discussed in section 4.2.2.2). This means that, for all our representations of objects, we are capable of actively generating the ingredient that represents these general features; call this ingredient ‘form.’

 ’    , ,  

317

would be to try to make something of the fact that, despite falling in the same basic ontological category of transcendentally ideal but empirically real intentional objects of human experience, empirical selves and empirical objects still ontologically differ from one another in various important ways. For example, in contrast to empirical objects, empirical selves are to a large extent private and neither in space nor material. And one could use these differences—in a move similar to the central move in Descartes’ proof for the existence of God in the third Meditation—to argue that, while each transcendental mind has enough creative power to constitute an empirical self, no transcendental mind has enough creative power to constitute empirical objects, let alone the entire empirical world.¹⁴⁰ While these considerations do not amount to a strict demonstration of P--—among other things, it is difficult to see how one could conclusively prove the claims just mentioned about the extent of the limitation of our creative power—they still succeed in showing that, for anybody who is committed to U, endorsing P-- is more plausible than not endorsing it, even if this person is also committed to P--. There is not a whole lot that can be said in order to justify upgrading P-, the thesis that inner sense can produce inner sensations only in response to being materially transcendentally affected by thing in themselves, to P--, the thesis that the material transcendental affections by means of which inner sense can produce inner sensations are self-affections. One thing to say is that by stopping short of P-- we would allow several possibilities that seem intuitively implausible or even quite unsettling. These possibilities include that my transcendental self and my transcendental mind can come apart, that is, that the thing in itself that materially transcendentally affects my inner sense and appears as my empirical self can be distinct from the thing in itself that is the ‘owner’ of my inner sense and of all other transcendental cognitive faculties that are involved in the generation of my representations. Also allowed are the scenarios that the same one empirical self is grounded in multiple distinct transcendental selves, or that the same one transcendental self is involved in grounding multiple distinct empirical selves.¹⁴¹ There is no trace in Kant’s text that these possibilities ever occurred to him, let alone that he took them seriously.¹⁴² More generally, Kant seems to be firmly committed to the assumption that each human being has one unique empirical self and one unique transcendental self, which is its transcendental mind. Call this thesis ‘S--.’ The step from P-- to P-- can be seen as justified by S--, which we may thus record as included among the basic commitments on which the weight of the justification for all versions of critical idealism rests that comprise P-. ¹⁴⁰ Roughly put, Descartes’ move is to argue that human minds, as finite, do not have enough (formal) reality to cause the idea of God as an infinite being, which has infinite (objective) reality. The only being that has enough (formal) reality to cause the idea of God is God himself. So, given that we have the idea of God as an infinite being, we can infer that God exists. See Descartes (1649), AT, 7:34–52. ¹⁴¹ One might think that another such unwanted possibility is that inner sense can be materially transcendentally affected by things in themselves that are not minds of any kind. But note that if we were to assume that only minds have ‘insides’ and that inner sense can be materially transcendentally affected only by the ‘insides’ of things in themselves, we would not need to assume P-- to rule out that things in themselves other than minds can affect inner sense. See note 105. ¹⁴² These issues are rather tricky, but one might also argue that at least some of these possibilities would lead to serious problems when it comes to assigning moral responsibility, praise, and blame.

318

       

G--, according to which Kantian outer appearances are essentially grounded in things in themselves that are distinct from all human minds and God, depends not only on P-- but also on the assumption that, in the normal course of events, God does not affect outer sense. This assumption is a special case of the more general thesis that God limits his involvement in worldly creative matters to the creation of things in themselves but does not otherwise meddle in the constitution of the empirical realm, a thesis that I take Kant to endorse.¹⁴³ Call this thesis ‘F.’ F- can be seen as counter-balancing U to some extent. U expresses the human mind’s impotence when it comes to creation, which is a direct consequence of its finitude. But F- entails that the constitution of the empirical realm directly depends only on finite entities, centrally among them human minds, whose representations ontologically specify empirical objects and empirical selves. This feature of critical idealism—that the constitution of the empirical realm directly depends only on finite entities, which, as a ‘team,’ supply both the ‘form’ and the ‘matter’ of appearances—sets it apart from ordinary idealism before Kant and some notable German idealist systems after Kant, all of which, in various ways, present God, or an infinite spirit, as directly and centrally involved in the constitution of the empirical realm. Since no independent argument for F- seems to be forthcoming, we will put it on the record as included among the basic commitments on which the weight of the justification for all versions of critical idealism rests that comprise G--. All versions of critical idealism, including timid critical idealism, depend for their justification on U. The justification for bold critical idealism, in addition, rests on S-- and F-C. It appears futile to attempt a justification for these basic commitments, commitments that one could characterize as being among critical idealism’s constitutive principles, so to speak. I want to conclude this section by stressing that the unjustifiability of its constitutive principles by further arguments is neither a special problem for critical idealism nor a real problem at all. All arguments for substantive views must start from some substantive assumptions. So, all philosophical positions must incorporate some basic commitments that are not justifiable by any further arguments. But unless we adopt the implausible epistemic norm that, under any circumstances, only claims that can be justified by an argument of some kind are epistemically permissible, or reasonable, we are within our epistemic rights to choose the most basic assumptions on which to build our philosophical positions according to what seems most agreeable to us, provided these assumptions are internally consistent and cohere with the rest of all claims that we accept, all necessary truths, and the totality of the available empirical evidence. Whether to opt for timid critical idealism or bold critical idealism or a version whose strength lies somewhere between these extremes ultimately is a matter of taste. Kant seems to like bold critical idealism best, and I do not blame him.

¹⁴³ See V-Met-K2/Heinze, 28:759: “Cartesius says: God immediately brings about representations if, e.g., my eye moves. The third thing, namely, e.g., the eye, is then completely obsolete since God could bring about the representations also without the eye. Leibniz regards these representations as pre-established by God, that is not much better.”

6 Kantian Things in Themselves, Leibniz-Wolffian Things in Themselves, and Fictionalism 6.1 Realism about Kantian Things in Themselves As noted in section 2.1, the term ‘thing in itself,’ like the term ‘appearance,’ can be used in a general and a more specific sense. In the general sense, it means an entity that is mindindependent, as defined in section 2.3. In the more specific sense, it refers to Kantian things in themselves, that is, things in themselves as characterized in Kant’s critical idealism. Minimally, in addition to being mind-independent, Kantian things in themselves have all of the features that timid critical idealism attributes to them (as described in section 5.9.2), plus whatever features are implied by them. The first core tenet of transcendental idealism (TI1) informs us that things in themselves are distinct from empirical objects. The second core tenet of transcendental idealism (TI2) contributes that things in themselves are nonspatial and non-temporal. Together with the strong version of the second core tenet (TI2-pure-strong), according to which space and time are (nothing but) forms of sensibility, this entails that things in themselves are non-sensible, or supersensible. P and G allow us to add that things in themselves transcendentally affect sensibility, bring about sensations, and thereby ground appearances. Mind-independent things that have all of these features and whatever features follow from them are Kantian things in themselves. The claim that there is at least one Kantian thing in itself thus understood that grounds empirical objects is also part of timid critical idealism, as is the claim that there is at least one Kantian thing in itself that grounds an empirical self. The characterization of Kantian things in themselves contained in bold critical idealism goes beyond the characterization contained in timid critical idealism just explicated in that it also includes that all Kantian things in themselves are distinct from God and thus finite. Furthermore, bold critical idealism goes beyond timid critical idealism as far as their claims about Kantian things in themselves are concerned in that it also includes that each inner appearance is grounded in a unique Kantian thing in itself that is a human transcendental mind, and that all outer appearances are grounded in Kantian things in themselves that are distinct from all human minds. Finally, bold critical idealism holds that there is at least one Kantian thing in itself that grounds empirical objects and is not a human mind, and at least one Kantian thing in itself that grounds an empirical self and is a human transcendental mind. As indicated, I take it that bold critical idealism, not only is Kant’s personal favorite version of critical idealism and believed by him to be true, but also should be regarded as his official position. But since the final determination of which version of critical idealism deserves this honor must be postponed until after we will have examined his epistemology and theory of cognition in more detail, in the following, when I talk about ‘Kant’s conception of things in themselves’ or ‘Kantian things in themselves’ I should be understood to mean the conception that is featured in whatever version of critical idealism represents Kant’s official The World According to Kant: Appearances and Things in Themselves in Critical Idealism. Anja Jauernig, Oxford University Press (2021). © Anja Jauernig. DOI: 10.1093/oso/9780199695386.003.0006

320

-     

position; ditto for when I talk about ‘Kant’s claims about things in themselves.’¹ It is also useful to keep in mind that, in addition to the theses about things in themselves comprised in critical idealism, there is one more claim about them to which Kant is committed in the context of his theoretical philosophy, namely, the unknowability doctrine, delivered by his epistemology and theory of cognition, according to which all of our substantive theoretical cognition of a certain kind, namely, (what I am calling) cognition in the strict sense, is restricted to the empirical realm, which means that things in themselves are unknowable or uncognizable for us in the strict sense. If we do not want to attribute to Kant a rather obvious howler, we must assume that his conception of things in themselves and his claims about things in themselves do not amount to cognitions in the strict sense. I will adopt this assumption for the following discussion.² At the same time, it is important to reiterate that, although not cognitions in the strict sense, Kant’s claims about things in themselves are still justified. Moreover, it can be shown that there is a thing to which his conception of things in themselves applies, as we saw in section 5.10.2. This also means that Kant’s conception of things in themselves is demonstrably ‘objectively real,’ in Kant’s technical terminology, that is, it can be shown to represent a really possible thing.³ A second assumption that I will make for the following discussion is that, in Kant’s assessment, the claims about things in themselves that are included in critical idealism, together with any claims implied by them, plus the unknowability doctrine, exhaust, or mark the outer limits of, what can be justifiedly asserted about things in themselves, at least from a theoretical point of view. All other claims about them are mere speculation and epistemically impermissible.⁴ Since many readers will probably be of the opinion that the claims about things in themselves that I am attributing to Kant already go well beyond what is justifiable and permissible according to his epistemology, this assumption should not be controversial. The reading of Kant on things in themselves sketched in the previous two paragraphs is in tension with two other readings that might also seem to be supported by the text, namely, the fictionalist reading and what I will call ‘the Leibniz-Wolffian reading.’ On my view, Kant is neither a fictionalist nor a Leibniz-Wolffian about things in themselves. But ¹ Recall that timid and bold critical idealism mark the extreme ends of a whole spectrum of possible critical idealist views that vary in strength. See section 5.9.2. ² For a sketch of how to defend this claim, see section 5.10.1; for a full-scale defense, see Jauernig (in preparation). ³ A representation is objectively real if, and only if, it represents a really possible thing. See FM, 20:326: “One can call the first possibility logical possibility, the second real possibility; the proof of the latter is the proof of the objective reality of the concept, which one is always within one’s right to demand.” See ÜE, 8:191: “ . . . its objective reality [of the concept of a parabola], i.e., the possibility that there could be a thing of those said properties . . . ” See B270/A223: “ . . . an arbitrary combination of thoughts, which, although it does not contain a contradiction, cannot lay claim to objective reality, i.e., to the possibility of such an object as one wants to think here.” For the distinction between logical and real possibility, see note 36. One might argue that Kant also sometimes uses ‘objectively real’ in a stronger sense, namely, to characterize representations of actual things. Cf. V-Th/Pölitz, 28:1015–1016: “Now the question is whether this idea of ours also has objective reality, i.e., if there actually is a being to which our idea of God corresponds. . . . But to give this concept of mine objective reality such that I can demonstrate that there actually is an object that conforms to my concept, for that surely more is required than that there is nothing in my concept that contradicts each other.” Kant’s conception of things in themselves is objectively real and objectively real in the indicated stronger sense. For the record, there is yet another sense of objectively real, about which, however, we need not worry in the present context. A representation R of a cognizer is (what I call) reflectively objectively real if, and only if, the cognizer is able to justify the claim that R is objectively real; see B146, B412. Kant’s conception of things in themselves is also reflectively objectively real. ⁴ I defend this assumption in Jauernig (in preparation).

-       

321

I agree that these views play an important role in his thinking about things in themselves overall. The purpose of this chapter is to explain this role and to show how the textual evidence that seemingly supports these other interpretations can be reconciled with reading Kant as a realist about Kantian things in themselves in the way just laid out.

6.2 Leibniz-Wolffian Things in Themselves and the Odd Argument for Transcendental Idealism It seems to speak against the reading of Kant on things in themselves just summarized that he appears, not infrequently, to ascribe various other features to things in themselves that are not among the features listed in this summary and do not seem to be implied by them either. These other features include that (1) things in themselves are such that, if they are composite, the parts are ontologically prior to the whole,⁵ (2) they are simple or composed of simples,⁶ (3) they can be fully cognized in terms of pure concepts alone,⁷ (4-a) they have intrinsic determinations, and, more strongly, (4-b) they are individuated by intrinsic determinations alone,⁸ (5) the ‘realities’ or positive determinations possessed by them are only opposed to their respective negations but not to each other⁹—which also means that no thing in itself can possess a determination and its opposite at the same time (for nothing can simultaneously be F and not F)¹⁰—and, (6) they are such that ‘matter’ precedes ‘form,’ which means, among other things, that they are ontologically prior to their relations.¹¹ All of these features are part of what Kant takes to be the LeibnizWolffian orthodoxy on things in themselves.¹² Accordingly, I will refer to them as ‘Leibniz-Wolffian features’ and to the conception of things in themselves that includes these Leibniz-Wolffian features, plus mind-independence and whatever features follow from these, as the ‘Leibniz-Wolffian conception of things in themselves.’¹³ The most important implied feature is that Leibniz-Wolffian things in themselves are endowed with representational capacities, and thus are simple minds or souls. This follows from their being individuated by intrinsic determinations alone and the assumption that representations are the only possible truly intrinsic determinations of things, an

⁵ See MAN, 4:506; R6330, 18:651. ⁶ See B330/A274; B339/A283; B469/A441; MAN, 4:507–508; ÜE, 8:201, 209–210; V-Met/Volckman, 28:436; V-Met/Schön, 28:517–518. ⁷ Kant does not explicitly state this feature of things in themselves but he presupposes it in several of his discussions; see B319–336/A263–280; Prol, 4:285–286; MAN, 4:483–484. ⁸ See B66–67; B319–320/A263–264; B327–328/A271–272; B333/A277; B335–336/A279–280; B341/A285; BJ, 8:153–154; FM, 20:280, 282. ⁹ See B320–321/A264–65; B328–329/A272–273. ¹⁰ See FM, 20:282–283; V-Met/Schön, 28:487–488. Also see BJ, 8:154. ¹¹ See B322–324/A266–268, B331–333/A275–277. ¹² This comes out particularly clearly in the so-called amphiboly chapter in the Critique, to be discussed later in this section, as well as in MAN, 4:505–508, and FM, 20:281–285. ¹³ In this context, I am using ‘Leibniz-Wolffian’ as an intentionally vague umbrella term. The question of which Leibniz-Wolffians exactly Kant has in mind when he talks about different aspects of the Leibniz-Wolffian philosophy is surprisingly tricky and goes beyond the scope of this book, as does the related question of whether Kant’s characterization of the Leibniz-Wolffian position adequately captures any of the actual views of Leibniz, Wolff, or any of their many followers. For discussion, see Jauernig 2011. For the present, we are only interested in what Kant has to say about what he takes to be the Leibniz-Wolffian philosophy.

322

-     

assumption to which Kant takes the Leibniz-Wolffians to be committed and with which he agrees, at least if it is understood as an assumption about noumena.¹⁴ Kant relies on the indicated Leibniz-Wolffian features in a specific kind of argument for one of the core theses of transcendental idealism, an argument that occurs in different variations, more or less explicitly, at several places in his work. The thesis that is supported by the argument in question is TI1, the transcendental ideality of empirical objects, or, more precisely, TI1a, the claim that empirical objects are not things in themselves.¹⁵ The different versions of the argument have in common that they include a premise that asserts that things in themselves possess a certain Leibniz-Wolffian feature and another premise that denies that empirical objects have this feature (or a feature implied by this feature). Schematically, the core of this kind of argument can be stated as follows: 1. Things in themselves are F, where ‘F’ stands for a Leibniz-Wolffian feature. (LeibnizWolffian premise.) 2. Empirical objects are not F. (Empirical-object premise; supported by empirical evidence or other independent considerations.) 3. Empirical objects are not things in themselves. (From 1 and 2, and Leibniz’s law, according to which things that have different properties are numerically distinct.) Call this the ‘odd argument for transcendental idealism.’ The argument is odd, at least in the context of the critical philosophy because, while apparently endorsing the argument, Kant also holds that, from a theoretical point of view, the Leibniz-Wolffian premises that are featured in the first line of the various incarnations of the argument are not justifiable or epistemically permissible. In what follows, we will take a look at some instances of the odd argument in Kant’s text, which rely on the Leibniz-Wolffian features listed above, in the order (1)–(6).¹⁶ The point of this exercise is not to evaluate the merits of these arguments as arguments. In order to do that, we also would have to examine the cogency of the support for the various versions of (what I have called) the empirical-object premise, which would lead us too far afield. The

¹⁴ See B330/A274: “Substances in general must have something inner, which is thus free from all relations, and thus also from all composition. The simple is thus the foundation for the inner of things in themselves. The inner of their states can also not consist in place, figure, touch or motion (which determinations are all outer relations), and accordingly we cannot attribute any other inner state to substances than that through which we ourselves determine our inner sense, namely, the state of representations. In that way, monads came about which are supposed to make up the basic material of the entire universe but whose active force consists only in representations, through which they are active basically only in themselves.” See B321 322/A265 266: “But as an object of the understanding every substance must have inner determinations and powers that pertain to inner reality. But what inner accidents can I conceive of except those which my inner sense offers to myself? namely, that which is either itself a kind of thinking or analogous to it. For this reason, Leibniz made all substances, even all constituent parts of matter, after having in thought taken everything from them what might signify outer relation, and thus composition, into simple subjects that are endowed with capacities of representation, in one word, into monads, since he conceived of them as noumena.” ¹⁵ Kant usually presents the conclusion of the argument as saying that empirical objects are not things in themselves but only appearances, even though in most cases the argument, strictly speaking, only establishes the first part of this claim—unless, of course, we understand ‘appearance’ in a more general sense as meaning no more than ‘mind-dependent entity’ (as opposed to ‘fully mind-dependent entity’). ¹⁶ Or, more precisely, the following arguments make explicit use of features (2) to (6), in that order, but the defense for the claim that things in themselves possess feature (2) depends on the premise that they possess feature (1); see the following note. So, feature (1) is at least implicitly involved in the following arguments too.

-       

323

main point of the exercise is to illustrate the odd argument and see it in action in Kant’s text, so to speak. An especially straightforward example is the following piece of reasoning, which depends on the assumption that things in themselves are simple or composed of simples. Now the Critique shows . . . that in the world of bodies, as the totality of objects of outer senses, there are composite things everywhere, but the simple is not encountered in it. At the same time, it demonstrates that reason, if it thinks for itself a composite out of substances, as a thing in itself . . . , must inevitably think it as consisting of simple substances. . . . Accordingly, there then remains no other way out than to admit that bodies are not at all things in themselves and that their sensible representation to which we affix the name body is nothing but the appearance of something, which as thing in itself alone can contain the simple but remains completely uncognizable for us . . . (ÜE, 8:209–210)

This argument fits the structure of the odd argument quite nicely: 1. Things in themselves are either simple or composed of simples. (Leibniz-Wolffian premise.)¹⁷ 2. All bodies and, hence, all empirical objects are composite. (From empirical evidence.) 3. Composite empirical objects are not composed of simples. (From the infinite divisibility of matter, as established, for example, in MAN, 4:503.) 4. Empirical objects are neither simple nor composed of simples. (Empirical-object premise; from 2 and 3.) 5. Empirical objects are not things in themselves. (From 1 and 4, and Leibniz’s law.) A more fanciful incarnation of the odd argument appeals to the phenomenon of incongruent counterparts. Incongruent counterparts, you will recall, are pairs of objects that are similar in shape and equal in size but cannot be put into each other’s places by rigid motions.¹⁸

¹⁷ On Kant’s view, the premise that things in themselves are simple or composed of simples follows from the premise that if things in themselves are composite, then their parts are ontologically prior to the whole, plus a few other assumptions, such as the claim that actual infinities are impossible. That is, the ascription of feature (2) to things in themselves follows from the ascription of feature (1) to them. See the textual evidence referenced in note 5. For a helpful discussion of the different arguments for the existence of simples proposed by Leibniz, Wolff, Baumgarten, and the pre-critical Kant, some of which are reflected in Kant’s critical discussion, see Watkins 2006. Also see Jauernig (forthcoming 2021b). ¹⁸ Incongruent counterparts appear at several different places in Kant’s writings throughout his career and, to the consternation of many commentators, are used to argue for a variety of different conclusions. The main arguments involving incongruent counterparts are to be found in the essay “On the First Ground of the Distinction of Regions in Space” (1768), the Inaugural Dissertation (1770), and the Prolegomena (1783). In the ‘Regions’ essay, incongruent counterparts provide the basis for an attack on Leibniz-Wolffian relationalism about space and a defense of the existence of substantival space, as briefly discussed in section 3.4.4; see GUGR, 2:377–383. In the Inaugural Dissertation, incongruent counterparts are used to illustrate the thesis that our representation of space is an a priori intuition as opposed to a concept; see MSI, 2:403. In the Prolegomena, incongruent counterparts are employed in versions of the odd argument for transcendental idealism, to be discussed presently, an argument to which Kant refers back in the Metaphysical Foundations of Natural Science (1786); see MAN, 4:484.

324

-     

What can be more similar and equal in all respects to my hand or my ear than their image in a mirror? And yet I cannot put such a hand as is seen in the mirror in the place of its original; for if this was a right hand, then that in the glass is a left one, and the image of the right ear is a left one, which never can take the place of the other. In this case there are no inner differences that any kind of understanding could merely think; and still the differences are inner as far as the senses teach, for the left hand cannot be enclosed within the same boundaries as the right one (they are not congruent), notwithstanding their mutual equality and similarity. The glove of the one hand cannot be used on the other. What is the solution? These objects are not representations of things as they are in themselves and as the pure understanding would cognize them, but they are sensible intuitions, i.e., appearances, whose possibility rests on the relation of certain things that are unknown in themselves to something else, namely, our sensibility. (Prol, 4:286)

Call an object whose shape is such that it can have an incongruent counterpart a ‘handed’ object. The argument turns, first, on the premise that things in themselves can be fully cognized in terms of concepts alone, which is underwritten by feature (3) of the LeibnizWolffian conception of things in themselves, and, second, on the premise that handed objects cannot be fully cognized in terms of concepts alone, which is illustrated by pairs of incongruent counterparts, such as a left and a right hand.¹⁹ The conceptual representation of a left hand is the same as the conceptual representation of a right hand; it includes, for example, the characterization that the hand comprises five fingers, one of them a thumb, that it is naturally attached to an arm, and that the distance from the tip of the thumb to its base is a certain fraction of the distance from the tip of the middle finger to its base. (This is assuming that there are no primitive properties of leftness and rightness, which we can pick out by the concepts ‘left’ and ‘right,’ which is an assumption that Kant quite clearly endorses.) But a right hand is inherently different from a left hand, as is shown by the fact that one cannot be put in the place of the other. This means that the conceptual representations of the hands leave something out and thus are not complete. Therefore, handed objects are not things in themselves. This does not quite show yet that empirical objects in general are not things in themselves. But if we assume that handed empirical objects have the same ontological status as non-handed empirical objects, we can draw the more general conclusion. More formally, the argument can be stated as follows: 1. Things in themselves can be fully cognized in terms of concepts alone. (LeibnizWolffian premise.) 2. Handed objects cannot be fully cognized in terms of concepts alone. (Premise supported by reflection on the incongruent counterpart phenomenon.) 3. Many empirical objects are handed. (From empirical evidence.) 4. Many empirical objects cannot be fully cognized in terms of concepts alone. (Empirical-object premise; from 2 and 3.)

¹⁹ According to Kant’s reading, the Leibniz-Wolffians do not only hold that things in themselves can be fully cognized in terms of concepts alone, but also that they can be fully cognized in terms of pure concepts alone—if not by us, then at least by God. But the former, weaker claim is all that we need for the present argument.

-       

325

5. Many empirical objects are not things in themselves. (From 1 and 4, and Leibniz’s law.) 6. All empirical objects have the same ontological status. (Uncontroversial Premise.) 7. Empirical objects are not things in themselves. (From 5 and 6.) The chapter in the Critique in which the line of reasoning that underwrites the odd argument is in evidence most prominently is the chapter on the amphiboly of the concepts of reflection, even though the argument itself is only implicit in Kant’s discussion there. The amphiboly chapter contains Kant’s diagnosis of a certain kind of mistake made by the Leibniz-Wolffians that originates in their confusion of phenomena with noumena, the latter of which consists in that, on account of their misconception of the relation between sensible and intellectual representations, they take phenomena to be confusedly perceived noumena, as already reported in sections 3.5 and 5.3. Kant uses the term ‘noumenon’ in several distinct senses. The main sense is the sense in which it is employed in the amphiboly chapter. Noumena in this sense are “objects of the pure understanding,” as Kant often puts it, that is, objects that can be cognized by the pure understanding alone or in terms of pure concepts alone, that is, objects whose determinations can be justifiably ascertained solely by means of logical reflections (broadly speaking) on pure concepts, their relations, and their function in thought.²⁰ In the following, this is how ‘noumenon’ should be understood unless indicated otherwise. Kant also sometimes uses the Latin term ‘intelligibilia’ instead of the Greek ‘noumena’ in this sense. The Leibniz-Wolffians believe that objects can be cognized in terms of pure concepts alone if, and only if, they are mindindependent, and, thus, that noumena are things in themselves and vice versa. So, by confusing phenomena with noumena, the Leibniz-Wolffians also confuse phenomena or appearances with things in themselves. Accordingly, even though the Leibniz-Wolffians correctly regard empirical objects as phenomena, they still end up with the erroneous transcendental realist view that empirical objects are things in themselves, albeit confusedly perceived ones. The mistake at issue in the amphiboly chapter consists in that, due to their conception of phenomena as confusedly represented noumena, they take certain metaphysical principles that are derived on the basis of pure concepts alone to be valid for phenomena and, thus, empirical objects, even though they are valid only for noumena. These principles include (i) the principle of the identity of intrinsic indiscernibles, according to which things are numerically identical if, and only if, all of their intrinsic properties are the same; (ii-a) the principle of the non-opposition of realities or positive properties, according to which two properties are opposed to each other if, and only if, one is the negation, or contradictory opposite, of the other; and (ii-b) the closely related principle that any thing whose determinations are opposed to each other is impossible; (iii-a) the principle that substances are, or are composed of, simples; and (iii-b) the principle that every substance must have some intrinsic determinations;²¹ and, finally,

²⁰ For this sense of ‘logical reflection,’ see B335/A279: “If we reflect merely logically, we merely compare our concepts with one another in the understanding, whether both contain the same, whether they contradict each other or not, whether something is inwardly contained in the concept or pertains to it, and which one of two is given and which one is supposed to count as a manner to think the given one.” Also recall that pure concepts are a priori, purely intellectual, non-sensible, unschematized concepts. See note 97, chapter 2. ²¹ In the assessment that the ‘inner’–’outer’ comparison yields these two quite different principles, I am in agreement with James Van Cleve, see Van Cleve 1988, esp. 232–3. But while I agree with Van Cleve that Kant was

326

-     

(iv-a) the principle that ‘matter’ is ontologically prior to ‘form’; and, more specifically, (iv-b) the principle that things are ontologically prior to space and time. According to Kant’s analysis, the indicated principles are based on so-called concepts of reflection, namely, the concepts of identity and diversity, harmony and opposition, inner and outer, and matter and form, which govern the comparison between things or their properties, by specifying and explicating how, or in what respect, the comparison is to be performed. For example, we could compare things with respect to identity and diversity, that is, we could ask whether they are identical or distinct or we could compare their determinations with respect to harmony or opposition, that is, we could ask whether these determinations are ‘opposed’ to each other or not. The problem is that, due to their failure to see that phenomena and noumena are fundamentally different kinds of entities, the LeibnizWolffians do not to realize that these concepts of reflection are ambiguous, or ‘amphibolous,’ and have different meanings depending on which kind of entities we are talking about. For example, the identity conditions for noumena differ radically from the identity conditions of phenomena; for noumena to be identical accordingly means something quite different from what it means for phenomena to be identical. Since they are unaware of this ambiguity, and since they take phenomena to be confusedly represented noumena, the Leibniz-Wolffians mistakenly believe that their metaphysical principles, which are derived by thinking about the comparison of things or their properties based on the concepts of reflection understood as it is appropriate for noumena, are universally valid and hold not only for noumena but also for phenomena.²² The question of how exactly Kant understands the derivations of these metaphysical principles is quite tricky and cannot be discussed in detail here.²³ Sometimes he apparently wants to say that the features of noumena can be directly read off the features of pure concepts. This is suggested by the formulation that “Leibniz compared all things to one another merely through concepts and found, as could naturally be expected, no other differences than those through which the understanding distinguishes its concepts from one another” (B326/A270).²⁴ For example, applying this method, one could derive the principle that noumena are individuated by their intrinsic determinations from the fact that pure concepts are individuated by the non-relational concepts contained in them.²⁵ At other times, the crucial assumption on which the derivations rely seems to be that whatever cannot be represented in terms of pure concepts alone cannot be true of noumena. For example, applying this method, the (alleged) fact that the only kind of opposition that can be represented in terms of pure concepts alone is the contradictory kind that obtains

not as careful as one could wish in clearly distinguishing them, I would not go as far as to say that he conflates them. For Kant, they are two manifestations of the same kind of reflective comparison. ²² See B326/A270: “ . . . deceived by the amphiboly of the concepts of reflection, the famous Leibniz created an intellectual system of the world or rather thought that he cognized the inner quality of things in that he compared all objects only by means of the understanding and the abstract formal concepts of its thinking. . . . He compared all things to one another merely through concepts and found, as could naturally be expected, no other differences than those through which the understanding distinguishes its concepts from one another. The conditions of sensible intuition, which carry their own differences with them, he did not regard as original; for sensibility was for him only a confused kind of representing and no special source for representations; appearance was for him the representation of the thing in itself, although distinguished from the cognition through the understanding according to its logical form. . . . Thus, Leibniz compared the objects of sense with each other only in the understanding as things in general.” ²³ For a detailed discussion, see Jauernig (in preparation). ²⁴ Also see B335/A279. ²⁵ See B328/A272.

-       

327

between a concept and its negation, no other opposition of determinations of noumena is possible than the opposition of a determination and its negation. The representation of “really repugnant” determinations, that is, determinations whose effects cancel each other either completely or in part, such as two moving forces that act in opposite directions, requires the representation of directions, which (Kant claims) we can accomplish only in intuitive terms.²⁶ For our present purposes, the main point to take away is that, on Kant’s reading, the Leibniz-Wolffian metaphysical principles are all derived based on pure concepts alone and thus apply only to objects that can be cognized in terms of pure concepts alone, that is, noumena. Turning to the specific mistakes with which Kant charges the Leibniz-Wolffians, here is his description of the mistake that arises in connection with the comparison of things with respect to their identity or diversity: Leibniz took appearances for things in themselves and thus for intelligibilia, i.e., objects of the pure understanding (although he called them phenomena due to the confusion of their representations), and there his principle of the indiscernible (principium identitas indiscernibilium) could indeed not be disputed; but since they are objects of sensibility, and the understanding is not of pure but merely empirical use with respect to them, the multiplicity and numerical distinctness is already indicated by space itself as the condition of outer appearances. (B320/A264)²⁷

Kant characterizes the mistake in the case of the comparison of determinations with respect to their harmony or opposition as follows: If reality is represented by the pure understanding only (realitas noumenon), no opposition between realities can be thought. . . . By contrast, the real in the appearance (realitas phenomenon) can indeed be opposed to one another, and united in the same subject one can destroy the effect of the other completely or partly, such as two moving forces in the same straight line insofar as they pull or push a point in opposite direction . . . (B320–321/A264–265)²⁸

The two closely related Leibniz-Wolffian mistakes concerning the comparison of what counts as ‘inner’ or ‘outer’ with respect to a substance are diagnosed as follows: With respect to an object of the pure understanding only that is inner which has no relation at all (as far as its existence is concerned) to something that is distinct from it. By contrast, the inner determinations of a substantia phaenomenon in space are nothing but relations, and it itself is entirely a sum total of relations. (B321/A265)²⁹

²⁶ See B329/A273; B337–342/A280–A286. Kant started to think about how to represent real (as opposed to merely logical) oppositions very early in his career. In the 1763 essay “Attempt to Introduce Negative Magnitudes in Metaphysics,” he suggests to represent them in terms of pairs of positive and negative magnitudes; see NG, 2:165 204. After thinking more about the metaphysical implications of incongruent counterparts, from 1768 onwards, he settled on the view that the representation of directions and directionality more generally, which is required to represent real oppositions, ultimately depends on intuition; see MSI, 2:403. ²⁷ See B327–328/A271–272; B337–338/A281–282; V-Met-L2/Pölitz, 28:569–570. ²⁸ See B328–329/A272–273; B338/A282; R5578, 18:238. ²⁹ See B330–331/A274–275; B339–342/A283–286; V-Met/Mron, 29:930. Also compare B67.

328

-     

Finally, here is how Kant presents the error of the Leibniz-Wolffians that arises in connection with comparisons of things with respect to whether they are ‘matter’ or ‘form’: For the understanding demands that something is given . . . in order to be able to determine it in some way. Therefore, in the concept of the pure understanding matter precedes form, and, for this reason, Leibniz first assumed things (monads) and inwardly their power of representation, in order to later base on this their outer relations and the community of their states (namely, of the representations). . . . That is how it, in fact, would have to be if the pure understanding could be immediately related to objects and if space and time were determinations of things in themselves. But if they are merely sensible intuitions, in which we determine all objects merely as appearances, the form of intuition (as a subjective quality of sensibility) precedes all matter (the sensations), and thus space and time precedes all appearances and all data of experience and rather makes them possible in the first place. (B323/A267)³⁰

These passages do not contain explicit examples of the odd argument, but the situation described by them can easily be re-described in the mold of the odd argument. The presentation in the amphiboly chapter highlights that, since things in themselves, on the one hand, and appearances and, thus, empirical objects, on the other hand, are fundamentally different kinds of entities, the Leibniz-Wolffian principles derived for the former cannot simply be assumed to hold for the latter, contrary to what the Leibniz-Wolffians think. And, indeed, it turns out that the Leibniz-Wolffian principles do not apply to empirical objects. Another way of describing the same situation is to say that, since the Leibniz-Wolffian principles derived for things in themselves do not apply to empirical objects, empirical objects are not things in themselves, not even confusedly represented ones, contrary to what the Leibniz-Wolffians think. Reformulated in this way, the argument of the first passage can be presented as follows: 1. Things in themselves are numerically identical if, and only if, all of their intrinsic properties are the same. (Principle of the identity of intrinsic indiscernibles, LeibnizWolffian premise.) 2. There can be numerically distinct empirical objects all of whose intrinsic properties are the same, namely, objects that are located at different positions. (Empirical-object premise; supported by the thought experiment of two intrinsically indiscernible water droplets at different positions.³¹) 3. Empirical objects are not things in themselves. (From 1 and 2, and Leibniz’s law.) Here is a possible reconstruction of the reasoning expressed in the second passage, cast in the form of the odd argument: 1. No thing in itself can have positive determinations that are opposed to each other. (Leibniz-Wolffian premise.)

³⁰ See B331–333/A275–278.

³¹ See B328/272.

-       

329

2. Each empirical object can have positive determinations that are opposed to each other, namely, determinations that are really repugnant, such as two moving forces that act in opposite directions. (Empirical-object premise; from empirical evidence.) 3. Empirical objects are not things in themselves. (From 1 and 2, and Leibniz’s law.) The third passage contains the materials for two versions of the odd argument, each one of which addresses one of the two metaphysical principles that the Leibniz-Wolffians draw from the ‘inner’–‘outer’ comparison. One of these arguments, which has as its LeibnizWolffian premise the principle that things in themselves are simple or composed of simples, is already familiar to us—it is the first incarnation of the odd argument that we looked at above—and so there is no need to reproduce it here again. The other one goes like this: 1. Things in themselves have intrinsic determinations. (Leibniz-Wolffian premise.) 2. Empirical objects have only extrinsic determinations. (Empirical-object premise.) 3. Empirical objects are not things in themselves. (From 1 and 2, and Leibniz’s law.) Finally, the version of the odd argument implicit in passage number four can be formulated as follows: 1. Things in themselves are ontologically prior to space and time. (Leibniz-Wolffian premise) 2. Space and time are ontologically prior to empirical objects. (Independent premise; from independent considerations, for example, an argument from incongruent counterparts such as the one in the ‘Regions’ essay.³²) 3. Empirical objects are not things in themselves (From 1 and 2, and Leibniz’s law.) One can raise various questions about all of these arguments, including questions about their soundness, as well as questions about my particular reconstructions of them. Interesting as these questions may be, they are not our present concern. All we need to take away from the foregoing discussion for present purposes is that the critical Kant offers a lot of arguments for a central thesis of transcendental idealism that rely on Leibniz-Wolffian premises concerning things in themselves. Confronted with this situation, one might be tempted to conclude that he endorses these LeibnizWolffian premises, and that his own conception of things in themselves includes the Leibniz-Wolffian features that figure in these arguments. Call this the ‘LeibnizWolffian reading.’³³

³² See GUGR, 2:377–378, 381–383, note 18, and the discussion in section 3.4.4. ³³ Note that (a) several commentators do conclude as much or, at least, they conclude that Kant’s conception of things in themselves includes some of the indicated Leibniz-Wolffian features, such as the feature that things in themselves have intrinsic properties, and (b) they conclude this in no small part based on Kant’s discussion in the amphiboly chapter. These commentators include Langton 1998, esp. 16, 20, 33–8, and Allais 2015, see esp. 234 43.

330

-     

6.3 Fictionalism about Things in Themselves According to the fictionalist reading, Kant’s critical idealism does not contain the claim that things in themselves actually exist. Rather, Kant regards the concept of a thing in itself as a kind of useful fiction that serves an explanatory function in our account of the world and our experience, as well as a need implicit in our epistemic aspirations, but we do not know whether there are any actual or really possible things that fall under the concept, or, more strongly, we know that there are no actual or really possible things that fall under it. Proponents of the reading that Kant is a fictionalist about things in themselves support their interpretation primarily by means of systematic considerations, in particular, the consideration that their reading provides an elegant solution to the problem of existence, discussed in section 5.10.1, which consists in the worry that by asserting the existence of things in themselves Kant appears to be in violation of both the unknowability doctrine and the result that the category of actuality meaningfully applies only to appearances and object of possible experience but not to things in themselves. On the fictionalist reading, Kant does not assert the existence of things in themselves, and so the problem of existence does not even arise. As far as textual support is concerned, the fictionalist interpreters tend to focus mainly on two areas in the Critique: first, on Kant’s remarks about noumena in the chapter on the distinction between phenomena and noumena that immediately precedes the amphiboly chapter at the very end of the Transcendental Analytic,³⁴ and, second, on his discussion of the ideas of reason in the Transcendental Dialectic and the Transcendental Doctrine of method.³⁵ Here is a representative passage from the phenomena-noumena chapter that may be put forward in support of the fictionalist interpretation: I call a concept problematic that contains no contradiction, that is also connected to other cognitions as a limitation of given concepts but whose objective reality cannot be cognized in any way. The concept of a noumenon, i.e., of a thing that is supposed to be thought not at all as an object of the senses but as a thing in itself (merely through the pure understanding) is not at all contradictory; for one cannot claim of sensibility that it is the only possible kind of intuition. Moreover, this concept is necessary so as to not extend sensible intuition to things in themselves and thus to limit the objective validity of sensible cognition (for the others, which intuition does not reach, are called noumena precisely so ³⁴ See Rescher 1981; Pogge 1991, esp. 506–7. ³⁵ See Vaihinger 1918, esp. 618–39; Schaper 1966, 241. Vaihinger argues that, in the Critique, Kant wavered between understanding things in themselves as fictional objects and as hypothetical objects. He started out conceiving of them in fictionalist terms, but since he made use of this fiction in the construction of his own theory, he unwittingly transformed it into a hypothesis, which he neglected to discharge once he was done. See Vaihinger 1918, 109–14, 118–19, 266–73. (On Kant’s view, the difference between a fiction and a hypothesis is that, in the case of the latter, we must know at least that its object is really possible. See, for example, B798–800/ A770–772.) Vaihinger claims that in the Opus Postumum Kant finally clearly recognized things in themselves for what they really are, namely, fictional objects. See Vaihinger 1918, 722–33. (On Vaihinger’s reading, things in themselves are not the only fictional objects in Kant’s philosophy. He sees them everywhere: in Kant’s account of mathematics, nature, ethics, right, beauty, religion . . . .) Vaihinger’s claim that in the Opus Postumum Kant emerges as a clear fictionalist about things in themselves is sharply criticized by Erich Adickes in Adickes 1920, which led to a disturbingly harsh exchange. See Vaihinger 1921, Adickes 1927, Schmid 1928, Adickes 1928. Adickes’s all-out assault on Vaihinger’s fictionalist interpretation of Kant, in Adickes 1927, is highly recommended reading.

    

331

that it is indicated thereby that these cognitions cannot extend their domain over everything that the understanding thinks.) But in the end the possibility of such noumena is not at all intelligible, and the extension outside of the sphere of appearances is (for us) empty . . . (B310)

Kant might be taken to say here that things in themselves are noumena, and that noumena are fictional in the sense that, although they are logically possible, we do not know whether they are really possible.³⁶ Relatedly, one might add that the apparent identification of things in themselves with noumena speaks for the fictionalist interpretation also in that Kant’s standard German term for a fictional object is ‘Gedankending,’ which literally means ‘object of thought,’ and one way of translating the Greek ‘noumenon’ into German would be as ‘Gedankending.’³⁷ Relatedly, Kant also uses ‘ens rationis’ to for fictional objects,³⁸ which is a possible Latin translation of ‘noumenon.’ Turning to the ideas of reason, in the Transcendental Dialectic Kant explains that we are subject to several “natural and inevitable illusions,” which arise as a result of our taking certain subjective rules of the use of reason for objective principles, that is,, for principles that apply to objects.³⁹ More specifically but without going into too much detail, the subjective rule of reason to always look for the ‘unconditioned’ with respect to anything conditioned that is given, together with the three main forms of syllogism of reason—the categorical, the hypothetical, and the disjunctive—is said to give rise to three kinds of concepts of reason, which, if understood as applying to objects, yield three classes of ideas.⁴⁰ First, the so-called psychological ideas, which concern the unconditioned unity of the thinking subject and, roughly, add up to the traditional idea of a soul;⁴¹ second, the cosmological ideas, which concern the unconditioned unity or totality of the series of conditions for empirical objects, or, in other words, different aspects of the empirical world as whole;⁴² and third, the theological idea or ideal, which concerns the unconditioned unity of the conditions for all objects of thinking in general and amounts to the idea of an omnitudo realitatis or a most real being, which is conceived of as the foundation for the real possibility of all objects.⁴³ With respect to the cosmological ideas, it is to be noted that they have the peculiarity that there are two specific ways in which the ³⁶ For the distinction between logical possibility and real possibility, see B624/A596, note: “The concept is always possible if it does not contradict itself. That is the logical mark of possibility, whereby its object is distinguished from the nihil negativum. Nevertheless, it can still be an empty concept if the objective reality of the synthesis through which the concept is generated is not specifically demonstrated. . . . This is a warning not directly to infer from the possibility of concepts (logical) to the possibility of things (real).” There is a lot to say about how to spell out the distinction between logical and real possibility for the case of supersensible things in more detail. For our present purposes, it is enough, however, to register that there is such a distinction, and that being really possible entails being logically possible but not vice versa. ³⁷ See B594/A566; B517/A489; Prol, 4:295–296; MS, 6:241, 292, 338, 442. ³⁸ See B347–349/A290–292; B394/A337; R5724, 18:336. ³⁹ See B349–355/A293–298; B670–672/A642–644. ⁴⁰ See B359–366/A303–309; B377–99/A321–333. Kant calls the explanation that and how the ideas of reason naturally and inevitably arise for us their ‘subjective derivation.’ See B393/A336: “Of these transcendental ideas no objective deduction is possible, properly speaking, such as we were able to provide for the categories. For they indeed have no relation at all to any object which could be given congruently to them, precisely because they are only ideas. But we could attempt a subjective derivation, and this also has been accomplished in the present chapter.” Also see B397/A339: “At least the transcendental (subjective) reality of the pure concepts of reason rests on the fact that we are led to such ideas through a necessary inference of reason.” ⁴¹ See B399–403/A341–345. ⁴² See B433–448/A406–420. ⁴³ See B599–611/A571–583.

332

-     

unconditioned with respect to the series of the conditions for empirical objects can be thought. First, one can regard certain elements of the series as unconditioned, in which case the series is understood to begin in these unconditioned elements. Second, one can regard the entire series of the conditions as unconditioned, in which case the series is understood to be infinite. This is why the cosmological ideas come in pairs, and why they give rise to antinomies. The theses of the antinomies assert the existence of the object of the relevant idea of the unconditioned conceived of in the first manner, while the antitheses assert the existence of the object of the relevant idea of the unconditioned conceived of in the second manner.⁴⁴ More specifically, the cosmological ideas include (1) the idea of the unconditioned with respect to the temporal and spatial conditions for empirical objects, an unconditioned that can be conceived of as (a) a beginning of the empirical world in time and a boundary of the empirical world in space or (b) an infinite duration and an infinite extension of the empirical world; (2) the idea of the unconditioned with respect to the material conditions for empirical objects, an unconditioned that can be conceived of as (a) a simple part of matter or (b) the totality of the series of smaller and smaller parts generated by dividing matter ad infinitum; (3) the idea of the unconditioned with respect to the causal conditions for empirical objects and events, an unconditioned that can be conceived of as (a) a free cause or (b) the totality of an infinite series of effects and their determined causes; and (4) the idea of the unconditioned with respect to the contingent existence of everything changeable in the empirical world, an unconditioned that can be conceived of as (a) a necessary being that is a part of or a cause of the empirical world or (b) the totality of an infinite series of contingently existing objects and events that depend for their existence on other contingently existing objects and events.⁴⁵ With respect to the theological idea, Kant adds that reason naturally combines the cosmological idea of a necessary being that is the cause of the world with the ideal, that is, the idea of a most real being, which yields the idea of God of traditional theology as a necessary and supremely intelligent (and supremely good and supremely powerful) being who creates the world according to a plan and is responsible for its systematic unity and purposeful arrangement.⁴⁶ Kant devotes the main bulk of the Transcendental Dialectic to showing that the arguments that can be put forward or traditionally have been put forward for the existence of the objects of these ideas are fallacious or based on untenable presuppositions (or both), and that taking them to be referring concepts, or concepts that apply to genuine existents, is thus illusory, at least in a theoretical context. He shows this for the psychological ideas in the paralogism chapter,⁴⁷ for the cosmological ideas in the antinomy chapter,⁴⁸ and for the ⁴⁴ See B445–46/A417–18: “This unconditioned can be thought either as consisting merely in the entire series, in which thus all elements without exception would be conditioned and only the whole of them would be absolutely unconditioned, and then the regress is called infinite; or the absolutely unconditioned is only a part of the series to which the other elements are subordinated but which itself does not stand under any other condition. In the first case, the series is without limits a parte priori (without beginning), i.e., infinite, and still entirely given, but the regress in it is never completed and can merely be called potentially infinite. In the second case, there is a first in the series, which with respect to past time is called the world-beginning, with respect to space the worldboundary, with respect to the parts of a whole that is given within its limits the simple, with respect to causes the absolute self-activity (freedom), with respect to the existence of mutable things the absolute necessity of nature.” ⁴⁵ See B454–488/A426–460. ⁴⁶ See B608/A580; B614–615/A586–587. ⁴⁷ See B406–431; A348–405. ⁴⁸ See B454–488/A426–460; B513–595/A485–567. As noted previously, Kant’s solution to the dynamical antinomies (that is, the third and fourth one) is to argue that on his transcendental idealist view, with its distinction between the empirical and the supersensible realm, the thesis and the antithesis turn out not to be contradictory because they can both be true. See note 171, chapter 2, and note 2, chapter 3. Nevertheless, Kant still

    

333

theological idea in the chapter on the ideal.⁴⁹ But in the appendix to the Dialectic he goes on to argue that there still is a legitimate use for these ideas after all, namely, a regulative use that is aimed at the expansion, unification, and systematization of the cognitions of empirical objects provided by the understanding.⁵⁰ This regulative use consists in our adopting the maxim to proceed in our investigations of the empirical realm as if the objects of these ideas existed. That is, the proper use of the psychological ideas in our empirical investigations of thinking nature is for us to proceed as if all inner states and actions of our mind were, or at least were the appearances of, the states and actions of a soul as traditionally understood, that is, of a simple substance that persists personally-identically over time.⁵¹ And the proper use of the idea of God in our empirical investigations of nature in general is to proceed as if the world had been purposefully created by a supreme intelligence.⁵² The case of the cosmological ideas is a bit more complicated due to the fact that they come in pairs, which appears to have as a consequence that we have conflicting rules of procedure on our hands, for example, to investigate nature both as if the world had a beginning in time and as if it was infinite a parte priori. Kant does not explicitly discuss this complication but he offers several explicit descriptions of what he takes the regulative use of the cosmological ideas to amount to, which suggest that he distinguishes two kinds of regulative uses of the cosmological ideas. First, a regulative use in our empirical investigations of nature in natural science, which relies on the ideas featured in the ‘empiricist’ antitheses of the antinomies. And, second, a regulative use in our non-empirical investigations of empirical objects and events from the point of view of their supersensible grounds, which is appropriate for practical philosophy and relies on the ideas featured in the ‘dogmatic’ theses of the third and fourth antinomies.⁵³ Since Kant’s main focus in the first Critique is holds that, in a theoretical context, there is no way for us to demonstrate that freedom or a necessary being that grounds the contingent world is really possible, let alone actual. ⁴⁹ See B611–670/A583–642. ⁵⁰ See B672/A644: “Thus I claim: the transcendental ideas are never of constitutive use, so that thereby concepts of certain objects would be given, and in the case where one understands them in this way they are merely speculative (dialectical) concepts. On the other hand, they have an excellent and indispensably necessary regulative use, namely, to direct the understanding to a certain goal, in expectation of which the focal lines of all its rules converge in one point, which, although it is merely an idea (focus imaginarius), i.e., a point from which the concepts of the understanding do not actually originate, in that it lies completely outside of the limits of possible experience, still serves to secure for them the greatest unity together with the greatest expansion.” Kant also describes this regulative use by saying that in it the ideas of reason are employed as quasi-schemata for the systematic unity of the concepts of the understanding. Just as the understanding needs schemata to unify the sensibly given manifold and bring it under concepts and empirical laws, so reason needs quasi-schemata to bring about a systematic unity of the use of the understanding. See B692–693/A664–665. ⁵¹ See B700/A672; B710–712/A682–684. ⁵² See B698–699/A670–671; B700–701/A672–673; B703–704/A675–676; B713–716/A686–688. ⁵³ See B713/A685: “The absolute totality of the series of these conditions, in the derivation of their elements, is an idea that can never be brought about completely in the empirical use of reason but still serves as a rule for how we should proceed with respect to them, namely, in the explanation of given appearances (in regressing or ascending) in such a way as if the series itself were infinite, i.e., in indefinitum, but where reason itself is considered as a determining cause (in freedom), that is, with respect to practical principles, as if we had before us not an object of the senses but of the pure understanding, where the conditions can be posited not anymore in the series of appearances, but outside of them, and the series of states can be regarded as if it were begun absolutely (through an intelligible cause).” See B700/A672: “Second, we must pursue (in cosmology) the conditions, of inner as well as outer appearances of nature, in such a never to be completed investigation, as if it were infinite and without a first or highest element, although, outside of all appearances, we must not therefore deny their intelligible first grounds, but still never bring them into the connection of the explanations of nature, since we do not know them.” See B536–543/A508–515. The alignment of empiricism with the antitheses and of dogmatism with the theses is due to Kant himself, see B493–494/A465–466: “One notices among the claims of the antithesis a complete homogeneity

334

-     

on theoretical matters, he tends to emphasize the first of these regulative uses of the cosmological ideas there, that is, that we are supposed to proceed in our empirical investigations of nature as if all of the various series of different kinds of conditions for empirical objects—spatial, temporal, material, causal, and existence conditions—that are the subject matter of natural science were infinite (totalities) and never terminated in ultimate first elements. There is a lot more to say about the regulative use of the ideas of reason, in particular, of the cosmological ideas,⁵⁴ but the sketch just presented is enough to

of the manner of thinking and complete unity of the maxim, namely, a principle of pure empiricism, not only in the explanation of the appearances in the world but also in the resolution of the transcendental ideas, of the universe itself. By contrast, the claims of the thesis, in addition to the empirical manner of explanation inside the series of appearances, also assume intellectual beginnings as grounds, and the maxim is insofar not simple. But I will call it, after its essential distinguishing criterion, the dogmatism of pure reason.” ⁵⁴ For example, it is not entirely clear if Kant understands the regulative use of the cosmological ideas in our empirical investigations to work in quite the same general fashion as the regulative use of the psychological and theological ideas. The regulative use of the cosmological ideas of an unconditioned causality (freedom) and of a necessary being that is the cause of the world does appear to be governed by the same general recipe as the regulative use of the psychological and theological ideas, that is, to proceed as if the objects of these ideas existed, albeit only in the supersensible realm. But does Kant also mean to say that in our empirical investigations of nature we are supposed to proceed as if the empirical world were a complex totality of various infinite series of conditions, or does he merely intend to say that we are supposed to proceed under the assumption that all of our regresses to more and more fundamental conditions will go on in indefinitum and that none of them will ever reach absolutely unconditioned first conditions? The text does not permit a clear verdict regarding this question, but one reason that might be taken to speak against the first option is that after realizing that empirical objects as well as their empirical conditions are appearances and not things in themselves, we are also in a position to see that a totality of an infinite series of empirical conditions is impossible, or, more precisely, that the notion of a totality of an infinite series of empirical conditions is nonsensical. (Recall our discussion in section 2.7.2.) This could be taken to speak against the first option on the grounds that it seems plausible to say that in order for something to be a useful fiction it cannot be known to be impossible. Another, related question concerns the ideas that are featured in the theses of the first and second antinomy, that is, the ideas of a beginning in time and of a boundary in space of the world and the idea of simple parts of matter. Is there a regulative use associated with them that is similar to the regulative use of the ideas of freedom and of a necessary being that is the cause of the world, which are featured in the theses of the third and fourth antinomy? The first thing to say is that, given that empirical objects and all of their empirical conditions are appearances, the notions of a first temporal condition, of an ultimate spatial boundary, and of ultimate simple parts of matter are just as nonsensical as the notion of a totality of an infinite series of empirical conditions. (Again, recall our discussion in section 2.7.2.) But could we not still get some kind of regulative use out of these ideas if, just as in the case of the ideas of freedom and of a necessary being that is the cause of the world, we consider their objects not as appearances but as supersensible things in themselves or properties of supersensible things in themselves? This proposal will not work due to a crucial difference between the kinds of series of conditions that are the subject matter of the mathematical (the first two) and the dynamical (the final two) antinomies, respectively. As Kant explains, in the mathematical case, the conditioned and the condition are always of the same kind, while in the dynamical case, they might be of different kinds, which is the reason why “in the mathematical connection of the series of appearances no other than sensible condition can be involved, i.e., such a condition that is itself a part of the series; while the dynamical series of sensible conditions still also admits a dissimilar condition that is not a part of the series but, as merely intelligible, lies outside of the series, whereby reason is satisfied and the unconditioned is placed before the appearances, without thereby confounding the series of the latter, as always conditioned, and to terminate it contrary to the laws of the understanding” (B558–559/A530–531). See FM, 20:292. That is, if a first temporal condition, an ultimate spatial condition, or ultimate material conditions existed, they would be part of the empirical world and, thus, could not be regarded as supersensible things in themselves or properties of supersensible things in themselves. So, does that mean that there is no regulative use associated with the ideas that are featured in the theses of the mathematical antinomies at all? I think it is fair to say that this is indeed the case for the ideas featured in the thesis of the first antinomy, that is, the ideas of an absolute beginning and of an ultimate spatial boundary of the empirical world. But one could try to argue that there is a regulative use of an idea that is closely related to the idea of a simple part of matter that is featured in the thesis of the second antinomy, namely, the idea of a simple thing in itself that is a supersensible ground of matter. Its regulative use would consist in our proceeding in our non-empirical investigations of empirical objects as if simple supersensible things in themselves were their grounds. In this way, Kant’s remark that in our empirical investigation of nature we must proceed “as if it were infinite and without a first or highest element, although, outside of all appearances, we must not therefore deny their intelligible first grounds but still never bring them into the connection of the explanations of nature, since we do not know them” (B700/A672), could be understood to apply not only to the third and fourth antinomy but also to the second.

, -,  

335

put us in a position to understand and later evaluate the claim of the proponents of the fictionalist interpretation that Kant’s theory of ideas supports their reading, which is all we need for now. Kant crisply summarizes the upshot of his theory of ideas in the following passage from the Doctrine of Method, which Vaihinger, the ‘godfather’ of fictionalism, regards as a “classic” statement of Kant’s fictionalism:⁵⁵ The concepts of reason are, as stated, mere ideas and, of course, have no object in any experience, but they still do not for this reason signify objects that are made-up and at the same time also assumed to be possible. They are thought merely problematically in order to ground, in relation to them (as heuristic fictions), regulative principles of the systematic use of the understanding in the field of experience. If one deviates from this, they are mere thought-things [Gedankendinge], whose possibility cannot be demonstrated and which thus cannot be used as the foundation for the explanation of real appearances by means of a hypothesis. (B799/A771)

Although Kant does not explicitly include the concept ‘thing in itself ’ among the ideas of reason, the suggestion is that the status and function of this concept in Kant’s philosophy is comparable to the status and function that he ascribes to the ideas of reason in our investigations of nature, even though he may not have been sufficiently clear on that point himself. Things in themselves—just like souls, indefinite series of empirical conditions, and God—are useful fictional objects that help us to unify, expand, and systematize the use of the understanding but cannot be demonstrated or justifiedly assumed to exist or be really possible.

6.4 Fictionalism, Leibniz-Wolffianism, and Noumena According to my reading, Kant endorses the Kantian conception of things in themselves, described in section 6.1, and is a realist about Kantian things in themselves. Furthermore, he is committed to the view that, from a theoretical point of view, no claim about things in themselves that goes beyond the claims contained in critical idealism, plus the unknowability doctrine, is justifiable or epistemically permissible. That is, he is a fictionalist about things in themselves conceived of in ways that go beyond this own conception of things in themselves. This reading is in tension with the Leibniz-Wolffian reading and the fictionalist reading, both of which have some foundation in the text. According to the former, Kant ascribes many Leibniz-Wolffian features to things in themselves, features that are not entailed by the Kantian conception nor by any other claims about things in themselves contained in critical idealism. According to the latter, Kant holds that things in themselves are useful fictional objects, which serve an explanatory function in our account of the world and ourselves but with respect to whose existence and real possibility we must remain agnostic. Something has got to give. The question is: what?

⁵⁵ See Vaihinger 1918, 619.

336

-     

On my view, both the fictionalist and the Leibniz-Wolffian reading must give. But I also acknowledge that they point to important elements of Kant’s views about things in themselves overall that are worth unearthing. To anticipate, while Kant is not a fictionalist about Kantian things in themselves, he is a fictionalist, not only about God and LeibnizWolffian souls, but more generally about noumena, that is, objects of the pure understanding alone. And while Kant does not endorse the Leibniz-Wolffian conception of things in themselves as a proper characterization of things in themselves, he does endorse it as a proper characterization of noumena. That Kant regards the Leibniz-Wolffian conception of things in themselves as a proper characterization of noumena, as well as a useful fiction, explain his appreciation for it and, together with the special importance of the Leibniz-Wolffian philosophy for Kant’s philosophical development, accounts for the presence of the odd argument for transcendental idealism in the critical writings. The project for the remainder of this section is to revisit the textual evidence that allegedly supports the fictionalist and Leibniz-Wolffian reading and show that it, in fact, does not support them, and to explain Kant’s fictionalism about noumena and noumenalism about Leibniz-Wolffian things in themselves, as one might playfully put it. We will return to the odd argument and the special place of the Leibniz-Wolffian conception of things in themselves in Kant’s heart and philosophy in section 6.5. Before re-evaluating the textual evidence offered by proponents of the traditional fictionalist reading, it is useful to remind ourselves that there are a lot of passages in which Kant unambiguously commits himself to or explicitly asserts the existence of things in themselves.⁵⁶ In the Critique of Practical Reason, he even goes as far as to state as an explicit result of the first Critique that fictionalism about things in themselves is false. . . . [i]t [the speculative Critique] regarded the objects of experience and, among them, our own subject only as appearances, but at the same time it insisted on giving them things in themselves as their grounds, and thus on not regarding everything supersensible as a fiction and its concept as empty. (KpV, 5:6)⁵⁷

This is to say that the fictionalist interpreters are fighting an uphill battle. Even if their evidence were strong, in order to be conclusive it would have to be weightier than the extremely weighty evidence for the reading that Kant is a realist about things in themselves. As it turns out, their evidence is quite weak. I agree that the cited passages from the phenomena-noumena chapter and the Doctrine of Method are useful illustrations of Kant’s conception of a fiction or Gedankending, by which he understands an alleged object (or determination of an object) of a so-called problematic concept, that is, a concept with respect to which we cannot demonstrate or

⁵⁶ Recall in particular Prol, 4:314–315; Prol, 4:350; and GMS, 4:451. Many other such passages are cited in sections 2.1 and 5.1. Also see Bxx: “ . . . the result of this first assessment of our a priori cognition of reason, namely, that it only applies to appearances, but, by contrast, sets the thing in itself aside as for itself real but uncognized by us.” Also recall B723 724/A695 696: “If one then asks (with one’s eyes on a transcendental theology) first: whether there is something distinct from the world that contains the ground of the order of the world and its connection according to general laws, the answer is: without a doubt. For the world is a sum of appearance, so there must be some transcendental ground of it, i.e., a ground that is thinkable only for the pure understanding.” ⁵⁷ Also see Prol, 4:322: “For we are not dealing with the nature of things in themselves, this is independent of both the conditions of our sensibility and [the conditions of] the understanding . . . ” If things in themselves were fictional objects or Gedankendinge their nature would depend on our understanding and its conditions.

, -,  

337

justifiedly assume that a really possible object, let alone an actually existing object, falls under it. More generally, the passages show that fictionalism is on Kant’s radar, so to speak, and that he is prepared to employ fictionalist strategies in his own philosophy. And they also helpfully illustrate that Kant takes some fictions to be useful or even necessary for us to accomplish our cognitive ends, that is, that he does not regard all fictional objects as mere fictional objects, or figments of the brain (Hirngespinste).⁵⁸ But I do not agree that the adduced texts or his account of the ideas of reason in general support the interpretation that Kant regards things in themselves as useful fictional objects. To begin with the ideas of reason, it must be admitted that there are certain similarities between Kantian things in themselves and the alleged objects of the ideas of reason, in particular, those objects that are conceived of as supersensible unconditioned conditions for empirical objects, such as the soul and God. In their capacity as mind-independent grounds of appearances, things in themselves also function as a kind of supersensible unconditioned conditions for empirical objects. The assumption of the existence of Kantian things in themselves also plays a unifying, heuristic, and systematizing role in Kant’s theory of experience that could be compared to the unifying, heuristic, and systematizing role (on Kant’s account) of the assumption of the existence of God in the theory of nature in general, or the unifying, heuristic, and systematizing role of the assumption of the existence of souls in psychology. But, of course, acknowledging these similarities does not automatically commit one to the fictionalist reading. Whether Kant is a fictionalist about Kantian things in themselves depends on whether, on his view, we can demonstrate or justifiedly assume their real possibility. This question cannot be answered by merely pointing out that there are certain similarities between Kantian things in themselves and God or souls. And, as it turns out, Kant offers several fairly detailed clues about how the actual existence of things in themselves, as characterized in critical idealism, is to be demonstrated, as discussed in section 5.10.2. So, clearly, he holds that we are justified in assuming their real possibility. In this context, it is also useful to highlight that Kant neither includes the concept of a thing in itself among the ideas of reason nor says anywhere that he regards Kantian things in themselves as epistemologically on a par with souls or God. By contrast, Leibniz-Wolffian things in themselves are explicitly featured among the objects of the ideas of reason. The conjunction of the psychological ideas corresponds to the traditional concept of a soul, and the thesis of the second antinomy features the idea of simple parts of matter as the unconditioned material conditions for empirical objects.⁵⁹ This is precisely what things in themselves are conceived to be in the Leibniz-Wolffian philosophy (according to Kant’s reading): simple substances or souls that function as ⁵⁸ For the contrast between a mere fiction and a useful fiction see, for example, B697–689/A669–670: “The ideas of reason do not permit any deduction of the kind that the categories permit; but if they are supposed to have at least some, if only indeterminate, objective validity and represent not mere empty thought-things [Gedankendinge] (entia rationis ratiocinantis), a deduction of them must be possible . . . ” For the equation of a mere Gedankending with a Hirngespinst, see B570–571/A542–543: “ . . . it is not permitted with respect to this law to exempt any appearance from it, since otherwise one would posit it outside all possible experience and thus distinguish it from all objects of possible experience and turn it into a mere Gedankendinge and a Hirngespinst.” Also see B123–124/A91; B196/A157; SF, 7:91. ⁵⁹ It is noteworthy in this context that even though the thesis of the second antinomy only talks about simple parts and, more generally, simple substances, in his later discussion Kant describes the thesis as saying that “my thinking self is of a simple and thus incorruptible nature” (B494/A466). That is, even though he does not make it explicit in the initial presentation of the second antinomy, the transcendental realist view that he has in mind as representing the thesis is the Leibniz-Wolffian view that the ultimate simple parts of matter are souls.

338

-     

ultimate ontological building blocks of matter.⁶⁰ Given these considerations, I think it is fair to say that the textual case for the traditional fictionalist interpretation mainly hangs on Kant’s fictionalist sounding remarks about noumena in the chapter on the distinction between phenomena and noumena, to which we will now turn. The phenomena-noumena chapter, just like the amphiboly chapter, is one of the less well researched corners of the Transcendental Analytic. As I see it, it is also one of the less well understood chapters, which seems to be due to a widespread lack of appreciation for Kant’s main overall concern there and a similar lack of attention to what exactly he takes the kind of noumena to be that are the primary focus of his discussion. Kant’s main overall concern in the phenomena-noumena chapter is closely connected to his main overall concern in the amphiboly chapter, which follows right on its heels and ends with a comprehensive summary of the former’s main results.⁶¹ The common thread that connects these two chapters is that, in both of them, Kant is using the results of the Transcendental Analytic to pinpoint crucial junctures at which he departs from the Leibniz-Wolffian orthodoxy that provided the general framework within which, and against the foil of which, he developed his own critical philosophy. More specifically, in the phenomenanoumena chapter, Kant’s main overall concern is to establish that, contrary to LeibnizWolffian orthodoxy, the distinction between appearances and things in themselves, that is, the distinction between fully mind-dependent and mind-independent things, cannot be assumed to map onto the traditional distinction between phenomena and noumena, that is, the distinction between objects of the senses and objects of the pure understanding. It is one of the central tenets of the Leibniz-Wolffian philosophy (as understood by Kant) that the senses furnish us with cognitions of appearances, while the pure understanding furnishes us with cognitions of things in themselves—a tenet that they share with a lot of other traditional metaphysicians going all the way back to Plato. More specifically, in Leibniz-Wolffian metaphysics, as well as in most other traditional metaphysical systems, it is assumed both that noumena are mind-independent and, thus, things in themselves, and that things in themselves can be cognized in terms of pure concepts alone and, thus, are noumena. The latter assumption is an expression of the rationalist nature of traditional metaphysics. Thus, according to traditional metaphysics, the concept ‘noumenon’ and the concept ‘thing in itself ’ are extensionally equivalent. Kant himself relied on this extensional equivalence as late as his Inaugural Dissertation (1770). And so, for a traditional metaphysician, the distinction between phenomena and noumena is basically the same as the distinction between appearances and things in themselves. The upshot of the phenomenanoumena chapter is that the assumption of an extensional equivalence between ‘noumenon’ and ‘thing in itself ’ is unjustified and unjustifiable because noumena are fictional objects. Accordingly, we can no longer assume that the traditional phenomena-noumena distinction maps onto the distinction between appearances and things in themselves. So, the main point of the phenomena-noumena chapter is almost the opposite of what the fictionalist interpreters take from it. On their view, the chapter shows that things in themselves are fictional objects because they are noumena, while, in truth, the chapter shows that we cannot simply

⁶⁰ For a discussion of the various changes in Kant’s views about the composition of matter vis-à-vis the views of the Wolffians, on the one hand, and Leibniz’s views, on the other hand, on his journey to critical idealism, see Jauernig (forthcoming 2021b). ⁶¹ See B342–346/A286–289.

, -,  

339

assume that things in themselves are noumena because the latter are fictional objects.⁶² Note that, in conformity with this reading, in the passage from B310 cited above, Kant actually does not say that things in themselves are fictional objects because they are noumena, but that things in themselves understood as noumena are fictional objects.⁶³ Why are noumena to be classified as fictional? Much of Kant’s discussion in support of this conclusion consists in summarizing some of the main results of the foregoing chapters in the Analytic and drawing out their implications for the question of what it would take for the pure understanding alone to cognize something. This is not the place to examine these results in any detail; I will just report them to the extent relevant for our present concerns. On Kant’s account, in order for a representation of a cognizer to count as a cognition, the cognizer must be able to justify the assumption that the object that the representation represents is really possible.⁶⁴ In the course of the Transcendental Analytic, we learn that, finite cognizers like us can justify assumptions of existence or real possibility only by relying on sensible intuitions in some way, that is, intuitions that are given to us through sensibility, which is a passive faculty separate from the understanding. This directly implies that noumena for us are impossible, that is, objects that are cognized by our pure understanding alone, or in terms of our pure concepts alone.⁶⁵ If there are any noumena, they would have to be objects of a pure understanding that is capable of producing intuitions on its own, that is, whose intuitions are not sensible but intellectual.⁶⁶ But since we do not know if such an intuitive intellect is really possible, let alone actual, it follows that we must remain agnostic about whether noumena are really possible and, thus, conclude that noumena are fictional objects.⁶⁷ All of this also means that things in themselves are definitely not noumena for us, and that we have no grounds for assuming that things in themselves are noumena at all. Things in themselves are not noumena for us because noumena for us are impossible. And we have no grounds for assuming that things in themselves are noumena at all because we have no grounds for assuming that an intuitive intellect is really possible, which, however, is a necessary condition for the possibility of things in themselves to be cognized in terms of pure concepts alone. Accordingly, we can no longer assume that the concept ‘thing in itself ’ and the concept ‘noumenon’ are extensionally equivalent. In the earlier parts of the Critique, Kant emphasizes that things in themselves are not phenomena—as he had argued already in the Inaugural Dissertation. In the phenomena-noumena chapter, he adds that things in

⁶² Misreadings of the phenomena-noumena chapter along these lines are not confined to proponents of the fictionalist interpretation. For example, Allais takes Kant’s arguments for the fictional status of noumena in the phenomenal-noumena chapter as evidence that he is not what she calls a ‘noumenalist,’ that is, somebody who is committed “to the existence of objects which are distinct from and entirely different from the things of which we have knowledge: objects which are not spatial, and which cannot be sensed” (Allais 2015, 60). See Allais 2015, 60 5. But that is not what the chapter shows. The objects whose fictional status is revealed in the phenomenanoumena chapter are not Kantian things in themselves (which is what Allais is talking about) but noumena, that is, objects that can be cognized in terms of pure concepts alone, based on mere logical reflection. ⁶³ See B310 (my emphasis): “The concept of a noumenon, i.e., of a thing that is supposed to be thought not at all as an object of the senses but as a thing in itself (merely through the pure understanding) is not at all contradictory. . . . But in the end the possibility of such noumena is not at all intelligible . . . ” ⁶⁴ See Bxxvi note: “To cognize an object it is required that I can demonstrate its possibility (be it according to the testimony of experience from its actuality or a priori through reason). But I can think what I want if only I do not contradict myself, if only my concept is a possible thought. . . . But in order to attribute objective validity to such a concept (real possibility for the former was merely the logical one), something more is required.” ⁶⁵ See B303 305/A246 248. ⁶⁶ See B308 309, A252 253. ⁶⁷ See B310 312/A254 256; B314 315/A258 259.

340

-     

themselves cannot be equated with noumena either, an addition that marks a radical departure from the view presented in the Inaugural Dissertation.⁶⁸ Given that the concept of a mind-independent thing is quite different from the concept of a thing that can be cognized in terms of pure concepts alone, it is natural to wonder why anyone would ever think that they pick out the same objects. In response to this question, Kant offers a kind of error theory by giving an account of the genesis of the traditional assumption that things in themselves are noumena. It is agreed all around that there are objects that we partly cognize by means of the senses, that is, that there are sensible objects. Since we have no reason to assume that sensible objects exhaust all objects that there are, we must introduce a conceptual distinction between them and those objects to which sensibility does not extend, that is, non-sensible objects. Philosophers who hold that sensibility furnishes us with cognitions of appearances only, such as the LeibnizWolffians or Plato, are also committed to a conceptual distinction between appearances and the things in themselves that appear as those appearances. So, on the view of these philosophers, things in themselves must be thought of as non-sensible objects. Now, given that our pure concepts are independent of sensibility and thus do not have a built-in restriction to sensible objects, it is tempting and natural for these traditional metaphysicians to assume that our a priori concepts furnish us with cognitions of non-sensible objects, and, hence, that things in themselves are not only non-sensible but intelligible objects or noumena. According to Kant’s analysis, this last step is where the error occurs. As he puts it, we are led “to regard the completely indeterminate concept of an object of the understanding as a something in general outside of our sensibility as a determinate concept of a being that we can cognize in some manner through the understanding” (B307).⁶⁹ The error in question can also be described as a confusion of two conceptions of a noumenon, only one of which is not fictional in the sense that sufficiently strong reasons can be adduced for the claim that the conception represents objects that actually exist or at least are really possible. Noumena in the positive sense are the kind of noumena that we have been talking about, that is, objects that can be cognized by the pure understanding alone, or objects that can be cognized in terms of pure concepts alone.⁷⁰ A noumenon in the negative sense is “a thing insofar as it is not an object of our sensible intuition” (B307), or a nonsensible object. The latter conception is a mere limiting concept whose primary function is to keep the pretentions of sensibility in check.⁷¹ The traditional identification of things in ⁶⁸ This sketch of the main line of reasoning in the phenomena-noumena chapter matches Kant’s own summary at the end of the amphiboly chapter. See B342/A286. ⁶⁹ See A249–253; B305–308; B345–346/A289. ⁷⁰ Some of Kant’s formulations suggest that he wants to define noumena in the positive sense as objects of a non-sensible intuition. (See B307: “If we understand by noumenon a thing insofar as it is not object of our sensible intuition in that we abstract from our manner of intuiting it, it is a noumenon in the negative understanding. But if we understand by it an object of a non-sensible intuition, we assume a special manner of intuiting, namely, the intellectual [manner], which, however, is not ours, and of which we can also not comprehend the possibility, and this would be the noumenon in the positive sense.”) But I think the noumena-phenomena chapter overall, as well as the summary of the phenomena-noumena chapter at the end of the amphiboly chapter, make it quite clear that Kant is working with a broader conception of a noumenon in the positive sense, namely, as an object that can be cognized by a pure understanding alone. Since it turns out that noumena thus understood would be possible only as objects of an intuitive intellect, it is understandable why Kant sometimes identifies noumena in the positive sense with the objects of an intuitive intellect, or of a non-sensible intuition, even though there is a conceptual distinction between them. ⁷¹ See B310 312/A255–256. Also see B345–346/A288. There are two oddities about the summary of the phenomena-noumena chapter at the end of the amphiboly chapter that are relevant for the question of how to understand Kant’s terminology of a negative and positive sense of the concept ‘noumenon,’ which are worth

, -,  

341

themselves with noumena in the positive sense is based on an illicit slide from the legitimate assumption that things in themselves are noumena in the negative sense, that is, that they are not objects of our (=human) sensible intuition, to the illegitimate assumption that things in themselves are noumena in the positive sense, that is, that they are objects that can be cognized by the pure understanding alone. (In the following, I will continue to use ‘noumenon’ without any further qualifications as before, that is, in the positive sense as meaning objects that can be cognized in terms of pure concepts alone.) Kant sometimes refers to Kantian things in themselves as noumena or intelligible objects. On these occasions, ‘noumenon’ is used in the negative sense. Kantian things in themselves are included in the class of objects that are not objects of our sensible intuition; they are a special kind of non-sensible objects.⁷² Granted, Kant’s conception of a thing in itself in his critical idealism is more specific than the concept of a noumenon in the negative sense. But Kantian things in themselves are still not noumena in the positive sense. Even if Kant’s conception of a thing in itself turned out to be a cognition of some kind (albeit not a cognition in the strict sense), it certainly is not a cognition in terms of pure concepts alone. All of the features that are included in the Kantian conception of things in themselves are underwritten by core tenets of critical idealism, none of which are justified on the basis of mere logical reflection, or the mere analysis or comparison of pure concepts. For example, the claim that things in themselves are non-spatial and nontemporal is underwritten by the thesis of the transcendental ideality of space and time, whose justification, as we saw in section 4.2.2.2, crucially depends on Kant’s analysis of the

commenting upon. First, in the summary Kant calls objects of a non-sensible intuition “noumena in a merely negative sense” (B342/A286), whereas in the B-edition of the phenomena-noumena chapter this is precisely how he characterizes noumena in the positive sense, as just reported in the previous footnote. Second, the summary in the amphiboly chapter sounds as if Kant regards the concept of an object of a non-sensible intuition as the exhaustive complement to the concept of an object of our sensible intuition, whereas in the B-edition of the phenomena-noumena chapter he is careful to distinguish three concepts, that is, the concept of an object of our sensible intuition, the concept of an object of a non-sensible intuition, and the concept of an object that is not an object of our sensible intuition. This suggests that between the publication of the A- and B-edition Kant conceptually sharpened his views about noumena, and that he made some attempts to record the results of this sharpening in the revised phenomena-noumena chapter but not in the corresponding section in the amphiboly chapter (which is unchanged compared to the A-edition). Kant came to realize that even though the concept of an object of a non-sensible intuition is a ‘negative’ concept of a noumenon in the sense that the intuition in question is not sensible, there is an even more negative concept of a noumenon, as it were, that is, a less informative and thus less problematic concept, namely, the concept of an object that is not an object of our sensible intuition. Such an object could be an object of a non-sensible intuition but it could also be an object that is not intuited at all. The following way of describing the situation might be helpful. There are several positive conceptions of a noumenon. The most general positive conception is that it is an object that can be cognized by a pure understanding alone. More specific positive conceptions result if we specify what kind of pure understanding we are talking about. First, there is the conception of a noumenon as an object that can be cognized by our pure understanding alone. Second, there is the conception of a noumenon as an object that can be cognized by an intuitive intellect alone, that is, as an object of a non-sensible or intellectual intuition. The negative conception of a noumenon is the conception of an object that is not an object of our sensible intuition. Based on the results of the Transcendental Analytic, we know that noumena in the first positive sense are impossible. By contrast, we have no reason to declare noumena in the second positive sense as impossible. This is the sense in which noumena in the positive sense “must indeed be admitted” (B342/A286), as Kant puts it in the summary passage in the amphiboly chapter. But since we have no way of knowing whether an intuitive intellect is really possible, we cannot know whether noumena in the second positive sense are really possible, which is why they must be classified as fictional objects. Noumena in the negative sense, by contrast, are demonstrably really possible, for they include things in themselves. ⁷² See A358: “ . . . the something that grounds outer appearances, what affects our sense in such a way that it receives the representations of space, matter, and figure, this something, considered as noumenon (or better, as transcendental object) . . . ” The correction in brackets is revealing; Kant knows that characterizing the things in themselves that ground empirical objects as noumena is misleading, given that many of his readers will understand the term in the positive sense.

342

-     

nature of intuition. Moreover, the existence of Kantian things in themselves can be demonstrated, as we saw in section 5.10.2, which means that if they were noumena in the positive sense they would constitute a counterexample to the claim that noumena in the positive sense are fictional objects, a claim to which Kant seems to be firmly committed, however. The sketched reading of the phenomena-noumena chapter also gives us the proper perspective from which to re-evaluate the amphiboly chapter, which, to reiterate, follows right on its tail. As already reported, the Leibniz-Wolffian metaphysical principles that Kant examines there, for example, the principle of the identity of intrinsically indiscernibles and the principle of the non-opposition of realities, are derived by means of logical reflection alone (broadly conceived), namely, by thinking about the comparison of things or their properties in terms of pure concepts alone. That is, these principles are principles about noumena. Since the Leibniz-Wolffians take things in themselves to be noumena they regard these principles to be valid for things in themselves. By contrast, Kant just made a point of emphasizing, in the phenomena-noumena chapter, that we have no reason to assume that things in themselves are noumena. So, it would be rather surprising, to say the least, if Kant were to follow the Leibniz-Wolffians and take their metaphysical principles to be valid for things in themselves. And, indeed, that is not what he does. Whenever Kant concurs with a Leibniz-Wolffian thesis in the amphiboly chapter, about identity and distinctness, harmony and opposition, the inner and outer, or matter and form, he makes sure to highlight that he agrees with the thesis understood as a claim about noumena or intelligible objects, not about things in themselves. Leibniz took appearances for things in themselves and thus for intelligibilia, i.e., objects of the pure understanding . . . and there his principle of the indiscernible (principium identitas indiscernibilium) could indeed not be disputed. (B320/A264) If reality is represented by the pure understanding only (realitas noumenon), no opposition between realities can be thought . . . (B320/A64) With respect to an object of the pure understanding only that is inner which has no relation at all (as far as its existence is concerned) to something that is distinct from it. . . . But as an object of the understanding every substance must have inner determinations and powers that pertain to inner reality. . . . For this reason, Leibniz made all substances . . . into monads, since he conceived of them as noumena. (B321–322/A265–266)⁷³ Therefore, in the concept of the pure understanding matter precedes form, and, for this reason, Leibniz first assumed things (monads) and inwardly their power of representation, in order to later base on this their outer relations and the community of their states (namely, of the representations) . . . (B323/A267)

That is, Kant agrees that the principles of Leibniz-Wolffian metaphysics are valid for noumena, or, more carefully put, he agrees that these principles are valid for the objects of human pure understanding, understood as objects that are represented (although not

⁷³ See V-Met-K2/Heinze, 28:731: “Since matter fills space and thus does not consist of simple parts, the material world is no monadatum; for if space is always divisible, so everything in it is divisible.—The world, considered as noumenon, does indeed consist of simple parts.”

, -,  

343

cognized) in terms of our pure concepts alone, objects that the Leibniz-Wolffians take to be noumena. (For future reference, call these objects ‘noumena in the weak positive sense.’) But since we have no way of knowing whether an intuitive intellect is really possible, let alone whether this intuitive intellect’s cognitive repertoire includes our pure concepts, we are certainly not justified in assuming that the principles in question are valid for things in themselves, pace the Leibniz-Wolffians.⁷⁴ Kant puts this point by saying that the LeibnizWolffian principles have no meaning with respect to things in themselves. The principle that realities (as mere affirmations) are never logically opposed to one another is an entirely true proposition concerning the relation of concepts but does not have the slightest meaning with respect to nature nor anywhere with respect to any thing in itself (of this we have no concept). (B328–239/A272–273)

All of this means that Kant’s discussion in the amphiboly chapter cannot be read as revealing further details about his own conception of a thing in itself—which is how it is read on the Leibniz-Wolffian interpretation.⁷⁵ Rather, what we get is an explication of how objects are represented by our pure understanding alone, which amounts to an exposition of how the Leibniz-Wolffians conceive of things in themselves, since they take things in themselves to be noumena, but which does not entail anything about Kantian things in themselves. It is also not difficult to see why Kant does not dwell on the Leibniz-Wolffian confusion of things in themselves with noumena in the amphiboly chapter. For one thing, given that the main result of the immediately preceding phenomena-noumena chapter is precisely that we are not justified in assuming that things in themselves are noumena, Kant probably thought that it would be obvious to his readers that he does not approve of the LeibnizWolffian inclusion of things in themselves in the scope of their purely intellectual, metaphysical principles about noumena. For another thing, his main focus in the amphiboly chapter lies elsewhere, namely, on showing that the Leibniz-Wolffians are also committed to a further mistake, namely, the mistake of taking their metaphysical principles to be valid for phenomena or empirical objects, to which they are led on account of their misconception of sensible representations as confused concepts, as explicated above. Kant’s main general concern in the phenomena-noumena chapter is to expose the untenability of the assumption that the traditional distinction between phenomena and noumena maps onto the distinction between appearances and things in themselves. His main general concern in the amphiboly chapter is to show that even if the traditional distinction between phenomena and noumena did map onto the distinction between appearances and things in themselves (which it does not), the Leibniz-Wolffians would still be wrong to extend their purely intellectual metaphysical principles to appearances.

⁷⁴ Note that a similar mistake—of ascribing certain features of objects of the pure understanding that are derived on the basis of logical reflection broadly conceived to things in themselves—also underlies the so-called paralogisms of reason, that is, the problematic syllogisms by means of which reason arrives at the conception of the soul as a simple substance that persists identically over time. See 349/B398: “In the syllogism of reason of the first kind I infer from the transcendental concept of the subject that does not contain anything manifold to the absolute unity of the subject itself, of which, in this way, I have not concept at all. I will call this dialectical inference the transcendental paralogism.” ⁷⁵ And on the interpretations referenced in note 33.

344

-     

But even if we could say something synthetically about things in themselves through the pure understanding (which, however, is impossible), this could still not at all be applied to appearances, which do not represent things in themselves. (B332/A276)

In sum, both Kant’s fictionalist sounding remarks about things in themselves in the phenomena-noumena chapter and his Leibniz-Wolffian characterizations of things in themselves in the amphiboly chapter do not apply to Kantian things in themselves but to things in themselves conceived of as noumena, that is, as objects of the pure understanding alone. These remarks and characterizations can thus not be used as evidence for either the fictionalist reading or the Leibniz-Wolffian reading of Kant on things in themselves.

6.5 The Special Status of Leibniz-Wolffian Things in Themselves and the Odd Argument Revisited The result that Kant does not endorse the Leibniz-Wolffian conception of things in themselves as unpacked in the amphiboly chapter both reinforces the impression of oddness caused by the odd argument for transcendental idealism discussed in section 6.2 and raises the question of how its dialectical function is supposed to be understood. The odd argument emerges as an argument for transcendental idealism, proposed by Kant, that is built on Leibniz-Wolffian premises, which he neither endorses nor regards as justifiable or epistemically permissible. What is up with that? A related, more general worry that might be lingering in the mind of some readers concerns Kant’s evident fondness for the Leibniz-Wolffian conception of things in themselves. If his realism about things in themselves is indeed restricted to Kantian things in themselves, and he regards all conceptions of things in themselves that go beyond his own as fictions, he should judge the Leibniz-Wolffians more sternly for being realists about Leibniz-Wolffian things in themselves and, more generally, should sound less appreciative of their conception of things in themselves than he in fact sounds, especially, in his discussion in the amphiboly chapter. The questions of how to make sense of the odd argument, and account for Kant’s apparent liking for the Leibniz-Wolffian conception of things in themselves and for the comparative mildness of his criticism of the Leibniz-Wolffians for endorsing it and understanding it in realist terms, are the topic for this section. Starting with the latter, the first thing to say is that although Kant judges it to be a mistake to endorse the Leibniz-Wolffian conception of things in themselves, he acknowledges it to be a mistake that is very hard not to make. Kant views the supposition that we can cognize things in themselves in terms of pure concepts as an “illusion that is difficult to avoid” (B305). The error detailed in the previous section of sliding from the observation that the pure concepts of the understanding do not come with a built-in restriction to sensible objects to the assumption that they can be used to cognize non-sensible objects is a natural one to which we are all prone.⁷⁶ In other words, Kant admits that it is hard to resist ⁷⁶ Also see in this sense Prol, 4:315–316: “Not only that our concepts of substance, power, action, reality etc. are completely independent of experience, and, similarly, do not contain any appearance of the senses, and thus seem to apply to things in themselves (noumena), they contain a necessity of determination in themselves to which

    -   

345

the temptation to think that things in themselves are noumena and to take ourselves to be within our rights to use our pure concepts to conceive of them in more determinate ways than is actually legitimate. Moreover, as we saw in our brief examination of Kant’s account of the ideas of reason, he holds that certain ideas and the supposition that the objects of these ideas exist are natural or even inevitable for us, given reason’s maxim to always look for the unconditioned for any given conditioned. In addition, Kant holds that the ideas of reason have a legitimate regulative use, and thus that their objects are useful fictions. And, as noted, the inevitable ideas whose objects are useful fictions include the concepts of a soul and of simple parts of matter, which are central ingredients in the Leibniz-Wolffian conception of things in themselves. More generally, Kant diagnoses in us a natural desire to cognitively access the unconditioned, in particular, the supersensible unconditioned, and thereby attain ultimate explanations and a stable cognitive resting place, as one might put it. This natural desire is grounded in a “peculiar principle of reason in general (in its logical use),” namely, the principle “to find for the conditioned cognition of the understanding the unconditioned whereby its unity is completed” (B364/A307). This principle also underlies our “natural disposition” to metaphysics,⁷⁷ whose ultimate aspiration is the cognition of the supersensible.⁷⁸ The ideas of reason of supersensible objects thus satisfy the “speculative interest” of reason that is expressed in our search for ultimate explanations in terms of the supersensible unconditioned.⁷⁹ Conceiving of things in themselves in Leibniz-Wolffian terms arguably serves this kind of speculative interest and caters to our natural metaphysical tendencies.⁸⁰ Finally, the Leibniz-Wolffian view that things in themselves are simple souls is also favored by our “practical interest,” which concerns matters of morality and religion, one of whose “cornerstones” is the claim that “my thinking self is of

experience never lives up, which additionally reinforces this supposition. . . . For this reason, the concepts of the understanding appear to have much more meaning and content than that the mere empirical use exhausted their entire determination, and thus the understanding without being noticed builds in addition to the house of experience a much more sprawling annex, which it fills with lots of thought-entities [Gedankenwesen] without even noticing that it has transgressed with its otherwise correct concepts beyond the bounds of their use.” ⁷⁷ See B21: “But this kind of cognition [metaphysics] is also to be regarded as given in a certain sense, and metaphysics is real as natural disposition (metaphysica naturalis), if not as a science. For human reason proceeds unstoppably, driven by its own need, without it being moved by mere vanity of knowing a lot, to such questions that cannot be answered by any empirical use of reason and principles taken from there, and so some kind of metaphysics has been real in all human beings . . . ” See Prol, 4:365–367; Avii–viii. ⁷⁸ See FM, 20:260: “This final end to which all metaphysics is directed is easily discovered and can ground a definition of it in this respect: ‘it is a science to progress through reason from the cognition of the sensible to the cognition of the supersensible’.” See B826/A798: “The final end in which the speculation of reason in the transcendental use eventually culminates concerns three objects: the freedom of the will, the immortality of the soul, and the existence of God.” See KU, 5:473; LB F2, 2:273–277, 23:134; V-Met/Dohna, 28:616. ⁷⁹ See B494–495/A466–467: “Second, a certain speculative interest of reason is also expressed on this side [the side of the thesis of the antinomies]. For if one takes and uses the transcendental ideas in this manner, one can comprehend the entire chain of conditions completely a priori and understand the derivation of the conditioned in that one begins from the unconditioned, which cannot be achieved by the antithesis . . . ” See Prol, 4:354–355: “Thus we ought to think an immaterial being, a world of the understanding, and a highest of all beings (all of them noumena), since reason finds completion and satisfaction only in them as things in themselves, which it can never hope to attain in the derivation of appearances from their grounds that are similar in kind . . . ” See also B694/ A666; B704/A676, See V-Th/Pölitz, 28:1082. For a very helpful in-depth investigation of Kant’s account of the sources of metaphysics in the nature of human reason, see Willaschek 2018. ⁸⁰ See Kant’s comments with respect to our compulsion to think of things in themselves as composed of simple parts in his notes for Schulz’s review of Eberhard, 20:390–391: “ . . . although we can say of them [things in themselves] that the understanding must think the simple in them, i.e., the non-composite, this is still not a cognition of them but merely the representation of reason of the unconditioned to whose cognition we cannot get any closer at all . . . ”

346

-     

a simple and thus incorruptible nature” (B494/A466).⁸¹ None of these considerations amounts to an exculpation of the Leibniz-Wolffians. But they show that Kant views their illegitimately over-determinate conception of things in themselves as grounded in the nature of human reason and, to that extent, as excusable. Similarly, they establish that Kant regards Leibniz-Wolffian things in themselves as useful fictional objects that serve various epistemic needs that are fundamental to us as human cognizers. Against this background, Kant’s light touch in criticizing the Leibniz-Wolffians for endorsing the Leibniz-Wolffian conception and understanding it in realist terms is no longer puzzling. The foregoing considerations also provide the beginning of an explanation of Kant’s apparent fondness for Leibniz-Wolffian things in themselves. But the main reason for Kant’s visibly positive attitude toward the Leibniz-Wolffian conception of things in themselves is, I take it, that he sees it as the ‘correct’ conception from the point of view of the pure understanding. It is not just natural and useful for us to think about things in themselves in Leibniz-Wolffian terms, in the framework of traditional metaphysics, that is, in a framework where it is assumed that we can cognize things in themselves in terms of pure concepts alone, adopting the Leibniz-Wolffian conception is mandatory if we want to get things ‘right.’ As we just saw in our reconsideration of the amphiboly chapter, Kant agrees with the Leibniz-Wolffians that their purely intellectual, metaphysical principles are valid for objects of the pure understanding, or, at least, of our pure understanding, that is, for (what I am calling) noumena in the weak positive sense. More precisely, I take Kant to be agreeing with the Leibniz-Wolffian account of how our pure understanding must represent objects when relying exclusively on its own resources. There is a particular way in which our pure understanding represents objects that can be unpacked by means of logical reflection on and analysis and comparison of our pure concepts. All rationalist metaphysicians are engaged in this project in one way or another. And, in Kant’s assessment, the Leibniz-Wolffian account of our pure understanding’s way of representing objects is right on the money. If things in themselves were objects of the pure understanding, and our pure understanding were sufficiently similar to God’s pure understanding, as is assumed in traditional metaphysics, Leibniz-Wolffian monadological metaphysics would be demonstrably true of the most fundamental level of reality.⁸² I take it that this kind of story is a central part of what Kant has in mind when he says that the Critique is “the true apology for Leibniz” by functioning as “the key for all interpretations of the products of pure reason based on mere concepts” (ÜE, 8:250–251). In thinking about how to properly understand the odd argument, it is useful to start by noting that it played an important role in Kant’s philosophical development and his journey to critical idealism, which he began in a philosophical environment that was dominated by the Leibniz-Wolffian philosophy.⁸³ This is not the place for a detailed discussion of Kant’s philosophical development, but, at a fairly high level of abstraction, one strand of this development can be characterized as the gradual emergence of the odd argument in between the lines or in the background, as it were, culminating in the Inaugural Dissertation in the distinction between sensible objects or phenomena and intelligible objects or noumena. Kant first discovers that several central principles of Leibniz-Wolffian metaphysics, which the Leibniz-Wolffians present as universally valid, ⁸¹ See B828–832/A800–804; FM, 20:309; KpV, 5:119–121. ⁸³ See Jauernig 2011.

⁸² See ÜE, 8:201.

    -   

347

are, in fact, not valid for sensible objects, as for instance, the principle of the identity of indiscernibles,⁸⁴ the principle that positive determinations are not opposed to each other,⁸⁵ and the principle that the objects in space are ontologically prior to space.⁸⁶ Initially, he takes this to show that these Leibniz-Wolffian principles are simply false. But eventually he comes to see how he can both give these principles their due and acknowledge that they do not apply to sensible objects, namely, by regarding sensible objects and noumena as fundamentally different kinds of entities and restricting the scope of the LeibnizWolffian principles to the latter. The final, critical step, which goes beyond the odd argument, is the realization that the assumption that things in themselves are noumena is unjustifiable, which means that we have no reason to think that the metaphysical principles of the Leibniz-Wolffians are valid for things in themselves.⁸⁷ By bringing up Kant’s philosophical development in this context, I do not mean to endorse a version of the notorious ‘patchwork theory,’ according to which Kant assembled the Critique by hastily and carelessly patching together various disparate arguments and strands of reasoning from different phases of his pre-critical philosophy, which accounts for the work’s near incomprehensibility and its many inconsistencies.⁸⁸ That is, I do not want to say that the presence of the odd argument in the critical writings is to be accounted for by the fact that it played an important role in the genesis of critical idealism and that Kant failed to notice, or noticed but then forgot again, that the argument is no longer defensible in the framework of the critical philosophy. My purpose is more modest and less hostile to Kant, namely, to point out that his philosophical development can shed light on how the odd argument came to him in the first place and why he puts so much stock in it as a pedagogical tool in the presentation of his new philosophical system. I submit that when Kant presents versions of the odd argument in his critical writings he does so for the special benefit of a particular group of his readers, namely, the LeibnizWolffians and other philosophers who are attracted by the Leibniz-Wolffian conception of things in themselves or who, like his younger self, are working in the LeibnizWolffian paradigm. Kant will have anticipated this group to make up a significant contingent of his readership, not only because of the lingering popularity of LeibnizWolffianism in late eighteenth-century Germany, but also because of his belief, just discussed, that the Leibniz-Wolffian conception of things in themselves is natural, useful, and ‘correct’ from the perspective of traditional metaphysics, which suggests that it is bound to retain a certain popularity for times to come. And so it is quite understandable, especially from a pedagogical perspective, why Kant would have employed the odd argument in mounting his case. His intention is to force the Leibniz-Wolffians and company to admit that, by their own lights, empirical objects

⁸⁴ See PND, 1:409–410. ⁸⁵ See NG, 2:167–204. ⁸⁶ See GUGR, 2:378, 381 383. ⁸⁷ This gradual emergence of the odd argument was certainly not the only important factor in the genesis of transcendental idealism. Transcendental idealism was developed as a solution to a whole range of problems that Kant was dealing with in the pre-critical period. Apart from the question of what to do with the mentioned Leibniz-Wolffian principles, these problems also include, for example, the question of how to reconcile libertarian freedom with determinism in nature and, more generally, of how to escape the antinomies as well as of how to explain the possibility of mathematics, understood as an a priori synthetic discipline that applies to empirical objects. ⁸⁸ Versions of the patchwork theory have been defended by commentators such as Vaihinger 1881/1892; Adickes 1929; and Kemp Smith 1918.

348

-     

cannot be identified with things in themselves, as conceived by them, not even confusedly perceived ones, but must be seen as entities of a fundamentally different kind. Once they have acknowledged that much, they are already one big step closer to transcendental idealism and, presumably, also more open to Kant’s other, non-odd arguments for it that were discussed in chapter 4. Understood in this way, one could compare the odd argument to Wittgenstein’s ladder, or, at least, a few rungs on the ladder, that is to be discarded after one has reached a new point of view with its help, in this case, the point of view of Kant’s critical philosophy.

6.6 The Indirect Argument for Transcendental Idealism Based on the Antinomies and Another Possible Feature of Kantian Things in Themselves As already noted, Kant holds that the antinomies allow us to construct an indirect argument for transcendental idealism, or, at least, for one of its core theses, namely, the claim that empirical objects are not things in themselves. Call this the ‘indirect argument’ for short. I want to wrap up by taking a closer look at the indirect argument and its dialectical function in Kant’s case for transcendental idealism. For, depending on how this function is to be understood, the indirect argument may make it incumbent upon us to acknowledge the inclusion of one more feature in Kant’s own conception of things in themselves that has gone unmentioned so far. Moreover, the indirect argument and the odd argument turn out to be close cousins. Here is how Kant characterizes the indirect argument in the Transcendental Dialectic: On the other hand, one can also reap some true benefit from this antinomy, if not a dogmatic one then a critical and doctrinal one: namely, to prove the transcendental ideality of appearances through it in an indirect way, in case somebody is not satisfied by the direct proof in the transcendental Aesthetic. The proof would consist in the following dilemma. If the world is a whole that exists in itself, it is either finite or infinite. Now the former as well as latter is false (according to the above proofs of the antithesis, on the one hand, and the thesis, on the other hand.) Thus, it is also false that the world (a sum total of all appearances) is a whole that exists in itself. Whence it follows that appearances in general are nothing outside our representations, which is precisely what we wanted to signify by their transcendental ideality. (B534–535/A506–507)

The way in which Kant presents the argument in this passage is tailored specifically to the mathematical antinomies, in particular, the first one. But since all antinomies depend on the assumption that empirical objects are things in themselves—albeit not all in quite the same way⁸⁹—arguably, they all can be used in an indirect argument for transcendental idealism along the general lines indicated in the passage. This general indirect argument may be formulated schematically as follows.

⁸⁹ See note 2, chapter 3.

     

349

1. Empirical objects are things in themselves. (TR1-ii, transcendental realist premise.) 2. Antinomies.⁹⁰ (From 1 and the nature of things in themselves.) 3. Empirical objects are not things in themselves. (By reductio ad absurdum of 1, given that it leads to 2.) Kant’s discussion of how and why the transcendental realist is driven to the antinomies reveals that he takes a certain feature of things in themselves to play an important role that we have not talked about yet, at least not in the present chapter. This feature is that things in themselves conform to the principle that if something conditioned exists, then its condition, and, indeed, the entire series of all of its conditions, exists as well. Call this principle the ‘conditioning principle’ for short.⁹¹ The realization that this principle does not apply to the conditioned aspects of empirical objects and their conditions that are thematized in the antinomies, that is, their spatial, temporal, material, causal, and modal conditions, because empirical objects are appearances and not things in themselves, is the key to the solution of the antinomies. For it means that the inferences from the existence of particular empirical objects to the existence of the various versions of the unconditioned— in the form of unconditioned first elements of the series or in the form of unconditioned infinite series—that is asserted in the theses and antitheses of the antinomies, are fallacious. In Kant’s words: If the conditioned as well as its condition are things in themselves, then if the first is given, not only the regress to the latter is set as a task, but it [the condition] is thereby really already given, and since this holds for all members of the series, the complete series of conditions, and thus the unconditioned, is given at the same time, or rather presupposed, through the fact that the conditioned is given, which was possible only through this series. Here the synthesis of the conditioned with its condition is a synthesis of the mere understanding, which represent the things as they are, without paying attention to whether and how we can get acquainted with them [zur Kennntnis derselben gelangen können]. On the other hand, if I am dealing with appearances which, as mere representations, are not at all given if I do not get to their acquaintance [Kenntnis] (that is, to them themselves, for they are nothing but empirical cognitions [Kenntnisse]), I cannot say in the same sense that if the conditioned is given, then all of its conditions (as appearances)

⁹⁰ That is, the world is both finite and infinite in space and time; composite substances are both composed of simples and not composed of simples; there are uncaused causes in the world and there are no uncaused causes in the world; there is a necessary being in the world and everything in the world is contingent. ⁹¹ Recall B364/A307–308: “ . . . if the conditioned is given then the entire series of subordinated conditions is given as well, which itself is thus unconditioned.” Also see B525/A497. Is ‘to be given’ the same as to exist? Maybe not in general, but in the present case this substitution seems both warranted and advisable. It is warranted because the problematic claims that issue from the misapplication of the conditioning principle and constitute the antinomies are all existence claims or depend on existence claims. And the replacement is advisable because talking about something ‘being given’ is rather vague. At B534/A506 (quoted in the following note), Kant himself uses ‘exist’ instead of ‘are given.’ Instead of talking about the ‘entire series of conditions’ Kant also often uses the expression ‘totality of the conditions’ or ‘totality of the series of conditions’; see the following quotation in the main text and the quotations in the subsequent note. It might be useful to highlight that, as I read him, none of these expressions are meant to signify anything more than just all the conditions. Granted, just as one might hold that a set of objects is a thing different from and in addition to all the elements of the set, one might hold that a series of conditions or a totality of conditions or a totality of a series of conditions is a thing different from and in addition to all the conditions. On my reading of Kant, all that he has in mind in the relevant passages, however, is all the conditions and not some additional thing.

350

-     

are also given, and thus cannot infer to the absolute totality of the series of them. . . . From this one sees that the major premise of the cosmological syllogism of reason takes the conditioned in the transcendental sense of a pure category, but the minor premise takes it in the empirical sense of a concept of the understanding applied to mere appearances, and thus that one finds in it the dialectical deception that is called sophisma figurae dictionis. (B526–528/A498–499)⁹²

Given this description of the fundamental error that gives rise to the antinomies, we can schematically formulate the indirect argument more precisely as follows. 1. Empirical objects are things in themselves. (TR1-ii, transcendental realist premise.) 2. Things in themselves conform to the principle that, if something conditioned exists, then its condition, and, indeed, the entire series of all of its conditions, exists as well. (Additional ‘odd’ premise about things in themselves.) 3. Empirical objects exist and are conditioned in various ways, namely, spatially and temporally, materially, causally, and modally. (From empirical evidence, possibly supplemented by some kind of refutation of idealism, and reflection on the properties of empirical objects.) 4. The entire series of all of the various conditions of empirical objects exist. (From 1, 2, and 3.) 5. Antinomies. (Arise on the basis of 4.)⁹³ 6. Empirical objects are not things in themselves. (By reductio ad absurdum of 1, given that it leads to 5.)⁹⁴ This argument has some striking similarities with the odd argument for transcendental idealism discussed in sections 6.2 and 6.5. Just like the odd argument, the indirect argument involves a premise about things in themselves that might strike one as odd, namely, premise 2, which asserts that the conditioning principle is true of them. This

⁹² See KpV, 5:107: “Pure reason always has its dialectic, one may consider it in its speculative or practical use; for it demands the absolute totality of the conditions to a given conditioned, and this simply can be found only in things in themselves. But since all concepts of things must be related to intuitions, which in the case of us human beings cannot be other than sensible, and thus allow the objects to be cognized not as things in themselves but only as appearances in whose series of the conditioned and the conditions the unconditioned can never be found, an inevitable illusion arises from the application of this idea of reason of the totality of the conditions (and thus of the unconditioned) to appearances as if they were things in themselves (for that is what they are always taken to be without a warning critique), an illusion that would never be noticed as illusory if it did not give itself away though an opposition of reason with itself in the application of the principle to presuppose the unconditioned to every conditioned with respect to appearances.” See B534/A506: “Thus, the antinomy of pure reason in its cosmological ideas is removed by showing that it is merely dialectical and an opposition of illusion, which arises from one’s application of the idea of an absolute totality, which is only valid as a condition for things in themselves, to appearances, which exist only in representation, and, if they make up a series, exist only in the successive regress but otherwise not at all.” Also see KpV, 5:107; and FM, 20:290. ⁹³ How exactly the different antinomies arise on the basis of the result recorded in line 4 is a complex story, which Kant himself does not outline for us. Telling this story involves spelling out how the various proofs of the theses and antitheses making up the antinomies depend on the conditioning principle as applied to empirical objects. Fortunately, for our present purposes, we need not worry about these details. ⁹⁴ The cosmological syllogism of reason that Kant talks about in the passage just quoted in the main text, basically, corresponds to lines 1–4 of this argument. See B525/A497: “The entire antinomy of pure reason rests on the dialectical argument: If the conditioned is given, then the entire series of all its conditions is given as well. Now objects of the senses are given to us as conditioned, therefore etc. . . . ” We will take a closer look at this syllogism and Kant’s diagnosis of why it is fallacious below.

     

351

premise might seem odd because it appears to go beyond Kant’s conception of things in themselves, as detailed in section 6.1, which might be taken as evidence that he regards it, or at least should regard it, as unjustifiable and epistemically impermissible. This raises the question whether the feature that things in themselves conform to the conditioning principle is included in Kant’s own conception of things in themselves, and of how we should understand the dialectical function of the indirect argument. I do not think that the textual evidence is fully conclusive, but I want to close by saying a few words about the two main interpretative options and indicating which one I favor. First, one could hold that Kant’s own conception of things in themselves does not include the feature that they conform to the conditioning principle, and that the indirect argument is, basically, on a par with the odd argument. While the critical Kant does not endorse it, it played an important role in the genesis of the critical philosophy and serves an important dialectical function in the critical period in Kant’s case against transcendental realism. In the critical period, the indirect argument is addressed mainly to a particular audience, namely, to those of his readers who accept the second, ‘odd’ premise—which, arguably, includes all (or most) transcendental realists, the Leibniz-Wolffians among them. The argument is employed as a pedagogical tool in order to force the transcendental realists to take their first step toward transcendental idealism by realizing that, by their own lights, empirical objects cannot be things in themselves, as conceived by them. In support of this reading, one can point to the fact that the antinomies and the question of how to escape from them indeed were chief among the problems with which Kant struggled in his younger years and that set him on the path of developing the critical philosophy.⁹⁵ And one can also draw attention to the fact that Kant describes the kind of synthesis of the conditioned and its condition that is expressed in premise 2 as “a synthesis of the mere understanding, which represents the things as they are” (my emphasis) and, at another place, characterizes the conditioning principle as “a principle of all things insofar as their connection is thought through pure reason, i.e., as a connection of things in themselves” (FM, 20:290). This could be taken as an acknowledgment that the conditioning principle is a principle about noumena, that is, objects of the pure understanding, which would put it on exactly the same footing as the metaphysical principles of the Leibniz-Wolffians that Kant discusses in the amphiboly chapter and that serve as premises in the different versions of the odd argument. Second, one could hold that Kant’s own conception of things in themselves includes the feature that they conform to the conditioning principle, and that Kant himself endorses the indirect argument in the critical period. In support of this reading, one could point to Kant’s rather enthusiastic characterizations of the probative force of the argument, for example, as “indisputably confirming” the results from the earlier sections of the Critique (FM, 20:291).⁹⁶ On this reading, a central question is whether the feature of conforming to ⁹⁵ See Letter to Garve, 12:257–258: “Not the investigation regarding the existence of God, of immortality etc. has been the point from which I started but the antinomy of pure reason: ‘the world has a beginning -: it has no beginning etc. to the fourth: there is freedom in the human being,—against this: there is no freedom but everything in him is natural necessity’; this antinomy was what first awoke me from the dogmatic slumber and drove me to the critique of reason itself in order to fix the scandal of the apparent contradiction of reason with itself.” ⁹⁶ The full passage from FM, 20:291 reads as follows: “What was earlier demonstrated a priori dogmatically in the Analytic, is here indisputably confirmed in the Dialectic through an experiment of reason, as it were, that it conducts on its own faculty.” The reason why Kant is here talking about a demonstration in the Analytic, rather

352

-     

the conditioning principle is already implicitly contained in Kant’s conception of things in themselves as described in section 6.1, or whether adding it to this conception amounts to making the conception more specific. In the latter case, a further important question would be how Kant can justify the assumption that things in themselves conform to the conditioning principle. Here is a somewhat radical suggestion, on which we get the conformity of things in themselves, including Kantian things in themselves, to the conditioning principle for free, so to speak. ‘Condition’ and ‘conditioned’ are among the concepts for which we need to distinguish a transcendental and empirical sense, in the way explicated in section 3.1. And if they are understood in the transcendental sense, the conditioning principle is a necessary conceptual truth. For something to be conditioned in the transcendental sense, or from the point of view of fundamental ontology, just means that, if it exists, then its condition exists as well. And since the same applies to each and every conditioned condition, it follows that, necessarily, if something exists that is conditioned in the transcendental sense, then all of its conditions, or the entire series of its conditions, exist as well.⁹⁷ On this reading, the conditioning principle trivially applies to all things whatsoever. It is thus not necessary to specifically include conformity to the conditioning principle in Kant’s conception of things in themselves, although it is, of course, true that Kantian things in themselves also conform to it. And since the conditioning principle is a necessary conceptual truth, no further justification for the claim that things in themselves conform to it is needed than the derivation just sketched. The obvious worry raised by this proposal is that it seems to cut the transcendental idealist off from his escape route from the antinomies. For if the conditioning principle applies to all things whatsoever, it also applies to appearances, which, in turn, seems to mean that not only the transcendental realist but also the transcendental idealist is forced to assume that the entire series of the various conditions of empirical objects exist, an assumption that leads straight into the antinomies. In response, on the proposed reading, the conditioning principle indeed also applies to appearances, but the transcendental idealist is not stuck with the antinomies because, on his view, the conditions of empirical objects that play a role in the genesis of the antinomies are not conditions in the transcendental sense but conditions in the empirical sense, precisely because he regards empirical objects as appearances and not things in themselves. That the antinomies rest on a confusion between transcendental and empirical conditions is precisely how Kant

than the Aesthetic, is that he is speaking more generally about transcendental idealism and the result that all of our substantive theoretical cognition (of a certain kind) is restricted to the empirical realm. But what is confirmed in the Dialectic, more specifically, is transcendental idealism—which, as we know, was earlier demonstrated in the Aesthetic. Also see Bxix–xx, note, and B534/A506, quoted at the beginning of this section in the main text. ⁹⁷ I am aware that at B364/A308 Kant says of the conditioning principle that it is “obviously synthetic; for even though the conditioned relates analytically to some condition, it does not analytically relate to the unconditioned.” I find this remark quite puzzling; if the conditioned relates analytically to some condition, why does it not analytically relate to all of its conditions (if those are not understood as constituting some additional thing)? And, as we know from Kant’s discussion, one of the ways in which the unconditioned figures in the antinomy is precisely as an infinite series of all conditions. But even if we take the claim that the conditioning principle is synthetic at face value, it need not be seen as conflicting with the view that this principle is a necessary conceptual truth. All that Kant’s claim means, strictly speaking, is that the conditioning principle is not an analytic judgment, as understood by Kant, that is, a categorical judgment whose predicate concept is contained in its subject concept, which, indeed, it is not. In order for the conditioning principle to count as something like an analytic judgment thus understood, the concept ‘conditioned’ would have to contain the concept ‘depends for its existence on the unconditioned’ or something like that. And I certainly agree that no such concept is contained in the concept ‘conditioned.’

     

353

himself characterizes the problem in the passage at B526–528/A498–499 quoted above, where he explains why he takes the cosmological syllogism to be fallacious, that is, the syllogism that has the conditioning principle as its major premise, the claim that empirical objects exist and are conditioned in various ways as its minor premise, and the claim that the entire series of all conditions of empirical objects exist as its conclusion. Kant identifies as the main problem with the cosmological syllogism an equivocation of the terms ‘conditioned’ and ‘condition.’ In the conditioning principle, ‘conditioned’ and ‘condition’ are understood in the transcendental sense, while in the claim that empirical objects are conditioned in various ways, ‘conditioned’ is understood in the empirical sense. On my reading, a condition in the transcendental sense is a condition for the existence of an object, or feature of an object, from the point of view of fundamental ontology; a condition in the empirical sense is a condition for the existence of an object, or feature of an object, with respect to the empirical level of reality, or from the point of view of the empirical level of reality. The transcendental conditions of empirical objects are the conditions for the possibility of experience because, from the point of view of fundamental ontology, empirical objects are appearances, which exist and have all of their determinations in virtue of being represented in experience. So, if empirical objects exist, the conditioning principle licenses us to infer that all conditions for the possibility of experience exist, for example, that there are human minds with certain cognitive faculties and things in themselves that affect sensibility.⁹⁸ But those are not the conditions whose existence is inferred in the cosmological syllogism. The cosmological syllogism focuses on what turn out to be merely empirical conditions of empirical objects, that is, the spaces and times that delineate them, the material parts that compose them, the causes that determine the changes in their states, and the other empirical objects that are involved in their generation. If empirical objects were things in themselves, that is, if, from the point of view of fundamental ontology, they did not depend for their existence and all of their determinations on being represented in experience, the conditions highlighted in the cosmological syllogism would be transcendental conditions for them, and the inference would go through. But since empirical objects are appearances, these conditions turn out to be merely empirical conditions, and the inference ends up being fallacious. If I had to choose, I would go with this second reading, on which the indirect argument for transcendental idealism indeed stands as a proper fallback demonstration in case “somebody is not satisfied by the direct proof in the Transcendental Aesthetic” (B534/A506).

⁹⁸ And, indeed, these inferences are among the arguments for various theses of critical idealism that we examined in section 5.10.2.

Concluding Postscript: The World According to Kant This completes my account of Kant’s critical idealism, understood as an ontological position, as developed in the Critique and associated theoretical writings. According to Kant, the world, understood as the sum total of everything that has reality, comprises several levels of reality, most importantly, the transcendental level and the empirical level. The transcendental level is a mind-independent level at which Kantian things in themselves exist; the empirical level is a mind-dependent level at which Kantian appearances exist. Things in themselves are mind-independent, appearances are fully mind-dependent. Things in themselves and appearances are numerically distinct and do not ontologically overlap in any way. Kantian outer appearances essentially are intentional objects of outer experience; Kantian inner appearances essentially are intentional objects of inner experience. Empirical objects are Kantian outer appearances, empirical space and time are constituted by the spatial and temporal determinations of outer appearances, pure space and time are (nothing but) forms of sensibility, and empirical selves, or empirical minds, are Kantian inner appearances. In contrast to other intentional objects, such as the intentional objects of fictions, dreams, hallucinations, illusions, and perceptions, Kantian appearances genuinely exist, that is, they exist from the point of view of fundamental ontology. This is due both to the special character of experience, in particular, the special character of outer experience and its conformity to Kant’s formal conditions of objectivity, and to the grounding of Kantian appearances in things themselves. Kantian things in themselves transcendentally affect sensibility and thereby bring about sensations, which provide the ‘matter’ for Kantian appearances and underwrite their existence. Kantian things in themselves are supersensible, non-spatial, and non-temporal, as well as distinct from God and thus finite. Each inner appearance is grounded in a unique Kantian thing in itself that is a human transcendental mind, and all outer appearances are grounded in Kantian things in themselves that are distinct from all human minds. What we commonly call ‘the external empirical world’ exists, including empirical space and time. Accordingly, there is also at least one Kantian thing in itself that is not a human mind. Moreover, there is at least one human being, that is, an entity whose ontologically basic parts include, minimally, a body (which is an empirical object), an empirical self (which is an empirical mind), and a transcendental self (which is a human transcendental mind). Since other intentional objects that are not Kantian appearances, although not genuine existents, are not nothing but have some reality and being, it is useful to conceive of Kantian reality as including yet another mind-dependent level to provide a home for these other fully minddependent entities—even if this conception goes beyond the direct textual evidence and may also go beyond Kant’s private, explicitly articulated thoughts on the matter. The ultimate basis for Kant’s case for transcendental idealism is the finitude of the human mind and, more specifically, its fundamentally uncreative nature in which this finitude manifests

356

 

itself. That the human mind is uncreative entails, among other things, that it necessarily includes a passive component, sensibility, which produces representations only in response to being affected. The essentially passive nature of sensibility and the corresponding essentially sensible nature of intuitions underwrite the main premises in Kant’s arguments for the existence of Kantian things in themselves, and in his master argument for transcendental idealism in the Transcendental Aesthetic. While Kant is a realist about Kantian things in themselves, he is a fictionalist about things in themselves understood as noumena or objects of the pure understanding, including things in themselves as conceived by the Leibniz-Wolffians. The Leibniz-Wolffian conception of things in themselves still occupies a special place in Kant’s thinking, however, in that he regards it as grounded in the nature of human reason, as a useful fiction that serves various fundamental epistemic needs of ours, and as ‘correct’ from the point of view of the pure understanding. This account of the world, in particular, the part of the account that concerns the supersensible realm, is fleshed out in various ways in Kant’s writings on moral philosophy, as well as the Critique of Judgment—which is just what one should expect, given Kant’s proclaimed aim to reach the “very same destination” as the Leibniz-Wolffians, namely, the cognition of the supersensible, “through the combination of the theoretical philosophy with the practical” (Letter to Kästner, August [?] 5, 1790, 11:186). The present investigation is limited to the context of Kant’s theoretical philosophy. The reconstruction of this more fleshed-out account in the context of Kant’s practical philosophy is a project for another day.

APPENDIX

Comprehension Tools A.1 Core Theses of Critical Idealism A.1.1 Core Thesis of Bold Critical Idealism Bold critical idealism lies at the strong end of the spectrum of a range of critical idealist positions. Kant seems to believe that bold critical idealism is true, and I also regard it as his official position. It comprises transcendental idealism (TI), empirical realism (ER), and several other claims about things in themselves and their relation to appearances. It is defined by the following theses.¹

(TI1-strong) (a) Empirical objects are not things in themselves in the transcendental sense, that is, they are not mind-independent; rather (b) empirical objects are Kantian outer appearances, that is, they essentially are intentional objects of outer experience as characterized in Kant’s theory of experience, and, thus, fully mind-dependent. (ER1) (i) There exist empirical objects; and (ii) empirical objects are things in themselves in the empirical sense, or with respect to the empirical level of reality, that is, they are mind-independent in the empirical sense, or with respect to the empirical level of reality. (TI2-pure-strong) (a) Pure space and pure time are neither things in themselves in the transcendental sense nor constituted by determinations of things in themselves in the transcendental sense (or in other words, things in themselves do not have any spatial or temporal determinations, which also means that things in themselves are not in pure space or pure time); rather (b) pure space and pure time are (nothing but) forms of sensibility and thus fully mind-dependent. (TI2-empirical-strong) (a) Empirical space and empirical time are neither things in themselves in the transcendental sense nor constituted by determinations of things in themselves in the transcendental sense; rather (b) empirical space and empirical time are constituted by determinations of appearances in the transcendental sense, and, thus, are fully mind-dependent. (ER2) (i) Empirical space and empirical time exist; and (ii) () empirical space and empirical time are mind-independent in the empirical sense, and () things in themselves in the empirical sense, namely, empirical objects, are in space and time and, thus, have spatial and temporal determinations. (TI3-strong) (a) Empirical selves are not things in themselves in the transcendental sense, that is, they are not mind-independent; rather (b) empirical selves are Kantian inner appearances, that is, they essentially are intentional objects of inner experience as characterized in Kant’s theory of experience, and, thus, fully mind-dependent. (ER3) (i) There exist empirical selves; and (ii) empirical selves are things in themselves in the empirical sense, or with respect to the empirical level of reality, that is, they are mind-independent in the empirical sense, or with respect to the empirical level of reality. (P--) Due to its essentially passive nature, outer sense can produce (a) outer sensations only in response to being materially transcendentally affected by things that are moderately strongly ‘outside us’ in the transcendental sense, that is, by

¹ Obviously, since the weaker versions of TI1, TI2, TI3, P, G, and E that we discussed in chapters 3, 4, and 5 are implied by the stronger versions listed here, these weaker versions are also included in bold critical idealism. But they need not be stated explicitly to characterize the position.

358

:  

things in themselves in the transcendental sense that are distinct from the thing in itself whose outer sense it is, and (b) a priori intuitions of space only in response to being formally transcendentally affected by the mind whose outer sense it is. (P--) Due to its essentially passive nature, inner sense can produce (a) inner sensations only in response to being materially transcendentally affected by the thing in itself whose inner sense it is and (b) a priori intuitions of time only in response to being formally transcendentally affected by the mind whose inner sense it is. (G--) Kantian outer appearances, that is, intentional objects of outer experience as characterized in Kant’s theory, are essentially grounded in finite things that are strongly ‘outside us’ in the transcendental sense, that is, things in themselves in the transcendental sense that are distinct from all human minds and God. (G--) Each Kantian inner appearance, that is, each intentional object of inner experience as characterized in Kant’s theory, is essentially grounded in one thing in itself in the transcendental sense that is (a) unique to the appearance and (b) a human transcendental mind. (E--) There is at least one thing in itself in the transcendental sense that grounds empirical objects and is neither a human mind nor God. (E--) transcendental mind.

There is at least one transcendental self that is a human

(E--) There is at least one human being, that is, an entity whose ontologically basic parts include, minimally, a body (which is an empirical object), an empirical self (which is an empirical mind), and a transcendental self (which is a transcendental mind). Additional foundational principles of bold critical idealism:

(U) Due to its essentially unoriginal nature, the imagination can produce pseudo-sensations only by reproducing and recombining sensations. (U) Due to its essentially finite nature, the human mind is ontologically uncreative and thus incapable of actively or spontaneously generating any ‘matter’ for the objects of its representations. (S--) Each human being has one unique empirical self and one unique transcendental self, which is its transcendental mind. (F-) God only creates things in themselves and does not otherwise meddle in the constitution of the empirical realm.

A.1.2 Core Theses of Timid Critical Idealism Timid critical idealism lies at the weak end of the spectrum of a range of critical idealist positions. It is the most guarded fallback position for a possible retreat in case bold critical idealism turns out to be problematic for reasons having to do with Kant’s epistemology and theory of cognition. Timid critical idealism differs from bold critical idealism in that it includes neither S- nor F- nor E-- and in that P-, P--, G--, G--,

:  

359

E--, and E-- are replaced by P--, P--, G--, G--, E-, and E--, respectively.

(P--) Due to its essentially passive nature, outer sense can produce (a) outer sensations only in response to being materially transcendentally affected by things in themselves in the transcendental sense and (b) a priori intuitions of space only in response to being formally transcendentally affected by the mind whose outer sense it is. (P--) Due to its essentially passive nature, inner sense can produce (a) inner sensations only in response to being materially transcendentally affected by things in themselves in the transcendental sense and (b) a priori intuitions of time only in response to being formally transcendentally affected by the mind whose inner sense it is. (G--) Kantian outer appearances, that is, intentional objects of outer experience as characterized in Kant’s theory, are essentially grounded in things in themselves in the transcendental sense. (G--) Kantian inner appearances, that is, intentional objects of inner experience as characterized in Kant’s theory, are essentially grounded in things in themselves in the transcendental sense. (E--) There is at least one thing in itself in the transcendental sense that grounds empirical objects. (E--) (a) There is at least one transcendental self, that is, a thing in itself in the transcendental sense that grounds an empirical self; and (b) there is at least one human transcendental mind.

A.2 Core Thesis of Transcendental Realism Transcendental realism is opposed to transcendental idealism and rejected by Kant; it is defined by the following theses.

(TR1) (i) There exist empirical objects; and (ii) empirical objects are (what a Kantian would call) things in themselves in the transcendental sense, that is, they are mind-independent. (TR2) (i) Space and time exist; and (ii) () space and time are mind-independent, and () (what a Kantian would call) things in themselves in the transcendental sense, namely, empirical objects, are in space and time and, thus, have spatial and temporal determinations.

A.3 Core Thesis of Ordinary Idealism Ordinary idealism (OI) is a rival form of idealism, which is rejected by Kant. It comprises empirical idealism and a certain view about mind-independent entities and their relation to appearances.

(OI1) Empirical objects are not (what a Kantian would call) things in themselves in the transcendental sense, or with respect to the transcendental level of reality, that is, they are not mind-

360

:  

independent; rather (b) empirical objects are (what a Kantian would call) appearances in the transcendental sense, that is, they are fully mind-dependent (or fully mind-dependent*).

(EI1) (a) Empirical objects are not (what a Kantian would call) things in themselves in the empirical sense; rather (b) empirical objects are (what a Kantian would call) appearances in the empirical sense, that is, they are (what a Kantian would call) mind-dependent in the empirical sense (or empirically mind-dependent*). (OI2) (a) Space and time are neither (what a Kantian would call) things in themselves in the transcendental sense nor constituted by determinations of (what a Kantian would call) things in themselves in the transcendental sense; rather (b) space and time are fully mind-dependent (or fully mind-dependent*). (EI2) (a) (What a Kantian would call) empirical space and empirical time are neither (what a Kantian would call) things in themselves in the empirical sense nor constituted by (what a Kantian would call) determinations of things in themselves in the empirical sense; rather (b) (what a Kantian would call) empirical space and empirical time are (what a Kantian would call) mind-dependent in the empirical sense. (G-OI) (What a Kantian would call) outer appearances are not grounded in any finite mind-independent things, except in the weak sense that they pseudo-exist and have all of their determinations and other ontological ingredients in virtue of being represented by finite minds.

A.4 Additional Theses, Definitions, and Terminology This section contains a collection of other important ‘named’ theses that make an appearance throughout the book, definitions of key technical Kantian terms, and explanations of special terminology introduced by me. The purpose of this section is not to provide a general glossary of all central Kantian terms. Affection Empirical affection: An empirical affection is an affection at the empirical level of reality of empirically real entities by empirically real entities. Transcendental affection: A transcendental affection is an affection at the transcendental level of reality of transcendentally real entities by transcendentally real entities. Material affection: A material affection is an affection by means of which sensations are generated. There are material affections at both the transcendental and the empirical level of reality. Material empirical affections are affections of our empirical mind, via our sense organs, by empirical objects; material transcendental affections are affections of sensibility by things in themselves. Formal transcendental affection: A formal transcendental affection is a self-affection, more specifically, an affection of sensibility by the mind’s own activity by means of which a priori intuitions are generated. Appearances Appearances in the general sense: An appearance in the general sense is an intentional and thus fully mind-dependent object. Kantian inner/outer appearances: A Kantian inner/outer appearance is an intentional object of inner/ outer experience as characterized in Kant’s theory of experience.

:  

361

Conditioning principle Conditioning principle: If something conditioned exists, then its condition and, indeed, the entire series of all of its conditions exist as well. Determinations, logical properties, and level-relative ontological properties Determinations of X: The determinations of X are X’s particular property instances and particular relation instances that determine X as the entity that it is at the level of reality at which, or the intentional ‘world’ in which, it exists. Logical properties: Logical properties are properties such as the property of being self-identical or being such that 2+2=4, which necessarily apply to all entities. Ontological properties: Ontological properties characterize the ontological status of objects and determine their ‘location’ in Kant’s scheme of reality, such as existing in the world of The Lord of the Rings, or being mind-independent with respect to the empirical level of reality. Entity and Existent X is an entity if, and only if, X has being, or reality. X is an existent if, and only if, X genuinely exists, or exists from the point of view of fundamental ontology. Empirical world The empirical world is the public, physical world that makes up part of the empirical level of reality. (Empirical selves also exist at the empirical level but they are not part of the empirical world.) The empirical world is the totality of all empirical objects and their determinations, and thus also includes empirical space and empirical time. Experience Formal conditions for the possibility of experience (= the ultimate a priori conditions for the possibility of us representing proper objects): The formal conditions for the possibility of experience are the forms of human cognitive faculties, including the forms of sensibility (space and time) and the forms of the understanding (the transcendental unity of apperception, and the pure categories). Formal conditions of experience/objectivity: The formal conditions of experience/objectivity are necessary conditions, reflecting the formal conditions for the possibility of experience, for a representation R of ours to (directly) represent a proper object, count as a piece of outer experience, and be objective, strictly speaking. They are that R presents its object as (1) a unified bearer of determinations, or a bearer of determinations that is one, (2) extended and in space, (3) an extensive magnitude and, hence, necessarily governed by the laws of mathematics, in particular, Euclidean geometry, and standard arithmetic, (4) having determinations that have intensive magnitudes and, as such, are also necessarily governed by the laws of mathematics, (5) persisting through time and remaining unchanged with respect to its quantum, (6) having states that are part of an objective temporal order and change only in response to the object’s being empirically affected by an external cause, and (7) mutually causally interacting with all objects that are simultaneous with it. Global formal condition of experience: It says that a collection of representations of which each conforms to the formal conditions of experience amounts to outer experience only if all of their intentional objects are compossible. Formal conditions of proper objecthood: The formal conditions of proper objecthood are necessary conditions, reflecting the formal conditions of objectivity, for an object to be a proper object for us. Ultimate material conditions for the possibility of experience: The ultimate material conditions for the possibility of experience are things in themselves. Material conditions for the possibility of experience: The material conditions for the possibility of experience are sensations and perceptions. Material conditions of experience: The material conditions of experience are necessary conditions, reflecting the material conditions for the possibility of experience, for a representation R of ours to be a piece of outer experience. They are that (1) (a) R presents its object O as having some ‘matter’, and (b) this ‘matter’ does not depend on special features of individual cognizers or special empirical

362

:  

circumstances, and (2) there is a perception P, and (a) R presents O as the proximate cause of the sensations contained in P, or (b) R presents O as causally connected to O*, which, in turn, it presents as the proximate cause of the sensations contained in P. Global material condition of experience: It says that a collection of representations of which each conforms to the material conditions of experience amounts to outer experience at time t only if, for all perceptions P available at t, one of the representations presents its intentional object as the proximate cause of the sensations contained in P. Ew-account at time t: An Ew-account at time t is a shared, or at least sharable, account of the empirical world that includes a list of all general laws and all specific empirical laws of nature that have been specified by science so far; a list of general commonsensical descriptions of causal laws proposed so far that are inductively based on observations and govern various different kinds of phenomena; a record of all appearances that have been perceived to date as well as of all appearances that have been inferred so far on the basis of known laws and the available empirical evidence; and a collection of empirical judgments and judgments of experience made so far concerning the determinations of these appearances, including their relations to each other. Experience: Experience is the conjunction of all actual ew-accounts. Pseudo-experience: Pseudo-experience is just like experience except that is based, not on genuine perceptions but on pseudo-perceptions. Forms of intuition, forms of sensibility Form of intuition in the broad sense: S is a form of intuition in the broad sense if, and only if, (a) S is responsible for all intuitions of finite objects of type T necessarily presenting these objects as having determinations of a corresponding type T*, and (b) S is a general scheme of order that necessarily governs all determinations of type T*. Form of sensibility/intuition: S is a form of sensibility/intuition if, and only if, (1) S essentially is an intentional object of a pure intuition of a certain type T, (2) S is responsible for all intuitions of type T of finite objects necessarily presenting these objects as having determinations of a corresponding type T*, and (3) S is a general scheme of order that necessarily governs all determinations of type T*. Form of experience: S is a form of experience if, and only if, (1) S essentially is an intentional object of a pure intuition of a certain type T, (2) S is responsible for experience necessarily presenting its objects as having determinations of a corresponding type T*, and (3) S is a general scheme of order that necessarily governs all determinations of type T*. Ideal, real Transcendentally ideal: An entity is transcendentally ideal if, and only if, it is fully mind-dependent from the point of view of fundamental ontology, or with respect to the transcendental level of reality. Transcendentally real: An entity is transcendentally real if, and only if, it exists at the transcendental level and from the point of view of fundamental ontology and is mind-independent from the point of view of fundamental ontology, or with respect to the transcendental level. Empirically real: An entity is empirically real if, and only if, it exists at the empirical level and from the point of view of fundamental ontology and is mind-independent with respect to the empirical level. Empirically ideal: An entity is empirically ideal if, and only if, it is mind-dependent with respect to the empirical level. Inner sense U---: No particular inner sense can belong to more than one thing in itself, and no thing in itself can have more than one particular inner sense. Levels of reality Transcendental level: The transcendental level of reality is the most fundamental, mind-independent level, at which transcendentally real entities exist; those entities are things in themselves (in the transcendental sense).

:  

363

Empirical level: The empirical level of reality is a less fundamental, mind-dependent level of reality, at which empirically real entities exist; those entities are Kantian appearances (in the transcendental sense). Mind-dependence and mind-independence Mind-(in)dependent being: The being of an entity E is mind-dependent if, and only if, necessarily, if E has being, there is a finite mind M, and E has being, at least partly, in virtue of the fact that M has [or could have] a possibly complex representation whose presentational content includes E; the being of an entity is mind-independent if, and only if, it is not mind-dependent. Strongly mind-dependent being: The being of an entity E is strongly mind-dependent if, and only if, necessarily, if E has being, there is a finite mind M, and E has being in virtue of the fact that M has [or could have] a possibly complex representation whose presentational content includes E. Mind-(in)dependence of an entity in the basic sense: An entity is mind-dependent in the basic sense if, and only if, its being is mind-dependent; an entity is mind-independent in the basic sense if, and only if, its being is mind-independent. Mind-(in)dependence of properties/ontological ingredients: A property P/an ontological ingredient OI of an entity E is mind-dependent if, and only if, either, necessarily, if E has P/comprises OI, then there is a finite mind M, and E has P/comprises OI, at least partly, in virtue of the fact that M has [or could have] a possibly complex representation whose presentational content includes E having P/comprising OI, or P/OI is a representation or feeling in inner sense; a property P/an ontological ingredient OI of an entity E is mind-independent if, and only if, it is not mind-dependent. Strong mind-dependence of properties/ontological ingredients: A property P/an ontological ingredient OI of an entity E is strongly mind-dependent if, and only if, either, necessarily, if E has P/comprises OI, then there is a finite mind M, and E has P/comprises OI in virtue of the fact that M has [or could have] a possibly complex representation whose presentational content includes E having P/comprising OI, or P/OI is a representation or feeling in inner sense. Strong mind-dependence* of properties/ontological ingredients: A property P/an ontological ingredient OI of an entity E is strongly mind-dependent* if, and only if, necessarily, if E has P/comprises OI, then there is a mind M, and E has P/comprises OI in virtue of the fact that M has a possibly complex representation R whose presentational content includes E having P/comprising OI, and there is a finite mind who could have a representation whose presentational content includes E having P/ comprising OI. Full mind-dependence: An entity E is fully mind-dependent if, and only if, all of E’s determinations and other ontological ingredients are strongly mind-dependent. Full mind-dependence*: An entity E is fully mind-dependent* if, and only if, all of E’s determinations and other ontological ingredients are strongly mind-dependent*. Mind-independence: An entity E is mind-independent if, and only if, E is mind-independent in the basic sense, and all of E’s determinations and other ontological ingredients are mind-independent. Empirical mind-independence: An entity is empirically mind-independent if, and only if, experience presents it as mind-independent, by presenting it as being in space. Empirical mind-dependence: An entity is empirically mind-dependent if, and only if, it is fully minddependent and not empirically mind-independent. Empirical mind-dependence*: An entity is empirically mind-dependent* if, and only if, it is fully mind-dependent* and not empirically mind-independent. Noumenon Noumenon in the positive sense: An object is a noumenon in the positive sense if, and only if, it can be cognized by the pure understanding alone, or in terms of pure concepts alone, based on nothing but logical reflection in the broad sense, such as analysis and comparison of concepts. Noumenon in the strong positive sense: An object is a noumenon in the strong positive sense if, and only if, it can be cognized by an intuitive intellect alone.

364

:  

Noumenon in the weak positive sense: An object is a noumenon in the weak positive sense if, and only if, it is an object of human pure understanding alone, that is, if, and only if, it is represented in terms of our pure concepts alone. Noumenon in the negative sense: An object is a noumenon in the negative sense if, and only if, it is not an object of human sensible intuition, that is, if, and only if, it is a non-sensible object. Objectively real Objectively real: A representation is objectively real if, and only if, it represents a really possible thing. Objectively real in the strong sense: A representation is objectively real in the strong sense if, and only if, it represents an actual thing. Reflectively objectively real: A representation R of a cognizer is reflectively objectively real if, and only if, the cognizer is able to justify the claim that R is objectively real. Ontological ingredient, ontologically basic part Ontological ingredient: The ontological ingredients of a thing are elements, constituents, or building blocks, that, in proper combination, make up the thing; they correspond to the entries in a recipe for making the thing in question in an ontology ‘cookbook.’ Possible ontological ingredients include bare particulars, substances, attribute instances, property instances, and various kinds of parts or elements (temporal, spatial, etc.). Ontologically basic part: A part P of an entity E is ontologically basic with respect to ontological kind K if, and only if, (1) P is of kind K, (2) either P does not contain any parts, or P contains only parts that are of kind K, and (3) P does not compose any larger part of E with other parts of E that are of kind K. Ontological overlap: A and B ontologically overlap if, and only if, (1) A is an ontological ingredient of B, (2) B is in an ontological ingredient of A, or (3) A and B share at least one ontological ingredient. ‘Outside us,’ ‘in us’ ‘In us’ in the transcendental sense: An object is ‘in us’ in the transcendental sense if, and only if, it is fully mind-dependent. ‘Outside us’ in the transcendental sense: An object is ‘outside us’ in the transcendental sense if, and only if, it is mind-independent. Strongly ‘outside us’ in the transcendental sense: An object is strongly ‘outside us’ in the transcendental sense if, and only if, the object is mind-independent and distinct from all human minds. Moderately strongly ‘outside me’ in the transcendental sense: An object is moderately strongly ‘outside me’ in the transcendental sense if, and only if, the object is mind-independent and distinct from my mind. ‘Outside us’ in the empirical sense: An object is ‘outside us’ in the empirical sense if, and only if, it is in empirical space, that is, if, and only if, experience presents it as being in space. Strongly ‘outside us’ in the empirical sense: an object is strongly ‘outside us’ in the empirical sense if, and only if, it is in empirical space and distinct from all human minds. Perceptions and sensations Confused perception of an appearance: An appearance A is confusedly perceived by means of perception P if, and only if, the intentional object of the unconscious ‘little’ empirical intuitions on which P supervenes is A’s intuitive counterpart. Intuitive counterpart: An object O1 is an intuitive counterpart of object O2 if, and only if, O1 and O2 both have intuitive properties, and O1 has all of the intuitive properties that O2 has. Intuitive property/determination: A property/determination is intuitive if, and only if, it can be represented in intuition. Genuine sensation: Genuine sensations are qualia-representations that are direct results of affections of sensibility by things in themselves.

:  

365

Pseudo-sensation: Pseudo-sensations are qualia-representations that are just like genuine sensations except for not being the direct result of affections of sensibility by things in themselves. Pseudosensations are generated by the imagination. Genuine perceptions: Genuine perceptions are the representations that our mind generates by means of the application of a certain arsenal of cognitive unification operations encapsulating the categories of quantity and quality to genuine sensations. Pseudo-perceptions: Pseudo-perceptions are representations that we mistakenly regard as genuine perceptions but that our mind generated in other ways, for example, by means of the application of cognitive operations to genuine sensations that differ from the operations employed in genuine perception, or by means of the application of various cognitive operations to pseudo-sensations. Problems Problem of affection: P and G are in tension with the unknowability doctrine, and conflict with the result of the Transcendental Analytic that the category of causality meaningfully applies only to appearances and objects of possible experience. Problem of existence: E is in tension with the unknowability doctrine, and conflicts with the result of the Transcendental Analytic that the category of actuality meaningfully applies only to appearances and objects of possible experience. Unknowability problem: That Kant makes all sorts of claims about things in themselves in his theoretical philosophy—for example, that they are supersensible, non-spatial, and non-temporal, and that they exist, affect us, and ground appearances—is in tension with the unknowability doctrine. Representations, and their features and objects Representation: A representation is a vehicle of content in the mind, a mental ‘item’ that represents something. Presentational content: The presentational content of a representation is the mental, that is, representation-immanent, content that is present to the mind of the cognizer who entertains the representation. Intentional object O of representation R: The intentional object O of a representation R is a particular kind of presentational content of R, which minimally satisfies the condition that O is a particular something that is (a) distinct from R and the cognizer that entertains R and (b) a bearer of determinations. Directly representing, and being represented in: A representation R directly represents object O (O is represented in R) if, and only if, O is R’s intentional object. Indirectly representing: A representation R indirectly represents an object O if, and only if, R is intentionally related to O but O is not R’s intentional object (that is, O is not captured by the presentational content of R or O exists ‘outside’ of R). Representing, and being represented by: A representation R represents object O (O is represented by R) if, and only if, R directly or indirectly represents O. Reference: A representation R refers to object O if, and only if, R is intentionally related to O, and O is a genuine existent. Ontological specification: A representation R ontologically specifies object O if, and only if, O has its being and has all of its determinations and other ontological ingredients in virtue of being represented in R. A representation R ontologically specifies a property P of object O if, and only if, O has P in virtue of R presenting O as having P. Immediate relation of a representation to its object: A representation relates immediately to an object if, and only if, it intentionally relates to the object without the mediation of any other representation; a representation relates mediately to an object otherwise. All intuitions relate immediately to objects; all concepts relate mediately to objects. Singularity/generality of a representation: A representation R is singular if, and only if, R’s presentational content is an individual, or, equivalently, if, and only if, R is completely determinate; a

366

:  

representation R is general if, and only if, R’s presentational content is not an individual, or, equivalently, if, and only if, R is not completely determinate. All intuitions are singular; all concepts are general. Completely determinate entity: An entity E is completely determinate if, and only if, (1) for every possible determinable, it is determinate whether E has it or not, and (2) for every determinable of E, it is determinate which corresponding determination E has. Completely determinate representation: A representation R of type T is completely determinate if and only if, (i) R represents a completely determinate entity E, (ii) R captures every determination of E that can be represented in representations of type T, and (ii) R is sufficient to individuate E in the sense of distinguishing it from all other entities. If a representation is completely determinate, then it is infinitely complex, in the sense of comprising infinitely many representations in it. Sameness Compositional sameness: An object O and objects X, Y, Z, . . . are compositionally the same if, and only if, O is the mereological fusion of X, Y, Z, . . . Collective sameness: A collection A and a collection B are collectively the same if, and only if, A and B are not numerically identical but share all of their most basic ingredients. Genetic sameness: Appearance E and thing in itself D are genetically the same if, and only if, (1) there is a thing T such that (a) E is derived from T by means of an ‘as it appears’ operation, (b) D is derived from T by means of an ‘as it is in itself ’’ operation, and (2) E and D share an ontological ingredient. The Self Empirical self: An empirical self is a Kantian inner appearance, that is, an intentional object of inner experience as characterized in Kant’s theory of experience. Empirical selves are empirical minds. Transcendental self: A transcendental self is a thing in itself that grounds an empirical self by affecting inner sense and bringing about inner sensations. Transcendental selves are human transcendental minds. Transcendental subject₁: The transcendental subject₁ is a projected bare particular = X that is presented as distinct from our representations of it and as a bearer of determinations. Each empirical self has the transcendental subject₁ as one of its ontological ingredients. (In the main text, ‘transcendental subject’ without any additional qualifications is to be understood as ‘transcendental subject₁’.) Transcendental subject₂: A transcendental subject₂ is a transcendental self. Human transcendental mind: A human transcendental mind is a transcendentally real part of the human being that is equipped with various transcendental cognitive faculties, including sensibility and the understanding, and is a subject of consciousness in self-consciously thinking. The Self: The Self is the one entity that each one of us is, that is, a particular human being that includes as its ontologically basic parts (at least) a transcendental self, which is a transcendental mind, an empirical self, which is an empirical mind, and a body. The multiple-parts view of human beings: A human being is a complex entity composed of several ontologically basic parts, namely, a sensible part and a supersensible part, the former of which contains two further, more specific ontologically basic parts, namely, an empirical self and a body, and the latter of which either is a transcendental self, or is composed of two further, more specific ontologically basic parts, namely, a transcendental self and the transcendental ground of its body, depending on whether the latter ground is indeed part of the human being or not. Space and time Pure space and time: Pure space and pure time are the objects represented by our most fundamental representations of space and time (that is, our pure intuitions of space and time); they are what we ultimately mean by ‘space’ and ‘time’. Empirical space and time: Empirical space and empirical time are the space and time in which empirical objects are in the empirical world.

:  

367

Fundamental feature of our ordinary representations of space: The fundamental feature of our ordinary representations of space is that we necessarily conceive of, and present, any spatial determination as being delineated in, and any determinate space, that is, any determinate figure and any determinate spatial volume, as a ‘cut-out’ part of the same all-encompassing space (that is, pure space). Spatial and temporal determinations: The spatial determinations of an object comprise (a) its extension, three-dimensionality, shape, and spatial size, where the latter is measured according to the Euclidean distance function (that is, the Pythagorean theorem) and specified with respect to a particular unit of measurement as well as (b) its spatial locations with respect to and its spatial distances from other appearances, where the former comprise instances of the relations of being in between and being next to, and the latter are of the form of being a particular Euclidean distance away from X specified with respect to a particular unit of measurement. The temporal determinations of an object comprise (a) its duration and temporal size (that is, the length of its duration), where the latter is determined according to a measure derived from the distance function of Galilean spacetime and specified with respect to a particular unit of measurement as well as (b) the temporal locations of its states with respect to and the temporal distances of its states from each other and the states of other appearances, where the former comprise instances of the relations of being before, being after, and being simultaneous with, and the latter are of the form of being a particular Galilean temporal distance before/after X specified with respect to a particular unit of measurement. Space-infected/time-infected properties: A property P is space-infected/time-infected if, and only if, P is essentially bound up with space/time in the sense that objects can have P only if they are in space/ time; a determination is space-infected/time-infected if, and only if, it is an instance of a spaceinfected/time-infected property. S-: Any determinate space, that is, any determinate figure and any determinate spatial volume, and, more generally, any spatial determination is necessarily contained in pure space, and any determinate time, that is, any determinate temporal duration, and, more generally, any temporal determination is necessarily contained in pure time. Accordingly, any space is either identical to, or necessarily contained in, pure space, and any time is either identical to, or necessarily contained in, pure time. S-: If x is in space S/time T, x is fully immersed in S/T such that x has some spatial/temporal determinations and, possibly, some space-infected/time-infected determinations, and the mode of being of all of x’s determinations and other ontological ingredients is the same as the mode of being of S/T. Things in themselves Thing in itself in the general sense: A thing in itself in the general sense is a mind-independent thing. Kantian thing in itself: A Kantian thing in itself is a thing in itself as characterized in Kant’s critical idealism. Leibniz-Wolffian thing in itself: A Leibniz-Wolffian thing in itself is a thing in itself as characterized in the Leibniz-Wolffian philosophy (as understood by Kant). Transcendental versus empirical distinction between things in themselves and appearances Transcendental distinction: The transcendental distinction is a distinction between things in themselves in the transcendental sense and appearances in the transcendental sense, that is, between (transcendentally) mind-independent entities and intentional objects, which are (transcendentally) fully mind-dependent. Empirical distinction: The empirical distinction is a distinction between things in themselves in the empirical sense and appearances in the empirical sense, that is, between empirically mindindependent entities, that is, empirical objects, and empirically mind-dependent entities, for example, the appearances of empirical objects in perception. Transcendental Object Transcendental object₁: The transcendental object₁ is a projected bare particular = X that is presented as (a) distinct from both the representations of it and the cognizer entertaining the representations

368

:  

and (b) a bearer of determinations. Each empirical object has the transcendental object₁ as one of its ontological ingredients. (In the main text, ‘transcendental object’ without any additional qualifications is to be understood as ‘transcendental object₁’.) Transcendental object₂: a thing in itself that grounds an empirical object. Two-world, two-aspect, one-object, multiple-object, and same-things views One object view: According to the one-object view, appearances and things in themselves are numerically identical. More precisely, every appearance is numerically identical to a thing in itself. Appearances are things (considered) as they appear to us, or qua bearers of appearance properties; things in themselves are the very same things (considered) as they are in themselves, or qua bearers of in-itself properties. Multiple-object view: According to the multiple-object view, appearances and things in themselves are numerically distinct. Same-things view: According to the same-things view, appearances and things in themselves are numerically distinct but the same things in some reasonable sense, where the latter minimally requires that they ontologically overlap. Two-world view: According to the two-world view, things in themselves and appearances are not the same things in any reasonable sense, which also implies that they are not numerically identical. Classic two-world view: According to the classic two-world view, things in themselves and appearances do not ontologically overlap in that the former are mind-independent, while the latter are fully mind-dependent. Empirical objects are understood to be appearances, and things in themselves are taken to exist and to affect sensibility and thereby produce sensations. Two-aspect view: According to the two-aspect view, things in themselves and appearances are the same things in some reasonable sense, where the latter minimally requires that they ontologically overlap. The two-aspect view comprises the one-object view and the same-things view. Methodological two-aspect view: According to the methodological two-aspect view, the distinction between appearances and things in themselves is not an ontological distinction but merely a distinction between different ways of considering, or different perspectives on, the same things. Unknowability Unknowability doctrine: All of our substantive theoretical cognition of a certain kind, namely, cognition in the strict sense, is restricted to the empirical realm, which means that things in themselves are unknowable or uncognizable for us in the strict sense.

A.5 The Foundational Structure of Bold Critical Idealism The following diagram depicts the foundational structure of critical idealism by displaying the order of derivation of the various core theses of bold critical idealism, summarized in section A.1.1. If an arrow points from A to Z, or A is located higher than Z in the diagram, then A is derived at an earlier stage than Z in Kant’s overall case for bold critical idealism. If A and Z are connected by an arrow that points from A to Z, either A implies Z, or A functions as one of several premises in an argument for Z. If several arrows that originate in different theses A, B, and C, come together and form a combined arrow that ends in Z, all of A, B, and C function as premises in an argument for Z. Note that no specific auxiliary premises are depicted. That is, the diagram is not a complete record of the overall argument. If an inference depends on other premises in addition to the theses at the origination points of the arrow that represents the inference, this is indicated in the diagram by the addition of a ‘+AP’’ diamond just before the endpoint of the relevant arrow, where ‘+AP’ stands for ‘plus additional premises.’² Thick arrows indicate that the

² So, if A and Z are connected by an arrow in the direction from A to Z, this does not necessarily mean that A is logically stronger than Z. A is logically stronger than Z if the arrow in question only connects A and Z, and there is

:  

369

inference in question is part of Kant’s main overall argument for bold critical idealism; thin arrows indicate that the inference in question is part of a supplemental argument. In addition to the core theses of critical idealism listed in section A.1.1, the diagram also features P-- and E--, listed in section A.1.2, as well as TI1, TI2-pure, TI2-empirical, TI3, and G--, since they constitute important intermediate conclusions in the overall argument. In order to facilitate the reading of the diagram, I re-state the latter theses here:

(TI1) (a) Empirical objects are not things in themselves in the transcendental sense, or with respect to the transcendental level of reality, that is, they are not mind-independent; rather (b) empirical objects are appearances in the transcendental sense, that is, they are fully mind-dependent. (TI2-pure) (a) Pure space and pure time are neither things in themselves in the transcendental sense nor constituted by determinations of things in themselves in the transcendental sense; rather (b) pure space and pure time are fully mind-dependent. (TI2-empirical) (a) Empirical space and empirical time are neither things in themselves in the transcendental sense nor constituted by determinations of things in themselves in the transcendental sense; rather (b) empirical space and empirical time are fully mind-dependent. (TI3) (a) Empirical selves are not things in themselves in the transcendental sense, or with respect to the transcendental level of reality, that is, they are not mind-independent; rather (b) empirical selves are appearances in the transcendental sense, that is, they are fully mind-dependent. (G--) Kantian outer appearances, that is, intentional objects of outer experience as characterized in Kant’s theory, are essentially grounded in things that are strongly ‘outside us’ in the transcendental sense, that is, things in themselves in the transcendental sense that are distinct from all human minds.

no ‘+AP’ diamond. But if there is a ‘+AP’ diamond, or if the arrow originates in more than one thesis, A need not be stronger than Z; indeed, there are even cases of this kind where Z implies A.

TI2-pure-strong-b

Master Argument part (B), step 2

+AP

Copernican Turn to explain the possibility of a priori congnitions

TI2-emprical-a/ TI2-empirical-strong-a

TI2-empirical-b

+AP

Argument from geometry/the general theory of motion

TI1-a/ TI1-strong-a

Indirect argument based on the antinomies

+AP

ER2-ii

+AP

ER1-ii

TI1-strong-b

+AP

+AP

TI2-empirical-strong-b

TI2-pure-a/ TI2-pure-strong-a

TI3-a/ TI3-strong-a

We want to avoid skepticism about the existence of the empirical world

TI1-b

+AP

TI2-pure-b

Master Argument part (A)/ Master Argument part (B), step 1

Uncreativity

TI3-b

+AP

ER3-ii

+AP

TI3-strong-b

+AP

+AP

Existence-inner-moderate-b

Existence-human-being

+AP

ER2-i

+AP

ER1-i

+AP

ER3-i

Self-consciousness

Existence-outer-bold

Grounding-outer-bold

Grounding-outer-strong

+AP

Passivity-outer-strong

Self-non-proliferation

Passivity-inner-weak

Existence-inner-strong

Grounding-inner-strong

+AP

Passivity-inner-strong

Finite-Creation

Bibliography Abela, Paul. (2002) Kant’s Empirical Realism, Oxford: Oxford University Press. Adams, Robert Merrihew. (1994) Leibniz—Determinist, Theist, Idealist, Oxford: Oxford University Press. Adams, Robert Merrihew. (1997) “Things in Themselves,” Philosophy and Phenomenological Research 62 (4): 801–25. Adickes, Erich. (1920) Kants Opus Postumum, dargestellt und beurteilt, Kant-Studien Ergänzungsheft 50, Berlin: Reuther und Reichert. Adickes, Erich. (1924) Kant und das Ding an sich, Berlin: Rolf Heise. Adickes, Erich. (1927) Kant und die Als-Ob-Philosophie, Stuttgart: Frommanns Verlag. Adickes, Erich. (1928) “Erwiderung auf den Aufsatz von R. Schmid: ‘Kant—Vaihinger—Adickes’,” Annalen der Philosophie und philosophischen Kritik 7, 165–8. Adickes, Erich. (1929) Kants Lehre von der doppelten Affektion unseres Ich als Schlüssel zu seiner Erkenntnistheorie, Tübingen: J.C.B. Mohr. Allais, Lucy. (2004) “Kant’s One World: Interpreting ‘Transcendental Idealism’,” British Journal for the History of Philosophy 12: 655–84. Allais, Lucy. (2006) “Intrinsic natures: A Critique of Langton on Kant,” Philosophy and Phenomenological Research 73: 143–69. Allais, Lucy. (2007) “Kant’s Idealism and the Secondary Quality Analogy,” Journal of the History of Philosophy 45: 459–84. Allais, Lucy. (2010) “Kant’s Argument for Transcendental Idealism in the Transcendental Aesthetic,” Proceedings of the Aristotelian Society 110: 47–75. Allais, Lucy. (2015) Manifest Reality, Kant’s Idealism and His Realism, Oxford/New York: Oxford University Press. Allison, Henry. (1983) Kant’s Transcendental Idealism: An Interpretation and Defense, New Haven, CT: Yale University Press. Allison, Henry. (1987) “Transcendental Idealism: the ‘two aspect view’,” in Bernard den Ouden (ed.), New Essays on Kant, New York: Peter Lang Publishing, 155–78. Allison, Henry. (1996) “Transcendental Idealism: A Retrospective,” in Idealism and Freedom: Essays on Kant’s Theoretical and Practical Philosophy, Cambridge/New York: Cambridge University Press, 3–26. Allison, Henry. (2004) Kant’s Transcendental Idealism: An Interpretation and Defense, Revised and enlarged edition, New Haven, CT: Yale University Press. Allison, Henry. (2006) “Transcendental Realism, Empirical Realism, and Transcendental Idealism,” Kantian Review 11: 1–28. Allison, Henry. (2007) “Debating Allison on Transcendental Idealism,” Kantian Review 12 (1): 24–39. Ameriks, Karl. (2003) Interpreting Kant’s Critiques, Oxford, New York: Oxford University Press. Ameriks, Karl. (2003a) “Introduction: The Common Ground of Kant’s Critiques,” in Interpreting Kant’s Critiques, Oxford, New York: Oxford University Press, 1–48. Ameriks, Karl. (2003b) “Recent Work on Kant’s Theoretical Philosophy,” in Interpreting Kant’s Critiques, Oxford, New York: Oxford University Press, 76–97. Ameriks, Karl. (2003c) “Kantian Idealism Today,” in Interpreting Kant’s Critiques, Oxford, New York: Oxford University Press, 98–111. Ameriks, Karl. (2003d) “Kant and Short Arguments to Humility,” in Interpreting Kant’s Critiques, Oxford, New York: Oxford University Press, 135–57.

372



Ameriks, Karl. (2011) “Kant’s Idealism on a Moderate Interpretation,” in Dennis Schulting and Jacco Verburgt (eds.), Kant’s Idealism. New Interpretations of a Controversial Doctrine, Dordrecht, Heidelberg, London, New York: Springer, 29–53. Aquila, Richard. (1979) “Things in Themselves and Appearances: Intentionality and Reality in Kant,” Archiv für Geschichte der Philosophie 61: 293–307. Aquila, Richard. (1983) Representational Mind: A Study of Kant’s Theory of Knowledge, Bloomington, IN: Indiana University Press. Aquila, Richard. (2003) “Hans Vaihinger and some Recent Intentionalist Readings of Kant,” Journal of the History of Philosophy 41 (2): 231–250. Aquinas, Thomas St. (1947) Summa Theologica, translated by Fathers of the Dominican Provice, 3 volumes, New York: Benziger Brothers. Aristoteles. (1831–1870) Aristotelis Opera, herausgegeben von I. Bekker, C.A. Brandis und H. Bonitz, Berlin: Deutsche Akademie der Wissenschaften zu Berlin. Bauch, Bruno. (1917) Immanuel Kant, Berlin und Leipzig: Göschensche Verlagshandlung. Baum, Manfred. (1986) Deduktion und Beweis in Kants Transzendentalphilosophie, Königstein: Athenäum. Baum, Manfred. (2011) “Objects and Objectivity in Kant’s First Critique,” in Dennis Schulting and Jacco Verburgt (eds.), Kant’s Idealism. New Interpretations of a Controversial Doctrine, Dordrecht, Heidelberg, London, New York: Springer, 55–70. Baumgarten, Alexander Gottlieb. (1757, fourth edition) Metaphysica, Halle/Magdeburg: Herman Hemmerde; cited according to the historical-critical edition by Günter Gawlick and Lothar Kreimendahl, Stuttgart: Frommann-Holzboog, 2011. Beck, Jacob Sigismund. (1793) Erläuternder Auszug aus den critischen Schriften des Herrn Prof. Kant, auf Anrathen desselben. Erster Band, welcher die Critik der spekulativen und practischen Vernunft enthält, Riga: Hartknoch. Beck, Jacob Sigismund. (1794) Erläuternder Auszug aus den critischen Schriften des Herrn Prof. Kant, auf Anrathen desselben. Zweiter Band, welcher die Kritik der Urtheilskraft und die metaphysischen Anfangsgründe der Naturwissenschaft enthält, Riga: Hartknoch. Beck, Jacob Sigismund. (1796a) Erläuternder Auszug aus den critischen Schriften des Herrn Prof. Kant, auf Anrathen desselben. Dritter Band: Einzig möglicher Standpunckt, aus welchem die critische Philosophie beurtheilt werden muss, Riga: Hartknoch. Beck, Jacob Sigismund. (1796b) Grundriss der critischen Philosophie, Halle: Rengersche Buchhandlung. Berkeley, George. (1710) A Treatise Concerning the Principles of Human Knowledge, in Works, 2: 20 115. Berkeley, George. (1713) Three Dialogues Between Hylas and Philonous, in Works, 2: 164 265. Berkeley, George. (1744) Siris: A Chain of Philosophical Reflexions and Inquiries Concerning the Virtues of Tar-water, and divers other Subjects connected together and arising One from Another, in Works 5: 26–168. Berkeley, George. (1781) Philosophische Werke. Erster Teil. Aus dem Englischen übersetzt und mit einigen Nachrichten aus dem Leben und den übrigen Schriften desselben versehen, Leipzig: Schwickert. Berkeley, George. (1948–1957) The Works of George Berkeley, Bishop of Cloyne, edited by A. A. Luce and T. E. Jessop, 9 vols., London: Thomas Nelson; cited as ‘Works’, by volume and page number. Bird, Graham. (1962) Kant’s Theory of Knowledge, New York: Humanities Press. Brandom, Robert. (1981) “Leibniz and Degrees of Perception,” Journal of the History of Philosophy 19: 447–479. Brastberger, Gebhard Ullrich. (1790) Untersuchung über Kants Kritik der reinen Vernunft, Halle: Jacob Gebauer. Brastberger, Gebhard Ullrich. (1792) “Schlüssel zu einer leichten und natürlichen Erklärung der Vernunft-Critik, Philosophen zur Prüfung vorgelegt,” Philosophisches Magazin, Band 4, Viertes Stück, herausgegeben von Johann August Eberhard, Halle: Gebauer, 381 405. Breitenbach, Angela. (2004) “Langton on things in themselves: a critique of Kantian Humility,” Studies in the History and Philosophy of Science 35: 137–48.



373

Brentano, Franz. (1874) Psychologie vom Empirischen Standpunkte, 2 Bände, Leipzig: von Duncker und Humboldt. Brewer, Bill. (2011) Perception and its Objects, Oxford/New York: Oxford University Press. Cassirer, Ernst. (1921) Zur Einsteinschen Relativitätstheorie, Erkenntnistheoretische Betrachtungen, Berlin: Cassirer Verlag. Cohen, Hermann. (1871) Kants Theorie der Erfahrung, Berlin: Harrwitz und Gossmann, Dümmlers Verlagsbuchhandlung. Cohen, Hermann. (1885) Kants Theorie der Erfahrung, second, revised and expanded edition, Berlin: Harrwitz und Gossmann, Dümmlers Verlagsbuchhandlung. Collins, Arthur. (1999) Possible Experience. Understanding Kant’s Critique of Pure Reason, Berkeley and Los Angeles: University of California Press. Descartes, René. (1644) Principia Philosophiae, in Oeuvres de Descartes, Volume 8. Descartes, René. (1649 tertia edition, *1641) Meditationes de Prima Philosophia in qua Dei existentia et animæ immortalitas demonstrator. His adjuncta sunt variae objections doctorum virorum in istas de Deo et anima demonstrationes; Cum Responsionibus Authoris, in Oeuvres de Descartes, Volume 7. Descartes, René. (1974–1989) Oeuvres de Descartes, edited by Charles Adam and Paul Tannery, 11 volumes, Paris: Libraire Philosophique J. Vrin; cited as ‘AT’, by volume and page number. Drobisch, Mortiz Wilhelm. (1885) Kants Dinge an sich und sein Erfahrungsbegriff, Hamburg und Leipzig: Leopold Voss. Emundts, Dina. (2008) “Kant’s Critique of Berkeley’s Concept of Objectivity,” in Daniel Garber and Béatrice Longuenesse (eds.), Kant and the Early Moderns, Princeton, NJ: Princeton University Press. Erdmann, Benno. (1878) Kants Kritizismus in der ersten und zweiten Auflage der Kritik der reinen Vernunft, Leipzig: Leopold Voss. Erdmann, Benno. (1881) Nachträge zu Kants Kritik der reinen Vernunft, aus Kants Nachlass, Kiel: Lipsius und Tischer. Ewing, Alfred Cyril. (1934) Idealism: A Critical Survey, London: Methuen. Falkenstein, Lorne. (1989) “Kant’s Argument for the Non-Spatiotemporality of Things in Themselves,” Kant-Studien 80: 265–83. Falkenstein, Lorne. (1995) Kant’s Intuitionism, Toronto: Toronto University Press. Falkenstein, Lorne. (2001) “Debate: Langton on Things in Themselves. Critique of Kantian Humility,” Kantian Review 5: 49–72. Fichte, Johann Gottlieb. (*1775–1814) Briefwechsel, herausgegeben von Hans Schulz, 2 Bände, Leipzig: Haessel, 1930; cited as Briefwechsel, by volume and page number. Fichte, Johann Gottlieb. (1792) “Recension des Aenesidemus,” in Sämmtliche Werke, 1: 1–26. Fichte, Johann Gottlieb. (1794) Grundlage der gesammten Wissenschaftslehre, in Sämmtliche Werke, 1: 83–328. Fichte, Johann Gottlieb. (1795) “Vergleichung des vom Herrn Prof. Schmid aufgestellten Systems mit der Wissenschaftslehre,” in Sämmtliche Werke, 2: 421–58. Fichte, Johann Gottlieb. (1797a) Erste Einleitung in die Wissenschaftslehre, in Sämmtliche Werke, 1: 417–49. Fichte, Johann Gottlieb. (1797b) Zweite Einleitung in die Wissenschaftslehre, in Sämmtliche Werke, 1: 451–518. Fichte, Johann Gottlieb. (1845–46) Sämmtliche Werke, edited by I. H. Fichte, 8 Bände, Berlin: Veit & Comp, cited as ‘SW’, by volume and page number. Fischer, Kuno. (1865, *1852) System der Logik und Metaphysik oder Wissenschaftslehre, zweite Auflage, Heidelberg: Friedrich Basserman. Fischer, Kuno. (1869, *1860) Immanuel Kant und seine Lehre. Erster Teil: Enstehung und Grundlegung der kritischen Philosophie, Band III der Geschichte der neueren Philosophie, zweite Auflage, Mannheim: Friedrich Basserman. Fischer, Kuno. (1870) Anti-Trendelenburg, Jena: Hermann Dabis. Friebe, Cord. (2007) “Über einen Einwand gegen die Zwei-Aspekte-Interpretation von Kants Unterscheidung zwischen Erscheinung und Ding an sich,” Zeitschrift für philosophische Forschung 61 (2): 229–35.

374



Friedman, Michael. (1983) Foundations of Space-Time Theories: Relativistic Physics and Philosophy of Science, Princeton, NJ: Princeton University Press. Friedman, Michael. (1992) Kant and the Exact Sciences, Cambridge, MA: Harvard University Press. Friedman, Michael. (2013) Kant’s Construction of Nature: A Reading of Kant’s Metaphysical Foundations of Natural Science, Cambridge/New York: Cambridge University Press. Garber, Daniel. (1985) “Leibniz and the Foundations of Physics: the Middle Years”, in K. Okruhlik and J. R. Brown (eds.), The Natural Philosophy of Leibniz, Dordrecht: Reidel, 27–130. Garber, Daniel. (1995) “Leibniz: Physics and Philosophy,” in Nicholas Jolley (ed.), The Cambridge Companion to Leibniz, Cambridge: Cambridge University Press, 270–352. Garve, Christian, and Feder, Johann Georg Heinrich. (1782) “Kritik der reinen Vernunft. Von Immanuel Kant. 1781”, Zugaben zu den Göttinger Gelehrten Anzeigen, June 19, Drittes Stück, 40–8. Guyer, Paul. (1987) Kant and the Claims of Knowledge, Cambridge, New York: Cambridge University Press. Guyer, Paul. (1989) “The Rehabilitation of Transcendental Idealism?,” in E. Schaper and W. Vossenkuhl (eds.), Reading Kant, Oxford: Blackwell. Guyer, Paul. (2007) “Debating Allison on Transcendental Idealism,” Kantian Review 12 (1): 10–23. Guyer, Paul. (2016) “Arguing for Transcendental Idealism: Lucy Allais on Manifest Reality,” Kantian Review 21 (2): 261–72. Hanna, Robert. (2006) Kant, Science, and Human Nature, Oxford: Clarendon Press. Hartmann, Eduard von. (1875) Kritische Grundlegung des Transcendentalen Realismus, second edition (title of the first edition from 1781: Das Ding an sich und seine Beschaffenheit) Berlin: Carl Duncker. Hartz, Glenn, and Cover, Jan. (1988) “Space and Time in the Leibnizian Metaphysic”, Noûs 22: 493–519. Heidegger, Martin. (1929) Kant und das Problem der Metaphysik, Freiburg: Vittorio Klostermann; cited according to the fifth edition from 1991. Heimsoeth, Heinz. (1924) “Metaphysische Motive in der Ausbildung des kritischen Idealismus,” Kant-Studien 29: 121–59. Herring, Herbert. (1953) Das Problem der Affektion bei Kant, Kant-Studien Ergänzungshefte 67, Köln: Kölner Universitätsverlag. Hoffman, Paul. (2002) “Descartes’s Theory of Distinction,” Philosophy and Phenomenological Research 64 (1): 57–78. Hogan, Desmond. (2009) “How to Know Unknowable Things in Themselves,” Noûs 43 (1): 49–63. Hortsmann, Rolf Peter. (1976) “Space as Intuition and Geometry,” Ratio 18: 17–30. Hume, David. (1978) Treatise of Human Nature, edited by L. A. Selby-Bigge and P. H. Nidditch, Oxford: Oxford University Press. Jacobi, Friedrich Heinrich. (1787) David Hume über den Glauben oder Idealismus und Realismus. Ein Gespräch. Beilage ‘Über den transzendentalen Idealismus’, Breslau: G. Loewe. Jakob, Ludwig Heinrich. (1786) Prüfung der Mendelssohnschen Morgenstunden oder aller spekulativischen Beweise für das Dasein Gottes, Leipzig: Johann Samuel Heinsius. Jarroush, Sami S. (2010) I am the Body, the Mind, and the Soul, Xlibris. Jauernig, Anja. (2010) “Disentangling Leibniz’s views on relations and extrinsic denominations,” Journal of the History of Philosophy 48 (2): 171–205. Jauernig, Anja. (2011) “Kant, the Leibnizians, and Leibniz,” in Brandon Look (editor), The Continuum Companion to Leibniz, London/New York: Thoemmes Continuum Press, 289–309. Jauernig, Anja. (2013) “The synthetic nature of geometry, and the role of construction in intuition,” in Stefano Bacin, Alfredo Ferrarin, Claudio La Rocca, and Margit Ruffing (eds.), Kant und die Philosophie in weltbürgerlicher Absicht. Akten des XI. Internationalen Kant Kongresses 2010 in Pisa, 5 Bände, Berlin/New York: Walter de Gruyter Verlag, Band 5, 89–100. Jauernig, Anja. (2019) “Finite minds and their representations in Leibniz and Kant,” Internationales Jahrbuch des Deutschen Idealismus/International Yearbook of German Idealism 17, 47 80. Jauernig, Anja. (forthcoming 2021a) “Kant on the (alleged) Leibnizian misconception of the difference between sensible and intellectual representations,” in Brandon Look (editor), Leibniz and Kant, Oxford/New York: Oxford University Press.



375

Jauernig, Anja. (forthcoming 2021b) “The Labyrinth of the Continuum: Leibniz, the Wolffians, and Kant on Matter and Monads,” in Karl Schafer and Nicholas Stang (ed.), The Sensible and Intelligible Worlds: New Essays on Kant’s Metaphysics and Epistemology, New York: Oxford University Press. Jauernig, Anja. (in preparation) Thought and Cognition according to Kant—Our Cognitive Access to Things in Themselves and Appearances in Kant’s Critical Philosophy. Kant, Immanuel. (1900–2020) Gesammelte Schriften, Band 1 22 herausgegeben von der Preußischen Akademie der Wissenschaften, Band 23 herausgegeben von der Deutschen Akademie der Wissenschaften zu Berlin, von Band 24 an herausgegeben von der Akademie der Wissenschaften zu Göttingen, Berlin: de Gruyter (Georg Reiner). Kant, Immanuel. (1924) Die philosophischen Hauptvorlesungen Immanuel Kants. Nach den neu aufgefundenen Kollegheften des Grafen Heinrich zu Dohna-Wundlacken, herausgeben von Arnold Kowalewski, München, Leipzig: Rösl & Cie; cited as ‘V-Anth/Dohna, Kowalewski’. Kemp Smith, Norman. (1918) A Commentary to Kant’s ‘Critique of Pure Reason’, London: Macmillan and Co. Kitcher, Patricia. (2011) Kant’s Thinker, Oxford, New York: Oxford University Press. Kuehn, Manfred. (2001) Kant. A Biography, Cambridge, New York: Cambridge University Press. Lange, Friedrich Albert. (1866) Geschichte des Materialismus und Kritik seiner Bedeutung in der Gegenwart, Iserlohn: J. Baedecker. Langton, Rae. (1998) Kantian Humility. Our Ignorance of Things in Themselves, Oxford, New York: Oxford University Press. Leibniz, Gottfried Wilhelm. (1684) Meditationes de Cognitione, Veritate, et Ideis, in Sämtliche Schriften und Briefe, VI.4A, 585 92. Leibniz, Gottfried Wilhelm. (1686) Discours de Métaphysique, in Sämtliche Schriften und Briefe, VI.4B, 1529 88. Leibniz, Gottfried Wilhelm. (mid-1680s) “De modo distinguendi phaenomena realia ab imaginariis,” in Sämtliche Schriften und Briefe, A VI.4B:1500–1504. Leibniz, Gottfried Wilhelm. (1695) “Systeme nouveau de la nature et de la communication des substances, aussi bien que de l’union qu’il y a entre l’ame et le corps,” in Die Philosophischen Schriften, 4: 477 87. Leibniz, Gottfried Wilhelm. (1698) “Eclaircissement des difficultés que Monsieur Bayle a trouvées dans le systeme nouveau de l’union de l’ame et du corps,” in Die Philosophischen Schriften, 4: 517 24. Leibniz, Gottfried Wilhelm. (1704) Nouveaux Essais sur l’Entendement Humain, in Sämtliche Schriften und Briefe, VI.6. Leibniz, Gottfried Wilhelm. (1714a) Principes de la Nature et de la Grace, fondés en raison, in Die Philosophischen Schriften, 6: 598 606. Leibniz, Gottfried Wilhelm. (1714b) “La Monadologie,” in Die Philosophischen Schriften, 6: 607 23. Leibniz, Gottfried Wilhelm. (1720) Merckwürdige Schrifften, welche zwischen dem Herrn Baron von Leibnitz und dem Herrn D. Clarcke . . . gewechselt, und mit einer Vorrede Herrn Christian Wolffens, nebst einer Antwort Herrn Ludwig Philipp Thümmigs, edited and translated by Heinrich Köhler, Jena. Leibniz, Gottfried Wilhelm. (1740) Des Freyherrn von Leibniz kleinere Philosophische Schriften . . . , edited by Caspar Jacob Huth, Jena. Leibniz, Gottfried Wilhelm. (1875–1890) Die Philosophischen Schriften von Gottfried Wilhelm Leibniz, edited by C. I. Gerhardt, 7 Bände, Berlin: Weidmann; cited as ‘G’, by volume and page number. Leibniz, Gottfried Wilhelm. (1923–) Sämtliche Schriften und Briefe, Berlin-Brandenburgische Akademie und Göttinger Akademie, Berlin: Akademie Verlag; cited as ‘A’, by series, volume, and page number. Liebmann, Otto. (1865) Kant und die Epigonen, Stuttgart: Carl Schober. Lipscomb, Benjamin, and Krueger, James (eds.). (2010) Kant’s Moral Metaphysics, Berlin, New York: De Gruyter.

376



Locke, John. (*1690) An Essay Concerning Human Understanding, edited by P. H. Nidditch. Oxford: Clarendon Press, 1970, cited by book, chapter, section, and page number. Longuenesse, Béatrice. (1998) Kant and the Capacity to Judge, Princeton, NJ: Princeton University Press. Longuenesse, Béatrice. (2017) I, Me, Mine: Back to Kant, and Back Again, Oxford, New York: Oxford University Press. Maaß, J. G. E. (1788) “Über die transcendentale Aesthetik,” Philosophisches Magazin, Band 1, Erstes Stück, 117–49. McDaniel, Kris. (2017) The Fragmentation of Being, Oxford/New York: Oxford University Press. McGuire, J. E. (1976) “ ‘Labyrinthus Continui’: Leibniz on Substance, Activity, and Matter,” in P. K. Machamer and R. G. Turnbull (eds.), Motion and Time, Space and Matter: Interrelations in the History and Philosophy of Science, Columbus: Ohio State University Press, 291–326. McRae, Robert. (1976) Leibniz: Perception, Apperception, and Thought, Toronto: University of Toronto Press. Maimon, Salomon. (1790) Versuch über die Transcendental-Philosophie, mit einem Anhang über die symbolische Erkenntnis, Berlin: Christian Friedrich Voß und Sohn. Maimon, Salomon. (1794) Versuch einer neunen Logik oder Theorie des Denkens, nebst angehängten Briefen des Philaletes an Aenesidemus, Berlin; Ernst Felisch. Maimon, Salomon. (1797) Kritische Untersuchungen über den menschlichen Geist oder das höhere Erkenntnis- und Willensvermögen, Leipzig: G. Fleischer. Marshall, Collin. (2013) “Kant’s Appearances and Things in Themselves as Qua-Objects,” Philosophical Quarterly 63 (252): 520–45. Martin, Gottfried. (1948) “Die metaphysischen Probleme der Kritik der reinen Vernunft,” Zeitschrift für philosophische Forschung 2 (2/3): 315 42. Martin, Gottfried. (1951) Immanuel Kant: Ontologie und Wissenschaftstheorie, Köln: Kölner Universitätsverlag. Matthews, H. E. (1969) “Strawson on Transcendental Idealism,” Philosophical Quarterly 19 (76): 204–20. Meinong, Alexius. (1899) “Über Gegenstände höherer Ordnung und deren Verhältnis zur inneren Wahrnehmung”, Zeitschrift für Psychologie und Physiologie der Sinnesorgane 21: 182–272. Meinong, Alexius. (1904) “Über Gegenstandstheorie” in Untersuchungen zur Gegenstandstheorie und Psychologie, Leipzig: J. A. Barth, 1–51. Meinong, Alexius. (1915) Über Möglichkeit und Wahrscheinlichkeit. Beiträge zur Gegenstandstheorie und Erkenntnistheorie, Leipzig: J. A. Barth. Melnick, Arthur. (1973) Kant’s Analogies of Experience, Chicago: University of Chicago Press. Moltmann-Wende. (1995) I Am My Body: A Theology of Embodiment, London/New York: Continuum Press. Natorp, Paul. (1912) “Kant und die Marburger Schule,” Kant-Studien 17: 193 221. Newton, Isaac. (*1668?) “De Gravitatione,” in A. R. Hall, and Hall, M. B. (editors and translators), Unpublished Scientific Papers of Isaac Newton, a selection from the Portsmouth Collection in the University Library, Cambridge: Cambridge University Press, 1962. Parkinson, G. H. R. (1982) “The Intellectualization of Thought and Appearances,” in Michael Hooker (ed.), Leibniz: Critical and Interpretative Essays, Minneapolis: University of Minnesota Press, 3 20. Paton, Herbert James. (1936) Kant’s Metaphysics of Experience, 2 vols., London: Allen & Unwin. Paulsen, Friedrich. (1899) Immanuel Kant―Sein Leben und seine Lehre, Stuttgart: Fr. Frommanns Verlag. Paulsen, Friedrich. (1900) “Kants Verhältnis zur Metaphysik,” Kant-Studien 4: 413 47. Pereboom, Derk. (1988) “Kant on Intentionality,” Synthese, 77: 321–52. Pistorius, Hermann Andreas. (1784) “[Review of] Kant, I.: Prolegomena zu einer jeden künftigen Metaphysik, die als Wissenschaft wird auftreten können, Riga: Hartknoch, 1783,” Allgemeine deutsche Bibliothek, 59 (1): 322–57.



377

Pistorius, Hermann Andreas. (1786) “[Review of] Erläuterungen über des Herrn Professor Kant Critik der reinen Vernunft von Joh. Schultze, Königl. Preußischem Hofprediger. Königsberg 1784,” Allgemeine Deutsche Bibliothek 66 (1): 92–123. Pistorius, Hermann Andreas. (1788) “[Review of] Prüfung der Mendelssohnschen Morgenstunden, oder aller spekulativen Beweise für das Daseyn Gottes in Vorlesungen von Ludwig Heinrich Jakob, Doctor der Philosophie in Halle. Nebst einer Abhandlung vom Herrn Professor Kant. Leipzig 1786,” Allgemeine Deutsche Bibliothek 82 (2): 427–70. Platner, Ernst. (1772) Anthropologie für Ärzte und Weltweise, Leipzig: Pogge, Thomas. (1991) “Erscheinungen und Dinge an sich,” Zeitschrift für philosophische Forschung 45 (4): 489–510. Posy, Carl. (1992) Kant’s Philosophy of Mathematics: Modern Essays, Dordrecht, Boston: Kluwer Academic Publishers. Prauss, Gerold. (1971) Erscheinung bei Kant: Ein Problem der “Kritik der reinen Vernunft,” Berlin: Walter de Gruyter. Prauss, Gerold. (1974) Kant und das Problem der Dinge an sich, Bonn: Bouvier. Prichard, Harold Arthur. (1909) Kant’s Theory of Knowledge, Oxford: Clarendon Press. Reichenbach, Hans. (*1920), Relativitätstheorie und Erkenntnis Apriori, in Die philosophische Bedeutung der Relativitätstheorie, vol. 3 of Gesammelte Werke in 9 Bänden, edited by A. Kamlah and M. Reichenbach, Braunschweig: Vieweg, 1979. Reicke, Rudolf. (1889) Lose Blätter aus Kants Nachlaß, first volume, Königsberg: Fred Beyer’s Buchhandlung, cited as ‘Reicke’, by volume and page number. Reinhold, Karl Leonhard. (1786–1787) “Briefe über die Kantische Philosophie,” Der Teutsche Merkur, later published in book form, volume 1 in 1790, Leipzig: Georg Joachim Göschen. Reinhold, Karl Leonhart. (1789) Versuch einer neuen Theorie des menschlichen Vorstellungsvermögens, Prag and Jena: C. Widtmann und J. M. Mauke. Reinhold, Karl Leonhard. (1790) Beyträge zur Berichtigung bisheriger Missverständnisse der Philosophen. Erster Band, das Fundament der Elementarphilosophie betreffend, Jena: Johann Michael Mauke. Reinhold, Karl Leonhard. (1791) Über das Fundament des Philosophischen Wissens, nebst einigen Erläuterungen über die Theorie des Vorstellungsvermögens, Jena: Johann Michael Mauke. Reinhold, Karl Leonhard. (1798), “[Combined review of] Fichte, J. G.: Über den Begriff der Wissenschaftslehre oder der sogenannten Philosophie, als Einladungsschrift zu seinen Vorlesungen über diese Wissenschaft; Fichte, J. G.: Grundlage der gesammten Wissenschaftslehre, als Handschrift für seine Zuhörer; Fichte, J. G.: Grundriss des Eigenthümlichen der Wissenschaftslehre in Rücksicht auf das theoretische Vermögen, als Handschrift für seine Zuhörer; Philosophisches Journal einer Gesellschaft teutscher Gelehrten, herausgegeben by J. G. Fichte und F. I. Niethammer,” Allgemeine Literatur-Zeitung, Numero 5, 6, 7, 8, 33–9, 41–7, 49–56, 57–63, 65–9. Reinhold, Karl Leonhard. (1799) Über die Paradoxien der neuesten Philosophie, Hamburg: Friedrich Perthes. Rescher, Nicholas. (1981) “On the Status of ‘Things in Themselves’ in Kant,” Synthese 47: 289–99. Riehl, Alois. (1876) Der philosophische Kritizismus und seine Bedeutung für die positive Wissenschaft, 2 volumes, Leipzig: Engelmann. Riehl, Alois. (1908) Der philosophische Kritizismus: Geschichte und System. Erster Band: Geschichte des philosophischen Kritizismus, second revised edition, Leipzig: Engelmann. Robinson, Hoke. (1994) “Two Perspectives on Kant’s Appearances and Things in Themselves,” Journal of the History of Philosophy 32 (3): 411–41. Robinson, Hoke. (1996) “Kantian Appearances and Intentional Objects,” Kant-Studien 87: 448–54. Roche, Andrew. (2011) “Allais on Transcendental Idealism,” Kantian Review 16 (3): 351–74. Roche, Andrew. (2013) “Transcendental Idealism: A Proposal,” Journal of the History of Philosophy 51 (3): 589–615. Rogerson, Kenneth. (1999) “Kant’s World(s) of Appearances and Things in Themselves,” Southwest Philosophy Review: The Journal of the Southwestern Philosophical Society 15 (2): 1–24. Rosefeldt, Tobias. (2007) “Dinge an sich und sekundäre Qualitäten,” in Jürgen Solzenberg (ed.), Kant in der Gegenwart, Berlin: Walter de Gruyter, 167–209.

378



Schaper, Eva. (1966) “The Kantian Thing-in-itself as a Philosophical Fiction,” Philosophical Quarterly 16: 233–43. Schelling, Friedrich Wilhelm Joseph von. (1836/37) Zur Geschichte der neueren Philosophie, Münchner Vorlesungen, in Sämtliche Werke, 10: 1 200. Schelling, Friedrich Wilhelm Joseph von. (1842/43) Philosophie der Offenbarung, Part 1, in Sämtliche Werke, 13: 1 530. Schelling, Friedrich Wilhelm Joseph von. (1856 1861) Sämtliche Werke, edited by Karl F. August Schelling, Erste Abteilung, volumes 1 10, Zweite Abteilung, volumes 11 14, Suttgart/Augsburg: Cotta; cited as ‘SW’, by volume and page number. Schmid, Carl Christian Erhard. (1788, second edition) Wörterbuch zum leichteren Gebrauch der kantischen Schriften nebst einer Abhandlung, Jena: Cröker. Schmid, Raymund. (1928) “Kant—Vaihinger—Adickes,” Annalen der Philosophie und philosophischen Kritik 7: 1–16. Schopenhauer, Arthur. (1847, second edition) Ueber die vierfache Wurzel des Satzes vom zureichenden Grunde, in Sämtliche Werke, volume 3, 101 268. Schopenhauer, Arthur. (1851) Parerga und Paralipomena. Kleine philosophische Schriften, in Sämtliche Werke, volumes 4 and 5. Schopenhauer, Arthur. (1859, third edition) Die Welt als Wille und Vorstellung, in Sämtliche Werke, volumes 1 and 2. Schopenhauer, Arthur. (1911–1941) Arthur Schopenhauers Sämtliche Werke, edited by Paul Deussen, München: Piper Verlag, cited as ‘Werke’, by volume and page number. Schulting, Dennis. (2011) “Kant’s Idealism: The Current Debate,” in Dennis Schulting and Jacco Verburgt (eds.), Kant’s Idealism. New Interpretations of a Controversial Doctrine, Dordrecht, Heidelberg, London, New York: Springer, 1–25. Schultz, Johann. (1785) “Erläuterungen über des Herrn Professor Kant Critik der reinen Vernunft,” Allgemeine Literaturzeitung, Band 3, Numero 162, 164, 178, 179, and 179 Beylage; reprinted in book form in 1791. Schultz, Johann. (1789) Prüfung der Kantischen Critik der reinen Vernunft, Teil I, Königsberg: Hartung. Schultz, Johann. (1790) “[Review of] Philosophisches Magazin, Bd. 2, St. 1–4. Hrsg. von J. A. Eberhard. Halle: Gebauer 1789–90,” Allgemeine Literatur-Zeitung, Band 3, Numero 281–285, 785–813. Schultz, Johann. (1792) Prüfung der Kantischen Critik der reinen Vernunft, Teil II, Königsberg: Nicolovius. Schulze, Gottlob Ernst. (1792) Aenesidemus oder Über die Fundamente der von dem Herrn Prof. Reinhold in Jena gelieferten Elementar-Philosophie; nebst einer Vertheidigung des Skeptizismus gegen die Anmaßungen der Vernunftkritik, Helmstedt. Schulze, Gottlob Ernst. (1801) Kritik der theoretischen Philosophie, two volumes, Hamburg: Karl Ernst Bohn. Schwab, Johann Christoph. (1796) “Ausführliche Erörterung der von der Königl. Akademie der Wissenschaften zu Berlin für das Jahr 1791 vorgelegten Frage: ‘Welches sind die wirklichen Fortschritte, die die Metaphysik seit Leibnizens und Wolffens Zeiten in Deutschland gemacht hat?’,” in Preisschriften über die Frage: Welche Fortschritte hat die Metpahysik seit Leibniztzens und Wolffs Zeiten in Deutschland gemacht? Von Johan Christoph Schwab, Karl Leonhard Reinhold und Johann Heinrich Abicht, edited by Königlich Preußische Akademie der Wissenschaften, Berlin: Friedrich Maurer, 1–170. Sellars, Wilfrid. (1968) Science and Metaphysics: Variations on Kantian Themes, New York: Humanities Press. Setiya, Kieran. (2004) “Transcendental Idealism in the ‘Aesthetic’,” Philosophy and Phenomenological Research, 68 (1): 63–88. Shabel, Lisa. (2004) “Kant’s ‘Argument from Geometry’,” Journal of the History of Philosophy 42 (2): 195–215. Smit, Houston. (2000) “Kant on Marks and Intuition,” Philosophical Review 109 (2): 235–66. Stang, Nicholas. (2014) “The Non-Identity of Appearances and Things in Themselves,” Noûs, 48 (1): 106–36.



379

Stang, Nicholas. (2016) Kant’s Modal Metaphysics, Oxford, New York: Oxford University Press. Strawson, Peter. (1966) The Bounds of Sense, London: Methuen & Con. Stuhlmann-Laeisz, Rainer. (1976) Kant’s Logik: eine Interpretation auf der Grundlage von Vorlesungen, veröffentlichten Werken und Nachlaß, Berlin, New York: De Gruyter. Trendelenburg, Friedrich Adolf. (1862, second, revised, and enlarged edition) Logische Untersuchungen, volume 1, Leipzig: S. Hirzel. Trendelenburg, Friedrich Adolf. (1867) “Über eine Lücke in Kants Beweis von der ausschließenden Subjectivität des Raumes und der Zeit: Ein kritisches und antikritisches Blatt,” in A. Trendelenburg, Historisch Beiträge zur Philosophie, vol. 3, no. 7, Berlin: G. Bethge, 215–76. Trendelenburg, Friedrich Adolf. (1869) Fischer und sein Kant—eine Entgegnung, Leipzig: S. Hirzel. Ueberweg, Friedrich. (1872, *1863) Grundriß der Geschichte der Philosophie. Dritter Band: Die Neuzeit, Berlin: E.S. Mittler & Sohn. Vaihinger, Hans. (1881, 1892) Commentar zu Kants Kritik der reinen Vernunft, 2 Bände, Stuttgart: W. Spemann; Stuttgart, Berlin, Leipzig: Union Deutsche Verlagsgesellschaft. Vaihinger, Hans. (1884) “Zu Kants Widerlegung des Idealismus,” in Strassburger Abhandlungen zur Philosophie. Edouard Zeller zu seinem siebenzigsten Geburtstage, Freiburg im Breisgau, Tübingen: J.C.B. Mohr. Vaihinger, Hans. (1918) Die Philosophie des Als Ob. System der theoretischen, praktischen und religiösen Fiktionen der Menschheit auf Grund eines idealistischen Positivismus. Mit einem Anhang über Kant und Nietzsche, dritte Auflage, Leipzig: F. Meiner; Nachdruck, 2007, VDM Verlag. Vaihinger, Hans. (1921) “Kants antithetische Geistesart, erläutert an seiner Als-ob-Lehre,” in Den Manen Fr. Nietzsches. Weimarer Weihnachtsgeschenke zum 75. Geburtstag von Frau Dr. E. FörsterNietzsche, 151–82. Van Cleve, James. (1988) “Inner States and Outer Relations: Kant and the Case for Monadism,” in Peter H. Hare (ed.), Doing Philosophy Historically, Buffalo: Prometheus Books, 231–47. Van Cleve, James. (1999) Problems from Kant, Oxford, New York: Oxford University Press. Van Cleve, James and Frederick, Robert E. (editors). (1991) The Philosophy of Right and Left. Incongruent Counterparts and the Nature of Space, The University of Western Ontario Series in Philosophy of Science, Dordrecht: Springer. Walker, Ralph C. S. (2010) “Kant on the Number of Worlds,” British Journal for the History of Philosophy 18 (5): 821–43. Walker, Ralph C. S. (2016) “Comments Lucy Allais, Manifest Reality: Kant’s Idealism and His Realism,” European Journal of Philosophy 24 (1): 267–74. Warda, Arthur. (1922) Immanuel Kants Bücher, Berlin: Verlag Martin Breslauer. Warren, Daniel. (1998) “Kant and the Apriority of Space,” Philosophical Review 107: 179–224. Watkins, Eric. (2005) Kant and the Metaphysics of Causality, New York: Cambridge University Press. Watkins, Eric. (2006) “On the Necessity and Nature of Simples: Leibniz, Wolff, Baumgarten, and the Pre-Critical Kant,” Oxford Studies in Early Modern Philosophy, edited by Daniel Garber and Steven Nadler, volume III, Oxford: Clarendon Press, 261–314. Westphal, Kenneth. (1997) “Noumenal Causality Reconsidered: Affection, Agency, and Meaning in Kant,” Canadian Journal of Philosophy 27 (2): 209–45. Westphal, Kenneth. (2004) Kant’s Transcendental Proof of Realism, Cambridge, New York: Cambridge University Press. Willaschek, Markus. (1997) “Der transzendentale Idealismus und die Idealität von Raum und Zeit, eine ‘lückenlose’ Interpretation von Kants Beweis in der ‘Transzendentalen Ästhetik’,” Zeitschrift für philosophische Forschung 52: 537–64. Willaschek, Marcus. (2001) “Die Mehrdeutigkeit der kantischen Unterscheidung zwischen Dingen an sich und Erscheinungen. Zur Debatte um Zwei-Aspekte- und Zwei-Welten-Interpretationen des transzendentalen Idealismus,” in V. Gerhard, R. P. Horstmann, R. Schumacher (eds.), Akten des IX. Internationalen Kant-Kongresses, Berlin, New York, Zweiter Band, 679–90. Willaschek, Marcus. (2018) Kant on the Sources of Metaphysics. The Dialectic of Pure Reason, Cambridge: Cambridge University Press.

380



Wilson, Catherine. (2005) “Confused Perceptions, Darkened Concepts: Some Features of Kant’s Leibniz-Critique,” in George MacDonald Ross (ed.), Kant and His Influence, New York: Continuum, 73–103. Windelband, Wilhelm. (1877) “Über die verschiedenen Phasen der kantischen Lehre vom Ding-ansich,” Vierteljahrschrift für wissenschaftliche Philosophie I (2): 224–66. Wolff, Christian. (1729) Vernünfftige Gedancken von Gott, der Welt und der Seele des Menschen, auch allen Dingen überhaupt, Halle, in Gesammelte Werke, Teil I: Deutsche Schriften, Bände 2.1 and 2.2. Wolff, Christian. (1734) Psychologia Rationalis Methodo Scientifica Pertractata Qua Ea, Que De Anima Humana, Frankfurt/Leipzig, in Gesammelte Werke, Teil II: Lateinische Schriften, Band 6. Wolff, Christian. (1962 2020) Gesammelte Werke, edited by J. École, H. W. Arndt, Ch. A. Corr, J. E. Hoffmann, and M. Thomann, Hildesheim: Olms, cited as ‘GW’, by volume and page number. Wood, Allen. (2007) “Debating Allison on Transcendental Idealism,” Kantian Review 12 (1): 1–10. Wundt, Max. (1924) Kant als Metaphysiker. Ein Beitrag zur Geschichte der deutschen Philosophie im 18. Jahrhundert, Stuttgart: Ferdinand Enke.

Index For the benefit of digital users, indexed terms that span two pages (e.g., 52–53) may, on occasion, appear on only one of those pages. actuality 19–20, 22n.69, 28, 39–40, 72–4, 100; see also categories; experience, actual Adams, Robert 15n.47, 18n.54, 286n.88, 307n.129 Adickes, Erich 22n.67, 22n.69, 23n.70, 78n.145, 169n.128, 305–6, 307n.129, 330n.35 affection 58–9 double 76–8 empirical 76–7 formal 214–15, 223, 237–40 material 214–15, 217–18, 237–40 problem of 19–20, 302–5 self- 276–7; see also formal affection transcendental 76–7 Allais, Lucy 8n.23, 24n.76, 30n.12, 43n.43, 46n.49, 212n.66, 218n.75, 252n.14, 253n.16, 254–5, 257–67, 306n.128, 329n.33, 339n.62 Allison, Henry 8n.22, 11–14, 13n.40, 67n.122, 119n.19, 195n.27, 241n.102 Ameriks, Karl 2n.5, 4n.8, 247n.6 amphiboly 171–2, 254–5, 325–9, 342–4 analogies of experience 61–3, 65–7 anticipations of perception 61–3, 65–7 antinomies 332–3, 348–53; see also ideas; idealism, transcendental, indirect argument for dynamical 90n.171, 111n.2 mathematical 84–94, 100 appearances 27–109; see also distinction between things in themselves and appearances versus illusion 250–7, 266–7 existence of 72–4 ‘form’ of 58–9, 229–30, 232–3, 238, 240; see also forms of sensibility; affection, formal; ‘matter’ empirical 70n.129, 116–17 inner 33, 103–9; see also self, empirical Kantian 27–8 ‘matter’ of 58–9, 70–2, 232–3, 238, 240, 245–7, 271–3 outer 33; see also object, empirical unconscious 79–80 Aquinas 31n.14, 315–16 Aristotle 31n.14, 315–16 axioms of intuition 61–3, 65–7 Baumgarten, Alexander Gottlieb 126n.36, 177n.151 bare particular 36, 45–8, 61–3, 65–7, 168–70; see also object, transcendental; subject, transcendental

Beck, Jacob Sigismund 20–1, 303n.117 Berkeley, George 34n.22, 51–2, 115–16, 124n.32, 125–9, 154, 174, 192, 217, 258–61, 269–76 categories 61–3, 65–7, 199 as applied to things in themselves 19–20, 302, 304–5 transcendental deduction of the 108n.211, 109n.214, 230–1, 234n.93 causality; see also categories empirical 76–7, 175–8 transcendental 76–7 cognition 281–2, 303–4, 339–40 color 251n.13; see secondary qualities concepts; see also representations generality of 207–9 mediacy of 79n.148, 81n.155 pure 61n.97 conditioning principle 349 Copernican Turn/Revolution 67n.122, 111–13, 228–9 Descartes, René 5–7, 33–4, 316–17 determinables 83n.158, 208–9 determinations 45–8 intuitive 79, 261 spatial/temporal 50n.58, 59n.87, 129–34 space-infected/time-infected 150–2, 166–8, 183 distinction between things in themselves and appearances empirical 70n.129, 116–17, 248, 250–7 transcendental 4, 27–50, 248 dualism versus monism at the empirical level of reality 117n.14, 121–2, 174–5 at the transcendental level of reality 275–8, 284 empiricism 313–14 entity 5–7, 28 epiphenomenalism 175–7; see also dualism versus monism Erdmann, Benno 20n.61, 22n.67 ether 147 ew-account 94–5 existence; see also actuality E 295–9, 302–18 genuine 37–9 is not a real property 158, 303–4

382



existence; (cont.) problem of 19–20, 22n.69, 302–5 pseudo- 37–9 experience 59–60, 64–81 actual 94–103 formal conditions (for the possibility) of 64–9 inner 107–9, 172–4 judgments of 67–9 material conditions (for the possibility) of 69–78, 96–7 outer 64–78, 81–3 possible 94–103 pseudo- 186–7 Falkenstein, Lorne 204n.44, 241n.102 finitude 58–9, 272–3, 313–16 Fichte, Johann Gottlieb 20–1 fictionalism; see things in themselves, fictionalism about Fischer, Kuno 22–3, 240–1 ‘form’; see appearances, ‘form’ of forms of intuition; see sensibility, forms of freedom and determinism 176–7, 286, 291–3 Friedman, Michael 89n.169, 144n.75 Garve-Feder review 128–9, 300 geometry 218–24, 227–37 argument from 194–229, 221n.80; see also transcendental exposition postulates of Euclidean 234n.93 German Idealism 315 God; see also intuitive intellect and ordinary idealism 272–3 idea of 331–2 Guyer, Paul 24n.74, 195n.26 Hartmann, Eduard von 307n.129, 314n.136 Heidegger, Martin 1–2, 235n.94, 314n.136 Heimsoeth, Heinz 1–2, 314n.136 human being 285–95 Hume, David 158n.112, 174n.143 ideal 37; see also real transcendentally 37–9, 115–16 empirically 37–9, 115–16, 125 idealism 13–14 critical 1, 3–4, 295–302 empirical or dogmatic 125–9 ordinary 174, 271–2; see also Berkeley problematic 186–7 refutation of 190–3; see also empirical realism, arguments for transcendental idealism 3–4, 38n.31, 110–59, 245 arguments for 86–7, 111–13, 153–4, 179–86, 192–3 indirect argument for 348–53 master argument for 194–244 ideality; see ideal ideas of reason 331–2, 335 material 177n.151 regulative use of 336–8

illusion 336–7; see also appearances, versus illusion imagination 61–3 unoriginality of 312–13 incongruent counterparts 153–4, 323–5 infinite whole 93–4 intelligibilia; see noumena intuition; see also perception; representations; sensibility; sense empirical; see perception formal 139n.67 form of; see form of sensibility form of in the broad sense 139–40 immediacy of 80–1, 212 intellectual 339–40; see also intuitive intellect singularity of 207–10 sensible nature of 212–14 intuitive counterpart 79–80 intuitive intellect 93–4, 314–16, 339–40; see also God ‘in us’ 116–17; see also ‘outside us’ inner versus outer 327–9 Jacobi, Friedrich Heinrich 19–20 Kemp Smith, Norman 22n.67, 195n.26, 196n.31 Kitcher, Patricia 280n.72 Langton, Rae 2n.5, 5–7, 12n.36, 13–14, 180n.3, 329n.33 Leibniz, Gottfried Wilhelm 30n.12, 31n.14, 34n.22, 37–9, 52–7, 131n.48, 174, 208–9, 272–3, 315–16 Leibniz’s law 164n.118, 181, 322–3, 325, 328–9 Leibniz-Wolffian philosophy 2–3, 171–2, 254–6, 321–30, 338–48 levels of reality 37–9, 114–19 Locke, John 171n.136, 258–61 Longuenesse, Béatrice 61n.98, 280n.72 Maimon, Salomon 20n.61 Marshall, Collin 8n.23, 10n.26 matter 144–5, 167n.122, 323; see also ether ‘matter’; see also appearances, ‘matter’ of as preceding ‘form’ 321–2, 328–9 Meinong, Alexius 34n.25 metaphysics 1–3 traditional 171–2, 338–9, 344–6; see Leibniz-Wolffian philosophy metaphysical exposition 203–11 mind; see also self human transcendental 284; see also self, transcendental; subject, transcendental empirical 174–5; see also self, empirical infinite; see intuitive intellect mind-(in)dependence 27–33, 42–50, 106–7 empirical 37–9, 123–4, 133 monism; see dualism motion 83–4, 89n.169, 131n.47, 144–5 general theory of 194–9 multiple-object view 4–7

 Natorp, Paul 22n.67 neglected third possibility 140, 240–4 neo-Kantianism 1–2, 22–4, 303n.117 Newton, Isaac 89n.169, 131n.47 noumena 325, 335–44 in the negative sense 340–1 in the positive sense 340–1 object intentional 33–42, 155–9 empirical 110–14, 119–29, 159–72; see also Kantian appearance fictional (Gedankending) 331 formal conditions for being a proper 65–7 of the pure understanding; see noumena proper 64–9 sensible 141; see also empirical object transcendental 36, 47n.51, 61–3, 65–7, 168–70 objectively real 320 objectivity; see formal conditions of experience one-object view; 4–7, 162–8, 268–9, 285–95, 300n.111; see also two-aspect view ontological ingredient 9–10, 45–8 ontologically basic part 287–9 ontological overlap 10–11 ontological priority 150–2, 183 ‘outside us’ 133 in the empirical sense 116–17, 122–3 in the transcendental sense 116–17, 121–2, 297–8 paralogisms 281–2, 332–3 partial identity 290 Paulsen, Friedrich 1–2 perception 50–64, 67–81 confused 52–3, 79–80 direct/indirect 78 judgements of 67–9 pseudo- 186–7 reference of 79–80 referent of 254–7 Pistorius, Hermann Andreas 19–20, 22n.66, 106n.204, 240–1 Pogge, Thomas 167n.123 point of view of fundamental ontology 37 possibility empirical 96–7; see also experience, possible; postulates of empirical thinking real 96–7, 331n.36 postulates of empirical thinking 61–3, 65–7, 72–4, 100, 102n.193, 103n.195 Prauss, Gerold 13–14, 118n.15, 169n.128 properties; see also determinations appearance 159–60 in-itself 159–60 logical 45–8 ontological 116–17, 159 relational or extrinsic 11–12, 321–2, 328–9; see also inner versus outer

383

rationalism; see metaphysics, traditional; Leibniz-Wolffian philosophy real 37, 40n.37; see also reality empirically 37–9, 115–16, 123 transcendentally 37–9, 115–16, 121–2 realism; see also things in themselves, realism about empirical realism 28, 119–25, 129–72 arguments for 186–94 transcendental realism 119–25, 129–59 reality; see also real as a category 39–40 Reinhold, Karl Leonard 21n.65 refutation of idealism; see idealism, refutation of regulative use of the ideas of reason; see ideas, regulative use relationalism; see space and time representations 28–9, 33–6 completely determinate 207–9 direct/indirect 35–6 immediacy of 80–1 objective 64–9 ontological specification of 33–4 presentational content of 34–6; see also intentional object reference of 34–5, 79–80 singularity of 207–9 Riehl, Alois 271n.50, 306n.128 Roche, Andrew 161n.114, 165n.119, 252n.14, 253n.16 Rosefeldt, Tobias 43n.43, 253n.16, 257–67 sameness 4–5, 9–10 same-things view 10–11 Schelling, Friedrich Wilhelm Joseph von 20n.58 schemata 61–3, 234n.93; see also imagination Schopenhauer, Arthur 72n.131, 306n.128 Schultz, Johann 22n.66 Schulze, Gottlob Ernst ‘Aenesidemus’ 19–20, 303n.117 science 82–3, 94–5, 143–5, 175–7, 333–5 secondary qualities 157–8, 251n.12, 257–67 self empirical 103–5, 172–8 transcendental 103–5, 276–85; see also transcendental subject, transcendental mind self-consciousness 280–4 sensation 58–9, 136–7 pseudo- 186–7, 313 sense 52 inner 33, 103–4, 106, 136–7, 238–9, 295–8 outer 33, 267–8, 295–8 sense organ 76–7 sensibility 58–9, 232–3, 264–5, 295–8 forms of 65–7, 136–41 passivity of 212–15, 237–40, 310–17 sensible 58–9, 141 versus supersensible 141 Setiya, Kieran 214n.69 Shabel, Lisa 196n.31

384



soul 331–2; see also mind; paralogisms; self as object of inner sense 103–4; see also appearance, inner; self, empirical space and time 129–59 empirical 132–3 original representation of; see space and time, pure intuition of pure intuition of 132, 203–12, 237–40 pure 132–45 relationalism about 131–2, 145–55, 186 substantivalism about 131–2, 145–55, 184, 186, 200n.36, 202 S- 182–3, 225–6 S-, 183–5 S-, 183 Spinoza, Baruch 131–2, 154n.102, 298n.107 Stang, Nicholas 97n.185, 162n.116 Strawson, Peter 23–4, 64–5, 195n.26 subject; see also self transcendental subject 104n.198, 282–4 substantivalism; see space and time, substantivalism about things in themselves 27–33, 42–50; see also distinction between things in themselves and appearances argument for the existence of 191, 193, 305–18 grounding of appearances by 74–6, 245–50, 267–9, 276–85 fictionalism about 14–15, 19–23, 330–44 Kantian 27–8, 319–20 Leibniz-Wolffian 321–30 possibly non-cooperating 83–4 realism about 14–15, 319–21; see also things in themselves, argument for the existence of transcendental exposition 211–12 transcendental principles of the understanding 61–3, 65–7 transcendental psychology 279–80 Trendelenburg, Friedrich Adolf 22n.69, 240–1

two-aspect view 4, 7–15, 24, 32–3, 46n.49, 87n.167, 117–18, 118n.15, 158n.112, 159–72, 170n.132, 257–67, 303, 305n.126; see also one-object view; same-things view methodological two-aspect view 11–12, 118–19, 163–4, 305n.126 two-world view 4, 10–11, 32–3; see also multiple-object view classic 12–15, 17, 19–25, 302 unconditioned, the 331–2, 344–6, 349–50 U 313–16 underdetermination 101–2 understanding; see also intuitive intellect discursive nature of 93–4, 208–9, 313–14 pure 336, 338–41 U--- 276–7, 310 unknowability doctrine 301–2 unknowability problem 303–4 U 312–13 Vaihinger, Hans 14n.44, 22n.69, 78n.145, 330n.35, 335 Van Cleve, James 5–7, 11n.31, 15n.48, 18–19, 162n.116, 163n.117, 195n.26, 325n.21 Walker, Ralph 167n.123, 269n.43, 289n.96 Warren, Daniel 204n.44 Watkins, Eric 2n.5, 323n.17 Willaschek, Markus 212n.66, 216n.72, 218n.75, 244n.109, 345n.79 Windelband, Wilhelm 22n.67, 22n.68 Wolff, Christian 30n.12, 56–7, 177n.151 world 16–17, 25–6, 355–6 empirical 36–7 intelligible 16–17 sensible 16–17 possible 96–7 Wundt, Max 1–2