Kant's Critique of Pure Reason and the Method of Metaphysics 1009172107, 9781009172103

Citation preview

KANT’S CRITIQUE OF PURE REASON AND THE METHOD OF METAPHYSICS

In two often neglected passages of the Critique of Pure Reason, Kant submits that the Critique is a ‘treatise’ or a ‘doctrine of method’. These passages are puzzling because the Critique is only cursorily concerned with identifying adequate procedures of argument for philosophy. In this book, Gabriele Gava argues that these passages point out that the Critique is the doctrine of method of metaphysics. Doctrines of method have the task of showing that a given science is indeed a science because it possesses ‘architectonic unity’ – which happens when it realizes the ‘idea’ of a science. According to Gava’s novel approach, the Critique establishes that metaphysics is capable of this unity, and his reading of the Critique from this perspective not only illuminates the central role of the Transcendental Doctrine of Method within it, but also clarifies the relationship between the different parts of the work. gabriele gava is Associate Professor of Theoretical Philosophy at the University of Turin. He is the author of Peirce’s Account of Purposefulness: A Kantian Perspective (2014).

bli h d

li

b C

b id

i

i

bli h d

li

b C

b id

i

i

K A N T ’ S C R I T IQU E OF PU R E R E A S O N A N D T H E M E T HOD OF M E TA PH Y S IC S G A BR I E L E G AVA University of Turin

bli h d

li

b C

b id

i

i

Shaftesbury Road, Cambridge cb2 8ea, United Kingdom One Liberty Plaza, 20th Floor, New York, ny 10006, usa 477 Williamstown Road, Port Melbourne, vic 3207, Australia 314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India 103 Penang Road, #05–06/07, Visioncrest Commercial, Singapore 238467 Cambridge University Press is part of Cambridge University Press & Assessment, a department of the University of Cambridge. We share the University’s mission to contribute to society through the pursuit of education, learning and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781009172103 doi: 10.1017/9781009172127 © Gabriele Gava 2023 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press & Assessment. First published 2023 A catalogue record for this publication is available from the British Library. Library of Congress Cataloging-in-Publication Data Names: Gava, Gabriele, 1981– author. Title: Kant’s critique of pure reason and the method of metaphysics / Gabriele Gava, University of Turin. Other titles: Critique of pure reason and the method of metaphysics Description: Cambridge, United Kingdom ; New York, NY, USA : Cambridge University Press, 2023. | Includes bibliographical references and index. Identifiers: LCCN 2022055873 (print) | LCCN 2022055874 (ebook) | ISBN 9781009172103 (hardback) | ISBN 9781009172127 (ebook) Subjects: LCSH: Kant, Immanuel, 1724–1804. Kritik der reinen Vernunft. | Metaphysics. | Knowledge, Theory of. Classification: LCC B2779 .G36 2023 (print) | LCC B2779 (ebook) | DDC 110–dc23/eng/20230324 LC record available at https://lccn.loc.gov/2022055873 LC ebook record available at https://lccn.loc.gov/2022055874 isbn 978-1-009-17210-3 Hardback Cambridge University Press & Assessment has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

bli h d

li

b C

b id

i

i

To Laura, Maddalena, and Greta

bli h d

li

b C

b id

i

i

bli h d

li

b C

b id

i

i

Contents

Acknowledgements page ix Citations of Kant’s Works xi Introduction

1

P a r t I  M e t a p h y s i c s a s a S c i e n c e a n d T h e R o l e o f t h e c r i t i q u e o f p u r e r e a s o n 15 Introduction to Part I

15

1 The Worldly Concept of Philosophy and the Possibility of Metaphysics as a Science

17

2

The Critique of Pure Reason as the Doctrine of Method of Metaphysics41

P a r t I I  T h e M e t h o d o f T r a n s c e n d e n t a l P h i l o s o p h y 65 Introduction to Part II

65

3

75

Metaphysical Deductions

4 Transcendental Deductions

122

P a r t I I I  T h e M e t h o d o f t h e C r i t i q u e of Pur e R e ason

169

Introduction to Part III

169

5 The Negative Side of the Critique of Pure Reason

174

vii

bli h d

li

b C

b id

i

i

Contents 

viii

6 Transcendental Philosophy and the Critique of Pure Reason in the B-Deduction

194

7 The Positive Side of the Critique of Pure Reason

207

Pa r t I V   K a n t on D o g m at i s m a n d S c e p t ic i s m

229

Introduction to Part IV

229

8 Kant on Wolff and Dogmatism

231

9 Kant on Hume and Scepticism

251

Conclusion

267

Bibliography 270 Index 282

bli h d

li

b C

b id

i

i

Acknowledgements

Working on this book has been a long journey. The idea of writing a book dedicated to Kant’s ‘method’ had been in my mind at least since I moved to Germany in 2012 to pursue a post-doc in Frankfurt. The present work and the way in which it addresses the question of ‘method’ is quite different from that initial idea. However, using an image dear to Kant, that idea can be considered its ‘seed’. My time in Frankfurt, which from the initially planned two years became an eight-year period, was essential to making that seed grow. In Frankfurt, I met many interesting philosophers, some of whom are now my friends. I am especially grateful to Marcus Willaschek, who has been and continues to be not only a constant source of insight and useful feedback but also a wonderful mentor and friend. While my time in Frankfurt has had an enormous effect on the theses defended in this book and how I argue for them, there is also a sense in which the view I develop here finds its roots in the Italian tradition of Kant scholarship in which I made my first steps as a scholar. What is distinctive in this tradition is that it emphasizes the importance of the Transcendental Doctrine of Method of the first Critique for understanding Kant’s project. In this respect, I am indebted to Alfredo Ferrarin, my PhD advisor, and Claudio La Rocca, both of whom greatly influenced my view.1 When I think of my time in Frankfurt and this Italian tradition, I see this book as a way of bringing together these two sources of influence. I would like to express my gratitude to numerous people who have read drafts of parts of this book and have provided useful feedback, as well as those who posed challenging questions during conferences and talks. Let me mention the following people in particular: Stefano Bacin, Stefano Bertea, Henny Blomme, Claudia Blöser, Karin de Boer, Luigi Filieri, Cord Friebe, Luis Garcia, Stephen Howard, Thomas Höwing, Colin 1

Other Italian scholars have dedicated important studies to the Transcendental Doctrine of Method, for example Massimo Barale and Giorgio Tonelli. Even though I have not collaborated directly with them, the approach I develop in this book inherits some of their insights.

ix

bli h d

li

b C

b id

i

i

Acknowledgements 

x

McQuillan, Sofie Møller, Luciano Perulli, Pavel Reichl, Elke Schmidt, Matthé Scholten, Dieter Schönecker, Thomas Sturm, Achim Vesper, Owen Ware and Lea Ypi. I also wish to thank the two anonymous reviewers for Cambridge University Press, who pushed me to make clearer how the different parts of the book hang together. I am extremely thankful to Eric Watkins, who organized a manuscript workshop on a previous draft of this book in the summer of 2021. The feedback I received from him and from the other participants, Lucy Allais, Andrew Chignell, Huaping Lu-Adler, Clinton Tolley and Marcus Willaschek, has been immensely helpful in making my view clearer and, hopefully, stronger. Special thanks also go to Marcello Garibbo, Mario Pati and Lorenzo Sala, who read the penultimate draft of the book and proposed various improvements, to Carolyn Benson, who corrected my English, and to Hilary Gaskin, who provided many valuable suggestions along the way. Needless to say, the remaining mistakes and inaccuracies are my responsibility. During the time I was working on this book, many things happened in my life. I married my wife, Laura, with whom I have been together since I was an undergraduate in Venice. She moved in Frankfurt with me, where our two daughters, Maddalena and Greta, were born. They have certainly made the process of writing this book less stressful than it would otherwise have been. This is also why I will always connect that process and Frankfurt to happy memories. It is to them that this book is dedicated. The research at the basis of this book has been partially supported by the Deutsche Forschungsgemeinschaft, to which I am grateful. A paragraph of Chapter 1 re-elaborates a passage from the chapter ‘The Doctrine of Method’, in The Kantian Mind, ed. by S. Baiasu and M. Timmons, London: Routledge, forthcoming. A small section from Chapter 7 appeared in ‘Kant and Crusius on Belief and Practical Justification’, Kantian Review, 24:53– 75, 2019. A shorter version of Chapter 8 was published as ‘Kant on Wolff and Dogmatism’, in Proceedings of the 13th International Kant Congress ‘The Court of Reason’, edited by C. Serck-Hanssen and B. Himmelmann, Berlin/Boston: De Gruyter, 2021, pp. 299–308. A paragraph of Chapter 8 re-elaborates a passage from ‘Kant, Wolff and the Method of Philosophy’, Oxford Studies in Early Modern Philosophy, 8:271–303, 2018. I thank the editors and publishers of these publications for permission to use these materials.

bli h d

li

b C

b id

i

i

Citations of Kant’s Works

Throughout the book, I have tried to keep abbreviations to a minimum and have adopted a simple system for citations of Kant’s works. Citations from the Critique of Pure Reason refer to its first and second editions by only using A and B, respectively (for example: A834/B862). Citations from the Akademie Ausgabe of Kant’s writings (Kants gesammelte Schriften, edited by the Berlin-Branderburgischen – formerly Preussischen – Akademie der Wissenschaften, Berlin: Reimer, De Gruyter, 1900–) only indicate volume and page number (for example: 9:70). Citations of Kant’s Reflexionen additionally include their number, preceded by the abbreviation Refl. and followed by the volume and page number in the Akademie Ausgabe (for example: Refl. 1656, 16:68). I cite one work, the Heschel Logic, that is not contained in this edition and instead appears in the volume LogikVorlesungen: Unveröffentliche Nachscriften, Vol. 2, edited by T. Pinder, Hamburg: Meiner, 1998. When citing this work, I indicate the page number in Pinder’s edition, followed by the page in the English translation by M. Young (Lectures on Logic, ed. and tr. by M. Young, Cambridge: Cambridge University Press, 1992) (for example: Heschel Logic, 488, Eng. tr. 416). English translations are from The Cambridge Edition of the Works of Immanuel Kant (edited by P. Guyer and A. Wood, Cambridge: Cambridge University Press, 1992–), adapted to British spelling. I specify in brackets or in notes when my translations deviate from this edition.

xi

bli h d

li

b C

b id

i

i

bli h d

li

b C

b id

i

i

Introduction

In the fifth and concluding section of the 1770 Inaugural Dissertation, Kant draws a distinction between mathematics and natural science, on the one hand, and metaphysics, on the other. He writes that while in mathematics and natural science ‘use gives the method ’ (2:410), in ‘pure philosophy’ ‘method precedes all science’ (2:411). In metaphysics, a preliminary clarification of the proper ‘method’ is needed because we risk improperly using principles that belong to ‘sensibility’ to represent objects of the ‘understanding’, which are its subject matter. The Critique of Pure Reason gives up the idea that metaphysics should comprise cognitions of objects of the understanding that are not given through sensibility. However, Kant’s claim that a clarification of the ‘method’ of metaphysics should precede the actual science foreshadows the idea that metaphysics requires a propaedeutic ‘critique’ of reason.1 Moreover, it is not the case that reference to a ‘critique’ completely substitutes talk of an inquiry focused on the ‘method’ of metaphysics. In fact, in two passages of the Critique of Pure Reason, Kant explicitly identifies his ‘critical’ investigation with such inquiry. In the 1787 Preface, he submits that the Critique ‘is a treatise on the method, not a system of the science itself’ (Bxxii). In a passage from the first edition that remains unaltered in the second, he labels the investigation he is pursuing a ‘doctrine of method’ (A82–3/B108–9). But what could it mean to say that the Critique is a doctrine or treatise on method?2 If we take a treatise on method to be an exposition of the procedures of investigation or argument that are appropriate in a particular discipline, it is difficult to understand Kant’s contention. The identification of such procedures is only partially pursued in the Discipline of Pure 1 2

As will become clear, to say that the Critique is ‘propaedeutic’ is not to say that it comes before any part of metaphysics has been established. Scholars who have taken this statement seriously include Barale (1988), Tonelli (1994), La Rocca (2003: Ch. 6), Ferrarin (2015) and McQuillan (2016).

1

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

2

Introduction

Reason, but this is insufficient to claim that the whole Critique has this aim. We find a similar problem if we focus on the second main part of the Critique in particular, the Transcendental Doctrine of Method: while the title suggests that it focuses on methodological issues, only the Discipline of Pure Reason fits this description. By contrast, it is not obvious why, for example, the Canon of Pure Reason or the History of Pure Reason are included in that part. Furthermore, an additional challenge comes from the fact that Kant characterizes both the whole Critique and one part of it as a ‘doctrine of method’, which is confusing. The main aim of this book is to provide a satisfactory answer to these questions. In short, my overarching answer is that the claim that the Critique is a ‘doctrine’ or ‘treatise’ on method signals that it is the ‘doctrine of method’ of metaphysics. According to Kant, ‘doctrines of method’ have specific characteristics. It is by attending to these characteristics that we can solve the difficulties I have just mentioned. First of all, the principal task of a doctrine of method is not to identify procedures of investigation or argument that are appropriate within a particular science. Rather, their principal task is to show that a set of cognitions can be considered a science because it forms a ‘system’ with a certain unity, which Kant calls ‘architectonic’. In this respect, the Critique can be a ‘doctrine of method’ even though it only cursorily singles out procedures of investigation or argument that are adequate for a particular kind of investigation. Additionally, Kant submits that ‘doctrines of method’ usually come ‘at the end’ of a science, ‘because only then am I acquainted with the nature of the science’ (24:795). In my view, this explains why both a part of the Critique is titled ‘Doctrine of Method’ and the entire book can be considered such a doctrine. Insofar as a doctrine of method rests on the previous establishment of at least some parts of the science of which it is a doctrine of method, the Critique of Pure Reason requires that at least some parts of metaphysics be established in its ‘doctrine of elements’. Only in this way will it be able to perform its task as a doctrine of method. Since I claim that the Critique of Pure Reason is the doctrine of method of metaphysics and that, as such, it rests on the establishment of some parts of that science, it is here useful to clarify what I mean by metaphysics and what its ‘parts’ are, according to Kant.

1  Kant on Metaphysics in the Critique of Pure Reason In the Critique of Pure Reason, Kant uses the term ‘metaphysics’ to refer to different things. For example, when he speaks of metaphysics as a ‘natural predisposition’ (B21–2), he has in mind reason’s tendency to ask questions

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Introduction

3

and advance claims about ‘unconditioned’ objects that lie beyond possible experience, such as God and the soul. However, he also speaks of metaphysics as including valid a priori cognitions of the ‘conditioned’ objects of possible experience, the validity of which is established in the Critique (Bxviii–xix). When I claim that the Critique of Pure Reason is the doctrine of method of metaphysics, I use the term ‘metaphysics’ in this second, more inclusive sense. In order to gain an overview of the aims and structure of metaphysics, let us focus on Kant’s sketch of this science in the Architectonic of Pure Reason. There, metaphysics is seen as a part of ‘pure’ philosophy, which he describes as ‘cognition from pure reason’ (A840/B868). Roughly, this means that pure philosophy collects a priori cognitions, which Kant calls ‘philosophical’ in order to distinguish them from ‘mathematical’ a priori cognitions (A841/ B869).3 Pure philosophy is either propaedeutic, which is the critique of pure reason, or metaphysics, which comprises ‘the whole (true as well as apparent) philosophical cognition from pure reason in systematic interconnection’ (A841/B869).4 Metaphysics, as the whole system of philosophical cognitions, is divided into the metaphysics of nature and the metaphysics of morals. The former identifies a priori concepts and principles for the theoretical cognition of objects. The latter singles out concepts and principles that determine ‘action and omission a priori’ (A841/B869). The metaphysics of nature is further divided into transcendental philosophy, which ‘considers only the understanding and reason itself in a system of all concepts and principles that are related to objects in general, without assuming objects that would be given’,5 and the physiology of reason, which considers ‘nature’ as ‘the sum total of given objects’ (A845/B873). Finally, physiology is either ‘immanent’ or ‘transcendent’. ‘The former pertains to nature so far as its cognition can be applied in experience (in concreto), the latter to that connection of the objects of experience which surpasses all experience’ (A845–6/B873–4). Kant includes rational physics and rational psychology within immanent physiology and rational cosmology and rational theology within transcendent physiology (A846–7/B874–5). There is much that remains obscure in this brief presentation of the aims and structure of metaphysics. Since our focus is the relationship 3 4

5

h

d i

Mathematical cognitions differ from philosophical cognitions because they rest on the ‘construction’ of concepts in intuition. In fact, Kant suggests that the term metaphysics can also be used to describe the whole of pure philosophy, including its propaedeutic part. This is clearly not the sense of ‘metaphysics’ I am using when I claim that the critique is the doctrine of method of this discipline. In Part II, I argue that transcendental philosophy also considers representations belonging to sensibility.

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

4

Introduction

between the critique of pure reason and the projected science of metaphysics, we can single out three problems in particular. First, it is not clear how we can distinguish between the critique of pure reason, as the propaedeutic to metaphysics, and transcendental philosophy, as that part of metaphysics that identifies a priori concepts and principles that determine our cognition of objects in general. After all, it seems that in the Critique Kant dedicates much effort to identifying and establishing the validity of concepts and principles of this kind. Second, if the Critique is propaedeutic to the whole of metaphysics, as Kant’s description suggests, in what sense is it also relevant to the metaphysics of morals? For example, it does not provide any justification for a priori principles of morality.6 Finally, how should we take Kant’s inclusion of rational physics, rational psychology, rational cosmology and rational theology within the system? Rational physics probably refers to what will become Kant’s Metaphysical Foundations of Natural Science. In it, Kant identifies principles belonging to the special metaphysics of corporeal nature, which form the a priori part of a science whose objects are given empirically (see 4:469–70). This agrees with Kant’s description of rational physics as included within ‘immanent’ physiology. At the time Kant was writing the first edition of the Critique, he still thought that an empirical psychology was possible and that it would similarly require an a priori part.7 Therefore, Kant’s reference to rational psychology as the second part of immanent physiology can be taken to refer to the ‘metaphysical foundations’ of psychology. But what should we make of Kant’s inclusion of rational cosmology and rational theology in the picture? Are Kant’s arguments in the Antinomy and the Ideal not designed to show that rational cosmology and rational theology cannot offer valid cognitions of objects and so cannot become sciences?

2  The Critique of Pure Reason, Transcendental Philosophy and the System of Metaphysics I submit that the aforementioned problems dissolve when we read the Critique of Pure Reason as the doctrine of method of metaphysics. The 6

7

h

d i

0 0

Of course, Kant did provide this justification in the second Critique. However, the passage in the Architectonic does not explicitly maintain that the first Critique only partially accomplished the propaedeutic to metaphysics. This suggests, first, that at the time he was writing the A-version of the first Critique Kant viewed the first Critique as completing the critical task; second, it suggests that the Critique’s relevance for moral metaphysics cannot lie in providing a ‘foundation’ for it. In a 1785 letter to Christian Gottfried Schütz, Kant still speaks of the ‘metaphysical foundations’ of psychology (10:406). The view that there can be such metaphysical foundations is rejected in the Metaphysical Foundations of Natural Science, however (4:471).

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Introduction

5

first problem I mentioned regarded the relationship between the critique of pure reason and transcendental philosophy. Even though Kant presents them as distinct in the Architectonic, much of what he does in the Critique seems to belong to transcendental philosophy. I have suggested that the Critique, as the doctrine of method of metaphysics, cannot perform its task if no part of metaphysics has already been established. But this gives us a tool for explaining the presence of arguments belonging to transcendental philosophy within the pages of the Critique. Put simply, the Critique contains elements of transcendental philosophy because they are instrumental to its role as the doctrine of method of metaphysics. This means that there are two disciplines that are established within the pages of the Critique: transcendental philosophy, as one part of metaphysics, and the critique of pure reason, as that discipline within the Critique that achieves the latter’s aim as the doctrine of method of metaphysics.8 In this respect, my approach turns the usual view on the relationship between these disciplines on its head. Typically, they are seen as fundamentally overlapping, or, alternatively, the critique of pure reason, as a propaedeutic, is thought to come before the establishment of any part of transcendental philosophy. I characterize transcendental philosophy as that part of the metaphysics of nature that investigates a priori concepts for the cognition of objects that do not contain anything empirical (see A845/B873).9 The part of transcendental philosophy that is established within the Critique takes into consideration not all of these concepts but only those that Kant calls ‘root concepts’ (Stammbegriffe) (see A14/B27–8),10 which are concepts that lie at the basis of synthetic a priori claims. With respect to these concepts, transcendental philosophy has two main tasks. First, it identifies concepts that are candidates for meeting this characterization and determines their origin. Second, it examines their validity. The first task is performed by metaphysical deductions, the second by transcendental deductions. This approach to transcendental philosophy is innovative in at least two respects. 8

9

10

h

d i

Let me here add a terminological note. Throughout this book, I will use ‘Critique of Pure Reason’ (in italics and with capitalization) to refer to Kant’s book in its entirety. By contrast, I will use ‘critique of pure reason’ (in roman font and without capitalization) to refer to that discipline within the Critique that achieves the latter’s aim as the doctrine of method of metaphysics. Admittedly, ‘a priori concepts for the cognition of objects’ is a vague formulation. I adopt it because if we consider the concepts studied by transcendental philosophy, they seem to be relevant to the cognition of objects in different ways. Some, like the categories, are constitutive of our cognition of objects. Others, like the ideas of reason used regulatively, are conditions for attaining some cognitions of objects but are not themselves constitutive of those cognitions. I here follow Pluhar’s translation of the first Critique in translating Stammbegriff as ‘root concept’.

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

6

Introduction

It characterizes the investigation into the validity of a priori cognition as only being concerned with establishing positive results. That is, it is not the task of transcendental philosophy to set limits to our cognitions. Rather, it determines where and why an a priori cognition is valid. Additionally, I argue that the distinction between the two tasks of transcendental philosophy can be drawn in all main parts of the Transcendental Doctrine of Elements, not only in the Transcendental Analytic. There is a metaphysical and a transcendental deduction of space and time in the Aesthetic, a metaphysical and a transcendental deduction of the categories in the Analytic, and a metaphysical and a transcendental deduction of the ideas of reason in the Dialectic. Let us now move to the second problem. In what sense is Kant’s first Critique relevant to the metaphysics of morals? Clearly, it does not provide an analysis of or foundation for moral principles. According to my approach, the Critique is relevant to the metaphysics of morals because it shows that the latter can form a coherent part of metaphysics as a whole. This does not necessarily involve providing a direct justification of moral principles, but it brings us to a characterization of the second discipline established in the Critique, namely, the critique of pure reason as the doctrine of method of metaphysics. As such, the critique must show that metaphysics as a whole (comprising both its theoretical and its practical parts) can become a science because it can achieve ‘architectonic unity’. In my account, ‘architectonic unity’, while being a condition of science, is different from mere systematicity. In order to attain architectonic unity, a science must realize what Kant calls its ‘idea’, which I take to be the correct description of the body of cognitions belonging to a science and the parts–whole relationships within it. In order to legitimately attribute architectonic unity to a body of cognitions, the latter must at least meet two minimal conditions. First, it must possess systematic coherence. I take a body of cognitions to be systematically coherent when: (a) the cognitions belonging to it are interconnected in a way that involves relations of either logical implication, explanatory support or both; and (b) it does not contain contradictions. Second, it must be possible to view the body of cognitions as realizing the fundamental ‘idea’ of a science, where this idea must (i) define the fundamental object of that science and (ii) prescribe the ordering of the body of cognitions that form that science. The main obstacles to the systematic coherence of metaphysics are the disputes in special metaphysics that form what Kant calls the ‘conflict of reason with itself ’. Kant’s solution to these disputes rests on drawing the negative consequences of the doctrines established in

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Introduction

7

transcendental philosophy. Accordingly, the critique of pure reason must first of all show how the very way in which the validity of root concepts is proved by transcendental philosophy implies that their validity is limited, which in turn can put a stop to the metaphysical disputes in question. Putting a stop to these disputes is insufficient for establishing that metaphysics can attain architectonic unity, however. The critique must also show that metaphysics can indeed be seen as realizing its fundamental ‘idea’. This clarifies, first, why the main task of the critique of pure reason is positive and, second, why the critique is relevant to the metaphysics of morals. For Kant, an essential part of the proper ‘idea’ of metaphysics is constituted by cognitions belonging to the practical part of metaphysics. The latter cognitions appear to demand a commitment to propositions in special metaphysics that the negative part of the critique deems beyond the scope of possible cognition. In this respect, the critique must show that we can meet the demands of the practical part of metaphysics without endangering the agnosticism regarding objects of pure reason that is established by the negative part of the critique of pure reason. The negative task of the critique, which is concerned with establishing limits to the use of root concepts, is thus best seen as merely subordinate to its main positive aim, that is, establishing that metaphysics can meet the two minimal conditions of architectonic unity. If this is right, we can see that the main task of the critique of pure reason, at least as it is conceived in the first Critique, is not to provide a justification of certain synthetic a priori principles (which is a task that is already part of metaphysics), but rather to show that a body of cognitions, including those very synthetic a priori principles, can form a whole with a proper unity that bestows the status of science to metaphysics. Let us move to the third problem identified in Section 1. Why does Kant include rational cosmology and rational theology in his sketch of the structure of metaphysics? An ‘easy’ way to put aside this worry is to stress that Kant’s sketch contains ‘true as well as apparent’ (A841/B869) cognitions. Accordingly, one could suggest that metaphysics includes analyses of the concepts of the world and God, but that it takes into account that they only constitute ‘illusions’, which is coherent with the results of the Dialectic. In my view, this explanation is unsatisfactory. It is unable to clarify why those ‘apparent’ cognitions should be part of metaphysics. I believe that a better answer can be obtained when we take the perspective of the critique in its relationship to the practical part of metaphysics. I have suggested that one chief task of the critique is to show that we can meet the demands of the practical part of metaphysics, where these demands

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

8

Introduction

concern commitments towards objects that customarily belonged to special metaphysics, such as God, the world, and the soul. Even though the commitments in question are ‘practically’ justified for Kant, since our justification rests on a moral argument, the propositions to which we commit ourselves are ‘theoretical’ because they are descriptive and do not concern ‘oughts’. In this sense, they belong to the ‘theoretical’ part of metaphysics, even though they are grounded in its practical part and do not constitute ‘cognitions’. I believe that we should read at least Kant’s inclusion of rational theology within his sketch of metaphysics in this sense.11

3  Kant and the ‘System’ of Metaphysics Reading the Critique of Pure Reason as the doctrine of method of metaphysics puts metaphysical concerns at the core of Kant’s project.12 But did Kant view the Critique as presenting his ultimate answer to at least some of these concerns, or did he take his investigation to be merely provisional? This question is not trivial, since many contemporaries of Kant, who took themselves to be furthering his approach, thought that the Critique only paved the way for a metaphysical investigation that ultimately needed a different and more fundamental foundation. Kant bitterly reacted to these attempts to ‘complete’ his plan. Famously, in his open letter on Fichte’s Wissenschaftslehre, published in 1799, he wrote: I must remark here that the assumption that I have intended to publish only a propaedeutic to transcendental philosophy and not the actual system of this philosophy is incomprehensible to me. Such an intention could never have occurred to me, since I myself, in the Critique of Pure Reason, have lauded the completed whole of pure philosophy as the best indication of the truth of that work. (12:370–71, translation altered)13

Kant’s open letter was motivated by an anonymous review of Johann Gottlieb Bohle’s Entwurf der Transcendental-Philosophie, in which the author of the review invited Kant to express his thoughts on Fichte’s Wissenschaftslehre, a project that, according to the reviewer, was the 11

12

13

h

d i

0 0

Unlike Ypi (2021: 16, 168–72), I do not see Kant’s inclusion of this discipline as involving a step back into dogmatism. As far as rational cosmology is concerned, recall that freedom is introduced as a cosmological idea and that it is essential from the perspective of the practical part of metaphysics. An emphasis on the relevance of these concerns has characterized various recent interpretations of Kant’s work. For example, Jauernig (2021) provides a ‘metaphysical’ interpretation of transcendental idealism; Willaschek (2018) provides an account of what we have called ‘metaphysics as a natural predisposition’; De Boer (2020) focuses on Kant’s projected system of metaphysics and its relationship to the Wolffian tradition. I here follow De Boer’s (2020: 251–2) suggested translation of the second sentence.

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Introduction

9

completion of the work that Kant initiated with his propaedeutic ‘critique’.14 The idea that Kant’s critical investigation was ‘provisional’ and needed to be brought to completion was not only the reviewer’s or Fichte’s but was shared by many philosophers who regarded themselves as ‘Kantians’, including Jakub Sigismund Beck and Karl Leonhard Reinhold. The common idea behind these projects was that Kant’s system remained incomplete because it lacked a unique and single principle that could serve as the ultimate foundation of metaphysics. In their view, without such a principle, metaphysics could not attain the unity of a system. We can understand Kant’s irritation with the claim that his position needed a further foundation. Still, his astonishment at the suggestion that he ‘intended to publish only a propaedeutic’ seems unwarranted. After all, he himself stressed at various points that the Critique was only a ‘preparation’ to metaphysics. Accordingly, many interpreters have taken Kant’s open letter to sharply contradict his earlier view. In contrast to this approach, Karin de Boer has recently argued that the open letter is coherent with Kant’s description of the Critique as a propaedeutic. She emphasizes that in this passage Kant is challenging not the contention that the system is incomplete but the contention that he did not intend to complete it (De Boer 2020: 251–2). In other words, he did not deny that the Critique was a propaedeutic but only that he was content with it and did not want to complete it himself. While De Boer is certainly right regarding what the passage literally says, I believe that Kant’s irritation was caused by a misrepresentation not only of what he intended to achieve, but also of what he took himself to have already achieved. This does not mean that we cannot make sense of his astonishment. I have already mentioned that the philosophers to whom Kant was responding viewed Kant’s position as incomplete because it lacked a proper foundation. This also applies to the analyses of ‘root concepts’ that, in my view, constitute the part of transcendental philosophy that is established in the Critique of Pure Reason. It is true that Kant did view transcendental philosophy as it is presented in the Critique as incomplete, but not because it lacked a proper ‘foundation’. Rather, what it still lacked was an identification of the ‘derivative’ concepts that rested on the root concepts singled out in the Critique. If we keep this in mind, we can understand Kant’s contentions in the passage without assuming any radical break with his earlier view. Kant took himself to have already established some parts of metaphysics within the pages of the Critique. 14

h

d i

The review appeared in the Literatur-Zeitung on Friday, 11 January 1799.

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

10

Introduction

These parts correspond to parts of transcendental philosophy, which are not in need of a further foundation. This is compatible with viewing the Critique as playing a propaedeutic role with respect to metaphysics as a whole, which Kant clearly did not think he had ‘completed’ in the Critique of Pure Reason. Therefore, we can make perfect sense of Kant’s open letter as confirming that the critique of pure reason rests on the previous establishment of some parts of transcendental philosophy.

4  The ‘Method’ of the Critique of Pure Reason Reading the Critique as the doctrine of method of metaphysics also has consequences for investigations aimed at determining the ‘method’ that Kant follows in the Critique of Pure Reason. Here, ‘method’ means a particular procedure of argument. These investigations often revolve around what a ‘transcendental argument’, a ‘transcendental proof’, or a ‘transcendental deduction’ is for Kant (see for example Strawson 1966; Henrich 1969; Henrich 1989; Carl 1992; Engstrom 1994; Cassam 1987; Ameriks 1978; Ameriks 2003; Hatfield 2003; G. Bird 2006b; Callanan 2006; Stapleford 2008; Moore 2010; Pereboom 2019). The purpose of these discussions is mainly to evaluate the aims, structure and validity of famous arguments from the Critique of Pure Reason, such as the transcendental deduction of the categories, the refutation of idealism and the second analogy, where these arguments are taken to be paradigmatic of Kant’s method. Alternatively, some interpreters have focused on the notion of ‘transcendental reflection’ that Kant introduces in the Amphiboly of the Concepts of Reflection as a key to unravelling what is distinctive about Kant’s argumentative strategy in the Critique (Longuenesse 1998: Chs. 5–6; Leitner 1994; Smit 1999).15 The task of transcendental reflection is to provide a ‘transcendental topic’ in which a concept is assigned its appropriate ‘transcendental place’ either in the understanding or in sensibility, according to the faculty through which the object of that concept is given to us (A268–9/324–5).16 If my suggestion that there are actually two disciplines established within the pages of the Critique is right, it seems to follow that the attempt to provide a unitary account of its method is doomed from the start, because we need at least to distinguish between the ‘method’ of transcendental 15

16

h

d i

0 0

One problem with this approach is that Kant’s position in the Amphiboly resembles his perspective in the Inaugural Dissertation. Accordingly, it is unclear whether the Amphiboly is completely coherent with Kant’s changed view in the Critique. See Willaschek (1998: 341–2). On transcendental reflection, see also De Boer (2020: Ch. 7), Merritt (2015) and Gava (2019a).

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Introduction

11

philosophy and the ‘method’ of the critique of pure reason. However, even if we do make such a distinction, it is clear that Kant’s approach is much more pluralistic than any attempt to identify a single pattern of argument could allow.17 I do not think that this ‘pluralism’ is a problem. What gives ‘unity’ to both transcendental philosophy and the critique of pure reason are the aims that these disciplines pursue. Reading the Critique of Pure Reason as the doctrine of method of metaphysics helps us to clearly identify these aims.

5  Structure of the Book The book is divided into four parts. The first part focuses on Kant’s claim that the Critique of Pure Reason must determine whether metaphysics can become a science. Chapter 1 argues that one of the chief ways through which this is established is by showing that metaphysics is capable of ‘architectonic unity’, which, for Kant, is a condition of science. I clarify what architectonic unity is and claim that it is different from mere systematicity. Furthermore, I read the discussion of the ‘worldly’ concept of philosophy within this framework and suggest that for Kant metaphysics can achieve architectonic unity only when it is construed according to this concept. Chapter 2 introduces one of the main theses of the book, namely, that the Critique of Pure Reason is the doctrine of method of metaphysics. I provide a characterization of what a doctrine of method is and distinguish between the doctrine of method of general logic and the doctrines of method of particular sciences. One of the tasks of ‘particular’ doctrines of method is to show that a body of cognition has ‘architectonic unity’, which clarifies their relevance to establishing that a science is indeed a science. I provide an overview of the Transcendental Doctrine of Method and show that it is meant to serve as the doctrine of method of metaphysics. Since, according to Kant, doctrines of method come ‘at the end of a science’, it is not clear why the Transcendental Doctrine of Method is placed within the Critique. I submit that while the Transcendental Doctrine of Method is an anomaly insofar as it comes before the ‘whole’ of metaphysics has been established, it does rely on the establishment of some parts of metaphysics, which are presented in the Transcendental Doctrine of Elements. These parts belong to ‘transcendental philosophy’. I show why this is compatible with characterizing the Critique as a ‘propaedeutic’. Parts II and III are dedicated to reconstructing the methods of the two disciplines established in the Critique, namely transcendental philosophy 17

h

d i

See Parts II and III.

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

12

Introduction

and the critique of pure reason, where method is here understood as a particular procedure of investigation or argument. As I mentioned before, I characterize transcendental philosophy as an investigation into the a priori concepts for the cognition of objects that do not contain anything empirical (see A845/B873). The Critique only considers a subclass of these concepts, which Kant calls ‘root concepts’. Space, time, the categories, and the ideas of reason all count among root concepts. The positive results that transcendental philosophy establishes regarding these concepts consist, first, in identifying them and tracking their origin and, second, in showing what validity they have. The first task is achieved by metaphysical deductions, which are analysed in Chapter 3, while transcendental deductions, the topic of Chapter 4, are responsible for the second. A chief claim in Chapter 3 is that metaphysical deductions do not simply assume an already presupposed distinction among faculties but contribute to making this distinction by tracking the different origins of root concepts. Furthermore, both Chapter 3 and Chapter 4 show how Kant in fact uses various argumentative strategies in his ‘deductions’. Part III considers Kant’s strategy for showing that metaphysics can attain architectonic unity. This involves, first, showing that metaphysics can achieve systematic coherence. In this respect, the critique of pure reason has first of all a negative task. It sets limits to the validity of the root concepts for the cognition of objects analysed by transcendental philosophy. Insofar as the use of these concepts is at the basis of metaphysical disputes around God, the world and the soul, establishing these limits is instrumental to achieving coherence. It is precisely in determining these limits that the critique of pure reason is dependent on the results of transcendental philosophy. In Chapter 5, I argue, first, that Kant does not follow a unique pattern when it comes to establishing these limits and, second, that the identification of these limits is pursued in each main part of the Transcendental Doctrine of Elements, namely the Transcendental Aesthetic, the Transcendental Analytic, and the Transcendental Dialectic. In Chapter 6, I find confirmation of my distinction between ‘positive’ and ‘negative’ arguments regarding the validity of ‘root’ concepts in passages where Kant makes this distinction when discussing the categories in particular. While Kant provides different accounts of how these arguments relate to one another, they all confirm that this distinction can be made and that it is the negative argument that belongs to the critique of pure reason. Chapter 7 explores the positive task of the critique of pure reason, namely establishing that metaphysics can achieve architectonic unity. I argue that it is in this framework that Kant’s practical arguments for the

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Introduction

13

existence of God and the immortality of the soul should be understood. These arguments clearly go beyond establishing that God and immortality are possible, logically or otherwise. They establish that we can form a justified commitment to these objects, even though that commitment cannot amount to cognition. In my view, reading the critique as the doctrine of method of metaphysics allows us to account for the inclusion of these arguments within the Critique. These commitments are fundamental for attaining the ‘idea’ that can grant architectonic unity to metaphysics. However, it is important to show that they do not conflict with the limits that the critique sets in order to attain systematic coherence. Part IV investigates the relationship between Kant’s project in the Critique of Pure Reason, on the one hand, and the dogmatism and scepticism that Kant attributes to Wolff and Hume, on the other. Chapter 8 reconstructs Kant’s critique of dogmatism and Wolff. It shows that this critique has two layers: one directed against Wolff’s metaphilosophical views and one attacking his actual procedures of argument. Additionally, I provide an account of Kant’s claim that metaphysics can indeed proceed according to Wolff ’s dogmatic procedure after the critique of pure reason has concluded its task. In Chapter 9, the distinction between transcendental philosophy and the critique of pure reason is again central. I distinguish between three alternative readings of Kant’s interpretation of Hume. I set aside one reading as implausible and show how the other two are compatible once we distinguish between the perspectives of transcendental philosophy and the critique of pure reason.

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Part I

Metaphysics as a Science and the Role of the Critique of Pure Reason

Introduction to Part I This first part of the book aims to determine in which sense the Critique of Pure Reason is tasked with determining whether and how metaphysics can become a science. Kant claims that this is the main objective of the Critique in both the A- and the B-Prefaces. In Chapter 1, I focus on one implication of this claim. Kant submits that having ‘architectonic unity’ is a condition for becoming a science, which means that the Critique must show that metaphysics is capable of this unity. I argue that architectonic unity does not simply consist in the achievement of a systematically ordered body of cognitions. Rather, Kant believes that there is only one way of organizing a body of cognitions that can give it architectonic unity. Moreover, Kant claims that architectonic unity is achieved according to an ‘idea’ given a priori by reason. I argue that the idea that is the best candidate for providing architectonic unity to metaphysics is what Kant calls the ‘worldly concept’ of philosophy, which shows that the relationship between the theoretical and the practical parts of metaphysics is already a central topic of the Critique of Pure Reason. In two often overlooked passages, Kant equates the Critique with a ‘doctrine of method’ (A82–3/B108–9) and a ‘treatise on method’ (Bxii). In Chapter 2, I argue that these claims should be taken literally. Kant understands the Critique of Pure Reason as the doctrine of method of metaphysics. Doctrines of method have a twofold task: first, they must provide object- or cognition-dependent methodological rules for how to proceed in a given science; second, they must show that a science in fact possesses ‘architectonic unity’. Moreover, in accomplishing this task, they necessarily rely on already established parts of a science. This has important consequences for how we understand the relationship between transcendental philosophy, as one doctrinal part 15

https://doi.org/10.1017/9781009172127.002 Published online by Cambridge University Press

16 Metaphysics as a Science and the Role of the Critique of Pure Reason of metaphysics, and the critique of pure reason, as that discipline which, within the Critique, achieves the latter’s aim as the doctrine of method of metaphysics. More precisely, in pursuing its objectives, the critique of pure reason relies on the establishment of doctrines belonging to transcendental philosophy.

https://doi.org/10.1017/9781009172127.002 Published online by Cambridge University Press

chapter 1

The Worldly Concept of Philosophy and the Possibility of Metaphysics as a Science

When Kant describes the project of the Critique of Pure Reason, he often stresses that its main aim is to establish whether and how metaphysics is possible as a science. In the words of the Prolegomena, the first Critique ‘contains within itself the whole well-tested and verified plan by which metaphysics as science can be achieved, and even all the means for carrying it out’ (4:365). Even though these remarks are well known to Kant scholars, it is not always clear what they imply. According to Kant, one condition for attaining the status of a science is possessing ‘architectonic unity’. A consequence of this is that, if it wants to show that metaphysics can become a science, the first Critique must also show that it can attain architectonic unity.1 In this chapter, I will attempt to clarify what this means.2 My first aim is to illustrate that architectonic unity does not simply consist in the achievement of a systematically ordered body of cognitions. Rather, there is only one way of ordering the cognitions belonging to a science that can provide it with architectonic unity. I will then go on to discuss the ‘idea’ that is a candidate for providing unity to metaphysics. There is a complication, though, since Kant identifies two candidate ‘ideas’ according to which metaphysics can be understood. Metaphysics can be construed according to either the ‘school concept’ (Schulbegriff ) or the ‘worldly concept’ (Weltbegriff ) of philosophy.3 I will argue that it is only according to the ‘worldly concept’ that metaphysics can attain architectonic unity and become a science. 1

2

3

Scholars who have insisted on the importance of architectonic unity for a proper understanding of what science is according to Kant include La Rocca (2003: Ch. 6), Manchester (2003), Manchester (2008), Goy (2007), Sturm (2009: Ch. 3), Ferrarin (2015: Ch. 1), Mensch (2013: Ch. 7), Fugate (2019) and Ypi (2021). Kant identifies additional conditions for regarding a body of cognitions as a science. For the conditions that apply to what Kant calls ‘proper’ natural science, see Pollok (2001: Ch. 3.1), Sturm (2009: 149–53), Van den Berg (2011) and Stan and Watkins (2014). I here follow Pluhar’s translation of the first Critique in translating Schulbegriff as ‘school concept’ and Young’s translation of the Lectures on Logic by rendering Weltbegriff ‘worldly concept’.

17

h

d i

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

18 Metaphysics as a Science and the Role of the Critique of Pure Reason That means that the Critique of Pure Reason must show how metaphysics according to the worldly concept can be regarded as capable of architectonic unity.

1  The Architectonic Unity of the Sciences In the chapter on the Architectonic of Pure Reason, Kant sometimes gives the impression that the demand for architectonic unity in the sciences is simply a demand for systematicity. For example, he begins the chapter by saying that ‘systematicity is that which first makes ordinary cognition into science, i.e., makes a system out of a mere aggregate of it’ (A832/B860). In this section, my main aim is to show that architectonic unity is not simply a demand for systematicity. I will first clarify the sense in which systematicity is a demand that our reason places on our cognitions in general, according to Kant. This general demand should be distinguished from what Kant calls the ‘architectonic unity’ of a science. Unlike Kant’s general demand for systematicity, which can be met in different ways, there is only one ordering of the cognitions belonging to a science that can grant architectonic unity. Additionally, the idea of a science does not necessarily need to remain a ‘focus imaginarius’, namely an ‘ideal’ that we cannot actually attain. There is a problem with Kant’s account, however: Kant does not provide criteria for conclusively determining when a body of cognitions has architectonic unity. I will suggest that we can nonetheless single out minimal criteria for architectonic unity. These are unable to conclusively determine when a body of cognitions possesses architectonic unity. Rather, they can simply identify with certainty cases in which architectonic unity cannot be attributed. In this section, I will discuss: (a) reason’s general demand for systematicity; (b) the architectonic unity of a science; and (c) the minimal criteria for architectonic unity. a. Systematicity as a demand of reason. In the Architectonic, Kant hints at the systematicity that reason demands of all our cognitions: ‘Under the government of reason our cognitions cannot at all constitute a rhapsody but must constitute a system, in which alone they can support and advance its essential ends’ (A832/B860). I will later discuss the role that essential ends play in the architectonic unity of metaphysics. Here, my focus is on the claim that reason demands systematicity of cognitions. The claim reflects Kant’s idea that reason, understood in a narrow sense as the faculty of inference, strives to draw inferential connections

h

d i

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

The Worldly Concept of Philosophy

19

among our cognitions (see Willaschek 2018: Chs. 1–2). In the words of the Appendix to the Transcendental Dialectic: ‘If we survey the cognitions of our understanding in their entire range, then we find that what reason quite uniquely prescribes and seeks to bring about concerning it is the system in cognition, i.e., its interconnection based on one principle’ (A645/B673; see also A302/B359). Under ‘principle’ in a broad sense, Kant understands a universal proposition that can serve as the major premise in a syllogism (see A300–1/B357–8). Therefore, the kind of systematic unity that reason can achieve is one in which different cognitions are shown to be syllogistically derivable from the same major premise. If we think of the propositions ‘Humans are mortal’, ‘Dolphins are mortal’ and ‘Horses are mortal’, we can see them as all derivable from the proposition ‘Mammals are mortal’. In order to establish this derivability relation, reason introduces the concept of a ‘mammal’, which allows us to see the concepts of a ‘human’, a ‘dolphin’ and a ‘horse’ as species concepts that fall under the same genus concept. Once we establish that ‘mortality’ is a characteristic that obtains at the genus level, we are able to represent the propositions attributing mortality at the species level as dependent on the proposition that attributes mortality at the genus level. Take the proposition ‘Humans are mortal’: it is now syllogistically obtainable from the major premise ‘Mammals are mortal’ and the minor premise ‘Humans are mammals’. The search for a ‘principle’ that provides unity to different cognitions by showing that they are equally derivable from a more general proposition can be carried further. We could show that the proposition ‘Mammals are mortal’ is further derivable from the proposition ‘Animals are mortal’. According to Kant, reason prescribes that we proceed in this way and try to bring our cognitions under the smallest possible number of ‘principles’ (A305/B361). This is not the only way in which reason increases the systematicity of our cognitions by finding inferential connections among them. Reason can also show how more specific cognitions are derivable from more general ones. From ‘Horses are mortal’ we can derive the propositions ‘Andalusian horses are mortal’ and ‘Arabian horses are mortal’ once we determine that ‘Andalusian horses’ and ‘Arabian horses’ are special horse breeds. In addition to introducing principles at the ‘top’ of a syllogistic chain, reason can thus also add conclusions at the ‘bottom’. But that’s not all. Reason can also increase the inferential connections within the chain. For example: instead of deriving ‘Horses are mortal’ directly from ‘Mammals are mortal’, we can introduce a further concept between ‘mammals’ and ‘horses’ and derive the proposition ‘Horses are mortal’ only mediately from ‘Mammals

h

d i

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

20 Metaphysics as a Science and the Role of the Critique of Pure Reason are mortal’ by first establishing that ‘equids are mortal’ through the proposition ‘Equids are mammals’. That ‘horses are mortal’ would than follow from the further premise that ‘Horses are equids’. The different ways in which reason can draw inferential relationships among cognitions find a correlate in the principles of homogeneity, specification and continuity of forms, which Kant discusses in the Appendix to the Transcendental Dialectic in the context of his remarks on the regulative use of reason (see A658/B686).4 These principles require that we investigate nature by presupposing, respectively, that there is ‘sameness of kinds’ (A657/B685), ‘variety of what is same in kind under lower species’ (A657/B685) and finally, ‘affinity’ through a ‘continuous transition from every species to every other through a graduated increase of varieties’ (A657–8/B685–6) in objects of nature. Let us first consider the principles of homogeneity and specification. Their combination instructs us to study nature as if it were always possible to find, on the one hand, a general kind to which different natural objects belong and, on the other, sub-kinds of this general kind. Given an already recognized set of genera and species, these two principles instruct us to look further for higher genera and lower subspecies: higher genera that single out common traits among the greatest possible number of the known genera and lower subspecies that introduce further varieties of the known species. What about the principle of the continuity of forms in nature? The principle says that there is no discrete leap between species (see A659–60/B687–8). The most direct way to read this claim is to see it as saying that it is always possible to find an intermediate species between two given species belonging to the same genus (for two readings along these lines, see Wartenberg 1979: 412 and Goldberg 2004: 407). Kant seems to have something different in mind, however: the principle first refers to the continuity in the increasing variety when we descend from higher kinds to lower sub-kinds. The principle postulates ‘a graduated increase of varieties’ (A658/B686) in the specification of the higher kinds in lower sub-kinds. This means that, given a genus and its species, we must always assume that we can introduce intermediate species that are more specific than the genus but more general than the species. Understood in this way, the principle of continuity does not necessarily imply that we should assume that there is always an intermediate species between two species of the same genus. Rather, it means that we should regard two proximate species of the same genus as continuous because they resulted from a continuous process of specification. In this process, 4

h

d i

0 0

See also Ch. 4.

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

The Worldly Concept of Philosophy

21

an increasing number of determinations is added to the highest genus in order to obtain different species, subspecies, and so on. In Kant’s words: ‘all manifolds are akin one to another, because they are all collectively descended, through every degree of extended determination, from a single highest genus’ (A658/B686). There is thus an important convergence between the ways in which reason promotes systematicity by increasing the inferential connections among cognitions and the principles of homogeneity, specification and continuity that guide research into nature. The best way to account for this convergence is to see the latter as dependent on the former: it is the fact that reason seeks to increase the inferential relationships among cognitions in the way I have described that brings us to see nature as if it were structured in a way that can satisfy our attempts at inferential systematization. Kant makes an additional point, though: reason’s search for inferential relationships among cognitions is one fundamental way in which we obtain genuine new knowledge. Think of the examples of natural classification provided above: looking for new inferential connections among cognitions can lead us to postulate completely new classes (either at the ‘top’, in the ‘middle’ or at the ‘bottom’), which, if backed by empirical evidence, can enter the set of our cognitions as genuine empirical knowledge. This suggests that following reason’s search for systematic inferential relationships is a condition for increasing our knowledge set in certain respects (that is, there is some knowledge that can only be obtained through attempts along these lines). But when the knowledge in question concerns nature, Kant submits, only if we assume that nature is structured in a way that allows for inferential systematization can we be rational in our attempt to increase inferential relationships among cognitions. Accordingly, we are not only inclined but also justified to regard nature as if it were designed to fit our efforts.5 Before turning to the architectonic unity of the sciences, I wish to emphasize three further issues. First, reason’s demand for systematization through increasing the inferential relationships among cognitions does not exclusively apply to natural knowledge. It is rather a demand that reason makes on every cognition we have. For example, it applies to cognitions 5

h

d i

By using the formulation ‘as if ’ here and above, I am not committing myself to a ‘fictionalist’ account of regulative ideas, that is, one according to which ideas involve false propositions that are nonetheless practically useful for improving our empirical knowledge. I take the ‘as if ’ terminology simply to imply that we can assume some propositions in our research into nature, even though we cannot (ever) determine whether they are true or false. Kant himself uses the as if terminology (see A616/B644; A619/B647; A670–1/B698–9; A672–3/B700–1; A678/B706; A681/B709, etc.). See also my account of the transcendental deduction of ideas in Ch. 4.

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

22 Metaphysics as a Science and the Role of the Critique of Pure Reason in the mathematical and the practical domains. Therefore, even though the principles of the homogeneity, specification and continuity of nature are certainly expressions of reason’s demand for systematicity, they only exemplify one instance in which this demand has consequences for how we represent objects of cognition and their relationships with one another. Second, both reason’s general demand for systematicity and the principles of the homogeneity, specification, and continuity of nature provide an ideal6 that we must try to approach but can never actually attain. These ideals are, respectively, a system in which we have an infinite set of inferential relationships among cognitions and a system of genera and species in which we start from a unique higher genus at the top and descend to a set of lower species through an infinite series of intermediate genera and species. It is possible to find a supreme principle at the top of our inferential system, but it seems impossible to reach a point at which our search for further inferential relationships in the ‘middle’ or at the ‘bottom’ of an inferential chain cannot in principle proceed further. Similarly, we can imagine that a unique higher genus of nature is possible, but not that lower species cannot in principle be specified further (see A655–6/B683–4). Additionally, given the principle of continuity, it seems that the possibility of searching for intermediate classes between given genera and species is in principle endless. These ideals of systematicity therefore have the status of a ‘focus imaginarius’, that is, a posited end that we cannot ever really actually achieve. In the Appendix to the Transcendental Dialectic, Kant explicitly attributes this status to the ideas of reason in their regulative use (A644/B672). Third, the ideals of systematicity provided by reason can (partly) be met in more than one way, which relates to their regulative status. Accordingly, reason prescribes what sort of inferential relationships, or what sort of genera/species relationships, we must seek in our cognitions, but it does not precisely determine a priori which system of cognitions satisfies those requirements. Rather, reason leaves open how the demand for systematicity will be met in concreto, such that different systems of cognitions can equally satisfy that demand. b. The architectonic unity of a science. How does reason’s general demand for systematicity relate to the architectonic unity of the sciences? I will argue that architectonic unity departs from the general requirement of systematicity of cognition in at least two respects. On the one hand, the 6

h

d i

0 0

Here, I do not use the word ‘ideal’ as Kant characterizes it in the Ideal of Pure Reason.

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

The Worldly Concept of Philosophy

23

ideas on which architectonic unity is based do not necessarily remain a ‘focus imaginarius’. On the other, for each science, there is only one system of cognitions according to which we can ascribe architectonic unity to it. But what is architectonic unity? Kant emphasizes that it depends on an ‘idea of reason’: I understand by a system, however, the unity of the manifold cognitions under one idea. This is the rational concept of the form of a whole, insofar as through this the domain of the manifold as well as the position of the parts with respect to each other is determined a priori. The scientific rational concept thus contains the end and the form of the whole that is congruent with it. The unity of the end, to which all parts are related and in the idea of which they are also related to each other, allows the absence of any part to be noticed in our knowledge of the rest, and there can be no contingent addition or undetermined magnitude of perfection that does not have its boundaries determined a priori. (A832–3/B860–1)

Kant does not explicitly speak of ‘architectonic unity’ in this passage, but since the latter is contained in the chapter on the Architectonic of Pure Reason and the ‘system’ that is at stake is one that encompasses cognitions belonging to a science, it is clear that he has architectonic unity in mind. Kant makes a series of substantial claims: first, the architectonic unity of a science is based on an idea that is given a priori by reason. Second, this idea is described as ‘the rational concept of the form of a whole’. As such, the idea should be able to determine the position of each part of the system and its relationship to the whole. It should also determine the boundaries of the system. Third, the idea functions as an ‘end’ in the sense that each part is considered in terms of its function in the realization of the whole.7 Fourth, this parts–whole relationship allows us to infer the existence of one part of the system from our knowledge of the other parts. These remarks can be taken as expressions of the need to avoid arbitrariness in our account of the structure of a science: when we consider a particular proposition of a science, or a particular sub-division of it, we regard the place of that proposition or that sub-division in the theory – its theoretical relation to other propositions and parts of the theory – not as arbitrary or accidental but as depending on internal relations that are objective and independent of our individual standpoint. In a science, some 7

h

d i

Kant often emphasizes this aspect with the use of biological metaphors. For example, he claims that the whole of a science is ‘like an animal body’ (A833/B861). On the relationship between architectonic and biological metaphors in the Architectonic of Pure Reason, see Ferrarin (2015: Ch. 1). On the historical background of Kant’s biological metaphors, see Mensch (2013).

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

24 Metaphysics as a Science and the Role of the Critique of Pure Reason propositions are more fundamental than others, and this is not a matter of choice. There are also propositions that clearly do not belong to that science and thus do not contribute to its unity and structure.8 Kant’s claim that a single idea given a priori by reason determines the form of the whole of a science captures the lack of arbitrariness that we attribute to the internal relationships among the doctrines of a science. To substantiate this view, let us consider for a moment Kant’s distinction between ‘architectonic’ and ‘technical’ unity. While ‘architectonic unity’ expresses the correct order of the cognitions belonging to a science because it ‘arises only in consequence of an idea’ (A833/B861) and is thus not arbitrary, ‘technical unity’ provides an order that arises from ‘aims occurring contingently’ (A833/B861), or ‘for all sorts of arbitrary external ends’ (A833/B861), which cannot be considered adequate for a science. Since one function of the ‘idea of reason’ which grants architectonic unity to a science is that of guaranteeing lack of arbitrariness when it comes to how its cognitions are organized, and since a different, equally systematic organization of those cognitions would only provide technical unity, only one way of ordering the body of cognitions of a science provides architectonic unity. Because the idea has this function, I submit that an adequate characterization of it is the following: it is a correct description of the body of cognitions that form a science and its parts–whole relationships. But how can the ‘idea of reason’ perform this function? What is its status? As we saw, Kant maintains that the idea is given ‘a priori’. This generates a worry: how can an idea that guarantees lack of arbitrariness in the organization of a body of cognitions be given a priori? Take the case in which the science and the cognitions in question are empirical. How can reason grasp the order of those cognitions independently of experience? To answer this worry, it is sufficient to note that the architectonic that is at stake when Kant makes these claims is an architectonic of pure reason. The cognitions whose order we can grasp a priori are therefore also a priori.9 Accordingly, the sciences that Kant discusses in the Architectonic chapter are all a priori: mathematics and the various parts of metaphysics.10 It 8 9

10

h

d i

0 0

Here, I re-elaborate claims I make with Marcus Willaschek in Gava and Willaschek (forthcoming). It is helpful to keep in mind that Kant has a rather demanding conception of ‘proper’ natural science, where he argues that proper natural sciences need to be based on a priori principles (see Pollok 2001: Ch. 3.1; Van den Berg 2011; Stan and Watkins 2014; McQuillan 2017b). It is true that he does also briefly speak of ‘empirical philosophy’ (A840/B868) and empirical psychology (A848–9; B876–7), but they are clearly not the focus of the chapter. Moreover, since these also rest on the a priori part of philosophy, that is, metaphysics, it is not implausible to think that the ‘order’ of their cognitions actually depends on a priori principles that they borrow from metaphysics.

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

The Worldly Concept of Philosophy

25

therefore seems accurate to say that if reason has a priori access to the cognitions that make up the sciences discussed in the Architectonic, it must also have a priori access to how they are related to one another. There is, however, a second worry concerning Kant’s claim that the idea of a science is given a priori by reason. This becomes apparent if we consider two objects in which the idea of a science is historically realized: its ‘schema’ and its ‘definition’. The schema represents the concrete way in which the doctrines of a purported science are ordered by its practitioners in a particular historical time. Accordingly, Kant contends: ‘[f]or its execution the idea needs a schema, i.e., an essential manifoldness and order of the parts determined a priori from the principle of the end’ (A833/B861). In other words, the schema represents the historical attempts to realize the idea in an actually existing ordered body of cognitions. Kant does not say much about what the definition of a science is, but we can assume that it should at least identify what the fundamental object of a science is.11 Let us now see what kind of relationship to these historical realizations the idea enjoys: But in its elaboration the schema, indeed even the definition of the science which is given right at the outset, seldom corresponds to the idea; for this lies in reason like a seed, all of whose parts still lie very involuted and are hardly recognizable even under microscopic observation. For this reason sciences, since they have all been thought out from the viewpoint of a certain general interest, must not be explained and determined in accordance with the description given by their founder, but rather in accordance with the idea, grounded in reason itself, of the natural unity of the parts that have been brought together. For the founder and even his most recent successors often fumble around with an idea that they have not even made distinct to themselves and that therefore cannot determine the special content, the articulation (systematic unity) and boundaries of the science. (A834/B862)

What does this passage tell us about the relationship between the idea and its historical realizations? It makes clear that Kant’s claim that the idea is given ‘a priori’ by reason cannot mean that we have unproblematic access to it prior to and independently of obtaining particular cognitions belonging to a science. Accordingly, the idea is something that remains unclear and obscure even to those who found a particular science. It is only in the process of developing the science that the idea gains clarity and distinctness. At this point, a second worry arises: if the idea can only be grasped after we are already involved in developing the cognitions belonging to a 11

h

d i

Here I disagree with Sturm (2009: 162), who claims that the idea depends on the definition and not the other way around. If it is in relation to the idea that we can determine whether the definition is adequate, it seems more accurate to say that the definition depends on the idea.

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

26 Metaphysics as a Science and the Role of the Critique of Pure Reason science, in what sense can it be given ‘a priori’? I believe that this worry can easily be answered when we take into account the fact that reason is not transparent to itself for Kant.12 We have seen that the sciences that Kant discusses in the Architectonic comprise a priori cognitions of reason. We have also seen that the ‘idea’ that provides unity to these different sciences is a correct description of the body of cognitions that form a science and its parts–whole relationships. Kant clearly believes that reason does not have unproblematic access to the a priori cognitions that are in its power. As the Transcendental Dialectic makes clear, reason has a natural tendency to make claims that it cannot legitimately make. Just as reason does not have unproblematic access to the a priori cognitions that are in its power, it lacks unproblematic access to the way in which those cognitions are ordered. This explains the sense in which the idea of a science can be both given a priori and only graspable through its historical realizations: the idea of a science describes the order of a body of a priori cognitions of reason that is potentially available to reason from the very beginning. Reason, however, is not transparent to itself. As a consequence, it takes time for it to finally organize that body of cognitions according to its correct order and thus realize the idea in an existing schema. Let me now draw two consequences of my characterization of architectonic unity. First, there is nothing in Kant’s description of the idea of a science that prevents this idea from being fully realized in an existing body of cognitions, which is then correctly seen as a science because it possesses architectonic unity. It seems perfectly possible for the schema of a science, that is, the actual order of a body of cognitions that purports to be a science, to adequately realize the idea of the whole given by reason. Therefore, there is nothing in Kant’s account of the idea of a science that compels us to see the latter as necessarily remaining a ‘focus imaginarius’.13 Moreover, Kant thinks that some bodies of cognition have in fact already been established as sciences. Chief examples in this respect are mathematics and physics, which are used as models for science both in the Prolegomena (4:279–80) and in the B-Introduction to the first Critique (Bx–xviii). If these sciences are established, and if architectonic unity is a condition for 12 13

h

d i

0 0

On reason’s lack of self-transparency, see Gava (2016). Against my reading, Sturm (2009: 168) suggests that the idea of a science must remain a ‘focus imaginarius’. I agree with much of what Sturm says on architectonic unity. On this particular issue, I believe that the difference between our interpretations depends on the fact that Sturm does not distinguish, as I do, between the regulative ideas of reason that play a role in reason’s search for systematicity in nature and the ideas that ground the unity of different sciences. In Gava (2014), I presented a position closer to Sturm’s, since I claimed that the idea of a science is regulative.

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

The Worldly Concept of Philosophy

27

being a science, this means that they must possess architectonic unity and fully realize the ideas of the whole that give them unity. Second, the architectonic unity of a science does not depend on the relationships between a body of cognitions and other existing sciences. In other words, to establish whether a body of cognitions is a science, we do not need to take into consideration whether it fits into a unitary system encompassing all sciences. The question whether a science has architectonic unity is answered only with reference to the fundamental idea of that particular science. It is for this reason that Kant can say that mathematics and physics are sciences even though, in his account, a complete and unitary system of all the sciences does not yet exist.14 Of course, this does not mean that the achievement of a complete and unitary system of all sciences is not an important issue for Kant. On the contrary, it is an aim that directly depends on reason’s general demand for systematicity with regard to cognition. Moreover, as we will see, Kant thinks that it is only metaphysics that can offer a principle that gives architectonic unity to the system of all sciences. This does not imply that individual sciences cannot achieve architectonic unity independently of such an encompassing system, however. I hope I have made Kant’s account of the ‘idea’ of a science a bit less opaque. It is important to stress once again that, on the one hand, there is only one system of cognition that can grant architectonic unity to a science and that, on the other, the idea on which the architectonic unity of a science is based does not need to remain a ‘focus imaginarius’. c. The minimal criteria for architectonic unity. It should be clear that according to my characterization of architectonic unity, the latter is not only necessary but also sufficient for science. Since a body of cognitions that possesses 14

h

d i

Against my interpretation, Sturm (2009: 136) claims that the ‘internal systematicity’ of a science is subordinated to its ‘external systematicity’, that is, its place in a system of all the sciences. It is only by determining the ‘external systematicity’ of a science that we can provide it with ‘internal systematicity’. Sturm bases this claim on his interpretation of the role of the definition of a science. According to him, the role of the definition is to make clear the differences between a particular science and other sciences. Sturm supports his interpretation with the following passage from the Prolegomena: ‘If one wishes to present a body of cognition as science, then one must first be able to determine precisely the differentia it has in common with no other science, and which is therefore its distinguishing feature; otherwise the boundaries of all the sciences run together, and none of them can be dealt with thoroughly according to its own nature’ (4:265). In my view, this implies that the definition of a science must provide a criterion for determining what its specific object of study is. Therefore, the definition provides us with a tool for distinguishing between different sciences when we compare them. With that said, this implies neither that a comparison between different sciences is necessary for the definition nor that a science must be placed in a system of all the sciences to be considered a science.

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

28 Metaphysics as a Science and the Role of the Critique of Pure Reason architectonic unity is one that realizes the idea that correctly describes the body of cognitions belonging to a science and the parts–whole relationships pertaining to it, achieving that unity is sufficient for attaining the status of a science. That architectonic unity is both necessary and sufficient for science does not mean that features that Kant links with science – think of what he says on the role of a priori principles in ‘proper’ natural science (4:468–9)15 – are not important. Rather, it means that if a feature is deemed necessary for science, a body of cognitions that possesses architectonic unity will possess that feature, too, exactly because that body of cognitions realizes the correct idea of a science. Even though architectonic unity is sufficient for science, we have seen that the idea of a science is not something to which we have unproblematic access. Rather, it often remains obscure even to the practitioners of a science. Given the difficulty of accessing the idea, it is plausible that we do not have conclusive criteria for determining whether an existing body of cognitions realizes the idea of a science and thus has architectonic unity. Therefore, as we saw, even if it is possible that the idea of a science might be fully realized in an existing body of cognitions, we cannot be in a position to tell with absolute certainty whether this is in fact the case. I submit that we nevertheless do have certain minimal criteria that should be met if we want to attribute architectonic unity to a body of c­ ognitions. These criteria cannot, of course, guarantee that the body of cognitions in question actually has architectonic unity, but they can at least ensure that our attribution of architectonic unity is not u ­ nwarranted. One might point out that minimal criteria for attributing architectonic unity should involve those features that Kant considers necessary c­ onditions for science. Again, the identification of a priori principles in natural science can serve as an example. Of course, if a priori principles are necessary for natural science, it seems legitimate to require that we be able to identify principles of this kind when we attribute architectonic unity to a body of cognitions that purports to be a natural science, given that attributing architectonic unity to it would mean considering it a science. However, in this study I will focus on the minimal criteria for attributing a­ rchitectonic unity that specifically track the ‘unity’ that is at stake here. In this respect, a first criterion is systematic coherence. I take a body of cognitions to be systematically coherent when: (a) the cognitions belonging to it are interconnected in a way that 15

h

d i

0 0

It is unclear whether what Kant says on proper natural science in the Metaphysical Foundations of Natural Science applies to proper science in general. As far as the role of a priori principles in science is concerned, I take it that it does.

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

The Worldly Concept of Philosophy

29

involves relations of either logical implication, explanatory support, or both; and (b) it does not contain contradictions.16 Only bodies of cognitions that are systematically coherent are candidates for architectonic unity. A second minimal criterion for architectonic unity is that a body of cognitions must be able to be considered an expression of an ‘idea’. This means that we should be able to describe the cognitions belonging to a science as dependent on an idea that (i) defines the fundamental object of that science and (ii) prescribes the ordering of the body of cognitions that form that science.

2  Two ‘Ideas’ of Metaphysics According to the minimal criteria for architectonic unity I have just spelled out, we can say that if metaphysics wants to become a science, it must at least attain systematic coherence and be able to identify an idea that defines its fundamental object and gives organization to the whole. I will here focus on the ‘idea’ of metaphysics that can play this role in making metaphysics a science. According to Kant, there is certainly one characterization of metaphysics that must figure in the fundamental idea: metaphysics encompasses all discursive rational cognitions that spring from the nature of pure reason (A841/B869).17 However, this is not all that Kant says about the idea that can provide architectonic unity to metaphysics. In fact, I submit that Kant’s distinction between the ‘school concept’ and the ‘worldly concept’ of philosophy identifies two candidate ideas according to which the architectonic unity of metaphysics can be sought. Kant rules out that metaphysics according to the school concept could indeed attain architectonic unity, which means that it is only by pursuing metaphysics according to the worldly concept that we have a chance of success. In this section, I will discuss: (a) the two concepts of philosophy just mentioned; (b) the architectonic unity of metaphysics; and (c) metaphysics and the unity of all the sciences. a. The school concept and the worldly concept of philosophy. In the Architectonic of Pure Reason, Kant suggests that one can provide two very different accounts of what philosophy is. On the one hand, one can understand philosophy according to its ‘school concept’ (Schulbegriff ). In this sense, philosophy is a 16 17

h

d i

For two attempts to define the systematicity that is distinctive of science, see Rescher (1979: Ch. 6) and Hoyningen-Huene (2013). On the basis of this characterization, metaphysics can be distinguished from mathematics, which assembles intuitive cognitions based on the construction of concepts (A713/B741; A837/B865).

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

30 Metaphysics as a Science and the Role of the Critique of Pure Reason ‘system of cognition that is sought only as a science without having as its end anything more than the systematic unity of this knowledge’ (A838/B866). Alternatively, philosophy can be represented according to its worldly concept (Weltbegriff ). In this sense, it ‘is the science of the relation of all cognition to the essential ends of human reason’ (A839/B867). What consequences does this distinction have for the possibility of metaphysics as a science? The first thing to keep in mind is that these two concepts are concepts of what philosophy is, not metaphysics. Philosophy comprises both metaphysics, which forms its ‘a priori’ part, and ‘empirical philosophy’, which is constituted by rational cognitions from empirical concepts (A840/B868). Even though philosophy is broader than metaphysics, it is legitimate to use the ‘school concept’ and the ‘worldly concept’ to characterize what metaphysics is. To begin with, as far as the school concept is concerned, metaphysics must certainly form an independent ‘system of cognitions’ with the status of a science for Kant. Additionally, it is in the ‘a priori’ part of philosophy that the essential ends of reason are made a topic of philosophical interest. Therefore, if the school concept and the worldly concept of philosophy are candidates for providing unity to philosophy in general, this is only because they are first of all concepts that can provide unity to metaphysics in particular. It is then no coincidence that Kant says of the a priori part of philosophy that it ‘alone constitute[s] that which we can call philosophy in a genuine sense’ (A850/B878). What kinds of metaphysics result from these two concepts, respectively? Let us begin with the school concept of philosophy. According to it, metaphysics is only concerned with achieving a ‘systematic unity of knowledge’ that is sought as a science. Now, since we know that metaphysics must encompass all discursive rational cognitions that spring from the nature of pure reason (A841/B869), the task of metaphysics according to the school concept is simply to provide a systematic ordering of these cognitions. An ordering along these lines is what governs Kant’s distinction between a metaphysics of nature and a metaphysics of morals, on the grounds that the former comprises theoretical cognitions from a priori concepts while the latter examines a priori principles of action (A841/B869). Once this distinction is in place, one can then hierarchically organize the cognitions belonging to each branch. This is what the distinction between transcendental philosophy and the ‘physiology of pure reason’ does, for example (A845/B873). What about metaphysics according to the worldly concept of philosophy? What are the ‘essential ends’ of human reason that are explicitly considered in the science on this conception? The claim that reason has ‘ends’ is central to Kant’s account of reason as a faculty. It can be taken as the view that reason is intrinsically teleological because it is directed towards

h

d i

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

The Worldly Concept of Philosophy

31

the realization of the states of affairs that would result from the perfect and complete application of this faculty. Accordingly, the achievement of complete rational knowledge is an end that derives from the theoretical use of reason. By contrast, the achievement of the highest good, that is, of a world in which morality is rewarded with the happiness one has made oneself worthy of, is an end that stems from the practical use of reason (on Kant’s teleological account of reason, see Kleingeld 1998 and Ferrarin 2015: Ch. 1; see also Gava and Willaschek forthcoming). In his account of the worldly concept of philosophy, Kant speaks of essential ends in the plural. He submits, however, that there is a hierarchy of such ends, with a single ‘final end’ at the top. It remains unclear what he has in mind when he speaks of ‘essential ends’. By contrast, we are in a better position to determine what the final end is. This end concerns ‘the entire vocation [Bestimmung] of human beings and the philosophy of it is called moral philosophy’ (A840/B868). Therefore, the final end of reason is a topic for moral philosophy. If we then consider that for Kant philosophy according to the worldly concept is a doctrine of wisdom (9:24) and that the chief object of a doctrine of wisdom is the highest good (5:108), we can submit that the final end of reason is the highest good (see 29:948) (for different readings of the highest good, see Yovel 1980: 48–78; Reath 1988; Höwing 2016b; and Marwede 2018). This is confirmed by Kant at the very end of the Architectonic of Pure Reason, where he claims that the ‘chief end’ of reason is ‘general happiness’ (A851/B879). Adherence to the moral law is the only way in which happiness can be ‘general’, that is, such that the happiness of one person can coexist with the happiness of others. Therefore, general happiness is dependent on virtue and simply another name for the highest good (for a view of this sort, see Marwede 2018). What would be distinctive about a metaphysics construed according to the worldly concept of philosophy? First of all, since the final end of human reason is a topic for moral philosophy, this concept provides a principle for stressing the priority of the practical part of metaphysics over its theoretical part. But the moral metaphysics at stake here is not one that only identifies a priori principles of action (arguably, this is what the metaphysics of morals according to the school concept would do). Rather, the need for guidance regarding how to attain virtue (as one component of the highest good) is made an essential topic of morals. In this sense, the moral part of metaphysics does not simply have to determine what is morally good. It also needs to instruct us concerning how this is achievable, which involves directing us with regard to how we can bring our partly sensible nature into conformity with morality. Moreover, the worldly concept of

h

d i

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

32 Metaphysics as a Science and the Role of the Critique of Pure Reason philosophy has consequences for the theoretical part of metaphysics. The latter must explicitly consider whether the highest good is possible given theoretical cognitions of reason.18 There seems to be a further important consequence for metaphysics when it is construed according to the worldly concept of philosophy. In this sense, metaphysics can never be seen as an accomplished task. The reason for this is simple: there cannot be a single answer to the question how we can bring ourselves to comply with morality. An answer to this question depends essentially on our individual nature as partly sensible beings. The answer would provide practical guidance on how we can make our individual nature, which is partly sensible, conform to the prescriptions of our rational nature, which is a universally valid principle in us.19 This clarifies why, for Kant, a doctrine of wisdom essentially involves a person’s capacity to be an example of perfect morality for other human beings. There is no other way of showing how a person’s partly sensible nature can be brought to conform to the general requirements of reason beyond providing an example of how this has been possible for someone in particular.20 Being able to do so means being a philosopher in the eminent sense, by showing the ‘infallible effect [of wisdom] in his own person as an example (in mastery of himself and the unquestioned interest that he preeminently takes in the general good), which the ancients also required for deserving that honourable title’ (5:109). The problem here is that no one, according to Kant, has the right to claim to have reached perfect morality: ‘It would be very boastful to call oneself a philosopher in this sense and to pretend to have equalled the archetype, which lies only in the idea’ (A839/867). Therefore, metaphysics according to the worldly concept of philosophy can only remain an ideal we cannot actually achieve, a ‘focus imaginarius’.21 18 19

20

21

h

d i

0 0

See Chapter 7. This issue is closely linked to the problem addressed by the Doctrine of Method of the second Critique, which determines ‘the way in which one can make objectively practical reason subjectively practical as well’ (5:151). On practical doctrines of method, see Bacin (2002) and Bacin (2010). This account of the role of metaphysics according to the worldly concept seems to violate the description according to which metaphysics only contains a priori cognitions. Kant is consistent in attributing this role to it, however. This marks a difference between metaphysics according to the worldly concept and other sciences. While, as we saw, it is generally possible for a science to completely realize its idea, the idea of metaphysics according to this concept cannot ever be fully realized. This clearly puts metaphysics according to the worldly concept at a disadvantage with respect to the other sciences. Nevertheless, metaphysics according to this concept has a privileged position in another sense. Without making reference to the highest good, scientists cannot appreciate the relevance of their discoveries to our moral vocation.

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

The Worldly Concept of Philosophy

33

To summarize, while metaphysics according to the school concept simply aims at organizing a priori discursive cognitions in a system with a hierarchical structure, metaphysics according to the worldly concept organizes these same cognitions with the chief purpose of answering the question whether and how the highest good is possible. b. The architectonic unity of metaphysics. Can metaphysics achieve architectonic unity when it is constructed according to the school concept of philosophy? Is it in a better position if interpreted according to the worldly concept? I submit that for Kant, it is only the worldly concept of philosophy that provides an idea of metaphysics according to which this science can attain architectonic unity. I base my claim on two considerations. First, Kant’s contentions regarding the school concept of philosophy suggest that the latter can only give technical unity to metaphysics. Second, there are reasons to think that metaphysics according to the school concept cannot even achieve systematic coherence. I will discuss these two issues in turn. In a brief footnote in the Architectonic, Kant contends: ‘A worldly concept here means one that concerns that which necessarily interests everyone. I determine the aim of a science in accordance with school concepts if it is regarded only as one of the skills for certain arbitrary ends’ (A839/B867n). A science built according to the school concept of philosophy is thus one whose aim is determined in connection with contingent and arbitrary ends. Now recall that it is in technical unity that the unity we achieve is obtained through reference to contingent and arbitrary ends (A833/B861). So, it seems that the school concept of philosophy can only provide technical unity to metaphysics. But if architectonic unity is necessary for science, this means that metaphysics cannot become a science according to the school concept. Read in these terms, Kant maintains that metaphysics construed according to the school concept is capable of achieving a degree of systematic coherence, but it is nonetheless unable to provide architectonic unity to a body of metaphysical cognitions since it fails to grasp the proper idea of the whole at the basis of the unity of this body. The point would then be that references to ‘essential ends’ of reason and to the moral vocation of human beings are elements that essentially belong to the proper idea of metaphysics. There is a complication here, though, for Kant describes not only metaphysics according to the school concept but also mathematics, logic and physics as sciences that are aimed at contingent and arbitrary ends: ‘The artist of reason or philodox, as Socrates calls him, is one who equips

h

d i

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

34 Metaphysics as a Science and the Role of the Critique of Pure Reason reason for any end one might wish. The mathematician is an artist of reason, likewise the logician and the physicist’ (24:798; see also A839/ B867). Therefore, if metaphysics according to the school concept can achieve only technical unity because it is directed towards contingent and arbitrary ends, why should mathematics, logic and physics be able to reach architectonic unity? If the latter sciences can achieve architectonic unity without any reference to necessary ‘essential ends’ of reason, then this must be possible for metaphysics according to the school concept as well. Alternatively, one must maintain that reference to ‘essential ends’ of reason is necessary for architectonic unity as such. But since essential ends of reason are only explicitly considered in metaphysics according to the worldly concept, this would mean that mathematics, logic and physics can in fact only achieve architectonic unity and become sciences if they are placed in a complete system of all the sciences, with metaphysics according to the worldly concept at the top. Kant sometimes suggests such a picture (9:48–9; 28:533), but this is inconsistent with his claim in the Prolegomena (4:279–80) and the B-Introduction to the first Critique (Bx–xviii) that mathematics and physics are sciences and can be a model for metaphysics. I have suggested elsewhere (Gava 2014) that the most plausible way out of this dilemma is to submit that Kant provides two different accounts of architectonic unity in the Architectonic without clearly distinguishing between them. In the most general sense, architectonic unity simply requires that a body of cognitions be organized according to the idea of the whole of a science that correctly describes the parts–whole relationships among its doctrines. This is the sense of architectonic unity discussed in Section 1.b. In a second sense, architectonic unity is described as requiring that reason ‘provides the ends a priori’ (A833/B861) and that the cognitions are organized around ‘a single supreme and inner end’ (A833/B861; see also A847/B875). These claims strongly recall the hierarchy of essential ends of reason that form the main topic of metaphysics according to the worldly concept. The best way to read this second account of architectonic unity is to say that it describes the conditions for reaching architectonic unity that only apply to metaphysics. That is, it is only for metaphysics that a consideration of the essential ends of reason is necessary for reaching architectonic unity (Gava 2014: 390–1). This is the case because essential ends of reason figure prominently, at least according to Kant, in the proper idea of metaphysics that grounds the architectonic unity of this science. If one considers that the aim of the Architectonic is to show how metaphysics can achieve architectonic unity, it is perhaps not so surprising that Kant does

h

d i

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

The Worldly Concept of Philosophy

35

not clearly distinguish between the ‘general’ sense of architectonic unity and the sense that only applies to metaphysics.22 Distinguishing between these two senses of architectonic unity allows us to understand why, on the one hand, metaphysics according to the school concept, mathematics, logic and physics all neglect necessary essential ends of reason in their system-building, while, on the other, it is only for metaphysics according to the school concept that this neglect undermines the possibility of achieving architectonic unity. One might object that my distinction between two senses of architectonic unity is based on systematic considerations only. These might well be plausible, but in the absence of unequivocal textual support there are no compelling reasons to adopt this distinction. Therefore, so the argument might continue, there are no compelling reasons to silence the claim that ‘if mathematics, logic and physics can reach architectonic unity without a reference to essential ends, this must also be possible for metaphysics according to the school concept’. I believe that the systematic considerations in favour of the distinction between two kinds of architectonic unity are strong,23 but let us assume that this objection is successful.24 Even granting this, there are independent reasons to maintain that metaphysics according to the school concept of philosophy is not in a position to attain architectonic unity. In fact, it seems that metaphysics according to the school concept cannot achieve any kind of unity, be it architectonic or technical, since it is unable to 22

23

24

h

d i

Note that a different sense of non-arbitrariness is connected to each of these accounts of architectonic unity. For the general sense of architectonic unity, lack of arbitrariness simply means that the unity of a body of cognitions is achieved in connection to the idea of the whole that correctly describes the body of cognitions belonging to a science and the parts–whole relationships pertaining to it. In this sense, the unity is not achieved according to organizing principles that are external to the science. By contrast, for the architectonic unity of metaphysics, lack of arbitrariness means that necessary and essential ends of reason are a fundamental tool for organizing the doctrines belonging to the science. In this way, these doctrines cannot be organized according to contingent ends that are external to the proper interest of metaphysics. For example, one can use the same line of argument to establish precisely the opposite: that ‘if metaphysics according to the school concept cannot reach architectonic unity without a reference to essential ends, this must also be impossible for mathematics, logic and physics’. The distinction between two senses of architectonic unity avoids the need to subscribe to either one of the positions at the extremes (‘all these sciences must be able to achieve architectonic unity’ or ‘all these sciences can only achieve technical unity’) in favour of a much more plausible position in the ‘middle’ (‘mathematics, logic and physics can achieve architectonic unity, while metaphysics according to the school concept can only achieve technical unity’). In fact, there is further textual support for the claim that metaphysics according to the school concept can only attain technical unity for Kant. In the Lectures on Philosophical Encyclopaedia, he claims that Christian Wolff, a paradigm of a philosopher working according to the school concept, cannot be considered a real philosopher because he was not an ‘architectonic philosopher’ but rather a ‘great artist’ (29:8).

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

36 Metaphysics as a Science and the Role of the Critique of Pure Reason achieve systematic coherence. We have seen that Kant names the philosopher who proceeds according to the school concept a ‘philodox’ (24:799; 9:24). In the Jäsche Logic, he then contends: The philosopher must thus be able to determine: 1. the sources of human knowledge, 2. the extent of the possible and profitable use of all knowledge, and finally, 3. the limits of reason. The last is the most necessary but also the hardest, yet the philodox does not bother himself about it. (9:25)

Kant makes a similar contention in the Vienna Logic, where he claims that it is when ‘the philosopher is to cognize the connection of all cognitions of reason with the final ends’ that ‘he must determine: 1. the sources of human knowledge, 2. the beginning of the use [of those cognitions], 3. their limits’ (24:799). Therefore, it is only the philosopher who follows the worldly concept of philosophy who is interested in determining the limits of human rational cognition. The ‘philodox’, the philosopher who follows the school concept of philosophy, does not care about these limits. This has fatal consequences for the project of a metaphysics according to the school concept, however. As is well known, according to Kant, it is only by respecting these limits that we can avoid the ‘conflict of reason with itself ’ addressed in the Transcendental Dialectic.25 The philodox, then, cannot achieve any unity with regard to metaphysical cognitions because he is not even able to attain systematic coherence (see also Gava 2014: 388–90). Putting together the two points I have made, we can conclude that metaphysics according to the school concept can at best reach technical unity, since it neglects an essential feature of the proper idea of the whole of metaphysics. Since metaphysics according to the school concept does not care about setting limits to rational cognition, however, it cannot even achieve technical unity, because it cannot avoid falling into the ‘conflict of reason with itself ’.26 25

26

h

d i

0 0

A consequence of these claims is that pursuing metaphysics according to the worldly concept is instrumental to the possibility of a critique of pure reason. One might object here that Kant sometimes suggests that the critique of pure reason should proceed ‘scholastically’ (see for example Bxliii). However, I do not think that these passages are in conflict with the idea that the critique of pure reason should incorporate a ‘worldly’ perspective. In these passages, a scholastic approach is characterized as rigorous and as contrasting with ‘popular’ philosophy, and therefore not as contrasting with a ‘worldly’ approach. I thank Colin McQuillan for raising this issue. My interpretation here differs from those of Tonelli (1994: 272) and Ypi (2011: 144), who maintain that metaphysics according to the school concept has technical unity, La Rocca (2003: 221), who claims that it has architectonic unity, and Ferrarin (2015: 81), who claims that its unity is neither technical nor architectonic. In my view, metaphysics according to the school concept has no unity at all.

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

The Worldly Concept of Philosophy

37

But why does Kant believe that the philodox does not care about the limits of human reason? We can make sense of this claim by acknowledging that the philodox also has an interest in the highest good, but he does not consciously or explicitly recognize it. We have already established that the ‘final end’ towards which the worldly concept of philosophy is directed is the highest good. Now for Kant, the highest good is possible only on the assumption that we are immortal and that God exists (A810–11/ B838–9). Freedom is a further condition of the highest good (insofar as it is a condition for acting morally). Kant believes that the fact that freedom, the immortality of the soul and God are conditions for the realization of the highest good explains why philosophers have always tried so hard to provide speculative proof that we are free, that we are immortal and that God exists: ‘If, then, these three cardinal propositions [the will is free, the soul is immortal and God exists] are not at all necessary for our knowing, and yet are insistently recommended to us by our reason, their importance must really concern only the practical’ (A799–800/B827–8). The highest good, as a necessary rational end, generates a practical interest in the conditions of its realization. Since past philosophers failed to acknowledge the practical interest animating their theoretical efforts, they made themselves unable to set limits to those efforts. Kant’s point can be put as follows: when we do not consciously realize that we are pursuing an investigation for the purposes of a specific interest of ours, we are not in a position to question the legitimacy of our investigation. That is to say, we are not in a position to ask ourselves whether that investigation can obtain results. This is why the philodox cannot set limits to rational cognition: since he does not explicitly consider the practical interest that drives his investigation, he cannot really consider whether his theoretical investigations regarding freedom, the soul and God have a chance of achieving results. To conclude, I submit that it is only when metaphysics is understood according to the worldly concept of philosophy that it is possible to regard it as capable of architectonic unity.27 Since essential ends of reason must 27

h

d i

Both Goy (2007) and Mensch (2013) argue that Kant’s account of architectonic unity is essential to understanding the fundamental problem of the Critique of Pure Reason and how this problem relates to the possibility that metaphysics might become a science. However, they fail to realize how the problem of the architectonic unity of metaphysics is closely related to the issue of the possibility of the highest good. Goy (2007: Ch. 1) reads the problem of the architectonic structure of the Critique as basically identical with the possibility of showing that the system of rational sciences sketched in the Architectonic chapter can be consistently derived from the system of a priori principles identified by Kant in the Transcendental Aesthetic and the Transcendental Analytic. By contrast, Mensch (2013: Ch. 7) argues that the Architectonic should be used to read the Transcendental Deduction anew. The Deduction’s determination of the ‘birthplace’ of the categories is nothing less than the determination of the ‘idea’ that can give unity to metaphysics. It should be clear that

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

38 Metaphysics as a Science and the Role of the Critique of Pure Reason figure prominently in the idea that gives architectonic unity to metaphysics, metaphysics according to the school concept can at best reach technical unity. But even conceding that metaphysics according to the school concept could in principle attain architectonic unity, there are reasons to stress that it cannot in fact reach any kind of unity because it cannot achieve systematic coherence. Notice here that while the orientation towards the highest good and wisdom is a condition for attaining architectonic unity in metaphysics, this does not mean that metaphysics can pick wisdom and the highest good as its only topic. If the ‘philodox’ is the philosopher who tries to establish a system of a priori rational cognitions, neglecting the highest good as a necessary end of reason, the ‘misologist’ is the person who pretends to directly pursue wisdom without establishing a scientific system of a priori rational cognitions: ‘He who hates science but loves wisdom all the more is called a misologist’ (9:26; see also 24:800). It is only as the pursuit of both science and wisdom together that metaphysics can obtain architectonic unity: ‘For science has an inner, true worth only as organ of wisdom. As such, however, it is also indispensable for it, so that one may well maintain that wisdom without science is a silhouette of a perfection to which we shall never attain’ (9:26; see also 5:108–9; Refl. 1656, 16:68).28 c. Metaphysics and the unity of all the sciences. An explicit reference to the highest good (and the essential ends of reason that are subordinated to it) is not only necessary for the architectonic unity of metaphysics. Kant believes that taking into account human reason’s pursuit of the highest good is the only way in which we can see all the sciences as parts of a unitary system of human knowledge moving congruently. Since it is metaphysics according to the worldly concept of philosophy that considers the

28

h

d i

0 0

my interpretation of the problem of architectonic unity for metaphysics essentially departs from their readings. The importance of the highest good for the idea that gives unity to metaphysics is emphasized by Ypi (2011; see also Ypi 2021). However, Ypi argues that the central role of the highest good for the system of metaphysics constitutes a problem for Kant, since it undermines certain achievements of the Critique (2011: 148; see also Ypi 2021). Ypi seems to believe that if the highest good must function as a unifying principle for the sciences, Kant needs a theoretical guarantee that this is possible, where this would contradict his arguments regarding the impossibility of theoretical cognition of the supersensible. I do not think that using the highest good as a unifying principle for metaphysics commits Kant to such a view. This relates to what Kant says on the cyclops as an ‘egoist of science’ (Refl. 903, 15:395). It is only the philosopher who can give the cyclops his second eye back and provide science with a moral orientation. See Ferrarin (2015: 60–5).

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

The Worldly Concept of Philosophy

39

relation of cognitions to the final end of reason, it is metaphysics that can provide unity to the sciences. In the Jäsche Logic, Kant contends that ‘philosophy [but we can read metaphysics here, for reasons I have given above] is the only science that knows how to provide for us this inner satisfaction [which can only be given by wisdom], for it closes, as it were, the scientific circle, and only through it do the sciences attain order and connection’ (9:26; see also 24:800). How should we read this claim? I have argued that sciences like mathematics, logic and physics can achieve architectonic unity and attain the status of sciences without bothering about wisdom and the highest good. However, once we have a set of different sciences that enjoy architectonic unity of their own, we must ask how we should regard the relationships that these sciences have to one another. More precisely, we must determine these relationships in a systematic and non-arbitrary way. How can this task be accomplished? Kant assumes that it is only in relation to some end that we can see different sciences as constituting a unity: the cognitions pertaining to each science can be organized in terms of how important they are in view of the end in question.29 If the end is arbitrary and contingent, however, the order it provides will only be one among many possible orders of the same sort. The unity in this case would only be technical. By contrast, organizing all the sciences around a necessary final end of reason, like the highest good, provides a principle of organization that is not arbitrary and is shared by every human being.

3 The Critique of Pure Reason and Metaphysics According to the Worldly Concept of Philosophy If what I have argued thus far is correct, then the Critique of Pure Reason, in its attempt to establish that metaphysics can become a science, must also show that metaphysics according to the worldly concept can achieve architectonic unity. This implies putting questions about the possibility of the highest good and the rationality of striving for it at the centre of the Critique. Since it is the possibility of metaphysics according to the worldly concept that the Critique must establish, its task seems paradoxical, for Kant claims that the idea of metaphysics according to the worldly concept must 29

h

d i

Kant does not argue for this assumption. Why can’t the unity of the different sciences be determined simply by looking at relationships among their doctrines? Why should we assume that the relationships between these doctrines themselves are unable to provide a unity that is not arbitrary?

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

40 Metaphysics as a Science and the Role of the Critique of Pure Reason remain a ‘focus imaginarius’: it cannot ever be completely realized as an existing system of doctrines. Does this not imply that metaphysics cannot ever become a science? Is the very question of the Critique undermined from the start? Not necessarily. Metaphysics can certainly establish the validity of discursive a priori cognitions and arrange them in a systematic way. It can also show that the highest good is a necessary idea of reason based on a priori moral principles and that theoretical cognitions of reason do not conflict with the possibility of the highest good. In sum, it can show that we can rationally pursue the highest good. These are all aspects of the ‘idea’ of metaphysics according to the worldly concept that can be realized in an existing body of cognitions. What cannot be realized is a doctrine of wisdom intended as practical guidance regarding how we, as individuals, can perfectly conform to morality. Kant’s point here is that even if this part of metaphysics cannot ever be completed, it is important not to lose sight of its fundamentality and to direct those parts of metaphysics that can be established towards furthering the continuous pursuit of virtue.

h

d i

0 0

9 8 009

2 2 003

bli h d

li

b C

b id

i

i

chapter 2

The Critique of Pure Reason as the Doctrine of Method of Metaphysics

In a passage in both editions of the Critique of Pure Reason, Kant describes what he wants to accomplish there as a ‘doctrine of method’ (A82–3/B108–9). In the B-Preface, he adds that the Critique ‘is a treatise on the method, not a system of the science [of metaphysics] itself ’ (Bxxii). These statements are often completely overlooked.1 After all, they appear to be merely a couple of odd passages that deviate from the most common descriptions of the Critique as an analysis of the faculty of reason and a propaedeutic to metaphysics. However, they do constitute a challenge. First, it is not at all clear what it means to say that the Critique is a ‘doctrine’ or ‘treatise’ of method. Moreover, this renders unclear why a part of the Critique is called the Transcendental Doctrine of Method if it is also the whole Critique that provides said doctrine. In this chapter, I argue that Kant’s claim that the Critique of Pure Reason is a doctrine of method must be read literally and indicates that the Critique of Pure Reason is the doctrine of method of metaphysics. Moreover, this has important consequences for how we understand the nature of the critical investigation, for reading the Critique of Pure Reason as a doctrine of method rules out the most natural, and arguably most common, interpretation of the claim that the critique of pure reason is a ‘propaedeutic’ to metaphysics. This cannot be taken as the view that the critique determines how metaphysics is possible before any part of metaphysics is established. To the contrary, the critique of pure reason rests on the establishment of at least some doctrines of metaphysics: in particular, some doctrines of transcendental philosophy.2 1 2

There are exceptions, however: see Barale (1988), Tonelli (1994), La Rocca (2003), Ferrarin (2015) and McQuillan (2016). De Boer (2020: Ch. 3) makes a distinction between ‘transcendental philosophy’ and ‘transcendental critique’ that comes close to my distinction between transcendental philosophy and the critique of pure reason. I take it that my account determines in more detail how they relate to one another.

41

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

42 Metaphysics as a Science and the Role of the Critique of Pure Reason This chapter is structured as follows. In Section 1, I clarify what a doctrine of method is for Kant. I distinguish between the doctrine of method of general logic and the doctrines of method of particular sciences. The former only provides general instructions regarding how to further the systematicity of the sciences. By contrast, the task of doctrines of method of particular sciences is twofold: first, they must provide objector ­cognition-dependent methodological rules regarding how to proceed in these sciences; second, they must show that a science in fact possesses ‘architectonic unity’. In a second step, I consider the Transcendental Doctrine of Method and argue that it is the particular doctrine of method of metaphysics. As such, it provides object- or cognition-dependent methodological rules regarding how metaphysics should proceed and shows that metaphysics can attain architectonic unity. In Section 3, I explain why the Critique of Pure Reason as a whole can be considered the doctrine of method of metaphysics even though it is only its second main part that explicitly bears that title. Since cognitions belonging to a science must already be established in order for a doctrine of method to perform its task, the Transcendental Doctrine of Method requires that at least some doctrinal parts of metaphysics be established in the Transcendental Doctrine of Elements. These doctrinal parts of metaphysics belong to ‘transcendental philosophy’. This has two main consequences. First, it makes clear in what sense providing a doctrine of method is the purpose of the Critique as a whole. Second, it highlights that two disciplines are established in the Critique of Pure Reason: transcendental philosophy, as one doctrinal part of metaphysics, and the critique of pure reason, as that discipline within the Critique that achieves the latter’s aim as the doctrine of method of metaphysics. In Section 4, I consider how my characterization of the Critique relates to more common descriptions that portray it as an analysis of our cognitive faculties or a propaedeutic to metaphysics. Finally, in Section 5, I submit that if there are two disciplines that are established within the Critique, clarifying its ‘method’, understood as a procedure of argument, requires clarifying the methods of both.

1  What is a ‘Doctrine of Method’ for Kant? Kant develops the distinction between ‘doctrines of elements’ and ‘doctrines of method’ as a reaction to and critique of the distinction between ‘theoretical’ and ‘practical’ logic, a distinction one can find, for example, in George Friedrich Meier. Accordingly, at the beginning of the Transcendental Doctrine of Method, he submits that what that part of

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

The Critique as the Doctrine of Method

43

the Critique accomplishes is related to what, ‘under the name of practical logic, […] the schools sought but accomplished only badly’ (A708/ B736). Since Kant developed his distinction between doctrines of elements and doctrines of method as a response to this traditional classification of the parts of logic, the former cannot be seen as a scholastic legacy, as Kemp Smith observes (Kemp Smith 1918: 563). Why did Kant think that the distinction between theoretical and practical logic had to be rejected? What are the reasons in support of his alternative proposal? In this section, I will first analyse Kant’s critique of the concept of practical logic. I will then introduce his account of what a doctrine of method consists of. a. Kant’s critique of the idea of practical logic. Both in the Vernunftlehre (Meier 1752b: § 13) and in the Auszug aus der Vernunftlehre (Meier 1752a: § 7), Meier distinguishes between theoretical logic (theoretische Vernunftlehre, lehrende Vernunftlehre, logica theoretica, logica docens) and practical logic (practische Vernunftlehre, ausübende Vernunftlehre, logica practica, logica utens). The former collects rules of learned cognition (gelehrte Erkentniß ) and learned exposition (gelehrter Vortrag) in general, avoiding consideration of the specific cognition or exposition to which these rules are to be applied. Practical logic, by contrast, determines its rules by taking into account the specific characteristics of particular cognitions and the requirements for their adequate exposition. The distinguishing feature of practical logic is thus that it does not limit itself to formulating abstract rules but applies its rules to particular cases. Kant advances two main objections against the idea of a practical logic (see also Tonelli 1994; Capozzi 2002: 257–64; La Rocca 2003: Ch. 6). First, he maintains that the concept is a ‘contradictio in adjecto’ (9:17; see also 24:700). Kant’s point is that logic, in the strict sense of the term, is by nature universal and does not take into account how its rules apply to this or that object of cognition. Rather, its rules are purely formal and apply to any cognition whatsoever. If practical logic must on the one hand advance a claim to universality insofar as it is a division of logic and on the other take into account particular forms of cognition insofar as it is the practical part of logic, there would seem to be something intrinsically odd about this science (see also 24:507–8).3 3

h

d i

Notice that when Kant introduces his distinction between pure and applied general logic in the Critique of Pure Reason, he emphasizes that applied logic is general because it ‘concerns the use of the understanding without regard to the difference of objects’ (A53/B77). Pure general logic is

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

44 Metaphysics as a Science and the Role of the Critique of Pure Reason The second objection against practical logic is that a practical general logic would necessarily be dialectical (9:16; Refl. 1579, 16:23; Refl. 1612, 16:36; Refl. 1613, 16:36; Refl. 1629, 16:47, 50). We can reconstruct this objection as follows: since practical logic, in developing its rules, does not abstract from the specific characteristics of different objects of cognition, its rules are not purely formal and analytic. Rather, they go over and above purely analytic principles and introduce synthetic ones. This means that once these rules are accepted, they can be used to directly derive certain propositions regarding objects.4 They are ‘object-’ or ‘cognition-dependent’ rules in the sense that they apply to a certain set of objects or cognitions precisely because they acquire their validity on the basis of specific features of those objects or cognitions. Since these rules are dependent on the features of objects or cognitions in this way, once you assume them, certain object- or cognition-related propositions immediately follow. After all, the rules already imply that objects or cognitions have the features in question. The problem is that there is no guarantee that an object- or cognition-dependent rule of reasoning, such as those of practical logic, will be universal. Rather, Kant thinks that the validity of an object- or cognition-dependent rule of reasoning is always restricted to a particular set of objects or cognitions, namely those on which the rule depends. As we have seen, however, practical logic pretends to be universal or general. This means that it treats object- or cognition-dependent rules of reasoning that are only locally valid as if they were universal. In this way, practical logic becomes dialectical, since by using its object- or cognition-dependent rules of reasoning outside of their proper domain of application, it makes claims that are not actually warranted and are based on only seemingly valid principles.5 To conclude, Kant believes that if practical logic is to become a coherent notion, it must either keep its claim to universality and give up its

4 5

h

d i

0 0

distinguished from applied general logic because the former singles out formal rules of reasoning and ‘abstract[s] from all empirical conditions under which our understanding is exercised’ (A52–3/ B77), while the latter considers how empirical features of our psychology constrain reasoning. Kant submits that the empirical focus of applied logic prevents it from being a science properly speaking (A53–4/B78; 9:18). Nonetheless, it is important to appreciate that this denial of scientific standing to applied logic is not due to a lack of generality, since Kant thinks that applied logic can in principle be general. Huaping Lu-Adler emphasizes that Kant directly criticizes Wolff ’s account of practical logic in his description of applied logic (2018: 113). Given this argument, it seems that Kant presupposes that general logic is analytic. For an argument against this view, see Tolley (2012). Lu-Adler also emphasizes that Kant believes that ‘practical’ logic becomes dialectical when it is used as a universally valid tool to obtain cognitions (2018: 165).

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

The Critique as the Doctrine of Method

45

appeal to object- or cognition-related rules of reasoning or keep its appeal to object- or cognition-related rules of reasoning and give up its claim to universality. In the former case, it would continue to be part of general logic, but it could not be used as an instrument for extending cognition. Its content would then be much thinner than was customary for practical logic. In the latter case, it would no longer be part of general logic. It would instead become what Kant calls a ‘particular’ logic, of which there can be many. As such, it would contain ‘the rules for correctly thinking about a certain kind of objects’ and would be ‘the organon of this or that science’ (A52/B76). Therefore, in order to rightfully be an organon and contain rules for the extension of cognition,6 practical logic needs to abandon its aspirations for universality. b. Kant on the concept of a doctrine of method. Kant’s development of the concept of a doctrine of method results from the two alternative directions I have just illustrated regarding how practical logic could become a coherent notion. Kant develops the concept of a doctrine of method in two distinct ways. On the one hand, he envisions a doctrine of method of general logic. Being general, this doctrine of method cannot contain object- or cognition-dependent rules. On the other hand, Kant submits that each science needs its own doctrine of method. The latter will contain object- or cognition-dependent rules that are only valid for the objects or cognitions that belong to the science in question.7 Let us see how Kant characterizes these two disciplines. Kant introduces the doctrine of method of general logic as part of a twofold division, the other member of which is a doctrine of elements (9:17; 24:700; 24:794; Refl. 1629, 16:50; Refl. 1703, 16:80). He proposes this partitioning as an alternative to the traditional division between theoretical and practical logic. He sometimes labels the doctrine of elements and the doctrine of method of logic ‘dogmatic’ and ‘technical’ logic, respectively (9:18; 24:508; 24:794). Kant submits that the doctrine of elements of logic ‘contains rules in general’ (24:700) and that it ‘has for its content the elements and conditions of the perfection of a cognition’ (9:139). 6

7

h

d i

Carboncini and Finster (1982) and Tonelli (1994: Ch. 1) identify different senses in which Kant uses the term ‘organon’. Some of these do not necessarily imply an extension of cognition. When referring to an ‘organon’ in this chapter, I will always understand it as implying an extension of cognition. In the context of his practical philosophy, Kant uses the expression ‘doctrine of method’ in yet another sense that is not relevant to the present chapter. For an account of the ‘practical’ doctrines of method, see Bacin (2002) and (2010).

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

46 Metaphysics as a Science and the Role of the Critique of Pure Reason Therefore, it can be associated with what Kant calls the ‘analytic’ part of logic and is ‘the negative touchstone of truth’ because it ‘analyses the entire formal business of the understanding and reason into its elements, and presents these as principles of all logical assessment of our cognition’ (A60/B84). As such, it is a ‘canon’ and not an ‘organon’ of cognition (A61/B85). Clearly, the distinctive feature of the doctrine of method of general logic cannot be that its rules are object- or cognition-dependent (9:18). By contrast, Kant stresses that it should be a discipline that ‘has to deal with the form of a science in general, or with the ways of acting so as to connect the manifold of cognition in a science’ (9:139; see also 24:700; Refl. 1629, 16:50; Refl. 1703, 16:88; Refl. 3332, 16:783). In other words, the role of the doctrine of method of logic is to provide instructions regarding how one can attain systematicity in a discipline (24:700; 24:779; 24:795). Kant does not say much on what these general instructions amount to. In the Jäsche Logic, he submits that the doctrine of method of logic must ‘expound the way for us to attain the perfection of cognition’ (9:139). He then clarifies that this perfection is obtained through ‘distinctness, thoroughness, and systematic ordering into the whole of a science’ (9:140) but only identifies the conditions for attaining distinctness in definitions, expositions and logical divisions (§ 98, 9:140). The rest of the Jäsche Logic is then devoted to these conditions for distinctness. The Jäsche Logic is certainly a problematic text,8 but it is not implausible to think that Kant believed that a general doctrine of method must contain a doctrine of definition and exposition, since this topic forms part of the first book of the Transcendental Doctrine of Method of the Critique of Pure Reason (see A727–32/B755–60). As I will argue below, the latter can be seen as a particular doctrine of method. As such, it provides specific instructions regarding how to define concepts in metaphysics. If this is right, it seems plausible to think that the doctrine of method of logic should instead provide general instructions for how to define concepts in any science. Although these remarks concerning the content of the doctrine of method of general logic are tentative, it is clear that Kant thinks that its instructions regarding how to build a science must be minimal. It is only when we consider the specific objects and cognitions of a particular science that we can obtain a doctrine of method that is informative. Accordingly, in the introductory section of the Transcendental Doctrine of Method, Kant stresses that a general doctrine of method cannot ‘do 8

h

d i

0 0

It was not written by Kant, and there is no evidence that Kant actually reviewed it.

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

The Critique as the Doctrine of Method

47

more than expound titles for possible methods and technical expressions that are used in regard to that which is systematic in all sorts of sciences, which first makes the novice familiar with names the significance and use of which he will only learn in the future’ (A708/B736). Again, it is unclear what ‘titles of possible methods’ and ‘technical expressions’ are in this context. It is nonetheless unquestionable that the contribution of a general doctrine of method to the development of different sciences will always be minimal. Unfortunately, Kant does not go into detail regarding what particular doctrines of method should do either. This may be due to the fact that each science has its own particular doctrine of method, ‘for in each we must have a form of thought’ (9:18; see also 24:508). Therefore, we cannot say in general what these different doctrines of method should deliver. What Kant does say is the following: first, particular doctrines of method can legitimately be considered the ‘organon’ of this or that science (24:795; Refl. 1579, 16:23). This means that the principles or rules they identify can be used to extend cognition in that science. Second, however, one should keep in mind that the rules or principles of particular doctrines of method are object- or cognition-dependent: ‘[t]he organon of the sciences can only be found in accordance with acquaintance with their nature, object, and sources of cognition’ (Refl. 1612, 16:36). This means that we must already be in possession of cognitions belonging to a science in order to single out these rules or principles. Moreover, since these rules or principles are valid only in connection to the objects or cognitions from which they are derived, they imply certain truths regarding those very objects or cognitions. It is in this sense that these rules or principles can ‘extend’ cognition. Therefore, the fact that a particular doctrine of method is an ‘organon’ cannot mean that we can first develop it and then magically obtain new cognitions thanks to the rules or principles that form that organon. Rather, third, the particular doctrine of method of a science must necessarily come at the end of its development. It ‘can only appear at the end of a science, because only then am I acquainted with the nature of the science’ (24:795; see also Heschel Logic, 488, Eng. tr. 416). But if the particular doctrine of method of each science comes at the end of its development, why do we need it after all? It does not seem to be needed for the extension of cognition in that science, since it rests on the already attained establishment of its cognitions. In the Hechsel Logic, Kant submits that method ‘is the unity of a whole of cognition according to principles’ (Heschel Logic, 488, Eng. tr. 416; see also 24:682). Method satisfies a particular need in science, which Kant puts as follows: ‘[t]here must be a certain connection

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

48 Metaphysics as a Science and the Role of the Critique of Pure Reason of cognitions in that they constitute a whole[;] there must be a rule of unity’ (Heschel Logic, 489, Eng. tr. 416). We have seen that a central aspect of the general doctrine of method is that it provides general instructions regarding how a body of cognitions can become a system and then a science (24:700; 24:779; 24:795). Assuming that the ‘method’ Kant refers to in the Hechsel Logic is the topic of a particular doctrine of method, we can then submit, fourth, that while the general doctrine of method provides general instructions on how to attain systematicity in a science, the task of particular doctrines of method is to show how a specific body of cognitions can hang together and constitute a unitary system. In this way, particular doctrines of method do not contribute, at least directly, to the establishment of particular cognitions within a science. Rather, they show how the unity of a science is possible. To use the words of Chapter 1, they show that a body of cognitions possesses architectonic unity. Since the latter is essential to science, their task is not a secondary achievement. It should be clear from what I have said thus far that Kant has a remarkably original understanding of the problem of method. First of all, at the level of both the general doctrine of method and the particular doctrines of method, he links the problem to the systematicity of the sciences. While it is certainly not uncommon to consider issues of systematicity under the label ‘method’, what is original is Kant’s positioning of the systematicity of the sciences as the main, if not the only, problem of a logical methodology. More importantly, one of Kant’s highly innovative claims is that it is only by considering the specific objects and cognitions of a particular science that we can determine informative methodological rules (see also Ferrarin 2015: Ch. 1 and Ferrarin 2019). We have seen that the object- or cognitiondependent character of these principles is inescapably connected to their local validity. In this way, Kant directly opposes those projects that assume that we can develop a general methodology of the sciences that is valid for any scientific discipline. Historical examples of this approach include Descartes’s Discours de la methode, which identifies universally valid rules of reasoning independently of any object of thought, and Christian Wolff ’s attempt to identify a general methodology of science (alternatively called the mathematical, scientific or philosophical method). It is true that Kant thinks that a general doctrine of method is a legitimate discipline, but its teachings are so minimal that at this level nothing is obtained that is useful for the development of a science. Perhaps the most original claim that Kant makes regarding particular doctrines of method is that they must come at the end of a science. It is natural to expect that a doctrine of method should precede a science and

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

The Critique as the Doctrine of Method

49

provide instructions regarding how to avoid error and arrive at the truth more quickly. Kant turns this traditional understanding of method upside down and submits that content must come at the beginning. There is much that is interesting and original in Kant’s concept of a doctrine of method, especially if one considers his observations on the particular doctrines of method of specific sciences. Nevertheless, it is difficult to get a sense of the precise purpose of these doctrines of method on the basis of his general remarks on these particular doctrines. The question is therefore how particular doctrines of method can perform their specific task. How can they show that a particular body of cognitions has architectonic unity? One of the tasks of the next section is to answer this question. In this respect, I will first show that the Transcendental Doctrine of Method can be considered the particular doctrine of method of metaphysics. In this respect, it is the culmination of the project of the Critique of Pure Reason, understood as such a doctrine. Second, I will provide an account of how this particular doctrine of method can perform its task.

2  The Transcendental Doctrine of Method of the First Critique Now that we have a better understanding of what a doctrine of method is for Kant, it is time to provide an account of the Transcendental Doctrine of Method of the Critique of Pure Reason. The first question to ask is whether the Transcendental Doctrine of Method is a general or a particular doctrine of method. If it is the latter, for which discipline, precisely, is it a doctrine of method? That the Transcendental Doctrine of Method cannot be a general doctrine of method is clear from its opening pages. Kant submits that its work on the ‘plan’ of the system of pure reason rests on the ‘building materials’ displayed in the Transcendental Doctrine of Elements (A707/ B735). Assuming, plausibly, that these materials are the a priori cognitions of reason singled out in the first part of the book, we can conclude that the rules and principles discussed in the Transcendental Doctrine of Method are object- or cognition-dependent and, as such, cannot be general. This is confirmed by what Kant says at the end of the introductory section of the Transcendental Doctrine of Method. There, he stresses that the latter accomplishes ‘in a transcendental respect’ what, ‘under the name of a practical logic, […] the schools sought but accomplished only badly’ (A708/B736). Kant claims that practical logic could not accomplish what the Transcendental Doctrine of Method must accomplish because the former advanced a claim to general validity ‘with regard to the use

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

50 Metaphysics as a Science and the Role of the Critique of Pure Reason of the understanding in general’ (A708/B736). Advancing such a claim is inadequate for the Transcendental Doctrine of Method, however, since it must identify rules that are distinctive of pure cognitions of reason. This perspective is different from that of general logic, which ‘is not limited to any particular kind of cognition of the understanding (e.g., not to the pure cognition of the understanding) nor to certain objects’ (A708/B736, my emphasis).9 The rules and principles discussed by the Transcendental Doctrine of Method are therefore cognition- and object-dependent. If the Transcendental Doctrine of Method is not a general doctrine of method, it must be the doctrine of method of a particular science. But which one? Kant’s claims regarding the object of the Transcendental Doctrine of Method are ambiguous. He says that it is concerned with the ‘system of pure reason’, but at one point he seems to understand the latter as only comprising the metaphysics of nature (at A707/B735, he only mentions the edifice of ‘speculative’ reason), while at another he seems to mean the complete system of pure reason (A707–8/B735–6), formed by the whole of metaphysics and pure mathematics. Moreover, given Kant’s general description of particular doctrines of method, one would expect all parts of the science for which the Transcendental Doctrine of Method provides a plan to be presented in the Transcendental Doctrine of Elements. After all, Kant says that particular doctrines of method come at the end of a science, when we are already acquainted with its cognitions. The Transcendental Doctrine of Elements, however, certainly does not establish either mathematical cognitions or the parts of metaphysics that belong to the special metaphysics of corporeal nature and to the metaphysics of morals. Therefore, if the ‘system of pure reason’ sketched in the Transcendental Doctrine of Method had to rest on the cognitions presented in the Transcendental Doctrine of Elements, it could not be either the whole metaphysics of nature or the complete system of pure reason. Kant’s explicit remarks on the object of the Transcendental Doctrine of Method are therefore not helpful. Nevertheless, looking at the actual contents of this part of the Critique clearly shows that the science for which the Transcendental Doctrine of Method provides a plan is metaphysics. First of all, even though the Architectonic of Pure Reason mentions mathematical cognitions (A837/B865), empirical philosophy (A840/B868) and 9

h

d i

0 0

Let me note here that saying that the Transcendental Doctrine of Method is a particular doctrine of method is compatible with a recent reading of Kant’s transcendental logic which argues that the latter is part of general logic (see Tolley 2012). While I maintain that the Transcendental Doctrine of Method is the particular doctrine of method of metaphysics, I remain neutral on the status of transcendental logic, which can plausibly be seen as a doctrinal part of metaphysics.

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

The Critique as the Doctrine of Method

51

empirical psychology (A848–9/B876–7), it is clear that the main task of the Architectonic is to sketch the system of metaphysics. Kant dedicates six full pages to this task in both original editions (A841–7/B869–75). He even offers an explanation of why, for the time being, empirical psychology could be given a place in metaphysics, even though it does not belong there (A848–9/B876–7). These are strong indications that the Transcendental Doctrine of Method is the particular doctrine of method of metaphysics, the specific task of which is to show that the body of cognitions belonging to metaphysics possesses architectonic unity. This initial impression is confirmed when we survey the contents of the three other chapters of the Transcendental Doctrine of Method, namely the Discipline of Pure Reason, the Canon of Pure Reason and the History of Pure Reason. My purpose is to show that they either provide objector cognition-dependent rules for obtaining metaphysical cognitions or are directly concerned with the question of the ‘architectonic unity’ of metaphysics. a. The Discipline of Pure Reason. Let us start with the Discipline of Pure Reason. The first section, the Discipline of Pure Reason in Dogmatic Use, is mainly dedicated to a comparison of the methodologies of philosophy and mathematics, with the chief purpose of obtaining negative rules and prescriptions regarding how philosophy should not proceed in its investigations. Kant bases his various remarks on his claim that while mathematics is rational cognition from the construction of concepts, philosophy is rational cognition from concepts (A713–14/B741–2). This fundamental difference between mathematical and philosophical cognitions gives rise to further differences between these disciplines in terms of how they must use definitions, axioms and demonstrations. Accordingly, it is clear that the first section of the Discipline provides cognition-dependent rules for how, in philosophy, we should proceed in clarifying our concepts, identifying fundamental synthetic a priori principles and structuring our proofs. The rules are cognition-dependent because the fact that we should proceed in this way and not another in philosophy rests on the discursive and non-intuitive character of the cognitions we deal with in philosophy. Kant follows a similar approach in the fourth section of the Discipline, which is dedicated to transcendental proofs. This is a special kind of proof within philosophy. They establish the validity of discursive synthetic a priori propositions (A782–3/B810–11) and do not assume anything empirical in their premises (A614/B642). Kant identifies three rules that apply to these proofs. First, ‘to attempt no transcendental proofs without

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

52 Metaphysics as a Science and the Role of the Critique of Pure Reason having first considered whence one can justifiably derive the principles on which one intends to build and with what right one can expect success in inferences from them’ (A786/B814);10 second, ‘for each transcendental proposition only a single proof can be found’ (A787/B815); and third, transcendental proofs ‘must never be apagogic but always ostensive’ (A789/B817). Apagogic proofs are indirect proofs that derive the truth of a proposition from the falsity of its opposite (9:71), whereas an ostensive proof is a direct proof ‘which is combined with the conviction of truth and simultaneously with insight into its sources’ (A789/B817). Similar to the first section of the Discipline, Kant characterizes these rules in direct opposition to the procedures that are appropriate in mathematics. Therefore, the section on transcendental proofs also provides rules that are cognition dependent. I do not have space here to consider the second section of the Discipline in any detail. Let me just say that it can also be seen as providing object- or cognition-dependent rules that are relevant to metaphysics. The section concerns the ‘polemical’ use of reason and submits that metaphysicians should not engage in this polemic, which Kant understands as ‘the defence of its [that is, reason’s] propositions against dogmatic denials of them’ (A739/767). Kant’s admonition not to engage in this polemic is based on a consideration of our inability to cognize objects of pure reason and can thus be considered an object- or ­cognition-dependent rule. Something similar can be said of the third section of the Discipline, dedicated to the use of hypotheses. The general question of the section is whether metaphysics can use hypotheses to make provisional claims about those objects that, lying beyond the limits of possible experience, are outside the grasp of our cognition. For Kant, a hypothesis is a kind of opinion, that is, a particular form of doxastic attitude in which we take a proposition to be true, recognizing that the grounds we have in support of our ‘taking-to-be-true’ are not conclusive (see A822/B850). A hypothesis is an opinion that provides a ‘ground of explanation’ of some phenomenon (A770/B798). When taken as a ground, the hypothesis must thus be ‘sufficient to explain other cognitions as consequences’ (9:85). In the Discipline, Kant identifies two general rules for the use of hypotheses.11 First and very generally, the hypothesis that is assumed must 10 11

h

d i

0 0

As Kant later clarifies, this condition expresses the need for a transcendental deduction of the principle that grounds the argument of a transcendental proof (A787/B815). In the Jäsche Logic, by contrast, one finds three rules (9:85–6).

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

The Critique as the Doctrine of Method

53

be able to provide a ground of explanation of some phenomenon (A770/ B798). Second, it must be sufficient to provide a complete explanation of the given phenomenon. It should not need additional hypotheses to provide a complete explanation (A774/B802). Given these general rules, Kant then derives specific reasons that prevent the use of hypotheses in the case of objects of pure reason. First, given that we in principle do not have any grasp of what an object of pure reason would be like, this can never contribute to the explanation of a given phenomenon (A772/ B800). Second, Kant submits that even if we grant that the assumption of an ‘unlimited perfect cause’ or of the ‘simple self-sufficiency of the human soul’ could have explanatory force with respect to the ‘purposiveness, order, and greatness that is found in the world’ (in the former case) or the appearances in the soul (in the latter case), one must always assume additional hypotheses to explain deviations, in the empirical realm, from the ideal cases the original hypotheses are supposedly able to explain (A774–5/B802–3). In addition to these two reasons for avoiding the use of hypotheses about objects of pure reason, Kant also provides a general consideration: since pure theoretical reason is constituted by a priori cognitions, its claims cannot have the nature of opinion but must be certain (A781/B809; see also A822–3/B850–1). Kant concedes, however, that hypotheses can be used not to extend our cognition in the field of objects of pure reason but to show that a dogmatic proof of a proposition that we oppose is invalid (A776–82/B804–10). Hypotheses can thus be used ‘polemically’ not to establish the truth of a claim, as in the polemical use of reason, but simply to show that another claim is invalid. Therefore, in the case of hypotheses, Kant derives a partly object-dependent, partly cognition-dependent rule. This rule says that we can make a very limited use of hypotheses in the field of objects of pure reason. In sum, all the rules identified in the four sections of the Discipline are object- or cognition-dependent. Are these rules meant for metaphysics? This is not immediately clear, since Kant often speaks of philosophy or philosophical cognitions in general. Nevertheless, various factors indicate that it is metaphysical investigations that are at stake. First, in the introductory remarks of the Discipline, Kant states that it is in its ‘transcendental use’, that is, the use directed towards objects of pure reason, that reason is in need of this discipline (A711/B739). Second, the section on proofs is dedicated to transcendental proofs, which do not assume anything empirical. Third, the conflicts discussed in the polemic of reason are clearly conflicts between theoretical claims of reason that concern objects of special metaphysics.

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

54 Metaphysics as a Science and the Role of the Critique of Pure Reason It is thus plausible to view the rules of the Discipline as intended for metaphysical investigations. Can these rules be seen as contributing to the chief purpose of a doctrine of method, that is, showing that a body of cognition is capable of architectonic unity? The rules of the discipline can all be seen as instrumental to preventing metaphysics from falling into seemingly unsolvable conflicts. Therefore, they certainly help to establish a minimal requirement for attributing architectonic unity, namely the requirement that the body of cognitions be coherent.12 b. The Canon of Pure Reason. In the opening paragraphs, Kant clarifies that once the Discipline has established that there is no legitimate theoretical use of pure reason, the task of the Canon of Pure Reason is to investigate whether there is a legitimate practical use of pure reason (A796–7/B824–5). The first section, On the Ultimate End of the Pure Use of our Reason, submits that metaphysical disputes concerning objects of pure reason have always concerned three objects: the freedom of the will, the immortality of the soul and the existence of God (A798/B826). Kant then argues that these objects are not particularly important from the standpoint of our theoretical reason. This is because even if we were able to establish freedom, immortality and the existence of God, this would not contribute in any way to our understanding of nature, since we would have no way of grasping how a free will, an immortal soul or God could relate to the sensible world of our experience (A798–9/B826–7). Since these objects are not particularly important for our theoretical reason but nonetheless appear to enjoy a special status in metaphysics, Kant suggests that they must be relevant from a practical point of view (A799/B827). In the rest of the first section of the Canon, Kant shifts his focus. The issue is no longer locating the source of our interest in freedom, immortality and God, but rather determining what sort of practical ground can justify claims regarding those objects. Since it is only moral commands that are not empirically conditioned and are ‘products of pure reason’ (A800/B828), they are the only appropriate practical grounds on which to advance claims regarding objects of pure reason. The implicit premise here is that claims regarding objects of pure reason need an a priori grounding and cannot be based on imperatives that rest on contingent features of the human being. 12

h

d i

0 0

See Chance (2015) for an account of the Discipline that is fundamentally in agreement with mine.

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

The Critique as the Doctrine of Method

55

The second section of the Canon, On the Ideal of the Highest Good, as a Determining Ground of the Ultimate End of Pure Reason, takes up the task left open by the first. The latter was only able to establish that if there are practical grounds for advancing claims about objects of pure reason, these must be based on moral commands. The second section shows that there are indeed practical grounds of this sort on which to advance claims about the existence of God and the immortality of the soul.13 Roughly, the section argues that there is a fundamental idea of reason, the highest good, which necessarily emerges from moral obligation. Since it would be impossible to regard the highest good as realizable if we did not assume that God exists and that there is a future life (A811/B839), the highest good gives us valid practical grounds on which to make these assumptions.14 After the second section of the Canon shows that there are indeed practical grounds, based on moral obligation, for advancing claims concerning the existence of God and the immortality of the soul, the third and last section, On Having an Opinion, Knowing, and Believing,15 clarifies the nature of the attitude we are justified in having when advancing these claims, which Kant calls belief (Glaube) and contrasts to knowledge (Wissen) and opinion (Meinung). This clarification is needed if we are to confirm that our commitment to the propositions stating the existence of God and the immortality of the soul is indeed justified and if we are to show that it does not conflict with the agnosticism regarding objects of pure reason that Kant defends in the Dialectic and the Discipline.16 Given this brief survey of the Canon, its chief aim appears to be that of establishing, first, that we can rationally hold that some conditions for the realization of the highest good obtain and, second, that this assumption does not contradict some of the negative results of the Discipline, according to which we cannot advance ‘dogmatic’ claims regarding objects of pure reason, where dogmatic claims are claims to cognition. If what I argued in Chapter 1 is correct, namely that the ‘worldly concept’ 13

14 15

16

h

d i

Kant does not provide an argument for freedom. He submits that ‘practical’ freedom, which is the ability to give imperatives to ourselves (A800/B828), is the only freedom we should care about from a practical standpoint (A803–4/B831–2). Since this freedom is definitive of the practical in general, we are not in need of a special argument to establish practical freedom. We simply must assume it once we take the practical standpoint and accept that we are able to give imperatives to ourselves. See Chapter 7 for a more detailed reading of Kant’s strategy in the Canon. For different accounts of this section, see Stevenson (2003), Chignell (2007), Willaschek (2010), Willaschek (2016), Pasternack (2011), Fonnesu (2015), Höwing (2016a), Gava and Willaschek (forthcoming) and Gava (2019b). See Chapter 7.

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

56 Metaphysics as a Science and the Role of the Critique of Pure Reason of philosophy is the only candidate ‘idea’ that can provide architectonic unity to metaphysics for Kant, the Canon clearly contributes to establishing that metaphysics can achieve this unity. It shows that it is rational for us to commit ourselves to the actuality of some of the conditions for the realizability of the highest good, where this commitment can find a place in a coherent system of metaphysics, comprising both its theoretical and its practical parts. c. The History of Pure Reason. In what sense can a History of Pure Reason contribute to establishing the architectonic unity of metaphysics? Put briefly, its task is to show that, once we have reached the critical standpoint in metaphysics, we can account for the history of this science and locate, in ‘the nature of pure reason’, the grounds for the positions that animated the disputes among philosophers (A852/B880; see also 20:341, 343). In such a way, the metaphysical positions held by historical figures are not regarded as contingently dependent on the views of this or that individual but are rather traced back to sources in our reason that explain why different people arrived at certain conclusions regarding objects of metaphysics, even though those conclusions were wrong. How can a history of reason understood in such terms contribute to the establishment of a system of metaphysics? Or better: why should the capacity to explain why metaphysics has a given history contribute to the establishment of the correct system of metaphysics as a science? Kant is not clear on this issue. There is, however, a way to make sense of his approach. A metaphysical system that, on the basis of its doctrines, can provide an explanation of why we held other views and can also account for the sources of the errors contained in those views has an obvious advantage over metaphysical systems that establish a body of interconnected claims and simply argue that other historically held views were wrong. Not only does the first type of system have greater explanatory power in comparison to the second, but it can also incorporate those historical views into the system, since it can provide a ‘diagnosis’ that explains why they arose. This survey of the Transcendental Doctrine of Method confirms that it can coherently be interpreted as the doctrine of method of metaphysics. In Section 1, we saw that the task of a doctrine of method is twofold: first, it must provide object- or cognition-dependent methodological rules for how to proceed in a given science; second, it must show that a science in fact possesses ‘architectonic unity’. Given our survey of the four chapters that make up the Transcendental Doctrine of Method, it

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

The Critique as the Doctrine of Method

57

should be clear that they all contribute to establishing that metaphysics is capable of architectonic unity. Moreover, it is the specific task of the Discipline to provide object- or cognition-dependent rules that help metaphysics to attain coherence, which is a minimal requirement of architectonic unity.

3 The Critique of Pure Reason as the Doctrine of Method of Metaphysics The previous section established that the Transcendental Doctrine of Method is the particular doctrine of method of metaphysics. We can now return to the puzzle that we identified at the beginning of this chapter: in what sense can Kant also claim that the Critique of Pure Reason as a whole is a ‘doctrine of method’? One way to make sense of this claim has been pursued by Giorgio Tonelli, who argued that the Critique of Pure Reason should be considered the particular logic of metaphysics (1994: 81) and that, as such, it is a methodology of this science ‘in a broad sense’ (1994: 92). Tonelli tries to solve the puzzle by blurring the boundaries between the Transcendental Doctrine of Elements and the Transcendental Doctrine of Method (1994: 91–2), on the one hand, and transcendental philosophy and the critique of pure reason – as the two disciplines that are established within the Critique (1994: 97) – on the other. He argues that since it is impossible to make a clear-cut distinction between the two main parts of the Critique and between transcendental philosophy and the critique of pure reason, there is nothing preventing us from treating the whole Critique as the particular logic of metaphysics, with a focus on methodology (or as the particular doctrine of method of this science, in my terminology). Even if there is much in Tonelli’s book that I admire, I don’t think that his strategy helps us to explain in what sense the whole Critique of Pure Reason can be considered the doctrine of method of metaphysics. In contrast to his view, I suggest that it is only by maintaining a distinction between the Transcendental Doctrine of Elements and the Transcendental Doctrine of Method, on the one hand, and transcendental philosophy and the critique of pure reason, on the other, that we can explain why the whole Critique of Pure Reason should be considered the particular doctrine of method of metaphysics.17 17

h

d i

Colin McQuillan (2017a) has advanced a criticism of Tonelli’s approach that complements mine. He argues that Tonelli wrongly characterizes the Critique of Pure Reason as only belonging to the tradition of ‘modern’ logics, which makes it impossible to understand what role the Transcendental Aesthetic plays in it.

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

58 Metaphysics as a Science and the Role of the Critique of Pure Reason In my view, there is one fundamental reason to maintain these distinctions. As we saw in Section 1, particular doctrines of method of specific sciences come at the end of the development of a science, once the doctrines that belong to it have already been established. Only in this way can they have cognitions and objects to refer to so as to determine their object- and cognition-dependent rules. Similarly, it is only when there is already a body of cognitions, the order of which is not clear, that a doctrine of method can show that it constitutes an architectonic whole. If we take into consideration the Transcendental Doctrine of Method, however, it clearly comes before metaphysics has been established as a science. We must then ask how it can establish its object- or cognition-dependent rules and how it can show that metaphysics can achieve architectonic unity. I submit that the Transcendental Doctrine of Method does not proceed in the total absence of doctrinal parts of metaphysics. Rather, it rests on the establishment of certain doctrines that belong to metaphysics. More specifically, it rests on the establishment of certain doctrines of transcendental philosophy. In the Introduction, Kant says explicitly that some parts of transcendental philosophy are already established in the Critique of Pure Reason: To the critique of pure reason there accordingly belongs everything that constitutes transcendental philosophy, and it is the complete idea of transcendental philosophy, but is not yet this science itself, since it goes only so far in the analysis as is requisite for the complete estimation of synthetic a priori cognition. (A14/B28)

Kant’s point here is that the Critique of Pure Reason establishes only those parts of transcendental philosophy that are instrumental to achieving its purpose, that is, showing that metaphysics can attain architectonic unity and become a science. The Critique of Pure Reason only focuses on those parts of transcendental philosophy that establish its synthetic a priori principles (A12/B25; A13–14/B27-8). How can we explain this focus? Taking the perspective of the Transcendental Doctrine of Method, one can say that it is exactly these synthetic a priori principles that need a ‘discipline’. They need the establishment of object- and cognition-dependent rules that can prevent metaphysics from falling into a conflict between equally unfounded metaphysical positions. In other words, if the Critique of Pure Reason is to set limits on the use of synthetic a priori principles belonging to transcendental philosophy, we first need to establish these principles as proper parts of transcendental philosophy.

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

The Critique as the Doctrine of Method

59

Therefore, the Transcendental Doctrine of Method cannot perform its task in the absence of doctrinal parts of metaphysics that belong to transcendental philosophy. This clarifies, first, why the Transcendental Doctrine of Method needs a Transcendental Doctrine of Elements to perform its task and, second, why it is the whole Critique of Pure Reason that must be considered a doctrine of method. The Transcendental Doctrine of Method needs the Transcendental Doctrine of Elements because it is in the latter that parts of transcendental philosophy are established. It is the whole Critique that must be considered the doctrine of method of metaphysics because the introduction of doctrinal parts of metaphysics in the Transcendental Doctrine of Elements has the sole purpose of enabling the Transcendental Doctrine of Method to do its job. This way of describing the relationship between the Transcendental Doctrine of Method and the project of the Critique of Pure Reason as a whole provides a rationale for Kant’s claim that the whole Critique is the doctrine of method of metaphysics. However, one might object that there is no reason to take this claim seriously. After all, there is only one passage where Kant describes the Critique in this way (A82–3/B108–9). As a response to this objection, I wish to emphasize that there are both systematic and textual reasons to take the claim literally. Let me begin with the systematic reason. Understanding the Critique as the doctrine of method of metaphysics fits well with Kant’s claim that the Critique must establish whether metaphysics can become a science. This question is central both to the B-Preface and to the B-Introduction, and it again takes centre stage at the beginning of the Transcendental Doctrine of Method. But establishing that a body of cognitions is a science is exactly one of the main tasks of a doctrine of method. Now the textual reasons. The passage where Kant identifies the Critique with a ‘treatise on method’ provides a description of its task that hints at the problem of determining whether metaphysics can attain ‘architectonic unity’. The context is Kant’s discussion of his ‘Copernican revolution’. He first suggests that the Critique shows that metaphysics must accomplish a revolution in the way of thinking on the model of geometry and physics. He then writes that the Critique, as a ‘treatise on the method’, ‘catalogues the entire outline of the science of metaphysics, both in respect of its boundaries and in respect of its entire internal structure’ (Bxxii). It seems appropriate to claim that determining these boundaries and this internal structure is only possible from the standpoint of the ‘architectonic unity’ of metaphysics.

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

60 Metaphysics as a Science and the Role of the Critique of Pure Reason Finally, Kant’s claim at the beginning of the Transcendental Doctrine of Method that its work on the ‘plan’ required a previous investigation into the ‘materials’ with which we can build (A707/B735) matches my contention that Kant’s work on the doctrine of method of metaphysics requires the establishment of at least some parts of transcendental philosophy. Note that we find fundamentally the same picture in the passage where Kant equates the whole Critique to a doctrine of method (A82–3/ B108–9). Kant submits that within the ‘system of pure reason’ it will be necessary to provide full definitions of the categories. When it comes to the Critique as a ‘doctrine of method’, however, it is not necessary to provide these definitions. Rather, it is sufficient to carry out the analysis up to the point that is required for achieving its aim. But this agrees with my view that the Critique carries out inquiries belonging to transcendental philosophy in order to fulfil its role as the doctrine of method of metaphysics.

4  The Different Characterizations of the Critique There may be systematic and textual reasons to take Kant’s claim that the Critique is the doctrine of method of metaphysics seriously. However, if this characterization were completely at odds with more common descriptions of it, it would be unclear why it should be preferred over the others. Accordingly, I will take into consideration the claim that the Critique is a propaedeutic to metaphysics and the claim that it provides an analysis of our faculties of cognition. a. The Critique as a propaedeutic. Kant often describes the Critique as a ‘propaedeutic’ to metaphysics (A841/B869). At first glance, this seems to conflict with my description of the Critique as the doctrine of method of metaphysics. How can a propaedeutic of metaphysics rest on one of its parts? First of all, let me note that this is the same Kant who submits that the particular logic of a science can both rest on already established doctrines belonging to that science and be a propaedeutic to that very science: In the schools the latter [the organon of this or that science, which is equivalent to its particular logic] is often stuck before the sciences as their propaedeutic, though in the course of human reason they are certainly the latest to be reached, once the science is already long complete, and requires only the final touch for its improvement and perfection. For one must already know the objects rather well if one will offer the rules for how a science of them is to be brought about. (A52/B76–7)

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

The Critique as the Doctrine of Method

61

Of course, the best way to make sense of this passage is to understand the ‘propaedeutic’ in question as a propaedeutic not for the establishment of a science but rather for the exposition of an already established science.18 That is to say, the particular logic of a science is first obtained after the establishment of that science and is later used as a propaedeutic for its exposition. We must concede that the Critique of Pure Reason cannot be a propaedeutic for the simple exposition of metaphysics, since its task is to enable metaphysics to become a science in the first place. And yet, what is interesting in the passage is that the propaedeutic of a science is closely linked to its particular logic, which, as we saw, Kant equates with its doctrine of method. Therefore, the fact that the Critique of Pure Reason is called a ‘propaedeutic’ can actually be read as evidence that Kant sees it as the particular doctrine of method of metaphysics. Moreover, even if the Critique is a propaedeutic for the establishment and not the exposition of metaphysics, this is not necessarily in conflict with the assumption that some already established parts of metaphysics are contained within it. After all, the task of the Critique of Pure Reason is to show that metaphysics as a whole can achieve architectonic unity. As I already suggested, this does not conflict with the idea that some parts of the science have already been established. Rather, the Critique makes use of one part of metaphysics, namely transcendental philosophy, to show that the whole of metaphysics can achieve architectonic unity. In this sense, it can both rest on certain doctrinal parts of metaphysics and be a propaedeutic to establishing the complete system as a science.19 b. The Critique as faculty analysis. When Kant describes the Critique as a propaedeutic, he often adds that it proceeds through an analysis of our ‘faculties’ of a priori cognition that should determine which cognitions are possible for us.20 Famously, in the A-Preface, Kant writes that it is not ‘a critique of books and systems, but a critique of the faculty

18 19

20

h

d i

For an account of Kant’s distinction between the analytic and the synthetic method of exposition, see Gava (2015). In the Introduction, Kant describes the Critique as a propaedeutic to transcendental philosophy (see A11–12/B25). This indeed suggests that the critique of pure reason, as the discipline established within the Critique, should come before any part of transcendental philosophy is established. Note, however, that the Critique does not rest on the establishment of the complete system of transcendental philosophy. Even with respect to that system, it is possible to submit that the Critique is propaedeutic while still resting on the establishment of parts of transcendental philosophy. On faculty analysis in Kant, see Land (2021).

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

62 Metaphysics as a Science and the Role of the Critique of Pure Reason of reason in general, in respect of all the cognitions after which reason might strive independently of all experience’ (Axii). Therefore, it seems that we have a clear way of distinguishing between the Critique and the system of metaphysics. The former is an investigation of the faculty of reason that determines whether metaphysics is possible by investigating which a priori cognitions are in our power. The latter is the inventory of those cognitions in a system. This description of the Critique seems at odds with describing it as the doctrine of method of metaphysics, and this for at least two reasons. First, given that faculty analysis is distinctive of critique, it seems to conflict with the idea that the Critique could ‘contain’ parts of transcendental philosophy. Second, faculty analysis appears to have little to do with showing that metaphysics can attain architectonic unity. However, it is not clear that Kant regards faculty analysis as distinctive of the critique of pure reason. To appreciate this, it is sufficient to consult the opening sentences of the first book of the Transcendental Analytic, where Kant clarifies what distinguishes his ‘analytic of concepts’ from the procedure of conceptual analysis of traditional metaphysics. Traditional analysis starts by focusing on the content of given concepts and analyses the latter into its component concepts so as to make the first concepts distinct. What the Analytic of Concepts should accomplish, by contrast, is an ‘analysis of the faculty of understanding itself, in order to research the possibility of a priori concepts by seeking them only in the understanding as their birthplace and analysing its pure use in general’ (A65–6/ B90). Kant glosses this remark by saying that this sort of faculty analysis ‘is the proper business of a transcendental philosophy’ (A66/B91). If we take Kant at his word, this means that faculty analysis cannot be distinctive of the critique of pure reason and is at least as fundamental to transcendental philosophy, understood as a doctrinal part of metaphysics. If this is true, it means that while the Critique certainly contains analyses of our faculties,21 this is not what characterizes its project. Rather, the Critique contains these analyses insofar as it contains parts of transcendental philosophy. What characterizes the critique of pure reason, understood as that discipline within the Critique that achieves the latter’s aim as the doctrine of method of metaphysics, is that it provides the doctrine of method of metaphysics. 21

h

d i

0 0

Accordingly, I will show in Chapter 3 that metaphysical deductions, which I take to be an essential part of transcendental philosophy, contribute to establishing Kant’s distinctions between different faculties of cognition.

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

The Critique as the Doctrine of Method

63

5  One or Two Methods? In this Chapter, I have argued that we can make sense of Kant’s claim that the Critique of Pure Reason as a whole is the doctrine of method of metaphysics. Moreover, this characterization is compatible with other descriptions of it. As the doctrine of method of metaphysics, the Critique must, first, show that metaphysics can attain architectonic unity and, second, provide rules that are instrumental to this purpose. Insofar as the Transcendental Doctrine of Method is chiefly dedicated to these tasks, it can be considered the culmination of the critical project within the Critique. In order to fulfil its role, however, the Critique must establish certain parts of metaphysics that belong to transcendental philosophy, which are presented in the Transcendental Doctrine of Elements. As we saw, one implication of this approach to the Critique is that two disciplines are established within it. On the one hand, the Critique establishes some parts of transcendental philosophy. On the other, it establishes the critique of pure reason, understood as that discipline that achieves the Critique’s aim as the doctrine of method of metaphysics. What consequences does this distinction between transcendental philosophy and the critique of pure reason have for my investigation? This brings us to the second sense of ‘method’ that is relevant here. When we speak of doctrines of method, the word ‘method’ is mainly understood as a way of arranging cognitions so that they can attain the status of a science (see 9:139). But reading the Critique as such a doctrine has consequences for how we should approach the procedures of argument that Kant follows within it. Since, according to my approach, transcendental philosophy and the critique of pure reason are two different disciplines, it is likely that they have different methods, understood as such procedures of argument. Therefore, in order to account for the ‘method’ pursued within the Critique of Pure Reason, we actually need to split our investigation into two parts: one directed towards the method of transcendental philosophy and another towards the method of the critique of pure reason. This is what I will do in the next two parts. Before I begin my analysis of the methods of these two disciplines, however, let me state clearly that when I say that the critique of pure reason is the doctrine of method of metaphysics, I do not mean to suggest that Kant’s presentation of this discipline is confined to the Transcendental Doctrine of Method. Rather, as will become clear in the following chapters, there is much that belongs to the critique of pure reason that is already established in the Transcendental Doctrine of

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

64 Metaphysics as a Science and the Role of the Critique of Pure Reason Elements. After Kant establishes these parts of the critique of pure reason in the Transcendental Doctrine of Elements, he draws their consequences in the Transcendental Doctrine of Method. This is especially true of those parts of the critique of pure reason that concern the limits of our capacity to have a priori cognition regarding objects of pure reason. While these parts are first established in the Transcendental Doctrine of Elements, they form the background of many of the negative rules that Kant singles out in the Discipline of Pure Reason.

h

d i

0 0

9 8 009

2 2 004

bli h d

li

b C

b id

i

i

Part II

The Method of Transcendental Philosophy

Introduction to Part II In Chapter 2, I argued that the Critique of Pure Reason should be understood as the doctrine of method of metaphysics. One of the main implications of this approach is that we must actually distinguish between two disciplines established within the Critique. First, the Critique needs to establish at least some parts of transcendental philosophy. Second, it needs to establish the critique of pure reason, understood as that discipline within the Critique that achieves the latter’s aim as the doctrine of method of metaphysics. The purpose of Parts II and III is to clarify the methods of each of these disciplines, respectively, where the word ‘method’ is here understood as a particular procedure of argument. In the present part, the focus will be on the method of transcendental philosophy. Before we provide an account of what is specific to the method of this discipline, it is important to clarify what it has in common with the other disciplines belonging to metaphysics as a whole. In this Introduction to Part II, I will suggest that what characterizes Kant’s approach to metaphysics as a whole is a moderate methodological conservatism regarding a priori concepts and principles. I will distinguish this form of conservatism from what I will call common sense conservatism, which, on my reading, is also characteristic of Kant’s metaphysics. This distinction will help situate my proposal with respect to readings that also attribute a form of conservatism to Kant, such as that offered by Karl Ameriks and, more recently, John Callanan. Furthermore, I will clarify what transcendental philosophy is for Kant, according to the characterization he offers in the Critique of Pure Reason. This is instrumental to outlining the general framework that I will follow in the next two chapters, where I approach the method of transcendental philosophy by distinguishing between metaphysical and transcendental deductions. 65

https://doi.org/10.1017/9781009172127.005 Published online by Cambridge University Press

66

The Method of Transcendental Philosophy

1  Moderate Methodological Conservatism and Common Sense Conservatism In a recent paper, John Callanan (2019) argues that Kant’s project in the Critique of Pure Reason can be described as methodologically conservative. Methodological conservatism is characterized as having ‘the aim of protecting certain first-order commitments against possible revision’ (Callanan 2019: 422).1 Furthermore, Callanan claims that one can be a methodological conservative by either assuming in the premises of a philosophical inquiry that some first-order commitments are unrevisable or by showing in the conclusions of the inquiry that commitments of this sort are not in need of revision (Callanan 2019: 423). In his account, Kant is a methodological conservative in both senses. He assumes as a premise of the Critique of Pure Reason that some first-order commitments are unrevisable, where these commitments are chiefly a priori claims in geometry and Newtonian physics (Callanan 2019: 430–2). But Kant is also conservative in his conclusions, since the latter confirm the validity of Newtonian physics while also preserving a familiar phenomenology and certain concepts of traditional ontology (Callanan 2019: 433). I here borrow Callanan’s terminology, but my characterization of Kant’s methodological conservatism departs in relevant ways from his approach. First, I do not think it is correct to say that Kant’s occasional use of a priori principles from geometry and Newtonian physics as premises is part of his methodological conservatism. Kant certainly sometimes takes it for granted that we have valid synthetic a priori claims in these disciplines and construes arguments on the basis of this assumption. But this assumption cannot be considered part of what I call his methodological conservatism, insofar as it is not their aim to show that a priori principles in geometry and Newtonian physics are valid. Kant simply treats certain principles in geometry and Newtonian physics as obviously true. For this reason, he relies on these principles to show that certain rival philosophical positions are inadequate insofar as they are unable to account for the possibility of those very principles. I take Kant’s assumption of a priori principles in geometry and Newtonian physics to express another sense in which his approach is ‘conservative’. I call this second form of conservatism common sense conservatism. This conservatism concerns not what we aim to establish in a philosophical investigation but rather what we can take for granted at the beginning of such an 1

I here prefer to use ‘conservatism’ rather than ‘conservativism’, which is the term used by Callanan.

https://doi.org/10.1017/9781009172127.005 Published online by Cambridge University Press

Introduction to Part II

67

investigation.2 It says that we can take for granted beliefs that are commonly held as true, unless we have good philosophical reason to challenge them.3 ‘Common sense’ should be construed in a broad sense to include scientific propositions that are generally recognized as true by the practitioners of a science. What I call common sense conservatism bears some similarity to what is often called epistemic conservatism. Roughly, epistemic conservatism is the view that, when we find ourselves holding a belief, we are prima facie justified in maintaining that belief, unless we find a reason to abandon it (for discussion, see Vahid 2004; McCain 2008). The main difference between epistemic conservatism and common sense conservatism is that while for the former it is sufficient that I personally hold a belief for that belief to be prima facie justified, for common sense conservatism a belief must be intersubjectively held as true. Moreover, while epistemic conservatism focuses on cases in which we find ourselves believing something even though we cannot offer a justification for that belief, beliefs that are taken for granted by Kant’s common sense conservatism include beliefs for which we do have justification, as in the case of beliefs regarding properties of space in geometry or regarding natural laws in physics. Kant takes these views for granted in the sense that he does not offer additional philosophical reasons for regarding them as true, at least at the beginning of some of his arguments. Returning to the differences between Callanan’s view and mine, second, Callanan characterizes Kant’s methodological conservatism as applying to different sorts of first-order commitments: a priori claims in geometry and physics, concepts belonging to traditional ontology and our familiar phenomenology (Callanan 2019: 433). By contrast, I take Kant’s methodological conservatism to apply solely to a priori claims. More specifically, its aim is to establish that a priori concepts and principles that govern different aspects of our knowledge and lives have at least some sort of validity. It is not the case, however, that the a priori concepts and principles whose validity is established 2

3

What I call common sense conservatism closely resembles Ameriks’s ‘regressive’ approach to philosophy, which is characterized by assuming that we have some knowledge in order to determine what its ‘conditions’ are. Ameriks makes a distinction between a ‘strongly regressive’ and a ‘modestly regressive’ approach, where the former takes for granted specific synthetic a priori claims and the latter only takes for granted that ‘there is some objectivity to our experience’ (Ameriks 2003: 8). Ameriks argues that Kant’s approach is ‘regressive’ in the second sense. In my view, Kant uses both approaches. For example, the transcendental exposition of space seems ‘strongly’ regressive in that it assumes that geometry contains synthetic a priori propositions. By contrast, the transcendental deduction of the categories appears to be regressive in Ameriks’s preferred sense. It is of course difficult to determine what ‘good philosophical reasons’ are. For example, are traditional sceptical scenarios enough to cast doubt on a belief? Since radical forms of external world scepticism would imply that Newtonian physics cannot be taken as knowledge, Kant’s common sense conservatism does not seem to take sceptical scenarios as sufficient grounds to cast doubt on a belief. For an account of Kant’s take on Humean scepticism, see Ch. 9.

https://doi.org/10.1017/9781009172127.005 Published online by Cambridge University Press

68

The Method of Transcendental Philosophy

in the course of Kant’s inquiry are clearly identifiable from the very beginning. One purpose of the investigation is therefore to bring these a priori concepts and principles completely to light and to trace their origin. Third, the conservatism is only moderate because even though Kant’s general tendency is to establish that the a priori concepts and principles uncovered by metaphysics have validity, it is not the case that all such concepts and principles are established without demanding the revision of our common understanding of them. At least in some cases – paradigmatically, in the case of the transcendental ideas – our common understanding of a priori concepts and principles needs to undergo fundamental revision. Fourth, I take Kant’s moderate methodological conservatism about a priori concepts and principles to apply not only to his project in the Critique of Pure Reason but also to his metaphysics as a whole. In this sense, for example, Kant’s main works in practical philosophy can also be seen as assuming that there must be an a priori principle of morality. The latter is clarified in the course of the investigation, which in turn establishes its validity (see for example 4:389–90). The critique of pure reason, understood as the doctrine of method of metaphysics, is also an expression of Kant’s moderate methodological conservatism about a priori concepts and principles. It will be the task of Part III to clarify the specific sense in which the critique is methodologically conservative. Fifth, while Callanan suggests that Kant is conservative regarding some commitments of traditional ontology, let me emphasize that the conservatism I attribute to Kant is completely compatible with the idea that Kant strongly criticizes traditional metaphysics. As I said, his methodological conservatism aims to establish that a priori concepts and principles that govern different aspects of our knowledge and lives have at least some sort of validity. In doing so, however, Kant can vehemently criticize how these concepts and principles have been used or accounted for in traditional metaphysics. To summarize, Kant’s moderate methodological conservatism regarding a priori concepts and principles starts with the general assumption that we commonly make use of a priori concepts and principles in various aspects of our lives, even though it is not exactly clear at the beginning of the investigation what these concepts and principles are.4 It is the aim of Kant’s metaphysics, first, to clarify what these concepts and principles are – to provide a complete catalogue of them and track their origin – and, second, to show that these concepts and principles have at least some sort of validity. 4

Of course, this is very different from assuming specific a priori principles, belonging to either g­ eometry or physics, as premises in an argument.

https://doi.org/10.1017/9781009172127.005 Published online by Cambridge University Press

Introduction to Part II

69

Validity can mean various things in this context, since the a priori concepts and principles that are at stake are of different kinds. As far as the theoretical part of metaphysics is concerned, validity means, first of all, that a priori concepts and principles involve valid cognition of objects. However, it could also mean that these concepts and principles are conditions for furthering our cognition of objects, even though they do not themselves involve any valid cognition of objects. Although Kant’s methodological conservatism aims to prove the validity of a priori concepts and principles, it can demand the radical revision of our common understanding of them. It is for this reason that the conservatism is only moderate. Finally, Kant’s conservatism is compatible with a fervent critique of traditional metaphysics.

2  Transcendental Philosophy It is now time to see what transcendental philosophy is for Kant and how it fits into his moderate methodological conservatism. Let me clarify from the start that throughout this book I concentrate my attention on Kant’s account of transcendental philosophy within the Critique of Pure Reason.5 Like the other disciplines belonging to metaphysics, transcendental philosophy aims at clarifying and validating a priori concepts and principles. It focuses on a particular subset of these concepts and principles: those that concern the cognition of objects. In the A-Introduction, Kant stresses that transcendental philosophy is the system of ‘our a priori concepts of objects in general’ (A11–12).6 In fact, this characterization seems overly narrow, since it suggests that the only concepts with which transcendental philosophy concerns itself are the categories, which are sometimes described as ‘concepts of objects in general’ (A290/B346). As I will argue in a moment, though, Kant includes within transcendental philosophy concepts that are not directly ‘concepts of objects in general’ but are nonetheless fundamental to explaining how the cognition of objects is possible. This is the case for the concepts of space and time and the transcendental ideas of reason. These are not ‘concepts of objects in general’ but apply to cognition ‘in 5

6

Kant’s views on transcendental philosophy evolve in later works, such as the Critique of the Power of Judgement. For a survey of Kant’s often contradictory claims regarding the role and place of transcendental philosophy in metaphysics, see Ferrarin (2015: Ch. 3); Förster (2018: Part I). This fundamentally agrees with the definition of transcendental philosophy that Kant gives in the Architectonic of Pure Reason, where he writes that transcendental philosophy ‘considers only the understanding and reason itself in a system of all concepts and principles that are related to objects in general’ (A845/B873). However, Kant seems here to exclude from transcendental philosophy a priori representations that belong to sensibility, whereas I think that these should be part of it, for reasons I will touch on presently.

https://doi.org/10.1017/9781009172127.005 Published online by Cambridge University Press

70

The Method of Transcendental Philosophy

general’ in the sense that the role they play is not limited to a particular object or subset of objects.7 For this reason, I submit that it is more accurate to say that transcendental philosophy analyses a priori concepts for the cognition of objects. These a priori concepts often have related principles; for example, the categories give rise to synthetic a priori principles of the possibility of experience, as discussed in the Analytic of Principles. Transcendental philosophy does not investigate all a priori concepts and principles that have to do with our cognition of objects, however. This approach characterizes what Kant calls the metaphysics of nature in general, but transcendental philosophy forms only one part of the latter, since it only focuses on concepts and principles that are pure, that is, those that do not have anything empirical mixed with them. In the Introduction, Kant accordingly stresses that ‘absolutely no concept must enter into it [transcendental philosophy] that contains anything empirical’ (A14/B28). He adds that these concepts must therefore be ‘entirely pure’ (A14/B28).8 This focus on pure concepts and principles distinguishes transcendental philosophy from the special metaphysics of nature. According to Kant, the latter depends on a partly empirical concept, like the concept of ‘matter’ in the special metaphysics of corporeal nature (see 4:469–70). It is only when this partly empirical concept is introduced that special metaphysics derives consequences a priori.9 It is important to note that Kant draws a distinction within pure concepts. He distinguishes between root (Stammbegriffe) and derivative pure concepts of objects (see A14/B27–8). Transcendental philosophy should provide a catalogue of all of these, but root concepts are more fundamental because they lie at the basis of synthetic a priori claims concerning objects (see A14/B27–8).10 This distinction is relevant because Kant claims that the part of transcendental philosophy that is established in the Critique of Pure 7

The concept of time applies indistinctly to all intuitions and that of space to all outer intuitions. Similarly, transcendental ideas, when taken regulatively, will play a fundamental role with respect to empirical cognition in general. 8 Transcendental philosophy is further characterized by the reflexive and second-order nature of its investigations. Accordingly, being a pure concept or principle for the cognition of objects is not sufficient to qualify as a ‘transcendental cognition’. In order to be such and be legitimately included in transcendental philosophy, one must be conscious of the pure a priori origin of the concept or principle in question. Kant remarks that ‘not every a priori cognition must be called transcendental, but only that by means of which we cognize that and how certain representations (intuitions or concepts) are applied entirely a priori, or are possible (i.e., the possibility of cognition or its use a priori)’ (A56/B80–1). 9 For a discussion of the relationship between transcendental philosophy and the special metaphysics of nature, see Ch. 8. 10 Accordingly, Kant suggests that the identification of the derivative concepts will be easy once the difficult task of clarifying the root concepts is completed.

https://doi.org/10.1017/9781009172127.005 Published online by Cambridge University Press

Introduction to Part II

71

Reason only concerns root concepts for the cognition of objects (see A81–2/ B107–8) and, arguably, related principles. What are the root concepts that are identified in the Critique of Pure Reason? Kant explicitly characterizes the categories as root concepts (A81/ B107), so there is no doubt that they count among them. Are the categories the only root concepts that the Critique analyses? Does this mean that the part of transcendental philosophy that the Critique develops is confined within the Transcendental Analytic, and perhaps even more narrowly within the Analytic of Concepts? I don’t think this is the case. Rather, each main section of the Transcendental Doctrine of Elements can be seen as containing an analysis of some of these root concepts. Accordingly, Kant remarks that ‘if sensibility were to contain a priori representations, which constitute the condition under which objects are given to us, it will belong to transcendental philosophy’ (A15/B29–30). Since space and time are a priori representations of this sort, his analysis of them in the Transcendental Aesthetic should be considered part of transcendental philosophy (see 29:802). Evidence that Kant considers space and time root ‘concepts’ is to be found in the Prolegomena.11 There, Kant calls both space and time and the categories ‘elementary’ concepts (Elementarbegriffe) (4:323), which is just a synonym for root concepts.12 It may seem more controversial to consider the Transcendental Dialectic part of transcendental philosophy, at least for two reasons. First, the ‘concepts of reason’ it investigates do not directly involve any valid cognition of objects. However, Kant certainly thinks that the transcendental ideas are pure concepts that naturally arise in connection to our cognition of objects. Moreover, as I will argue in Chapter 4, even though the transcendental ideas do not involve any direct cognition of objects, Kant nonetheless wants to establish that they have a certain validity as concepts that guide the process of perfecting empirical cognition. This suggests that they should also count as root concepts for the cognition of objects and so 11

12

As I will clarify in Chapter 3, I think that we do have ‘concepts’ of space and time. However, since in this case it is the pure intuitions of space and time that are the fundamental representations, it is these representations that should be described as root representations, not the concepts. Therefore, in the case of the concepts of space and time, it is preferable to speak of concepts of root representations. That ‘elementary concept’ is a synonym for ‘root concept’ is clear from the fact that, both in the Critique of Pure Reason and in the Prolegomena, Kant uses the two terms to express the same idea. More specifically, while in the Critique (in discussing the categories) root concepts are contrasted to derivative concepts or predicables (A81–2/B107–8), both the Critique and the Prolegomena (also speaking of the categories) similarly contrast elementary concepts to derivative concepts or predicables (A64/B89; 4:324). Derivative concepts are obtained through the combination of root or elementary concepts.

https://doi.org/10.1017/9781009172127.005 Published online by Cambridge University Press

72

The Method of Transcendental Philosophy

belong to the part of transcendental philosophy that is established within the Critique. Evidence that transcendental ideas are root concepts for Kant is provided in a passage from the Paralogisms of Pure Reason, where Kant distinguishes the pure ‘elements’ of a rational doctrine of the soul from the derivative concepts that can be obtained from a combination of such elements (A344–5/B402–3).13 A second worry might be that if ‘concepts of reason’ have some validity and deserve a place within metaphysics, it is not clear why they are to be located within transcendental philosophy and not the special part of metaphysics. After all, Kant clearly separates transcendental philosophy from special metaphysics (see A845–6/B873–4). The ideas of the soul, the world and God that he analyses in the Dialectic traditionally belonged to rational psychology, rational cosmology and rational theology, respectively, and these were all considered parts of special metaphysics.14 But Kant is explicit that transcendental philosophy should consider ‘concepts of reason’. For example, according to lecture notes from the Mrongovius Metaphysics: ‘Transcendental philosophy is that which considers the pure use of the understanding – and to that belong all the concepts and principles of pure understanding and of pure reason. But some concepts are of immanent use, others of transcendent. Transcendental philosophy considers the entire faculty of pure understanding and of pure reason for cognizing something a priori. The concept of ground and cause and [that] of God belong to transcendental philosophy, but the former is immanent’ (29:765; see also 29:807). Therefore, Kant explicitly includes ‘transcendent’ concepts of reason – and more specifically, the concept of God – within the objects of study of transcendental philosophy. Presumably, he included them within transcendental philosophy because, on the one hand, they are pure concepts that arise independently of empirical content and, on the other, they play an important role in the cognition of objects within possible experience.15 13

14 15

The concepts that constitute these pure elements are expressed in the four sentences ‘The soul is substance’, ‘The soul is simple’, ‘The soul is numerically identical (a unity)’ and ‘The soul is in relation to possible objects in space’. Kant submits that other concepts can be derived from the combination of the ‘root concepts’ expressed in these sentences. Derivative concepts that can be obtained in this way include the concepts of immateriality, incorruptibility, personality, spirituality, interaction with bodies, animality and immortality. The relationship between the pure elements and the derivative concepts resembles the relationship between the categories, as root concepts, and the derivative concepts that result from their combination (A81–2/B107–8). See Klimmek (2005: 123) for a similar reading. I thank Clinton Tolley and Huaping Lu-Adler for raising this issue. It is an open question whether space for these concepts is also left within special metaphysics. I address this issue in Chapter 7.

https://doi.org/10.1017/9781009172127.005 Published online by Cambridge University Press

Introduction to Part II

73

3  Metaphysical and Transcendental Deductions According to my characterization of Kant’s moderate methodological ­conservatism in metaphysics, Kant starts with the assumption that we commonly make use of a priori concepts and principles. With respect to these, metaphysics has two main aims. First, it aims to provide a complete catalogue of these concepts and principles and to track their origin. Second, it aims to establish that these concepts and principles have at least some sort of validity. I have suggested that the Critique of Pure Reason establishes those parts of transcendental philosophy that focus on root concepts for the cognition of objects and related principles. It is therefore to be expected that the Critique contains some investigations that clarify and catalogue root concepts, while also tracking their origin, and some investigations that establish what kind of validity these concepts have. The fundamental hypothesis that structures Part II is that there is a division of labour in each main part of the Transcendental Doctrine of Elements between metaphysical and transcendental deductions. Metaphysical deductions identify and clarify pure root concepts for the cognition of objects and account for their origin. Transcendental deductions reveal the sense in which these concepts are valid. The concepts that are first identified and then legitimated through metaphysical and transcendental deductions are, respectively, the concepts of space and time in the Transcendental Aesthetic, the categories in the Transcendental Analytic and the transcendental ideas of reason in the Transcendental Dialectic. The focus of all of these deductions is concepts. However, analysis of these concepts can be the basis of an investigation into related principles. One could object that a distinction between a metaphysical and a transcendental deduction is only central to the Transcendental Analytic. It has no role in the Transcendental Aesthetic and is only cursorily mentioned in the Transcendental Dialectic. True, Kant does not speak of ‘deductions’ in the Transcendental Aesthetic, and he only introduces a distinction between metaphysical and transcendental expositions of space and time in the B-edition of the Critique. In the introductory remarks on the transcendental deduction of the categories, however, he refers back to a transcendental deduction of space and time in the Transcendental Aesthetic, which allegedly proves the origin and objective validity of these concepts (A87/ B119–20; see also A85–6/B118, where Kant suggests that a transcendental deduction is needed both for the categories and for the concepts of space and time). Regarding the Dialectic, Kant draws a comparison between the metaphysical deduction of the categories and his discussion of the

https://doi.org/10.1017/9781009172127.005 Published online by Cambridge University Press

74

The Method of Transcendental Philosophy

transcendental ideas in the second section of the first book of the Dialectic (see A321/B377; A329/B386). Similarly, Kant mentions a transcendental deduction of the ideas of reason twice in the Appendix to the Dialectic. In one passage, he suggests that a transcendental deduction is never possible for ideas (A663–4/B691–2), even though in the same sentence he concedes that the principles of homogeneity, specification and continuity of nature, which also rest on ideas of reason, ‘have objective but indeterminate validity’ (A663/B691). In a second passage, he provides a transcendental deduction of the transcendental ideas (A669–71/B697–9), although he grants that this deduction proceeds differently than the transcendental deduction of the categories (A669/B697).

4  Main Objectives of Part II The first objective of Part II is to show that it is possible to disentangle those investigations that belong to transcendental philosophy from other parts of the Critique of Pure Reason. The second objective is to shed light on the nature of these investigations, highlighting what they have in common and what is specific to them in each main part of the Transcendental Doctrine of Elements. One feature that they share is that they only establish positive results concerning the clarification, origin and validity of root concepts for the cognition of objects and related principles. It is not their task to set limits on our use of root concepts and related principles. The latter task belongs to the critique of pure reason and will be clarified in Part III of this book. One consequence of this approach is that positive results concerning root concepts for the cognition of objects and related principles are identified in each main part of the Transcendental Doctrine of Elements. It is perhaps more common to view the Transcendental Aesthetic and the Transcendental Analytic as establishing positive results and the Transcendental Dialectic as providing the negative part of Kant’s investigation. By contrast, in my account, each main part of the Transcendental Doctrine of Elements will be shown to involve both positive and negative results. The former establish that we possess root concepts for the cognition of objects (and related principles) and determine what sort of validity they have – these results belong to transcendental philosophy. The latter set limits to the validity of those concepts and principles and belong to the critique of pure reason. This is not to say that the critique of pure reason only establishes negative results. As I will show in Part III, it also has positive results, but these are not achieved in the Transcendental Doctrine of Elements.

https://doi.org/10.1017/9781009172127.005 Published online by Cambridge University Press

chapter 3

Metaphysical Deductions

In the Introduction to Part II, I argued that there is at least one metaphysical deduction in each main part of the Transcendental Doctrine of Elements. Their aim is to catalogue pure root concepts for the cognition of objects and to track their origin. In this chapter, I will analyse these deductions, focusing in particular on their methods. I will consider Kant’s metaphysical deductions of space and time in the Transcendental Aesthetic, his metaphysical deduction of the categories in the Transcendental Analytic and his metaphysical deduction of the transcendental ideas in the Transcendental Dialectic. One natural way to interpret these metaphysical deductions is to say that they start by assuming Kant’s distinction between different faculties of cognition in order to single out which root concepts belong to each of them. Accordingly, the metaphysical deductions in the Aesthetic would aim to catalogue the root concepts of sensibility, the metaphysical deduction in the Analytic would aim to catalogue the root concepts of the understanding in the narrow sense, as the faculty of concepts, and the metaphysical deduction in the Dialectic would aim to catalogue the root concepts of reason in the narrow sense, as the faculty of inference. If this were Kant’s approach, a criticism that is often made against the Critique of Pure Reason in general, namely that it starts by assuming a division between faculties without providing an argument for it (see Falkenstein 1995: Ch. 1), could be extended to the metaphysical deductions in particular. In other words, in devising different metaphysical deductions for sensibility, the understanding and reason, Kant would be guilty of approaching a classification of root concepts by illicitly presupposing a ready-made classification of faculties from the start. In contrast to this approach, my aim is to show that the metaphysical deductions do not simply assume a distinction between faculties. Rather, they contribute to establishing this distinction by identifying the origin of the root concepts they clarify and catalogue. That the metaphysical 75

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

76

The Method of Transcendental Philosophy

deductions contribute to the establishment of Kant’s distinction between sensibility, the understanding and reason is reflected in their respective methods. Since one of the purposes of these deductions is to point out differences between the origins of root concepts, the metaphysical deductions in each of the main sections of the Transcendental Doctrine of Elements use different argumentative strategies that better serve the purpose of ­highlighting the root concepts on which they focus.

1  The Metaphysical Deductions of Space and Time I will start by considering Kant’s metaphysical deductions of the root concepts belonging to sensibility to see how these deductions indeed help to establish sensibility as a separate faculty. It may seem odd to speak of ‘root concepts belonging to sensibility’. How can a concept belong to sensibility, according to Kant? Aren’t space and time intuitions and not concepts for Kant? In fact, when it comes to the concepts of space and time, it is more accurate to say that these are ‘concepts of root representations belonging to sensibility’. Still, as I will argue, the fact that Kant speaks of ‘concepts’ of space and time in the expositions is important, and he should be taken literally. Returning to the metaphysical deductions of these concepts, the first question to ask is where these deductions are located, since, as we saw, Kant does not use the expression ‘metaphysical deduction’ himself in the Aesthetic. The metaphysical expositions of space and time are prima facie the best candidates for playing the role of metaphysical deductions. They are contrasted to transcendental expositions, and this parallels Kant’s distinction between metaphysical and transcendental deductions.1 According to my characterization of metaphysical deductions, to be such a deduction the metaphysical expositions must (a) establish that space and time are root concepts for the cognition of objects and (b) track the origin of these concepts. I have already noted, however, that it is difficult to understand whether the metaphysical expositions deal with space and time as concepts or intuitions (or both). I will remain neutral for the moment on whether concepts or intuitions are at stake and simply ask whether the expositions investigate whether space and time are root representations, while tracking their origin. From Kant’s characterization, it is 1

h

d i

0 0

In contrast to this approach, Vaihinger (1881–1892: Vol. 2, 151–4) and Falkenstein (1995: 394–5 n12) claim that the metaphysical expositions are transcendental deductions.

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

77

clear that a metaphysical exposition can at least show that a ­representation is a priori: I understand by exposition (expositio) the distinct (even if not complete) representation of that which belongs to a concept; but the exposition is metaphysical when it contains that which exhibits the concept as given a priori. (B38)

According to this passage, what makes an exposition metaphysical is its ­capacity to show that a concept is ‘given a priori ’.2 Metaphysical expositions can accordingly show that a representation is a priori. This is still insufficient to establish that such a representation is a ‘root representation’. Root representations are pure representations that, in addition to not containing anything empirical, lie at the basis of synthetic a priori claims to cognition of objects. The metaphysical expositions indeed appear to establish that space and time are pure representations. The second argument in each exposition maintains that space and time can be represented independently of any empirical intuition, even though all intuitions are necessarily in time (A31/B46) and one cannot represent the absence of space (A24/B38–9). Since space and time can be represented independently of empirical intuitions, our representations of space and time do not contain anything empirical and are accordingly pure. Can they also be considered root representations? In the fourth argument in the metaphysical exposition of time, Kant explicitly says that the representation of time serves as the basis of the synthetic proposition that ‘different times cannot be simultaneous’ (A32/B47). The representation of time thus lies at the basis of a synthetic a priori claim. Do the metaphysical expositions of space and time track the ‘origin’ of the root representation of space and time? In the last argument of both metaphysical expositions, Kant explicitly uses the expression ‘original representation’ (B40; A32/B48) to describe space and time, claiming that in the case of space and time this original representation can only be intuitive, not conceptual. This suggests that we have two different ways of representing space and time: one ‘original’, the other derivative. Therefore, by pointing out that space and time are primarily intuitions and not concepts, Kant also wanted to capture the specific origin of these representations. If this is right, it is plausible to regard the metaphysical expositions as the metaphysical deductions of space and time. I will now consider the method that Kant uses in these expositions. I will start by asking whether 2

h

d i

Again, for now, I remain neutral regarding whether the metaphysical expositions have to do with the concepts or intuitions of space and time and so take ‘concept’ to mean representation in general.

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

78

the expositions are expositions of concepts or intuitions and will submit that they start from an analysis of the given concepts of space and time. In contrast to James Messina (2015), however, I will argue that this ‘analysis’ should not be taken as ‘conceptual analysis’ in the strict sense. Thanks to a consideration of the singularity argument for space, it will become apparent that analysis here involves the examination of the use of concepts in synthetic a priori judgements, a process that can in turn track representations that are not conceptual. a. Expositions of concepts or expositions of intuitions? Many commentators have not taken Kant’s claim that the metaphysical expositions are expositions of the concepts of space and time literally, suggesting that concepts in this context could only mean ‘representations’ in general, thus encompassing both concepts and intuitions. Kemp Smith’s Commentary is quite explicit in this regard: The use of the term Begriff in the title of the section [the Metaphysical Exposition of the Concept of Space], and also in this sentence [the first sentence of the first argument of the metaphysical exposition of space], is an instance of the looseness with which Kant employs his terms. It is here synonymous with the term representation (Vorstellung) which covers intuitions as well as general or discursive concepts. Consequently, the contradiction is only verbal, not real, when Kant proceeds to prove that the concept of space is an intuition, not a concept. (Kemp Smith 1918: 99; for a more recent defence of the same view, see Janiak 2016: Section 3.1).

The view, in short, is that since the metaphysical expositions aim to prove that space and time are intuitions and not concepts, they cannot start by considering space and time concepts in the strict sense; otherwise, they would be aiming to establish the absurd claim that a concept is an intuition. This has consequences for how we interpret the method of the expositions. According to Kemp Smith, the method is psychological and consists of a psychological investigation into representations of space and time, which are shown to be a priori and intuitive (see Kemp Smith 1918: 102). Other commentators have emphasized, however, that Kant’s claim that the expositions of space and time concern our concepts of them does not necessarily imply a contradiction. Space and time are certainly first and foremost intuitions, but that does not mean that we cannot form concepts of them from those intuitions. Moreover, philosophy in general has to do with concepts, and so it is no surprise that Kant speaks of concepts of space and time in the Aesthetic (see G. Bird 2006a: 106). The view that

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

79

the metaphysical expositions have to do with concepts has recently been defended by James Messina (2015). Messina focuses on space in particular and claims not only that the metaphysical exposition of space starts by considering the concept of space, but also that the results Kant obtains are only conceptual and rest on the analysis of the concept in question.3 Accordingly, he claims that the kind of knowledge that is at stake in the metaphysical exposition of space is analytic knowledge, based solely on analysis of the concept of space (Messina 2015: 423 n18–19). Messina supports his reading by appealing to the technical meaning of the term ‘exposition’ in Kant. Expositions are the kinds of definitions that are attainable in philosophy, where definitions can be obtained not through the construction of concepts but only through analysis of the constituent marks of a concept. The fact that Kant points out that metaphysical expositions have to do with a concept that is ‘given a priori’ (B38) confirms this view, since ‘given’ concepts are the proper object of philosophical inquiry, which cannot ‘make’ its own concepts, as mathematics does (Messina 2015: 420–2; see A729–32/B757–60 for Kant’s account of philosophical expositions).4 Messina’s reading of the metaphysical exposition of space is interesting for two main reasons. On the one hand, he is clearly interested in drawing the methodological consequences of viewing the metaphysical exposition of space as dedicated to its concept. On the other, he offers an original proposal regarding how to interpret the kind of conceptual analysis that is at stake in the exposition.5 He suggests that Kant’s analytic procedure makes use of ‘modal intuitions (in the contemporary philosophical sense of the term, not Kant’s)’ (Messina 2015: 418). These intuitions involve ‘intuitions […] about what is and is not possible with respect to an instance of a given concept’ (Messina 2015: 429). They imply ‘imagining a concrete object as instantiating the concept’ (Messina 2015: 428). In what follows, I will argue that Messina is right to emphasize that the metaphysical exposition of space is about the concept of space, but I will 3

4 5

h

d i

That the metaphysical expositions proceed by conceptual analysis has been argued by McGoldrick (1985) and Leirfall (2004). Messina only attributes an anticipation of his view to Falkenstein (1995: 148–9). However, while Falkenstein emphasizes that the metaphysical expositions are concerned with the concepts of space and time, he does not seem to hold the view that conceptual analysis is sufficient for accounting for their results. I here focus on Messina’s reading because he is the ­interpreter who most strongly emphasizes the conceptual nature of Kant’s argument. For the distinction between mathematical and philosophical definitions, see De Jong (1995), WolffMetternich (1995) and Gava (2015). A second interesting proposal is that Kant’s account of analysis was influenced by Christian August Crusius (Messina 2015: 434).

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

80

challenge his claim that it only establishes analytic knowledge obtained through the analysis of this concept. To illustrate this point, I will now focus on an argument of Kant’s that is based on the singularity of space and time. b. The singularity of space and time. In both editions of the Critique of Pure Reason, Kant presents two arguments based on the claim that we represent space and time as singular objects. Here, my aim is to show that Kant viewed the propositions asserting that space and time are singular as expressing not analytic but rather synthetic knowledge. In this and the following subsections, I will mainly be focusing on Kant’s singularity argument for space, but I will also ­sometimes refer to the singularity argument for time. Kant’s argument for the singularity of space appeals to the way in which we represent the relationship between different spaces. When we think of different spaces, these are necessarily regarded as parts and limitations of a singular space that encompasses all of them (see A24–5/B39). How should we account for Kant’s strategy? Messina would probably say that Kant is simply analyzing the concept of space, which in this case involves an appeal to what Messina calls ‘modal intuitions’. Accordingly, Kant is showing that singularity is an essential mark of our concept of space by imagining concrete objects that instantiate the concept (see Messina 2015: 428). As a result, we can obtain analytic knowledge that space is singular. The problem with Messina’s reading is that what he calls ‘modal intuitions’ do not seem to be compatible with Kant’s idea of analytic knowledge in the Critique. Kant is clear that the need to appeal to objects that instantiate a concept is paradigmatic of synthetic, not analytic judgements. Take Kant’s ‘containment’ account of analytic judgements in the Introduction, where he claims that to see the connection between the subject concept and the predicate concept in an analytic judgement, we do not need to go ‘beyond’ the subject concept (see A6–7/B10–11). This means that if it is necessary to appeal to objects that instantiate the subject concept of a judgement in order to see the connection between this subject concept and the predicate concept, the judgement is synthetic, according to Kant. This view is made explicit in a Reflexion from the mid-1770s: ‘In analytical judgements, the predicate properly pertains to the concept a, in synthetic ones to the [crossed out: condition of the] object of the concept, and the predicate is not contained in the concept’ (Refl. 4684, 17:671). Since what Messina calls a ‘modal intuition’ depends on imagining concrete objects

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

81

that instantiate the concept, judgements that necessarily appeal to these intuitions cannot be analytic in Kant’s sense.6 At this point, one could point out that if imagining the possible instantiation of an object is only used to clarify ‘marks’ that analytically belong to it, the procedure will provide analytic knowledge after all. If we frame the appeal to ‘modal intuitions’ in a vocabulary closer to the logic of concepts with which Kant was familiar, we might say that it is often the case that the concepts we use are not ‘distinct’ for us, which means that we do not have a clear grasp of the ‘marks’ that form them. It is plausible to think that imagining concrete objects that instantiate a concept is instrumental to making our concept ‘distinct’.7 This is correct. However, it is important to keep in mind that we go through the procedure in question due to contingent features of our psychology. It is not the case that, ‘in principle’, analytic knowledge requires modal intuitions for Kant, whereas it is only synthetic knowledge that, ‘in principle’, requires appealing to objects that instantiate the subject concept of a synthetic judgement. Therefore, even if we concede that the appeal to objects that instantiate a concept does not rule out the knowledge which is at stake being analytic, an account of analytic knowledge in Kant that relies on such an appeal owes us some criteria for distinguishing when the appeal to objects is a mark of synthetic knowledge (as Kant often stresses), and when it is not (because it is only a means to obtain a ‘distinct’ concept). 6

7

h

d i

One might object that my point appeals to the ‘containment’ account of analytic judgements. Since Kant provides different accounts of the analytic/synthetic distinction, Messina’s claim that modal intuitions provide analytic knowledge might be reconcilable with one of these alternative accounts. However, it is plausible that the ‘containment’ account is more fundamental than other accounts (for a view of this kind, see Anderson 2015: Ch. 1.3). This is confirmed by Kant’s insistence that what is distinctive about synthetic judgements is that they need a ‘third thing’ to establish a connection between their subject and predicate concepts (see A155/B194). Moreover, he sometimes describes the ‘third thing’ that is necessary for synthetic a priori judgements as a ‘pure object’ (see A157/B196), which provides further evidence that a judgement that depends on imagining an object that instantiates a concept must be synthetic. I thank Marcus Willaschek for raising this issue. Another possible objection might refer to Kant’s characterization of analytic judgements in the Jäsche Logic (9:111), where Kant characterizes analytic judgements by taking into account how they refer to objects (for an interpretation that builds on this passage, see Longuenesse 1998: 87–8). I believe we can account for Kant’s appeal to objects in his characterizations of analytic judgements by taking into account two issues, the first of which is the issue of how we form concepts and determine the analytic relationships between them and their marks. In this sense, it is plausible to think that we consider objects when we form a particular empirical concept. However, when we have formed that concept, it is not necessary to take into consideration objects that fall under it in order to justify an analytic judgement that depends on the contents of that concept. A second issue is whether our concepts, and the analytic relationships subsisting in them, can count as cognition of objects. It might be the case that, given our concept of ‘body’, the judgement ‘all bodies are extended’ is true. However, it does not follow from this that there are objects that correspond to our concept of body. I thank Marcello Garibbo for pointing out the passage in the Jäsche Logic. I thank Marcus Willaschek and Andrew Chignell for this objection.

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

82

More importantly, even if we do not treat the appeal to objects that instantiate an object as a mark of synthetic knowledge, Kant is in fact explicit that some of the judgements that are taken into account in the metaphysical expositions are synthetic. In the singularity argument for time, Kant remarks that ‘the proposition that different times cannot be simultaneous cannot be derived from a general concept. The proposition is synthetic, and cannot arise from concepts alone’ (A32/B47). Kant hints at the fact that the way in which we arrive at the proposition ‘Time is singular’ is similar to the way in which we arrive at the proposition ‘Different times cannot be simultaneous’.8 In both cases, we attribute properties to time that are not analytically contained in its concept. Therefore, the metaphysical exposition of the concept of time considers its use in synthetic a priori judgements. I submit that in analysing the ‘given’ concepts of space and time, the metaphysical expositions perform an ‘analysis’ of these concepts in a broader sense. This analysis also considers the use of a concept in judgements where the concept is not analytically connected to the predicate. In taking into account this use of the concept, metaphysical expositions need to account for how we arrive at the synthetic a priori judgements in question. c. The singularity argument for space. Let me return to Kant’s singularity argument for space. What I aim to show is, first, that Kant starts his argument from the concept of space and shows that from this concept alone we cannot grasp space as singular. Space is in fact regarded as necessarily singular, however. From this Kant concludes that our original representation of space cannot be conceptual but must be intuitive. The argument proceeds as follows: Space is not a discursive or, as is said, general concept of relations of things in general, but a pure intuition. For, first, one can only represent a single space, and if one speaks of many spaces, one understands by that only parts of one and the same unique space. And these parts cannot as it were precede the single all-encompassing space as its components (from which its composition would be possible), but rather are only thought in it. It is essentially single; the manifold in it, thus also the general [allgemeine] 8

h

d i

0 0

Since two simultaneous times would not belong to a singular time, one might regard these two propositions as equivalent. They are not, however. To see this, compare the case of space, where two simultaneous spaces can (and must) belong to the same space. This points to the fact that, to see that two simultaneous times cannot belong to the same time (and are thus impossible), we need to appeal to our intuition of time, which grounds both the proposition that ‘time is singular’ and the proposition that ‘different times cannot be simultaneous’.

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

83

concept of spaces in general [überhaupt], rests merely on limitations. From this it follows that in respect to it an a priori intuition (which is not empirical) grounds all concepts of it. Thus also all geometrical principles, e.g., that in a triangle two sides together are always greater than the third, are never derived from general concepts of line and triangle, but rather are derived from intuition and indeed derived a priori with apodictic c­ ertainty. (A24–5/B39)

The first thing to note about the argument is that Kant clearly thinks that we can have a concept of space. This is expressed in the sentence claiming that ‘an a priori intuition […] grounds all concepts of it’ or that the ‘general concept of spaces’ rests on limitations. Since concepts of space are here contrasted to space as an intuition, and since Kant speaks of a ‘general’ concept of spaces (where ‘generality’ is a distinctive characteristic of concepts for Kant), the term ‘concept’ must be taken in Kant’s specific sense and not as meaning ‘representation’ in general. Moreover, Kant claims that there must be an intuition at the basis of these concepts. Accordingly, his argumentative strategy appears to assume that we have concepts of space, his aim being to show that we also need an a priori intuition of space in order to explain how we can attribute singularity to it. Therefore, the singularity argument for space does not proceed by first assuming we have a representation of space that is either a concept or an intuition and then determining that it must be an intuition (see Falkenstein 1995: 218). Rather, it starts by assuming that we have a concept of space and then determines that we must also have an intuition of space and that the latter is more fundamental than the former. Let me make two additional points before I propose my reconstruction of the argument. First, in the singularity argument for space, Kant presupposes that he has already established that our representation of space is a priori in the first two arguments of the metaphysical exposition. Since the first two arguments are not aimed at showing that the representation of space is originally an intuition, we can take Kant as presupposing that the concept of space is a priori.9 Second, Kant speaks of a ‘concept of spaces’ and ‘concepts’ of space. How should we interpret the fact that Kant refers to ‘concepts’ and ‘spaces’ in the plural? One possibility is to say that it is not really the general concept of space that is at stake in these expressions, but rather determinate concepts describing particular regions of space. This is implausible, however, since the generality of the concept Kant is 9

h

d i

I remain neutral here regarding whether this result of the first two arguments can be achieved solely by means of conceptual analysis.

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

84

considering is emphasized twice when he uses the expression ‘the general [allgemeine] concept of spaces in general [überhaupt]’. Because the concept is general, it must be equally applicable to different regions of space. I submit that Kant speaks of a ‘general concept of spaces’ here because he wants to emphasize that singularity is not something that can be attributed to space on the basis of our concept of it. For Kant, concepts are general by definition (see 9:91), which means that they can in principle be applied to different objects. As a consequence, the general concept of space must be applicable to different ‘spaces’, which means that the singularity of space cannot be appreciated by considering that concept alone. Keeping these preliminary remarks in mind, how can we reconstruct the singularity argument for space? Here is my attempt: (1) We possess an a priori concept of space (from the first two arguments in the metaphysical exposition of space). (2) We cannot represent space as singular on the basis of our a priori concept of space. (3) We regard space as necessarily singular. (4) The necessary singularity of space cannot be derived from empirical intuitions. (5) We must be in possession of a non-conceptual a priori representation of space that allows us to represent space as necessarily singular (from 2, 3 and 4). (6) The only way we can represent space as necessarily singular through a non-conceptual a priori representation is by means of intuition. (7) We must have an a priori intuition of space (from 5 and 6). (8) The representation through which we see space as singular is more fundamental than any other representation of space. (9) Our a priori intuition of space is more fundamental than any other representation of space (from 7 and 8). The preliminary remarks provided above should be enough to clarify the general direction of the argument. Still, steps (6) and (8) need further elucidation. Thanks to (6), Kant can go from a negative characterization of the a priori representation of space that supports the claim of singularity (it is non-conceptual) to a positive characterization of it (it is intuitive). How can Kant justify this step? The easiest way to account for it would be to say that Kant simply presupposes that the a priori representation of space through which we represent its singularity must be either a concept or an intuition. Since it is not the former, it must be the latter. But he does not only do that. Rather, he also provides positive evidence for the

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

85

claim that the representation through which we represent the singularity of time must be an intuition. He does so by appealing to the content of our representation of the singularity of space and argues that the best way to describe that content is to say that it is intuitive. This is the upshot of Kant’s remarks concerning the way in which we represent parts–whole relationships in space. Kant argues that the fact that we represent space as necessarily singular is essentially connected to the way in which we represent the relationships between the parts and the whole in it. More precisely, we represent the totality of space as necessarily given prior to its parts, where the parts are only limitations of the originally given singular space. Given the radically different way in which parts–whole relationships are represented through concepts, Kant’s point is that the best way to capture this way of representing singularity is to say that it is intuitive.10 Let us now turn to step (8). The idea behind it is that any representation that gives access to a priori knowledge that another representation would not provide, while also giving access to all a priori knowledge obtainable through this second representation, must be more fundamental than this second representation. This seems to be the case with the intuition of space. It provides a priori knowledge (for instance, ‘space is singular’ or ‘particular spaces are only limitations of singular space’) that is not derivable from other representations of space (namely, its concept). By contrast, it is plausible to suppose that any a priori knowledge that we can obtain through the concept of space is also obtainable through the intuition of space. Note that my reading of Kant’s singularity argument for space makes intelligible his use of the example of the triangle in the last sentence of the argument. Just as we cannot derive the judgement ‘space is singular’ from the concept of space, we cannot derive the judgement ‘in a triangle two sides together are always greater than the third’ from ‘general concepts of line and triangle’ (A25/B39). In order to see how we can obtain the latter proposition, we need to go beyond our concept of a triangle and attend to our capacity to represent the relationships between the sides of a triangle through our intuition of space.11 10

11

h

d i

Kant thinks that parts–whole relationships are represented through composition in concepts. One could perhaps object that when we use concepts to cognize organisms, we must also presuppose that the whole comes before the parts. Therefore, the priority of the whole to the parts is not a distinctive feature of intuition. It might be true that in concepts of organisms the whole precedes the parts, but organisms certainly display parts–whole relationships that are different from those of space, since in organisms these relationships are not based on limitations. There is nonetheless a problem with the analogy between how we attribute singularity to space and how we attribute certain properties to triangles. In Kant’s account of geometry, seeing that the length of two sides of a triangle is always greater than the third would require a construction

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

86

d. The method of the metaphysical expositions. What conclusions can we draw regarding the method of the metaphysical expositions of space and time? First of all, we have seen that it is correct to consider the metaphysical expositions as analyses of our given concepts of space and time. This does not mean that all knowledge with which the expositions are concerned is analytic knowledge that can be obtained from these concepts. Rather, ‘analysis’ should here be understood in a broad sense, as also including ‘analysis’ of the use of given concepts in synthetic a priori judgements that we commonly perform with those concepts.12 An analysis of this kind is at stake in Kant’s singularity arguments for space and time. The arguments establish that we need to ‘go beyond’ the given concepts of space and time in order to see how we can regard space and time as necessarily singular. Furthermore, they establish that we must have an a priori non-conceptual way of representing space and time that gives us access to their singularity. This explains why, in the case of the concepts of space and time, it is better to speak of ‘concepts of root representations for the cognition of objects’. Root concepts are pure concepts for the cognition of objects that lie at the basis of synthetic a priori judgements. As is clear from my analysis of the judgements that ascribe singularity to space and time, these

12

h

d i

0 0

in intuition. Does this mean that the singularity of space is also established through construction? This would contradict Kant’s claim that the use of constructions is distinctive of mathematics and is denied to philosophers (see A712–38/B740–66). Indeed, the way in which Kant invites us to attend to our representation of parts–whole relationships in space does suggest a procedure of construction (for a similar claim with regard to other arguments in the Aesthetic, see Falkenstein 1995: Ch. 6). Kant invites us to imagine different spaces so as to realize how we regard those spaces as necessarily part of a singular encompassing space. I believe the problem is only apparent, however. This becomes clear when we realize that Kant’s argument is an analysis of the ‘given’ concept of space. As we saw, analysis can here be understood in a broad sense, as including analysis of the use of a concept in synthetic a priori judgements that we commonly perform with that concept (such as ‘Space is singular’). The purpose of the appeal to intuition and the use of construction within this ‘analysis’ is not to obtain new synthetic a priori knowledge regarding the object described in the concept in question, but rather to clarify how we arrive at a synthetic a priori judgement that we assume at the start, one that contains the concept under analysis. Something similar can be said regarding Kant’s philosophical account of mathematics. In that context, the task of an appeal to construction is not to obtain new synthetic a priori judgements based on construction. Rather, an appeal to construction is necessary to explain how we obtain some synthetic a priori judgements in mathematics, while these synthetic a priori judgement are assumed ‘as given’ at the beginning of our philosophical analysis. One might ask why it is appropriate to speak of ‘analysis’ in this case. One reason to keep this term is that Kant himself sometimes uses it in this broader sense. In the Introduction to the Critique, for example, Kant submits that the critique of pure reason ‘goes only so far in the analysis as is requisite for the complete estimation of synthetic a priori cognition’ (A14/B28). In Gava (2015) and Gava (2018a), I argued that the analysis of given a priori concepts is important for understanding the method of the Critique of Pure Reason. What I say about analysis there should be taken to imply this broader sense of analysis.

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

87

synthetic a priori judgements can be explained only by attending to our intuitive representations of space and time. Therefore, it is these representations that must be seen as ‘root’ representations, not the concepts of space and time. As far as the singularity argument for space is concerned, we have seen that it also provides a positive characterization of our non-conceptual representation of space, which is obtained by pointing at the specific manner in which we represent parts–whole relationships in it. Accordingly, Kant starts from an ‘analysis’ of the given concept of space, but this analysis is able to track a specific way of representing space that is non-conceptual and to isolate an essential feature of this representation, namely its way of grasping parts–whole relationships. This means that Kant’s ‘analysis’ in the metaphysical expositions is also able to go beyond the given concepts and isolate a different form of representation. It does so by identifying features of the content of our representations that, while distinguishable a priori, can only pertain to a non-conceptual representation of objects. By tracking and positively characterizing this original non-conceptual manner of representation, the arguments in the metaphysical expositions do not simply assume a distinction between concepts and intuition, on the one hand, and understanding and sensibility, on the other. Rather, they contribute to establishing what intuition is and to identifying sensibility as a distinct faculty of cognition.13 One might argue that according to my reconstruction of the metaphysical exposition of space, the latter only determines that we have a pure intuition of space, not that it is sensible. As I will argue in the next chapter, it is in the transcendental exposition of space that Kant shows that our pure intuition of space essentially determines the form of outer sense. Therefore, it seems that it is only at that point that the characterization of the intuition of space that we obtain in the metaphysical exposition can be said to contribute to determining what sensibility is. This is true. But this does not mean that the metaphysical exposition does not contribute to determining what sensibility is. It only means that it is only at a later stage that we discover that the faculty 13

h

d i

To be clear, I am not suggesting that the metaphysical expositions do not assume any previous characterization of intuition. They obviously do that. As the singularity argument shows, however, they do not simply conclude that the original representations of space and time must be intuitive from the inability of the concepts of space and time to ground certain of our synthetic a priori judgements. Rather, the expositions provide positive evidence that these original representations must be intuitive, such as when they consider the way in which we represent parts–whole relationships in them. In this way, they allow us to learn something new about intuition, and they accordingly contribute to establishing what intuition is.

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

88

of cognition that the metaphysical exposition helps us to determine is sensibility.

2  The Metaphysical Deduction of the Categories I will now turn my attention to the metaphysical deduction in the Transcendental Analytic. Unlike the metaphysical deductions of space and time, in this case we do not have the problem of locating the deduction. In the B-edition of the Critique, Kant explicitly contrasts a metaphysical with a transcendental deduction of the categories (B159). There is agreement among Kant scholars that the metaphysical deduction takes place in the first chapter of the Transcendental Analytic, which is entitled On the Clue to the Discovery of all Pure Concepts of the Understanding (hereafter ‘the Clue chapter’). Additionally, what Kant does in this chapter matches my general characterization of metaphysical deductions. Kant explicitly says that the metaphysical deduction of the categories provides a complete catalogue of the root concepts belonging to the understanding (A81–2/B107–8). Moreover, in the B-edition, he emphasizes that the characteristic feature of the metaphysical deduction of the categories is that it focuses on their origin (B159). Although determining where the metaphysical deduction is located may be simple enough, understanding how it proceeds is a different matter. This will be the task of this section. Similar to my approach to the metaphysical deductions in the Aesthetic, I will show that, by establishing the origin of the categories, the metaphysical deduction of the categories contributes to determining what the understanding (in the narrow sense) is as a faculty. Since much of the discussion concerning the method of the metaphysical deduction in the Analytic has revolved around the problem of ‘completeness’ for Kant’s list of categories, however, I will first introduce this problem in some detail. a. The problem of ‘completeness’. In the introductory remarks in the Clue chapter (A66–7/B91–2), Kant makes a distinction between a ‘rhapsodic’ (see A81/B106) and a systematic presentation of the fundamental concepts belonging to a faculty of cognition. While the former proceeds by observing our use of a faculty and isolating the fundamental concepts when they become apparent in this use, the second identifies a ‘principle’ from which the table of concepts can be derived. This principle allows us to systematically order the concepts and guarantees that the table is complete. Kant

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

89

later clarifies that the principle that guarantees the completeness of his table is ‘the faculty for judging’ (A81/B106). As is well known, in the metaphysical deduction Kant provides two ‘tables’, each containing twelve items divided into four groups. The first table represents forms of judegments (A70/B95), while the second lists the categories (A80/B106). At least one of the purposes of the metaphysical deduction is to establish the parallelism of these two tables, such that precisely one category corresponds to each form of judgement. The parallelism extends to the way in which forms of judgements and categories are grouped. In both tables, items are arranged into four groups of three, and the ‘titles’ that characterize the groups are ‘quantity’, ‘quality’, ‘relation’ and ‘modality’. Kant’s claim that the table of the categories is complete and systematically ordered expresses his idea that the categories are precisely twelve in number and that they are arranged according to the model just outlined. There is a famous objection to Kant’s claim to completeness in the metaphysical deduction. The objection has it that Kant takes for granted a classification of forms of judgement that was customary in the logic textbooks of his time and ‘derives’ the categories by showing that a parallel table of categories could be obtained. This appears to conflict with Kant’s contention that he was able to derive the categories from a ‘principle’. Moreover, since the classification of the forms of judgement is simply presupposed, Kant does not offer a proof that this classification is complete, which means that the table of the categories is not shown to be complete either. The objection dates back at least to Hegel but has been repeated by many other figures in the history of philosophy (see M. Wolff 1995: 36–7 for an overview of some of these figures). As a response to this objection, Klaus Reich (1986) and Michael Wolff (1995) have famously tried to show that Kant indeed wanted to offer an argument for the completeness of the table of the forms of judgement. Reich claims that Kant does not provide such an argument in the Critique of Pure Reason but that he planned to do so within his metaphysical system (Reich 1986: § 6). By contrast, Wolff reads Kant as offering such an argument within the Clue chapter, in particular its first and second sections. Wolff’s approach has obvious consequences for how he reads the method of the metaphysical deduction. Moreover, his reading has been highly influential. For these reasons, I will briefly present his interpretation here. While I think that Wolff’s book is an invaluable read for anybody interested in the metaphysical deduction, he sometimes reads more into the Clue chapter than it actually contains.

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

90

The Method of Transcendental Philosophy

Wolff concedes that the table of the forms of judgement is used as a heuristic for establishing that there are twelve logical ‘functions’ of the understanding (in the broad sense) that lie at the basis of both the twelve forms of judgement and the twelve categories (see M. Wolff 1995: 40). Nevertheless, he challenges the idea that Kant simply assumes that there are four classes of forms of judgement. Rather, Kant establishes that there are precisely four classes of forms of judgement by deriving these classes from four ‘fundamental functions’ of the understanding that are exercised in judgement. Since Kant is able to establish that there are only these four ‘fundamental functions’ and that the four classes of the forms of judgement can be traced back to them, Kant does have a principle that guarantees the completeness of his table (see M. Wolff 1995: 39–40). What are these fundamental functions? Three of these functions correspond to three fundamental ways of using concepts in judgements. First, concepts are used predicatively when they are used as predicates in a judgement. When they are not used predicatively but play the role of the subject of the judgement, they can refer to their object either ‘immediately’ or ‘mediately’ (see M. Wolff 1995: 106). In the first case, the subject concept is taken to express the objects to which it refers explicitly. In the second case, the subject concept is taken to express an object to which it refers only implicitly. Using the same example on which Wolff draws, in the judgement ‘All bodies are divisible’, the concept ‘body’ can be taken to refer immediately to bodies, but only mediately to metals. How is this mediate reference established? It is established through the judgement ‘Every metal is a body’, that is, through a judgement in which ‘body’ plays the role of the predicate. From this judgement and the judgement attributing divisibility to bodies, we can inferentially derive the judgement ‘All metals are divisible’. In this sense, the concept ‘body’ in the judgement ‘All bodies are divisible’ can be taken to mediately refer to metals (see A68–9/B93–4; M. Wolff 1995: 96–102). To these three fundamental ways of using concepts in judgements correspond the three ‘faculties’ that Kant sees as the constituent capacities of the understanding in the broad sense: the understanding in the narrow sense as the faculty of concepts corresponds to the predicative use of concepts; the faculty of judgement corresponds to the non-predicative use of concepts with immediate reference to objects; and reason in the narrow sense as the faculty of inference corresponds to the non-predicative use of concepts with mediate reference to objects (see M. Wolff 1995: 91–3, 109–10). These three faculties can thus be traced back to the three fundamental ways of

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

91

using concepts in judgements, which, in turn, express three fundamental functions of the understanding in the broad sense (see M. Wolff 1995: 112). Wolff then adds a fourth fundamental function of the understanding that corresponds not to a way of using concepts in judgements but to judging as such (see M. Wolff 1995: 112–13). Therefore, what the first section of the Clue chapter establishes, according to Wolff, is that there are exactly four fundamental functions of the understanding in the broad sense that correspond to four ‘functions of unity in judgements’ (A69/B94).14 It is the identification of these four fundamental functions independently of any given logical table that guarantees the completeness of Kant’s table of the forms of judgement. Once the first section has identified these four fundamental functions, the second section first establishes a connection between the four fundamental functions and the four ‘titles’, according to which the forms of judgement are divided into classes (quantity, quality, relation and modality). The complete table of the forms of judgement, which is indeed presupposed by Kant, is a heuristic tool for determining that three more specific functions are subordinated to each fundamental function of the understanding, so that the complete table of the functions of the understanding contains twelve of these functions (see M. Wolff 1995: 175–6). The completeness of the table of the functions is the basis for claiming that the table of the categories is equally complete. As I said, I consider Wolff ’s detailed reconstruction illuminating in various respects. Nonetheless, I do not agree that the main aim of the first section of the Clue chapter is to establish that there are exactly four fundamental functions of the understanding in the broad sense and that these functions correspond to four ‘functions of unity in judgements’. A first problem with this approach is that it projects onto Kant’s argument a desideratum which was first identified in later criticisms of Kant. It is not clear that the best way to defend Kant from these criticisms is to say that the desideratum of demonstrating that the table of the forms of judgement is complete was Kant’s as well. Indeed, when Kant discusses the problem of completeness, he usually connects it to the categories and not to the forms of judgements or the functions they express (see A66–7/B91–2; A80–1/B106–7). Moreover, it is the connection that Kant establishes between the table of the categories and the 14

h

d i

For a recent useful reading of the first section of the Clue chapter that similarly identifies four fundamental functions that are exercised in cognition through concepts, where these functions in turn correspond to the four classes of the categories, see Hoeppner (2021: Ch. 2).

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

92

‘faculty for judging’ that guarantees the completeness of the latter table on his view (see A80–1/B106–7). This suggests that the first task of the Clue chapter, as far as the completeness of the table of the categories is concerned, is to establish a connection between the categories and the activity of judging. A second problem concerns Wolff’s claim that all fundamental functions of the understanding in the broad sense are identified in the first section of the Clue chapter, which in turn allows us to establish a complete list of all the functions of this faculty in the second section. This implies, on the one hand, that there is no function of the understanding in the broad sense that is left out of the Clue chapter and, on the other, that there is no other root concept of the understanding in the broad sense that corresponds to this function. We have already seen that reason in the narrow sense as the faculty of inference is one constituent element of the understanding in the broad sense. Moreover, Wolff thinks that the function of unity pertaining to this faculty is already established in the first section of the Clue chapter, in connection to the non-predicative use of concepts with a mediate reference to objects. In the section On the Transcendental Ideas in the first book of the Transcendental Dialectic, however, Kant speaks of a ‘function’ that is performed by reason in the narrow sense in its inferences (A321/B378). This cannot be a function that is already identified in the Clue chapter, since in the section in question Kant brings up this function in connection to the purpose of identifying forms of inference that are different from the forms of judgement identified in the metaphysical deduction of the categories (A321/B378).15 Because there are functions of the understanding in the broad sense that are first analysed in the Dialectic, and because there are classes of root concepts that correspond to these functions, the Clue chapter cannot present an exhaustive list of all of the functions that pertain to the understanding in the broad sense. b. The essential link between concepts and judgements. In what follows, I will present my interpretation of the metaphysical deduction. On my reading, the main purpose of the first section of the Clue chapter is to establish an essential connection between our use of concepts and judgements. By contrast, the second section provides a table of forms of judgement that 15

h

d i

0 0

As we will see below, an appeal to the forms of inference lies at the basis of one of the two strategies Kant follows for establishing that there are exactly three classes of root concepts that pertain to reason in the narrow sense.

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

93

corresponds to fundamental ways in which different concepts are brought under other concepts in judgements. The table belongs to transcendental logic because it distinguishes between forms of judgement according to the role they play in the cognition of objects. Finally, the third section establishes that the categories are fundamental ways in which a manifold of intuition is united in a single intuition. As such, they are simply a different application of the forms of judgement. Let us start with the first section of the Clue chapter. Commentators on the metaphysical deduction rightly emphasize that the concept of a function plays a key role in it (see Longuenesse 1998: Part 2; Longuenesse 2006; Caimi 2000; M. Wolff 1995: 19–32). They sometimes view Kant as identifying these ‘functions’ as items that are clearly distinguishable from the forms of judgement. In this way, once Kant singles out these functions as independent items, he can show that the forms of judgement and the categories are different ways of exercising them.16 If it was Kant’s purpose to single out functions of the understanding that are clearly distinguishable from the way they are exercised in judgements, he was certainly unable to make this purpose clear. I think it is more charitable to attribute a different aim to Kant in connection with his use of the concept of function. More precisely, I think that by speaking of functions, Kant wanted to take a different perspective on concepts, namely, one that focuses on what we can do with concepts. On my reading, a function is the activity through which different representations, usually concepts, are brought ‘under’ a concept. This closely resembles Kant’s own definition, according to which a function is ‘the unity of the action of ordering different representations under a common one’ (A68/B93),17 where a ‘common’ representation, as a general representation that is superordinate with respect to other representations, can only be a concept. How can we characterize this activity? What does it mean to say that it brings representations ‘under’ a concept? In order to 16

17

h

d i

This approach is linked to the purpose of showing that Kant does not start his argument by assuming a given set of forms of judgement. Rather, he bases the argument on the identification of certain fundamental functions. As we saw, this is certainly the case in Michael Wolff’s reading of the Clue chapter, but he is not alone in this respect. In a way that is similar to Wolff, Caimi (2000) thinks that what he calls ‘original functions’ can be identified as independent items. Unlike Wolff, however, he thinks that the table in the second section of the Clue chapter lists these functions, not forms of judgement. It must be noted, however, that Kant here defines a function not simply as an action but as the ‘unity of an action’, which might suggest that a function is not an action after all. I here follow Till Hoeppner (2011), who reads Kant’s reference to the ‘unity of an action’ as pointing out that a ­function is a complex action constituted by more than one simple action.

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

94

make this clear, let me quote a long passage from the first section of the Clue chapter: Now the understanding can make no other use of these concepts than that of judging by means of them. Since no representation pertains to the object immediately except intuition alone, a concept is thus never immediately related to an object, but is always related to some other representation of it (whether that be an intuition or itself already a concept). Judgement is therefore the mediate cognition of an object, hence the representation of a representation of it. In every judgement there is a concept that holds of many, and that among this many also comprehends a given representation, which is then related immediately to the object. So in the judgement, e.g., ‘All bodies are divisible,’ the concept of the divisible is related to various other concepts; among these, however, it is here particularly related to the concept of body, and this in turn is related to certain appearances that come before us. These objects are therefore mediately represented by the concept of divisibility. All judgements are accordingly functions of unity among our representations, since instead of an immediate representation a higher one, which comprehends this and other representations under itself, is used for the cognition of the object, and many possible cognitions are thereby drawn together into one. (A68–9/B93–4)

Let me point out a few things about this passage. First, Kant’s example clearly describes a judgement in which the concept ‘divisible’ is analytically contained in the concept ‘body’. In Kant’s terminology, concepts that are ‘in’ another concept are simply the constituent marks of that concept. Concepts that are ‘under’ a concept are concepts that contain the latter concept as a mark. In the example, the concept ‘body’ stands under the concept ‘divisible’ because it contains ‘divisibility’ as one of its marks. By standing under the concept ‘divisible’, the concept ‘body’ has all the constituent marks of the former, plus additional marks that further specify the properties that bodies must have in comparison to the properties possessed by all divisible things. Kant accordingly says that there are ‘various other concepts’ that stand under the concept of the divisible. In the judgement in question, however, the concept of the divisible is related to one particular concept that stands under it, that is, the concept ‘body’. In this way, the concept ‘body’ is brought ‘under’ the concept ‘divisible’. Second, Kant says that all judgements are functions. I take this claim literally. Because Kant here speaks of judgements in general and not only of analytic judgements, it means that the activity of bringing concepts under another concept cannot only express analytic relationships between concepts, where the concept ‘A’ is brought ‘under’ the concept ‘B’ simply because it contains ‘B’ as one of its constituent marks. Moreover,

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

95

since functions are activities through which different representations are brought under a concept, and since in a judgement we bring one concept under another, judgements are particular cases of functions. They are those functions that bring concepts under another concept. Third, by bringing one concept under a more general concept, judgements use the more general concept to cognize something about the objects that are picked out by the more specific concept. In Kant’s words, in a judgement, ‘instead of an immediate representation a higher one, which comprehends this and other representations under itself, is used for the cognition of the object’.18 In the example, the concept ‘divisible’ is used to cognize something about the objects picked out by the concept ‘body’. This describes the particular way in which we cognize objects through concepts. While intuitions represent their objects immediately, concepts can only represent their objects by being related to another representation of the same object. Fourth, Kant seems to think that judgements, as the functions that bring concepts under another concept, are the only way in which concepts are used for the cognition of objects. Accordingly, Kant says that ‘the understanding can make no other use of these concepts than that of judging by means of them’. If my reading is right, then Kant uses the concept of a function to establish a necessary connection between concepts and judgements.19 Judgements simply are the functions through which concepts are brought under another concept. It is through this operation that we obtain what we can properly call ‘cognition through concepts’. The primary result of the first section of the Clue chapter is thus a positive characterization of the faculty of understanding as the faculty of ‘cognition through concepts’ (A68/B93). Since cognition through concepts can only be obtained by 18

19

h

d i

Unlike Michael Wolff (1995: 96–102), I do not think that the example Kant discusses later in the first section of the Clue chapter is fundamentally different from that discussed in the quoted passage. Kant says that ‘concepts, as predicates of possible judgements, are related to some representation of a still undetermined object. The concept of body thus signifies something, e.g., metal, which can be cognized through that concept. It is therefore a concept only because other representations are contained under it by means of which it can be related to objects. It is therefore the predicate for a possible judgement, e.g., “Every metal is a body”’ (A69/B94). What Kant is saying is that in order to be possibly used for the cognition of objects, the concept of a body must have other concepts under it. In this way, the concept of a body can be used to cognize objects when one concept that stands under it is explicitly brought under it in a judgement, as in the judgement ‘Every metal is a body’, where the concept ‘metal’ is brought under the concept ‘body’. Therefore, I do not think that Kant is describing an inference in this passage or hinting at the faculty of reason in the narrow sense. In this way, Kant denies that we can have a simplex apprehensio through concepts, namely, a cognition through a concept that does not take the form of a judgement. For an analysis of the novelty of this thesis, see M. Wolff (1995: 75, 77).

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

96

The Method of Transcendental Philosophy

bringing concepts under other concepts in judgements, the faculty of the understanding, as a faculty of cognition through concepts, is also a ‘faculty for judging’ (A69/B94). Notice that I here take the first section of the Clue chapter as providing a characterization of the understanding in the narrow sense as the faculty of concepts. This contrasts with Wolff ’s reading, which views this section as describing the four fundamental functions of the understanding in the broad sense, as comprising the understanding in the narrow sense, the faculty of judgement, reason in the narrow sense and judgement as such (see M. Wolff 1995: 89–90; see also Brandt 1991: 52). This reading is supported by Kant’s use of the expression ‘understanding in general’ (Verstand überhaupt) at A69/B94, which suggests that he has the understanding in the broad sense in mind. At the beginning of the Analytic of Principles, however, Kant uses the same expression, ‘understanding in general’ (Verstand überhaupt) (A132/B171), to characterize the understanding as ‘the faculty of rules’, which he contrasts in that context with the faculty of judgement as ‘the faculty of subsuming under rules’. However, the understanding in the broad sense also contains the faculty of judgement. Moreover, in another passage from the Introduction to the Dialectic, Kant contrasts the ‘faculty of rules’ with reason in the narrow sense as the ‘faculty of principles’ (A299/ B346), which shows that the understanding, as the faculty of rules, cannot include reason as the faculty of inference. This means that in the Analytic of Principles Kant uses the expression ‘understanding in general’ in a way that is not equivalent to the understanding in the broad sense, since the ‘faculty of rules’ does not contain either the faculty of judgement, as the ‘faculty of subsuming under rules’, or reason, as the ‘faculty of principles’, within it, whereas the understanding in the broad sense should include these. Therefore, the fact that Kant speaks of the understanding in general in the first section of the Clue chapter is not evidence that he means the understanding in the broad sense. Moreover, even though Kant speaks of a ‘faculty for judging’ (Vermögen zu urteilen), this seems to be different from his description of the ‘faculty of judgement’ (Urteilskraft) in the Analytic of Principles. In that context, the faculty of judgement is characterized as a ‘talent’ (A133/B172), according to which we are able to recognize that a particular case falls under a rule or a concept. In the Clue chapter, by contrast, the faculty for judging is responsible for the activity of bringing concepts under another concept in a judgement. This suggests that the first section of the Clue chapter is only meant to provide a positive characterization of the understanding in the narrow sense and that it does so by highlighting an essential connection between concepts and judgements.

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

97

Kant closes the first section of the Clue chapter by claiming that ‘the functions of the understanding can therefore all be found together if one can exhaustively exhibit the functions of unity in judgements’ (A69/B94). I read this sentence as making an additional point on the basis of the connection between concepts and judgements that Kant has just established. This is not immediately clear, however. To make this apparent, consider first that ‘understanding’ should here be taken to mean understanding in the narrow sense, which is the faculty of cognition through concepts (see A68/B93). Second, the passage introduces a new characterization of functions. As we saw, Kant had called judgements functions because they instantiate the activity through which concepts are brought under concepts. Here, however, Kant speaks of ‘functions of unity in judgements’. What is a ‘function of unity’, and how can a function be in a judgement if that judgement is a function? I take it that Kant is here referring to fundamental kinds of functions. If judgements are functions because they bring concepts under another concept, one can distinguish among fundamental ways in which they perform this task. These fundamental ways in which they bring concepts under another concept correspond to different kinds of functions. Accordingly, when Kant speaks of the ‘functions of unity in judgements’, he simply means the kinds of function that specific judgements instantiate. Specific judgements are functions. Insofar as they belong to a kind, one can say that this kind is instantiated in the specific judgement. Keeping these preliminary remarks in mind, what does the passage at the end of the first section of the Clue chapter say? It says that if one wants to identify the fundamental ways in which concepts are brought under another concept for the sake of cognition through the understanding – or, in other words, if one wants to identify the fundamental kinds of cognition through concepts – one simply needs to identify the fundamental kinds of judgements. The point is simply that since there is no cognition through concepts that is not at the same time a cognition in judgements, once one has identified the fundamental kinds of judgements, one has at the same time identified the fundamental kinds of cognition through concepts. c. The forms of judgement for the cognition of objects. If my reconstruction of the first section of the Clue chapter is right, its purpose is to legitimate the appeal to fundamental kinds of judgements in order to identify the fundamental kinds of cognition through concepts. Of course, this fits perfectly with Kant’s use of the table of the forms of judgement as a tool for the discovery of the categories. One possible objection against

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

98

this reading is that it cannot escape the ‘completeness’ problem discussed above. That is to say, it may be argued that although Kant indeed justifies appealing to the forms of judgement in the discovery of the categories, he does not prove that the table of the forms of judgement is in fact complete, which in turn invalidates the claim to the completeness of the table of the categories as well. I have already suggested that a detailed proof of the completeness of the table of the forms of judgement is more a desideratum that has been attributed to Kant (due to a classical criticism of the metaphysical deduction) than a task that Kant himself saw as essential to his own purposes. Kant seems to think that the forms of judgement are by nature easier to catalogue in a systematic way than concepts. Kant might have thought that a systematic table of the forms of judgement was easier to obtain because concepts can only be catalogued by considering their content. This means that if one wants to single out concepts that are pure or ‘root’ concepts, one must analyse the content of a multiplicity of concepts and show that a group of them possesses specific characteristics that justify seeing its members as constituting a special kind. The problem with this approach is that one can never be sure that one has included all potentially relevant concepts in the initial analysis. In cataloguing forms of judgement, by contrast, the content of the concepts that figure in the judgements is not crucial. One simply focuses on the different ways in which concepts are connected in a judgement. Kant’s idea might thus have been that, given the ‘formal’ character of a catalogue of judgements, it is easier to find out that one fundamental form of judgement is missing than to realize that a ‘root’ concept has been left out. Leaving these speculations aside, I wish now to show that even if Kant does not provide a detailed argument for the completeness of the table of the forms of judgement, it is wrong to say that, in the second section of the Clue chapter, he simply took for granted a classification of forms of judgement that was customary in the logic textbooks of his time. As other commentators have emphasized (M. Wolff 1995: 28–32; Caimi 2000; Schulting 2018a: 6), Kant clearly states that the classification of forms of judgement that he offers in the second section of the Clue chapter belongs not to general but to transcendental logic. Accordingly, when he explains why, under the title of quality, he distinguishes between infinite and affirmative judgements, he claims: ‘in a transcendental logic infinite judgements must also be distinguished from affirmative ones, even though in general logic they are rightly included with the latter and do not constitute a special member of the classification. General logic abstracts from all content of the

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

99

predicate (even if it is negative), and considers only whether it is attributed to the subject or opposed to it. Transcendental logic, however, also considers the value or content of the logical affirmation made in a judgement by means of a merely negative predicate, and what sort of gain this yields for the whole of cognition’ (A71–2/B97). Infinite judgements are judgements with a negative predicate, like the judgement ‘The soul is non-mortal’ (see A72/B97). Kant says that general logic would simply consider this judgement an affirmative one, since we affirm the ‘non-mortality’ of the soul. From the point of view of transcendental logic, however, an infinite judgement of this kind must be distinguished both from the negative judgement ‘The soul is not mortal’ and from the affirmative judgement ‘The soul is mortal’. Unlike the negative judgement, which only denies that the predicate of mortality pertains to the soul without making any positive claim regarding the soul, the infinite judgement makes a positive claim, since it says that ‘the soul is one of the infinite multitude of things that remain if I take away everything that is mortal’ (A72/B97). Still, the infinite judgement is different from an affirmative one, because it does not determine what the soul is but only places the soul in the infinite sphere of possible things that remain once one determination (mortality) is taken away.20 What is distinctive about the transcendental logic’s approach to the forms of judgement? In the passage quoted above, Kant says that while general logic abstracts from all content of the predicate of judgements when considering their quality, transcendental logic does consider one aspect of this content – that is, whether the content of the predicate contains a negation. Of course, this should not be taken to mean that, when it catalogues forms of judgement, transcendental logic takes into consideration the actual content of the concepts that are used in judgements. Like general logic, transcendental logic only considers formal aspects of judgements. The relevant difference between general logic’s and transcendental logic’s analyses of judgements is rather that while the distinctions introduced in the latter are intended to illuminate the contribution that different forms of judgement make in the cognition of objects, this aim does not belong to general logic. In other words, given the particular aim of transcendental logic, there are certain distinctions among forms of judgement that are important within it, although they are irrelevant to general logic.21 20

21

h

d i

For an account of the importance of infinite judgements in the Critique, see Serck-Hanssen (2013). See also Stang (2012). Capozzi (2018) shows the importance of the distinction between negative and infinite judgements for understanding the Antinomy of Pure Reason. Kant makes a similar point regarding the quantity of judgements, even though he does not ­explicitly mention transcendental logic. He explains why he distinguishes between universal and

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

100

The Method of Transcendental Philosophy

Since the table of forms of judgement of the second section of the Clue chapter belongs to transcendental logic and deviates from classifications of general logic, it cannot be said that Kant simply took a traditional table for granted. Rather, he carefully reflected on the status of the classification he offered. He could begin with distinctions among forms of judgement customarily made within logic. However, since transcendental logic implies a new perspective on the forms of judgement, and since this new perspective sometimes needs additional distinctions with respect to the one that pertains to general logic, the table of forms of judgement of the second section of the Clue chapter must be taken as the fruit of Kant’s own work on transcendental logic. d. Connecting the forms of judgement to the categories. Let me summarize the results of the first two sections of the Clue chapter according to my reconstruction. In the first section, Kant establishes that cognition through concepts needs judgements, which are the functions through which concepts are brought under another concept. Given this essential connection between cognition through concepts and judgements, Kant submits that once one has identified the fundamental kinds of judgement, one has at the same time identified the fundamental kinds of cognition through concepts. The table of the forms of judgement in the second section of the Clue chapter identifies these fundamental kinds of judgements. How does Kant get from the forms of judgement to the categories? We know that judgements are the only way in which we can obtain cognition through concepts. This happens when a concept is brought under another concept in a judgement, so that the more general concept is used to indirectly cognize something regarding the objects picked out in the less general concept. The question concerns how we are able to see that the more particular concept falls under the more general concept. Again, using singular judgements as follows: ‘The logicians rightly say that in the use of judgements in syllogisms singular judgements can be treated like universal ones. […] If, on the contrary, we compare a singular judgement with a generally valid one, merely as cognition, with respect to quantity, then the former relates to the latter as unity relates to infinity, and is therefore in itself essentially different from the latter. Therefore, if I consider a singular judgement (judicium singulare) not only with respect to its internal validity, but also, as cognition in general, with respect to the quantity it has in comparison with other cognitions, then it is surely different from generally valid judgements (judicia communia)’ (A71/B96). Even in this case, what characterizes the perspective of Kant’s table is the contribution of different forms of judgement to cognition. For an account of why the category of unity corresponds to universal judgements and the category of totality to singular judgements, see Capozzi (2013).

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

101

the example we borrowed from Kant, how can we see that the concept ‘metal’ falls under the concept of ‘body’ in order to judge that ‘every metal is a body’? If we only operate at the level of concepts and judgements, the only way in which we can justify the judgement is if the concept of ‘body’ is already analytically contained as a mark in the concept of ‘metal’. In the present example, this seems to be the case. Even in this example, however, how can we explain the fact that we are in possession of a concept of ‘metal’ that is so constituted as to analytically contain the mark ‘body’? Kant’s answer is that we must assume a previous ‘synthesis’ of the manifold of intuition through which we are able to see that the objects that fall under the concept of metal are necessarily bodies. Of course, once we have obtained this concept, together with its analytical marks, every judgement that spells out those marks can be made without any reference to a synthesis of the manifold of intuition. Still, this synthesis is needed to explain how we first developed the concept as a concept that can provide actual cognition of objects: Prior to all analysis of our representations these must first be given, and no concepts can arise analytically as far as the content is concerned. The synthesis of a manifold, however, (whether it be given empirically or a priori) first brings forth a cognition, which to be sure may initially still be raw and confused, and thus in need of analysis; yet the synthesis alone is that which properly collects the elements for cognitions and unifies them into a certain content; it is therefore the first thing to which we have to attend if we wish to judge about the first origin of our cognition. (A77–8/B103)

What does Kant mean when he says that ‘no concepts can arise analytically as far as the content is concerned’? In my view, he means that if the analytic relationships displayed between a concept and its constituent marks are to provide some kind of cognition of objects, these relationships must be traceable to relationships among characteristics of the relevant objects cognizable in intuition. To provide an example, the analytic judgement ‘Every metal is a body’ can constitute a cognition of what actual metals are only because it rests on a previous synthesis of the manifold of intuition which supports the analytic unity we see between the concept ‘metal’ and the concept ‘body’. One could perhaps clarify Kant’s point by saying that we can certainly arbitrarily define a concept and obtain an analytic judgement on the basis of that definition. However, if the analytic judgement is to provide some sort of cognition regarding actual objects, this means that the concept on which our judgement is based must have some sort of connection with a synthesis in intuition, through which we can see the characteristics we attribute to

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

102

objects by means of that concept as instantiated together in the objects in question.22 If appealing to a synthesis of the manifold of intuition is essential to explaining how analytic judgements can provide cognition of objects, appealing to a synthesis in intuition is even more essential to accounting for synthetic judgements. Take the judgement ‘All bodies are heavy’, which Kant regards as a synthetic judgement (see A7/B11). According to Kant’s account of judgement in the first section of the Clue chapter, this judgement provides a cognition through concepts by bringing the concept ‘body’ under the concept ‘heaviness’. Since the judgement is synthetic, however, this cannot mean that the concept ‘body’ already contains the concept ‘heaviness’ as an analytic mark. Therefore, in this case more than in the previous one, it is only through appealing to a synthesis in intuition that we can explain how we can appreciate the connection between the concepts that the judgement expresses. What is Kant’s first move in the third section of the Clue chapter? He starts by assuming what he already established in the first section, that is, that the only way in which we can obtain cognition through concepts is through a judgement that brings a concept under another concept. He then argues that in order to understand how judgements can yield actual cognition of objects through this activity, we must assume a synthesis of intuitions that is able to provide the concepts used in those judgements with some content and a connection to objects. In a second move, Kant submits that the activity through which a sensible manifold is united in a single intuition expresses, at another level, the same activity through which different concepts are brought under another concept in a judgement (for a similar claim, see Young 1992: 105). This means that to the fundamental ways in which we connect concepts in judgements for the sake of cognition there correspond fundamental ways in which we connect a manifold of intuition in a single intuition: the former are the forms of judgement identified in the second section of the Clue chapter; the latter are the categories that Kant introduces in the third section: The same function that gives unity to the different representations in a judgement also gives unity to the mere synthesis of different representations in an intuition, which, expressed generally, is called the pure concept 22

h

d i

0 0

Notice that this is compatible with what I say above regarding analytic judgements as not requiring reference to objects. It is one thing to say that I do not need a reference to objects to see that the relationship between concepts that is expressed in an analytic judgement obtains given a certain concept. It is another thing to say that through an analytic judgement that makes the contents of a concept explicit I can cognize something that pertains to objects.

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

103

of understanding. The same understanding, therefore, and indeed by means of the very same actions through which it brings the logical form of a judgement into concepts by means of the analytical unity, also brings a transcendental content into its representations by means of the synthetic unity of the manifold in intuition in general, on account of which they are called pure concepts of the understanding that pertain to objects a priori. (A79/B104–5)

How should the categories be understood in this picture? The categories are simply the fundamental ways in which we order the manifold of intuition to provide the necessary grounding for the connections between concepts that we establish in judgements for the sake of cognition. Let me add one last remark before I move on to the concluding section on the metaphysical deduction of the categories. In the third section of the Clue chapter, Kant often emphasizes that the synthesis guided by the categories is ‘pure’ because the manifold that we order through them is given not empirically but rather a priori through the pure forms of space and time (see A76–7/B102; A78–9/B104; Caimi 2000: 274–5 insists on the importance of these claims). This is an important point in the economy of the Critique of Pure Reason. It plays an essential role in the transcendental deduction of the categories when it comes to demonstrating that to any synthesis of an empirical manifold of intuition there must correspond a synthesis of a pure manifold in space and time. As we will see in the next chapter, this is necessary for explaining how the categories can provide a priori knowledge that constrains possible experience. However, I do not think that the reference to a pure manifold of intuition is essential to the metaphysical deduction. The latter aims to catalogue the categories as ‘root’ concepts for the cognition of objects. It does so by arguing that the categories correspond to fundamental ways of ordering the manifold of intuition in accordance with the fundamental forms of judgement. It is not the task of the metaphysical deduction, however, to explain how the categories can provide a priori knowledge that constrains possible experience. e. The method of the metaphysical deduction of the categories. What can be said of the method of the Clue chapter on the basis of my analysis? First, Kant’s approach fundamentally rests on an investigation of the way in which we use concepts for the cognition of objects. In this respect, Kant proposes an original thesis and claims that the only way in which concepts can be used for cognition is in judgements that bring one concept under another. Second, through this original thesis Kant provides a

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

104

The Method of Transcendental Philosophy

positive characterization of the understanding in the narrow sense. The latter, as the faculty of cognition through concepts, is at the same time and fundamentally a ‘faculty for judging’. Therefore, even in the case of the metaphysical deduction of the categories, the deduction contributes to positively characterizing the faculty of cognition to which the ‘root’ concepts that are being deduced belong. It does so by tracing the origin of the categories in the forms of judgement and the kinds of functions they represent. In my analysis, I have remained neutral concerning whether the Clue chapter can be considered a successful argument and have only reconstructed what I consider to be its essential steps. In concluding this analysis, let me point out one thing that remains obscure and one problem. One issue that remains unclear is how Kant’s general characterization of the way in which concepts are used in judgements fits with the specific forms of judgement. The examples that Kant provides to clarify the general characterization are the judgements ‘All bodies are divisible’ and ‘Every metal is a body’. As we saw, both judgements express an analytic relationship between concepts, in the sense that the subject concept is under the predicate concept because it contains it as an analytic mark. I have already suggested that Kant’s general description must be applicable to synthetic judgements that do not express this analytic relationship, such as the judgement ‘All bodies are heavy’. There are further problems regarding Kant’s general characterization of judgements, however. The judgements in Kant’s examples are ‘universal’ judgements. But consider particular judgements, such as the judgement ‘Some bodies are metals’. In this case, if we look at the analytic relationship between the concepts involved, it is the predicate concept that stands under the subject concept, since it is the concept ‘metal’ that contains the concept ‘body’ as a mark. But particular judgements do not necessarily rest on an analytic relationship of this kind. Take the judgement ‘Some dogs are brown’. In this judgement, the subject concept does not contain the predicate concept as an analytic mark, but nor does the predicate concept contain the subject concept. One can perhaps say that the concept ‘dog’ is brought under the concept ‘brown’ because there are more things that are brown than things that are dogs, and only in this sense is brown more ‘general’ than ‘dog’. Alternatively, one could also say that ‘brown’ stands under the concept ‘colour’, which is a property of material things such as ‘dogs’. These suggestions notwithstanding, determining what the activity of ‘bringing one concept under another’ could be in the case of particular judgements remains difficult. Similar

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

105

considerations can be made with r­ eference to other forms of judgement in Kant’s table. Finally, let me point out a problem with Kant’s line of reasoning. Kant does not seem to provide an argument for the claim that the categories simply are the forms of judgement in another guise. What he does argue is that we must assume a synthesis of intuitions in order to make sense of how judgements can provide cognition of objects. However, the simple fact that there must be such a synthesis does not imply that it must be performed according to the same procedures by which concepts are united in judgements.

3  The Metaphysical Deduction of the Transcendental Ideas Where the metaphysical deduction of the transcendental ideas occurs and what it establishes might seem obvious. After all, Kant explicitly draws a comparison between the metaphysical deduction of the categories and what he does in the section On the Transcendental Ideas in the first book of the Transcendental Dialectic (see A321/B377; A329/B386). In this section, Kant invokes a parallel between categorical, hypothetical and disjunctive syllogisms on the one hand, which he derives from the three relational categories, and three forms of ‘the unconditioned’ on the other (A323/B379). Additionally, because Kant alludes to three types of the unconditioned, it is natural to assume, as many interpreters have (see for example Renaut 1998: 356), that the metaphysical deduction of the transcendental ideas only identifies three ideas: the ideas of the soul, the world and God. Things are not so simple, however. Nikolai Klimmek (2005: 51) and Marcus Willaschek (Willaschek 2018: 171–5) have plausibly argued that Kant does not offer a derivation of the transcendental ideas in the first book of the Transcendental Deduction. Rather, he describes three classes of ideas (see A334/B391), which serve the purpose of systematically organizing them. The actual derivation of the ideas is accomplished in the second book of the Transcendental Dialectic, where Kant analyses the dialectical inferences of traditional metaphysics. Moreover, there are more than three transcendental ideas, although they are organized according to the three classes identified in the first book. Taking the second class as an example, it collects four cosmological ideas that respectively represent the complete composition of appearances, the complete division of appearances, the complete causal chain explaining the origin of an appearance and the complete set of conditions of something that is

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

106

The Method of Transcendental Philosophy

contingent (I take this to be a clearer reformulation of Kant’s classification at A415/B443).23 The problem of locating the metaphysical deduction of the ideas does not only arise when one compares what Kant accomplishes in the first and second books of the Dialectic, however. Even assuming that the deduction takes place in the first book, it is not clear where it is located and where it begins and ends, for Kant appeals to two different trichotomies in order to derive his distinction between psychological, cosmological and theological ideas. The first trichotomy is the one I have already mentioned and classifies types of syllogisms according to the three relational categories. It is central to the second section of Book I. The second trichotomy is first introduced in the third section and distinguishes between three fundamental relationships in a representation, which are relationships to the subject and to the object considered as either appearance or object in general (A333–4/B390–1). It is unclear what the relationship between these two trichotomies is. One can accordingly ask which is responsible for the metaphysical deduction of the classes of transcendental ideas, or whether they both play a role in it.24 In what follows, I will present my interpretation of the metaphysical deduction of the transcendental ideas. While I agree with Klimmek and Willaschek that Book I of the Dialectic only presents a classification of three classes of ideas, I regard its second and third sections as still presenting essential steps in their metaphysical deduction. I have suggested that metaphysical deductions contribute to the identification of the faculties of cognition to which the concepts that they deduce belong. Sections 2 and 3 of the first book of the Dialectic present Kant’s attempt to establish a firm link between how the transcendental ideas originate and reason as the faculty of inference. In other words, they support Kant’s claim to systematicity and completeness by linking the transcendental ideas to a unitary and systematic characterization of reason. I will start by presenting Kant’s characterization of reason in the section On the Transcendental Ideas. As I will suggest, Kant in fact offers 23

24

h

d i

0 0

Even though Klimmek and Willaschek agree on these general points, they give different accounts of how many transcendental ideas there are (twelve for Klimmek and nine for Willaschek) (see Klimmek 2005: 74; Willaschek 2018: 169) and of Kant’s derivation of the ideas in the second book of the Dialectic. While Willaschek sees this derivation as being obtained through the dialectical inferences of reason (Willaschek 2018: 174), for Klimmek the derivation is independent of such inferences (Klimmek 2005: 118–21). For example, Renaut (1998) attributes the responsibility for carrying out the deduction to the first trichotomy. By contrast, Wundt (1924: 222) tries to establish a connection between the two trichotomies. Finally, Bröcker (1970: 95) and Klimmek (2005: 50) claim that it is the second trichotomy that does the work in the metaphysical deduction.

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

107

two conflicting accounts of how reason arrives at the idea of the unconditioned. Next, I will discuss Kant’s two different strategies for showing that the transcendental ideas are divided into exactly three classes, as they play out in the sections On the Transcendental Ideas and System of the Transcendental Ideas, respectively. In a further step, I will consider how Kant brings the derivation of the cosmological ideas to completion in the chapter on the Antinomy of Pure Reason. This will help to shed light on the difficulties I identified regarding Sections 2 and 3 of the first book of the Dialectic. While only one of the two strategies for deriving the classes of ideas fits what Kant is doing in the Antinomy of Pure Reason, I will suggest that the derivation of the cosmological ideas serves as a basis for Kant’s account of the transcendental ideas in general. In turn, this explains why Kant provides two conflicting accounts of the unconditioned in the second section of Book I. This is a consequence of Kant’s attempt to make the psychological and theological ideas fit into this general account. a. Two characterizations of reason and the unconditioned in On the Transcendental Ideas. Kant opens the second section of Book I of the Dialectic by clearly stating that, in analogy with what he did for the categories in relation to the understanding in the Analytic, the derivation of the ‘pure concepts of reason’ goes hand in hand with a characterization of the structure and nature of reason as a faculty (see A321/B378). Reason is described as a faculty of cognition through inference and as naturally driven towards the concept of the totality of conditions for a given conditioned, which, in turn, gives us the idea of the unconditioned. While Kant seems consistent if we stop at this very general characterization of reason and the unconditioned, when we consider Kant’s line of reasoning in On the Transcendental Ideas in more detail, it is my contention that he provides two contrasting accounts of reason and its relation to the unconditioned, where it is unclear how they relate to one another. The first account can be found at the very beginning of Section 2. Kant accordingly claims that ‘[t]he function of reason in its inferences consisted in the universality of cognition according to concepts, and the syllogism is itself a judgement determined a priori in the whole domain of its condition’ (A321–2/B378). What does this mean? Kant provides clarification through an example. He writes that what makes a sentence like ‘Caius is mortal’ a cognition of reason is its inferential derivation from a more general premise, such as ‘All humans are mortal’, through the minor premise ‘Caius is a man’ (A322/B378). In Kant’s vocabulary, ‘humans’ is the ‘condition’ under which mortality is asserted in the major premise. The

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

108

minor premise brings the ‘condition’ of the conclusion, i.e. Caius, under the ‘condition’ of the major premise, so that what is asserted regarding all humans in the major premise can be asserted of Caius in the conclusion. Kant further explains what happens in this categorical syllogism by saying: Accordingly, in the conclusion of a syllogism we restrict a predicate to a certain object, after we have thought it in the major premise in its whole domain under a certain condition. This complete magnitude of the domain, in relation to such a condition, is called universality (universalitas). (A322/B378–9)

What is the function of reason in an inference? In an inference like the one in the example, the conclusion is derived from a major premise that ascribes mortality to all of the objects that fall under a certain condition (being human).25 It is in this sense, then, that reason aims at ‘the universality of cognition according to concepts’ and asserts something ‘in the whole domain of its condition’ (A321–2/B378). In the example, Kant only considers a particular form of inference, that is, a categorical syllogism with a universal judgement as one of its premises.26 This is problematic in many respects given that he wants to offer a general characterization of reason as a faculty of inference. I do not want to discuss this issue here, however. After having characterized reason in this way, Kant derives the concept of the unconditioned from the universality he attributes to the major premise of the syllogism. He writes: In the synthesis of intuition this [the universality of the major premise] corresponds to allness (universitas), or the totality of conditions. So the transcendental concept of reason is none other than that of the totality of 25

26

h

d i

0 0

Klimmek (2005: 25–6) and Willaschek (2018: 177) read Kant’s claim to universality as a reference to the entire domain of objects of which mortality can be asserted (living beings). This reading is not obligatory, however. Kant could also mean that in every premise of a universal syllogism a certain property is universally asserted for a certain class of objects, identified by the ‘condition’ that delimits the assertion. According to this approach, Kant would be attributing universality to the major premise even though the latter would not define the entire domain of objects of which a certain property can be predicated. Thus, according to this second reading, one could attribute universality to the premise ‘All humans are mortal’ (where humanity is the condition under which mortality is universally asserted), even though the premise does not identify the entire domain of objects that are mortal. For a reading along these lines, see Allison (2004: 315). A categorical syllogism with the structure of the example could have a universal judgement as a conclusion, as when we derive the sentence ‘All scientists are mortal’ from the premises ‘All humans are mortal’ and ‘Scientists are humans’. Universality therefore cannot be the defining feature of the major premise. Even in this case, though, the syllogism would identify in the major premise a ‘condition’ with a domain of objects that is broader than the domain of objects captured by the ‘condition’ in the conclusion. For this reason, Kant would probably say that in a syllogism of this kind, the universality with which the conclusion is asserted is only derivative compared to the ­universality of the major premise.

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

109

conditions to a given conditioned thing. Now since the unconditioned alone makes possible the totality of conditions, and conversely the totality of conditions is always itself unconditioned, a pure concept of reason in general can be explained through the concept of the unconditioned, insofar as it contains a ground of synthesis for what is conditioned. (A322/B379)

In the passage, Kant first connects the universality of the major premise to the category of allness or totality. He then submits that a syllogism like the one under analysis appeals, through the universality of the major premise, first, to the idea of the totality of conditions for a given conditioned and, second, to the concept of the unconditioned, which Kant claims is indispensable to accounting for such a totality of conditions. As various interpreters have pointed out, there are many problems with Kant’s argument in the quoted passage. Let me just list a couple of these. First, it is unclear how Kant can establish a connection between the universality of the major premise and the category of totality.27 Second, Kant seems to use the concept of a ‘totality of conditions’ in at least two different senses. When this concept is linked to the universality of the major premise in the example, this means the totality of objects in the domain of the subject concept, of which a predicate is universally asserted. As we saw, Kant calls this subject concept the ‘condition’ under which the predicate is asserted. Accordingly, the major premise does not encompass a ‘totality of conditions’ but only a ‘totality of objects’ falling under the condition of the major premise. In addition to this sense of ‘totality of conditions’, ‘totality of conditions’ is used to hint at the ‘real grounds’ that together can explain the possibility of a given object (see Allison 2004: 316; Willaschek 2018: 179). For example, this latter meaning is at stake when Kant speaks of the ‘totality of conditions to a given conditioned thing’ (my emphasis). Given this ambiguity in Kant’s use of the idea of the ‘totality of conditions’, it is unclear how he can get from the former to the latter. Leaving these difficulties aside, what I want to emphasize is that, according to this first account of the unconditioned, reason obtains the idea of a totality of conditions and the unconditioned without appealing to a regress in the series of premises through prosyllogisms. Rather, the idea of the unconditioned is obtained in connection to the universality with which a property is predicated under a certain ‘condition’ in the major premise of a single universal categorical syllogism. 27

h

d i

One way to emphasize this difficulty is to note that totality is the third category under the title of quantity, whereas universal judgements are the first types of judgements in the corresponding title in the table of the forms of judgement. See Heimsoeth (1966–1971: Vol. 1, 44n); Willaschek (2018: 178).

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

110

Let us now turn to the second account of reason and the unconditioned that we find in the second section of Book I of the Dialectic. This can be found towards the end of the section and more closely recalls Kant’s description of reason in the Introduction to the Dialectic. Kant starts by offering a description of reason that basically agrees with the characterization we have just taken into account: Reason, considered as the faculty of a certain logical form of cognition, is the faculty of inferring, i.e., of judging mediately (through the subsumption of a condition of a possible judgement under the condition of something given). The given judgement is the universal rule (major premise, major). The subsumption of the condition of another possible judgement under the condition of the rule is the minor premise (minor). The actual judgement that expresses the assertion of the rule in the subsumed case is the conclusion (conclusio). The rule says something universal under a certain condition. Now in a case that comes before us the condition of the rule obtains. Thus what is valid universally under that condition is also to be regarded as valid in the case before us (which carries this condition with it). (A330/B386–7)

What Kant says here matches what we have discussed above almost p ­ erfectly. If we continue reading, however, it becomes clear that, in the present context, Kant does not regard the universality of the ‘rule’, that is, the major premise, as sufficient to obtain an idea of the totality of conditions and the unconditioned. Rather, in the text that immediately follows the quotation, Kant appeals to reason’s regress in the series of premises through prosyllogisms to derive this idea. He first notes that the series of concatenated ­sentences that every syllogism is ‘can be continued to an indeterminate extent either on the side of the conditions (per prosillogismos) or on the side of the conditioned (per episyllogismos)’ (A331/B387–8). He then submits that the series of prosyllogisms has a significance that the series of episillogisms cannot have: ‘For since in the first case [the series of prosyllogisms] the cognition (the conclusio) is given only as conditioned, we cannot reach it by means of reason except at least on the presupposition that all members of the series are given on the side of the conditions (totality in the series of premises), because only under this presupposition is the judgement before us possible a priori’ (A331/ B388). In other words, we cannot regard a given cognition as a cognition of reason unless we regard it as inferentially ‘conditioned’, namely, as derived from premises. But since these premises can be seen as ‘conclusions’ of further syllogisms, and so on, reason takes not only the first premises as given but also the complete series of premises that can be obtained through a series of valid prosyllogisms.

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

111

In this second account of reason, Kant derives the idea of the totality of conditions for a given conditioned from the series of premises in a chain of prosyllogisms: Now it may or may not be that on the side of the conditions, the series of premises has a first [member] as the supreme condition, and hence that it is without bound a parte priori; nevertheless it must still contain the ­totality of the condition, assuming that we could never succeed in grasping it; and the whole series must be unconditionally true if the conditioned, which is regarded as a consequence arising from it, is supposed to count as true. (A332/B389)

Kant stresses that independently of whether reason takes the totality of conditions of a given conditioned as finite (and so as having a first ‘supreme’ condition) or as infinite, it surely needs to assume this totality as given. As I’ve already suggested, this second characterization of reason more closely resembles Kant’s account of reason in the Introduction to the Transcendental Dialectic, where he also insists on the importance of prosyllogisms. Recall that the ‘maxim’ of reason in its logical use urges reason to ascend in the series of conditions through prosyllogisms and to seek the unconditioned (A307/B364). How are Kant’s two accounts of reason and the unconditioned in the section On the Transcendental Ideas related to one another? Why does Kant fall into this apparent inconsistency? I will answer this latter question below, but I will first analyse another inconsistency in the metaphysical deduction of the ideas. b. Two derivations of the classes of ideas. Kant scholars have emphasized another ‘conflict’ in the metaphysical deduction of the ideas (see Guyer 2000: 80–84; Rohlf 2010: 205–6; Klimmek 2005: 48–50). As noted above, I regard the part of the metaphysical deduction contained in Book I of the Dialectic as only providing a distinction among three classes of ideas. The conflict to which I am referring arises when one compares the way in which the three classes of ideas are obtained in the second and third sections of Book I, respectively. Since this conflict has already been discussed in the literature, I will present it only briefly. Let me start with the first strategy for obtaining the three classes of ideas. Kant distinguishes between three general types of ‘the unconditioned’, namely, an ‘unconditioned, first, for the categorical synthesis in a subject, second for the hypothetical synthesis of the members of a series, and third for the disjunctive synthesis of the parts in a system’ (A323/B379). These three types are derived from three kinds of syllogisms,

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

112

The Method of Transcendental Philosophy

that is, categorical, hypothetical and disjunctive syllogisms. These types are in turn related to the three relational categories, given that these categories correspond to categorical, hypothetical and disjunctive judgements and given that it is by means of the distinction between these judgements that Kant distinguishes between the corresponding types of syllogisms: categorical, hypothetical and disjunctive syllogisms are simply syllogisms with either a categorical, a hypothetical or a disjunctive judgement as the major premise. There are various problems with this strategy for deriving the three main classes of transcendental ideas, the most evident of which is the difficulty of obtaining classes of ideas that have to do with the soul, the world and God, respectively, simply from the form of these three kinds of syllogism. Leaving these difficulties aside, what I wish to emphasize is that, according to this first strategy, Kant’s concept of the unconditioned is from the very start divided into three fundamental types. By this I mean that there is no general idea of the unconditioned which is then specified into different kinds. Rather, since there are three fundamental ways of proceeding syllogistically and it is in relation to these procedures that we develop our ideas of the unconditioned, there must be three original forms of the unconditioned, which in turn correspond to the three classes of ideas.28 But let us see how the second derivation of the classes of ideas proceeds. Kant starts by distinguishing between two fundamental relationships in one representation. These are the relationship to the subject and the relationship to the object. He then further divides the second relationship, saying that it is either to objects considered as appearances or to objects in general that a representation is related. From these two dichotomies, he obtains a trichotomy according to which the fundamental relationships in a representation are: ‘1) the relation to the subject, 2) to the manifold of the object in appearance, and 3) to all things in general’ (A334/B391).29 After Kant has introduced the tripartition of 28 29

h

d i

0 0

This can be said independently of whether one views the regress through prosyllogisms as essential or superfluous to obtaining an idea of the unconditioned. There are, of course, various problems concerning this trichotomy. To begin with, it is unclear how it relates to the trichotomy of syllogisms. Furthermore, if one assumes that the two trichotomies are related, it is unclear which one is more fundamental. By contrast, if one assumes that they are independent of one another, it remains to be determined which is responsible for the actual derivation of the three classes of transcendental ideas. But even leaving the relationships between the two trichotomies to the side, in considering the second one gets the impression that it comes out of the blue and was arbitrarily conceived only to serve Kant’s purposes in the third section (see Schmucker 1990: 63–4).

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

113

relationships in a representation, he proceeds to derive the classes of transcendental ideas as follows: Now all pure concepts have to do generally with the synthetic unity of representations, but concepts of pure reason (transcendental ideas) have to do with the unconditioned synthetic unity of all conditions in general. Consequently, all transcendental ideas will be brought under three classes, of which the first contains the absolute (unconditioned) unity of the thinking subject, the second the absolute unity of the series of conditions of appearance, the third the absolute unity of the condition of all objects of thought in general. (A334/B391)

Kant distinguishes between three types of the unconditioned even in this quote. However, the way in which he gets to this distinction is different from the strategy based on the trichotomy of kinds of syllogism. Of course, there is the obvious difference that one strategy starts from a subdivision of relationships in a representation while the other starts from kinds of syllogism. The difference I want to emphasize is another one. As we saw, in the derivation based on the kinds of inference, the concept of the unconditioned is immediately divided into three fundamental types. Kant immediately distinguishes between three fundamental kinds of inferential relationship to which three original forms of the unconditioned correspond. In the derivation based on the relationships in a representation, by contrast, Kant first assumes a general idea of the unconditioned that does not distinguish between three fundamental kinds. This is clear when Kant says that concepts of pure reason ‘have to do with the unconditioned synthetic unity of all conditions in general’ (my emphasis), where no distinction between kinds of the unconditioned is yet made. This general idea of the unconditioned is presumably obtained from an account of the inferential nature of reason that does not distinguish between three fundamental kinds of syllogism. It is only in a second step that Kant distinguishes between types of the unconditioned when possible domains of application of reason are taken into account. This means that a distinction between these types is not directly obtained through a consideration of the forms of inference but is only obtained in a second move.30 c. The deduction of the cosmological ideas. I will now consider how Kant completes the deduction of the cosmological ideas in the Antinomy of Pure Reason. This is useful for three reasons. First, it will allow us to appreciate 30

h

d i

Rohlf (2010: 205) also interprets the second derivation as being based on an application of the idea of the unconditioned in three different domains.

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

114

The Method of Transcendental Philosophy

how the deduction of the ideas that fall under one of the classes identified in Book I is carried to completion in Book II. Second, this analysis may shed some light on the first part of the metaphysical deduction and clarify what strategy for deriving the three classes of ideas is most appropriate. Third, it may also help to identify one source of the conflict between two accounts of the unconditioned in the second section of Book I. The derivation of the cosmological idea is located in the second section of the Antinomy of Pure Reason, which is entitled System of the Cosmological Ideas. Kant starts by providing a characterization of reason in general. He first submits that reason, as the faculty of inference, cannot obtain its concepts on its own. Rather, reason ‘can at most only free a concept of the understanding from the unavoidable limitations of a possible experience, and thus seek to extend it beyond the boundaries of the empirical, though still in connection with it’ (A409/B435–6). This extension of the categories beyond their empirical use is required by reason according to the principle: ‘If the conditioned is given, then the whole sum of conditions, and hence the absolutely unconditioned, is also given, through which alone the conditioned was possible’ (A409/B436). This closely resembles the ‘supreme’ principle that Kant, in the Introduction to the Dialectic, ascribes to the ‘real’ use of reason (A307–8/B364). On the basis of these considerations, Kant submits that ‘transcendental ideas will really be nothing except categories extended to the unconditioned’ (A409/B436). Notice here that Kant still speaks of ‘transcendental ideas’ without specification, so that what he says seems to apply to reason’s pursuit of the unconditioned in general. Without (explicitly) signalling a change in focus from the transcendental ideas in general to the cosmological ideas in particular, Kant then explains that it is only the categories that give rise to a ‘series’ of subordinated conditions that demands an extension towards the unconditioned. Furthermore, this extension takes place not in the ‘descending’ series of consequences but rather in the ‘ascending’ series of conditions of a given conditioned (A409–10/B436–7). It is only at a later point that Kant connects the use of the categories in ascending the ‘regressive’ series of conditions to the cosmological ideas in particular (A411/B438). Accordingly, he derives the four cosmological ideas (namely, the complete composition of appearances, the complete division of appearances, the complete causal chain explaining the origin of an appearance and the complete set of conditions of something contingent) from the regress generated by the categories of totality, reality,

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

115

causality and necessity/contingency, respectively.31 The categories play a different role here in comparison to their use in Kant’s first strategy for deriving the three classes of ideas. In that context, the relational categories were used to distinguish between three forms of syllogism and three corresponding classes of ideas (A323/B379). Here, by contrast, the categories are not related to forms of syllogism and are employed to obtain individual ideas in a way that is directly related to the specific content of each idea.32 Let me briefly illustrate the idea behind each of these derivations. First, in determining the totality of time and space as extensive magnitudes, we regard any given time as preceded and conditioned by the time preceding it, and we view any given space as surrounded by a space that encloses it (A411–12/B438–9). Second, Kant identifies the ‘reality in space’ (A413/ B440) with matter and suggests that we view the parts of matter as the conditions for its existence, where each part will in turn have its own parts. Third, the category of causality gives rise to a regress, because we regard any given event as following from a previous event according to causal laws, where this previous event is also causally determined, and so on (A413–14/B441–2). Fourth, we regard any contingently existing thing as depending on a series of conditions for its existence, where this series is necessary when taken in its totality (A415/B442). Reason’s pursuit of the unconditioned in each of these regresses gives rise to four different specific concepts of the unconditioned that correspond to the four cosmological ideas. In each of these ideas, however, it is still undetermined whether the unconditioned it denotes is finite or infinite (A417–18/B445–6). Can the derivation of the cosmological ideas help to shed light on how the tripartition of three classes of ideas is carried out in Book I of the Dialectic? I believe it can, first of all regarding which strategy for deriving these classes does the real work. Recall that Kant pursues two strategies to obtain the class of cosmological ideas: one based on the form of the hypothetical syllogism, which he connects to the idea of a series (A323/B379; see also A335/B392), and one based on the objects 31

32

h

d i

According to Klimmek (2005: Ch. 2) the cosmological ideas are instead derived from the categories of plurality, negation, causality and existence/non-existence. This corresponds to his attempt to show that there are exactly four psychological ideas derived from the first category in each of the titles in Kant’s table of the categories, four cosmological ideas derived from the second category in each title and four theological ideas derived from the third category in each title (see Klimmek 2005: 73–4). Kant identifies a similar role for the categories in the derivation of the individual psychological ideas at different points in the Paralogisms (see A334/B402; A404). By contrast, he is not explicit regarding how the categories contribute to the derivation of the individual idea of an ens realissimum in the Ideal. Kant hints at this role for the categories in the derivation of individual ideas twice in Book I of the Dialectic (see A326/B383; A335/B392).

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

116

The Method of Transcendental Philosophy

considered as appearances, which he describes as one of the fundamental relata of a representation (A334/B391). Now, if we consider the derivation of the cosmological ideas in the Dialectic, there is no clear reference to each of these derivation strategies. There is no reference to a hypothetical syllogism, unless we take the appeal to the ‘supreme’ principle of reason (A409/B436) as pointing to a premise of such a syllogism. However, the supreme principle is a principle of reason in general. If it grounds hypothetical reasoning, this should characterize every ascent towards the unconditioned, not only those resulting in the cosmological ideas. Similarly, Kant submits that the only categories that are relevant to the derivation of the cosmological ideas are those that give rise to a series (A409–10/B436–7). There is thus a reference to a series, but this is not a clear bridge to the first strategy for deriving the class of cosmological ideas, where the generation of a series should be specific to the derivation of cosmological ideas. By contrast, Kant justifies his appeal to the idea of a series by hinting at the supreme principle of reason which, again, characterizes reason in general: ‘[a]bsolute totality is demanded by reason only insofar as reason is concerned with the ascending series of conditions for a given conditioned’ (A409–10/B436). In other words, Kant does not explain why it is only the cosmological ideas that have to do with series and instead suggests that any ascent from the conditioned to the unconditioned has to do with a series. Things look better for the derivation of the classes of ideas based on the kinds of relations in a representation. Kant says that the regress that leads to cosmological ideas starts from a conditioned given within possible experience (A409/B435–6) and extends an ‘empirical synthesis’ (A409/B436). Moreover, all four cosmological ideas describe totalities of appearances (A415/B443). Kant accordingly emphasizes that, for cosmological ideas, ‘the idea of an absolute totality concerns nothing other than the exposition of appearances, hence it does not concern the understanding’s pure concept of a whole of things in general’ (A416/B443). This recalls Kant’s derivation based on the three fundamental relationships in a representation. In that context, the class of cosmological ideas is obtained from the application of the idea of the unconditioned to the objects considered as appearances. Further evidence in this direction comes from the way in which Kant introduces the idea of the unconditioned in the derivation of the individual cosmological ideas. Recall that Kant’s derivation of the classes of ideas based on kinds of syllogism divides the idea of the unconditioned into three from the very beginning. There are three essential forms of inference and

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

117

three corresponding original forms of the unconditioned. By contrast, the strategy based on the relationships in a representation starts from a general characterization of the unconditioned and only subdivides this idea into three when the application of reason is taken into account. This second route matches more closely what Kant does in the section on the System of Cosmological Ideas. Accordingly, it is only through the application of reason to objects as appearances, an application that follows the lead of some of the categories, that we get from the idea of the unconditioned in general to the cosmological unconditioned. Let me now turn to the conflict between two accounts of the unconditioned in the second section of Book I of the Dialectic. Recall that these conflicting accounts respectively submit that the series of prosyllogisms is either superfluous or fundamental to obtaining the idea of the unconditioned. Can the derivation of the individual cosmological ideas help us to ‘diagnose’ the origin of this conflict? I submit that the cosmological ideas play a primary role in Kant’s account of reason and the unconditioned in general (see Willaschek 2018: 91–2n42 for a similar suggestion). Accordingly, one way to account for the conflict in Section 2 of the first book is to say that it is due to Kant’s attempt to make other forms of the unconditioned, namely those corresponding to the psychological and theological ideas, fit into a framework that is first obtained through the cosmological ideas. Why do I claim that the cosmological ideas play this central role in Kant’s account of reason? One point where this is apparent is in the importance of ‘series’ of conditions both in Kant’s derivation of the cosmological ideas and in his account of reason in general, whereas series do not appear to be essential to obtaining the psychological and theological ideas. Take Kant’s account of the ‘supreme’ principle of reason. Its formulation in the Introduction to the Dialectic explicitly refers to a ‘series’ of conditions: ‘when the conditioned is given, then so is the whole series of conditions subordinated one to the other, which is itself unconditioned, also given’ (A307–8/B364). This idea of a series is directly related to Kant’s derivation of the concept of the unconditioned through the regressive series of prosyllogisms. While reason in its logical use is naturally driven to pursue a regress to ever more general premises for a given conclusion, reason in its real use assumes that there is a series of real conditions that corresponds to the series of logical conditions. That the idea of a series of conditions is also central to the cosmological ideas is evident, for example, in the regress in the series of causes for a given event that characterizes the third cosmological idea (see A414/B441–2 and A444–51/B472–9). By contrast, the

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

118

idea of a ‘series’ of conditions for a given condition does not appear to be essential to Kant’s derivation of the psychological ideas in the Paralogisms (see A343–4/B401–2) and the idea of an ens realissimum in the Ideal (see A571–83/B599–611). These ideas certainly involve the thought that there must be something unconditioned to explain a given conditioned, but their postulation of the unconditioned does not necessarily take the route of an ascent in a ‘series’ of conditions.33 Confirmation that the idea of a series of conditions plays little role in the derivation of the psychological ideas comes from Kant himself when, in the Antinomy, he explains why the category of substance does not give rise to a regress. He claims: What might still seem to be an idea of transcendental reason here would be the concept of the substantial. Only since this signifies nothing other than the concept of a subsisting object in general, insofar as one thinks in it merely the transcendental subject without any predicates, but here only the unconditioned in a series of appearances is under discussion, it is clear that the substantial cannot constitute a member of that. (A414/B441)

This seems to be a clear reference to the use of the category of substance in the derivation of one of the psychological ideas. Kant states that while it is legitimate to speak of a transcendental idea in that context because the substantial subject is thought as an ‘unconditioned’, there is no need for the idea of a series of conditions to get to it. If this is right, Kant’s attempt to derive the idea of the unconditioned in the section On the Transcendental Ideas without appealing to the series of prosyllogisms, appealing only to the universality of the major premise of a universal categorical syllogism, might respond to the need to provide an account of the unconditioned that does not essentially imply a series of conditions, where this is necessary to make room for forms of the unconditioned corresponding to the psychological ideas and the ens realissimum. 33

h

d i

0 0

It is not only the close link between the supreme principle and the idea of a series of conditions that provides evidence for the primary role of the cosmological ideas in Kant’s general characterization of reason. Recall that, in discussing the second characterization of reason in the section On the Transcendental Ideas, Kant submits that the idea of the unconditioned leaves undecided whether the unconditioned involves a first unconditioned condition or an infinite series of conditions (see A332/B389). It is only in the cosmological ideas that the ‘unconditioned’ in each of them comes in two versions, one corresponding to a finite and the other to an infinite series. By contrast, for both the psychological ideas and the Ideal of an ens realissimum, the unconditioned always takes the form of a ‘first’ unconditioned condition. The fact that Kant, in a general characterization of the unconditioned, describes it as containing this essential duplicity is further evidence of the importance of the cosmological ideas in Kant’s general account of reason.

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

119

d. Kant’s method in the metaphysical deduction of the transcendental ideas. As should be clear, I consider the parts of the metaphysical deduction of the transcendental ideas contained in Book I of the Dialectic to be deeply problematic, mainly because Kant works with a general account of reason modelled on the cosmological ideas, where this model does not always fit well with the other transcendental ideas. This does not mean, however, that we can bypass those parts and simply focus on Kant’s derivation of the individual transcendental ideas in Book II. In my view, the parts of the metaphysical deduction contained in Book I have one fundamental task. They contribute to obtaining a fuller picture of what reason as a faculty is. Kant might presuppose at the beginning of the metaphysical deduction that what is typical of reason in the narrow sense is that it proceeds through syllogisms. The parts of the metaphysical deduction contained in Book I extend this picture, and they do so by means of their characterization of reason’s ascent towards the ‘unconditioned’. True, Kant provides two different accounts of how this ascent takes place (one based on the universality of the major premise of a universal categorical syllogism and another based on the regress through prosyllogisms). This reveals an inconsistency in Kant. However, it does not disprove the idea that both accounts of the unconditioned are intended to expand our understanding of what reason is. Like the other metaphysical deductions, the deduction of the ideas contributes to positively characterizing the faculty of cognition to which the ‘root’ concepts that are being deduced belong. It does so by describing the way in which we get to the idea of the ‘unconditioned’ through an inferential procedure.

4  Methodological Pluralism in Kant’s Metaphysical Deductions In this chapter, I have shown how the metaphysical deductions contained in the Aesthetic, the Analytic and the Dialectic of the Critique of Pure Reason, respectively, catalogue pure root concepts for the cognition of objects and track their origin. My starting hypothesis was that these deductions do not assume a readymade distinction between faculties and then proceed to determine which root concepts belong to sensibility, understanding and reason, respectively. Rather, by pointing at the specific origin of the corresponding root concepts, the deductions contribute to establishing what sensibility, understanding and reason are. I have found confirmation of this hypothesis, first, in Kant’s singularity argument for space. This argument contributes to characterizing what sensibility is by identifying the specific way in which parts–whole relationships are

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

120

The Method of Transcendental Philosophy

represented in the intuition of space. Similarly, the metaphysical deduction of the categories contributes to determining what the understanding is by highlighting an essential connection between concepts and judgements, which is the key to tracking the origin of the categories. Finally, it is through the deduction of the transcendental ideas that Kant shows how important the ascent to the unconditioned is for reason, where this characterization goes beyond any description that simply connects reason to the forms of inference. The metaphysical deductions therefore have something in common: they all point to the origin of root concepts for the cognition of objects, and they contribute to offering a characterization of different faculties by emphasizing what is specific to these origins. However, since the metaphysical deductions emphasize the differences between these origins, their commonalities soon come to an end. Tracking the specific origin of the root concepts belonging to sensibility, the understanding and reason, respectively, requires different methodologies. For example, while establishing the origin of the concept of space involved starting from some ‘given’ synthetic a priori judgement (such as ‘Space is singular’) and showing that a non-conceptual representation was needed to arrive at that judgement, neither the deduction of the categories nor the deduction of the ideas started from particular ‘given’ judgements; instead, they started from a general characterization of how we use concepts and judgements in the cognition of objects. The metaphysical deduction of the categories built on the activity through which we bring concepts under other concepts in judgements, whereas the metaphysical deduction of the ideas started from the way in which we can see a judgement as a particular case of a more general rule in an inference. Since both the metaphysical deduction of the categories and the metaphysical deduction of the ideas focus on activities in which concepts play a role, one might be led to think that there is a stronger link between them than the connection they both have to the metaphysical deductions of the concepts of space and time. This should not be taken as revealing a strong methodological continuity between the deductions in the Analytic and the Dialectic, however. For instance, while in the former it is essential to establish a parallel between the way we connect concepts in a judgement and the way we synthesize contents in intuition, establishing a parallel between different procedures is not at all at stake in the latter. In a further attempt to display other similarities between the metaphysical deductions of the categories and the transcendental ideas, one could point out that they both start from considerations belonging to general logic and proceed

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

Metaphysical Deductions

121

to considerations that have some kind of ‘transcendental’ validity. But even in this case, the differences need to be emphasized. While the derivation of the classes of transcendental ideas, in starting from a consideration either of prosyllogisms in general or of three original kinds of prosyllogisms, does seem to begin from purely logical considerations, I have suggested that the table of forms of judgement that provides the starting point for the derivation of the categories already belongs to transcendental logic. It is for this reason that it is more helpful to focus on the differences between the methods of the metaphysical deductions. Highlighting these differences helps us to understand how they can achieve their purpose, namely, tracking what is distinctive about the origins of different root concepts in a way that allows us to distinguish between our faculties of cognition.

h

d i

0 0

9 8 009

2 2 006

bli h d

li

b C

b id

i

i

chapter 4

Transcendental Deductions

Whereas Chapter 3 identified different metaphysical deductions in each main part of the Transcendental Doctrine of Elements, this chapter is dedicated to transcendental deductions. Metaphysical deductions have the purpose of identifying and cataloguing root concepts for the cognition of objects. Additionally, by tracking the origins of these concepts, they contribute to establishing a distinction among faculties. Transcendental deductions, by contrast, have the task of determining the type of validity that can be attributed to root concepts. More specifically, they show that root concepts are ‘objectively valid’. How can we characterize the ‘objective validity’ of a concept? I take it that the main sense in which Kant uses the term is the following: concepts have objective validity when through them we cognize something that really pertains to objects. This means that there exist (or have existed) objects with features that are represented through those concepts. For empirical concepts, this means that there are (or have been) objects that instantiate them. That my concept of red is ‘objectively valid’ means that there are (or have been) things that are red. A priori concepts like ‘space’ and ‘cause’ are objectively valid because objects that are given in experience will present features that are represented through these concepts. Objects of experience will have these features because a priori concepts like space and cause are conditions for cognizing them in the first place.1 Accordingly, Kant submits that space and time are objectively valid because they ‘contain a priori the conditions of the possibility of objects as appearances’ (A89/ B121–2). Similarly, in the context of the transcendental deduction of the categories, he submits that having objective validity means ‘yield[ing] conditions of the possibility of all cognition of objects’ (A89–90/B122). This is the sense of ‘objective validity’ that is central to two of the transcendental 1

To be precise, it is the intuition, not the concept of space, that is such a condition. Nonetheless, as we will see, we can ascribe ‘objective validity’ to the concept of space as well.

122

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

123

deductions that I will analyse in this chapter. As we will see, Kant uses a different understanding of ‘objective validity’ in the context of the transcendental deduction of the transcendental ideas.2 When one thinks of the concept of a ‘transcendental deduction’, the first thing that comes to mind is the transcendental deduction of the categories. The interpretative issues that remain open in this chapter of the first Critique are numerous. Among other things, they include: the relation between the first and the second versions of the transcendental deduction; the meaning of the concept of a ‘subjective’ and an ‘objective’ deduction, to which Kant refers in the A-Preface (Axvi–xvii); the relation between the so-called first and second steps of the B-deduction; whether the transcendental deduction commits Kant to a form of conceptualism; whether the argument is ‘progressive’ or ‘regressive’; whether the argument depends on or contributes to establishing transcendental idealism; and the implications of Kant’s ‘legal’ characterization of the term ‘deduction’ for the structure of his arguments.3 Since the present chapter is only partially dedicated to the transcendental deduction of the categories, I will inevitably remain silent on many of these questions. My central aim is to provide an account of the role that transcendental deductions play in transcendental philosophy, their structure and the kind of results they establish. I will thus highlight the similarities and differences between transcendental deductions, but I will inevitably have to neglect many of the details of Kant’s arguments. In what follows, I will consider the transcendental exposition of space in the B-edition of the Aesthetic,4 the B-version of the transcendental deduction of the categories,5 and the transcendental deduction of the ideas 2

3

4

5

h

d i

Kant sometimes also speaks of the ‘objective reality’ of concepts (see for example B150–1; A155–7/ B194–6). Here, I will only focus on objective validity, for two reasons. First, it is unclear whether there is a straightforward and unquestionable way to distinguish between objective reality and objective validity. Second, the issue of objective validity is more central to determining the aim of the transcendental deductions. For useful surveys of the interpretative problems that are still open, see Conant (2016) and Schulting (2018b). Important contributions on the transcendental deduction of the categories that discuss one or more of these problems include Strawson (1966), Henrich (1969), Henrich (1989), Ameriks (1978), Guyer (1987), Engstrom (1994), Longuenesse (1998), Proops (2003), Allison (2004), Allison (2015), Schulting (2018a) and Møller (2020: Ch. 2). I focus on space for two reasons. First, Kant’s transcendental exposition of space is longer than his transcendental exposition of time. For this reason, there are fewer points that remain implicit in Kant’s line of reasoning. Second, this parallels my focus on space in the section on the Aesthetic in the previous chapter. I favour the B-version of the Aesthetic because it is only there that Kant clearly distinguishes between metaphysical and transcendental expositions. I focus on the B-deduction of the categories because I take the structure of its argument to be clearer. Moreover, this focus is instrumental in view of Chapter 6, where I will distinguish between a ‘positive’ and a ‘negative’ argument regarding the validity of the categories within the B-deduction. I remain neutral concerning the relationship between the A- and the B-version of the deduction.

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

124

in the Appendix to the Transcendental Dialectic. I will reconstruct these arguments with the following issues in mind. First, I will clarify the sense in which Kant’s transcendental deductions express what I have called his moderate methodological conservatism and common sense conservatism. Recall that the former aims to establish that a priori concepts and principles have validity, where the concepts that are at stake in the case of the metaphysical and transcendental deductions are what Kant calls ‘root’ concepts. By contrast, the latter concerns not what a philosophical investigation aims to establish, but rather what it can take for granted. It submits that we can take for granted beliefs that are commonly held as true unless we have good philosophical reason to challenge them. I have characterized the term ‘common sense’ broadly to include scientific propositions that are generally recognized as true. Second, I will consider whether the different transcendental deductions I analyse have a common structure or whether they present different forms of argument. I will suggest that their structures are in fact different. Third, and finally, I will consider whether my thesis that transcendental philosophy only establishes positive results and is not concerned with setting limits to our cognition finds confirmation in the transcendental deductions.

1  The Transcendental Deduction of the Concept of Space Scholars who believe that there is a transcendental deduction of the concepts of space and time in the Aesthetic tend to take one of three main interpretative routes. Hans Vaihinger (1881–1892: Vol. 2, 151–4) and Lorne Falkenstein (1995: 394–5 n12) believe that the transcendental deduction is contained in the metaphysical expositions of space and time. Karl Ameriks (1978), by contrast, reads the transcendental expositions as providing the deduction in question. Finally, Melissa Merritt (2010) has more recently argued that the transcendental deduction of the concepts of space and time is located in the sections of commentary that follow each transcendental exposition. She draws a distinction between the legitimacy and the objective validity of the concepts of space and time and argues that while the transcendental expositions establish the former, the sections of c­ ommentary that follow them establish the latter.6 6

h

d i

0 0

The logical possibilities that have been explored concerning where to locate the transcendental deduction of the concepts of space and time include locating them in the metaphysical expositions, the transcendental expositions or neither. There seems to be at least one possibility that has been neglected: Kant might have meant that the transcendental deduction of the concepts of space and

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

125

As should already be clear at this point, I side with Karl Ameriks’s approach, even though my reconstruction of the argument in the transcendental exposition of space will depart from his. Since I disagree with Vaihinger, Falkenstein and Merritt regarding the location of the transcendental deduction of the concepts of space and time, I owe an explanation of why I think that the arguments they present in support of their views fail. I will not say much about Vaihinger’s and Falkenstein’s views. I hope to have provided sufficient grounds for reading the metaphysical expositions of the concepts of space and time as their metaphysical deductions in the previous chapter. Here, let me just add an additional consideration against a point made by Falkenstein. He argues that a transcendental exposition cannot be considered a transcendental deduction because the former shows that certain synthetic a priori cognitions are only possible when a particular understanding of space and time is presupposed, whereas the latter shows how a concept can relate to objects a priori. He further claims that if the transcendental expositions correspond to anything in the Analytic, it is to the Analytic of Principles, since it is in this section that Kant explains the possibility of certain synthetic a priori judgements (Falkenstein 1995: 394–5 n12). This argument does not consider a passage from the Prolegomena, where Kant speaks of a transcendental deduction of the concepts of space and time and characterizes this deduction as offering an explanation of the possibility of synthetic a priori cognitions in mathematics (4:285). The passage shows that a transcendental deduction can do exactly what the transcendental expositions of the concepts of space and time do. time is in both the metaphysical and the transcendental expositions. There are two considerations that speak in favour of this hypothesis. First, Kant speaks of a transcendental deduction of the concepts of space and time in the A-version of the Critique (A87/B119–20), where no distinction between metaphysical and transcendental expositions has yet been made. Second, in the passage in question, Kant submits that the transcendental deduction of these concepts both traces them to their sources and proves their objective validity. These two tasks correspond closely to the division of labour between metaphysical and transcendental expositions in the B-edition. These considerations notwithstanding, I will argue that it is only the transcendental expositions that should be considered transcendental deductions. In my account, the passage at A87/B119–20 can be explained as follows. When Kant introduces the concept of a metaphysical deduction of the categories in the B-edition, he explicitly links this deduction to the question of the origin of the categories, while the transcendental deduction addresses their objective validity (B159). Given this passage, it seems odd that Kant links the transcendental deduction of the concepts of space and time to both the issue of origin and objective validity. I submit that Kant connected these two issues in the A-edition because at that point the distinction between metaphysical and transcendental expositions hadn’t yet been made. In the B-edition, however, where this distinction had already been established, it would have been more correct to claim that the transcendental deduction of the concepts of space and time only concerned their objective validity.

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

126

Merritt’s proposal is much subtler, and it is indeed plausible if one only considers the transcendental exposition of the concept of space and the following commentary. For this reason, in what follows I will first discuss her view in more detail. As will become clear, I think her approach ultimately fails because it is not applicable to the corresponding sections on the concept of time. I will suggest that a better way to account for the relationship between the transcendental expositions and the commentary sections that follow them is to say that the former establish that the concepts of space and time are objectively valid with respect to appearances. In addition to this, the commentary sections establish that space and time cannot be features of objects in themselves. After my discussion of Merritt’s view, I will present my own interpretation of Kant’s argument in the transcendental exposition of space, with an eye to the issues outlined in the introductory section above. a. Merritt on transcendental expositions and transcendental deductions. In Merritt’s reconstruction of the Transcendental Aesthetic, this part of the Critique proceeds in three steps. The first is accomplished in the metaphysical expositions. These provide a conceptual analysis of the concepts of space and time. The analysis shows that the way in which we use these concepts commits us to the idea that they refer to the pure intuition of space and time (Merritt 2010: 10).7 The second step is achieved in the transcendental expositions. While metaphysical expositions show that our use of the concepts of space and time commits us to the idea that they are pure intuitions, they do not prove that we do in fact have such intuitions.8 They leave open the possibility that our concepts of space and time are nonreferential. The transcendental expositions show that we do in fact have pure intuitions of space and time. In this sense, they prove the legitimacy of our concepts (Merritt 2010: 19–25). Merritt claims that proving this legitimacy nevertheless falls short of showing that our concepts of space and time are objectively valid. In order to show this, one needs to prove that the concepts of space and time are universally applicable to all external objects and to all experiences, respectively. Simply demonstrating that we 7

8

h

d i

0 0

I argued in the previous chapter that the metaphysical expositions cannot simply express analytic knowledge obtained through conceptual analysis. For this reason, I will not take issue with Merritt’s account of the metaphysical expositions here. I disagree with Merritt on this point, but my purpose here is first to reconstruct her view and then to show why it ultimately fails. As should be clear from the previous chapter, I take metaphysical expositions to already show that we have an a priori intuition of space and time, where these intuitions are more fundamental than our concepts because we cannot explain how we are able to make synthetic a priori judgements such as ‘Space and time are singular’ without making reference to them.

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

127

do indeed have pure intuitions of space and time is insufficient to establish this (Merritt 2010: 25). Kant’s argument needs a third step, which expresses its transcendental deduction. This is accomplished in the commentary sections following the transcendental expositions. These commentaries show that the concepts of space and time have objective validity. Because the pure intuitions of space and time to which these concepts refer are ‘constitutive of our receptivity’, they are also ‘necessarily valid of all appearances’ (Merritt 2010: 26). It is no coincidence that Merritt supports her reading with an analysis of the sections on space. Kant’s transcendental exposition of the concept of space builds on the idea that synthetic a priori cognition in geometry would be inexplicable if we were not in possession of a pure intuition of space. Since the transcendental exposition only shows the possibility of cognition in geometry, Merritt can support her reading with a passage from the Prolegomena in which Kant suggests that we need an explanation of why geometry applies to experience (see 4:287 and Merritt 2010: 25–6). If it cannot be assumed that geometry applies to objects of experience, it seems to follow that even if we can prove that our concept of space is a correct description of the space in which geometry operates, we will not have proven that the concept of space necessarily applies to all outer objects. Even though Merritt’s reading is both subtle and supported by Kant’s claims about the relationship between pure mathematics and its applications, her interpretation ultimately fails. A first problem with Merritt’s view is that in the commentary that follows the transcendental exposition of space, Kant does not suggest that we need to prove that the concept of space that applies to geometry necessarily applies to all outer objects as well. Rather, in the transcendental exposition of space Kant already treats space as ‘constitutive of our receptivity’, to use Merritt’s words (Merritt 2010: 26). Accordingly, in that context, Kant characterizes space as the form of outer sense (B41). The biggest problem for Merritt’s reading, however, comes from the transcendental exposition of time and the commentary that follows. Recall that Merritt believes that the transcendental expositions only prove the legitimacy of our concepts of space and time. This means that they prove that we have pure intuitions that correspond to our concepts of space and time. This does not imply that these concepts necessarily apply to empirical objects, however. It seems difficult to maintain that the transcendental exposition of time does not have implications for how we experience objects in empirical intuition. Kant starts from our concepts of alteration

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

128

and motion and argues that these concepts cannot be explained if we do not assume that we have a pure intuition of time (B48–9). It would be highly implausible to maintain that Kant does not take what he says on the concepts of alteration and motion to have implications for how we experience alteration and motion in empirical intuition. Accordingly, Kant concludes the transcendental exposition of time by saying that his account of time explains the possibility of a ‘general theory of motion’ (B49), which clearly must apply to empirical objects. Finally, a similar point can be made regarding the closing sentence of the third argument of the metaphysical exposition of time. This argument actually belongs to the transcendental exposition of time in Kant’s account (B48). The sentence reads: ‘These principles [that is, “axioms” of time such as the judgement “Different times are not simultaneous, but successive”] are valid as rules under which alone experiences are possible at all, and instruct us prior to them, not through it’ (A32/B47). If a certain concept of time (according to which time is originally a pure intuition) is the only one that can explain how these principles are possible, and if these principles are conditions for experiencing objects, it seems that the transcendental exposition of time does indeed prove that our concept of time is empirically applicable and, consequently, objectively valid. How can we account for the relationship between the transcendental expositions and the commentary sections that follow them? I suggest that a better way to account for their relationship is to say that the transcendental expositions only establish that our concepts of space and time are objectively valid with respect to appearances, because appearances necessarily agree with space and time as described in those concepts. The commentary sections add the negative point that our concepts of space and time cannot also describe features that objects have as things in themselves. This approach finds confirmation when we take three passages into consideration. The first is contained in the transcendental exposition of space, while the second and the third are in the commentary section. In the first passage, Kant equates space with the form of outer sense: ‘Now how can an outer intuition inhabit the mind that precedes the objects themselves, and in which the concept of the latter can be determined a priori? Obviously not otherwise than insofar as it has its seat merely in the subject, as its formal constitution for being affected by objects and thereby acquiring immediate representation, i.e., intuition, of them, thus only as the form of outer sense in general’ (B41). The second passage is the first sentence in the commentary and reads: ‘Space represents no property at all of any things in themselves

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

129

nor any relation of them to each other, i.e., no determination of them that attaches to objects themselves and that would remain even if one were to abstract from all subjective conditions of intuition’ (A26/B42). Finally, the third passage states: ‘Space is nothing other than merely the form of all appearances of outer sense, i.e., the subjective condition of sensibility, under which alone outer intuition is possible for us’ (A26/B42, emphasis mine). Starting from the assumption that the transcendental ­exposition must establish something concerning both the concept and the ­intuition of space, it is clear that in the first passage it is only the objective ­validity of these representations with respect to objects of outer sense that is at stake. The passage does not rule out, at least explicitly, that the concept of space might also be valid for objects in themselves. It is only the ­second and third passages that explicitly make this point. This ­distinction lines up well with my characterization of ­transcendental philosophy and the critique of pure reason. Recall that I claim that the former only ­establishes positive results, while it is the latter that is charged with setting limits to our cognition. Accordingly, the ­transcendental expositions of the concepts of space and time, which I read as their transcendental deductions, belong to ­transcendental ­philosophy because they only establish a positive result concerning the objective validity of these concepts (and the corresponding intuitions) with respect to appearances. Since I focus on transcendental d ­ eductions in this chapter, I will only consider the transcendental expositions, and more precisely the transcendental exposition of space. I will take into account how the commentaries set limits on the validity of our ­concepts of space and time in the next chapter. b. Kant’s argument in the transcendental exposition of space. Let me recall that in my account, the metaphysical expositions of space and time establish that we do have pure concepts of space and time and that these concepts depend on the pure intuitions of space and time. However, the metaphysical expositions do not establish that these concepts or the intuitions on which they depend are ‘objectively valid’, in the sense that through these representations we cognize something that really pertains to objects. The task of the transcendental expositions is to prove that our concepts and intuitions of space and time have this validity. How can they achieve this result? From this point on, I will focus on the exposition of space.

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

130

As many commentators have emphasized, Kant’s argument builds on the presence of synthetic a priori propositions in geometry. The first thing we must determine is what kind of validity Kant ascribes to geometry. As we saw, Merritt bases her claim that the transcendental exposition only establishes the legitimacy of the concept of space on the assumption that Kant’s argument does not yet take geometry to necessarily apply to empirical objects (Merritt 2010: 25–6). This contrasts, however, with the conclusion Kant wants to establish. Kant takes the transcendental exposition of the concept of space to already determine that space is the form of outer sense and that, as such, it necessarily applies to all outer appearances (B41). Since the argument must establish this strong result, it is more plausible to read it as already granting that geometry necessarily applies to outer appearances. Starting from this assumption, the argument in the transcendental exposition can be reconstructed as proceeding in two fundamental steps. The first occurs in the following passage: Geometry is a science that determines the properties of space synthetically and yet a priori. What then must the representation of space be for such a cognition of it to be 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., prior to all perception of an object, thus it must be pure, not empirical intuition. For geometrical propositions are all apodictic, i.e., combined with consciousness of their necessity […]. (B40–1)

Kant takes into consideration three possible explanations for how we can arrive at synthetic a priori propositions in geometry. Two of these fail, while the third succeeds. The explanations that fail contend that synthetic a priori propositions depend either on the content of geometrical concepts or on empirical intuitions, respectively. The former account fails because it does not explain how geometrical propositions can be synthetic. The latter fails because it does not explain how geometrical propositions can be apodictic. The successful explanation submits that geometrical propositions depend on a pure intuition of space. Kant assumes that these three explanations exhaust the scope of possible explanations. Since the third is the only one that accounts for the possibility of synthetic a priori cognition in geometry, it is to be preferred. Can this argument be used to establish that the concept and intuition of space identified in the metaphysical exposition have objective validity? It can do so on the assumption that the pure intuition of space that lies at the basis of geometrical synthetic a priori propositions is the same intuition

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

131

on which the concept of space analysed in the metaphysical exposition depends. The argument can be reconstructed as follows: (1) There are synthetic a priori propositions in geometry. (2) The possibility of these propositions can be explained in three ways: they depend either on the content of geometrical concepts, or on empirical intuitions, or on a pure intuition of space. (3) The first two explanations fail. (4) Synthetic a priori propositions in geometry must depend on a pure intuition of space (from 1, 2 and 3). (5) The pure intuition of space on which geometry depends is the same intuition on which our concept of space depends. (6) A representation on which true synthetic a priori propositions depend is objectively valid. (7) The pure intuition of space on which geometry depends is objectively valid (from 4 and 6). (8) The pure intuition of space on which our concept of space depends is objectively valid (from 4, 5 and 6). Let me clarify a few points. First, how can we assume in (5) that the pure intuition of space on which geometry depends is the same as that on which our concept of space depends? The thought might simply be that it would be absurd to think that we can have different pure intuitions of space. After all, both the intuition on which our concept of space is based and the intuition at the basis of geometry represent space as singular. It would be odd to claim that we can have different pure intuitions of the same singular space. Further support for the identity of these intuitions comes from the analogy Kant draws between the way in which we arrive at synthetic a priori propositions starting from our concept of space, such as ‘Space is singular’, and the way in which we arrive at synthetic a priori propositions in geometry (A25/B39). Since both kinds of synthetic a priori propositions appeal to a pure intuition of space in a similar way, the best way to account for the relationship between them is to say that they depend on the same pure intuition of space.9 9

h

d i

Clearly, I take the claim that space is singular to mean that it is one. Falkenstein (1995: 219–22) acknowledges that Kant frames his argument by appealing to the ‘oneness’ of space, but he argues that the core of the argument insists on the fact that representing space involves representing a particular (which contains other particulars). This agrees with what I say in Chapter 3 regarding the importance of Kant’s account of how we represent parts–whole relationships in space. However, we must keep in mind that Kant sees the issue of how we represent parts–whole relationships in space as fundamentally related to the issue of how we see that space is necessarily ‘one’.

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

132

Second, in what sense can we speak of objective validity in the context of this argument? I submitted that ‘objective validity’ in the context of the transcendental exposition of space equates to the notion that through a representation we cognize something that really pertains to objects. Kant here only speaks of synthetic a priori propositions in geometry and does not take into consideration, at least explicitly, the necessary validity of geometry for all external objects. However, Kant thinks that in order to attribute objective validity to mathematical propositions, those propositions must be valid for empirical objects as well. Accordingly, Kant writes in the B-transcendental deduction: ‘Consequently all mathematical concepts are not by themselves cognitions, except insofar as one presupposes that there are things that can be presented to us only in accordance with the form of that pure sensible intuition’ (B147). Therefore, since the first step assumes the objective validity of geometrical propositions, it ‘presupposes’ that geometrical synthetic a priori propositions have validity for empirical objects. The second step of the transcendental exposition of space makes this assumption explicit. I have already quoted the passage in which we find the second step above, but let me quote it again: Now how can an outer intuition inhabit the mind that precedes the objects themselves, and in which the concept of the latter can be determined a priori? Obviously not otherwise than insofar as it has its seat merely in the subject, as its formal constitution for being affected by objects and thereby acquiring immediate representation, i.e., intuition, of them, thus only as the form of outer sense in general. (B41)

Notice that the passage moves from the claim that we must assume a pure intuition of space in order to account for geometry (which was the result of the first step) to the claim that the kind of cognition we can obtain through this pure intuition presupposes that space is also the form of outer intuition in general, which of course includes empirical outer intuitions. Since Kant does not argue for the transition from geometrical space to outer intuition in general, he assumes that the synthetic a priori cognitions we obtain through the pure intuition of space in geometry necessarily apply to all outer objects. How should we account for the division of labour between the two steps? The first step of the argument shows that synthetic a priori cognitions in geometry are only explainable by assuming a pure intuition of space. By contrast, the second step illustrates how the synthetic a priori cognitions we obtain through the pure intuition of space can set constraints on all outer intuition, including intuition of empirical objects. In order to explain how this is possible, Kant submits that our pure intuition of space must be the form of all outer intuitions. As such, space is

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

133

something that relates to the subjective constitution of human beings and constrains how objects must appear to them in spatial relationships. Only on this assumption can we explain how the a priori cognition of properties of space that we obtain through pure intuition must constrain all outer intuitions. The second step of the argument proceeds as follows: (9) Through our pure intuition of space we obtain cognitions that constrain the spatial properties of all outer objects. (10) This can only be explained if our pure intuition of space is the form of all outer intuitions. (11) Our pure intuition of space must be the form of all outer intuitions (from 9 and 10). As I said, the second step makes explicit that the objective validity of the pure intuition of space extends to all objects of outer intuition in the sense that through that intuition we are able to cognize something that really pertains to those objects. One first thing to note is that the first and second steps only determine the objective validity of the pure intuition of space and not the concept of space. Kant seems to assume that the argument that establishes the objective validity of the pure intuition of space also determines that our concept of space is objectively valid. But why should this be the case? The thought might be the following: through our pure intuition of space we cognize spatial properties of objects of outer sense. This is why our intuition of space is objectively valid. Through our concept of space we attribute some properties to space, that is, those that are analytically derivable from our concept. On the basis of that concept, we can also indirectly attribute some properties to objects of outer sense. But can we say that we cognize something about these objects when we use our concept in this way? I think that Kant’s idea is that we do cognize something about objects of outer sense through our concept of space. This is possible because the concept of space derives from our pure intuition of space. This means, first, that we would not have any concept of space if we had no intuition of space. Second, it means that all the properties that we can attribute to objects of outer sense through our concept of space are first of all properties of those objects that we cognize through our pure intuition of space. In this sense, we can say that the concept of space is only an expression of our pure intuition of space. But since when we attribute spatial properties to objects through our concept of space we are only expressing cognitions that we have first of all through our pure intuition of space, we can say that through our concept of space we can cognize something that really pertains to objects.

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

134

A second issue that must be emphasized concerns the relationship between the pure intuition of space and space as the form of outer intuition. (10) and (11) simply equate the two, but in view of what Kant says later in the B-transcendental deduction of the categories (see B160–1), it is not clear that this equation can be made. I will discuss the relationship between space as a form of intuition and the pure intuition of space in the context of my analysis of the transcendental deduction of the categories. For now, I will remain neutral regarding whether space as pure intuition and space as a form of intuition can be equated. For this reason, I will simply say that the pure intuition of space is identical to, or depends on, space as the form of outer intuition. In concluding my analysis of the argument in the transcendental exposition of space, let me observe that it is the second step of Kant’s transcendental exposition that establishes the transcendental ideality of space. Kant submits that space is the form of outer sense and that this form ‘has its seat merely in the subject’ (B41). Kant’s claim already entails a distinction between outer appearances (objects as far as they are intuited through the form of outer sense) and things in themselves (objects as seen independently of this intuition).10 It also entails that the spatial properties of objects that we do cognize are properties that objects have only as they are intuited by us. In the transcendental exposition of space, however, Kant does not rule out, at least explicitly, that objects considered in themselves might also have ‘spatial’ properties that somehow correspond to the spatial properties they acquire by appearing to us. What is ruled out is that in cognizing the spatial properties of objects, we might be cognizing features that objects have independently of our outer intuition. c. The method of the transcendental expositions. At this point, I wish to connect my reconstruction of the transcendental exposition of space to the issues listed at the beginning of this chapter. First, how is my account of the transcendental exposition of space connected to what I have called methodological and common sense conservatism? As far as the former is concerned, the transcendental exposition of space is methodologically conservative because it aims to establish that our concept and pure intuition of space have objective validity with respect to appearances, where objective validity here means that through these representations we cognize 10

h

d i

0 0

I here use a formulation that leans towards a ‘two aspects’ or ‘two perspectives’ reading of transcendental idealism, but I take my reconstruction to be compatible with a ‘two worlds’ reading as well. I wish to remain neutral on this point.

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

135

something that really pertains to objects (as appearances). Since Kant’s proof that these representations are objectively valid is based on the claim that our pure intuition of space is identical to, or depends on, space as the form of outer sense, it certainly implies a fundamental revision of our common way of understanding what space is. Therefore, Kant’s argument is methodologically conservative to the extent that it confirms that an a priori concept and a corresponding ‘root’ intuition have objective validity, but it is only moderately conservative to the extent that it involves a revision of our common way of characterizing space. Turning to common sense conservatism, Kant’s transcendental exposition of space clearly takes for granted propositions that were commonly held to be true by scientists and scholars of his time, such as ‘Only one straight line lie[s] between two points’ (A24),11 or ‘Space has only three dimensions’ (B41). His first move is to claim that these propositions must be synthetic a priori. After that, he argues that they can only be explained by appealing to a pure intuition of space. The second issue I want to discuss is the form of argument that Kant adopts in the transcendental exposition of space. There are various elements that suggest that Kant’s argument is best understood as an inference to the best explanation, where the explanations that are ruled out are not simply worse but completely unsuccessful.12 Kant explicitly considers three possible accounts of synthetic a priori judgements in geometry in claim (2) in the first step of the transcendental exposition. Moreover, claim (10) in the second step submits that identifying our pure intuition of space with the form of outer intuition is the only way to explain how we can have a priori cognition of features of space that constrain all outer objects. That the transcendental exposition of space is a sort of inference to the best explanation is confirmed in two other passages. The first introduces Kant’s general characterization of a transcendental exposition. This is an ‘explanation of a concept as a principle from which insight into the possibility of other synthetic a priori cognitions can be gained. For this aim it is required (1) that such cognitions actually flow from the given concept, and (2) that these cognitions are only possible under the presupposition of a given way of explaining this concept’ (B40). From what Kant says under (2), it is clear that transcendental expositions aim to establish that a certain 11 12

h

d i

Kant mentions this proposition in the argument of the A-exposition of space, which on his account corresponds to his transcendental exposition in the B-edition. In this sense, the inference might be called an inference to the only explanation. In recent literature in the philosophy of science, this kind of inference is sometimes regarded as a limit case of an inference to the best explanation (see A. Bird 2007; A. Bird 2010; Niiniluoto 2018).

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

136

explanation of the possibility of some synthetic a priori claims is the only one that works. Kant once again emphasizes his focus on competing explanations in the closing sentence of the transcendental exposition of space, which reads: ‘Any kind of explanation that does not accomplish this [rendering the possibility of geometry as a synthetic a priori cognition comprehensible], even if it appears to have some similarity with it, can most surely be distinguished from it by means of this characteristic’ (B41).13 To conclude, let me again emphasize that my reading of the transcendental exposition of space aligns with my characterization of transcendental philosophy as a discipline that only establishes positive results regarding the objective validity of root concepts for the cognition of objects (although as I argued in Chapter 3, in the case of the concept of space, it is better to speak of the concept of a root representation for the cognition of objects). More precisely, the argument establishes that our concept of space and the pure intuition on which it depends are objectively valid with respect to appearances. It also establishes that the spatial properties of objects we do cognize are properties that objects have only as they are intuited by us. This means that in cognizing these spatial properties, we are not cognizing those spatial properties that the objects might still have independently of our intuition – properties that may nevertheless somehow agree with the spatial properties we do cognize. In other words, the argument establishes that what we directly cognize through our concept and intuition of space are only properties that objects have as appearances. If we could present an argument according to which objects in themselves have spatial properties that necessarily correspond to the spatial properties we cognize through our pure intuition and concept of space, however, this argument would grant indirect objective validity to our concept and pure intuition of space with respect to objects in themselves. As I suggested above, establishing that this is impossible is the task of the commentary that follows the transcendental exposition of space, which I will consider in the next chapter.

2  The B-Transcendental Deduction of the Categories Similar to the transcendental deductions of the concepts of space and time, the transcendental deduction of the categories is charged with showing the kind of validity held by the root concepts identified in the corresponding 13

h

d i

0 0

Recently, Thomas Vinci (2015: Ch. 4) has also claimed that the transcendental expositions are inferences to the best explanation. He also claims that the transcendental deduction of the categories should be read as one such inference (2015: Ch. 5). As will become clear, I disagree with him on the latter point.

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

137

metaphysical deduction. As in the case of the transcendental expositions in the Aesthetic, objective validity will here mean that through the categories we cognize something that really pertains to objects. Kant introduces the problem of the transcendental deduction of the categories by saying that this should show ‘the way in which concepts can relate to objects a priori’ (A85/ B117). He also emphasizes a difficulty that is unique to the transcendental deduction of the categories. Kant submits that it was relatively ‘easy’ to prove that the concepts of space and time are objectively valid once he proved that these concepts depend on space and time as forms of intuition. Since all objects appearing to us through intuition must conform to these forms, and since the concepts of space and time only express cognitions we can obtain through space and time as these forms, these concepts necessarily apply to all objects of intuition and are thus objectively valid (A89/B121–2). But the categories are neither forms of intuition nor simply derived from these forms, like the concepts of space and time. Rather, the metaphysical deduction has shown that they are the fundamental ways in which we order the manifold of intuition in agreement with the fundamental kinds of judgements. This means, at least prima facie, that the categories do not constrain all objects of intuition in the same way that the forms of intuition and the corresponding concepts do. In other words, it seems possible that objects can be given to us in intuition without the intervention of the categories (A89–90/B122). It is at this point that ‘conceptualist’ and ‘non-conceptualist’ readings of the transcendental deduction part ways. The former usually argue that when Kant introduces the possibility of having intuitions without the categories, he is simply playing with a scenario the possibility of which he in fact rejects.14 The latter argue that Kant does believe that we can have intuitions of objects without the categories. Accordingly, the proof of the validity of the categories does not show that they are necessary for all intuitions of objects, but rather that they are necessary for ‘empirical cognition’, where this implies more than simple intuition, such as the use of empirical concepts (see Allais 2015: Ch. 11 for a recent view of this kind).15 I have mentioned this dispute between conceptualist and non-­conceptualist readings because it helps us to identify two desiderata that a successful interpretation of the transcendental deduction of the categories should meet. The first concerns Kant’s characterization of the B-transcendental deduction as 14 15

h

d i

For relevant conceptualist readings, see Longuenesse (1998), Ginsborg (2008), Grüne (2009), Land (2015), Conant (2017) and Williams (2018). Other non-conceptualist readings include those by Hanna (2011a), Hanna (2011b), Golob (2016) and Schulting (2018a).

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

138

proceeding in two steps (B144–5).16 The second concerns how we should account for the unity of space and time considered as wholes. I will begin by showing how the dispute between conceptualists and non-conceptualists clearly reveals these desiderata, and I will clarify how my reading, which can be described as a weak form of non-conceptualism, meets them. I will then analyse the two steps of the B-deduction in turn. I will conclude by considering the issues I listed in the introductory section. Before proceeding with my discussion, I would like to call attention to one point at the outset. It is a central tenet of my reading of transcendental deductions that they only establish positive results and are not concerned with setting limits to cognition. It is clear that in the chapter on the transcendental deduction of the categories, Kant also establishes negative results concerning the limits of our cognition (see, for example, B165–6). It is not my aim to challenge this. Rather, my aim is to show that an argument that only establishes positive results regarding the validity of the categories can be singled out in the chapter dedicated to the transcendental deduction. It is this positive argument that constitutes the transcendental deduction of the categories belonging to transcendental philosophy. I will discuss the relationship between the positive and negative results established within the sections of the transcendental deduction of the categories in Chapter 6. a. Conceptualists, non-conceptualists and two desiderata for interpretations of the B-deduction. In § 21 of the B-deduction, Kant stresses that what he accomplished up to § 20 was only ‘the beginning of a deduction’ (B144). He also submits that it is in § 26 that ‘the aim of the deduction will first be fully attained’ (B145). Kant distinguishes between the two steps by saying that the first abstracts from our particular way of intuiting objects (B144). This means that what it establishes would also be valid for beings who have an understanding similar to ours, namely, one that depends on sensibility to receive objects, but who have a different form of sensibility (B145–6; see also B138–9 and B150). By contrast, the second step takes into account the particular way in which we intuit objects in space and time (B144–5). Conceptualists have a fairly straightforward way of accounting for the two steps. Recall that they claim that the categories are necessary for ­having intuitions. One key move in the B-deduction is to claim that the categories are necessary for providing unity to the manifold of intuitions. 16

h

d i

0 0

Famously, Henrich (1969) was the first to insist on the relevance of this distinction.

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

139

Conceptualists have different ways of accounting for the first step. They sometimes claim that this step only establishes that the categories are ­necessary for bringing different intuitions under the unity of a concept (see, for example, Williams 2018). Alternatively, they stress that the first step only applies to intuitions that already have unity (see Henrich 1969). What is interesting for my purposes is that conceptualists can e­ asily account for what is achieved in the second step; they can say that the first step does not rule out that we can have intuitions without the ­categories and that it is the task of the second step to do so. The ­typical ­conceptualist move is to claim that the second step proves that the c­ ategories are ­necessary for representing space and time as unitary wholes containing a manifold, where these representations of space and time are in turn presupposed by all outer intuitions and all intuitions, respectively. This implies that every intuition is determined by the categories because they fundamentally determine the a priori forms of intuition in a way that precedes all possible intuitions. A conceptualist reading of the B-deduction is appealing for two main reasons. First, it can easily account for what is gained in the second step. Second, it has a relatively straightforward way of stressing the objective validity of the categories: they simply must be used to obtain intuitions in the first place. This does not mean that conceptualists do not have their own problems. These arise when we try to account for the relationship between the B-deduction and the Transcendental Aesthetic; as we have seen, conceptualists usually claim that it is through a categorial synthesis that space and time obtain their unity as wholes. This seems problematic, however, when one takes into account how Kant appeals to specific features of the unity of space and time in the Aesthetic. Recall my reconstruction of the singularity argument for space in the previous chapter. The argument proves that space must originally be represented through a pure intuition because this is the only way to explain how we can cognize a priori that every particular space is a limitation of a singular encompassing space. In this way, the argument appeals to the particular way in which we represent the unity of space as a whole. The whole of space precedes its parts, which are obtained through limitations. Claiming that the unity of space is obtained through a categorial synthesis seems to undermine Kant’s strategy. On the one hand, it undermines Kant’s claim that the unity of space and time displays something specific.17 If this unity is obtained through 17

h

d i

Messina (2014), McLear (2015) and Onof and Schulting (2015) insist on the special character of the unity of space and time.

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

140

the categories, which are concepts, it must share characteristics with other forms of conceptual unity. On the other hand, it undermines Kant’s argument in the Aesthetic. If there is nothing specific about the unity of space and time, this unity cannot be used to prove that they must originally be represented through pure intuitions. Non-conceptualists clearly have an advantage over conceptualists here. On their account, the B-deduction does not establish that the categories are necessary for having intuitions. Accordingly, they are happy to concede that space and time are unitary wholes independently of the categories and that we can have intuitions within these wholes without applying the categories. It is in dealing with the two-step structure of the B-deduction that they encounter difficulties, however. As we saw, they usually claim that the objective validity of the categories is proven by showing that they are necessary for having cognitions of objects in a stronger sense, where either the use of empirical concepts (see Allais 2015: Ch. 11) or the capacity to recognize objects as objects (see Schulting 2018a: Ch. 11) is already at stake. Clearly, this view makes it difficult to claim that the second step is charged with extending the validity of the categories. Rather, the non-conceptualist seems to be committed to the view that the same kind of cognition is at stake in both the first and the second step. How do non-conceptualists account for the second step? A typical move is to suggest that it is only explanatory and does not extend the validity of the categories (see Schulting 2018a: 296–7). Recall that it is only in the second step that our particular forms of intuition are taken into account. Non-conceptualists typically claim that the first step already proves that the categories are necessary for having cognitions of objects in the stronger sense just mentioned. Starting from this result, the explanatory role of the second step consists in showing how, exactly, the categories perform their role in beings with the forms of intuition we have. This approach is plausible, but it goes against Kant’s claim that it is only in the second step that the purpose of the B-deduction is achieved. To sum up, the main difficulty for conceptualist readings of the B-deduction is that they cannot easily account for Kant’s appeal to the specific form of unity that space and time have in his proofs that our original representations of space and time must be pure intuitions. The main difficulty for the non-conceptualist, by contrast, is to account for the twostep structure of the B-deduction in a way that does not make the second step appear trivial. My first purpose in introducing this dispute was to clearly identify two desiderata for any interpretation of Kant’s argument: a

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

141

successful interpretation must be able both to account for Kant’s appeal to the unity of space and time in the Aesthetic and to non-trivially explain the two-step structure of the B-deduction. How does my reading meet these desiderata? Let’s start with the twostep structure of the argument. I will argue that the second step of the B-deduction is not only explanatory but makes a substantial point. On my reading, while the first step proves that the categories are necessary for having unitary representations of the manifold of intuition, it does not rule out their being used to synthesize an empirical manifold of intuition in the absence of a corresponding synthesis of a pure manifold. The second step establishes that to any synthesis of an empirical manifold of intuition there must correspond a synthesis of a pure manifold. It is the omnipresence of this latter synthesis that explains how the categories can provide a priori cognition that constrains possible experience. Turning to the other desideratum, this synthesis of a pure manifold of intuition concerns only the unity of particular spaces and times, not the unity of space and time as wholes, which means that the categories are not responsible for this unity.18 How can my reading be classified? I grant that we may have intuitions without the categories in some sense and that the categories are not responsible for the unity of space and time as wholes. Accordingly, my reading can broadly be categorized as non-conceptualist. It is a weak form of nonconceptualism, however, because it argues not that we do have intuitions without the categories but only that we cannot rule out that we don’t. Let us now turn to my reconstruction of the two steps. b. The first step of the B-deduction. Both steps of the B-deduction are complex. In the sections in which they appear, Kant often brings up issues that are obscure and whose connection to the main line of argument is unclear: the distinction between the objective and the subjective unity of apperception in § 18 and § 19 and Kant’s discussion of figurative synthesis in § 24 are key examples in this regard. Since I am dedicating only a portion of one chapter to the B-deduction, it is impossible to give due attention to all of these issues and to determine how they connect to the main line of reasoning.19 What I will do is focus on Kant’s summary 18

19

h

d i

I do believe, however, that it is only through cognizing the unity of particular spaces and times (which requires the categories) that we can cognize the specificities of the unity of space and time as wholes. For recent book-length discussions of the transcendental deduction of the categories, see Allison (2015) and Schulting (2018a).

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

142

reconstructions of the two steps in § 20 and § 26. I will only refer to other sections when they illuminate Kant’s reconstructions. Let us start with Kant’s summary of the first step in § 20 (B143), which proceeds as follows: (1) The manifold of intuition necessarily belongs to the synthetic unity of apperception because through it the unity of intuition first becomes possible. (2) The understanding brings the manifold of representations (either intuitions or concepts) under the synthetic unity of apperception through the logical functions of judgement. (3) Every manifold that belongs to a unitary intuition must be determined through (at least) one of the logical functions of judgement (from 1 and 2). (4) The categories are nothing other than these logical functions of judgement. (5) Every manifold that belongs to a unitary intuition must be determined through (at least) one of the categories (from 3 and 4). What is the synthetic unity of apperception that Kant mentions in (1)? In § 16, Kant describes it as the activity through which I combine different representations in my consciousness and recognize them as mine (B133–4). What does it mean to say that the manifold of intuition necessarily belongs to the synthetic unity of apperception? Kant is not saying that the manifold of intuition must actually be brought under the synthetic unity of apperception. He is rather claiming that because all of my representations must be recognizable as mine, they must in principle be combinable in a single consciousness. This is the point Kant wants to make when he famously writes that ‘[t]he I think must be able to accompany all my representations’ (B131). This expresses the idea that the manifold of intuition must stand under a principle of synthesis because all intuitions, as mine, can at least be united in my consciousness of them. In (1), Kant adds the additional point that the activity through which I recognize different representations as mine is a condition for finding unity in the manifold of intuitions. What is the unity that is here at stake? It is a unity that results from an activity of synthesis and where I recognize the manifold that composes it. This clarification is important because it points out that this part of the argument cannot establish that the unity of space and time as wholes depends on the unity of apperception or the categories. As we have already seen there seems to be something relevantly different in the unity of space and time as wholes, which means that it is not clear whether

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

143

we can consider this unity as the result of a synthesis. Of course, one could add that it is also not clear whether the unity of particular spaces and times is the result of a synthesis. However, while there are textual reasons to maintain that the unity of space and time as wholes is not the result of an intellectual synthesis – namely, the singularity arguments for space and time – the same cannot be said for the unity of particular spaces and times. Therefore, I take it that the condition for the ‘synthetic’ unity of intuitions identified in step (1) extends to all cases in which we recognize a unity in intuition except for the case of the unity of space and time as wholes. At this point of the argument, it cannot be decided whether this unity is the result of a synthesis or not. Notice that in (3) Kant refers not to a manifold of intuition in general but to a manifold that belongs to a unitary intuition. I take this to mean that Kant is considering intuitions that are actually brought under the synthetic unity of apperception: intuitions that, rather than merely potentially possessing the unity that comes with the latter, actually are unitary representations. This highlights an important shift of perspective from (1) to (3). While the former only determines that, necessarily, it is possible for all intuitions to be united through the synthetic unity of apperception, the latter submits that, necessarily, all intuitions that actually possess unity stand under the synthetic unity of apperception (and consequently the logical functions of judgement). The focus on intuitions that actually possess unity is carried over to (5).20 Therefore, the argument establishes that intuitions that possess unity must stand under the categories. Since, necessarily, all manifolds of intuition can be brought under the synthetic unity of apperception, however, it seems to follow that, necessarily, all manifolds of intuition can be brought under the categories. This is compatible with the view that we might have intuitions that are not actually brought under the synthetic unity of apperception and in this sense do not require the application of the categories (even though, necessarily, they are combinable through the categories). (4) states that the categories simply are the logical functions of judgements. This aligns with my analysis of the metaphysical deduction of the categories in the previous chapter. In that context, I characterized the categories as the fundamental ways in which we order the manifold of intuition in a way that provides the necessary grounding for the 20

h

d i

Henrich (1969) famously argued that the first step of the B-deduction considers only intuitions that already have unity. Similarly, I argue that in (3) and (5) Kant makes claims that only apply to unitary intuitions. Unlike Henrich, however, I claim, first, that (1) applies to all intuitions and, second, that the second step of the B-deduction only discusses unitary intuitions.

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

144

connections between concepts that we establish through the fundamental forms of judgement. These fundamental forms are those listed in Kant’s table of the forms of judgement. They correspond to what Kant calls ‘logical functions’ in (4). Therefore, while the argument in the metaphysical deduction establishes that we rely on the categories as fundamental ways of ordering the manifold of intuition according to the forms of judgement, the first step of the transcendental deduction establishes that these categories must be used in every act of giving unity to the manifold of intuition. Since both the metaphysical deduction and this step of the transcendental deduction draw on the relationship between forms of judgement and categories, one might wonder what kind of cognition is at stake in the present argument. Forms of judgement have to do with the unity of concepts (unlike the categories, since the categories are just those forms in another guise). Accordingly, one might claim that the cognition that is at stake is one that involves both concepts (united through the forms of judgement) and intuitions (united through the categories). This view finds support in various passages from § 17 to § 19. In that context, Kant notes that under ‘judgement’ he only includes those acts in which, by establishing a connection between representations, we determine something pertaining to an object. This excludes those cases in which a connection among representations only concerns sensations (see B142). Kant defines an object as ‘that in the concept of which the manifold of a given intuition is united’ (B137). Moreover, he often speaks of a manifold of intuition that is united in the concept of an object (B139). These passages suggest that the cognition that is at stake is one that involves both concepts (over and above the categories) and intuitions.21 Since Kant only speaks of the unity of a manifold of intuition in § 20, however, I will assume that what is at stake is a unity of intuitions that does not necessarily require concepts (except for the categories, of course). In (5), Kant concludes that the categories are necessary for giving unity to a manifold of intuition. But how should this manifold be understood in the context of the first step? It is important to keep in mind that the first step abstracts from our human forms of intuition (B144). Therefore, it disregards the fact that we have a priori forms of intuition and that these 21

h

d i

0 0

Moreover, the passages in question can support both those conceptualist readings according to which the first step of the B-deduction only establishes the necessity of the categories for cognitions that involve both concepts and intuitions (see Williams 2018; I provide a reading along these lines in Gava 2015) and those non-conceptualist readings according to which this kind of cognition is at stake in the whole B-deduction (see Allais 2015: Ch. 11; Schulting 2018a: Ch. 11).

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

145

are space and time. The results it establishes should be valid for any understanding that, like ours, depends on sensibility for receiving a manifold that needs to be combined (B145–6; see also B138–9 and B150). What consequences does this have for the results obtained in (5)? Clearly, as many interpreters have noted, the first step cannot determine anything specific about how the categories synthesize a given manifold. Rather, it establishes that the categories are necessary for providing unity to any sensible manifold, where this possibly includes the sensible manifold experienced by beings with a different form of sensibility. What interpreters have failed to see is that this has an important consequence for the economy of the B-deduction, since the first step cannot establish a relationship between how the categories are used to synthesize empirical and pure manifolds. In other words, the first step establishes that the categories are necessary for giving any unity to a manifold of intuitions, whether that manifold is empirical or pure. This leaves open the possibility that the categories can be used to give unity to an empirical manifold without the need for a corresponding synthesis of a pure manifold. On my account, this result is sufficient to establish that the categories have objective validity, since they are required for giving unity to intuition. But what it cannot explain is how the categories provide a priori cognition that constrains possible experience. What I mean by this is that the categories can provide this a priori cognition by determining what relationships in space and time must look like. In order to do this, they must perform an a priori synthesis on a pure manifold in space and time. Since the first step does not establish that a categorial synthesis of a pure manifold in time, space or both must correspond to any categorial synthesis of an empirical manifold, it cannot account for how the categories can provide a priori cognition. What does it mean, however, to say that the first step can still establish that the categories have objective validity? Recall that objective validity here means that through a representation we cognize something that really pertains to objects. Since the first step shows that the categories must be used in providing any unity to intuition, and since unitary representations are fundamental to cognition, through the categories we cognize something that really pertains to objects (of intuition). The categories do not have any determinate content when they are considered independently of the way in which they determine space and time, however. Since the first step abstracts from this determination of space and time, the categories, as considered in it, cannot involve substantial a priori cognition concerning possible experience. In this sense, it is only the second step that shows how they can provide this substantial cognition.

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

146

c. The second step of the B-deduction. As we know, what characterizes the second step of the B-deduction is that it takes into consideration the fact that we have particular a priori forms of intuition, that is, space and time. How can the results of the first step be reinterpreted once we attend to these forms? The first step established that the categories are necessary for providing any unity to a pure manifold in space and time (for example in mathematics) and any unity to an empirical manifold, which, of course, must also be in space and time. As I just said, this leaves open the possibility of the categories’ being used to unify an empirical manifold in the absence of a corresponding synthesis of a pure manifold in space and time. On my account, the second step of the B-deduction rules out this possibility, since otherwise we could not explain how the categories can deliver a priori cognitions that constrain possible experience. The main task of the second step is thus that of establishing a necessary connection between the synthesis of every empirical manifold and a corresponding synthesis of a pure manifold in space and time. My reconstruction of Kant’s argument in § 26, which I see as occurring at B160–1, runs as follows:22 (6) The synthesis of apprehension is responsible for combining an empirical manifold of intuition into a unitary intuition. (7) The empirical manifold of intuition must conform to our a priori forms of intuition, space and time. (8) The synthesis of apprehension must conform to our a priori forms of intuition, space and time (from 6 and 7). (9) Every synthesis of an empirical manifold in space and time presupposes a corresponding synthesis of a pure manifold in space and time. (10) The synthesis of apprehension presupposes a corresponding synthesis of a pure manifold in space and time (from 8 and 9). The role of (6) is to make clear what is at stake in the second step of the B-deduction: the question is how we can combine an empirical manifold of intuition to form a unitary intuition. In (7), Kant makes explicit what we know is a distinctive feature of the second step: the argument takes into account that space and time are the forms of our intuition and that 22

h

d i

0 0

The reading I provide here fundamentally departs from my interpretation of the B-deduction in Gava (2015). Even though I now have a different view on the deduction, I believe that the general point of the article, which concerns the use of the ‘analytic’ and ‘synthetic’ methods in the Critique of Pure Reason, is correct.

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

147

every empirical manifold must conform to them. (8) draws the first consequence: since the empirical manifold that is combined in a synthesis of apprehension necessarily stands under the forms of space and time, the latter synthesis must also stand under these forms. It is in (9) that the most relevant premise is introduced. Kant’s point is that in order to use the categories to synthesize an empirical manifold in space and time, the categories cannot simply be used to synthesize what is empirically given in intuition. They must at the same time be used to synthesize space and time themselves and the pure manifold they contain.23 This is the only way in which they can serve as rules that constrain possible experience. They are, firstly, rules that constrain what relationships in space and time must look like. I take it that Kant introduces the distinction between space and time as forms of intuition and space and time as intuitions resulting from a synthesis to make this point – namely, Kant wants to single out an operation that the categories must perform on space and time themselves in order for those categories to count as rules that can constrain the combination of an empirical manifold: But space and time are represented a priori not merely as forms of sensible intuition, but also as intuitions themselves (which contain a manifold), and thus with the determination of the unity of this manifold in them (see the Transcendental Aesthetic). Thus even unity of the synthesis of the manifold, outside or within us, hence also a combination with which everything that is to be represented as determined in space or time must agree, is already given a priori, along with (not in) these intuitions, as condition of the synthesis of all apprehension. (B160–1)

In a footnote at the end of the first sentence, Kant clarifies that space as a form of intuition already contains a pure manifold (B160n). We can presumably say the same for time, such that both space and time as forms of intuition are capable of offering a pure manifold by themselves. This pure manifold is always available to us a priori, which means that we can combine elements of it and obtain unitary representations that have nothing empirical within them. This is what happens with geometrical figures, for example. If I trace the shape of a triangle in my imagination, I perform a synthesis of a pure manifold in space. I thus obtain, completely a priori, a unitary representation in which I can determine relationships that hold necessarily. The main point Kant makes in the second step of the 23

h

d i

According to Kant, this pure synthesis is performed by the imagination. Kant clarifies how this synthesis works in § 24. For excellent discussions of the role of the synthesis of imagination, see Ferrarin (1995), Land (2014), Schulting (2018a: Ch. 11) and Rosefeldt (2021).

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

148

B-deduction is that a similar operation on the pure manifold provided by the forms of space and time is always needed for the categories to be able to synthesize what is empirical in our intuition. How should we characterize the synthesis of the pure manifold provided by the forms of space and time? Is this synthesis necessary to account for the unity of space and time as wholes, as the conceptualists maintain? In fact, Kant does sometimes suggest that the synthesis of the pure manifolds of space and time is necessary for providing unity to space and time as wholes. At the end of § 26, he gives the following example: ‘Thus if, e.g., I make the empirical intuition of a house into perception through apprehension of its manifold, my ground is the necessary unity of space and of outer sensible intuition in general, and I as it were draw its shape in agreement with this synthetic unity of the manifold in space’ (B162). Here, Kant speaks of the necessary unity of space in general, and he seems to attribute this unity to a synthesis performed according to the categories. We have seen, however, that attributing this view to Kant creates problems for his argument in the Aesthetic to the effect that space and time must originally be intuitions given the particular unity they possess as wholes. As I made clear above, one of the two main desiderata for any interpretation of the B-deduction is to preserve the plausibility of this argument in the Aesthetic. Because my account allows us to satisfy the second desideratum (the one that requires a substantial reading of the second step) without claiming that the purpose of the latter is to extend the validity of the categories to all intuitions, it is better not to read the second step as arguing that the categories are necessary for the unity of space and time as wholes. But how does my reading of the second step satisfy the second desideratum identified above? As we saw, conceptualists usually claim that the second step of the B-deduction is not trivial because it extends the validity of the categories. On my reading, this is not the purpose of the second step. The first step already proves, first, that every unitary representation of a manifold of intuition, whether pure or empirical, must be the result of a synthesis performed according to the categories and, second, that every intuition that is mine is possibly synthesizable because it can in principle be brought under the synthetic unity of apperception. The second step does not extend the validity of the categories because the necessity of the latter still concerns manifolds of intuition that are combined, not necessarily all intuitions. I will say more on this in a moment. That the second step does not extend the validity of the categories does not mean that it does not make a substantial point, however. As I

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

149

have argued, the second step shows that a categorial synthesis of a pure manifold in space and time must always accompany the synthesis of an empirical manifold. First of all, this makes clear that a condition for the categories being used to synthesize an empirical manifold is that they are at the same time used to synthesize space and time themselves and the pure manifold they contain. Additionally, this explains how the categories can provide a priori cognitions that constrain possible experience. Before I move to my concluding remarks on the method of the transcendental deduction of the categories, let me make one last point. I have suggested that both steps of the B-deduction claim that every intuition that is mine must in principle be synthesizable according to the categories because it can be brought under the synthetic unity of apperception. I have also suggested that this is compatible with claiming that we might have intuitions without the categories in the sense that we might have intuitions that are not actually synthesized according to the categories. This commits me to a form of non-conceptualism. Let me emphasize, however, that it is a weak form of it, since I do not claim that we in fact have intuitions that are not synthesized and thus do not have any unity. Of course, it is hard to conceive of what an intuition without categorial unity could be. Plausibly, one could argue that for any sensible representation we can have, it must at least have either spatial or temporal unity, where the unity at stake is the unity of particular spaces and times. For example, one could argue that if I have a spatial representation, it must have a determinate shape. In saying that we might have intuitions without the categories, I do not want to suggest that one can make sense of the idea of an intuition that does not have the unity provided by the categories. I only maintain that the B-deduction does not rule out the possibility of our having intuitions that do not actually have unity. It is likely that Kant was simply not interested in this issue. d. The method of the transcendental deduction of the categories. Similar to my reading of the transcendental exposition of space, I now wish to connect my reconstruction of the B-deduction to the issues listed at the beginning of this chapter. The first issue concerns what I have called methodological and common sense conservatism. In what sense does the B-deduction exemplify these forms of conservatism? The B-deduction is methodologically conservative because it aims to establish that the categories, as root concepts for the cognition of objects, have objective validity. In the context of the transcendental deduction, this means that through the categories we cognize

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

150

something that really pertains to objects. The categories can and must be applied to objects a priori, otherwise we would not be able to obtain unitary representations, where unitary representations are fundamental to cognition. Even in the case of the categories, the conservatism is only moderate because it involves a revision of our common way of characterizing the cognitions expressed by the categories. The transcendental deduction adopts the characterization of the categories given in the metaphysical deduction, according to which the categories are the essential ways in which we organize the manifold of intuition in accordance with the forms of judgement. If we take a category like causality as an example, characterizing the latter as depending on a form of judgement clearly involves a radical reconceptualization of what a cause is. In what sense is the B-deduction common sense conservative? Does the B-deduction start from common sense beliefs or from propositions that were commonly held to be true by scientists and scholars of Kant’s time? What the argument takes for granted is that we have unitary representations of the manifold of intuition. This claim can hardly qualify as a ‘common sense belief’ since it involves a level of abstraction that is more common to the sciences. Was it commonly held to be true in psychological investigations of Kant’s time? This might be the case, but I want to emphasize another sense in which the B-deduction is common sense conservative. By proving that the categories are necessary for unifying both empirical and pure manifolds, the B-deduction demonstrates that they relate to objects a priori. There seem to be two implicit assumptions in this argumentative strategy. The first is that unitary representations are necessary for cognizing empirical objects. The second is that we do cognize empirical objects.24 It is this latter implicit assumption that clearly commits Kant to a common-sense belief and to a related common sense conservatism. Now turning to the second issue, what kind of argument does the B-deduction present? What is its structure? As many interpreters have already emphasized, the B-deduction can be characterized as a transcendental argument. Transcendental arguments are normally described as deductive arguments that argue for a claim q by showing that q is a necessary condition for the claim p. Moreover, the ‘necessity’ that is at stake in the claim ‘q is a necessary condition for p’ is not causal or physical. Rather, transcendental arguments normally build on the idea that p would not be conceivable if q were not the case. The attribution of inconceivability 24

h

d i

0 0

In this respect, my reading agrees with so-called ‘regressive’ accounts of the B-deduction. See Ameriks (1978).

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

151

can be based on different considerations. It might simply rest on logical necessity, such that claiming p and ¬q at the same time would be a logical contradiction. It might also rest on a more broadly construed conceptual necessity, which appeals to the idea of a conceptual schema and the constraints it sets on our representations of objects. One further feature of transcendental arguments is that p customarily describes features of experience. ‘Experience’ can either be understood in subjective terms, as designating the private representational states of a self-conscious subject, or in objective terms, as consisting of intersubjectively accessible representations that at least sometimes correctly represent the world (I have provided a similar characterization of transcendental arguments in Gava 2019c; see also Stern 2019). Given this description, it is plausible to describe the B-deduction as a transcendental argument. First, the B-deduction builds on the idea that combination or synthesis is necessary for having unitary representations of a manifold of intuition. Accordingly, it can be reconstructed as starting from the claim ‘We have unitary representations of a manifold of intuition’ and arguing that if this claim obtains, the claim ‘We are able to combine or synthesize the manifold of intuition’ must also obtain. The same procedure is reiterated when the necessity of the claim ‘The categories guide the synthesis of the manifold of intuition’ is established. Second, when Kant argues that the categories are necessary for combining the manifold of intuition, he claims that the unity within this manifold would otherwise remain inconceivable.25 Third, the claim that ‘we have unitary representations of a manifold of intuition’ clearly expresses features of our experience. Is this experience to be understood in subjective or objective terms? Since two implicit assumptions in Kant’s argument are that unitary representations are necessary for the cognition of empirical objects and that we do cognize empirical objects, experience should be taken in objective terms. The last issue we must discuss is whether the B-deduction agrees with my characterization of transcendental philosophy as a discipline that only establishes positive results regarding the objective validity of root concepts for the cognition of objects. As I have noted above, I do not want to claim 25

h

d i

Notice that this is different from the appeal to explainability in the transcendental exposition of space. The exposition starts from a phenomenon that is taken for granted, namely that there are synthetic a priori cognitions in geometry, and asserts that this can only be explained by assuming another fact, namely that we have a pure intuition of space. By contrast, by saying that the unity of the manifold would be inconceivable without the categories, the transcendental deduction argues that without the categories we could not form the representation of that unity. Perhaps one good way to mark this difference is to say that explainability takes a ‘third-person’ perspective on a ­phenomenon, whereas ‘conceivability’ takes a ‘first-person’ view.

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

152

that within the sections dedicated to the B-deduction, Kant does not make negative points stressing that the validity of the categories has limits. As we will see more clearly in Chapter 6, Kant evidently makes claims of this sort. What I want to say is that within those sections, there is a positive argument about the validity of the categories that can be reconstructed independently of these negative points. It is this positive argument that I have reconstructed here and that on my account belongs to transcendental philosophy.

3  The Transcendental Deduction of the Transcendental Ideas Strange as it may seem, the Critique of Pure Reason contains a ­transcendental deduction of the ideas of pure reason (A670–1/B698–9), which is presented in the Appendix to the Transcendental Dialectic. This deduction proves that the ideas have some objective validity, even though this objective validity is characterized as only being ‘indeterminate’ (A669–70/697–8).26 The claim that the ideas have objective but ‘indeterminate’ validity has been used against so-called ‘fictionalist’ interpretations of the validity of the ideas.27 According to these interpretations, the ideas can only be used ‘regulatively’ because they involve false propositions, which are nonetheless useful as heuristic devices for the study of nature (see Vaihinger 1911: Part III; Grier 2001: Ch. 8). Against this view, some interpreters have argued that if the ideas have objective validity, this means that they must in a sense provide cognition of objects (Wartenberg 1979; Wartenberg 1992; O’Shea 1997; Ypi 2017). In turn, this seems to imply that the ideas are at least partially constitutive, which contrasts with Kant’s contention that the only legitimate use of the ideas is ‘regulative’. My aim in this section is, first, to provide an interpretation of the objective validity of the ideas that can coherently fit with the claim that they only have a regulative use. As we will see, ‘objective validity’ in the case of the ideas will mean something very different in comparison to the objective validity of our representations of space and time and the categories. I will begin by introducing what I call the ‘strong reading’ of the objective validity of the ideas, which I will reject because it ultimately treats them as constitutive. In a further step, I will propose a weaker reading of the 26

27

h

d i

0 0

That Kant ascribes objective validity to the ideas is strange because the main point of the Dialectic appears precisely to be that the ideas cannot be considered cognitions of objects, where this appears to entail that they cannot have objective validity. Additionally, Kant denies that a transcendental deduction of the ideas is possible just a few pages before he provides one (see A663–4/B691–2). I borrow this term from Marcus Willaschek (2018: 117).

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

153

objective validity of the ideas, according to which the objective validity of the ideas captures two thoughts. The first is that the ideas are necessary means of perfecting our capacity to obtain empirical cognitions through the understanding. This is what Kant calls indirect objective validity. The second thought is that in the process of inquiry in which we use the ideas as heuristic devices, we are in a certain sense justified in taking them as having objects, even though this justification does not imply that we have knowledge of those objects. This second thought expresses what I will call the practical validity of the ideas.28 After having characterized the objective validity of the ideas in these terms, I will consider what it means to say that this validity is ‘indeterminate’. I will then move on to the second main aim of this section, which is to provide a reconstruction of Kant’s argument in the transcendental deduction of the ideas. Finally, I will consider the issues outlined in the introductory section. One point needs to be emphasized from the outset. As in the transcendental deduction of the categories, in the transcendental deduction of the ideas Kant clearly voices negative contentions regarding cases in which the use of the ideas is illegitimate. This seems to conflict with my claim that transcendental deductions only establish positive results and are not concerned with setting limits to the use of concepts. My strategy here will be different from the one I used for the transcendental deduction of the categories. In that context, I conceded that Kant does make negative points within the sections dedicated to the transcendental deduction. I argued that the positive argument that belongs to transcendental philosophy can be reconstructed independently of these negative points. Here, by contrast, I will suggest that when the transcendental deduction of the ideas makes negative claims concerning their validity, it simply relies on the negative arguments already presented in the main chapters of the Dialectic. a. The strong interpretation of the objective validity of the ideas. As mentioned above, I will start by presenting what I have called the strong interpretation of the objective validity of the ideas. I will analyse two variants of this interpretation offered by Thomas Wartenberg (1979; 1992) and James O’Shea (1997), respectively. While I think both variants ultimately fail to deliver a satisfactory interpretation of the objective validity of the ideas, 28

h

d i

My use of the term ‘practical’ may seem strange in the context of the Appendix to the Transcendental Dialectic. I use it for two reasons. First, the justification of the validity of the ideas appeals to a certain practice, namely the practice of inquiry. Second, I want to mark the similarities with Kant’s justification of belief in the Canon.

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

154

O’Shea’s approach introduces an important insight that I will use in my own reading. Let me start by analysing Wartenberg’s position. On his account, the task of the Appendix is to identify necessary conditions for the practice of hypothesis formation, which is essential to natural science. This practice, which Wartenberg calls methodology M, is guided by the principles of genera, species and affinity, which instruct us to build scientific hypotheses that, respectively: (a) subsume empirical concepts under more general genus concepts; (b) specify existing empirical concepts by identifying different species concepts of those concepts; and (c) increase the number of species concepts under the same genus concept (Wartenberg 1979: 412).29 Wartenberg reads Kant as arguing that following methodology M and assuming the corresponding principles is the only way in which we can rationally pursue scientific inquiry into nature. He further maintains that it would be irrational to pursue methodology M if nature itself did not correspond to the principles that are expressions of M. In this sense, the Appendix offers an argument to the effect that nature itself agrees with these principles, which in turn means that nature must be systematic. Here is how Wartenberg reconstructs Kant’s line of reasoning: Science is rational in that it is possible via the adoption of methodology M to reach our goal, the establishment of true empirical theories. But such rationality is established only when we have proved that nature itself has the sort of unity that science will enable us to uncover. For only with this item of metaphysical knowledge can we guarantee the possibility of reaching our aim by engaging in scientific research. Thus we see Kant’s justification of scientific procedure involves two distinct parts. On the one hand, he shows that, in order for science to be possible, nature must be systematic. It is such metaphysical knowledge that I have claimed is embodied in (SN) [the principle that nature is systematic]. On the other hand, if we are to come to know that order which we know on a priori grounds nature must have, we need to proceed in accord with methodology M. (Wartenberg 1979: 423)

As it is clear from the quote, Wartenberg takes the Appendix to establish ‘metaphysical knowledge’ that nature is systematic. On his account, it is this metaphysical knowledge that provides objective validity to the idea of the systematicity of nature. Wartenberg denies that his reading commits him to a ‘constitutive’ use of the idea of the systematicity of nature. To 29

h

d i

0 0

I provide an interpretation of these principles in Chapter 1. It is in my reading of the third principle in particular that I depart from Wartenberg.

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

155

support this claim, he compares the idea to the principles of the understanding. While the latter are constitutive in the sense that they constitute the objective domain of appearances, the idea of systematicity is not required to constitute that domain (Wartenberg 1979: 420–1; Wartenberg 1992: 237–8). From this it certainly follows that the idea of the systematicity of nature cannot be constitutive in the same way that the categories or the principles of the understanding are. It does not follow that the idea of systematicity as described by Wartenberg is not constitutive, however. In fact, the claim that the idea of the systematicity of nature provides metaphysical knowledge cannot but commit him to a constitutive use of this idea. Kant connects the constitutive use of a concept to its capacity to determine an object (see A670/B698). I cannot see what ‘metaphysical knowledge’ of nature might involve beyond a determination of nature as an object, and this is what Kant explicitly excludes. While O’Shea’s reading of objective validity moves in a similar direction, it differs from Wartenberg’s approach in at least two respects. First, it does not read the argument as being based on what is required to make scientific inquiry rational. Rather, it regards the Appendix as establishing a more general point concerning what is necessary for obtaining empirical concepts and laws.30 Second, O’Shea bases the objective validity of the ideas on requirements that follow from the way in which the understanding legislates over nature but that cannot be fulfilled by the understanding itself. Here is how O’Shea reconstructs his reading: I have suggested that Kant’s solution to this aspect of the problem is to argue, first, that the genuinely objectively valid requirements of understanding issue directly in the demand for genuine lawfulness in the empirical realm […]; second, that the pure understanding cannot itself legislate a priori what particular forms this empirical lawfulness will take; third, that the global empirical assumptions spelled out in the regulative maxims are conditions that are necessary for the possibility of understanding (and so of experience) itself; that is, they are necessary for the possibility of meeting the empirical demands of understanding that issue from the transcendental laws of pure understanding themselves; and finally, that the regulative ideal of systematic unity that is thus warranted does not flout the sensible limits on understanding set by the critical philosophy, since we are warranted in presupposing that systematic unity obtains in nature only to an a priori indeterminable degree. (O’Shea 1997: 241–2) 30

h

d i

Of course, in this context one needs to distinguish between empirical concepts and laws and determine the kind of relation that holds between them. Since O’Shea himself does not clarify this relationship, however, I will not take up this task. It goes beyond the point I want to make in the remainder of this section.

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

156

O’Shea justifies the first two claims by appealing to an interpretation of the Second Analogy. On his account, the latter establishes not only that possible experience necessarily conforms to a general law of causality but also that there are specific empirical causal laws. While the understanding guarantees that there must be such laws, it cannot determine a priori which laws these are. It is at this point that the necessary role of the ideas of reason comes into play. Only by following the idea of systematic nature can we fulfil the requirements regarding empirical laws set by the understanding. In O’Shea’s account, the idea of the systematicity of nature is objectively valid because it is a necessary condition for finding empirical concepts and laws that we know must obtain in nature, even though the understanding cannot determine precisely which laws obtain. O’Shea’s approach finds support in various passages in which Kant claims that regulative ideas are necessary for experience or empirical cognition (A651/B679; A654/B682; A663/B692). He takes the fact that the ideas are necessary conditions for experience as guaranteeing that nature is in fact systematic. Furthermore, he maintains that this does not equate to making the idea of systematicity constitutive of nature, first, because the idea is not used beyond phenomenal nature and, second, because systematicity is only attributed in an indeterminate way. I find O’Shea’s proposal subtle and elegant, but I think he nonetheless does not succeed in avoiding making the idea of the systematicity of nature constitutive. On the one hand, if one wants to respect the limits to our cognition set by the Dialectic, it is not sufficient to say that the idea of systematicity is only used for objects of nature as appearances. Rather, if the idea of the systematicity of nature is regulative and cannot provide cognition of an object, this means that it is not a cognition of either appearances or things in themselves. Moreover, it does not help to say that the idea of the systematicity of nature remains regulative because it is indeterminate. If we can say we know that nature is systematic, no matter how indeterminately, it seems to follow that our idea of the systematicity of nature is at least in some respects constitutive of it. b. What is the objective validity of the ideas? Even though I have suggested that O’Shea’s reading is ultimately unsuccessful, I think he provides an important insight for grasping a first aspect of the objective validity of the ideas, which I will call indirect objective validity. The insight I am referring to is the view that ideas of reason are a necessary means of obtaining cognitions that in fact belong to the understanding. O’Shea spells out this

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

157

thought by claiming that the idea of the systematicity of nature is necessary for finding empirical laws and concepts that we know must obtain through the understanding’s a priori legislation. My take on this insight is more modest in two respects. First, I do not claim that the understanding would be utterly unable to find empirical laws and concepts without the guidance of the ideas of reason. Second, on my account, the necessity of appealing to the ideas does not issue from the a priori legislation of the understanding. In what sense are ideas of reason necessary means of obtaining cognitions that in fact belong to the understanding? I take this claim to mean that the ideas of reason are a necessary means of perfecting our capacity to obtain objective empirical cognitions through the understanding. That is to say, it is not the case that without the contribution of the ideas of reason the understanding would be unable to obtain empirical concepts and laws. However, guidance from these ideas is necessary if the understanding is to perform its proper tasks in the best possible way. Kant points towards such a view in various passages from the Appendix. For example, he submits that through the ideas ‘the empirical and determinate use of the understanding in experience can be brought into thoroughgoing agreement with itself’ (A665–6/B693–4). Moreover, he concludes the transcendental deduction of the ideas by claiming that they are only vindicated ‘as regulative principles for the systematic unity of the manifold of empirical cognition in general, through which this cognition, within its proper boundaries, is cultivated and corrected more than could happen without such ideas, through the mere use of the principles of understanding’ (A671/B699, my emphasis). How should we read these passages? Kant is saying that one of the main tasks of the understanding is to provide empirical concepts and to identify empirical laws. In thinking of this task, there are criteria internal to the understanding that determine what it means to perfect empirical cognition. In this sense, one simple way in which the understanding could make progress is by increasing the number of empirical concepts it provides.31 If we start from these considerations, we can maintain, first, that providing empirical concepts and laws is a fundamental task of the understanding and, second, that there are criteria of the understanding that determine when the understanding is perfecting the way in which it performs this 31

h

d i

In the quote, Kant also speaks of ‘correcting’ the cognition of the understanding. This could refer to the need to avoid inconsistences in the empirical concepts and laws we identify through the understanding. Kant would thus be suggesting that trying to increase the systematic relationships between our concepts is instrumental to discovering and correcting those inconsistences.

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

158

task. At this point, one can argue that while the understanding possesses a concept of perfection with its own criteria, it is guidance by the ideas of reason, and by the idea of systematicity in particular, that is a condition for the understanding’s fulfilment of its criteria of perfection in the best possible way. That is to say, following the criteria for the perfection of cognition provided by reason (criteria that instruct us to systematically integrate the empirical cognitions of the understanding as far as we can) is a condition for fulfilling the criteria for the perfection of the understanding (again taking up our example: increasing the number of our empirical concepts) in the best possible way. Accordingly, the ideas of reason can be regarded as necessary means of perfecting our capacity to obtain objective empirical cognitions through the understanding. This line of reasoning is key to grasping the first sense in which ideas of reason can have ‘objective validity’, for when we follow the ideas of reason in the direction of perfecting the performance of the understanding, we do not simply systematically organize the empirical cognitions we already have. We also obtain cognitions that we would not have obtained without the guidance of those ideas. This happens, for example, when in studying natural kinds, we successfully introduce a new genus that covers different species (see Kant’s example of pure water, pure earth and pure air at A646/ B674). Therefore, there are some empirical cognitions of the understanding that we would not have obtained had we not followed the ideas. It is in this sense that the ideas have indirect objective validity. They are necessary means of obtaining certain empirical cognitions of the understanding that we would not have otherwise obtained. To see this, let me again quote a key passage, this time in full: Now since every principle that establishes for the understanding a thoroughgoing unity of its use a priori is also valid, albeit only indirectly, for the object of experience, the principles of pure reason will also have objective reality in regard to this object, yet not so as to determine something in it, but only to indicate the procedure in accordance with which the empirical and determinate use of the understanding in experience can be brought into thoroughgoing agreement with itself, by bringing it as far as possible into connection with the principle of thoroughgoing unity. (A665–6/B693–4)

I take this to mean that the ideas of reason are necessary for perfecting our capacity to obtain objective empirical cognitions through the understanding. When they perform this task, they can be taken to have indirect objective validity because they are necessary means to obtaining actual empirical cognitions we would not have obtained without their guidance.

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

159

This is what I call the indirect objective validity of the ideas. As suggested above, Kant connects the objective validity of the ideas to a second thought. This expresses what I call the practical validity of the ideas. Famously, Kant submits that the regulative use of the ideas exhibits a logical requirement of reason that instructs us to organize our empirical cognitions systematically. However, reason cannot avoid interpreting this logical requirement transcendentally, as not only requiring that our empirical cognitions are organized systematically but also presupposing that nature itself is systematically ordered. In what sense can the idea of systematicity, when used regulatively, also be interpreted transcendentally? Defenders of the strong interpretation of the objective validity of the ideas treat the transcendental presupposition of systematicity as confirmation that the Appendix proves that nature is in fact systematic, even though we cannot determine the extent to which this is so. As we saw, in so doing, they in fact make the idea of systematicity constitutive. In contrast to this view, I believe that the transition from systematicity understood as a solely logical principle to systematicity understood objectively can be read as based on practical considerations. I here follow a proposal put forward by Andrew Chignell (2007), who suggests that the way in which Kant justifies the regulative use of the ideas in the Appendix to the Dialectic is sometimes linked to the justification of what Kant calls ‘doctrinal’ belief in the Canon. Belief (Glaube) is a technical term in Kant’s philosophy. It describes a doxastic attitude in which we fully accept a proposition as true and we are justified in doing so, even though we consciously recognize that our ‘taking-to-be-true’ (Fürwahrhalten) is not based on sufficient evidence. Rather, the evidence we have leaves undecided whether the proposition in question is true or false.32 Why can we be rational in fully accepting a proposition for which we lack sufficient evidence? Because that proposition plays an important role in our practice. The main idea behind the justification of belief, in Kant’s technical sense of the term, is the following: when the evidence we have leaves it undecided whether a proposition is true or false, we are rationally required to believe that proposition if it identifies conditions that are necessary for the realization of the ends we are pursuing.33 Chignell suggests that Kant uses a similar strategy to justify the assumption that nature is systematic in the Appendix (Chignell 2007: 351–4).34 32 33 34

h

d i

On Kantian belief, see also Stevenson (2003), Pasternack (2011), Fonnesu (2015), Willaschek (2016) and Höwing (2016a). I provide a more detailed interpretation of Kantian belief in Gava (2019b). See also Chapter 7. Another interpreter who draws a connection between the Canon of Pure Reason (where Kant ­discusses belief) and the Appendix is Frederick Rauscher (2010).

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

160

The end that is at stake in this case is epistemic. Using the formulation I introduced in discussing the indirect validity of the ideas, the end can be identified with perfecting the empirical cognition we obtain through the understanding. Following the logical requirement of reason that instructs us to organize our empirical cognitions systematically can be seen as a necessary condition for meeting that end. Kant then submits that a condition of success when we proceed in this way is that nature is in fact systematic. That is to say, we would not be able to perfect our empirical cognition by following the logical requirement of systematicity if nature were not systematic. Therefore, since we do not have sufficient evidence that nature is systematic, when we do follow that requirement we can only be rational if we believe, in Kant’s sense of the term, that nature is systematic. Understood in such terms, the transition from systematicity understood as a solely logical principle to systematicity understood objectively does not involve a claim to cognition or knowledge. It only involves believing that systematicity actually applies to nature. From this perspective, what is ‘objective’ in the idea of the systematicity of nature is simply the fact that we take the idea to apply to an ‘object’, namely nature. It is not objective in the sense that it constitutes knowledge (or cognition, for that matter), however. This is what I call the practical validity of the ideas.35 I do not have enough space to argue for this view further.36 Let me just note that in the Appendix Kant often uses terms that have strong practical connotations, such as ‘interest’ or ‘maxim’ (see, for example, A666/B694). This suggests that at least part of his argument rests on identifying the process of perfecting empirical cognition as a particular practice and specifying the conditions for the rationality of that practice.37 c. What is the indeterminate validity of the ideas? On my account, the indirect and practical validity of the ideas captures what is ‘objective’ about this validity. As we saw, Kant characterizes the validity of the ideas by also saying that it is ‘indeterminate’. What does this mean? Kant is quite clear 35

36 37

h

d i

0 0

I here disagree with Willaschek (2018: 115–16), who argues that the regulative use of ideas cannot involve belief and instead involves hypothesis. At A670/B698, there is a passage that speaks against Willaschek’s interpretation, where Kant submits that the object of an idea cannot be given ‘hypothetically’. I consider this interpretative hypothesis in Gava (2018b). Of course, that I call this sense of the validity of the ideas ‘practical’ does not mean that it is obtained with reference to moral or prudential considerations. I simply wish to point out that it can be characterized with reference to an end that we pursue in the ‘practice’ of perfecting empirical cognition.

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

161

in this respect. He submits that the validity of the ideas is indeterminate because we lack schemata of sensibility for those ideas: The actions of the understanding, however, apart from the schemata of sensibility, are indeterminate; likewise the unity of reason is also in itself indeterminate in regard to the conditions under which, and the degree to which, the understanding should combine its concepts systematically. (A664–5/B692–3, translation altered)

Kant draws a comparison between unschematized categories and the ideas. For my purposes, the important point to keep in mind is that unschematized categories do not constitute cognitions of objects. It is the capacity to determine time relationships a priori that renders them cognitions. From this, it follows that the ideas, as indeterminate in a similar way, cannot provide cognitions of objects. Just as the unschematized categories, insofar as they are indeterminate, cannot constitute a cognition, so the ideas, insofar as they are indeterminate in a similar way, cannot constitute a cognition. This speaks against the way in which strong readings of the objective validity of the ideas appeal to their indeterminateness in order to safeguard the claim to regulativity. As we saw, these readings claim that the idea of the systematicity of nature provides an actual cognition. Since this cognition is ‘indeterminate’, however, the idea can coherently be described as only ‘regulative’. If my reconstruction of the indeterminateness of the ideas is correct, it is simply contradictory to claim both that an idea ­provides ­cognition and that it is indeterminate. d. The argument of the transcendental deduction of the ideas. Now that we have clarified what is ‘objective’ and what is ‘indeterminate’ in the validity of the ideas, we can reconstruct the transcendental deduction, which is charged with establishing this validity. The deduction takes place between A670/B698, where Kant suggests that a transcendental deduction of the ideas ‘must definitely be possible’, and A671/B699, where Kant closes the argument by saying ‘[a]nd this is the transcendental deduction of all the ideas of speculative reason […]’.38 38

h

d i

Even though Kant explicitly says that he provides a transcendental deduction of the ideas, some interpreters claim that this cannot be the case. They usually focus on the passage where Kant says that a transcendental deduction of the ideas is impossible (see for example Horstmann 1989 and Goldberg 2004). Those interpreters who do take Kant’s contention that he provides a transcendental deduction seriously (see, for example, Zocher 1958; Caimi 1995; La Rocca 2011; Ypi 2017) do not agree on very basic issues, such as where the deduction is located or how many deductions there are. I here simply take Kant at his word and locate the deduction where Kant explicitly says it occurs. For a useful overview of the literature on the transcendental deduction of the ideas, see Meer (2019: Ch. 5).

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

162

Thus far, in discussing the validity of the ideas of reason, I have mainly taken into account the idea of the systematicity of nature. There are various reasons for this choice, but let me just mention two. First, the idea of systematic unity is more widely discussed in the literature. This is probably because Kant’s argument for the regulative use of the ideas sounds more plausible when it is taken to refer to this idea. Second, it is in discussing the idea of the systematicity of nature that Kant elaborates more extensively on the meaning of ‘objective’ and ‘indeterminate’. In reconstructing the transcendental deduction of the ideas, it is important to keep in mind that the ideas that are at stake are those of the soul, the world and God.39 Does this constitute an obstacle to using our clarification of the ‘objective indeterminate validity’ of the ideas to illuminate Kant’s deduction? I do not believe it does, since the ideas of the soul, the world and God are also justified by their role in the systematization of the empirical cognitions of the understanding. Let us turn to my reconstruction of the argument: (1) The understanding can only meet its own criteria for perfecting empirical cognition when it systematically integrates its empirical cognitions by following the ideas of the soul, the world and God. (2) The use of an idea as a heuristic device is legitimate when this use is a condition for perfecting empirical cognition. (3) The ideas of the soul, the world and God can legitimately be used as heuristic devices (from 1 and 2). (4) An idea that is a condition for perfecting empirical cognition has objective indirect validity with respect to the empirical cognitions that we would not have obtained without following that idea. (5) The ideas of the soul, the world and God have objective indirect validity with respect to the empirical cognitions that we would not have obtained without following those ideas (from 1 and 4). (6) We cannot rationally use an idea as a heuristic device for attaining an epistemic end (like perfecting our empirical cognition) if we do not believe (in Kant’s technical sense of the term) that the conditions for attaining that end obtain.40 39

40

h

d i

0 0

In Chapter 3, we saw that there are actually more than three transcendental ideas. Kant only speaks of three ideas in the transcendental deduction, however. We can understand his talk of ‘the idea of the soul’ as a concise way of capturing the role the different psychological ideas can play in empirical cognition. Something similar can be said of ‘the idea of the world’ and the different cosmological ideas. As far as the idea of God is concerned, I agree with Willaschek (2018: 169) that there is only one theological idea. One might object that this premise is implausible when the epistemic aim we are pursuing is proving a particular theory. For example, we might have the epistemic aim of proving that life on Earth

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

163

(7) The existence of the soul, the world and God are conditions for perfecting our empirical cognition by following the ideas of the soul, the world and God.41 (8) We cannot rationally use the ideas of the soul, the world and God as heuristic devices for perfecting our empirical cognition if we do not believe that the soul, the world and God exist (from 6 and 7). (9) We can believe that the soul, the world and God exist (from 3 and 8). Let me emphasize three points regarding my reconstruction. First, the argument starts by establishing the legitimacy of the ideas of the soul, the world and God in their use as heuristic devices. This is connected to what I have previously called the logical requirement of reason and contrasts with the transcendental interpretation of this requirement. This means that the ideas are justified by their role in systematizing cognition. In this role, they simply offer the ideal of a maximally integrated system of empirical cognitions. It is only later in the argument (steps 6 to 9) that Kant justifies treating the ideas as having objects in a certain sense. Second, the justification of the use of the ideas as heuristic devices is sufficient to establish that they have indirect objective validity in the way I have defined it. They have indirect objective validity because they are conditions for obtaining empirical cognitions of the understanding that we could not have obtained had we not attempted to perfect our empirical cognition by following the ideas. Saying that the ideas have objective validity in this sense is of course different from saying that the ideas can be taken ‘transcendentally’ as referring to existing objects. Third, the argument does establish that the ideas can also be taken ‘transcendentally’ as referring to existing objects, but only in the sense that we are justified in ‘believing’, in Kant’s technical sense of the term, that the soul, the world and God exist, not in the sense that we can know or cognize that they exist. Kant does not refer explicitly to believing in the

41

h

d i

came from a comet. In rationally pursuing this hypothesis, we do not need to believe (in Kant’s sense of the term) that the hypothesis is true. Notice, however, that in this example it is odd to say that our aim is to prove a particular theory. It is more plausible to say that our aim is to find out where life on Earth came from. In pursuing that aim, we might follow different hypotheses (including the one claiming that it came from a comet) without believing them. But if our aim is really to prove a particular theory (such as the theory that life on Earth came from a comet), it does not seem so implausible to say that we can only rationally pursue that aim if we believe in the truth of the theory. I thank Marcus Willaschek for this objection. Let me here point out a problem with this step when we consider the idea of the world. The step assumes that the ‘world’ can exist. However, one can read the Antinomy of Pure Reason as establishing that the world, understood as a ‘totality of appearances’, cannot in fact exist. This problem only applies to the idea of the world. By contrast, it is not problematic to assume that God and the soul can exist (as things in themselves).

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

164

transcendental deduction, but I would argue that this is the best way to interpret his claim that the objects of the ideas are given only as ‘object[s] in the idea’ (A670/B698). Kant later connects this way of having objects to what he calls the ‘relative assumption’ of an object. He explains that this assumption involves thinking ‘as existing a being that corresponds to a mere and indeed transcendental idea’ (A676/B704). The relative assumption is contrasted with an ‘absolute’ assumption. I would argue that the best way to interpret this distinction is to read the relative assumption of objects as based on an argument like the one presented in steps (6) to (9).42 By contrast, the claim that the existence of the objects of the ideas cannot be assumed absolutely simply means that we cannot know or cognize that these objects exist. e. The method of the transcendental deduction of the ideas. Let me now discuss the three issues I raised at the beginning of the chapter. I will start by discussing methodological and common sense conservatism. Is the transcendental deduction of the ideas conservative in these two senses? It is methodologically conservative because it aims to confirm that the ideas have some sort of validity. We have seen that Kant calls this validity ‘objective’ even though what he means by this term is their indirect and practical validity. The indirect and practical validity of the ideas does not involve ‘objective’ validity in the strong sense of the term – namely, ‘objective validity’ here does not mean that through a representation we cognize something that really pertains to objects, as in the case of the categories and the representations of space and time. In fact, it is difficult to explain why Kant describes indirect and practical validity as ‘objective’. First of all, practical validity expresses the kind of validity that beliefs have, and beliefs for Kant only have subjective validity because they are not based on evidence or ‘objective grounds’.43 Moreover, it is true that the indirect validity of the ideas connects them to actual cognitions. It shows that 42

43

h

d i

0 0

It is true that in the context of the deduction, Kant explains that assuming an object in the idea ‘is really only a schema for which no object is given, not even hypothetically’ (A670/B698). Since Kant denies that we can hypothetically assume that the objects of the ideas exist, and since belief involves a stronger commitment to the truth of a proposition than hypothesis does, one might take the passage as evidence against my reading: if we cannot legitimately hypothesize that the ideas have objects, we certainly cannot believe they do. Kant describes hypotheses as particular cases of opinion (Meinung) (A770/B798), however, and opinions are only rational when they are based on knowledge we already have (A822/B850). We do not have any knowledge on which to base a hypothesis regarding the existence of the soul, the world and God. By contrast, belief is an attitude that does not need to be based on knowledge. Therefore, the fact that Kant rules out hypothesizing about the existence of the soul, the world and God does not rule out its being rational for us to believe in that existence. I make a similar point in Gava (2018b).

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

165

without the ideas we would not have obtained certain cognitions of the understanding. However, this is very different from saying that the ideas themselves are cognitions. So why does Kant claim that the ideas have objective validity? RolfPeter Horstmann (1989: 168) has suggested that the Critique of Pure Reason assumes that, in order to be considered a ‘transcendental principle’, a ­representation must involve actual cognition of objects. If this is accurate – and because, as we saw, Kant does want to prove that the ideas are transcendental conditions for obtaining certain cognitions of the ­understanding – one might argue that in attributing objective validity to the ideas Kant is attempting to justify viewing them as ‘transcendental principles’ in some sense. Note that this would mean that Kant ultimately failed in his attempt. If, given the conceptual means of the first Critique, the ideas can only be transcendental conditions of cognition if they imply the actual cognition of objects, then proving that they are such conditions requires more than proving their indirect and practical validity, since neither their indirect nor their practical validity establishes that they involve the ­cognition of objects (see Gava 2018b). Leaving this issue aside, it is clear that by establishing the indirect and practical validity of the ideas Kant was attempting to show that they have at least some validity. In this sense, the transcendental deduction can be described as methodologically conservative. Moreover, because saying that the ideas have indirect and practical validity clearly involves a revision of our common ways of characterizing the concepts of the soul, the world and God, Kant’s methodological conservatism regarding these concepts is moderate. Is the transcendental deduction of the ideas also common sense conservative? It certainly does not start by taking for granted common sense beliefs or propositions that were commonly held to be true by scientists of Kant’s time. There is a minimal sense in which it can be considered common sense conservative, however, since it starts from the assumption that the practice of perfecting our empirical cognition is rational. On Kant’s account, this practice is essential to the pursuit of science, even though it is not limited to it. Accordingly, the Appendix can be read as taking the rationality of the pursuit of scientific knowledge for granted, which is an assumption that can plausibly be viewed as part of our common sense. The second issue concerns what kind of argument the transcendental deduction of the ideas employs. In fact, the transcendental deduction seems to combine three kinds of argument. The first establishes the

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

The Method of Transcendental Philosophy

166

legitimacy of assuming the ideas as heuristic devices for perfecting our empirical cognition. It draws on assumptions regarding the rationality of investigations aimed at improving and extending empirical cognition. The second argument – that which establishes the indirect validity of the ideas – proves that the ideas are, in a sense, conditions for obtaining empirical cognitions that we would not otherwise have obtained. It resembles a transcendental argument since it points to a claim p (which expresses an empirical cognition that we obtained thanks to the guidance of reason) and then maintains that this cognition would have been impossible without assuming at least one of the ideas of reason (which is expressed by a claim q) as a heuristic device. However, the argument only resembles a transcendental argument because it does not prove that p is inconceivable if we do not assume q. It merely proves that we can only get to p if we pursue a course of investigation under the guidance of q. The third argument is a practical argument that establishes the rationality of a belief, in Kant’s technical sense of the term, by treating empirical inquiry as a particular kind of practice the aim of which is epistemic. The final issue we must address is whether the transcendental deduction of the ideas can be read as only establishing positive results regarding the validity of root concepts for the cognition of objects. If my reconstruction of the above argument is right, the transcendental deduction is an argument that only establishes the positive sense(s) in which the ideas have validity. It is certainly true that, in the transcendental deduction of the ideas, Kant claims that the ideas cannot be taken constitutively, as providing actual cognition of objects (see A671/B699). However, this negative claim is not established in the transcendental deduction of the ideas (or in the Appendix, for that matter). Rather, both the transcendental deduction of the ideas in particular and the Appendix in general rely on what Kant has already established in the main parts of the Dialectic. The Appendix merely proves that the ideas can be used as heuristic devices and that they have indirect and practical validity. On its own, it cannot rule out the ­possibility of there being other senses in which the ideas are valid.

4  Transcendental Deductions and their Aims Now that we have reconstructed the transcendental deductions of space, the categories and the ideas, it is time to consider what, if anything, they have in common. Do they follow a similar method? Do they have similar aims? One consequence of the analysis provided in this chapter is that transcendental deductions are not identifiable as a specific kind of

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Transcendental Deductions

167

argument. What transcendental deductions have in common is not their structure but rather their aim. They aim to confirm that root concepts for the cognition of objects have some sort of validity, but they do not say anything about how this validity must be proved. The transcendental deductions use very different forms of argumentation. As we saw, the deduction of space can be understood as an inference to the best explanation, the B-deduction presents the form of a transcendental argument, and the deduction of the ideas combines three different arguments, one of which draws on practical considerations. Moreover, even though it is true that they all aim to confirm that root concepts for the cognition of objects have some sort of validity, when they speak of the ‘objective validity’ of these concepts they refer to very different things. By demonstrating the ‘objective validity’ of the categories and the representations of space and time, the corresponding deductions establish that, through those representations, we cognize something that really pertains to objects. By contrast, the ‘objective validity’ of the ideas that is proved by means of their deduction only expresses what I have called their indirect objective validity and their practical validity. Let me add one last consideration. In this Chapter, I have provided reconstructions of three transcendental deductions. I have used them as a guide to draw certain conclusions regarding transcendental deductions in general. These conclusions concerned methodological and common sense conservatism, the ‘pluralism’ in Kant’s strategies of argument and the view that the deductions only establish ‘positive’ results. What I want to point out is that one can disagree with the details of my reconstructions but still accept the conclusions to which they have brought me. Among these, two in particular are important for the reading of transcendental philosophy defended in this book: (a) the deductions are methodologically conservative in that they establish the validity of ‘root’ concepts, and (b) they only display ‘positive’ results regarding the validity of these concepts.

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

h

d i

0 0

9 8 009

2 2 00

bli h d

li

b C

b id

i

i

Part III

The Method of the Critique of Pure Reason

Introduction to Part III In this part of the book, I consider the method of the critique of pure reason, namely the second discipline contained in the Critique that is charged with accomplishing its task as the doctrine of method of metaphysics. In Chapter 2, I suggested that the chief aim of this discipline is to show that metaphysics can attain architectonic unity. Moreover, in Chapter 1, I maintained that there are two minimal conditions for attributing architectonic unity to a body of cognition. First, the body of cognition in question must possess systematic coherence. Second, it can be viewed as realizing the fundamental ‘idea’ of a science, where this idea must (a) define the fundamental object of that science and (b) prescribe the ordering of the body of cognitions that form that science. In this Introduction, I will say something more regarding how the critique shows that metaphysics can meet these conditions. I will suggest that Kant’s strategy involves both a negative and a positive task. The positive task, however, is not completely captured by Kant’s identification of the positive ‘utility’ of the critique (Bxxiv).

1  The Critique of Pure Reason and the Minimal Conditions of Architectonic Unity Let us consider the first minimal condition of architectonic unity. As I submitted in Chapter 1, I take a body of cognitions to be systematically coherent when: (a) the cognitions belonging to it are interconnected in a way that involves relations of either logical implication, explanatory support or both, and (b) it does not contain contradictions. We can take as an example of characteristic (a) Kant’s description of the relationship between the ‘root’ and ‘derivative’ pure concepts within transcendental philosophy. He writes that the former are more fundamental because they lie at 169

https://doi.org/10.1017/9781009172127.008 Published online by Cambridge University Press

170

The Method of the Critique of Pure Reason

the basis of synthetic a priori claims. Moreover, the latter are analytically derivable from the former, which gives systematicity to how they relate to one another (see A14/B27–8). However, it is in relation to characteristic (b) that it is doubtful whether metaphysics can achieve systematic coherence. This is doubtful because it is not clear whether metaphysics can avoid contradictions. Kant thinks that the history of metaphysics is clear evidence of this problem. In the Discipline of Pure Reason in Polemical Use, Kant identifies two types of contradiction that are problematic for the possibility of metaphysics. The first type comprises the contradictions analysed in the Antinomy of Pure Reason. Kant writes: We had such an apparent antithetic of reason before us above, to be sure, but it turned out that it rested on a misunderstanding, namely that of taking, in accord with common prejudice, appearances for things in themselves, and then demanding an absolute completeness in their synthesis, in one or another way (which were both equally impossible), which could hardly be expected in the case of appearances. (A740/B768)

As we saw in Chapter 3 and will consider again in Chapter 5, the antinomies arise when we assume that there are ‘totalities of appearances’, or, using the words of the quote, when we demand an ‘absolute completeness in their synthesis’. When we make this assumption and try to determine what those totalities look like, we end up with couples of contradictory propositions, where one describes the totality as finite and the other characterizes it as infinite. The antinomies are not the only contradictions that threaten the possibility of metaphysics, though. Kant points to a second type of contradiction: However, such a misunderstanding cannot be alleged and the conflict of reason thereby set aside if, say, it is asserted theistically There is a highest being and asserted atheistically, on the contrary, There is no highest being, or when it is asserted, in psychology, ‘Everything that thinks is of absolutely persistent unity and therefore distinct from all transitory material unity,’ against which someone else asserts, ‘The soul is not an immaterial unity and cannot be exempted from all transitoriness.’ (A741/B769)

Clearly, Kant here identifies two ‘conflicts’ of reason that arise in connection to theological and psychological ideas. He submits that they cannot be solved in the same way as the antinomies, whose solution rest on the claim that the contradiction is illusory because there cannot be any ‘totality of appearances’. Another relevant difference in this second type of conflict is that in this case only one claim within the conflict arises

https://doi.org/10.1017/9781009172127.008 Published online by Cambridge University Press

Introduction to Part III

171

necessarily from reason. If we consider the arguments in the Paralogisms and the Ideal, it is evident that reason is naturally driven to affirm the substantiality and immateriality of the soul and the existence of God but not the opposite of these claims. However, since the arguments that drive reason to these affirmations are fallacious, it cannot use these arguments to legitimately silence anybody who contends exactly the opposite. Therefore, this is also a conflict that is relevant to reason because, on the one hand, it is naturally driven to affirm one horn of the conflict and, on the other, it does not actually have the means to decide the conflict when one affirms the opposite. Taken together, these conflicts constitute the main obstacle that prevents metaphysics from attaining coherence. Famously, Kant’s strategy for avoiding them is to set limits to our cognition. The first task of the critique of pure reason is therefore a negative one. In my account, this negative task is achieved by showing that the understanding of the root concepts for the cognition of objects that was established and legitimated within transcendental philosophy is the only legitimate understanding of these concepts when they are used for the purposes of cognition. I will investigate the negative task of the critique in Chapters 5 and 6. Let us now consider the second minimal condition of architectonic unity. According to it, a body of cognition must be able to be viewed as realizing the fundamental ‘idea’ of a science, where this idea must (a) define the fundamental object of that science and (b) prescribe the ordering of the body of cognitions that form that science. As I argued in Chapter 1, the idea of a science is the correct description of the body of cognitions that form that science and its parts–whole relationships. Given this account, no body of cognition can be viewed as realizing the fundamental idea of a science if it leaves out certain propositions that appear to be essential to the idea of that science. A consequence of this is that coherence, as the first condition for attributing architectonic unity, cannot be achieved by simply leaving out propositions that, while seemingly essential to the idea of a science, are nonetheless a source of contradictions within a body of cognitions purporting to become a science. As we saw in Chapter 1, it is clear that the propositions asserting freedom, the immortality of the soul and the existence of God are essential to the idea that is the only candidate for giving architectonic unity to metaphysics. Recall that it is the ‘worldly’ concept of metaphysics that can provide this unity, which is the concept of metaphysics that gives centre stage to the highest good as the ‘final end’ of our reason. Recall, also, that freedom, the immortality of the soul and God are conditions for the

https://doi.org/10.1017/9781009172127.008 Published online by Cambridge University Press

172

The Method of the Critique of Pure Reason

realization of the highest good. This explains why metaphysicians have always tried to provide speculative proofs that we are free, that we are immortal, and that God exists. Given the importance of these objects for metaphysics, how can the critique show that it can achieve coherence? It cannot simply set limits to our use of root concepts in a way that prevents us from claiming cognition that we are free, that the soul is immortal and that God exists. Since we have a practical interest in these objects, we will simply come back to attempting theoretical proofs of their existence. Therefore, the critique includes arguments to the effect that we can justifiably commit ourselves to the existence of these objects, even though we cannot cognize their existence. This is the upshot of the practical argument contained in the Canon of Pure Reason. The positive task that rightfully belongs to the critique is showing that metaphysics, including its theoretical and practical parts, can achieve coherence while still including everything that is essential to the ‘idea’ that we at least have in view. With respect to our commitments regarding freedom, immortality and the existence of God,1 this means that it must show that these commitments do not conflict with the limits imposed on cognition by the ‘negative’ part of the critique. If there were such a conflict, these commitments would endanger the very instrument that the critique has devised to avoid the contradictions of reason identified above. Kant’s characterization of belief (Glaube) has the purpose of avoiding this conflict and belongs to the tools of the critique. This will be the topic of Chapter 7.

2  The Negative and the Positive ‘Utility’ of the Critique I now wish to spell out how my way of characterizing the positive and negative tasks of the critique relates to Kant’s description of its negative and positive ‘utility’ in the B-Preface. Let us begin with the former. Kant equates it with the injunction ‘never to venture with speculative reason beyond the boundaries of experience’ (Bxxiv). In the Introduction and the Canon of Pure Reason, by contrast, it is identified with the erection of boundaries of cognition and the avoidance of error (A11/B25; A795/B823). While Kant does not speak explicitly of limiting the use of root concepts to the understanding that transcendental philosophy legitimates here, there is fundamental agreement between what I call the negative ‘task’ of the critique and what Kant calls its negative ‘utility’. This is so because limiting our use of root concepts to the understanding that transcendental philosophy 1

As we will see in Chapter 7, in the Canon Kant actually only focuses on God and immortality.

https://doi.org/10.1017/9781009172127.008 Published online by Cambridge University Press

Introduction to Part III

173

legitimates does set boundaries to our cognition. Transcendental philosophy establishes that root concepts have objective validity with respect to appearances. When we say that this is the only legitimate understanding of these concepts for cognition, we are clearly setting boundaries of this kind. Moreover, the ‘errors’ Kant mentions when speaking of negative ‘utility’ arise when we use root concepts in a way that goes beyond what transcendental philosophy legitimizes. In addition to the negative ‘utility’ of the critique, the B-Preface also identifies a much more relevant ‘positive’ utility (Bxxv). First, in speaking of freedom, Kant argues that the distinction between appearances and things in themselves allows us to regard the concept of freedom as noncontradictory and at least logically possible (Bxxvii–xxviii). When he adds God and immortality to the picture, he submits that the critique establishes that the existence of these objects must necessarily remain undecidable from a theoretical point of view. Therefore, the positive ‘utility’ of the critique can be equated with establishing that any attempt to theoretically disprove that either freedom, God or immortality obtain must fail. While this achievement is obviously important from the perspective of the practical part of metaphysics, this account of the positive ‘utility’ of the critique cannot explain the inclusion in the Canon of Kant’s practical argument for the commitment to God and immortality and his characterization of belief. Accordingly, in Chapter 7, I will argue that the positive task of the critique is not limited to what Kant calls its positive ‘utility’. One might ask how this picture fits into Kant’s moderate methodological conservatism. The critique is conservative because it aims to preserve everything that seems essential to our idea of metaphysics.2 It is moderately conservative because, given the constraint of coherence, it sometimes requires a radical revision of our understanding of some of the elements that we regard as essentially belonging to metaphysics in order to make them fit into a coherent whole.3 2

3

Let me emphasize, as I did in the Introduction to Part II, that I am not suggesting that Kant is conservative with respect to the claims of traditional metaphysics. Rather, he thinks that we are in possession of a priori concepts and principles that we regularly use in various aspects of our lives. These concepts and principles constitute the proper object of metaphysics, and it is with respect to them that we must be conservative. Also here, let me prevent a possible misunderstanding. In speaking of coherence, I do not want to attribute to Kant the view that in philosophy we must start with our common sense commitments and adjust those commitments when they give rise to contradictions. Rather, metaphysics must first isolate a priori concepts and principles that play a central role in different aspects of our lives. It is only after these concepts and principles have been identified that a coherence constraint emerges, since we must guarantee that the system of these concepts and principles is coherent.

https://doi.org/10.1017/9781009172127.008 Published online by Cambridge University Press

chapter 5

The Negative Side of the Critique of Pure Reason

According to my reconstruction, the first task of the critique of pure r­ eason is to set limits as to the validity of the root concepts for the c­ ognition of objects analysed by transcendental philosophy. The first aim of this ­chapter is to investigate how Kant sets these limits. More precisely, I wish to make clear the sense in which the critique of pure reason uses the results of ­transcendental philosophy. The second aim of the chapter is to ­determine where Kant argues for these limits within the Critique of Pure Reason. In this respect, I will show that each main part of the Transcendental Doctrine of Elements contains arguments that are essential to establishing that the root concepts analysed there have a limited validity. I will start by analyzing Kant’s remarks on what made a transcendental deduction of the concept of space ‘unavoidably necessary’, since these remarks provide a first clue for determining how the negative part of the critique of pure reason depends on transcendental philosophy. These remarks are ambiguous, however. I will therefore examine how Kant supports the claim that things in themselves do not have spatial properties in the Aesthetic. Because Chapter 6 considers in detail how Kant sets limits to the valid use of the categories in the Analytic, I move on to examine the limits to cognition that are established in the Dialectic. This examination will first show that the ‘clue’ provided by Kant’s remarks on the unavoidable necessity of the transcendental deduction of the concept of space cannot offer a general model for determining how the negative side of the critique of pure reason depends on transcendental philosophy. I will then analyse a passage in which Kant introduces a distinction between ‘transcendent principles’ of reason and the ‘transcendental use’ of the categories. The passage suggests that the Dialectic does not contribute to setting limits to cognition. Rather, its role is to diagnose the natural disposition of our reason to use the categories in a way that violates those limits. In a last move, I will show that, at least when we consider the cosmological ideas, the function of the Dialectic is not only to diagnose 174

h

d i

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Negative Side of the Critique of Pure Reason

175

this natural tendency but to introduce essential materials for establishing that the cosmological ideas are unfit to provide cognition of the objects they represent.

1  The ‘Unavoidable Necessity’ of a Transcendental Deduction of the Concept of Space A central tenet of the interpretation I propose in this book is that the critique of pure reason, as that discipline within the Critique that is responsible for accomplishing the latter’s task as the doctrine of method of metaphysics, rests on the establishment of some parts of transcendental philosophy: more precisely, it rests on those parts that have to do with the origin and validity of root concepts for the cognition of objects. But how is this relationship of dependence to be understood? What are the results of transcendental philosophy that the critique uses for its purposes? In a passage introducing the transcendental deduction of the categories where Kant explains what makes a ‘transcendental deduction’ of the concept of space ‘unavoidably necessary’ (A87/B119), Kant gives the impression that it is specifically the results of transcendental deductions that the critique of pure reason uses to set limits to cognition. Kant first explains that a deduction of the concepts of space and time and the categories can only be transcendental. Because these concepts are a priori, an ‘empirical’ deduction is inadequate for them (A85–6/B118). This, however, is insufficient to explain why a transcendental deduction of the concept of space is ‘unavoidably necessary’: We have above traced the concepts of space and time to their sources by means of a transcendental deduction, and explained and determined their a priori objective validity. Geometry nevertheless follows its secure course through strictly a priori cognitions without having to beg philosophy for any certification of the pure and lawful pedigree of its fundamental concept of space. Yet the use of the concept in this science concerns only the external world of the senses, of which space is the pure form of its intuition, and in which therefore all geometrical cognition is immediately evident because it is grounded on intuition a priori, and the objects are given through the cognition itself a priori in intuition (as far as their form is concerned). With the pure concepts of the understanding, however, there first arises the unavoidable need to search for the transcendental deduction not only of them but also of space, for since they speak of objects not through predicates of intuition and sensibility but through those of pure a priori thinking, they relate to objects generally without any conditions of sensibility; and since they are not grounded in experience and cannot exhibit any object in

h

d i

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

176

a priori intuition on which to ground their synthesis prior to any experience, they not only arouse suspicion about the objective validity and limits of their use but also make the concept of space ambiguous by inclining us to use it beyond the conditions of sensible intuition, on which account a transcendental deduction of it was also needed above. (A87–8/B119–21)

At the beginning of the quote, Kant submits that he has provided a transcendental deduction of the concepts of space and time in the Transcendental Aesthetic. Kant explains that it was not the role of the transcendental deduction of space to provide a validation of geometrical knowledge. Geometry can be considered legitimate a priori knowledge even in the absence of a transcendental deduction. As we saw in Chapter 4, the transcendental exposition of the concept of space, which I read as its transcendental deduction, starts from the assumption that geometry is valid cognition. In this sense, it seems obvious that it cannot offer a validation of geometry. What did the transcendental deduction of the concept of space offer for our understanding of geometry? It offered an explanation of how it is possible for us to have a priori geometrical knowledge that also constrains all outer intuition. Providing this explanation is certainly important from the standpoint of a philosophical account of geometry. However, it is not ‘unavoidably necessary’ in the sense that it is not the case that our geometrical knowledge would be threatened if we were unable to provide this deduction. What makes the transcendental deduction of space ‘unavoidably necessary’ is that we sometimes use the concept in ways that are problematic. Kant suggests that the categories, which are concepts of objects in general and so can be used for thinking objects independently of the conditions of sensibility, drive us to use the concept of space beyond the limits of possible intuition, thus making this concept ‘ambiguous’. These remarks are interesting for two main reasons. First, they make clear that it is not the purpose of transcendental deductions (or of metaphysics) to offer a validation of knowledge claims in the sciences. If the concept of space is central to geometry (and, we might add, the concept of a cause central to physics), it is not the case that without a transcendental deduction this science would be in danger.1 1

h

d i

0 0

This is of course different from saying that a transcendental deduction is useless if the concepts that are being deduced are not sometimes used problematically. Transcendental deductions, together with the metaphysical deductions that they presuppose, provide a philosophical explanation of how certain cognitions within a science are possible. Even though they do not offer a foundation of that science, they are nonetheless genuine pieces of (philosophical) knowledge.

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Negative Side of the Critique of Pure Reason

177

The second reason why Kant’s remarks are interesting is that they distinguish between two roles played by transcendental deductions. At a first level, transcendental deductions are ‘explanation[s] of the way in which concepts can relate to objects a priori’ (A85/B117).2 It is at this level that they can provide an explanation of the possibility of cognition in other sciences. What makes a transcendental deduction ‘unavoidably necessary’ is that it can contribute to setting limits to the use of a concept in ways that prevent its illegitimate application. In my account, it is this role that belongs to the critique of pure reason. If my reconstruction of transcendental deductions in the previous chapter is correct, transcendental deductions are essential parts of transcendental philosophy, and, as such, their arguments do not directly establish limits to the valid use of concepts. Therefore, I submit that when Kant says that transcendental deductions are ‘unavoidably necessary’ for setting those limits, this should be read as stating that transcendental deductions are instrumental to identifying those limits within the critique of pure reason. Let us return to the concept of space, which is the main focus of the passage. Kant is suggesting that there is something in the transcendental deduction of the concept of space that can be put to work in establishing that the concept of space can only be legitimately used for appearances. At this point, there is a further complication that needs to be considered. When Kant speaks of the ‘transcendental deduction’ of the concept of space in the above passage, he seems to also encompass what I have characterized as its metaphysical deduction. This is the case because the transcendental deduction of the concept of space is said to concern not only its validity but also its origin (A87/B119–20). Bear in mind that the passage was already contained in the A-edition of the Critique, where Kant had not yet distinguished between metaphysical and transcendental expositions. Therefore, it is not surprising that the question of the origin and the question of the validity of the concepts of space and time were not clearly separated. This has an important consequence for my investigation. Since Kant mentions both of these questions in reference to the ‘transcendental deduction’ of the concept of space, it is unclear whether his argument concerning the limits of the legitimate use of the concept of space builds on the conclusions of the metaphysical exposition (which I read as the 2

h

d i

Of course, this characterization of transcendental deductions is only applicable to the transcendental deductions of the concepts of space and time in the Aesthetic and the transcendental deduction of the categories in the Analytic. It does not square well with what Kant does in the transcendental deduction of the ideas, unless in this latter case one takes the ‘relation to objects a priori’ to simply imply what I have characterized as the ‘practical validity’ of the ideas (see Chapter 4).

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

178

The Method of the Critique of Pure Reason

metaphysical deduction of the concept of space), the transcendental exposition (which I read as its transcendental deduction), or both. In what follows, I will first determine how Kant’s argument for the non-applicability of the concept of space to things in themselves works. In particular, I will consider whether this argument builds only on the results of the transcendental exposition of space or whether it also fundamentally rests on points that Kant makes in the metaphysical exposition. Adopting Kant’s argument regarding the limits of the valid use of the concept of space as paradigmatic of the Aesthetic, I will consider whether a similar procedure is used for the categories and the ideas. a. Kant’s argument against the applicability of the concept of space to things in themselves. In the section entitled Conclusions from the Above Concepts that directly follows the transcendental exposition of space (hereafter Conclusions), Kant argues that things in themselves cannot have spatial properties (A26/B42). How does Kant support his claim? According to the so-called ‘neglected alternative’ objection, Kant does not have a valid argument in support of that claim (for useful reconstructions of the origin of the objection, see G. Bird 2006a and Specht 2014). Roughly, the objection has it that Kant illegitimately assumed that space is either a form of intuition and something that pertains to appearances or something that pertains to things in themselves. Since Kant was able to prove that space is a form of intuition, he concluded that it cannot pertain to things in themselves; the conclusion is unwarranted, however, since space can be both a form of intuition and something that pertains to things in themselves. There have been various attempts to save Kant from this objection. Some have tried to find indirect support for his claim about the non-spatiality of things in themselves in sections of the Aesthetic where he does not explicitly make the point.3 Others have appealed to a particular interpretation of transcendental idealism (see Allison 2004: 3

h

d i

0 0

Falkenstein (1995: 301–4) identifies an argument for the non-spatiality of things in themselves in the metaphysical exposition. It rests on the distinction between different kinds of order: presentational order, causal order and comparative order. The argument claims that the spatial order of appearances is presentational, while the order of things in themselves cannot be presentational. According to Falkenstein, even this argument is unable to establish that there is no sense in which things in themselves might have a spatial order. In their rejection of the neglected alternative objection, Marcus Willaschek (1997) and Lucy Allais (2015: Ch. 8) rely on Kant’s characterization of intuition in the introductory section of the Aesthetic, where intuitions are said to be ‘singular’ and ‘immediate’ representations. From these features and the additional claim that our intuitions of space and time are not causally derived from anything extra-subjective they conclude that the pure intuitions of space and time cannot represent features of things in themselves.

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Negative Side of the Critique of Pure Reason

179

130–2; Rosefeldt 2016). My purpose here is to reconstruct how Kant arrived at the claim that things in themselves do not have spatial properties in the Conclusions. Furthermore, I will try to understand what that claim exactly means. Let me first recall the results of the transcendental exposition of the concept of space. The argument established that what we cognize through our concept and intuition of space are properties objects have as appearances. This followed from two main considerations. First, geometry is synthetic cognition that constrains all outer intuitions. We could not account for that cognition if we did not assume a pure intuition of space that is identical to (or depends on) space as the form of outer intuition. Second, our concept of space does not have meaning independently of our intuition of space, because the content of the former depends on cognitions of space we obtain through the latter. Since the argument concerned what we know through our concept and intuition of space, it did not say anything about whether things in themselves have spatial properties. I also suggested that the argument did not rule out, at least explicitly, the possibility of our indirectly cognizing properties that objects have in themselves through our concept and pure intuition of space. Proving that this is possible would require showing that objects in themselves have spatial properties and that these necessarily align with the spatial properties we cognize through our pure intuition and concept of space. The way Kant introduces the notion that things in themselves do not have spatial properties in the Conclusions gives the impression that he is simply boldly making this claim without providing an argument: (a) Space represents no property at all of any things in themselves nor any relation of them to each other, i.e., no determination of them that attaches to objects themselves and that would remain even if one were to abstract from all subjective conditions of intuition. For neither absolute nor relative determinations can be intuited prior to the existence of the things to which they pertain, thus be intuited a priori. (b) Space is nothing other than merely the form of all appearances of outer sense, i.e., the subjective condition of sensibility, under which alone outer intuition is possible for us. Now since the receptivity of the subject to be affected by objects necessarily precedes all intuitions of these objects, it can be understood how the form of all appearances can be given in the mind prior to all actual perceptions, thus a priori, and how as a pure intuition, in which all objects must be determined, it can contain principles of their relations prior to all experience. (A26/B42)

h

d i

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

180

The Method of the Critique of Pure Reason

Is Kant simply proclaiming a bold thesis in these passages, or do they contain an argument? In order to answer this question, let me first point out that ‘space’ is here considered not as an object or a property of an object but as a representation (for a similar reading, see Allais 2015: 198). Kant accordingly starts the first sentence by writing ‘space represents’, where it is plausible to read this as ‘the representation of space represents’. This is confirmed in a passage a few lines after the quote, where Kant writes: ‘If we depart from the subjective condition under which alone we can acquire outer intuition […], then the representation of space signifies nothing at all’ (A26/B42, my emphasis). Starting from this assumption, we can now interpret (a). The passage can be taken to argue that our concept and intuition of space cannot provide indirect cognition of properties of objects in themselves. Kant maintains that because our representation of space is a priori, it cannot be derived from properties that pertain to objects in themselves. If it were, our representation would be a posteriori. In what sense does this support the claim that our concept and intuition of space cannot provide indirect cognition of properties that things have in themselves? The point follows if one adds the following considerations. We know from the transcendental exposition of space that our concept and intuition of space provide us with valid a priori cognition regarding appearances. The argument in (a) submits that any cognition of the features of space derived from the properties of things in themselves would be a posteriori. It follows from this that if there is any agreement between the cognition of the features of space we do have (cognition that is a priori) and the properties that things in themselves might have (properties that would yield a posteriori cognition), this agreement is accidental. But an accidental agreement between cognition of appearances, on the one hand, and properties of objects in themselves, on the other, would be insufficient to establish indirect cognition of objects in themselves. Therefore, since the relationship between the cognition of appearances that we get through our concept and intuition of space and the properties of things in themselves is at best one of accidental agreement, it is in principle impossible for our concept and intuition of space to yield indirect cognition of the properties of objects in themselves. Note that up to this point, the argument does not establish that things in themselves cannot have spatial properties of their own. It only establishes that through our concept and intuition of space we cannot in principle obtain cognition of those spatial properties, if they exist. Kant does make bolder claims regarding the spatial properties of things in themselves, however. For example, he writes that if we were to take away our forms of sensibility, ‘space and time themselves would disappear’ (A42/B59),

h

d i

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Negative Side of the Critique of Pure Reason

181

which suggests that things in themselves cannot have spatial properties. Kant does not present an explicit argument for this additional point in the Aesthetic.4 One can help him out, however, by drawing on his account of the concept of space. In Kant’s general account of representations, concepts are the only representations that can be used to discuss things in themselves. This is because all our intuitions must agree with the a priori forms of space and time, which, as we have seen, are the forms of our sensibility. Even though concepts need to be connected to intuition to yield cognition of objects, they can nonetheless represent objects independently of the conditions of sensibility. It is for this reason that, if we have to speculate about what things in themselves might be like, we can only do so through concepts. One assumption that Kant makes in the transcendental exposition of space is that the concept of space derives from the pure intuition of space. I suggested that this means, first, that we would not have any concept of space if we had no pure intuition of space and, second, that all cognitions of spatial properties of objects that we have through the concept of space are first of all cognitions that we have through the pure intuition of space. In this sense, the concept of space is solely an expression of our pure intuition of space. This implies that what we represent through the concept of space lacks the required independence from our sensible intuition that allows other concepts to be used to speculate about what objects might be like independently of the conditions of our sensibility. As a consequence of this, the concept of space cannot be used to speculate about properties that things in themselves might have. It is in this sense that things in themselves cannot have spatial properties. Even if things in themselves had properties that in a certain way agreed with or were similar to the spatial relationships of appearances, it wouldn’t make sense to call those properties spatial.5 It wouldn’t make sense because our concept of space essentially depends on our intuition of space. Accordingly, it cannot be used to speculate about things in themselves. Of course, this does not rule out their (accidentally) having properties that in a certain way agree with or are similar to spatial properties of appearances (for a similar concession, see Allais 2015: 198). 4 5

h

d i

An argument to this effect can be found in the indirect argument for transcendental idealism in the Antinomy of Pure Reason. This comes close to a point made by Willaschek (1997: 554–5). We take different paths to this position, however. Willaschek bases the contention that it makes no sense to use the term ‘spatial’ for things in themselves on a comparison between Kant’s account of intuition and causal theories of reference. Moreover, he only makes the point with respect to the intuition of space. By contrast, I have tried to extend the point to the concept of space.

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

182

The Method of the Critique of Pure Reason

It is sufficient, however, to prove that things in themselves cannot have spatial properties. Now that we have a reconstruction of Kant’s argument in the Conclusions, it is time to return to our question regarding the parts of transcendental philosophy on which Kant’s argument for the limits of the validity of our concept of space rests. More precisely, we can now answer the question of whether Kant’s argument builds only on the results of the transcendental exposition of space or whether it also rests on the metaphysical exposition. In my reconstruction, I have repeatedly appealed to claims that Kant makes in the transcendental exposition of space. Kant’s argument regarding the limits of the concept of space therefore clearly rests on it. Notice, however, that Kant’s contention in the transcendental exposition that the concept of space is merely an expression of cognitions that we obtain through our pure intuition of space is partly based on a point he makes in the metaphysical exposition. The point in question is that the pure intuition of space is more fundamental than its concept because the former gives access to a priori knowledge that the latter would not provide (like ‘space is singular’), while also giving access to all a priori knowledge obtainable through the concept of space. Therefore, it seems appropriate to claim that Kant’s argument concerning the limits of the legitimate use of our concept of space rests on both the metaphysical and the transcendental expositions of this concept. b. The limits of the validity of the categories and the ideas. The question at this point is whether the relationship between the metaphysical and the transcendental deductions of space,6 on the one hand, and the argument regarding the limits of the validity of the concept of space, on the other, provides a general model for how the ‘negative’ side of the critique of pure reason relates to the parts of transcendental philosophy on which it depends. I will assume that what Kant does for space is paradigmatic of the entire Aesthetic (and thus that Kant proceeds in a similar way when it comes to establishing limits for our use of the concept of time). A similar relationship between the negative side of the critique and transcendental philosophy might be found in the context of the categories. After all, in the passage in which Kant discusses the ‘unavoidable necessity’ of a transcendental deduction of the concept of space – the passage that provided the 6

h

d i

0 0

Recall that in my account the metaphysical and transcendental expositions of space and time are its metaphysical and transcendental deductions, respectively.

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Negative Side of the Critique of Pure Reason

183

clue for our analysis above – Kant also speaks of the ‘unavoidable necessity’ of the transcendental deduction of the categories (A87–8/B119–21). It is therefore reasonable to expect that Kant’s argument regarding the limits of the validity of the categories rests at least on the transcendental deduction of these concepts.7 It is an open question whether the argument also relies on the metaphysical deduction in a way that is similar to the argument regarding the concept of space. Since I will discuss how transcendental philosophy and the negative side of the critique are related in the B-deduction of the categories in Chapter 6, I will not consider the issue in further detail here. The problems associated with using Kant’s arguments for space as a general model for determining the relationship between transcendental philosophy and the critique of pure reason become evident when one considers transcendental ideas. In this case, the transcendental deduction of the ideas cannot possibly provide an essential tool for determining how the ideas cannot be used. This is because, as we saw in the previous chapter, Kant places the transcendental deduction of the ideas at the end of the Dialectic. At this point, he has already established that the ideas cannot yield direct cognition of objects. Kant does not first establish the sense in which these concepts are valid and then determine the limits of this validity. Instead, he starts by establishing that the ideas are unfit to provide cognition of the objects they represent. It is only after this point is made that Kant identifies the limited validity of the ideas as heuristic principles that guide our attempts to ‘perfect’ the cognitions of the understanding. ­ octrine Therefore, if the negative points of the Dialectic rest on a positive d belonging to transcendental philosophy, the relationship between this ­latter discipline and the critique of pure reason will be different from the one we found in the arguments for the origin, validity and limits of the concept of space. In Sections 2 and 3, I will consider how Kant uses the results of transcendental philosophy to demonstrate that ideas are unfit to provide cognition of the objects they represent. Furthermore, I will show that, at least when we consider the cosmological ideas, the Dialectic 7

h

d i

Even though Kant first explicitly identifies a metaphysical deduction of the categories in the B-edition of the Critique of Pure Reason (B159), it is clear that when he speaks of the ‘unavoidable necessity’ of the transcendental deduction of the categories in the A-edition, he cannot have intended to include what in the B-edition is called their metaphysical deduction. This is simply because, in both editions, the sections containing what in the B-edition is called the metaphysical deduction are clearly separated from the transcendental deduction. Therefore, in the case of the ‘unavoidable necessity’ of the latter deduction, we do not find the ambiguity that we identified regarding Kant’s reference to the transcendental deduction of space.

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

184

does not simply rest on the result of the Analytic when it comes to determining the limits of cognition. Rather, the Dialectic contributes to establishing the unfitness of ideas for the cognition of objects.

2  ‘Transcendent Principles’ of Reason and the ‘Transcendental Use’ of the Categories In Chapter 3, we saw that Kant gives contrasting accounts of the origin of the transcendental ideas. By focusing on how Kant completes the derivation of the cosmological ideas in the Antinomy of Pure Reason, we established that, at least in this case, Kant first provides a general characterization of what the ‘supreme principle’ of reason is, namely, the principle that states that when a conditioned is given, the complete series of its conditions is also given. The individual cosmological ideas are obtained by using different categories to think about these totalities of conditions leading up to the unconditioned for a specific domain of objects (that is, appearances). In short, the cosmological ideas are the result of an application of the categories that is encouraged by the ‘supreme principle’. Extending this characterization to all the transcendental ideas, Kant writes that ‘transcendental ideas will really be nothing except categories extended to the unconditioned’ (A409/B436). Kant provides an account of the relationship between the ‘supreme principle’ and the categories that is fundamentally consistent with this picture in the following passage: We will call the principles whose application stays wholly and completely within the limits of possible experience immanent, but those that would fly beyond these boundaries transcendent principles. But by the latter I do not understand the transcendental use or misuse of categories, which is a mere mistake of the faculty of judgement when it is not properly checked by criticism, and thus does not attend enough to the boundaries of the territory in which alone the pure understanding is allowed its play; rather, I mean principles that actually incite us to tear down all those boundary posts and to lay claim to a wholly new territory that recognizes no demarcations anywhere. Hence transcendental and transcendent are not the same. The principles of pure understanding we presented above should be only of empirical and not of transcendental use, i.e., of a use that reaches out beyond the boundaries of experience. But a principle that takes away these limits, which indeed bids us to overstep them, is called transcendent. If our critique can succeed in discovering the illusion in these supposed principles, then those principles that are of merely empirical use can be called, in opposition to them, immanent principles of pure understanding. (A295–6/B352–3)

h

d i

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Negative Side of the Critique of Pure Reason

185

The passage introduces two distinctions. The former distinguishes between immanent and transcendental uses of the categories, while the latter differentiates between immanent and transcendent principles. By immanent use of the categories, Kant means their employment within the boundaries of possible experience. By contrast, their transcendental use surpasses those boundaries and employs the categories to determine what objects in themselves are like. If we now move to the second distinction, Kant characterizes transcendent principles as those that incite us to surpass the boundaries of experience. This is consistent with what Kant says about the supreme principle of reason, which, as we saw, drives us to use the categories to determine the idea of the unconditioned. An immanent principle, by contrast, is a principle of the understanding that remains within its legitimate use after the illusion created by transcendent principles of reason is revealed. One interesting thing about this passage is that it affirms that, if it were not for the incitement produced by transcendent principles of reason, our transcendental and illegitimate use of the categories would be ‘a mere mistake of the faculty of judgement when it is not properly checked by criticism’. In other words, the transcendental use of the categories would not arise from any necessary ‘disposition’ characteristic of our reason. Since we know that it is specifically the categories that the supreme principle of reason incites us to use illegitimately, the function of identifying a transcendent principle of reason is specifically that of showing that the transcendental use of the categories is not simply the effect of an accidental misuse of these concepts but is rather due to a misapplication that arises with a certain necessity.8 From this characterization of the transcendent principles of reason, we get the impression that the Dialectic is not really responsible for showing that the ideas are unfit to provide cognition of the objects they represent. Since the ideas are just ‘categories extended to the unconditioned’ (A409/ B436), the critique of the transcendental use of the categories does the real limiting work, which presumably occurs in the Analytic. The main responsibility of the Dialectic is to provide a diagnosis of why we do not simply accidentally fall into the illegitimate use of the categories but are rather encouraged to use them transcendentally due to a certain disposition of our reason.

8

h

d i

For a recent helpful account of the necessity with which questions about the unconditioned arise according to Kant, see Willaschek (2018: esp. Part I). For an alternative view, see Grier (2001).

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

186

3  Why are the Cosmological Ideas Invalid for Cognition? In this section, I will reconstruct Kant’s argument for the unfitness of the cosmological ideas to provide cognition of objects. I focus on the cosmological ideas because they make clear that, at least in this case, the Dialectic does not simply rest on the Analytic when it comes to establishing certain limits of cognition. In particular, I wish to show that without the identification and critique of the ‘supreme principle’ of reason, Kant would be unable to claim that the cosmological ideas are unfit to provide direct cognition of objects. That is to say, the identification of the limits of the valid use of the categories would be insufficient to prove that the cosmological ideas are illegitimate when used as cognitions. Therefore, at least in the case of the cosmological ideas, the identification and critique of the supreme principle of reason provides not only a diagnosis of the natural tendency of our reason to surpass the boundaries of experience but also an essential tool for determining that the cosmological ideas cannot be cognitions of objects. a. Totalities of appearances. What is distinctive about the cosmological ideas? As we saw in Chapter 3, Kant is quite clear that these ideas describe totalities of appearances (A408/B434–5; A416/B444). Recall that the cosmological ideas represent the complete composition of appearances, the complete division of appearances, the complete causal chain explaining the origin of an appearance and the complete set of conditions of something contingent, respectively (A415/B443). Recall also that the cosmological ideas arise from a joint application of the ‘supreme principle’ and one of the categories of totality, reality, causality and necessity/contingency, respectively, to the determination of what one must assume to make sense of a given empirical conditioned (more precisely: an experienced moment in time or an experienced portion of space; a complex substance in space; a causal event; a contingent alteration). When Kant stresses that cosmological ideas describe totalities of appearances, he means that we represent this totality as formed by members that are fundamentally analogous to the ‘conditioned’ we assumed at the start. Since this conditioned is empirically given and stands under the conditions of our sensibility, this means that we think of each member of the assumed totalities as standing under the same conditions. This is evident when one considers the idea of the ‘complete composition of appearances’ that lies behind the first antinomy (A426–33/B454–61). As is well known, the idea has two versions, one concerning the existence in time of the world-whole, and one concerning its extension in space. The given ‘conditioned’ that are assumed in the two cases are an experienced

h

d i

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Negative Side of the Critique of Pure Reason

187

moment in time and an experienced portion of space. The totalities of conditions that are obtained starting from those conditioned are the totality of elapsed moments in time preceding the ‘given’ moment and the totality of portions of space enclosing the ‘given’ portion. In these totalities, all the members are thought to be fundamentally analogous to the initially given ‘conditioned’.9 Obviously, all members of the two totalities are either in time or in space. Moreover, even if they are not all actually experienced moments of time and portions of space, all members of the two totalities are thought to be fundamentally connected to the actually experienced moments in time and portions of space. In this sense, each portion of space is thought to be in principle experienceable. Similarly, each elapsed moment of time is thought to have been in principle experienceable as it constituted present time. One can here object that when Kant explains the difference between his resolutions of the first and second antinomies, on the one hand, and the third and fourth antinomies, on the other, he claims that it is only the ideas of totalities lying behind the ‘mathematical’ antinomies that contain ‘a synthesis of homogeneous things’ (A530/B558), whereas the ideas grounding the ‘dynamical antinomies’ contain ‘a synthesis of things not homogeneous’ (A530/B558). What Kant means by this is that the first two antinomies have to do with magnitudes in which each member of the series must be thought as fundamentally analogous. A magnitude is by definition numerically measurable. It is a condition of its being so measurable that it is homogeneous. By contrast, the third and fourth antinomies have to do with dynamical relationships between things, where the paradigmatic example is the causal relationship between cause and effect. In these relationships, the relata need not necessarily be analogous.10 At 9

10

h

d i

True, the thesis of the first antinomy submits that we must assume that the world has a beginning in time and that it has limits in space (A426/B454). In this sense, the beginning of the world in time and the portion of space at its limits are in one respect disanalogous to the other members of the relevant totalities. Unlike any other moment in elapsed time, the beginning of the world is not preceded by any state of the world. Similarly, unlike any other portion of space, the portion of space enclosing the world is not enclosed by any other portion of the world in space. Notice, however, that the argument of the thesis submits that assuming these limiting members of the respective totalities is the only way to make the very idea of these totalities coherent. When one disregards the differences between the limiting members that are assumed for the purposes of preserving coherence, the thesis agrees that these members must be fundamentally analogous to the other members. This distinction is crucial for understanding Kant’s claim that while in the mathematical antinomies both the theses and the antitheses are false, in the dynamical antinomies they can both be true. More precisely, the claim supports the contention that the third and fourth theses can be true. Since the series at stake are not necessarily homogeneous in this case, a first member of a series can be postulated as existing in the ‘intelligible’ world (A530–2/B560).

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

188

this point, one can object that if the series at stake in the third and fourth antinomies arise from the synthesis of things that are not homogeneous, and if they can contain members existing in the ‘intelligible’ world, this means that the ideas behind these antinomies cannot really be formed by members that are fundamentally analogous or by totalities of only appearances. To answer this objection, it is important to keep in mind that it is when Kant is working on his solution to the antinomies that he stresses that the series at stake in the dynamical antinomies are not necessarily syntheses of homogeneous things. Kant is equally clear, however, that in the antinomies themselves, all the totalities at stake are thought to arise from a synthesis of homogeneous things: When we represented the antinomy of pure reason in a table through all the transcendental ideas, […] we in all cases represented the conditions for their conditioned as belonging to relations of space and time, which is the usual presupposition of common human understanding, on which, therefore, the conflict entirely rested. In this respect all dialectical representations of totality in the series of conditions for a given conditioned were of the same kind throughout. There was always a series, in which the condition was connected with the conditioned as a member of the series, and thereby was homogeneous […]. (A528/B556)

Therefore, all the antinomies arise in connection with ideas of reason that describe totalities of appearances in space and time. This has important consequences for how we should understand Kant’s argument that the cosmological ideas are unfit to provide direct cognition of objects. In this case, the limits to the valid use of the categories established in the Analytic are insufficient to determine that the ideas are illegitimate for direct cognition of objects. Since the totalities at stake in the cosmological ideas are thought to be totalities of appearances, when we use the categories to determine the form and shape of these totalities, it seems prima facie that we are not violating any restrictions concerning the valid use of the categories. Therefore, in order to prove that the cosmological ideas are unfit for the direct cognition of objects, we need a further tool with respect to the one already established in the Analytic in the context of the limited validity of the categories. b. The unfitness of the cosmological ideas for cognition. What is the tool that the Dialectic provides for proving the unfitness of the cosmological ideas for the direct cognition of objects? Put briefly, the Dialectic proves that the ‘supreme principle’ of reason is invalid when the ‘conditioned’ that is

h

d i

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Negative Side of the Critique of Pure Reason

189

assumed at the start is an appearance. Through this critique Kant can show that it is wrong to assume that anything like a ‘totality of appearances’ is possible. When, in the cosmological ideas, we assume that a totality of appearances is given together with the ‘conditioned’ from which we start, we surpass the limits of our possible experience, even though the members of the totalities we are considering are all thought to be analogous to that ‘conditioned’: Above I have called the ideas with which we are now concerned ‘cosmological ideas,’ partly because by ‘world’ is understood the sum total of all appearances, and our ideas are also directed only toward the unconditioned among appearances, but partly too because in the transcendental sense the word ‘world’ signifies the absolute totality of the sum total of existing things, and we are directing our attention only to the completeness of the synthesis (though properly only in the regress toward its conditions). Considering, moreover, that taken collectively these ideas are all transcendent and, even though they do not overstep the object, namely appearances, in kind, but have to do only with the sensible world (not with noumena), they nevertheless carry the synthesis to a degree that transcends all possible experience; thus in my opinion one can quite appropriately call them collectively world-concepts. (A419–20/B447)

In this quote, Kant distinguishes between two explanations of why it is appropriate to call the cosmological ideas ‘world-concepts’. Relatedly, he identifies two senses of the term ‘world’. In a first sense, ‘world’ means ‘the sum total of all appearances’, which aligns well with the fact that the cosmological ideas are thought to describe totalities of appearances. In a second, transcendental sense the term ‘world’ means ‘the sum total of existing things’, which is a totality that is not any more bound to the conditions of our sensibility. Kant concludes that it is legitimate to consider the cosmological ideas ‘world-concepts’ because they – in agreement with the first meaning of the term ‘world’ – are thought to encompass totalities of appearances. Since a totality of appearances cannot in fact be given in possible experience, however, they surpass its limits. Therefore – in agreement with the second meaning of the term ‘world’ – through these ideas we do not really think objects of possible experience. What needs to be clarified now is, first, why Kant thinks that the ‘supreme principle’ of reason is invalid when the ‘conditioned’ that is assumed at the start is an appearance and, second, how Kant’s argument concerning the unfitness of the cosmological ideas for the direct cognition of objects is best reconstructed. Let us start with the first issue. Kant submits that the ‘supreme principle’ would be valid if the given conditioned

h

d i

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

190

The Method of the Critique of Pure Reason

were a thing in itself.11 He further submits that when a conditioned is given as an appearance, it is legitimate to consider the totality of conditions ‘given as a problem’ (aufgegeben) (A498/B526). What he means by this is that, given a conditioned appearance, it is legitimate to look for its conditions and to ‘ascend’ in the series of these conditions. As we saw in Chapter 4, proceeding in this way is a condition for perfecting the cognitions we obtain through the understanding. What it is illegitimate to do is to take the ‘totality of conditions’ as actually given. Because the given ‘conditioned’ is not a thing in itself but an object as it is represented by us in agreement with the conditions of our sensibility, we must take the idea of the totality of conditions not ontologically, as referring to the conditions for the existence of the object considered in itself, but heuristically, as a ‘regulative’ maxim for perfecting the cognitions obtained through the understanding (A508–15/B536–43).12 It is not clear whether Kant can actually support the claim that, in the case of appearances, the assumption that a complete series of conditions must actually be given when a conditioned is given is illegitimate (for recent discussions, see Willaschek 2018: 155; Walden 2019). But let us grant that Kant’s claim is justified. How should we reconstruct his argument concerning the unfitness of the cosmological ideas to provide direct cognitions of objects? The argument builds on two theses. On the one hand, we have the thesis, defended in the Analytic, that the categories can only legitimately be used for cognition within the boundaries of possible experience. On the other hand, we have the thesis that when a conditioned appearance is given, the assumption that the totality of its conditions is also given constitutes a transgression of the limits of possible experience. It is only when this second thesis is in place that the thesis concerning the limitation of the validity of the categories to possible experience can do the work in the Antinomy. Only at this point does it become clear that the use of the categories to determine totalities of appearances in the cosmological ideas implies a violation of their legitimate use. Indeed, the Dialectic does provide an important tool for determining the limits of our cognition. At least in the case of the cosmological ideas, its role is not simply to diagnose the natural disposition of our reason to use the categories in an illegitimate way. In determining that the ‘supreme principle’ is invalid for 11 12

h

d i

0 0

See Willaschek (2018: 152–6) for an explanation of why the supreme principle holds for things in themselves. Even if Kant speaks of the ‘regulative’ use of the ideas in the Antinomy and in other parts of the Dialectic, it is only in the Appendix that this use receives a proper justification.

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Negative Side of the Critique of Pure Reason

191

appearances, the Dialectic offers a tool without which it would be impossible to stress that the cosmological ideas arise from an application of the categories that cannot yield cognition. c. Transcendental philosophy and the critique of pure reason in the Dialectic. It is now time to draw conclusions about the relationship between the critique of pure reason and transcendental philosophy when it comes to establishing Kant’s claim that the ideas, as root concepts, are unfit to provide cognition of the objects they represent. How does this claim depend on doctrines that belong to transcendental philosophy? To answer this question, I will again focus on the cosmological ideas. Recall the two theses on which the claim about the unfitness of the cosmological ideas for cognition rests. The first thesis submits that the categories can only legitimately be used for cognition within the boundaries of possible experience. Arguably, this thesis is already established in the Analytic. Moreover, because it concerns the limits of the valid use of these concepts, it belongs not to transcendental philosophy but to the negative side of the critique of pure reason. Insofar as the Dialectic relies on this thesis, it simply uses a negative point established by the critique of pure reason in the Analytic in another context. If we then ask how this thesis depends on transcendental philosophy, the answer will simply be identical to the answer regarding how the Analytic sets limits to the valid use of the categories. Since I discuss this problem in Chapter 6, I will not consider it further here. The second thesis states that when a conditioned appearance is given, the assumption that the totality of its conditions is also given is invalid. On the most basic level, since the thesis concerns the validity of the ‘supreme principle’ of reason, it rests on the identification of this principle, which plays an essential role in the metaphysical deduction of the transcendental ideas. However, the metaphysical deduction does not offer any tools for demonstrating that the supreme principle is invalid. What is decisive in establishing the lack of validity of the principle is a characterization of the ‘conditioned’ to which the ‘supreme principle’ is applied. Recall that Kant contends that it is for appearances that the supreme principle is invalid, while it would be valid if the ‘conditioned’ were a thing in itself. Because the determination of what an appearance is is obtained in the transcendental expositions of the concepts of space and time,13 which I read as belonging to transcendental philosophy, Kant’s negative point regarding the invalidity of the ‘supreme principle’ 13

h

d i

See Chapter 3.

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

192

The Method of the Critique of Pure Reason

fundamentally rests on parts of transcendental philosophy established in the Aesthetic. Kant’s idea is that by attending to some features that characterize objects considered as appearances, we can see that the supreme principle cannot be applied to them.

4  The Negative Side of the Critique of Pure Reason and Transcendental Philosophy This chapter had two main aims. The first was to investigate how the critique of pure reason depends on transcendental philosophy in its arguments regarding the limits of the validity of the root concepts for the cognition of objects. In this respect, I used Kant’s remarks regarding the ‘unavoidable necessity’ of the transcendental deduction of the concept of space as providing a clue as to how his argument concerning the limits of the validity of this concept relies on claims established by transcendental philosophy. I suggested that Kant’s argument rests on claims he makes in both the metaphysical exposition and the transcendental exposition of the concept of space. I then showed that the way in which Kant establishes the invalidity of the concept of space for things in themselves cannot offer a general model for determining how the negative side of the critique of pure reason depends on transcendental philosophy. This is evident when one considers transcendental ideas. In this case, Kant’s argument regarding the unfitness of ideas to provide direct cognition of objects cannot rest on their transcendental deduction, since that deduction comes after Kant’s argument that the ideas are unable to deliver cognitions. But this means that Kant’s strategy in the negative part of the critique of pure reason is as pluralistic as the strategy that he follows in the metaphysical and transcendental deductions. The second aim of the chapter was to determine where Kant establishes that there are limits to our cognition. I claimed that each main part of the Transcendental Doctrine of Elements contains arguments that establish that the root concepts analysed there have limited validity. I took Kant’s argument that things in themselves do not have spatial properties to be paradigmatic of how those limits are established in the Aesthetic. Since I will discuss the way in which Kant sets limits to the use of the categories in Chapter 6, and since this is not a controversial claim, I have here simply assumed that the Analytic does set limits to the use of the categories. The rest of the chapter was dedicated to the limits established in the Dialectic. Kant describes ideas of reason as ‘categories extended to the unconditioned’ (A409/B436). Given this description, one might suppose

h

d i

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

The Negative Side of the Critique of Pure Reason

193

that the limits to the valid use of the categories established in the Analytic are sufficient for proving that the ideas cannot be cognitions of objects. I have shown that this impression is wrong, at least when one considers the cosmological ideas. Since these ideas are thought to describe ‘totalities of appearances’, they do not prima facie violate the limits of the use of the categories established in the Analytic. It is only after the Dialectic shows that the use of the ‘supreme principle’ of reason for determining totalities of appearances is illegitimate and involves trespassing the limits of possible experience that the cosmological ideas can be proved invalid for the direct cognition of objects. In this sense, the Dialectic also makes an essential point for establishing the unfitness of the ideas to deliver cognition. In conclusion, let me recall a claim I made at the end of Chapter 2. The chapter argued that the critique of pure reason is the doctrine of method of metaphysics. In that context, however, I suggested that the fact that the critique of pure reason can be so described does not mean that this discipline is confined to the Transcendental Doctrine of Method. Rather, both the present chapter and the next single out parts belonging to the ‘negative side’ of the critique of pure reason that are defended within the Transcendental Doctrine of Elements. Chapter 2 also provided an account of the Discipline of Pure Reason, according to which one of its functions is to identify cognition- or object-dependent ‘negative’ rules for enabling metaphysics to become a science. I suggest that the arguments establishing the limits of the valid use of root concepts that I analyse in this and the following chapter form the background for those ‘negative’ rules.

h

d i

0 0

9 8 009

2 2 009

bli h d

li

b C

b id

i

i

chapter 6

Transcendental Philosophy and the Critique of Pure Reason in the B-Deduction

In Chapter 4, I provided a reconstruction of Kant’s transcendental deduction of the categories in the B-edition of the Critique of Pure Reason. I suggested that in the sections where Kant provides the deduction, it is possible to disentangle a positive argument that proves that the categories have objective validity and ‘can relate to objects a priori ’ (A85/B117). The argument in question belongs to transcendental philosophy and is not charged with establishing that the categories have limited validity. I conceded, however, that Kant does claim in the same sections that the categories are only valid for appearances, where this claim belongs to the negative side of the critique of pure reason, on my account. At this point, the question arises concerning how Kant supports that negative claim. On what basis can he submit that the categories can only be legitimately applied to appearances? Does he argue for this claim in those sections of the B-deduction, or is he repeating a point made elsewhere? Moreover, how is the negative claim related to the positive argument reconstructed in Chapter 4? Does the negative claim somehow rest on that argument, or is it established independently of it? In this chapter, my aim will be to show, first, that Kant does argue for the negative claim in the sections dedicated to the B-deduction and, second, that the latter argument depends on the positive argument reconstructed in Chapter 4. More precisely, the ‘negative’ argument partly depends on the result of the first step of the ‘positive’ argument of the B-deduction. This reading is important because it sheds new light on the structure of the sections dedicated to the B-deduction. As we saw in Chapter 4, there is vast agreement among Kant scholars that the B-deduction is divided into two steps. In my reconstruction, both steps are part of the positive argument belonging to transcendental philosophy. What is often overlooked,1 however, 1

Often, but not always: I take this approach to give credence to those readings of the transcendental deduction that argue that its main purpose is to set limits to the valid use of the categories in order

194

h

d i

0 0

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

Transcendental Philosophy & the Critique in the B-Deduction

195

is that within these sections it is also possible to disentangle a negative argument that belongs to the critique of pure reason and holds that the categories are only valid for appearances.2 One might worry that differentiating between a positive and a negative argument within the sections of the B-deduction is arbitrary and only designed to confirm my thesis regarding the division of labour between transcendental philosophy and the critique of pure reason. In order to answer this worry, in Section 1 I will show that it is Kant himself who insists on the importance of this distinction, especially in two passages that illuminate his thinking during the preparation of the B-deduction. While these passages are helpful for illustrating the importance of distinguishing between a ‘positive’ and a negative ‘argument’ concerning the validity of the categories, they are problematic in another respect since they provide two conflicting accounts. Therefore, in order to properly understand how the positive argument and the negative argument are related to one another, in Section 2 I will reconstruct Kant’s negative argument in the B-deduction and determine whether and how it rests on the positive argument.

1  The ‘Positive’ and the ‘Negative’ Argument Regarding the Validity of the Categories In this section, my purpose is to substantiate the claim that the B-deduction contains both a positive and a negative argument. The claim is supported by two passages, contained in the Preface to the Metaphysical Foundations of Natural Science and On the Use of Teleological Principles in Philosophy, respectively. In these passages, Kant clearly distinguishes between two arguments that concern the validity of the categories. The aim of the first argument is to prove that the categories are valid for appearances. The aim of the second argument is to show that the categories have no other

2

h

d i

to prepare Kant’s resolution of the dialectical conflicts of reason in the Dialectic (see Engstrom 1994 and Hatfield 2003; see Gava 2019c for an attempt to develop a model of transcendental argument that builds on the basic intuition of these readings). The advantage of my approach is that we do not need to downplay the positive achievements of the transcendental deduction. By distinguishing between a positive argument belonging to transcendental philosophy and a negative argument belonging to the critique of pure reason, it is possible to emphasize the importance of the latter in the economy of the Critique of Pure Reason without denying that in the B-deduction there is also an argument the chief aim of which is to prove that the categories have validity. Kostantin Pollok (2008) is an exception. He draws on the same texts that I discuss in Section 1, where Kant distinguishes between the ‘negative’ and the ‘positive’ tasks of the Critique of Pure Reason. His aim is not to insist on the importance of this distinction, however. Rather, he uses these texts to support his reading of the second step of the B-deduction. He sees this step as introducing an important novelty in comparison to the A-deduction.

0 0

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

196

The Method of the Critique of Pure Reason

possible valid use for cognition. Although these texts agree on drawing this distinction, they disagree about how to depict the relationship between the two arguments. The picture is further complicated if we take into consideration the passage in the first Critique where Kant speaks of the ‘unavoidable necessity’ of the transcendental deduction of the categories. While the passage does not explicitly distinguish between a positive and a negative argument regarding the validity of the categories, it can be read as maintaining this distinction while hinting at yet another way of understanding the relationship between these arguments. I will now analyse these passages in turn, following the order in which they were published. a. The ‘unavoidable necessity’ of the transcendental deduction of the categories. In Chapter 5, I discussed the passage in which Kant explains why a transcendental deduction of the categories and the concept of space was ‘unavoidably necessary’. Recall that what makes the transcendental deduction of these concepts ‘unavoidably necessary’ is that we sometimes use them in problematic ways and therefore need to set limits to this use (see A88/B120–1). Why is this passage important for our purposes? It is important because it clearly differentiates between two roles played by the transcendental deduction. Its ‘positive’ role is the one it would still have even if the concepts that are deduced were not sometimes used problematically. In these cases, the only purpose of a transcendental deduction would be to prove how ‘concepts can relate to objects a priori’ (A85/B117). It would not need to identify limits for the use of such concepts. This need, which characterizes the negative role of the deduction, first kicks in when we are faced with problematic uses of the concepts. The passage does not explicitly distinguish between a positive and a negative argument about the validity of the categories. It hints at a specific way of depicting their relationship, however, namely one in which the ‘negative’ argument depends on the ‘positive’ argument. This is clear because positively determining how the categories can relate a priori to objects is said to be essential to putting a stop to their problematic use. b. The limits of the validity of the categories and the ‘beginning’ of the transcendental deduction. In a footnote to the Preface of the Metaphysical Foundations of Natural Science, Kant explicitly distinguishes between two arguments within the transcendental deduction, where it is only one of these that plays a role in determining the limits of the categories. The footnote provides an answer to a criticism of the first edition of the

h

d i

0 0

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

Transcendental Philosophy & the Critique in the B-Deduction

197

Critique of Pure Reason advanced by Johann Schultz in his review of Johann August Heinrich Ulrich’s Institutiones Logicae et Metaphysicae from 1785. In brief, Schultz complains that, given the obscurity of the transcendental deduction, Kant was unable to provide an adequate foundation for his claim that our capacity to obtain cognition through reason has specific limits (Schultz 2000 [1785]). Kant’s answer to Schultz proceeds as follows: I assert, on the contrary, that the system of the Critique must carry apodictic certainty for whoever subscribes (as the reviewer does) to my propositions concerning the sensible character of all our intuition, and the adequacy of the table of categories, as determinations of our consciousness derived from the logical functions in judgements in general, because it is erected upon the proposition that the entire speculative use of our reason never reaches further than to objects of possible experience. For if we can prove that the categories which reason must use in all its cognition can have no other use at all, except solely in relation to objects of possible experience (insofar as they simply make possible the form of thought in such experience), then, although the answer to the question how the categories make such experience possible is important enough for completing the deduction where possible, with respect to the principal end of the system, namely, the determination of the limits of pure reason, it is in no way compulsory, but merely meritorious. For the deduction is already carried far enough for this purpose if it shows that categories of thought are nothing but mere forms of judgements insofar as they are applied to intuitions (which for us are always sensible), and that they thereby first of all obtain objects and become cognitions; because this already suffices to ground with complete certainty the entire system of the Critique properly speaking. (4:474n)

The passage is interesting for various reasons. First, it identifies the main role of the Critique of Pure Reason as making a ‘negative’ point, namely, the role of proving that pure speculative reason cannot cognize objects beyond possible experience. Second, it directly connects the discussion of the ‘negative’ and ‘positive’ roles of the Critique to the issue of the validity of the categories. Third, Kant clearly distinguishes the question whether the categories are valid from the question how they can be valid. Fourth, he maintains that to establish that the categories are valid it is not necessary to bring the transcendental deduction to completion, which suggests that the beginning of the transcendental deduction is nonetheless essential to this purpose. Fifth, he submits that establishing that the categories have validity is sufficient for determining their limits. In turn, this means that it is only the beginning of the transcendental deduction that is needed to establish these limits.

h

d i

0 0

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

198

The Method of the Critique of Pure Reason

The passage leaves undetermined whether the beginning of the deduction directly establishes this ‘negative’ point or whether it is only instrumental for devising an argument to that effect. This issue is clarified further below, where Kant reconstructs how one can arrive at the negative claim concerning the validity of the categories (4:475n). He describes the claim as following from three premises. The first states that the table of the categories contains all pure concepts of the understanding. The second asserts that the understanding contains synthetic a priori principles, in which all objects that may be given are brought under these categories. Moreover, since objects can only be given through intuition, the categories need intuition to fulfil their function. The third submits that through intuition we only cognize objects as appearances. From these premises, Kant writes, it follows ‘that all use of pure reason can never extend to anything other than objects of experience, and, since nothing empirical can be the condition of a priori principles, the latter can be nothing more than principles of the possibility of experience in general’ (4:475n). Because the first and third premises clearly reflect the results of the metaphysical deduction of the categories and the Transcendental Aesthetic, respectively, it is the second premise that should be obtained at the ‘beginning’ of the transcendental deduction. Kant thinks that this premise is insufficient to establish the full negative point concerning the limited validity of the categories; otherwise, he would not have indicated two additional premises to sustain that point. True, the function of the first premise is only to warrant that we have the full list of the categories, but the third premise introduces a claim that is essential to positively determining the limits of the categories, namely, that through intuitions we only cognize appearances. To summarize, the passage distinguishes between the positive argument of the transcendental deduction and a negative argument. The former establishes that and how the categories are valid, while the latter establishes that they are only valid for appearances. Kant depicts the negative argument as depending on the beginning of the positive argument of the transcendental deduction, which only proves that the categories are valid. c. The ‘positive’ and the ‘negative’ intentions of the Critique. Another passage in which Kant distinguishes between a positive and a negative argument regarding the validity of the categories is to be found in On the Use of Teleological Principles in Philosophy. At the very end of the text, Kant

h

d i

0 0

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

Transcendental Philosophy & the Critique in the B-Deduction

199

responds to two charges of contradiction that were raised against the Critique of Pure Reason. I will here focus only on Kant’s answer to the second of these charges. Kant does not explicitly say who was responsible for it, but it is possible to trace it back to Karl Leonhard Reinhold, who, in a letter to Kant dated 12 October 1787, wrote the following: ‘In the Note beneath the text of the Preface to the Metaphysical Foundations of Natural Science you write very pointedly that the main foundation of your system is secure “even without a complete deduction of the categories” – on the other hand in both the first and second editions of the Critique of Pure Reason in Chapter II of the Transcendental Analytic, Section I, “the unavoidable necessity” of that deduction is asserted and demonstrated’ (10:500, translation altered). Reinhold points out an apparent contradiction between the passages discussed in Sections 1.a and 1.b. He correctly frames the contradiction as one between the claim that the transcendental deduction is ‘unavoidably necessary’ for the foundation of Kant’s system and the claim that its completion is superfluous for that foundation. Given this way of framing the contradiction, it seems that a fairly easy way out was available to Kant. He could have simply submitted that it is only part of the transcendental deduction, namely what he calls its ‘beginning’ in the Metaphysical Foundations, that is ‘unavoidably necessary’ for establishing the limits of the valid use of the categories. Kant did not make this point, however. Rather, he provided yet another account of the relationship between the ‘positive’ and the ‘negative’ argument regarding the validity of the categories: One can easily see that in the former work [the Metaphysical Foundations of Natural Science] the deduction was considered only with a negative intention (Absicht), namely in order to prove that with the categories alone (without sensible intuition) no cognition of things could come about – which becomes clear already if one turns only to the exposition of the categories (as logical functions applied merely to objects in general). Yet since we also engage in a use of the categories in which they actually pertain to the cognition of objects (of experience), the possibility of an objective validity of such concepts a priori in relation to the empirical had to be proven separately, so that they would not be judged to be without meaning or to have originated empirically. And that was the positive intention with respect to which the deduction is indeed indispensably necessary. (8:184)

Kant distinguishes between the negative and the positive ‘intentions’ of the Critique of Pure Reason. He characterizes the former as an attempt to show that the categories are insufficient for providing cognition when

h

d i

0 0

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

200

The Method of the Critique of Pure Reason

they are not connected to a possible sensible intuition. He also submits that the ‘exposition’ of the categories is sufficient to establish this point, and he seems to have the metaphysical deduction of the categories in mind (for a similar view, see Pollok 2008: 340). By contrast, the positive ‘intention’ consists in showing that the categories can indeed provide valid cognition and are not derived from experience. It is in relation to this positive ‘intention’ that the transcendental deduction is ‘unavoidably necessary’. The most apparent difference between this response and the passage from the Metaphysical Foundations is that Kant does not identify a single main purpose of the Critique of Pure Reason. He instead submits that the Critique has both a negative and a positive aim and that it is with regard to the latter that the whole transcendental deduction is indeed necessary. What is striking is that Kant provides a completely different account of how the negative aim is achieved. More precisely, he claims that the metaphysical deduction of the categories provides everything that is needed to prove that the categories cannot provide cognition in the absence of a possible sensible intuition. Therefore, the transcendental deduction is not at all required for setting limits to the use of the categories. d. Overview. How are the positive and the negative arguments concerning the validity of the categories related to one another? The Critique, the Metaphysical Foundations, and On the Use of Teleological Principles in Philosophy present three different and conflicting answers to this question. The passage from the Critique hints at the dependence of the negative argument on the whole positive argument. The passage from the Metaphysical Foundations presents the negative argument as only relying on the beginning of the positive argument.3 Finally, the passage from On the Use of Teleological Principles in Philosophy characterizes the negative argument as independent of the positive one. Given these conflicting characterizations, Kant is certainly not helpful when it comes to understanding how the relationship between the 3

h

d i

0 0

In relation to the latter passage, one might plausibly ask what Kant could have had in mind when he distinguished between the beginning and the completion of the transcendental deduction. Given that the Metaphysical Foundations appeared in 1786, only one year before the publication of the B-edition of the Critique, it is possible that Kant was thinking of the two-step structure of the B-deduction. This would fit his claim that it is only the conclusion of the transcendental deduction that answers the question how the categories can be valid. Arguably, Kant’s account of synthesis speciosa in § 24 of the transcendental deduction provides the main insight of Kant’s answer to that question.

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

Transcendental Philosophy & the Critique in the B-Deduction

201

‘positive’ and the ‘negative’ argument concerning the validity of the categories should be reconstructed. In the next section, I will argue that, when we focus on the B-deduction, Kant’s account in the Metaphysical Foundations is to be preferred. The sections dedicated to the B-deduction contain a ‘negative’ argument concerning the validity of the categories. This negative argument rests on the first step of the ‘positive’ argument of the B-deduction.

2  The ‘Negative’ Argument in the B-deduction When one surveys the sections composing the B-deduction of the categories, there is prima facie evidence that the negative argument concerning the validity of the categories rests on part of the transcendental deduction.4 This evidence is provided by § 22 and § 23 of the B-deduction, where Kant clearly speaks of the limits of the valid use of the categories. Since these sections come immediately after the first step of the B-deduction, it is plausible to assume that they at least partially rest on that step, while they do not presuppose the second step. In this section, I will first analyse § 22 and § 23 and provide a reconstruction of Kant’s negative argument. I will then clarify the sense in which this argument depends on the first section of the B-deduction. a. The negative argument in the B-deduction. That § 22 and § 23 of the B-deduction are charged with setting limits to the valid use of the categories is evident. Accordingly, the title of § 22 reads: ‘The category has no other use for the cognition of things than its application to objects of experience’ (B146). Moreover, Kant closes the paragraph with the following sentence: ‘Consequently, the categories have no other use for the cognition of things except insofar as these are taken as objects of possible experience’ (B147–8, translation altered), which clearly shows that establishing that the categories have a limited valid use for cognition is meant to be the conclusion of the paragraph. Now turning to § 23, it starts by specifying why the conclusion of § 22 is important: ‘The above 4

h

d i

Of course, when Kant introduced the first interpretative option through his remarks on the ‘unavoidable necessity’ of the transcendental deduction, he could not have meant it to be a description of the structure of the B-deduction, since the passage in question was already contained in the A-edition of the Critique. This does not pose a problem for my reading, however, since I am not claiming that that interpretative option is correct, as far as the structure of the B-deduction is concerned. By contrast, as I have already suggested, it is possible to assume that in 1786, when the Metaphysical Foundations of Natural Science was published, Kant was already thinking of the structure of the B-deduction.

0 0

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

202

proposition is of the greatest importance, for it determines the boundaries of the use of the pure concepts of the understanding in regard to objects, just as the Transcendental Aesthetic determined the boundaries of the use of the pure form of our sensible intuition’ (B148). The paragraph draws the consequences of § 22 regarding the use of the categories in relation to sensible intuitions that are different from ours and non-sensible intuitions. In both cases, we cannot use the categories to characterize objects, because it is only in relation to our sensible intuition that the categories contribute to cognition. Therefore, the purpose of § 22 and § 23 is indeed to set limits to the valid use of the categories. Moreover, it seems that the main argument to that effect is already attained in § 22, whereas § 23 draws further consequences. How does the argument of § 22 work? Its core is contained in the first sentence of the paragraph: To think of an object and to cognize an object are thus not the same. For two components belong to cognition: first, the concept, through which an object is thought at all (the category), and second, the intuition, through which it is given; for if an intuition corresponding to the concept could not be given at all, then it would be a thought as far as its form is concerned, but without any object, and by its means no cognition of anything at all would be possible, since, as far as I would know, nothing would be given nor could be given to which my thought could be applied. Now all intuition that is possible for us is sensible (Aesthetic), thus for us thinking of an object in general through a pure concept of the understanding can become cognition only insofar as this concept is related to objects of the senses. (B146)

Let me first note that when Kant speaks of sensible intuitions here, he is considering not sensible intuition in general 5 but rather sensible intuitions that are given through our particular form of sensibility, whose a priori forms are space and time. This becomes apparent later in the text, where Kant explicitly discusses space and time as our forms of intuition. With this point in mind, the argument appears to proceed as follows: (1) In order to cognize an object, we need both concepts and intuitions. (2) Categories are concepts. (3) Categories need to be related to intuitions in order to be used for the cognition of objects (from 1 and 2). (4) Our particular forms of intuition are space and time. 5

h

d i

0 0

Recall that the first step of the B-deduction abstracts from our particular forms of sensibility. Its argument is valid for any understanding that, like ours, depends on sensibility for receiving objects but could have a different form of sensibility (B145–6; see also B138–9; B150).

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

Transcendental Philosophy & the Critique in the B-Deduction

203

(5) All our intuitions necessarily agree with the forms of space and time. ( 6) For us, the categories can only provide cognition in relation to (possible) intuitions that agree with the forms of space and time (from 3, 4 and 5). What is striking about this argument is that it is not clear why it is located at the end of the first step of the B-deduction. More precisely, it does not seem to rest on any point made in that step. It is based, first, on the claim that we need both concepts and intuitions to obtain cognition and, second, on the claim that all our intuitions are necessarily in space and time. Kant makes the first claim at the beginning of the Transcendental Analytic, where he famously submits that ‘[t]houghts without content are empty, intuitions without concepts are blind’ (A51/B75). In relation to this claim, what the argument does is simply to draw its consequences for a particular set of concepts, that is, the categories. By contrast, as Kant signals in the quoted passage, the second claim is already established in the Aesthetic. b. Why does the ‘negative’ argument need the first step? Kant points out at various points that the first step of the B-deduction assumes that an understanding like ours essentially depends on sensible intuitions for receiving objects (B145–6; B150; see also B138–9). Of course, this claim is immediately related to the contention that we need both concepts and intuitions for cognition, or, what amounts to the same thing, that concepts cannot refer to objects without intuitions. It is not clear, however, whether Kant also assumes that the categories need intuitions in order to refer to objects. Rather, he takes the categories as ostensibly constituting an anomaly because they do seem to be able to refer to objects without the need for a corresponding intuition. Take again the passage where Kant considers the ‘unavoidable necessity’ of the transcendental deduction of the categories. The latter arises because the categories ‘speak of objects not through predicates of intuition and sensibility but through those of pure a priori thinking’, and, consequently, ‘they relate to objects generally without any conditions of sensibility’ (A88/B120).6 6

h

d i

One might ask why the transcendental deduction begins by assuming that the categories are ‘concepts of objects in general’ (A93/B126), given that the metaphysical deduction already characterizes them as the fundamental ways in which we order the manifold of intuition in a single intuition. I believe the two descriptions are compatible. The former focuses on the representational content of the categories, while the latter focuses on an activity that we perform through them. Still, one might object that the characterization of the categories in the metaphysical deduction already suggests

0 0

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

204

The Method of the Critique of Pure Reason

Since the categories do not seem to be related to corresponding intuitions through which they can refer to objects, they appear to violate one general requirement for attaining cognition of objects, which might be used to question their validity. But the same consideration might provide grounds for waiving the general requirement as far as the categories are concerned. Therefore, the following apparent disjunction arises: either the validity of the categories extends over and above the limits that customarily apply to concepts or they have no validity in the first place.7 This is a direct consequence of the categories’ being a priori concepts. While it is clear that empirical concepts need (empirical) intuitions to refer to their objects (it is only through empirical intuitions that the objects of an empirical concept are first given to us), if the categories were valid, they would of course be able to refer to objects without depending on any particular empirical intuition. But this generates the (false) impression that they can refer to objects in the absence of any relation with intuition whatsoever (or, using Kant’s words, ‘they relate to objects generally without any conditions of sensibility’). Now if one, starting with this disjunction, were to appeal to Kant’s general principle that ‘for cognitions of objects we need both concepts and intuitions’ and apply it to the categories, one would have to draw the unwelcome conclusion that the categories do not have any validity at all.

7

h

d i

0 0

that the categories must be used in connection to possible intuitions, in which case we might ask why Kant starts by contemplating the possibility of the categories’ referring to objects without intuitions. One reason for Kant’s assumption may lie in the idea that this approach represents one of the two dominant ways in which the categories were used by metaphysicians before him. Famously, Kant characterizes the history of metaphysics as being dominated by two fundamental approaches. The first is dogmatism, which assumes synthetic a priori principles without ‘critique’ and uses them without any determination of their limits. The second is scepticism, which expresses general doubt about our capacity for synthetic a priori cognition. Given that, as root concepts, the categories lie at the basis of synthetic a priori principles, we can take dogmatism and scepticism as making unlimited use of the categories and doubting that they have any validity, respectively. But this aligns well with the disjunction with which Kant begins the transcendental deduction, which submits that the categories either are able to refer to objects without any condition arising from intuition or have no validity at all. That Kant thinks the categories present this apparent disjunction is clear from how he describes the category of causality in the context of his discussion of the ‘unavoidable necessity’ of the transcendental deduction of the categories. He writes: ‘It is not clear a priori why appearances should contain anything of this sort [the concept of a cause] (one cannot adduce experiences for the proof, for the objective validity of this a priori concept must be able to be demonstrated), and it is therefore a priori doubtful whether such a concept is not perhaps entirely empty and finds no object anywhere among the appearances’ (A90/B122). We can read the passage as follows: because the concept of cause is a priori, it requires a proof that does not rest on particular experiences. These would be cases in which the concept is connected to empirical intuitions. But since we cannot appeal to these empirical intuitions to show that the concept refers to objects, it seems that the concept should be able to refer to objects without any relationship to intuition, which is cause to doubt its validity.

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

Transcendental Philosophy & the Critique in the B-Deduction

205

We are now in a position to explain why Kant places the ‘negative’ argument of § 22 at the end of the first step of the B-deduction. In view of the characterization of the categories that Kant assumes at the beginning of the transcendental deduction, the first step of the B-deduction shows that the categories do refer (a priori) to objects given in (sensible) intuition. They do refer (a priori) to objects given in (sensible) intuition because without the categories we would not be able to have any unitary representation of a manifold of intuition. In this way, the argument of the first step overcomes the illusory disjunction according to which either the categories are able to refer to objects without intuitions or they have no validity at all. The first step shows that the categories are indeed objectively valid and that they can be so without violating Kant’s general requirement according to which cognition requires both concepts and intuitions. The categories do not violate that requirement because they refer a priori to objects only by determining conditions for having unitary intuitions of objects. In this way, they are necessarily related to (possible) intuitions. Therefore, by appealing to the same general requirement after the first step of the B-deduction, the negative argument in § 22 can argue that the categories cannot be used in the absence of possible intuitions without having to conclude that they have no validity at all. Notice, however, that if this explains why Kant located the ‘negative’ argument in § 22, it also shows that § 22 does not ‘depend’ on the first step of the B-deduction in a strong sense of the term. Kant could equally first have used his general requirement for cognition to establish that the categories cannot be used for cognition in the absence of any relation to intuition. At that point, he could have established, in a move that is similar to what he does in the first step of the B-deduction, that the categories do indeed have validity in relation to (possible) sensible intuitions. This means that § 22 ‘depends’ on the first step of the B-deduction only from the perspective of the particular argumentative strategy that Kant pursues to overcome the illusory disjunction with which he starts. It does not depend on the first step of the B-deduction in the sense that it could not have established the same point by taking another route.

3  How the ‘Negative’ Argument Rests on Results of Transcendental Philosophy What I hope to have shown in this chapter is that Kant clearly distinguishes between a ‘positive’ and a ‘negative’ argument regarding the validity of the categories, where the former establishes that and how the categories

h

d i

0 0

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

206

The Method of the Critique of Pure Reason

are valid and the latter sets limits to their valid use. In my account, the ‘positive’ argument belongs to transcendental philosophy and the ‘negative’ argument belongs to the critique of pure reason. Notice that even though Kant provides contrasting accounts of how the two arguments are related to one another, all of these accounts align with my claim that the ‘negative’ argument rests on results belonging to transcendental philosophy. As we saw in Section 1, Kant describes the negative argument as either resting on the whole positive argument, resting on part of it, or being totally independent of it. Since in my account the positive argument belongs to transcendental philosophy, it is clear that the first and second options are consistent with my general approach. The third option fundamentally agrees with it too, however. It maintains that the ‘negative’ argument rests on the results of the metaphysical deduction of the categories. Because the metaphysical deduction belongs to transcendental philosophy in my account, the third option is also consistent with my claim that the ‘negative’ argument regarding the validity of the categories rests on claims established within transcendental philosophy. Even though all three interpretative options offered by Kant are consistent with my general approach, I have not remained neutral with respect to them. Rather, I have suggested that, at least when one considers the B-deduction, the second option seems to be correct. Finally, let me recall my reconstruction of the ‘negative’ argument for the concept of space analysed in Chapter 5. I argued that the negative argument for space rested on both the metaphysical and the transcendental deductions of that concept. Kant’s description of the negative argument for the categories in the Metaphysical Foundations resembles this negative argument in some respects, since it also rests on both the metaphysical and the transcendental deduction of the concepts that are at issue. If, as I suggested, it is Kant’s account in the Metaphysical Foundations that more accurately describes what Kant does in the B-deduction, this would mean that Kant’s negative argument for the categories is somehow in continuity with his negative argument for space.

h

d i

0 0

9 8 009

2 2 0 0

bli h d

li

b C

b id

i

i

chapter 7

The Positive Side of the Critique of Pure Reason

In the Preface to the B-edition of the Critique, Kant famously submits that he had ‘to deny knowledge [Wissen] in order to make room for belief [Glaube]’ (Bxxx, translation altered).1 This passage points to the Canon of Pure Reason, where Kant argues that we can form ‘beliefs’ – a technical term that refers to rational positive attitudes towards propositions not based on evidence – regarding the existence of God and the immortality of the soul. Kant’s argument for sustaining these beliefs, both in the Critique of Pure Reason and in other works, has been the object of meticulous analysis by Kant scholars, yet an adequate account of the systematic place of this argument within the Critique of Pure Reason has yet to be given. In this chapter, I will argue that reading the Critique as the doctrine of method of metaphysics puts us in a position to provide such an account. Specifically, this account can be achieved when we take the perspective of the ‘positive’ side of the critique of pure reason in its attempt to establish that metaphysics can attain architectonic unity. Of course, many would concede that Kant took the first Critique to be relevant to the ‘practical’ part of metaphysics, but the question of how to understand this relevance remains open. The most common way to construe it is as follows: the critique of pure reason is an analysis of our faculty of cognition that determines what we can and cannot cognize a priori. In turn, this analysis is relevant to the practical part of metaphysics because it shows that we must necessarily remain agnostic regarding objects of pure reason that are important for that part, which means that such objects cannot be theoretically proved impossible. This approach is backed by the B-Preface, where Kant, in speaking of freedom, God and immortality, oscillates between suggesting that the Critique shows that these objects are logically possible and claiming that it shows that their existence is theoretically 1

I here translate Glaube as belief and not faith, which is the most common translation, to better emphasize the connection with Kant’s discussion of moral belief in the Canon.

207

h

d i

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

208

undecidable.2 The problem with this account is that it does not explain why the Canon contains an argument to the effect that we can commit ourselves not only to the possibility but also to the actuality of God and immortality. Therefore, it seems that the story regarding the relevance of the Critique to the practical part of metaphysics needs to be developed further.3 In this chapter, I intend to show that reading the critique of pure reason as the doctrine of method of metaphysics provides a straightforward explanation of the systematic place of Kant’s practical argument in the Canon and the related characterization of belief. Once one characterizes the critique of pure reason as the doctrine of method of the whole of metaphysics, the critique cannot be content with proving that God and immortality are possible or theoretically undecidable. It cannot be content with this because, given Kant’s general moderate methodological conservatism, Kant assumes that moral principles that in his mind will form part of the practical part of metaphysics have at least some validity. Since these principles, according to Kant, require a commitment to the existence of God and immortality,4 the critique must show that, within metaphysics as a whole, there is space to accommodate that commitment without endangering the results of the ‘negative’ side of the critique. I will start in Section 1 by looking at the passage of the B-Preface where Kant presents the ‘positive utility’ of the critique. Kant links this positive utility to either the ‘logical possibility’ or the ‘theoretical undecidability’ of objects of pure reason. I will clarify this distinction while maintaining that it is unable to account for the inclusion of the practical argument in the Canon and the related characterization of belief. In Section 2, I will present Kant’s argument that we should commit ourselves to the actual existence of God and immortality and consider whether the commitment to actuality that it requires is justified. Section 3 will address Kant’s account of ‘taking-to-be-true’ (Fürwahrhalten) in the Canon, giving particular attention to his characterization of belief (Glaube). Finally, in Section 4 I will show how my account of the ‘positive’ side of the critique of pure reason provides a straightforward explanation for the inclusion in the Critique of 2

3

4

h

d i

0 0

In this vein, Michael Wolff (2018) claims that the first Critique is only concerned with establishing the logical possibility of a practical use of pure reason. See also G. Bird (2006b: 32), Guyer (2010: 8–9) and Frierson (2013: 22). Of course, it is no coincidence that the passage in the B-Preface appeared in 1787. One way to account for the discrepancy with the Canon is to say that the latter was already contained in the 1781 edition of the Critique. The passage in the B-Preface could simply reflect the fact that, in 1787, Kant changed his mind regarding the role of the practical argument in the Canon for the project of the first Critique. Here, I only refer to God and immortality and disregard freedom for reasons that will become apparent below.

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Positive Side of the Critique of Pure Reason

209

all the elements discussed in this chapter, that is, the passage on the ‘positive utility’ of the critique, Kant’s practical argument in the Canon and his characterization of belief. These elements can be considered part of a three-step strategy for establishing that metaphysics can achieve ‘architectonic unity’.

1  The Theoretical Undecidability of Freedom, God and Immortality Let us begin by taking a closer look at the passage on the ‘positive utility’ of the critique in the B-Preface. The main focus of Kant’s discussion is freedom. He argues that the critique establishes that the concept of freedom is not contradictory. Since we can define the logical possibility of an object in terms of a lack of contradiction in its concept, we can take Kant’s discussion as emphasizing the logical possibility of freedom. Kant writes: I can think freedom to myself, i.e., the representation of it at least contains no contradiction in itself, so long as our critical distinction prevails between the two ways of representing (sensible and intellectual), along with the limitation of the pure concepts of the understanding arising from it, and hence that of the principles flowing from them. (Bxxviii)

Freedom is spontaneous causation. The passage submits that freedom would be contradictory if our concept of cause were to apply to things in themselves. However, since our concept of cause is only valid for appearances, there is conceptual space for claiming that freedom is logically possible when it comes to things in themselves. Kant adds that a similar argument can be presented for God and immortality. However, it is not the case that the concepts of God and immortality would be self-contradictory if we did not assume transcendental idealism. Therefore, it is not surprising that Kant shifts the focus away from logical possibility: Just the same sort of exposition of the positive utility of critical principles of pure reason can be given in respect to the concepts of God and of the simple nature of our soul, which, however, I forgo for the sake of brevity. Thus I cannot even assume God, freedom and immortality for the sake of the necessary practical use of my reason unless I simultaneously deprive speculative reason of its pretension to extravagant insights; because in order to attain to such insights, speculative reason would have to help itself to principles that in fact reach only to objects of possible experience, and which, if they were to be applied to what cannot be an object of experience, then they would always actually transform it into an appearance, and thus declare all practical extension of pure reason to be impossible. (Bxxix–xxx)

h

d i

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

210

The point here is not that freedom, God and immortality are logically possible. Rather, the point is that the existence of these objects is theoretically undecidable for us. Any attempt to attain cognition of them would involve an illegitimate use of concepts and principles that are only valid within possible experience.5 Notice that saying that these objects are theoretically undecidable goes beyond claiming that they are logically possible. It means not only that their concepts are not contradictory but also that any attempt to theoretically disprove their actual existence must fail. In the Discipline of Pure Reason in its Polemical Use, in speaking of dogmatic arguments for the existence of God and immortality, Kant accordingly stresses: it is also apodictically certain that no human being will ever step forward who could assert the opposite with the least plausibility, let alone assert it dogmatically. For since he could only establish this through pure reason, he would have to undertake to prove that a highest being or the thinking subject in us as pure intelligence is impossible. But whence will he derive the cognitions [Kentnisse, translation altered] that would justify him in judging synthetically about things beyond all possible experience? (A742/B770)

Kant points out that any argument that attempts to disprove the actual existence of God and immortality must proceed through pure reason. This means that the argument must show that they necessarily do not exist, which is equivalent to showing that they are impossible.6 But the impossibility that is at stake here is not merely logical impossibility. Kant suggests that the argument would have to be based on synthetic cognitions, which means that he already rules out that the argument could simply move from a presumed contradiction within the concepts of the objects in question. Yet, the negative part of the critique of pure reason has established that these objects are theoretically undecidable for us. Therefore, we cannot ever attain the required synthetic cognitions.

2  The Actuality of God and Immortality I will now turn my attention to Kant’s argument that we must commit ourselves to the actual existence of God and the immortality of the soul. 5 6

h

d i

0 0

Obviously, this is a consequence of the ‘negative’ side of the critique analysed in Chapters 5 and 6. Here, one might object that Kant does not rule out arguments that establish the non-existence of God starting from empirical premises, such as the argument from evil. In this case, the argument aims to establish the non-existence of God but does not do so by proving that God necessarily does not exist. This may be true. However, Kant does not consider this possibility in the quoted passage. He clearly states that an argument establishing that God and immortality do not exist must do so through pure reason. I take this to imply that the argument would have to show that these objects necessarily do not exist. I thank Andrew Chignell for this point.

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Positive Side of the Critique of Pure Reason

211

My focus in this chapter is on the systematic relevance of Kant’s practical argument to the project of the critique of pure reason. However, since many have doubted that this argument is successful, I will include some considerations on its cogency. Given that this argument has a peculiar place within Kant’s system of metaphysics, it is useful to begin by providing a general framework for locating it within the latter. a. The place of Kant’s practical argument within metaphysics. Kant often distinguishes between theoretical and practical cognitions by saying that the former describe what exists while the latter determine what ought to exist (A633/B661; see also A802/B830). Here, it is convenient to differentiate, first, between theoretical and practical propositions, where the former describe what exists and the latter describe what ‘ought’ to exist. As far as theoretical propositions are concerned, that they concern what ‘exists’ does not mean that they only describe what is given in the mode of actuality. Rather, Kant’s distinction is meant to point out that they lack the normative dimension of propositions that express ‘oughts’. However, in addition to affirming ‘it is the case that…’, they can also take the form ‘it is possible that…’ or ‘it is necessary that…’. As far as practical propositions are concerned, ‘ought’ has a very specific meaning. It expresses a prescription regarding an action that a rational subject is required to perform due to an imperative to which that subject has submitted herself (see A547/B575). One can then distinguish between theoretical and practical cognitions. Theoretical cognitions are expressible through theoretical propositions because they also concern what exists. Since they are cognitions, however, they are valid representations of what exists. Moreover, to speak of cognition in the theoretical domain, the object of the cognition in question should be either given in intuition or concern a priori features that determine how objects are given in intuition. For example, theoretical cognitions of this sort are obtainable through perception or, when we deal with a priori cognition, through the joint work of the categories and the forms of intuition. We might draw a similar distinction between practical propositions and practical cognitions. Accordingly, practical cognitions incorporate ‘oughts’ that are valid and are actually binding. For my purposes, however, it is more important to distinguish between theoretical propositions and theoretical cognitions because, as we will see, we can sometimes be justified in committing ourselves to a theoretical proposition even if this commitment concerns objects of which we cannot attain theoretical cognition. Among theoretical cognitions, those that derive from the

h

d i

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

212

theoretical use of reason are accessible a priori and concern what necessarily exists.7 Similarly, practical cognitions that derive from the practical use of reason are accessible a priori (A633/B661). At this point, we can say that the theoretical part of metaphysics concerns theoretical cognitions that derive from the theoretical use of reason, while the practical part of metaphysics concerns practical cognitions that derive from the practical use of reason (A633/B661). One might assume that metaphysics only comprises theoretical and practical cognitions, but this would be misleading.8 Rather, there are some theoretical propositions that demand a place within metaphysics even though they are not cognitions. This is the case because cognitions that determine what ought to exist do sometimes require us to make assumptions concerning what exists: Now if it is indubitably certain, but only conditionally, that something either is or that it ought to happen, then either a certain determinate condition can be absolutely necessary for it, or it can be presupposed as only optional and contingent. In the first case the condition is postulated (per thesin), in the second it is supposed (per hypothesin). Since there are practical laws that are absolutely necessary (the moral laws), then if these necessarily presuppose any existence as the condition of the possibility of their binding force, this existence has to be postulated, because the conditioned from which the inference to this determinate condition proceeds is itself cognized a priori as absolutely necessary. (A633–4/B661–2, translation altered)

Kant’s point is that when there is a binding moral law that prescribes that something ought to happen, and when realizing what the law prescribes is ‘conditioned’ in the sense that it is only possible on the assumption that a theoretical proposition describing something that ‘exists’ obtains, this assumption must be made, and not only in a problematic or provisional way (it must not be supposed ‘per hypothesin’) but rather assertorically (it must be postulated ‘per thesin’).9 7 8 9

h

d i

0 0

Reason is here understood in the broad sense as the faculty of a priori cognition. In fact, we already know that the theoretical part of metaphysics contains concepts and principles that are not themselves cognition. This is what regulative ideas and principles of reason are. Kant explicitly connects this commitment to theoretical propositions based on practical cognition to assertoric judgements. See for example 9:66; 24:732; 24:851. The line of reasoning Kant presents in this passage clearly reflects a version of the ‘ought implies can’ principle. The moral law tells us what we ought to do. If we ought to do something, it must be possible for us to do it. As a consequence, when we are under an obligation and performing the relevant action is only possible if a certain state of affairs obtains, we can in a certain sense reason to the conclusion that the state of affairs does obtain. Notice, however, that in order to use the ‘ought implies can’ principle in this way, the theoretical proposition in question must be undecided on theoretical grounds. That is to say, we cannot reason from the ‘ought’ to the ‘is’ if we have evidence that the theoretical proposition in question does not obtain. In this case, it seems that it would instead be appropriate to conclude that there was no ‘ought’ in the first place. The ‘ought implies can’ principle is notoriously

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Positive Side of the Critique of Pure Reason

213

Kant applies reasoning of this sort in the most straightforward way when discussing freedom. In discussing the third antinomy, for example, he presents an argument to the effect that we must assume ‘transcendental’ freedom, namely, the possibility of ‘spontaneous’ causes that are not determined by previous states of affairs. He first submits that we are capable of setting ‘oughts’ for ourselves (A534/B562), where in this case the ‘oughts’ can be based on either prudential or moral imperatives (A548/B576). These ‘oughts’ are valid even in cases where we do not follow them. Given the ‘ought implies can’ principle, this means that we could have acted according to these oughts even if we do not (A534/B562). From this Kant concludes that we must assume ‘transcendental’ freedom (A548/B576), for otherwise we could not account for the validity of these oughts, especially in cases where they are not adhered to. Since we are here focusing on the Canon, and since in that context Kant suggests that the question of ‘transcendental’ freedom is irrelevant to the practical domain (see A803–4/B831–2), I will not discuss his position on freedom further.10 For our purposes, freedom was important because it clearly exemplifies how we can reason from an ‘ought’ to an ‘is’ within metaphysics, or, using the terminology I introduced above, from a practical cognition to a theoretical proposition. The systematic role that theoretical propositions based on practical cognitions play within metaphysics remains an open matter. I will return to this issue in Section 4. b. Kant’s argument for the actuality of God and immortality. Let us see how Kant argues for a similar reasoning from an ‘ought’ to an ‘is’ in the case of the existence of God and the immortality of the soul. In fact, for these propositions, the reasoning to the ‘is’ is much more indirect. The need to assume them does not follow directly from the ‘ought’, which in this case is provided by a moral law. Rather, Kant first introduces the concept of the highest good as the concept of a world in which happiness is rewarded

10

h

d i

problematic and controversial, and I do not want to get into the difficulties it generates. Let me point out, however, that Kant can be interpreted as using a modest version of it (for a similar reading, see Stern 2004; see also Gava 2019b: 65). Kant usually stresses that, given a certain ought, we are required to assume or postulate a certain proposition describing a state of affairs. This vocabulary describes an attitude towards the ‘theoretical proposition’ in question that does not involve knowledge. Rather, we are simply justified in taking the proposition to be true because doing so secures the rationality of our action (we cannot be rational in acting according to the ought if we simultaneously hold that the action in question is impossible). On the problems arising from Kant’s seemingly contradictory claims on the importance of transcendental freedom for the possibility of morality, see Schönecker (2005).

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

214

proportionally to a person’s moral worth. Once this concept is secured, it is the need to grant the possibility of the highest good that requires us to assume the existence of God and the immortality of the soul. How do we get from the moral law to the highest good and from the highest good to God and immortality? Kant starts from the assumption that, as sensible beings, we all strive for happiness and, as rational beings, we stand under binding moral laws. Given this twofold constitution, the only way we can rationally pursue happiness is when the latter is subordinated to the condition of moral worth. This means that we ought to strive to become ‘worthy’ of happiness.11 Actual happiness is something we can only hope for as a consequence of our moral actions: I say, accordingly, that just as the moral principles are necessary in accordance with reason in its practical use, it is equally necessary to assume in accordance with reason in its theoretical use that everyone has cause to hope for happiness in the same measure as he has made himself worthy of it in his conduct […]. (A809/B837)

In this passage, Kant states clearly that the concept of the highest good generates a transition from the practical cognition of an ‘ought’ to a theoretical proposition. The status of the theoretical proposition is unclear, however, since it does not seem to simply describe a particular state of affairs. Rather, the proposition in question submits that we have ‘cause to hope’ for happiness in proportion to our moral conduct.12 Notice, moreover, that the happiness that we hope for is not something that can contingently happen to be proportional to our morality. Rather, we hope for happiness in proportion to morality insofar as the former is a consequence of the latter. Accordingly, Kant adds that we hope that ‘the system of morality is […] inseparably combined with the system of happiness’ (A809/B837). It is this hope that warrants the assumption of two additional theoretical propositions, that is, those that assert the existence of God and the immortality of the soul. It is unclear how a ‘hope’ for happiness that we somehow relate to our adherence to the moral law can justify a commitment to the existence of God and the immortality of the soul. Andrew Chignell has suggested that 11

12

h

d i

0 0

Here, I am only considering Kant’s characterization of the highest good in the first Critique. The concept evolved significantly during Kant’s critical period, and Kant in fact seemed to introduce different characterizations of it (see Yovel 1980: Ch. 1; Reath 1988). For a recent collection dedicated to this concept, see Höwing (2016b). The proportionality between happiness and moral worth that characterizes the highest good is notoriously obscure. For a recent attempt to make sense of it, see Marwede (2018: Ch. 6), who claims that proportionality should be understood in qualitative terms.

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Positive Side of the Critique of Pure Reason

215

Kant’s aim in the argument is to identify conditions for rational hope (Chignell 2014), where in the case of the highest good our hope can be characterized as ‘life-structuring’ (Chignell forthcoming: Ch. 10). A condition for the rationality of this type of hope is that we view the object of our hope not only as logically possible but as really possible (Chignell 2014: 106–7; see also Wood 2020: 33–4). This means that in the case of our hope for the highest good, in order for our hope to be rational, we cannot merely view the concept of the object of our hope as non-contradictory. We must additionally assume that certain grounds that can bring about that object are given (Chignell 2014: 106; Chignell forthcoming: Ch. 10). The only grounds that can play this role are God and the immortality of the soul (Chignell 2014: 107; Chignell forthcoming: Ch. 10). Kant’s practical argument in the Critique of Pure Reason is generally regarded as problematic and unsuccessful. Chignell also presents various worries. Of course, one can challenge the idea that God and immortality are the only grounds that can make the highest good ‘really’ possible. Moreover, Kant sometimes suggests that if we were to lack hope for the highest good, we would lack the proper motivation to act morally (see A811/B39; A812/B840), which generates problems for his deontological approach to ethics. For our purposes, however, another problem is more relevant: it is unclear whether claiming that hope for the highest good is ‘life-structuring’ is sufficient to justify the assumption that God and immortality actually exist. Even though we require ‘real possibility’ for such hopes, ‘hope’ might remain an attitude that is too weak to demand any commitment to actuality (see Guyer 2000: 334, 336). To solve this problem, Chignell argues that life-structuring hopes require real practical possibility, that is, the realizability of what we hope for, where this form of real possibility entails the existence of adequate grounds for that possibility (Chignell forthcoming: Ch. 10). But why are God and immortality needed to grant the real practical possibility of the highest good? Kant explains that in an intelligible world in which everybody acts morally, happiness would necessarily follow from morality, which would guarantee that happiness is a consequence of our morality, as the idea of the highest good requires (see A809/B837).13 This world is not our world, however, and even if we act morally, we cannot be sure that others will do the same. In a world like ours, if there were any proportionality between happiness and morality for a subject, this proportionality would be contingent, which means that happiness could not be 13

h

d i

Of course, one might ask why a moral world would warrant this fair distribution of happiness.

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

216

regarded as a consequence of moral action. Given this situation, the only way we can (really practically) enjoy happiness in proportion to our moral worth and as a consequence of it is if there is a supreme being that is the cause of nature and, as such, is able to distribute happiness according to its will. In Kant’s words, this ‘may be hoped for only if it is at the same time grounded on a highest reason, which commands in accordance with moral laws, as at the same time the cause of nature’ (A810/B838). As for immortality, Kant submits that in the sensible world we live in, we do not see the connection between morality and happiness that we postulate in the concept of the highest good. This connection is only (really practically) possible if there is a future life where we can be rewarded in proportion to our moral actions in the sensible world (see A810–11/B838–9).14 While Chignell’s approach is elegant and promising, one might still doubt whether characterizing the highest good as a life-structuring hope is sufficient to support a commitment to the actuality of God and immortality. As I said above, my aim here is not to determine whether Kant’s practical argument in the Canon is successful. Given that Kant seems to believe that licensing a commitment to the actuality of these objects is important to the aim of the critique, however, it is worth investigating the argument further. My strategy will be different from Chignell’s. I will not characterize our attitude towards the highest good as a life-structuring hope. Rather, I will try to make sense of the idea that when we act morally, we see ourselves as in a certain sense ‘entitled’ to expect no harm as arising from our action. To see this, let us focus on two characteristics of our attitude towards the highest good: first, it arises from a categorically binding obligation; s­ econd, it rests on the assumption that it would be in a certain sense ‘unfair’ if, in acting morally, we were systematically exposing ourselves to harm. Because we would be systematically exposing ourselves to harm in acting morally if there were no necessary connection between morality and happiness, and because we do not know whether this necessary connection is given, we see o­ urselves as in a certain sense entitled to assume that this necessary connection obtains. In what sense would we be systematically exposing ourselves to harm if, in acting morally, there were no necessary connection between morality and happiness, and in what sense do we take this to be unfair? As for the first ­question, we saw that the pursuit of happiness is a fundamental feature of our constitution. Morality requires that we subordinate the pursuit of happiness to its demands. This often means choosing not to pursue a course of action that 14

h

d i

0 0

Kant provides different arguments for establishing that the existence of God and the immortality of the soul are conditions for the highest good in the second and third Critiques. For a useful overview of this evolution, see Wood (2020: Ch. 2) and Chignell (forthcoming: Ch. 10).

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Positive Side of the Critique of Pure Reason

217

would be a more immediate route to satisfying our desires. But crucially, while in acting morally we sometimes have to renounce the immediate pursuit of happiness, at the same time we expose ourselves to harm due to the immoral actions of others. Therefore, in a world without a necessary connection between morality and happiness, it could be the case that, on the one hand, I resist the satisfaction of a desire due to a moral command and, on the other, I am harmed as a consequence of the morally wrong actions of others. Since this could happen systematically, I would be systematically exposing myself to harm in relying on moral imperatives to determine how to act. This is what Kant has in mind when he stresses that ‘the mere worthiness to be happy, is also far from being the complete good’ (A813/B841). As for the second question, given that the pursuit of happiness is a fundamental feature of our nature, it would be in a certain sense ‘unfair’ from our perspective if morality were to require us to totally disregard this characteristic of ours. I submit that the assumption that it would be unfair if acting morally involved systematically exposing ourselves to harm plays a crucial role in justifying belief in the actual existence of God and immortality. The thought is that, in following the obligation to act morally, I cannot rationally regard my action as implying that I am systematically exposing myself to harm. But God and immortality are conditions in the absence of which acting morally would involve systematically exposing myself to harm. This is because, without God and immortality, a necessary connection between morality and happiness would be ‘really impossible’. Therefore, as long as these objects are theoretically undecided, the obligation to act morally justifies my taking them as given.15 At this point, one might object that subordinating our pursuit of h ­ appiness to morality cannot be construed as exposing ourselves to harm.16 This ­subordination certainly requires us to give up the pursuit of happiness when the latter goes against a moral obligation, but it does not require us to refrain from pursuing happiness altogether. This is true. As suggested above, however, a world in which no necessary connection between morality and happiness 15

16

h

d i

Note that this way of constructing Kant’s argument does not build on Kant’s suggestion in the Canon that without the assumption of proportionality between happiness and moral worth, the motivation to act morally, or even the validity of the moral law itself, could waver (see A811/B39; A812/B840). This has often been regarded as the most serious problem in the Canon version of Kant’s practical argument, since claiming that the validity of the law, or the motivation to follow it, is dependent on securing hope that happiness will be distributed proportionally to moral worth would be catastrophic for Kant’s deontological ethics. Allison (1986) has accounted for these claims by saying that, at the time of the first edition of the first Critique, Kant had not yet developed his view that the moral law is sufficient to motivate us. More recently, Marwede (2016) has tried to make these claims consistent with Kant’s mature position, and Wood (2020: Ch. 2) has argued that Kant’s position in the Canon is not eudaimonistic or heteronomous. I thank Andrew Chignell for this objection.

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

218

obtained would be a world in which, in many particular cases, in renouncing happiness for the sake of moral commands we would at the same time expose ourselves to experiencing harm as a consequence of others’ moral failures. In this way, morality would require us to set aside the motive of happiness without guaranteeing that this will not result in harm. In such a world, renouncing the possibility of happiness in many relevant cases would seem to constitute exposing ourselves to harm. The case is structurally analogous to the following situation. Suppose I am on trial for a crime I did not commit, where I face a serious sentence. Suppose also that I have the opportunity to escape and enjoy life on a remote island. You might say that in this situation I have an obligation (at least a legal one) to show up for the trial. But when I adhere to that obligation, I cannot regard it as requiring that I expose myself to harm. Now, if the judge or the jury were corrupt or biased against me, in showing up for the trial I would be harming myself. Therefore, as long as I have no evidence of corruption or bias, the obligation justifies my assuming that the judge or the jury will perform their task as they should. Returning to God and immortality, of course, their existence would not guarantee that we will not experience harm due to the immoral actions of others. However, it would guarantee that the harm will somehow be ‘repaired’ with adequate happiness in the future. Two questions arise. First, why does this way of constructing Kant’s main point make any difference to the legitimation of our assumption of the actuality of God and immortality? In this framework, if God and immortality were not given, this would not simply make a life-structuring hope impossible; we would have to regard adherence to moral obligation as involving exposing ourselves to harm, which would give us a stronger reason to commit to their actuality. Second, why, if we are indeed justified to assume that these objects are given, are we not also justified to take a firmer attitude towards the highest good? In fact, it seems that if we can take God and immortality as given, and if God and immortality warrant a necessary connection between morality and happiness, this would be enough to support the same kind of attitude towards happiness as well. I believe this is right and shows that characterizing the attitude we are justified in having towards the highest good as a hope can be misleading. In this respect, Kant’s choice to only legitimate hope in the highest good may lie in the fact that while our commitment to the existence of God and immortality is directed towards objects and states that presently exist, the happiness promised in the highest good would be realized in the future. However, the fact that Kant only insists on hope does not mean that the argument he presents cannot justify a stronger attitude.

h

d i

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Positive Side of the Critique of Pure Reason

219

I have presented some considerations with the aim of making sense of Kant’s claim that moral obligation licenses an inference from an ‘ought’ to an ‘is’ in the case of God and immortality. This is not meant to suggest that Kant’s argument is successful. More importantly, even if these considerations are unconvincing, what is important for me in this Chapter is that Kant does believe that he has a valid practical argument that supports belief in God and immortality and that he includes that argument in the pages of the first Critique.

3  Kant’s Characterization of Belief In the Critique of Pure Reason, Kant not only presents an argument that justifies a commitment to the actuality of God and immortality but also makes sure to carefully distinguish the kind of attitude we can have when making that commitment. Famously, Kant submits that we can only have an attitude of ‘belief’ (Glaube). Therefore, before we move to determining the systematic place of Kant’s argument, we need to provide a characterization of ‘belief’.17 Kant describes ‘belief’ as a kind of ‘taking-to-be-true’ (Fürwahrhalten).18 This can be described as a positive attitude towards a proposition, such that we take it to truly represent some state of affairs (see A822/B850; 9:65–6). Kant identifies two other kinds of taking-to-be-true in addition to belief: opinion (Meinung) and knowledge (Wissen). He claims that opinion is both subjectively and objectively insufficient, while belief is subjectively sufficient and objectively insufficient and knowledge is both subjectively and objectively sufficient (see A822/B850; 9:66).19 A subjectively sufficient taking-to-be-true is based on grounds that are sufficient to produce full 17

18 19

h

d i

In the last two decades, the literature dedicated to this attitude and, more broadly, Kant’s concept of ‘taking-to-be-true’ (Fürwahrhalten) has increased exponentially (see, for example: Stevenson 2003; Chignell 2007; Pasternack 2014; Höwing 2016a). The aim of this section will be mainly expository. The following paragraphs are a re-elaboration of a section in Gava (2019b). Scholars have provided very different interpretations of these concepts (cf. Stevenson 2003; Chignell 2007; Pasternack 2014; Höwing 2016a; Gava and Willaschek, forthcoming). One way to display the difficulties faced by the interpreter in this respect is to compare the Jäsche Logic with the Canon of Pure Reason. In the former, Kant distinguishes between subjective and objective grounds and suggests that these grounds can be either sufficient or insufficient for taking-to-be-true (see 9:66–80). In the Canon, by contrast, he refrains from speaking of subjective grounds, which may suggest that grounds are by definition ‘objective’ and that one cannot consequently identify a subjectively sufficient taking-to-be-true by saying that it is based on ‘subjective’ grounds. As is clear from my reconstruction, I consider the Jäsche Logic compatible with the Canon in this respect. Accordingly, in my view, it is correct to refer to ‘subjective grounds’ in order to clarify Kant’s account of a subjectively sufficient taking-to-be-true in the Canon.

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

220

approval of a proposition in a particular subject,20 whereas an objectively sufficient taking-to-be-true is based on grounds that are sufficient to warrant the truth of the proposition we take to be true.21 We can call the grounds that are at issue in the latter case ‘objective’. They identify different kinds of evidence (mathematical, empirical, etc.) that are intersubjectively recognizable as valid. The grounds that are relevant in the first case can instead be called ‘subjective’. They can be either what we subjectively take to be objective grounds or distinctively subjective grounds that do not aim at objective validity.22 According to this characterization of subjective and objective sufficiency, we can describe belief as a taking-to-be-true in which, although we do not have sufficient objective grounds that warrant the truth of the proposition we take to be true, we nonetheless have subjective grounds that not only produce but also warrant full approval of the proposition. Therefore, in the case of belief, subjective grounds have not only ‘causal’ determining force over our approval but also justificatory force, which I will clarify in a moment. Importantly, in order for belief to obtain, we need to be conscious that the grounds we have are not objective and, consequently, that our ­taking-to-be-true is not a candidate for knowledge (see A822/B850). The only subjective grounds that are permissible in cases of belief are then genuinely subjective grounds, not what we falsely take to be objective grounds. This explains why subjective grounds of belief have not only a ‘causal’ determining force but also a justificatory force over our approval. In this case, it is the subjective grounds themselves that do the justificatory work. For Kant, these grounds are practical in nature. They license belief in a state of affairs because the latter is a condition for the realization of certain ends we pursue in practice. Kant’s thought seems to be the following: when the evidence we have leaves undecided whether a proposition is true or false (such that we lack sufficient objective grounds to decide the issue), we are rationally required to believe in that proposition if it identifies conditions that are necessary for the realization of ends we are pursuing. In other words, for Kant we cannot rationally pursue an end if we do not believe that the conditions for its attainment obtain. For this reason, once we actually have this end and follow it in our actions, we are rationally required to believe that 20 21 22

h

d i

0 0

I use ‘approval’ instead of the more natural ‘acceptance’ to avoid confusion with the contemporary use of ‘acceptance’ picked up by Chignell (2007). Objective sufficiency can, however, be read in a more modest way if one follows Kant’s characterizations of conviction (Überzeugung). See for example Gava (2016). For a similar distinction within subjective grounds, see Stevenson (2003) and Chignell (2007). For an attempt to portray subjective grounds in a more unitary way, see Höwing (2016a).

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Positive Side of the Critique of Pure Reason

221

the conditions of its realization are satisfied even if we lack decisive evidence on the matter (see A823–4/B851–2). Now that we have sketched the subjective grounds that are at play in cases of belief, let me turn to a specific kind of belief, that is, moral belief. These are beliefs where the ends we pursue are necessary (we are rationally required to pursue them) and we cannot but regard certain states of affairs as conditions for their realization (see A823–4/B51–2; A828/B856). Let us first consider necessary ends. The ‘necessary’ end that Kant connects to moral beliefs is the highest good (A828/B856). But why does Kant describe the highest good as a necessary end in this context? In Section 2, the highest good was characterized as the object of a hope that our moral conduct will be compensated with an adequate amount of happiness. In what sense can the object of this hope be an ‘end’ for us? Sure, Section 2 also raised doubts about whether the adequate attitude towards the highest good is hope. But this does not mean that the highest good can easily be construed as an end. Kant argues that God is a condition for the realization of the highest good precisely because establishing proportionality between moral worth and happiness is not up to us. The best way to make sense of Kant’s claim is to characterize the highest good, when construed as an end, as primarily being concerned with moral worth.23 The moral law directly determines how we ought to act, and it does so independently of our contingent ends. Because it can determine our actions in this way, the moral law issues ends that are necessary, since these ends directly arise from moral obligation. Here, the thought is that when I act because an action is morally prescribed, the moral law determines the end I pursue with my action. If this is correct, it means that the ‘end’ we pursue is specifically moral worth, as a component of the highest good (this closely agrees with what Kant says at A828/B856, where he submits: ‘It is entirely otherwise in the case of moral belief. For there it is absolutely necessary that something must happen, namely, that I fulfil the moral law in all points. The end here is inescapably fixed […]’). Let us now consider the second characteristic of moral belief, namely, that we cannot but regard certain states of affairs as conditions for the 23

h

d i

Of course, this does not mean that, as rational beings, we cannot pursue happiness as a direct end and must only pursue morality. We can of course pursue happiness through our actions as long as this activity does not violate the moral law. But note, first, that even in this case the pursuit of happiness is morally constrained and, second, that this is not the kind of moral constraint on happiness that is central to Kant’s account of the highest good. The case in question is one in which happiness is in our power but we must nonetheless pursue it by respecting the moral law. By contrast, the happiness that is at stake in the highest good is not up to us.

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

222

realization of these ends. This circumstance appears to be a direct consequence of what we have called the theoretical undecidability of God and immortality. Kant’s idea is that we cannot but view the existence of God and the immortality of the soul as conditions for the realization of this end (see A811/B839). That we cannot evaluate these conditions differently depends on the fact that our epistemic situation cannot change in relevant ways with respect to these objects. We see these objects as conditions of reaching the highest good on the basis of rational reflection, which is not dependent on contingent empirical factors. Moreover, because God and the soul are objects of pure reason, we cannot expect that our evidence for them will ever change. Kant emphasizes that the grounds that are at stake in moral belief are only ‘subjective’ and, as such, are not communicable like the objective grounds that can deliver knowledge (A828–9/B856–7). We find similar claims in the Jäsche Logic, where Kant submits that the ‘conviction’ we attain through moral belief is not communicable (9:70), and in lecture notes, where he suggests that believing is ‘private’ (24:732). Of course, that the grounds of belief are not communicable does not mean that the moral law on which they are based is not universally binding. As we just saw, however, in the case of moral belief, the moral law lies at the basis of an end that we pursue as individual subjects, where that end consists in acquiring moral worth. Since this is an end we follow as individuals, it can only serve as a subjective ground that we cannot share with others. Our aim of attaining moral worth is not their aim of attaining moral worth, even though the condition for attaining moral worth (agreement with the moral law) is identical for all. To sum up, Kant’s claim that the attitude we are justified in having towards the existence of God and immortality is one of moral belief is meant to provide further details regarding how we should interpret his practical argument in the Canon. It specifies what sort of attitude towards those objects is legitimated by that argument. Because the grounds that are at stake in moral belief are only ‘subjective’, the attitude cannot be taken as implying theoretical cognition.

4  The Systematic Place of Kant’s Practical Argument and Characterization of Belief At the beginning of this chapter, I suggested that Kant’s account of the ‘positive utility’ of the Critique in the B-Preface is insufficient to explain the inclusion in the Canon of Kant’s practical argument for belief in God

h

d i

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Positive Side of the Critique of Pure Reason

223

and immortality and his discussion of types of ‘taking-to-be-true’. I will now try to provide a satisfactory account of this inclusion. Of course, one way to account for it is to say that Kant wanted to counterbalance the negative consequences of the Critique. Since the Dialectic denied that we can have cognitions of God and immortality, Kant showed that there was a way to justify a commitment to these objects anyway, especially considering their importance for the practical part of metaphysics.24 This suggestion is unconvincing, however, for at least two reasons. First, it makes the inclusion of the argument superfluous to the project of the critique. Second, it seems that Kant’s account of the regulative use of ideas in the Appendix to the Dialectic already provided a counterbalance to the negative results of the critique. So why did he need to also add the argument in the Canon? Let us now take the perspective of the critique of pure reason, understood as the doctrine of method of metaphysics. I have suggested that the ‘positive task’ of this discipline is to show that metaphysics can achieve architectonic unity. As I argued in Chapter 1, there are two minimal conditions for legitimately attributing architectonic unity. First, the body of cognition in question must possess systematic coherence. Second, it must be able to be viewed as realizing the fundamental ‘idea’ of a science.25 The propositions asserting the existence of God and the immortality of the soul have a central importance as far as these conditions are concerned. On the one hand, they seem to be one of the main reasons why metaphysics cannot achieve coherence, given the unending disputes to which they give rise. On the other hand, however, they appear to deserve a spot in metaphysics if we want to realize its ‘idea’. This is the case because they are theoretical propositions that we regard as rationally required given a ‘ought’ that categorically binds us. More precisely, we see ourselves as rationally required to accept these propositions because otherwise we would have to regard the obligation to act morally as requiring that we systematically hinder our pursuit of happiness, which would involve exposing ourselves to harm in a world without a necessary connection between morality and happiness. Therefore, the critique, as the doctrine of method of metaphysics, must devise a strategy that is able both to put a stop to speculative conflicts over these objects and to establish that there is space for accommodating a commitment to them. In this vein, the critique does not provide evidence that every concept, principle or proposition that will form part of 24

25

h

d i

This picture is suggested by Andrew Chignell, who writes that the goal of the practical arguments in the Canon is to satisfy ‘the rational needs that the Discipline quashes in the epistemic sphere’ (Chignell 2017: 261). For a fuller account of these conditions, see Chapter 1 and the Introduction to Part III.

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

224

metaphysics will be given a proper foundation. It does not show that all of these concepts, principles and propositions will coherently hang together either. Rather, the critique focuses on those concepts and propositions that, while appearing essential to metaphysics, have been the source of hindrances that have prevented metaphysics from becoming a science. The propositions asserting the existence of God and the immortality of the soul have exactly this character. Taking this viewpoint, we can account for the elements of Kant’s position analysed in the previous sections as parts of a three-step strategy for establishing that metaphysics can achieve ‘architectonic unity’. Let us see how these three steps work. a. The theoretical undecidability of God and immortality. The first step of Kant’s strategy is identical with proving that God and immortality are theoretically undecidable. This is the means that Kant uses to put a stop to the speculative disputes that have prevented the achievement of coherence. Even though we have a natural tendency to make claims to cognition with respect to these objects, we must give up this aspiration insofar as these claims are simply illegitimate.26 Furthermore, theoretical undecidability not only puts a stop to these theoretical disputes but also creates a ‘safe’ space for practical arguments: because theoretical undecidability implies that we cannot ever theoretically disprove that God exists or that our souls are immortal, it establishes that any ‘practical’ argument that would be able to justify a commitment to these objects cannot be invalidated by a theoretical argument. b. Kant’s practical argument for the commitment to God and immortality. The second step consists in actually providing the practical argument we considered in Section 2. With respect to this argument, one thing to consider is whether it is part of the critique of pure reason, understood as the doctrine of method of metaphysics, or rather a component of metaphysics. I believe that the best way to ‘locate’ the practical argument and the corresponding commitment is to say that they form a ‘transition’ from the practical to the theoretical part of metaphysics. My idea is that Kant places the propositions to which we commit ourselves within the theoretical part of metaphysics, precisely because the propositions are ‘theoretical’. However, our commitment to these propositions is grounded in an obligation that is 26

h

d i

0 0

In Chapter 4, in discussing regulative ideas, we saw that there is a commitment to these objects that can be justified from the perspective of the theoretical part of metaphysics. This commitment is not a claim to cognition, however.

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Positive Side of the Critique of Pure Reason

225

analysed within the practical part of metaphysics. In this sense, the practical argument constitutes a transition from the practical to the theoretical part of metaphysics. Support for this view can be gathered from Kant’s claim in the Canon that the ‘moral theology’ established by his practical argument ultimately gives rise to both a ‘physico-theology’ and a ‘transcendental theology’ (A814–6/B842–4). Kant’s idea is the following: the practical argument provides reasons to assume God as the ground of a moral world where there is harmony between moral agents. Kant describes this as ‘a systematic unity of ends in this world of intelligences’ (A815/B843). In turn, thinking of this moral world as possible requires us to regard nature as itself created by God. This is because the actions of moral agents are performed in the sensible world, so it seems that there must be conditions that apply to the latter in order for the moral world to be realizable. Ultimately, this brings us to the concept of God as considered by transcendental theology, which is the concept of God that is obtained by simply taking into consideration ‘the inner possibility of things’ (A816/B844). These claims are interesting for at least two reasons. First, since physicotheology and transcendental theology were traditionally part of the theoretical part of metaphysics, Kant is arguing that his practical argument for the existence of God can be the ground of a commitment to a theoretical proposition which finds its place within that part of metaphysics. However, because the commitment is based on a practical argument, it cannot count as a theoretical cognition, understood as a ‘valid’ theoretical proposition about what there is.27 Second, the claims shed light on Kant’s inclusion of rational theology in the plan of metaphysics that he sketches in the Architectonic chapter, where rational theology is seen as belonging to ‘transcendent physiology’ (A845–7/B873–5). The inclusion of this discipline in the plan seems to conflict with the results of the Transcendental Dialectic, where it is denied that rational theology can become a science.28 But if there is a route from ‘moral’ to ‘transcendental’ theology, where commitment to God is ultimately grounded in a practical argument, it seems that we can make sense of the inclusion of ‘rational theology’ in 27 28

h

d i

Here, I read Kant’s reference to transcendental theology differently from Lea Ypi (2021: 16, 168–72), who claims that it displays a residual dogmatism on Kant’s part. The inclusion of ‘rational psychology’ is less challenging, even though Kant is obviously critical of the discipline in the Dialectic. Rational psychology is described as part of ‘immanent’ physiology. Accordingly, it can be taken to refer to the ‘metaphysical foundations’ of psychology. At the time Kant wrote the first edition of the Critique, he still believed that empirical psychology needed to be grounded in some metaphysical a priori principle on the model of the Metaphysical Foundations of Natural Science. See the Introduction on this point.

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

226

Kant’s sketch without assuming this means that rational theology is capable of ‘theoretical cognition’.29 If I am right, we find here a relationship between a part of metaphysics and the Critique that is analogous to what I suggested regarding transcendental philosophy. We have seen that the Critique contains parts of transcendental philosophy that are instrumental to pursuing its project as the doctrine of method of metaphysics. Similarly, the Critique contains the practical argument of the Canon, where this belongs to metaphysics and constitutes a transition from its practical to its theoretical part. However, while Kant suggests that the most important concepts and principles of transcendental philosophy are all established within the pages of the Critique and that only derivative concepts and principles are left out (see the Introduction to Part II), Kant does not establish the most important principle of the practical part of metaphysics within its pages, that is, the moral law. Rather, Kant simply assumes that there is such a binding law: ‘I assume [nehme an, italics mine] that there are really pure moral laws, which determine completely a priori (without regard to empirical motives, i.e., happiness) the action and omission, i.e., the use of the freedom of a rational being in general, and that these laws command absolutely (not merely hypothetically under the presupposition of other empirical ends), and are thus necessary in every respect’ (A807/B835).30 Once this assumption is made, Kant establishes that a commitment to God and immortality can follow from that law. Therefore, within the Critique Kant assumes that the moral law, which will be a topic of the practical part of metaphysics, is valid. On the basis of that assumption, he provides an argument that justifies a commitment to God and immortality, where this argument also belongs to metaphysics but constitutes a transition from its practical to its theoretical part. c. The characterization of belief. If the justification of the moral law and the practical argument in the Canon belong to metaphysics, what is the role of the critique of pure reason with respect to them? From Kant’s presentation 29

30

h

d i

0 0

Since the Critique of Pure Reason does not present a practical argument in support of cosmological ideas, it does not seem that we can make a similar suggestion to explain the inclusion of ‘rational cosmology’ in the picture (A845–7/B873–5). One could remark that rational cosmology may include theoretical propositions regarding transcendental freedom as based on a practical argument. However, we have already seen that Kant’s position in the Canon does not allow us to take this route, since he explicitly claims that we do not need to assume transcendental freedom from a practical standpoint. This claim directly opposes Michael Wolff’s interpretation of the Canon (2018), according to which the latter does not assume that there is anything like practical reason and, consequently, a priori moral laws. On his view, the Canon only depicts what would follow if there were practical reason.

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Positive Side of the Critique of Pure Reason

227

of his practical argument, it is not yet clear whether it violates the limits he himself imposed on our theoretical cognition of objects. Since the argument licences a commitment to theoretical propositions, we must establish that it does not conflict with the tool that Kant devised to put a stop to metaphysical disputes regarding God and immortality. This is what constitutes the third step in Kant’s strategy for establishing that metaphysics can achieve ‘architectonic unity’. The core of this third step is Kant’s characterization of belief. Since belief is a ‘taking-to-be-true’ that is based on subjective grounds, the commitment it licences cannot be a theoretical cognition. If we could claim cognition of God and immortality, we would be in possession of objective grounds for a ‘takingto-be-true’, where those grounds could possibly justify a claim to knowledge.31 But while belief in general can be partially based on objective grounds, in the case of moral beliefs like belief in God and immortality, our taking-to-be-true lacks any objective grounds, precisely because the objects of our belief are theoretically undecidable for us. Accordingly, when discussing moral belief in the Jäsche Logic, Kant submits that it ‘relates to objects in regard to which we not only cannot know anything but also cannot opine anything, indeed, cannot even pretend there is probability, but can only be certain that it is not contradictory to think of such objects as one does think of them’ (9:67). Objective grounds are those grounds that lend at least some degree of probability to a proposition’s being true, which means that in the case of moral beliefs we completely lack objective grounds and only have subjective ones. Since theoretical cognition would involve the possession of objective grounds that could justify a claim to knowledge, in the case of moral beliefs we cannot take ourselves to have theoretical cognition of the objects of our ‘taking-to-be-true’. We can summarize Kant’s three steps for establishing that metaphysics can achieve architectonic unity as follows. First, he establishes the theoretical undecidability of objects of pure reason such as God and immortality. This is instrumental to putting a stop to conflicts in metaphysics that have prevented the discipline from obtaining coherence. Second, Kant presents 31

h

d i

Notice that I am not conflating theoretical cognition with knowledge here. Rather, I am only saying that the objective grounds we have in theoretical cognition can be the basis of a claim to knowledge. Above, I suggested that theoretical cognitions are valid representations concerning what there ‘is’, where the object of the cognition in question is either given in intuition or an a priori feature of how objects are given in intuition. Kant sometimes suggests that we can have knowledge regarding objects we cannot cognize. For example, he argues that we know that things in themselves are not spatiotemporal (on the distinction between cognition and knowledge, see Willaschek and Watkins 2020). The claim that the objects of pure reason are theoretical undecidable is a claim that concerns the impossibility of cognizing those objects. Therefore, in order to safeguard this claim, it is important for Kant’s characterization of belief to make clear that belief does not involve cognition. Of course, it also makes clear that it does not involve knowledge.

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

The Method of the Critique of Pure Reason

228

a practical argument to support a commitment to God and immortality, which serves the purpose of including within metaphysics something that its proper ‘idea’ appears to require. Third, the identification of ‘belief’ as a particular kind of ‘taking-to-be-true’ shows that the commitment to God and immortality that the practical argument licences is not based on ‘objective grounds’ and does not constitute a case of theoretical cognition. In this way, the three steps together establish that consideration of God and immortality does not constitute a hindrance to meeting the two minimal conditions for legitimately attributing architectonic unity to a body of cognitions.32 While the first and the third steps properly belong to the critique of pure reason as the doctrine of method of metaphysics, the second step presents parts that belong to metaphysics. However, similar to what the critique does with transcendental philosophy, it uses those parts to establish that metaphysics can become a science.

5  The Critique of Pure Reason and Kant’s Practical Argument for God and Immortality In this chapter, I have tried to make sense of Kant’s inclusion of his practical argument for belief in God and immortality in the Critique of Pure Reason. I have suggested that reading the critique of pure reason as an analysis of our faculty of cognition that establishes what we can and cannot theoretically cognize a priori cannot account for that inclusion. Understood in this sense, the critique only establishes the theoretical undecidability of God and immortality. While establishing the theoretical undecidability of these objects makes room for a practical extension of pure reason, it does not by itself determine that this extension is indeed within our reach. Reading the critique of pure reason as the doctrine of method of metaphysics and understanding its ‘positive’ task as that of showing that metaphysics can attain architectonic unity provides a straightforward explanation for the inclusion of both Kant’s practical argument and his characterization of belief. Kant includes the former because it shows that metaphysics can realize what seems to be required from the perspective of its proper ‘idea’. He includes the latter because it shows that the practical argument can form a coherent part of metaphysics without endangering the ‘negative’ results of the critique. 32

h

d i

0 0

Clearly, since theoretical propositions concerning God and immortality are not (either theoretical or practical) cognitions, the ‘whole’ of metaphysics does not only contain cognitions. However, I think one can still speak of a ‘body of cognitions’ because the inclusion of these theoretical propositions serves the purpose of showing that the practical and theoretical cognitions belonging to metaphysics can form a whole and can coherently hang together.

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

Part IV

Kant on Dogmatism and Scepticism

Introduction to Part IV At the end of the last chapter of the Transcendental Doctrine of Method, the History of Pure Reason, Kant identifies dogmatism and scepticism as the two ‘methods’ in the history of philosophy that constitute the only alternative to his ‘critical’ method if one wishes to pursue metaphysics scientifically. In this context, he describes Christian Wolff and David Hume as key defenders of the dogmatic and the sceptical approaches, respectively (A856/B884). This last part of the book is dedicated to examining how Kant understood the relationship between his project in the Critique of Pure Reason, on the one hand, and the dogmatic and sceptical approaches of Wolff and Hume, on the other. Chapter 8 is dedicated to Kant’s critique of dogmatism and Wolff. It shows that this critique is much more complex and subtle than is usually assumed. In addition, I provide an account of Kant’s claim that metaphysics can indeed proceed according to Wolff’s dogmatic procedure once the Critique has done its job. I show how this claim relates to Kant’s critique of dogmatism and how it is compatible with my distinction between transcendental philosophy and the critique of pure reason. The latter distinction plays a fundamental role in Chapter 9, which is dedicated to Kant’s interpretation of Hume’s scepticism. In particular, I distinguish between three different readings of Kant’s interpretation of Hume. I reject the first as inadequate and then show that my distinction between transcendental philosophy and the critique of pure reason can be used to clarify how the other two are compatible. Importantly, when considering Wolff and Hume, my intention is neither to provide an interpretation of these philosophers nor to evaluate Kant’s criticisms of them in the light of such an interpretation. Rather, my

229

https://doi.org/10.1017/9781009172127.012 Published online by Cambridge University Press

230

Kant on Dogmatism and Scepticism

purpose is simply to clarify how Kant himself read these philosophers so that we might better understand the positions that Kant viewed as ‘rivals’ to his own critical approach.1 1

I consider whether Kant’s critique of Wolff’s application of the mathematical method to philosophy conceals elements of continuity between the two in Gava (2018a). See also McQuillan (2017c), who claims that Kant’s contention that Wolff proceeds without ‘critique’ is unwarranted.

https://doi.org/10.1017/9781009172127.012 Published online by Cambridge University Press

chapter 8

Kant on Wolff and Dogmatism

This chapter has two aims. First, I want to provide an interpretation of Kant’s critique of dogmatism, with Christian Wolff as the chief defender of that approach. Second, I aim to offer an account of Kant’s views on the relationship between his ‘critical’ method and Wolff’s ‘dogmatic’ method, since Kant suggests that the latter is not without its merits. I will start in Section 1 by considering Lanier Anderson’s (2015) recent account of Kant’s criticism of Wolff and the project of rationalist metaphysics.1 Anderson’s main claim in this regard is that Kant bases his attack on Wolff and rationalist metaphysics on his distinction between analytic and synthetic judgements. By means of this distinction, Kant is able to show that the kind of claims that Wolff defends in metaphysics are inevitably synthetic. Since Wolff’s argumentative strategy tries to establish that these claims are analytic conceptual truths, however, his project is destined to fail. In Section 2, I will distinguish between three different characterizations of dogmatism that Kant provides, where it is only the first two of these that portray dogmatism methodologically. I show that Kant regards Wolff as a dogmatist in these first two senses and explain in Section 3 how Kant can consistently hold that view. Once we distinguish between the two senses of dogmatism that Kant ascribes to Wolff, it becomes clear that Anderson’s account of Kant’s criticism of Wolff is insufficient. More precisely, Anderson provides an adequate account of Kant’s criticism of Wolff’s metaphilosophy, that is, of what Kant took Wolff to believe and say about the method philosophy should adopt. This does not completely capture Kant’s critique of Wolff’s actual line of argument, however. In Section 4, I will try to make sense of Kant’s claim that while we should avoid Wolff’s dogmatism, metaphysics should nonetheless adhere to Wolff’s dogmatic procedure once the ‘critique of reason’ is accomplished 1

For a useful discussion of Anderson’s book, see the symposium hosted in Studi Kantiani, with contributions by Lucy Allais (2017), Robert Hanna (2017), Giuseppe Motta (2017) and a response by Anderson (2017).

231

h

d i

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

232

(cf. Bxxxv–xxxvi). Finally, Section 5 will show that this claim is consistent with my account of the relationship between transcendental philosophy and the critique of pure reason.

1  Anderson on Kant’s Rejection of Wolffian Metaphysics In his recent book The Poverty of Conceptual Truth (2015), Lanier Anderson provides a careful and well-argued analysis of Kant’s distinction between analytic and synthetic judgements. The book makes two fundamental claims. First, it argues that Kant’s notion of analyticity, based on the idea of containment, is defensible if read against the background of the rationalist logic of concepts to which Kant was responding. Second, it maintains that Kant’s introduction of a distinction between analytic and synthetic judgements provides the fundamental materials for a powerful argument against the project of a rationalist metaphysics. A consequence of this reading is that the Critique of Pure Reason is seen as a ‘two-step argument against the Wolffian paradigm’ (Anderson 2015: 205). The first step is negative. It is based simply on the distinction between analytic and synthetic judgements and shows that the Wolffian approach, which is dependent on purely analytic judgements, is destined to fail. The second step is positive. It shows what kind of metaphysics is possible, given Kant’s account of the conditions of validity for synthetic a priori judgements. This second step is not itself part of Kant’s critique of Wolff and rationalism, which is already accomplished in the first step. It is rather an investigation that shows that abandoning Wolff’s paradigm does not entail abandoning all hope for metaphysical cognitions. Anderson describes his two-step approach as follows: In the first step, the critical philosopher identifies the expressive limitations of purely conceptual truth, and thereby of the traditional metaphysics itself. That step, based on the insight that there is a basic logical distinction between analytic and synthetic judgments, and that the important judgments of metaphysics will fall on the synthetic side, raises a deep problem: how are the synthetic judgments of metaphysics possible? Kant’s positive theory of synthesis in the ‘Transcendental Analytic’ of the Critique and his mature positive theory of the role of pure intuition in mathematical knowledge are meant to answer that question, and thereby make the second step. (Anderson 2015: 205; see also Anderson 2017: 130–1)

My concern here is with Anderson’s interpretation of the first step of Kant’s argument. Anderson needs to secure three fundamental claims to make his reading of the first step work. First, he needs to show that Wolff

h

d i

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Wolff and Dogmatism

233

is committed to a ‘containment’ account of truth, according to which every true judgement is an analytic judgement where the subject concept contains the predicate concept within it. Second, he needs to indicate how Kant demonstrates that there are some truths that cannot be reduced to this model. Third, he needs to show how Kant establishes that the truths that metaphysics tries to prove necessarily fall under this latter category. Only once these claims are secured can Anderson argue that Kant is able to show that Wolff’s metaphysical project is destined to fail, since the latter can only establish analytic truths while the truths that metaphysics pursues are instead essentially synthetic. Anderson gathers evidence for the first claim in Wolff’s idea that a true system of philosophy provides a ‘logically ordered hierarchy of adequate concepts standing in genus/species containment relations’ (Anderson 2015: 76). In Wolff’s logic of concepts, an ‘adequate’ concept is a concept in which we have not only a clear grasp of the constituent marks of the concept, as in the case of a ‘distinct’ concept, but also a ‘distinct’ grasp of them. This means that we are able to bring our analysis of the concept one step further. We not only clearly identify the marks that constitute the concept but are able to isolate the ‘sub-marks’, as it were, that form those very marks (see Wolff 1965 [1754]: 130–1). In this way, we can achieve a complete representation of the internal conceptual structure of our concepts.2 With this representation at our disposal, we can then organize our concepts into genus/species relationships, where the higher genus concepts are ‘contained in’ lower species concepts and the latter are ‘contained under’ the higher genus concepts (see Anderson 2015: 49–50). Here, ‘contained in’ means that the genus concepts figure as marks in the species concepts. By contrast, ‘contained under’ indicates that a group of species concepts belong to the same genus concept because they all contain that genus concept in them. Anderson’s point here can be put as follows: if a scientific body of knowledge is achieved by means of the systematic organization of our concepts obtained thanks to conceptual analysis, it seems that every truth in this system should be regarded as analytic, since every true judgement in the system can be traced back to a containment relationship between the subject concept and the predicate concept of the judgement in question. The second claim that Anderson needs to secure is that Kant successfully demonstrates that there are some truths that cannot be reduced to this model of conceptual containment. Kant’s philosophy of mathematics 2

h

d i

For reservations about whether complete adequacy can in fact be achieved for Wolff, however, see Wolff (1965 [1754]: 131–32).

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

234

plays an essential role for Anderson in this respect. Accordingly, he maintains that by means of his account of mathematics, Kant is able to ‘establish the substantive point that some serious knowledge is irreducibly synthetic’ (Anderson 2015: 264). The essential point here is that being able to represent an equivalence between non-identicals is essential to mathematics (Anderson 2015: 231). Only thanks to this relation of equivalence can we make sense of a quantitative composition that yields more of the same (A + A = 2A) (Anderson 2015: 228–9). By contrast, the relations of equivalence that are achievable by purely conceptual means are ‘strict’ (Anderson 2015: 231) and never yield more of the same, but only the same concept over again (A + A = A) (Anderson 2015: 230). The fact that a fundamental part of our scientific knowledge like mathematics is synthetic does not imply that metaphysics is necessarily synthetic as well, however. For this reason, Anderson needs to secure a third claim according to which Kant can establish precisely the latter point. The contention that metaphysics is necessarily synthetic is established by what Anderson calls the ‘master argument’ of the Transcendental Dialectic (see Anderson 2015: Ch. 10). Roughly, the argument can be summarized as follows: metaphysics aims to establish truths regarding what must exist. Judgements based on containment analyticity are unable to establish such truths, since they are true simply in virtue of their constituent concepts, independently of whether those concepts refer to real objects or not. It is only synthetic judgements that can establish that an object must exist, because in this case the ‘truth-sustaining’ relationship between the concepts that form the judgement obtains not because of a containment relation but because the two concepts essentially figure in a judgement that correctly describes an existing object. That is to say, it is only through the object that the relationship between the two concepts can be established. For this reason, the synthetic judgement, if true, can establish that the object must exist (see Anderson 2015: 271–7). When these three claims are secured, Anderson can successfully argue that Kant provides a powerful argument against Wolffian ­metaphysics. The argument establishes that metaphysical truths are necessarily s­ ynthetic, and, since the Wolffian approach can only establish analytical truths, it is ­inadequate for our metaphysical aspirations. Anderson’s r­econstruction of Kant’s rejection of the Wolffian paradigm is an illuminating and useful tool. It captures an essential strand of Kant’s critique of Wolff.3 Nevertheless, 3

h

d i

0 0

This is not the dominant strand in the Transcendental Dialectic, however. Anderson reads Kant’s strategy there as one that displays the limits of conceptual analysis (see Anderson 2015: Ch. 10). Since Wolff and the rationalists are committed to conceptual analysis, they cannot reach the synthetic

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Wolff and Dogmatism

235

it is insufficient to account for Kant’s critique of Wolff in all its complexity. I will now turn to my analysis of Kant’s criticism of dogmatism and Wolff as a defender of this approach, where other aspects of Kant’s critique of Wolff become apparent.

2  Three Senses of Dogmatism in the Critique of Pure Reason Kant describes dogmatism as one of the chief critical targets of the first Critique and Wolff as the main representative of that approach (A856/ B884). But how does Kant characterize dogmatism? In the Critique of Pure Reason, Kant provides at least three different characterizations of dogmatism. Two of these are directly related to his views on Wolff’s method, but it is only one of the latter two that is adequately captured by Anderson’s reconstruction of Kant’s criticism of Wolff. A second and arguably more fundamental criticism is instead connected to what I will here discuss as Kant’s second characterization of dogmatism. The three characterizations are: (1) dogmatism as the pursuit of a demonstration ‘from concepts’; (2) dogmatism as the absence of critique and the unwarranted use of synthetic a priori principles; and (3) dogmatism as the affirmation of the theses of the Antinomy of Pure Reason. a. Dogmatism as the pursuit of a demonstration ‘from concepts’. According to Kant’s first characterization of dogmatism (hereafter dogmatism1), the latter is the attempt to establish metaphysical truths by the sole means of conceptual analysis. This account is to be found in the B-Introduction to the first Critique, where Kant claims that one cannot acquire synthetic metaphysical truths by proceeding dogmatically: Thus one can and must regard as undone all attempts made until now to bring about a metaphysics dogmatically; for what is analytic in one or the other of them, namely the mere analysis of the concepts that inhabit our reason a priori, is not the end at all, but only a preparation for metaphysics proper, namely extending its a priori cognition synthetically, and it is useless for this end, because it merely shows what is contained in these concepts […]. (B23)

Kant formulates a similar objection to dogmatism twice in the Analogies. In the context of the First Analogy, he equates dogmatic proofs with conclusions they are aiming at by their own standards. By contrast, I think that Kant’s move against Wolff and the rationalists in the Transcendental Dialectic aims to show that they make an illegitimate use of synthetic a priori principles by Kant’s own critical standards.

h

d i

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

236

demonstration ‘from concepts’ and laments the inadequacy of this procedure for establishing the persistence of substance (A184–5/B227–8). Kant argues that the proposition asserting that substance persists is synthetic a priori and is valid only for objects existing in space and time, that is, objects of possible experience. Kant’s point is that it is only when we think of the concept of substance as a condition for making sense of temporal relationships like simultaneity and succession that we must see substance as persistent. Persistence is not something that belongs to the concept of substance as such, but is rather something that we necessarily connect to that concept when we use it in judgements about objects of possible experience (in time).4 It is for this reason that, according to Kant, we cannot establish the persistence of substance by analysing the latter concept. Subsequently, Kant extends this consideration to the Analogies in general and stresses that, since the principles defended there are all synthetic a priori, [i]f we had wanted to prove these analogies dogmatically, i.e., from concepts – namely, that everything that exists will only be encountered in that which persists; that every occurrence presupposes something in the previous state, which it follows in accordance with a rule; finally, that in the manifold that is simultaneous the states are simultaneous in relation to each other in accordance with a rule (stand in community) – then all effort would have been entirely in vain. For one cannot get from one object and its existence to the existence of another or its way of existing through mere concepts of these things, no matter how much one analyses them. (A216–7/B263–4)

Therefore, the error committed by dogmatism1 is that it tries to establish synthetic metaphysical principles by analytic means alone. In this respect, the role of the critical philosopher is to show the dogmatist1 that her attempts are ‘in vain’. To do that, what is needed is simply a formulation of the distinction between analytic and synthetic judgements, paired with evidence that the dogmatist can only establish analytic judgements and that metaphysical truths are irremediably synthetic. This description of dogmatism and its main mistake matches almost perfectly with Anderson’s reconstruction of Kant’s argument against Wolffian metaphysics. In fact, in the Prolegomena, Kant expresses a very similar point directly against Wolff. In that context, he complains that dogmatists1 do not distinguish between synthetic and analytic judgements and, for this reason, try to establish synthetic principles as if they were analytic 4

h

d i

0 0

On Kant’s account of substance in the First Analogy, see Guyer (1987: Ch. 9), Van Cleve (1999: Ch. 8), Ward (2001) and Hall (2011).

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Wolff and Dogmatism

237

truths. According to Kant, this is exactly what Wolff and Baumgarten do when they try to establish the validity of the principle of sufficient r­ eason by deriving it from the principle of contradiction (4:270).5 b. Dogmatism as the absence of critique. Whereas dogmatism1 fits very well with Anderson’s reading of Kant’s criticism of Wolff, this is not true of the second characterization of dogmatism that we find in the Critique of Pure Reason, hereafter dogmatism2. Dogmatism2 is ‘the presumption of getting on solely with pure cognition from (­philosophical) concepts according to principles, which reason has been using for a long time without first inquiring in what way and by what right it has obtained them’ (Bxxxv). Here, Kant still connects dogmatism to a procedure ‘from concepts’. He adds an important qualification, however: in proceeding from concepts, the dogmatist makes use of certain principles without having a clear grasp of their origin and v­ alidity. Dogmatism is thus equated to the use of these principles without an ­antecedent critique (cf. Bxxx; Bxxxv; B7; 9:83–4). But precisely what principles does Kant have in mind? Kant makes this explicit in On a Discovery, where he begins his response to Eberhard with a clarification of what he means when he uses the terms ‘dogmatism’ and ‘scepticism’ in the Critique of Pure Reason. In this context, he first defines dogmatism as ‘the general trust in its principles [that is, of metaphysics], without a previous critique of the faculty of reason itself, merely because of its success’ (8:226). Kant clarifies what it means to say that these metaphysical principles are used ‘successfully’ in a footnote, where he writes that ‘[s]uccess in the use of principles a priori lies in their constant confirmation in ­application to experience’ (8:226n). Therefore, the principles assumed by ­ riori the dogmatist are metaphysical because they are used to determine a p features of objects. They are used ‘successfully’ because we constantly use them to determine features of objects of experience. It is because of this ‘success’ that the dogmatist assumes them without critique. In the same footnote, Kant explicitly says that the principles that the ­dogmatist unduly assumes are those that the Critique of Pure Reason considers in the Analytic. As we know, Kant thinks that these principles are synthetic a ­priori and valid only within the boundaries of possible experience. The error of the dogmatist is thus that of assuming these synthetic a priori ­principles (because of their successful application within possible experience) 5

h

d i

Unsurprisingly, Anderson cites the passage from the Prolegomena several times (see Anderson 2015: 11, 13, 21, 89, 136–7).

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

238

while failing to clarify the conditions of their legitimate use (8:227n). As a consequence, the dogmatist unjustifiably applies these principles to objects that cannot be given in possible experience, so that ‘a dogmatism arises in regard to the supersensible’ (8:227n). Dogmatism2 is related to quite a different account of the role of the critical philosopher in responding to such a view. The critical philosopher does not have to show that the attempt to ‘analytically’ build a metaphysics ‘from concepts’ is doomed to fail, as in the case of the critical response to dogmatism1. Rather, she needs to show that the dogmatist2 makes an illegitimate use of certain synthetic a priori principles that are assumed without critique. To perform the latter task, it is not sufficient to introduce a distinction between synthetic and analytic judgements and to prove that while the dogmatist can only establish analytic judgements, metaphysical truths are irremediably synthetic. In addition, we must already have a clear notion of a synthetic a priori judgement and the conditions of its validity. Only with such a notion at our disposal can we show that the dogmatist2’s use of synthetic a priori principles is illegitimate. As should now be clear, this account of the critical response to dogmatism does not fit well with Anderson’s reconstruction of Kant’s main argument against Wolffian metaphysics. Assuming that Kant had Wolff (among others) in mind when presenting his response to dogmatism2, Kant’s argument against Wolff would not be that he lacks the means to establish synthetic metaphysical truths, but rather that he illegitimately assumes certain synthetic a priori principles. In a Reflection from 1777–1778, Kant explicitly makes this point against Wolff: ‘Wolff did great things in philosophy; but he got ahead of himself and extended cognition without securing, altering, and reforming it through a special critique’ (Refl. 5035, 18:68; cf. Refl. 4866, 18:14). I take it that when Kant talks about the extension of cognition here, he has synthetic a priori judgements in mind. It therefore seems that Anderson’s reconstruction of Kant’s argument against Wolff does not cover what Kant critically says about Wolff as a dogmatist2. We might rescue Anderson’s strategy by claiming that it in fact has everything it needs to support Kant’s move against Wolff as a dogmatist2. Anderson’s reconstruction of Kant’s argument against Wolff does establish that Wolff can only accept and demonstrate analytic truths. It could be argued that this is enough to demonstrate that any assumption of synthetic a priori principles is illicit within the Wolffian paradigm. Kant’s strategy against Wolff as a dogmatist2 would thus consist in revealing an internal inconsistency in Wolff’s system: while Wolff, according to his

h

d i

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Wolff and Dogmatism

239

own standards, can only accept analytic truths, he in fact assumes synthetic a priori principles as well.6 This is not how Kant frames his critique of dogmatism2, however. Kant’s point is not that dogmatism2 makes an illicit use of synthetic a priori principles according to the standards set by the dogmatist2 herself. Rather, the problem is that dogmatism2 makes an illicit use of synthetic a priori principles according to Kant’s critical standards. This becomes evident if one takes into consideration two sets of remarks made by Kant. First, Kant often suggests that assuming synthetic a priori principles without critique is not a serious problem so long as we use these principles within their proper boundaries. In the Discipline of Pure Reason, for example, he criticizes those scholars who, in trying to address philosophical problems by means of a mathematical approach, ‘merely’ use the pure concepts of the understanding (and the synthetic principles they ground) without investigating their ‘origin’ and ‘the scope of their validity’ (A725/B753). He concedes, however, that ‘[i]n all of this they proceed quite correctly, as long as they do not overstep their appointed boundaries, namely those of nature’ (A725/B753). This suggests that it is fine for the dogmatist2 to assume certain synthetic a priori principles without critique so long as she ‘by chance’ uses them within their proper boundaries. This is acceptable even if, according to her own self-understanding, the dogmatist2 can only accept analytic truths. This means that Kant’s criticism of dogmatism2 is not directed at displaying an internal inconsistency. Rather, Kant is interested in showing that the dogmatist2 makes improper use of synthetic a priori principles according to Kant’s critical standards. Second, Kant argues that we cannot really put a stop to dogmatism2 without a clear grasp of the conditions and limits of the validity of synthetic a priori principles. Kant makes this point in his criticism of the sceptic’s strategy against dogmatism2, which he describes as follows: in order to put a stop to the dogmatist2’s unwarranted assumption of synthetic a priori principles, the sceptic questions our capacity for synthetic a priori judgements tout court. Kant contends that this strategy is ultimately unsuccessful against the dogmatist2. Given that some synthetic a priori judgements are fundamental to our experience, an unqualified prohibition against assuming any such judgements is insufficient to stop the dogmatist2. The dogmatist2 will be driven back to them and build inferences on them without any understanding of the conditions and limits of their 6

h

d i

This is in fact how Anderson frames the ‘master argument’ of the Transcendental Dialectic against the Wolffian paradigm (cf. Anderson 2015: Ch. 10).

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

240

validity (cf. A768/B796).7 This suggests that simply establishing that the assumption of synthetic a priori principles is inconsistent with the dogmatist2’s self-understanding may be insufficient to prevent her from using such principles. She might adjust this self-understanding in a way that allows for this assumption but still lacks a critical grasp of the conditions of the validity of synthetic a priori judgements. Therefore, Anderson’s construal of Kant’s argument against Wolff cannot account for Kant’s critique of Wolff as a dogmatist2. c. Dogmatism as affirmation of the theses of the Antinomy. There is one last characterization of dogmatism in the Critique of Pure Reason, and it is discussed in the Antinomy of Pure Reason. More precisely, Kant introduces the term in the section dedicated to the different ‘interests of reason’ that govern the arguments in the antinomies. The purpose of the section is to clarify why one might hope that each of the claims analysed in the Antinomy are true and so be inclined to argue for them. Kant thinks that there are underlying common ‘interests’ that explain why we are led to affirm each of the theses. Similarly, there are underlying common interests on the side of the antitheses, as well. The common interests on each side are at the basis of two ‘maxims’ that guide the arguments of the theses and the antitheses, respectively. The maxim guiding the arguments for the antitheses is a ‘principle of pure empiricism, not only in the explanation of appearances in the world, but also in the dissolution of the transcendental ideas of the world-whole itself’ (A466/B494). Kant suggests that the maxim motivating the arguments for the theses is more complex and involves both an empiricism ‘within the series of appearances’ and the assumption of ‘intellectualistic starting points’ (A466/B494). Kant calls the maxim of the theses a principle of dogmatism (hereafter dogmatism3) (A466/B494). Therefore, dogmatism3 is understood as the maxim governing the affirmation of each of the theses in the Antinomy of Pure Reason, according to which: the world has a beginning in time and space is enclosed in boundaries (A426/B454); substances are made of simple parts (A434/B462); causality through freedom must be assumed (A444/B472); and there must be an absolutely necessary being (A452/B480). Kant does not clarify the principles of pure empiricism and dogmatism3 further. What he seems to have in mind with the former is that the antitheses simply affirm a claim that is central to experience and gives rise 7

h

d i

0 0

For an analysis of Kant’s critique of Hume and scepticism, see the following chapter.

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Wolff and Dogmatism

241

to a ‘regress’ from a ‘conditioned’ to a ‘condition’, and from this ‘condition’ to a further ‘condition’, and so on (every given time is preceded by another time/every given region of space is enclosed within another space; an object in space is always divisible; every event in time has a preceding cause; every contingently existing being in the world has a condition for its existence). By contrast, the theses argue that in order to make sense of such a claim, one must assume the truth of another claim which holds that there is an absolute first condition where the regress stops. It is in this sense that dogmatism3 involves both empiricism ‘within the series of appearances’ and the assumption of ‘intellectualistic starting points’.8 Therefore, what is distinctive about dogmatism3 is, on the one hand, the assumption of these ‘intellectualist starting points’ and, on the other, the fact that this assumption is encouraged by certain ‘interests’, which the existence of these starting points would permit us to satisfy.9 With this overview of the different senses in which Kant characterizes the concept of dogmatism in the Critique of Pure Reason in hand, we can now examine how they relate to each other. In this respect, ascribing dogmatism3 involves a different standpoint in comparison to dogmatism1 and dogmatism2. What is distinctive about dogmatism3 is its defence of a particular set of claims, which are important from the standpoint of certain of our interests. By contrast, dogmatism1 and dogmatism2 involve following a certain procedure in arguing for a claim. We either try to establish the claim by arguing analytically ‘from concepts’ (dogmatism1) or uncritically assume certain synthetic a priori principles without inquiring into the conditions of their application and validity (dogmatism2). Of course, since the arguments analysed in the Antinomy follow a dogmatic procedure for Kant, we arrive at the claims made by dogmatism3 by proceeding according to either dogmatism1 or dogmatism2. But this is not something that characterizes dogmatism3 as the form of dogmatism that it is, for Kant believes that both the theses and the antitheses in the Antinomy chapter are obtained by following a dogmatic procedure (cf. A420/B448; 4:340). That means that when one follows the ‘principle of pure empiricism’ and affirms the antitheses, one proceeds according to either dogmatism1 or dogmatism2, too. But since dogmatism3 is contrasted to the principle of pure empiricism, which also follows either dogmatism1 or dogmatism2, 8

9

h

d i

That there are such intellectual starting points should nonetheless be compatible with Kant’s claim that the totalities of conditions that are at stake in the Antinomy are thought as ‘totalities of appearances’ (see Chapter 5). Kant identifies both a ‘practical’ (A466/B494) and a ‘speculative’ interest (A467/B495) related to dogmatism3. He also submits that the latter has the ‘merit of popularity’ (A467/B495).

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

242

that dogmatism3 is called ‘dogmatism’ has little to do with the fact that it pursues a certain procedure of argument. What about the relationship between dogmatism1 and dogmatism2? They are both characterizations of dogmatism that describe the procedure we follow in arguing for philosophical claims, but they seem incompatible at first glance. Dogmatism1 proceeds only analytically ‘from concepts’ and tries to establish a system of analytic truths. Dogmatism2, by contrast, proceeds synthetically, since it uncritically assumes synthetic a priori principles. One way to account for this difference is to say that Kant simply identifies two different ways in which one can proceed dogmatically in philosophy, such that dogmatism1 and dogmatism2 need not be compatible. As we have seen, however, Kant reads Wolff as being both a dogmatist1 and a dogmatist2. This might not mean much, since it is of course possible that Kant understands some arguments put forward by Wolff as displaying dogmatism1, while some other arguments are instead guilty of dogmatism2. In fact, this is exactly the view that Kant suggests in the context of a lecture on metaphysics from the 1780s, when he discusses Wolff’s arguments for the principle of sufficient reason. We have already seen above that Kant complains in the Prolegomena that Baumgarten and Wolff tried to establish the truth of this principle analytically, by deriving it from the principle of contradiction (4:270). In the lecture, Kant suggests that Wolff certainly realized that his attempt at an analytic proof of the principle of sufficient reason was unsuccessful. For this reason, Kant continues, Wolff conceded that the principle could be assumed on the basis of common sense (29:788). This suggests that while Wolff’s attempt at an analytic proof of the principle of sufficient reason was dogmatic in the sense of dogmatism1, his assumption of the principle on the basis of common sense was dogmatic in the sense of dogmatism2. This is, of course, a conceivable way in which one can be both a dogmatist1 and a dogmatist2. Nevertheless, dogmatism1 and dogmatism2 are related in a more complex way in Kant’s critique of Wolff’s dogmatism. This is what I will explore in the next section, where I will draw a distinction between Kant’s critique of the metaphilosophical views held by Wolff and his critique of Wolff’s actual method of argument.

3  Two Levels of Critique Given that what characterizes dogmatism3 has little to do with method, which is the focus of this book, and given that, as we saw, dogmatism3 has no direct connection to dogmatism1 or dogmatism2, in what follows I

h

d i

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Wolff and Dogmatism

243

will focus on the relationship between dogmatism1 and dogmatism2. We know that dogmatism2 is distinctive because it is a procedure that, in trying to establish philosophical claims, assumes synthetic a priori principles without critique. In the Introduction to the Critique of Pure Reason, Kant provides an explanation of why we have a ‘natural tendency’ to proceed in this way. Kant’s remarks concern philosophical claims about objects that cannot be given in experience (A3/B6–7), which, in the B-edition, are specified as God, freedom and immortality (B7). So why do we have a ‘natural tendency’ to argue for claims about these objects by assuming synthetic a priori principles without critique? The first explanation that Kant gives appeals to the success of mathematics. Since mathematics provides ‘a splendid example of how far we can go with a priori cognition independently of experience’ (A4/B8), we are led to assume that we can easily have the same success in pursuing metaphysical questions regarding supersensible objects (see also A712–13/ B740–1; A724–5/B752–3). In doing so, we fail to realize that there is a fundamental difference between mathematics and philosophy. While mathematics can construct its objects in pure intuitions and so have ‘confirmation’ that its a priori claims are valid for those objects, this possibility is not open to philosophy, especially when it makes claims concerning God, freedom and immortality – objects that, as supersensible, cannot be given in sensible intuition by definition. Kant’s critique of the use of the mathematical method in philosophy deserves close attention,10 but in the context of the present chapter, I am more interested in the second explanation that Kant provides. According to the latter, in philosophy we are naturally led to assume synthetic a priori principles without critique because the majority of cognitions that we gain through pure reason are analytic, and this leads us to falsely believe that synthetic a priori truths are also analytic. As Kant puts it: A great part, perhaps the greatest part, of the business of our reason consists in analyses of the concepts that we already have of objects. This affords us a multitude of cognitions that, although they are nothing more than illuminations or clarifications of that which is already thought in our concepts (though still in a confused way), are, at least as far as their form is concerned, treasured as if they were new insights, though they do not extend the concepts that we have in either matter or content, but only set them apart from each other. Now since this procedure does yield a real a priori cognition, which makes secure and useful progress, reason, without itself 10

h

d i

On Kant’s distinction between the methods of philosophy and mathematics, see Gava (2015) and Gava (2018a).

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

244

noticing it, under these pretences surreptitiously makes assertions of quite another sort, in which reason adds something entirely alien to given concepts and indeed does so a priori, without one knowing how it was able to do this and without such a question even being allowed to come to mind. (A5–6/B9–10)

These remarks add a new element to our characterization of dogmatism2. When the dogmatist2 builds her philosophical arguments by assuming synthetic a priori principles the validity of which she has not confirmed, she does so on the false belief that she is proceeding analytically, solely on the basis of the analysis of concepts. Therefore, if we only consider what she believes she is doing, she thinks she is following a method that matches what Kant calls dogmatism1. She believes she is establishing metaphysical truths by simply arguing ‘from concepts’. Where does all this bring us? It provides a more complex picture of how dogmatism1 and dogmatism2 can be combined, according to Kant. In this combination, dogmatism1 applies first of all to the metaphilosophical views held by a particular philosopher, that is, to her beliefs and theory concerning the method she is following. By contrast, dogmatism2 pertains to the procedure she actually employs. Of course, the fact that the method that she follows does not always reflect her metaphilosophical views does not mean that her beliefs about her method and her actual method must always diverge. It is possible for her to be a dogmatist1 both in terms of her metaphilosophical views and in terms of the arguments she in fact proposes. To also be a dogmatist2 in practice, she only needs to sometimes assume synthetic a priori judgements without noticing it and to build ­philosophical arguments on their basis. What is interesting about this characterization of the relationship between dogmatism1 and dogmatism2 is that it has obvious consequences for how we should account for Kant’s critique of dogmatism, broadly construed. That is, we must grant that Kant’s criticisms operate on two levels: a metaphilosophical level that concerns the views on method consciously held and defended by a given philosopher, and a methodological level that points to the procedure that the philosopher actually follows, consciously or not.11 As we saw above, Kant’s criticism of dogmatism1 argues that, 11

h

d i

0 0

It may seem odd to distinguish between a metaphilosophical and a methodological level in Kant’s critique of dogmatism, since the method of philosophy is usually considered a topic within metaphilosophy. However, the dogmatist’s metaphilosophical views and her actual method come apart when her dogmatism proceeds in a way that does not correspond to her theory of method. Therefore, while the metaphilosophical level of Kant’s criticism targets the dogmatist’s theory of method, the methodological level also takes into account those elements of the dogmatist’s method that do not reflect or agree with her theory.

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Wolff and Dogmatism

245

since the metaphysical truths that the dogmatist1 aims to establish are all synthetic, her method, which is only able to prove analytic propositions, is destined to fail. It seems plausible to regard such a criticism as being primarily directed at the metaphilosophy of the dogmatist1, for what Kant is saying is that the dogmatist1’s philosophical project is ill-conceived, given certain views on method that are essential to the latter. Of course, this criticism also applies to the actual method followed by the dogmatist1, as long as her actual method coincides with her views. By contrast, Kant’s critique of dogmatism2 seems only to concern the second dimension of Kant’s critical rejection of dogmatism. Recall that Kant’s point against the dogmatist2 is that she makes illegitimate use of synthetic a priori principles. We know that Kant explains this illegitimate use by saying that the dogmatist2 assumes synthetic a priori principles without evaluating the conditions and scope of their validity. It would be extremely odd to say that this diagnosis and criticism operate at the metaphilosophical level. Doing so would entail ascribing to the dogmatist2 both an understanding of what synthetic a priori judgements are and, simultaneously, the view that assuming such principles does not require justification. Kant certainly believes that there are schools of thought that argue for the uncritical assumption of certain principles that, for him, are synthetic a priori. This is what he takes the Scottish common sense philosophers to be doing, for example (cf. 4:258–9). He does not think that these philosophers recognize these principles for what they are, however; that is, they do not have a clear notion of what a synthetic a priori judgement is.12 Therefore, it seems much more plausible to view Kant’s criticism of dogmatism2 as being directed at the actual methodology used by the dogmatist2, a methodology that is in part explained by her neglect of the problem of synthetic a priori judgements. Introducing this distinction between a metaphilosophical and a methodological level in Kant’s critique of dogmatism is extremely helpful for understanding Kant’s criticism of Wolff as a dogmatist. As Lanier Anderson has shown, Kant characterizes Wolff’s metaphysical project as an attempt to make every metaphysical truth a conceptual truth. It therefore seems appropriate to read Kant’s criticisms of Wolff’s dogmatism1 as being primarily directed at his metaphilosophical views. Of course, these criticisms 12

h

d i

In the Prolegomena, Kant suggests that while Hume partially recognized the problem of synthetic a priori judgements (see 4:260–1), this is what led him to sceptical conclusions. As far as common sense philosophers are concerned, Kant complains that in their response to Hume’s doubt they ignored the central point of Hume’s investigation, which aimed to determine whether and how a synthetic proposition can be known a priori (4:258–9).

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

246

also apply to Wolff’s actual methodology, as long as the latter agrees with the project that Wolff consciously pursued. By contrast, Kant’s remarks on Wolff’s dogmatism2 are only directed at his actual argumentative procedure. One of the advantages of approaching Kant’s critique of Wolff’s dogmatism from this angle is that Kant’s objection to Wolff’s metaphilosophy contributes to explaining why Wolff fails to provide a warrant for the synthetic a priori principles he uses in his arguments. More precisely, Wolff’s conscious attempt to build a system of analytic metaphysical truths underlies a tendency to treat every a priori truth as an analytic truth. It is this tendency that explains why Wolff does not provide an adequate grounding for synthetic a priori judgements. He simply takes such principles to be analytic and not in need of special justification. A diagnosis of this kind is suggested by the passage at A5–6/B9–10 quoted above. Of course, Kant does not explicitly mention Wolff there, but it is striking how well this diagnosis fits with other remarks that Kant makes and with the fact that Kant considers Wolff both a dogmatist1 and a dogmatist2.

4  The Dogmatic Procedure of Reason Kant’s characterizations of dogmatism are all negative. In the Preface to the B-edition of the Critique of Pure Reason, however, he draws a distinction between dogmatism, which we must avoid, and the ‘dogmatic procedure of reason’, which we must follow after having completed the critical investigation of our capacity for synthetic a priori judgement (Bxxxv). Kant picks Wolff as a positive model for this dogmatic method (Bxxxvi). So, while we should not follow Wolff’s metaphilosophy or method so long as it leads to dogmatism1 and dogmatism2, there is still something good about his argumentative procedure that we might wish to keep and that is ‘dogmatic’ in a certain sense. These considerations further complicate Kant’s account of Wolff’s dogmatism and introduce new questions. What does the legitimate dogmatic procedure of reason amount to? How is it related to Wolff’s dogmatism1 and dogmatism2? Kant introduces this dogmatic procedure in direct opposition to dogmatism2. I have already quoted part of the relevant passage in my discussion above, but let me quote it here again: Criticism is not opposed to the dogmatic procedure of reason in its pure cognition as science (for science must always be dogmatic, i.e., it must prove its conclusions strictly a priori from secure principles); rather, it is opposed only to dogmatism, i.e., to the presumption of getting on solely with pure cognition from (philosophical) concepts according to principles,

h

d i

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Wolff and Dogmatism

247

which reason has been using for a long time without first inquiring in what way and by what right it has obtained them. Dogmatism is therefore the dogmatic procedure of pure reason, without an antecedent critique of its own capacity. […] [C]riticism is the preparatory activity necessary for the advancement of metaphysics as a well-grounded science, which must necessarily be dogmatic, carried out systematically in accordance with the strictest requirement, hence according to scholastic rigour (and not in a popular way); for this requirement is one that it may not neglect, since it undertakes to carry out its business wholly a priori and thus to the full satisfaction of speculative reason. (Bxxxv–xxxvi)

This passage focuses on what distinguishes the dogmatist2 from the metaphysician who legitimately follows a dogmatic method in metaphysics. We already know that the dogmatist2’s error consists in uncritically assuming synthetic a priori judgements. But what does it mean to say that the dogmatist2 can proceed dogmatically once this error is corrected? It is at this point that Kant presents Wolff as an example to be emulated due to the ‘spirit of well-groundedness’ he introduced in Germany. This spirit is realized in the ‘lawful ascertainment of the principles, the clear determination of concepts, the attempt at strictness in the proofs, and the prevention of audacious leaps in inferences’ (Bxxxvi, translation altered). At least some of these characteristics of Wolff’s dogmatic procedure point towards Wolff’s deductive model of science, which is based on analytic definitions of concepts and on syllogisms (cf. Shabel 2006; Gava 2018a). Therefore, when Kant says that we must proceed dogmatically once synthetic a priori principles have been secured by the critique, he seems to mean that each of our metaphysical claims must be deductively derived from such principles. But how does this work, exactly? Unfortunately, Kant does not provide much detail in the passage under discussion. A passage from the Metaphysical Foundations of Natural Science, however, may be of greater assistance: [The metaphysics of nature] must always contain solely principles that are not empirical […], but it can still either: first, treat the laws that make possible the concept of a nature in general, even without relation to any determinate object of experience, and thus undetermined with respect to the nature of this or that thing in the sensible world, in which case it is the transcendental part of the metaphysics of nature; or second, concern itself with a particular nature of this or that kind of thing, for which an empirical concept is given, but still in such a manner that, outside of what lies in this concept, no other empirical principle is used for its cognition (for example, it takes the empirical concept of matter or of a thinking being as its basis, and it seeks that sphere of cognition of which reason is capable a priori

h

d i

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

248

concerning these objects), and here such a science must still always be called a metaphysics of nature, namely, of corporeal or of thinking nature. However, [in this second case] it is then not a general, but a special metaphysical natural science (physics or psychology), in which the above transcendental principles are applied to the two species of objects of our senses. (4:469–70; see also 5:181)13

In this passage, Kant introduces a distinction between transcendental and metaphysical principles, where the former belong to transcendental philosophy and the latter to the special metaphysics of nature. Transcendental principles do not take as given anything empirical in the concept of an object. Rather, they consider the object only as an object of possible experience in general that is thought by means of the categories. In doing so, they use synthetic a priori principles vindicated by transcendental philosophy to form judgements about it. Metaphysical principles, by ­contrast, do not consider objects only as objects of possible experience in general and instead attribute further characteristics to them that depend on a partly empirical concept, like the concept of ‘matter’. Once this ‘thicker’ concept of the object is introduced, we obtain metaphysical principles by specifying the synthetic a priori principles that are vindicated by transcendental philosophy for the object so defined. In doing so, however, we do not introduce completely new synthetic a priori principles. Rather, we assume the synthetic a priori principles vindicated by transcendental philosophy and deductively derive further principles from them when a richer concept of the object is assumed.14 The dogmatic procedure of reason must therefore be understood as a method for deriving truths that belong to the special metaphysics of nature from principles that instead belong to transcendental philosophy.15 How does this dogmatic procedure differ from Wolff’s dogmatism1 and dogmatism2? I have suggested that dogmatism1 applies primarily to Wolff’s metaphilosophy. In this respect, when we legitimately follow the dogmatic procedure of reason, we do not believe we are (or claim to be) establishing 13

14 15

h

d i

0 0

See also 4:473–77, where Kant presents the task of the Metaphysical Foundations of Natural Science as that of specifying the functions of the pure concepts of the understanding through the concept of matter. This paragraph re-elaborates materials presented in Gava (2018a). If we consider Kant’s position in the Metaphysical Foundations of Natural Science (4:471), it seems that devising a special metaphysics of thinking nature is actually impossible (see Pollok 2001: 94–5), which suggests that the only possible derivation concerns the special metaphysics of corporeal nature. On Kant’s account and critique of rational psychology, see Klemme (1996), Ameriks (2000) and Dyck (2014). Another issue is whether Kant thinks that empirical psychology can become a science. For conflicting accounts of the possibility of empirical psychology in Kant, see Sturm (2009), Frierson (2014) and Kraus (2018). See also the discussion of Frierson’s work in Kraus and Sturm (2017).

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Wolff and Dogmatism

249

metaphysical truths that are purely analytic (as Wolff did, on Kant’s view). We proceed syllogistically by the simple application of the principle of contradiction, but we consciously assume synthetic a priori principles and consequently hold the conclusions of our arguments to be equally synthetic. Dogmatism2 instead applies to Wolff’s actual argumentative procedure. Since, as I argued in Chapter 2 and Part III, one of the tasks of the critique of pure reason (as distinguished from transcendental philosophy) is to establish clear limits to the valid use of certain concepts that ground synthetic a priori judgements, in proceeding dogmatically after the critique has been accomplished, we have a grasp of the limited validity of such concepts and synthetic a priori principles. As a consequence, there is no risk of our using them to deductively derive conclusions that take us beyond their legitimate scope of application. That means that if the derivation that is at stake in the dogmatic procedure of reason is one that concerns the special metaphysics of nature, as I have argued here, it cannot extend to objects for which the critique of pure reason has prohibited any claim to cognition. Disciplines that formed the backbone of special metaphysics, such as rational cosmology and rational theology, are therefore ruled out as possible applications of the dogmatic procedure of reason because they concern objects that lie beyond the legitimate scope of application of synthetic a priori principles – that is, possible experience.

5  Transcendental Philosophy, the Critique of Pure Reason and Special Metaphysics The previous section provided a characterization of how the ‘dogmatic procedure’ of reason relates to prior inquiry into the validity and limits of certain synthetic a priori principles. In this last section, I wish to show that this characterization is compatible with my distinction between transcendental philosophy, as a doctrinal part of metaphysics, and the critique of pure reason, as that discipline within the Critique that achieves the latter’s aim as the doctrine of method of metaphysics. This compatibility becomes evident when one considers, first, that the ‘dogmatic procedure’ of reason is only applicable to special metaphysics, i.e. when it comes to deriving particular metaphysical principles from synthetic a priori principles that are established and justified elsewhere.16 16

h

d i

Although I have focused on the metaphysics of nature here, one might ask whether there is a similar relationship of derivation between the categorical imperative, as a synthetic a priori principle, and other moral principles introduced in the Metaphysics of Morals. This is not my concern here, however.

0 0

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

250

Second, recall that in my account both transcendental philosophy (at least the parts that are established in the first Critique) and the ‘negative’ side of the critique of pure reason focus on ‘root’ concepts of reason, which for Kant are concepts that lie at the basis of synthetic a priori judgements (see the Introduction to Part II). Accordingly, third, both transcendental philosophy and the critique of pure reason take into account the validity of synthetic a priori principles, but from different perspectives: the former aims to establish that some of these principles are valid, while the latter sets clear limits to the valid use of root concepts and the corresponding synthetic a priori principles. Given these remarks, the compatibility of the ‘dogmatic procedure’ of reason that Kant borrows from Wolff and my account of the relationship between transcendental philosophy and the critique of pure reason should now be clear. More precisely, when Kant submits that the use of the dogmatic procedure should be preceded by an investigation focused on the validity of synthetic a priori principles, what he has in mind is the analysis of principles that belong both to transcendental philosophy and to the critique of pure reason. Evidently, when it comes to deductively deriving further conclusions from these principles, it is equally important that they be established as valid (in transcendental philosophy) and that they have clear limits of application (identified by the critique).17 17

h

d i

0 0

Let me emphasize here that the derivation pursued by the dogmatic procedure can only be carried out in relation to constitutive concepts and the corresponding principles. Even though Kant ascribes some validity to the ideas as ‘root concepts’ with a valid regulative use, they cannot constitute the basis of a procedure of deductive derivation.

9 8 009

2 2 0 3

bli h d

li

b C

b id

i

i

chapter 9

Kant on Hume and Scepticism

Scepticism is the second approach to metaphysics that Kant considered an alternative to his ‘critical’ investigation (A856/B884). In this context, it is David Hume who plays the role of the paradigmatic representative of the sceptical viewpoint (see A764/B792; A855/B883). Furthermore, Kant describes Hume’s scepticism as playing a central role in the development of his critical position. As is well known, Kant claims in the Prolegomena that Hume awakened him from his dogmatic slumber (cf. 4:260), and he there presents a basic question of his critical philosophy – ‘How are synthetic judgements a priori possible?’ – as a generalized version of Hume’s inquiry regarding the origin of the concept of a cause. Kant accordingly equates the Critique of Pure Reason with ‘the elaboration of the Humean problem in its greatest possible amplification’ (4:261). This certainly provides prima facie evidence that: (1) Kant considered scepticism a position that had to be taken seriously (after all, it is one of the two approaches to philosophy to which criticism must provide an alternative), and (2) he understood scepticism ­primarily in terms of Humean scepticism regarding causal laws. Given Kant’s remarks, it is no surprise that much has been written on Kant’s answer to Hume’s scepticism about causality. Although there is vast agreement that Hume’s scepticism was important for Kant,1 interpretations diverge widely concerning how Kant understood the problem posed by Hume. There is disagreement regarding two questions in particular: (1) how did Kant interpret Hume’s doubts about causality? And (2) how close did Kant think Hume’s project was to his own? My purpose in this chapter is to use my distinction between transcendental philosophy and the 1

There are also scholars who downplay the role of Hume’s scepticism about causality in the development of Kant’s critical views, however. See for example Carl (1989) and G. Bird (2006b). In this respect, Eric Watkins’s position deserves particular attention. He shows that it is much more plausible to regard Kant’s account of causality as responding to the debate in eighteenth-century Germany rather than to Hume (Watkins 2005: Chs. 1–4). On the other hand, he also emphasizes how, in the Critique of Pure Reason, Kant presents his project of a critique of pure reason as sharing certain goals with Hume (Watkins 2005: 374–81).

251

h

d i

0 0

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

252

critique of pure reason to show that two apparently contrasting readings of Kant’s interpretation of Hume are in fact compatible. I begin, in the first three sections, by analysing three different readings of Kant’s understanding of Hume’s scepticism about causality. Hume’s scepticism is seen as posing a challenge to natural science and ordinary knowledge, as a problem that puts into question the possibility of general metaphysics, or as a ‘dialectical’ form of scepticism primarily directed against special metaphysics. While I reject the first reading as implausible, I suggest in Section 4 that the second and third readings are compatible. The second reading reconstructs Kant’s interpretation of Hume from the perspective of transcendental philosophy, while the third reading takes the standpoint of the critique of pure reason. My approach faces a problem, however. It has been argued that the third reading lacks a sufficient textual basis. More precisely, the texts by Hume to which Kant had access do not contain evidence of the ‘dialectical’ dimension of Humean scepticism. But if the third reading is implausible, there is no need to show that it is compatible with the second. In Section 5, I present attempts to find a textual basis in Hume that can support the third reading, and I consider objections to each of them. Finally, in Section 6, I argue that the third reading can be defended even in the absence of a textual basis. This can be done if one takes the perspective of Kant’s ‘history of pure reason’, namely a reconstruction of the history of philosophy that traces its sources to the nature of reason itself.2

1  A Challenge for Natural Science and Ordinary Knowledge? A very traditional way of understanding the challenge that Kant detected in Hume’s scepticism is the following: Hume’s view that our reasoning from causes to effects ultimately depends on a habit of the mind puts into question the certainty and necessary validity of the principle of causality,3 2

3

h

d i

0 0

I will focus on the Critique of Pure Reason and on the texts that appeared in the period between and following the publication of its two editions. Kant read Hume’s Enquiry in translation in the pre-critical period, and it is plausible to think that this text exerted an influence on him as early as the 1760s. Therefore, an important issue concerning the relationship between Kant and Hume is how Kant’s views developed between the pre-critical and the critical period (for a recent and welldocumented account of this development, see Chance 2012). Given that my aim here is to clarify Kant’s account of the method of philosophy within the Critique of Pure Reason, however, I will not consider Kant’s early remarks on Hume. Some interpreters have convincingly argued that Hume does not in fact deny that there are necessary causal connections in the world. Rather, he simply denies that we can determine whether these connections exist (see, for example, G. Strawson 2014). With this note, my aim in this chapter is not to determine the correct interpretation of Hume but to consider how Kant interpreted Hume.

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Hume and Scepticism

253

thus menacing one of the very foundations of our ordinary picture of the world and of natural science. In the Critique of Pure Reason, Kant attempts to save natural science and ordinary knowledge from the sceptical consequences of Hume’s position by showing that the concept of a cause, like other a priori concepts that are fundamental to our ordinary and scientific picture of the world, does not rest on habits of associations but is rather a condition of experience in the first place. According to this reading, because Kant has to provide an answer to Hume’s scepticism that safeguards natural science and ordinary knowledge, his argument, to be successful, must start from premises that do not beg the question posed by Hume.4 This account of Kant’s relation to Hume is well exemplified in the following passage from Peter Strawson: The central problem of classical empiricism was set by the assumption that experience really offers us nothing but separate and fleeting senseimpressions, images and feelings; and the problem was to show how, on this exiguous basis, we could supply a rational justification of our ordinary picture of the world as containing continuously and independently existing and interacting material things and persons. Hume, it is true, rejected the problem in this form, holding that such a justification was impossible, but also unnecessary, since the gaps found, and left, by reason were filled by the helpful fictions of the imagination. Between the views of Hume, the most sophisticated of the classical empiricists, and those of Kant, there is a subtle and interesting parallelism. But there is also a great gap. For Kant rejected the basic empiricist dogma which Hume never questioned. […] His rejection took the form […] of a proof that the minimal empiricist conception of experience was incoherent in isolation, that it made sense only within a larger framework which necessarily included the use and application in experience of concepts of an objective world. […] Any philosopher who invites, or challenges, us to justify our belief in the objective world by working outwards, as it were, the private data of individual consciousness thereby demonstrates his failure to have grasped the conditions of the possibility of experience in general.5 (Strawson 1966: 18–19; see also Robert Paul Wolff 1963 or, more recently, Arthur Melnick 2006)

Passages that support this interpretation of Kant’s reading of Hume are to be found, for example, in the Critique of Pure Reason (see A89–90/B122) and in the Critique of Practical Reason (see 5:53–4). If these passages are put in the context of other remarks, however, it seems implausible that Kant 4 5

h

d i

In fact, this reading is almost as old as the first Critique itself. See, for example, Reinhold (1791: 135–8), Schulze (1792: 131–80), Vaihinger (1881–1892: Vol. 1, 3–11) and Kemp Smith (1918: 61–4). It should be noted that in this passage Strawson views Hume’s scepticism as directed against both the existence of causal relationships and the numerical identity of non-continuously perceived objects.

0 0

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

254

considered our use of the concept of cause in ordinary knowledge and in natural science to be in need of philosophical justification. Accordingly, as he claims in the Discipline of Pure Reason: ‘[n]o critique of reason in empirical use was needed, since its principles were subjected to a continuous examination on the touchstone of experience; it was likewise unnecessary in mathematics, whose concepts must immediately be exhibited in concreto in pure intuition, through which anything unfounded and arbitrary instantly becomes obvious’ (A710–11/B738–9). For these reasons and others, the reading of Kant that portrays him as deliberately trying to ‘save’ ordinary knowledge and natural science from Hume’s doubt has been criticized by many interpreters (cf. Carl 1989; Engstrom 1994; Hatfield 2001; G. Bird 2006b; Watkins 2005: Ch. 6) and is now largely regarded as questionable.6 Accordingly, we do not need to scrutinize the motives for its rejection in further detail.

2  A Challenge for General Metaphysics? According to a second reading of Kant’s relationship to Hume, Kant views Hume not as posing a challenge to ordinary knowledge and natural science, but rather as putting into question the use of the principle of causality and, by extension, the use of other synthetic a priori propositions in the investigation of metaphysics. This means that, so the story goes, Hume had no interest in requiring a philosophical foundation for the application of the concept of cause in ordinary knowledge and natural science, an application that was instead left intact by Hume’s scepticism. By contrast, Kant read Hume’s scepticism as directed against the use of the concept of cause in metaphysics. At this point, it is important to introduce a distinction within metaphysics, since Kant distinguishes between general metaphysics, which is sometimes characterized as another name for ontology, and special metaphysics, which comprises rational psychology, cosmology and theology. Let us focus on general metaphysics first. Interpreters who read Kant as taking Hume’s doubt regarding causality as directed specifically against the possibility of general metaphysics have the following picture in mind:7 Kant interprets Hume’s scepticism 6

7

h

d i

0 0

Notice that Kant’s claims regarding the superfluousness of critique to mathematics and the empirical use of reason perfectly agree with the common sense conservatism that I attributed to Kant in the Introduction to Part II. Gary Hatfield (2001), Paul Guyer (2008: Intro., Ch. 1) and Michael Forster (2008: Ch. 5) have all emphasized in different ways that Kant viewed Hume’s scepticism on causality as mainly directed towards general metaphysics.

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Hume and Scepticism

255

as challenging the possibility of a properly metaphysical account of causality, that is, an account that is able to offer an a priori explanation of the necessity of causal relationships. In what sense can this reading consistently hold that Hume’s doubt about causality only raises problems for general metaphysics? Doesn’t the failure to provide a metaphysical a priori justification of causal laws imply that the use of causal principles in natural science and everyday life is equally unjustified? Not necessarily. Kant could simply have thought that the use of ‘causal talk’ in natural science and everyday life was not in need of a philosophical foundation.8 In metaphysics, by contrast, where an a priori account of causality is required,9 failing to explain how the concept of cause can legitimately be used in synthetic a priori judgements would be equal to admitting that a metaphysical account of that concept is impossible. In other words, failure to provide a metaphysical explanation of causality would prove an incapacity on the side of the metaphysician, not the lack of validity of the law of causality. Evidence that Kant read Hume’s doubt about causality as being primarily directed at the a priori use of this concept in general metaphysics is to be found, for example, in the B-version of the Introduction of the first Critique, where Kant claims that Hume, with his account of causality, purported to prove that ‘everything that we call metaphysics would come down to a mere delusion of an alleged insight of reason into that which has in fact merely been borrowed from experience and from habit has taken on the appearance of necessity’ (B20, my emphasis).

3  A Challenge for Special Metaphysics? A third possible reading of Kant’s interpretation of Hume has it that Kant took Hume’s doubt regarding causality to primarily pose a challenge to special metaphysics. This is suggested, for example, in a passage from the 8

9

h

d i

Here, one might object that Kant does indeed think that a philosophical foundation for ‘causal talk’ in natural science and everyday life is in order. His account of the regulative use of principles of the systematicity of nature (where these principles justify the search for particular empirical causal laws in nature) is designed to legitimate this ‘causal talk’ when it comes to regarding particular empirical causal laws as displaying a necessary connection. In response to this objection, I do not think that reference to regulative principles of reason does any justificatory work when we consider particular empirical causal laws that are already part of our knowledge set. What the regulative principles legitimize is regarding nature as systematic in our search for new particular causal laws that we do not yet know. I thank Achim Vesper for this objection. Recall that for Kant, the metaphysical foundation of natural science is strictly speaking part of ­metaphysics, not natural science.

0 0

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

256

Prolegomena, where Kant criticizes the way in which Reid, Oswald, Beattie and Priestley responded to Hume’s scepticism: One cannot, without feeling a certain pain, behold how utterly and completely his opponents, Reid, Oswald, Beattie, and finally Priestley, missed the point of his problem, and misjudged his hints for improvement – constantly taking for granted just what he doubted, and, conversely, proving with vehemence and, more often than not, with great insolence exactly what it had never entered his mind to doubt – so that everything remained in its old condition, as if nothing had happened. The question was not, whether the concept of cause is right, useful, and, with respect to all cognition of nature, indispensable, for this Hume had never put in doubt; it was rather whether it is thought through reason a priori, and in this way has an inner truth independent of all experience, and therefore also a much more widely extended use which is not limited merely to objects of experience: regarding this Hume awaited enlightenment. (4:258–9)

There are at least three issues that need to be pointed out in this passage. First, Kant complains that Reid, Oswald, Beattie and Priestley failed to understand that Hume did not want to challenge the ordinary use of the concept of cause. Accordingly, Kant claims that Hume never doubted that ‘the concept of cause is right, useful, and, with respect to all cognition of nature, indispensable’.10 Second, Kant explicitly states that Hume’s doubt was instead directed towards the a priori application of the concept of cause independently of all experience. Both points agree with what I have suggested in the previous section. Third, however, the main application of the concept of cause that Hume had in mind, according to Kant, is one ‘which is not limited merely to objects of experience’ and in which the concept has ‘a much more widely extended use’. This ‘extended use’ points towards the employment of the concept of cause in special metaphysics. Interpreters who propose this third interpretation of Kant’s take on Hume claim that for Kant, Hume was preoccupied with the endless conflicts in special metaphysics, where the concept of cause was used, for example, to prove the existence of God or to investigate the nature of the world as a whole. Hume’s doubts regarding the concept of cause were thus generated by the seemingly insoluble controversies into which philosophers interested in the latter problems never ceased to fall. Moreover, this doubt was not only a reaction to these controversies but also an attempt to terminate them, since the claim that our reason has no a priori insight into the nature of causality was meant to put a hold on this sort of investigation. According to this approach, Hume’s scepticism had a dialectical or 10

h

d i

0 0

Clearly, this is evidence against the first reading discussed in Section 1.

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Hume and Scepticism

257

Pyrrhonian character. Its essential motivation was an attempt to put a stop to endless controversies in metaphysics, where all parties seem to advance perfectly rational arguments.11 There are numerous passages that support this approach to Kant’s views on Hume. In the Discipline of Pure Reason, Kant presents scepticism as a reaction to dogmatism and as a purported ‘cure for dogmatic self-conceit’, a cure that tries to ‘end the conflict of reason with itself ’ by means of an admission of ignorance (A757/B785; cf. A758–69/B786–97). In this context, Hume is called ‘the most ingenious of all sceptics’ (A764/B792), and it is said that his sceptical attacks are chiefly directed against the ‘dialectical pretensions of reason’ (A768/B796). In a similar way, in the Prolegomena, Kant states that ‘[s]cepticism originally arose from metaphysics and its unpoliced dialectic’ (4:351; see also 4:258n; 5:13). But if Hume thought that metaphysical controversies were based on a problematic application of the concept of cause, what was his strategy for putting an end to this application? His strategy was to reveal (rightly, according to Kant) that judgements regarding the necessity of the effect could not be based on an analysis of the concept of cause (see 4:257–8). He realized that if a priori judgements based on the concept of cause were to be valid, they could not be analytic judgements but had to be synthetic. In a vague way, Hume grasped the notion of a synthetic a priori judgement, but he held the latter to always be due to a deception of our imagination and thus always invalid. Therefore, Hume’s strategy was to show that the principle of cause, if it had to be a priori and necessary, had to be synthetic a priori, which was impossible (see A764–5/B792–3).

4  Hume’s Challenge and the Distinction between Transcendental Philosophy and the Critique of Pure Reason Which of these readings of Kant’s interpretation of Hume is correct? I have already suggested that there are various reasons to regard the first reading as inadequate. What about the second and the third readings? In this section, I wish to show that they are compatible, and that Kant was able to consistently depict Hume according to both readings at the same time, but from two different perspectives, namely transcendental philosophy and the critique of pure reason, respectively.12 11 12

h

d i

Versions of this reading have been put forward by Kuehn (1983), Kreimendahl (1990), Ertl (2002), Stern (2006) and Chance (2012). Two interpreters who seem to regard the second and third readings as incompatible are Guyer (2008) and Forster (2008), who distinguish between two forms of scepticism to which Kant

0 0

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

258

Let me begin with the second reading. Recall that it submits that Kant takes Hume to pose a challenge to the possibility of general metaphysics. This means that by doubting that we can have a priori cognition of the validity and necessity of the principle of causality, Hume puts into question the possibility of a metaphysical treatment of that principle (and of other synthetic a priori principles). Now notice, first, that Kant sometimes equates transcendental philosophy with general metaphysics or ontology (see A845/B873).13 Second, the chief purpose of transcendental philosophy is to investigate ‘root’ concepts and to show that they are valid, which implies that synthetic a priori judgements based on them are also valid.14 Accordingly, it should be clear why characterizing Hume as challenging the possibility of an a priori account of causality involves depicting him as a direct antagonist of Kant’s project in transcendental philosophy. Therefore, I submit that the second reading simply describes Kant’s ­interpretation of Hume from the perspective of this discipline. Let us now move to the third reading. According to this interpretation, Kant viewed Hume’s doubt regarding causality as having been devised to put a stop to the endless disputes in special metaphysics, where that concept was used in arguments that led to seemingly undecidable controversies. Clearly, this interpretation does not depict Hume as pursuing a project in direct opposition to Kant’s.15 Since Kant also wanted to put a stop to the endless disputes in special metaphysics, Hume and Kant can be pictured as having similar projects, which they pursued using two different strategies. While Hume raised doubts about all synthetic a priori judgements, Kant tried to set limits to their validity. I submit that this way of characterizing Hume’s doubt captures its relevance from the perspective of the critique of pure reason. Recall that in my account, the critique has the chief aim of showing that metaphysics as a whole can achieve architectonic

13

14 15

h

d i

0 0

responded. The first is Hume’s scepticism, which poses a challenge to general metaphysics. The second is a form of Pyrrhonian scepticism generated by the dialectical conflict of reason. Since they both contrast this latter form of scepticism with Hume’s, it is clear that neither Guyer nor Forster would agree that Kant could have attributed a dialectical dimension to Hume’s scepticism. Of course, this equation does not mean that Kant took transcendental philosophy to be fundamentally the same as traditional ontology. Rather, Kant might have taken transcendental philosophy to take the place of traditional ontology as general metaphysics. Here, I am only focusing on root concepts that are constitutive. As we saw in Chapter 4, ideas of reason and related principles also have some validity, but not in the same sense as constitutive root concepts. Scholars who see Kant as attributing an anticipation of important features of the critique of pure reason to Hume include Kuehn (1983), Watkins (2005: 374–81), Waxman (2005), Chance (2011) and De Boer (2019). Still another view is defended by Wolfgang Carl (1989: 151–3), who claims that Kant’s characterization of Hume evolved dramatically between the first and the second editions of the Critique. While in the first edition Kant depicted Hume as an opponent, in the second he emphasized his connection to his critical project.

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Hume and Scepticism

259

unity. Recall, also, that this involves showing that metaphysics is at least capable of systematic coherence. Overcoming reason’s ‘conflict with itself’ is essential to achieving this coherence. Therefore, the third reading can plausibly be described as expressing Kant’s take on Hume from the perspective of the critique of pure reason. With these considerations in mind, I submit that the second and third readings of Kant’s interpretation of Hume are compatible. While the former emphasizes the relevance of Hume’s doubt to transcendental philosophy (and its aim of proving the validity of root concepts and synthetic a priori judgements), the second reading focuses on the relationship between Hume’s project and the project of a critique of pure reason (where the aim of putting a stop to disputes in special metaphysics is key).

5  Hume and Kant’s Slumber While the approach sketched in the previous section is plausible, there is a possible objection that I need to address. In particular, certain scholars have offered reasons for resisting the third reading of Kant’s interpretation of Hume. They have argued that because the texts by Hume to which Kant had access do not contain evidence of the ‘dialectical’ dimension of Humean scepticism, there is no sufficient textual basis for the third reading. Attempts to find textual support for a ‘dialectical’ reading of Hume’s scepticism in the texts available to Kant began in connection to two seemingly conflicting statements that Kant makes regarding the cause of his ‘awakening’ from his ‘dogmatic slumber’. As is well known, Kant remarks in the Prolegomena that ‘the remembrance of David Hume was the very thing that many years ago first interrupted my dogmatic slumber and gave a completely different direction to my researches in the field of speculative philosophy’ (4:261). This isn’t the only story that Kant tells about the reasons for his awakening, however. According to this second narrative, Kant contends that ‘[i]t was not the investigation of the existence of God, immortality, and so on, but rather the antinomy of pure reason – “The world has a beginning; it has no beginning, and so on, right up to the 4th: There is freedom in man, vs. there is no freedom, only the necessity of nature” – that is what first aroused me from my dogmatic slumber and drove me to the critique of reason itself, in order to resolve the scandal of ostensible contradiction of reason with itself’ (12:257–8). Since the first story appears in the Prolegomena, which was published in 1783, while the second account is found in a 1798 letter to Garve, it is tempting to assume that Kant simply changed his mind. In another passage in the Prolegomena, however, Kant attributes the awakening, this period in

h

d i

0 0

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

260

philosophy, to the cosmological ideas (that is, the ideas responsible for the antinomy of reason) (4:338). It thus seems implausible to simply attribute the two views to two different periods of Kant’s life. This fact has led scholars such as Thomas Kuehn (1983), Lothar Kreimendahl (1990; see also Gawlick and Kreimendahl 1987: Ch. 9) and, more recently, Wolfgang Ertl (2002) to suggest that the two accounts of Kant’s awakening must refer to a single cause, for otherwise we would be forced to see Kant as having voiced two inconsistent views in the same text (that is, the Prolegomena).16 Accordingly, they maintain that reading Hume’s Treatise gave Kant the fundamental insight that led him to discover the antinomy. A German translation of the Conclusion of Book I of the Treatise was published by Kant’s friend Johann Georg Hamann in July 1771 in two issues of the Königsbergische gelehrte und politische Zeitung, with the title ‘Nachtgedanken eines Zweiflers’. Kuehn, Kreimendahl and Ertl all agree that the inspiration for Kant’s mature notion of an antinomy came from reading this translation, even though they disagree on precisely when this reading took place.17 They locate the source of the inspiration in passages such as the following: No wonder a principle so inconstant and fallacious [that is, imagination] shou’d lead us into errors, when implicitly follow’d (as it must be) in all its variations. ’Tis this principle, which makes us reason from causes and effects; and ’tis the same principle, which convinces us of the continu’d existence of external objects, when absent from the senses. But tho’ these two operations be equally natural and necessary in the human mind, yet in some circumstances they are directly contrary, nor is it possible for us to reason justly and regularly from causes and effects, and at the same time believe the continu’d existence of matter. How then shall we adjust those principles together? Which of them shall we prefer? Or in case we prefer neither of them, but successively assent to both, as is usual among philosophers, with what confidence can we afterwards usurp that glorious title, when we thus knowingly embrace a manifest contradiction? (T:1.4.7.4)18 16 17

18

h

d i

0 0

Alternatively, scholars have accounted for these seemingly contrasting claims by identifying two different steps of Kant’s awakening. See, for example, Forster (2008: Chs. 3–5). In particular, Kreimendahl (1990; see also Gawlick and Kreimendahl 1987: Ch. 9 for a slightly different dating) and Ertl (2002) date Kant’s reading of Hamann’s translation to a period before the latter was published, in an attempt to make Kant’s claims concerning the role of Hume and the antinomies in his awakening coherent with his dating of the beginning of the critical project, a dating that we find, for example, in a letter to Mendelssohn from 1783 (cf. 10:345). In contrast to this approach, Karin de Boer (2019) has recently claimed that Kant’s reading of Hume was based on the Enquiry and that Kant took the analysis of causality that we find there to be relevant primarily because it undermines proofs of the existence of God. In other words, it is on the Enquiry, not the Treatise, that Kant based his ‘dialectical’ reading of Hume. Quotations from Hume (2007) follow the standard method: T followed by book, part, section and paragraph number(s).

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Hume and Scepticism

261

Here we find Hume suggesting that imagination leads us both to reason from cause to effect and to believe in the continued existence of external objects. These two operations of the imagination sometimes end up in opposition, however, supporting contradictory propositions. In these cases, we have a single principle that supports two contradictory views, and we are not in a position to choose between the two. Hume also suggests that philosophers are often responsible for uncritically embracing contradictory propositions of this sort. Moreover, a few lines later, he hints at a use of the concept of cause that bears certain similarities to Kant’s account of the regress of reason towards the unconditioned in the third antinomy: ‘Nothing is more curiously enquir’d after by the mind of man, than the causes of every phænomenon; nor are we content with knowing the immediate causes, but push on our enquiries, till we arrive at the original and ultimate principle’ (T:1.4.7.5). These passages have been offered as evidence that Hume’s Treatise was a fundamental inspiration for Kant’s concept of antinomy in the investigations, which eventually resulted in the Critique of Pure Reason. If this interpretation is correct, it provides a straightforward explanation for why Kant connected Hume’s scepticism to the dialectical conflict of reason. This approach has been strongly criticized, however. In particular, it has been claimed that there is not enough evidence to support the thesis that Kant took his concept of antinomy from Hume. Accordingly, Wolfgang Carl (1988: 213) has argued, against both Kuehn (1983) and Gawlick and Kreimendahl (1987), that the concept of antinomy does not even appear in the Conclusion of Book I of the Treatise. Moreover, the ‘manifest contradiction’ that Hume considers at T:1.4.7.4 appears to be very different from the conflict of reason with itself that Kant describes in the antinomies. On the one hand, we have a conflict between different empirical principles that sometimes collide when they orient us in experience. On the other, we have a conflict that arises from reason of necessity and that strictly determines the nature of certain metaphysical inquiries. In a related way, Reinhard Brandt (1992: 104) has argued, against Kuehn (1983) and Kreimendahl (1990), that Kant does not mention Hume together with the problem of the antinomy.19 Moreover, Brandt continues, if one pays attention to the way in which Kant discusses Hume, it is more plausible to believe that the Enquiry was the main source of Hume’s influence on 19

h

d i

In truth, this claim does not seem to be accurate. In the Critique of Practical Reason, Kant does discuss Hume in connection with the mathematical antinomies (see 5:13–14).

0 0

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

262

Kant, which would rule out the possibility of using Hume’s Treatise as an inspiration for Kant’s notion of antinomy.20 While I think that Carl and Brandt are correct to say that there is ultimately no textual basis in Hume for attributing a ‘dialectical’ interpretation of Humean scepticism to Kant, this does not force us to jump to the conclusion that Kant could not have interpreted Hume in that way. In the following section, I will suggest that the correct way to view the connection that Kant draws between Hume’s scepticism and the dialectic of pure reason is from the standpoint of what he calls the history of pure reason.

6  Scepticism and the History of Reason I submit that the Transcendental Doctrine of Method, and in particular its chapter on the History of Pure Reason, provides a useful tool for interpreting Kant’s remarks on Hume. It makes clear why Kant was able to attribute a particular view to Hume while recognizing that Hume did not explicitly defend that view in his writings. But what is the history of pure reason for Kant? In short, he uses this label to describe the history of philosophy. In the Critique of Pure Reason, he only provides a ‘cursory glance’ (A852/B880) of his take on this history. At the beginning of the chapter dedicated to this issue, he only maintains that the ‘cursory glance’ he is about to cast will be given ‘from a merely transcendental point of view, namely that of the nature of pure reason, on the whole of its labours hitherto’ (A852/B880). He then says that he ‘will not now distinguish the times in which this or that alteration of metaphysics occurred, but will present in a cursory outline only the difference of the ideas which occasioned the chief revolutions’ (A853/B881).21 It would be 20

21

h

d i

0 0

As a partial response to Brandt, De Boer (2019) argues that it is not necessary to refer to the Treatise in order to argue that there is a textual basis in Hume for a ‘dialectical’ reading of his scepticism. This textual basis can in fact be found in the Enquiry. In her own reading of Kant’s interpretation of Hume, which maintains that Kant read Hume’s Enquiry as primarily challenging proofs of the existence of God, the concept of an antinomy plays only a secondary role. He divides his analysis of these ideas by taking three different points of view with regard to the object and origin of pure cognitions of reason and the method of philosophy. What we find in this tripartite analysis is a classification of historical figures within very broad historical categories. Accordingly, when the object of pure cognitions is taken into account, Kant mentions Epicurus as the foremost representative of the ‘sensualist philosophers’ and Plato as that of the ‘intellectual philosophers’ (A853/B881). By contrast, Aristotle, Locke and Epicurus are all categorized as ‘empiricists’ when the origin of the cognitions of pure reason is considered. From this perspective, Plato and Leibniz are instead called ‘noologists’ (A854/B882). Finally, when Kant takes into consideration the method of philosophy, he first distinguishes between the naturalistic and the scientific method (cf. A855/B883). Under the former he understands something very different from what is today called naturalism. A naturalist, for Kant, is somebody who is content to base her claims on common sense.

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Hume and Scepticism

263

wrong to think that Kant’s decision to disregard the details of the views held by particular philosophers at specific times is due to the brevity of the chapter. Rather, it points to an essential feature of Kant’s approach to the history of philosophy. This becomes clear if we analyse some of Kant’s remarks on the history of philosophy in his unpublished writings. In some drafts of Progress, he clarifies the way in which the history of philosophy should be drawn from the nature of reason. He notes that ‘[a] philosophical history of philosophy is itself possible, not historically or empirically, but rationally, i.e., a priori. For although it establishes facts of reason, it does not borrow them from historical narrative, but draws them from the nature of human reason, as philosophical archaeology’ (20:341). He continues: ‘[a] history of philosophy is of such a special kind, that nothing can be told therein of what has happened, without knowing beforehand what should have happened, and also what can happen. […] For it is the history, not of the opinions which have chanced to arise here or there, but of reason developing itself from concepts’ (20:343). Four ideas need to be emphasized in these remarks. First, Kant views the history of philosophy as a ‘history of reason developing itself from concepts’. Philosophical systems existing in history represent the ­various attempts of reason to gain self-knowledge. This is achieved when the c­ognitions we can obtain on the basis of pure reason are appropriately laid out and ordered in a system.22 Second, given that these attempts at ­self-­knowledge are ­expressions of reason’s own efforts, they are best ­understood as a ­manifestation of reason’s nature and structure. This is why the history of philosophy should be told by conveying ‘what should have happened, and also what can happen’. In other words, a history of reason should show that past philosophical theories are not totally ­contingent ­historical ­outputs but have their necessary ground in reason. Moreover, a clear grasp of r­ eason’s structure gives us access to what its p ­ ossible ­expressions might be, even though the latter have not yet appeared in history. Third, this implies

22

h

d i

By contrast, the scientific method is characterized by its systematicity. It is at this point that Kant distinguishes between dogmatism and scepticism as two alternatives within the scientific method. As we have already seen, he lists Wolff and Hume as the two most representative followers of these methods, to which only the critical method is a genuine alternative (cf. A855/B883). The superficiality with which Kant groups philosophers from very different traditions and eras under very broad historical categories is something that any historian of philosophy will certainly find irritating. However, the question here is not whether Kant was a good historian of philosophy (for illuminating considerations in this respect, see Ferrarin 2015: 76–80) but what his approach to the history of philosophy can teach us about his interpretation of Hume’s scepticism. This system would constitute the correct ordering of metaphysical cognitions that provides architectonic unity to metaphysics (see Chapter 1).

0 0

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

264

that we cannot comprehend the history of philosophy if we limit ourselves to reconstructing philosophical theories in their succession. Accordingly, Kant notes that we should not borrow ‘facts of reason’ from a ‘historical narrative’.23 Fourth, Kant’s remarks indicate that the only perspective from which the history of philosophy can properly be told is that of a reason that has already gained self-knowledge. For if reason’s history needs to be clarified by locating its sources in the structure and nature of reason, only when we have grasped this nature and structure can we fully understand the history of philosophy. What does Kant’s account of the history of reason teach us about how we should approach his views on Hume? It tells us that the connection that Kant draws between Hume’s scepticism and reason’s conflict with itself does not need to rest on an explicit combination of these matters that Kant found in Hume’s own texts. Rather, Kant may have seen this connection as the best way to account for Hume’s scepticism from the point of view of the nature of reason. Taking this point of view, Kant would have been in a position to say that reason’s conflict with itself was the actual source of Hume’s scepticism, while at the same time recognizing that Hume did not himself draw this connection between his scepticism and the conflict in question in his texts. Therefore, if our goal is to maintain that Kant read Hume’s scepticism as a response to reason’s dialectic, to sustain this view we do not need to show that Kant found something similar to his notions of an antinomy or a conflict of reason directly in Hume. What we need to do is to provide evidence that Kant thought this was the correct way to interpret Hume’s position from the point of view of the history of reason. Evidence for this view can be found in a Reflexion that probably dates from the mid to late 1780s. In this context, Kant first defines dogmatism as ‘the manner of thinking that is attached to assertions without critique’ (Refl. 5645, 18:293–4). He then maintains that dogmatism is ‘[t]he most natural tendency of mankind’ (18:294). Scepticism, by contrast, ‘is a principle adopted to break with dogmatism, but not […] with the aim of introducing […] true conviction against it, but rather only in order to topple the persuasions of others’ (18:294). Kant then notes that the inclination to scepticism ‘is not natural but artificial and can only arise from displeasure with the usurpation of dogmatism’ (18:294). 23

h

d i

0 0

This need to refer to facts of reason even for a history of pure reason points towards the problematic nature of Kant’s account of the history of philosophy. See Yovel (1980: Epilogue) and Ferrarin (2015: 76–80).

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Hume and Scepticism

265

In a way that is similar to the Discipline of Pure Reason, Kant characterizes both dogmatism and scepticism as rooted in our nature. What is new in this Reflexion is how Kant describes the proper way to react to the dangers implied by dogmatism and scepticism. These dangers are, respectively, ‘that of rousing up a cloud of errors among a small number of truths and of bringing contempt upon the latter themselves because of their relation to the former’, and ‘the denial of our duty of always serving our reason and a laziness in this that is excused by its [that is, scepticism’s] plausible objections’ (18:294). Predictably, Kant thinks that the only possible strategy for avoiding these dangers is a critical investigation of reason. Here, however, he describes the latter as involving a twofold task: This danger can only be averted through the greatest critical diligence, on the empirical side in tracking down the sources of history and its derivation from us and on the rational side in tracking down the nature and the capacity of human reason in its speculative use in metaphysics as well as its practical use in morality, and in determining their boundaries, likewise their scope and the necessary principles of the latter. (Refl. 5645, 18:294–5)

While the ‘rational side’ of the critical task aligns with Kant’s usual description of the project of a critique of pure reason, the ‘empirical side’ introduces something new. Accordingly, a tracing back of philosophical theories to their sources in human reason, which is what Kant understands under a history of pure reason, is here described as an essential part of the critical investigation. What is more, the need for such a critical history of pure reason is emphasized in connection with the dangers presented by dogmatism and scepticism. Therefore, given that Kant portrays the history of pure reason as essential to his critical investigation, and given that the need for such a history is expressed in relation to dogmatism and scepticism, it is plausible to expect Kant to approach historical expressions of dogmatism and scepticism from the standpoint of the history of pure reason. Since, as we know, Kant viewed Hume as ‘perhaps the most ingenious of all sceptics’ (A764/B792), what reason is there to think that he wouldn’t have considered Hume’s position from the standpoint he thought was the most appropriate for the ‘empirical side’ of his critique of pure reason? More precisely, what reason is there to think that he wouldn’t have applied to Hume the interpretative strategy he envisioned for historical expressions of scepticism in general? I think that the only sensible way to answer these questions is to conclude that the correct way to approach Kant’s remarks on Hume and, more specifically, on the relationship between Hume’s scepticism and reason’s

h

d i

0 0

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Kant on Dogmatism and Scepticism

266

conflict with itself, is to see them as instantiations of what Kant calls the history of pure reason.24

7  Two Perspectives on Hume’s Scepticism If what I have argued in Section 6 is right, the reading of Kant’s take on Hume that submits that Kant interpreted Hume’s scepticism as having a ‘dialectical’ dimension can be defended even though it is impossible to find a textual basis in Hume that could have prompted Kant’s interpretation. Since this reading is plausible and can be defended, the question of its relation to the second reading presented in Section 2 arises once again. As I suggested in Section 4, the best way to account for how these readings are related is to say that they provide two different perspectives from which Kant regarded Hume’s scepticism. The former is the perspective of transcendental philosophy, which depicts Hume as an antagonist of Kant’s project who put into question the possibility of vindicating synthetic a priori judgements in general metaphysics. The latter is the perspective of the critique of pure reason, which regards Hume as a fellow traveller and as similarly interested in putting a stop to controversies in special metaphysics. 24

h

d i

0 0

One might worry that approaching Kant’s interpretation of Hume from the perspective of the history of reason makes his reading uninteresting. For Kant, the name ‘Hume’ may just have been a label for a position that the actual Hume had very little to do with. However, I do not think that claiming that Kant read Hume from the perspective of the history of reason means that he did not take Hume’s actual position seriously. Rather, he thought that the framework of the history of reason gave him the means to go deeper in his analysis of Hume’s view, in a way that enabled him to appreciate features of that view that were not evident on the surface. I thank Achim Vesper for raising this issue.

9 8 009

2 2 0 4

bli h d

li

b C

b id

i

i

Conclusion

In this book, I have argued that the Critique of Pure Reason is the doctrine of method of metaphysics. This does not mean that its main task is to provide a clarification of the procedures of argument that are adequate within metaphysics.1 Rather, the Critique determines that metaphysics can attain the status of a science because it is capable of achieving ‘architectonic unity’. What I want to point out is that this understanding of ‘critique’ is distinctive of the first Critique and is not shared by the second and third Critiques. Let me recall two of the consequences of my reading. First, my approach provides a straightforward way to distinguish between transcendental philosophy, as one part of metaphysics that is partly established in the Critique, and the critique of pure reason, as that discipline within the Critique that achieves the latter’s aim as the doctrine of method of metaphysics. While both disciplines are interested in assessing reason’s capacity for a priori cognition, they take a different perspective when evaluating it. What is distinctive about the critique of pure reason is not that it identifies and validates a priori concepts and principles of reason – this is what transcendental philosophy and metaphysics more broadly do. Rather, the critique considers how these concepts and principles relate to one another and can form a coherent whole. Its strategy for accomplishing this aim essentially rests on setting limits to the validity of such concepts and principles. Second, I explain the sense in which the Critique is relevant to the practical part of metaphysics. Clearly, it does not provide an analysis or foundation of a priori principles of practical reason. Rather, it assumes that there are such principles, without offering a justification for them. Given that these principles require a commitment to God and immortality, the Critique shows that we can coherently adopt that commitment without endangering the limits to our cognition set by the negative side of the critique. 1

Although it does so in the Discipline of Pure Reason.

267

h

d i

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

Conclusion

268

If this is right, it means that the Critique of Pure Reason is not tasked with identifying and justifying a priori principles of reason. It contains analyses that do that, but only insofar as it contains parts of transcendental philosophy or, as far as the practical part of metaphysics is concerned, presupposes that there are a priori principles of practical reason. Therefore, the distinctive feature of the critique of pure reason is that it considers the ‘architectonic plan’ of metaphysics2 in a way that shows that it can become a science. There is an important shift of emphasis when one considers how the second and third Critiques are characterized. In the preface to the Critique of Practical Reason, Kant famously submits that his investigation is not called the ‘Critique of Pure Practical Reason’ because it cannot presuppose but must establish ‘that there is pure practical reason’ (5:3). Presumably, this requires showing that there is an a priori principle of morality. A ­similar characterization of the tasks of the critique is provided in the ­preface to the Critique of the Power of Judgement. Kant motivates its need by ­writing that it must investigate whether there is an a priori principle of the power of judgement (5:168–9).3 In other words, the main task of the critique becomes establishing that there are a priori principles belonging to ­different faculties. We can speculate on the reasons that motivated this shift in the characterization of critique. As far as the third Critique is concerned, one possibility might be the following: according to Kant, the a priori principle of purposiveness is not a principle that properly belongs to metaphysics. In the Introduction to the Critique of the Power of Judgement, Kant submits that while metaphysics is divided into two parts, because there are only two ‘domains’ where the legislation of reason is exerted, namely, the domain of nature and the domain of freedom (5:175–6), the critique of pure reason must be divided into three parts, because there are three different faculties that possess a priori principles (5:179). It is precisely the principle of purposiveness investigated in the third Critique that does not have a corresponding part in metaphysics. But this means that its identification and justification could only happen within the third Critique. A similar explanation is not available for the Critique of Practical Reason. Kant is explicit that a priori principles of practical reason have a place 2 3

h

d i

0 0

At A13/B27, Kant explicitly stresses that a chief task of the Critique is to ‘outline’ a ‘plan’ ‘architectonically’, even though the focus is on transcendental philosophy in particular. Curiously, in this context Kant limits the contribution of the first Critique to the identification of the a priori principles of the understanding, which completely neglects the investigations that the book dedicates to sensibility and reason as the faculty of inference. On the fluctuations within Kant’s representation of the critical project, see Ferrarin (2015: Ch. 3); Förster (2018: Part I).

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

Conclusion

269

in metaphysics. In both the Critique of Pure Reason and the Critique of the Power of Judgement, he emphasizes that a priori principles of morality are ‘metaphysical’ and not ‘transcendental’ because they are a priori but not pure, since they presuppose empirically given concepts (A15/B29; 5:181–2).4 More importantly, the Metaphysics of Morals evidently identifies metaphysical principles of right and virtue (6:205) and a fundamental principle of morals that lies at the basis of both (6:226).5 One can ask whether the categorical imperative identified in the second Critique (5:30) can be equated with any of these metaphysical principles. But even if these principles were different in the end, there would still be no reason to claim that the task of singling out the categorical imperative, as it is formulated in the second Critique, had to be attributed to the critique and not to a part of metaphysics. So it seems that the shift in Kant’s characterization of critique in the second Critique is more difficult to explain. Even though we cannot properly explain the shift in Kant’s account of critique in the second and third Critiques, what I want to emphasize is that describing the critique of pure reason as the doctrine of method of metaphysics, as I do, singles out a distinctive way of portraying its aims. It is to clarifying this understanding of critique, which is specific to the first Critique, that this book has been dedicated. 4

5

h

d i

Notice, however, that Kant’s explanation of why the principles of morality are not transcendental changes radically when one compares the A-version and the B-version of the passage from the first Critique. The question of the relationship between right and virtue in Kant is vividly debated in the literature. Here, I do not want to take sides and suggest that they have a common foundation. This is how Kant presents the issue, but it remains an open question whether he is successful in attaching the two parts of the metaphysics of morals to a unique root.

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

Bibliography

Allais, Lucy. 2015. Manifest Reality: Kant’s Idealism and His Realism. Oxford: Oxford University Press. Allais, Lucy. 2017. ‘Empirical Concepts, the Role of Intuition, and the Poverty of Conceptual Truth’. Studi Kantiani 30: 99–114. Allison, Henry E. 1986. ‘The Concept of Freedom in Kant’s “Semi-Critical” Ethics’. Archiv für Geschichte der Philosophie 68 (1): 96–115. Allison, Henry E. 2004. Kant’s Transcendental Idealism: An Interpretation and Defense. 2nd ed. New Haven, CT: Yale University Press. Allison, Henry E. 2015. Kant’s Transcendental Deduction: An Analytical–Historical Commentary. Oxford: Oxford University Press. Ameriks, Karl. 1978. ‘Kant’s Transcendental Deduction as a Regressive Argument’. Kant-Studien 69 (1–4): 273–87. Ameriks, Karl. 2000. Kant’s Theory of Mind: An Analysis of the Paralogisms of Pure Reason. 2nd ed. Oxford: Oxford University Press. Ameriks, Karl. 2003. Interpreting Kant’s Critiques. Oxford: Clarendon. Anderson, R. Lanier. 2015. The Poverty of Conceptual Truth: Kant’s Analytic/ Synthetic Distinction and the Limits of Metaphysics. Oxford: Oxford University Press. Anderson, R. Lanier. 2017. ‘On the Two-Step Interpretation of Kant’s Dialectical Strategy in Theoretical Philosophy: Replies to Allais, Hanna, and Motta’. Studi Kantiani 30: 129–51. Bacin, Stefano. 2002. ‘Sul rapporto tra riflessione e vita morale in Kant: Le Dottrine del metodo nella filosofia pratica’. Studi Kantiani 15: 65–91. Bacin, Stefano. 2010. ‘The Meaning of the “Critique of Practical Reason” for Moral Beings: The Doctrine of Method of Pure Practical Reason’. In Kant’s Critique of Practical Reason: A Critical Guide, edited by Andrews Reath and Jens Timmermann, 197–215. Cambridge: Cambridge University Press. Barale, Massimo. 1988. Kant e il metodo della filosofia. I. Sentire e intendere. Pisa: Edizioni ETS. Bird, Alexander. 2007. ‘Inference to the Only Explanation’. Philosophy and Phenomenological Research 74 (2): 424–32. Bird, Alexander. 2010. ‘Eliminative Abduction: Examples from Medicine’. Studies in History and Philosophy of Science Part A 41 (4): 345–52. 270

h

d i

0 0

9 8 009

2 2 0 6

bli h d

li

b C

b id

i

i

Bibliography

271

Bird, Graham. 2006a. ‘The Neglected Alternative: Trendelenburg, Fischer, and Kant’. In A Companion to Kant, edited by Graham Bird, 486–99. Oxford: Blackwell. Bird, Graham. 2006b. The Revolutionary Kant: A Commentary on the Critique of Pure Reason. 1st ed. Chicago: Open Court. Brandt, Reinhard. 1991. Die Urteilstafel: Kritik der reinen Vernunft A 67–76; B 92–101. Hamburg: Meiner. Brandt, Reinhard. 1992. ‘Review of Kant: Der Durchbruch von 1769, by Lothar Kreimendahl’. Kant-Studien 83 (1): 100–11. Bröcker, Walter. 1970. Kant über Metaphysik und Erfahrung. Frankfurt am Main: Klostermann. Caimi, Mario. 1995. ‘Über eine wenig beachtete Deduktion der Regulativen Ideen’. Kant-Studien 86 (3): 308–20. Caimi, Mario. 2000. ‘Einige Bemerkungen über die Metaphysische Deduktion in der Kritik Der Reinen Vernunft’. Kant-Studien 91 (3): 257–82. Callanan, John J. 2006. ‘Kant’s Transcendental Strategy’. Philosophical Quarterly 56 (224): 360–81. Callanan, John J. 2019. ‘Methodological Conservativism in Kant and Strawson’. British Journal for the History of Philosophy 27 (2): 422–42. Capozzi, Mirella. 2002. Kant e la logica. Vol. 1. Napoli: Bibliopolis. Capozzi, Mirella. 2013. ‘The Quantity of Judgments and the Categories of Quantity: A Problem in the Metaphysical Deduction’. In Kant und die Philosophie in weltbürgerlicher Absicht: Akten des XI. Internationalen KantKongresses, edited by Stefano Bacin, Alfredo Ferrarin, Claudio La Rocca and Margit Ruffing, 2:65–76. Berlin: De Gruyter. Capozzi, Mirella. 2018. ‘La Prima antinomia di Kant: questioni logiche’. Studi Kantiani 31: 11–42. Carboncini, Sonia, and Reinhard Finster. 1982. ‘Das Begriffspaar Kanon-Organon: Seine Bedeutung für die Entstehung der kritischen Philosophie Kants’. Archiv für Begriffsgeschichte 26 (1): 25–59. Carl, Wolfgang. 1988. ‘Review of Hume in der Deutschen Aufklärung. Umrisse einer Rezeptionsgeschichte, by Günter Gawlick and Lothar Kreimendahl’. Philosophische Rundschau 35 (3): 207–14. Carl, Wolfgang. 1989. Der schweigende Kant: Die Entwürfe zu einer Deduktion der Kategorien vor 1781. Göttingen: Vandenhoeck & Ruprecht. Carl, Wolfgang. 1992. Die Transzendentale Deduktion der Kategorien in der ersten Auflage der Kritik der reinen Vernunft: Ein Kommentar. Frankfurt am Main: Klostermann. Cassam, Quassim. 1987. ‘Transcendental Arguments, Transcendental Synthesis and Transcendental Idealism’. Philosophical Quarterly 37 (149): 355–78. Chance, Brian A. 2011. ‘Sensibilism, Psychologism, and Kant’s Debt to Hume’. Kantian Review 16 (3): 325–49. Chance, Brian A. 2012. ‘Scepticism and the Development of the Transcendental Dialectic’. British Journal for the History of Philosophy 20 (2): 311–31. Chance, Brian A. 2015. ‘Kant and the Discipline of Reason’. European Journal of Philosophy 23 (1): 87–110.

h

d i

0 0

9 8 009

2 2 0 6

bli h d

li

b C

b id

i

i

272

Bibliography

Chignell, Andrew. 2007. ‘Belief in Kant’. Philosophical Review 116 (3): 323–60. Chignell, Andrew. 2014. ‘Rational Hope, Possibility, and Divine Action’. In Kant’s Religion within the Boundaries of Mere Reason, edited by Gordon Michalson, 98–117. Cambridge: Cambridge University Press. Chignell, Andrew. 2017. ‘Knowledge, Discipline, System, Hope: The Fate of Metaphysics in the Doctrine of Method’. In Kant’s Critique of Pure Reason, edited by James R. O’Shea, 259–79. Cambridge: Cambridge University Press. Chignell, Andrew. Forthcoming. Knowledge and Belief in Kant: Making Room for Practical Metaphysics. Oxford: Oxford University Press. Conant, James. 2016. ‘Why Kant Is Not a Kantian’. Philosophical Topics 44 (1): 75–125. Conant, James. 2017. ‘Kant’s Critique of the Layer-Cake Conception of Human Mindedness in the B Deduction’. In Kant’s Critique of Pure Reason: A Critical Guide, edited by James R. O’Shea, 120–39. Cambridge: Cambridge University Press. de Boer, Karin. 2019. ‘Kant’s Response to Hume’s Critique of Pure Reason’. Archiv für Geschichte der Philosophie 101 (3): 376–406. de Boer, Karin. 2020. Kant’s Reform of Metaphysics: The Critique of Pure Reason Reconsidered. Cambridge: Cambridge University Press. de Jong, Willem R. 1995. ‘How Is Metaphysics as a Science Possible? Kant on the Distinction between Philosophical and Mathematical Method’. The Review of Metaphysics 49 (2): 235–74. Dyck, Corey W. 2014. Kant and Rational Psychology. Oxford: Oxford University Press. Engstrom, Stephen. 1994. ‘The Transcendental Deduction and Skepticism’. Journal of the History of Philosophy 32 (3): 359–80. Ertl, Wolfgang. 2002. ‘Hume’s Antinomy and Kant’s Critical Turn’. British Journal for the History of Philosophy 10 (4): 617–40. Falkenstein, Lorne. 1995. Kant’s Intuitionism: A Commentary on the Transcendental Aesthetic. Toronto: University of Toronto Press. Ferrarin, Alfredo. 1995. ‘Construction and Mathematical Schematism: Kant on the Exhibition of a Concept in Intuition’. Kant-Studien 86 (2): 131–74. Ferrarin, Alfredo. 2015. The Powers of Pure Reason: Kant and the Idea of Cosmic Philosophy. Chicago: University of Chicago Press. Ferrarin, Alfredo. 2019. ‘Method in Kant and Hegel’. British Journal for the History of Philosophy 27 (2): 255–70. Fonnesu, Luca. 2015. ‘Kant on Private Faith and Public Knowledge’. Rivista Di Filosofia 106: 361–90. Förster, Eckart. 2018. Die 25 Jahre der Philosophie. 3rd revised edition. Frankfurt am Main: Klostermann. Forster, Michael N. 2008. Kant and Skepticism. Princeton, NJ: Princeton University Press. Frierson, Patrick R. 2013. What Is the Human Being? Kant’s Questions. New York: Routledge.

h

d i

0 0

9 8 009

2 2 0 6

bli h d

li

b C

b id

i

i

Bibliography

273

Frierson, Patrick R. 2014. Kant’s Empirical Psychology. Cambridge: Cambridge University Press. Fugate, Courtney. 2019. ‘Kant’s World Concept of Philosophy and Cosmopolitanism’. Archiv für Geschichte der Philosophie 101 (4): 535–83. Gava, Gabriele. 2014. ‘Kant’s Definition of Science in the Architectonic of Pure Reason and the Essential Ends of Reason’. Kant-Studien 105 (3): 372–93. Gava, Gabriele. 2015. ‘Kant’s Synthetic and Analytic Method in the Critique of Pure Reason and the Distinction between Philosophical and Mathematical Syntheses’. European Journal of Philosophy 23 (3): 728–49. Gava, Gabriele. 2016. ‘The Fallibilism of Kant’s Architectonic’. In Pragmatism, Kant and Transcendental Philosophy, edited by Robert Stern and Gabriele Gava, 46–66. New York: Routledge. Gava, Gabriele. 2018a. ‘Kant, Wolff, and the Method of Philosophy’. Oxford Studies in Early Modern Philosophy 8: 271–303. Gava, Gabriele. 2018b. ‘Sind die regulativen Ideen ein doktrinaler Glaube? Über die Rechtfertigung des regulativen Gebrauchs der Ideen im Anhang zur transzendentalen Dialektik’. In Natur und Freiheit: Akten des XII. Internationalen Kant-Kongresses, edited by Violetta Waibel, Margit Ruffing, and David Wagner, 2:1207–16. Berlin, Boston: De Gruyter. Gava, Gabriele. 2019a. ‘C. I. Lewis, Kant, and the Reflective Method of Philosophy’. British Journal for the History of Philosophy 27 (2): 315–35. Gava, Gabriele. 2019b. ‘Kant and Crusius on Belief and Practical Justification’. Kantian Review 24 (1): 53–75. Gava, Gabriele.2019c. ‘Kant, the Third Antinomy and Transcendental Arguments’. Pacific Philosophical Quarterly 100 (2): 453–81. Gava, Gabriele, and Marcus Willaschek. Forthcoming. ‘The Doctrine of Method’. In The Kantian Mind, edited by Sorin Baiasu and Mark Timmons. New York: Routledge. Gawlick, Günter, and Lothar Kreimendahl. 1987. Hume in der Deutschen Aufklärung: Umrisse einer Rezeptionsgeschichte. Stuttgart: Fromann-Holzboog. Ginsborg, Hannah. 2008. ‘Was Kant a Nonconceptualist?’ Philosophical Studies 137 (1): 65–77. Goldberg, Nathaniel Jason. 2004. ‘Do Principles of Reason Have Objective but Indeterminate Validity?’ Kant Studien 95 (4): 405–25. Golob, Sacha. 2016. ‘Why the Transcendental Deduction Is Compatible with Nonconceptualism’. In Kantian Nonconceptualism, edited by Dennis Schulting, 27–52. London: Palgrave-Macmillan. Goy, Ina. 2007. Architektonik oder die Kunst der Systeme: Eine Untersuchung zur Systemphilosophie der Kritik der reinen Vernunft. Paderborn: mentis. Grier, Michelle. 2001. Kant’s Doctrine of Transcendental Illusion. Cambridge: Cambridge University Press. Grüne, Stefanie. 2009. Blinde Anschauung: Die Rolle von Begriffen in Kants Theorie sinnlicher Synthesis. Frankfurt am Main: Klostermann. Guyer, Paul. 1987. Kant and the Claims of Knowledge. Cambridge: Cambridge University Press.

h

d i

0 0

9 8 009

2 2 0 6

bli h d

li

b C

b id

i

i

274

Bibliography

Guyer, Paul. 2000. Kant on Freedom, Law, and Happiness. Cambridge: Cambridge University Press. Guyer, Paul. 2008. Knowledge, Reason, and Taste: Kant’s Response to Hume. Princeton, NJ: Princeton University Press. Guyer, Paul. 2010. ‘Introduction’. In The Cambridge Companion to Kant’s Critique of Pure Reason, edited by Paul Guyer, 1–18. Cambridge: Cambridge University Press. Hall, Bryan. 2011. ‘A Dilemma for Kant’s Theory of Substance’. British Journal for the History of Philosophy 19 (1): 79–109. Hanna, Robert. 2011a. ‘Beyond the Myth of the Myth: A Kantian Theory of NonConceptual Content’. International Journal of Philosophical Studies 19 (3): 323–98. Hanna, Robert. 2011b. ‘Kant’s Non-Conceptualism, Rogue Objects, and The Gap in the B Deduction’. International Journal of Philosophical Studies 19 (3): 399–415. Hanna, Robert. 2017. ‘Richer than You Think: Kant, Conceptual Truth, and the Discursive Structure of the Manifest World’. Studi Kantiani 30: 115–22. Hatfield, Gary. 2001. ‘The Prolegomena and the Critiques of Pure Reason’. In Kant und die berliner Aufklärung: Akten des IX. Internationalen KantKongresses, edited by Volker Gerhardt, Rolf-Peter Horstmann, and Ralph Schumacher. 1:185–208. Berlin: De Gruyter. Hatfield, Gary. 2003. ‘What Were Kant’s Aims in the Deduction?’ Philosophical Topics 31 (1/2): 165–98. Heimsoeth, Heinz. 1966–1971. Transzendentale Dialektik: Ein Kommentar zu Kants Kritik der reinen Vernunft. 4 vols. Berlin: De Gruyter. Henrich, Dieter. 1969. ‘The Proof-Structure of Kant’s Transcendental Deduction’. The Review of Metaphysics 22 (4): 640–59. Henrich, Dieter. 1989. ‘Kant’s Notion of a Deduction and the Methodological Background of the First Critique’. In Kant’s Transcendental Deductions, edited by Eckart Förster, 29–46. Stanford, CA: Stanford University Press. Hoeppner, Till. 2011. ‘Kants Begriff der Funktion und die Vollständigkeit Der Urteils- und Kategorientafel’. Zeitschrift für philosophische Forschung 65 (2): 193–217. Hoeppner, Till. 2021. Urteil und Anschauung: Kants metaphysische Deduktion der Kategorien. Boston: De Gruyter. Horstmann, Rolf-Peter. 1989. ‘Why Must There Be a Transcendental Deduction in Kant’s Critique of Judgment?’ In Kant’s Transcendental Deductions, edited by Eckart Förster, 157–76. Stanford, CA: Stanford University Press. Höwing, Thomas. 2016a. ‘Kant on Opinion, Belief, and Knowledge’. In The Highest Good in Kant’s Philosophy, edited by Thomas Höwing, 201–22. Berlin: De Gruyter. Höwing, Thomas, ed. 2016b. The Highest Good in Kant’s Philosophy. Berlin, Boston: De Gruyter. Hoyningen-Huene, Paul. 2013. Systematicity: The Nature of Science. New York: Oxford University Press.

h

d i

0 0

9 8 009

2 2 0 6

bli h d

li

b C

b id

i

i

Bibliography

275

Hume, David. 2007. David Hume: A Treatise of Human Nature. 2 vols, edited by David Fate Norton and Mary J. Norton. Clarendon Hume Edition Series. Oxford: Oxford University Press. Janiak, Andrew. 2016. ‘Kant’s Views on Space and Time’. In The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta, Winter 2016. https://plato.stanford.edu/archives/win2016/entries/kant-spacetime/. Jauernig, Anja. 2021. The World According to Kant: Appearances and Things in Themselves in Critical Idealism. Oxford: Oxford University Press. Kemp Smith, Norman. 1918. A Commentary to Kant’s Critique of Pure Reason. London: Macmillan. Kleingeld, Pauline. 1998. ‘The Conative Character of Reason in Kant’s Philosophy’. Journal of the History of Philosophy 36 (1): 77–97. Klemme, Heiner. 1996. Kants Philosophie des Subjekts: Systematische und entwicklungsgeschichtliche Untersuchungen zum Verhältnis von Selbstbewusstsein und Selbsterkenntnis. Hamburg: Meiner. Klimmek, Nikolai. 2005. Kants System der transzendentalen Ideen. Berlin: De Gruyter. Kraus, Katharina T. 2018. ‘The Soul as the “Guiding Idea” of Psychology: Kant on Scientific Psychology, Systematicity, and the Idea of the Soul’. Studies in History and Philosophy of Science Part A 71: 77–88. Kraus, Katharina T., and Thomas Sturm. 2017. ‘“An Attractive Alternative to Empirical Psychologies Both in His Day and Our Own”? A Critique of Frierson’s Kant’s Empirical Psychology’. Studi Kantiani 30: 203–23. Kreimendahl, Lothar. 1990. Kant: Der Durchbruch von 1769. Cologne: Dinter. Kuehn, Manfred. 1983. ‘Kant’s Conception of “Hume’s Problem”’. Journal of the History of Philosophy 21 (2): 175–93. La Rocca, Claudio. 2003. Soggetto e mondo. Studi su Kant. Venezia: Marsilio. La Rocca, Claudio. 2011. ‘Formen des Als-Ob bei Kant’. In Über den Nutzen von Illusionen: Die regulativen Ideen in Kants theoretischer Philosophie, edited by Bernd Dörflinger and Günter Kruck, 29–45. Hildesheim: Olms. Land, Thomas. 2014. ‘Spatial Representation, Magnitude and the Two Stems of Cognition’. Canadian Journal of Philosophy 44 (5–6): 524–50. Land, Thomas. 2015. ‘Nonconceptualist Readings of Kant and the Transcendental Deduction’. Kantian Review 20 (1): 25–51. Land, Thomas. 2021. ‘Epistemic Agency and the Self-Knowledge of Reason: On the Contemporary Relevance of Kant’s Method of Faculty Analysis’. Synthese 198: 3137–54. Leirfall, Anita. 2004. ‘Kant’s Metaphysical Exposition: On Philosophical Expositions Considered as Analysis of Given Concepts’. SATS 5 (2): 34–46. Leitner, Heinrich. 1994. Systematische Topik: Methode und Argumentation in Kants kritischer Philosophie. Würzburg: Königshausen & Neumann. Longuenesse, Béatrice. 1998. Kant and the Capacity to Judge: Sensibility and Discursivity in the Transcendental Analytic of the Critique of Pure Reason. Princeton, NJ: Princeton University Press.

h

d i

0 0

9 8 009

2 2 0 6

bli h d

li

b C

b id

i

i

276

Bibliography

Longuenesse, Béatrice. 2006. ‘Kant on A Priori Concepts: The Metaphysical Deduction of the Categories’. In The Cambridge Companion to Kant and Modern Philosophy, edited by Paul Guyer, 129–68. Cambridge: Cambridge University Press. Lu-Adler, Huaping. 2018. Kant and the Science of Logic: A Historical and Philosophical Reconstruction. New York: Oxford University Press. Manchester, Paula. 2003. ‘Kant’s Conception of Architectonic in Its Historical Context’. Journal of the History of Philosophy 41 (2): 187–207. Manchester, Paula. 2008. ‘Kant’s Conception of Architectonic in Its Philosophical Context’. Kant-Studien 99 (2): 133–51. Marwede, Florian. 2016. ‘Kant on Happiness and the Duty to Promote the Highest Good’. In The Highest Good in Kant’s Philosophy, edited by Thomas Höwing, 51–70. Berlin: De Gruyter. Marwede, Florian. 2018. Das höchste Gut in Kants deontologischer Ethik. Berlin: De Gruyter. McCain, Kevin. 2008. ‘The Virtues of Epistemic Conservatism’. Synthese 164 (2): 185–200. McGoldrick, P. M. 1985. ‘The Metaphysical Exposition: An Analysis of the Concept of Space’. Kant-Studien 76 (1–4): 257–75. McLear, Colin. 2015. ‘Two Kinds of Unity in the Critique of Pure Reason’. Journal of the History of Philosophy 53 (1): 79–110. McQuillan, J. Colin. 2016. Immanuel Kant: The Very Idea of a Critique of Pure Reason. Evanston, IL: Northwestern University Press. McQuillan, J. Colin. 2017a. ‘Does Kant’s Critique of Pure Reason Belong to the Tradition of Modern Logic’. In La Modernidad En Perspectiva, edited by Roberto Casales Garcia and Martin Castro Manzana, 45–54. Granada: Editorial Comares. McQuillan, J. Colin. 2017b. ‘Kant on the Science of Aesthetics and the Critique of Taste’. Kant Yearbook 9 (1): 113–32. McQuillan, J. Colin. 2017c. ‘Wolff’s Logic, Kant’s Critique, and the Foundations of Metaphysics’. In 300 Years of Christian Wolff’s German Logic: Sources, Significance and Reception, edited by Arnaud Pelletier and Karin de Boer, 217–39. Hildesheim: Olms. Meer, Rudolf. 2019. Der transzendentale Grundsatz der Vernunft: Funktion und Struktur des Anhangs zur Transzendentalen Dialektik der Kritik der reinen Vernunft. Berlin: De Gruyter. Meier, Georg Friedrich. 1752a. Auszug auf die Vernunftlehre. Halle. Meier, Georg Friedrich. 1752b. Vernunftlehre. Halle. Melnick, Arthur. 2006. ‘The Second Analogy’. In A Companion to Kant, edited by Graham Bird, 169–81. Oxford: Blackwell. Mensch, Jennifer. 2013. Kant’s Organicism: Epigenesis and the Development of Critical Philosophy. Chicago: University of Chicago Press. Merritt, Melissa McBay. 2010. ‘Kant on the Transcendental Deduction of Space and Time: An Essay on the Philosophical Resources of the Transcendental Aesthetic’. Kantian Review 14 (2): 1–37.

h

d i

0 0

9 8 009

2 2 0 6

bli h d

li

b C

b id

i

i

Bibliography

277

Merritt, Melissa McBay. 2015. ‘Varieties of Reflection in Kant’s Logic’. British Journal for the History of Philosophy 23 (3): 478–501. Messina, James. 2014. ‘Kant on the Unity of Space and the Synthetic Unity of Apperception’. Kant-Studien 105 (1): 5–40. Messina, James. 2015. ‘Conceptual Analysis and the Essence of Space: Kant’s Metaphysical Exposition Revisited’. Archiv für Geschichte der Philosophie 97 (4): 416–57. Møller, Sofie. 2020. Kant’s Tribunal of Reason. Cambridge: Cambridge University Press. Moore, Adrian W. 2010. ‘The Transcendental Doctrine of Method’. In The Cambridge Companion to Kant’s Critique of Pure Reason, edited by Paul Guyer, 310–26. Cambridge: Cambridge University Press. Motta, Giuseppe. 2017. ‘One Step? Two Steps? Reflections on R. L. Anderson’s The Poverty of Conceptual Truth’. Studi Kantiani 30: 123–28. Niiniluoto, Ilkka. 2018. Truth-Seeking by Abduction. Cham: Springer International Publishing. Onof, Christian, and Dennis Schulting. 2015. ‘Space as Form of Intuition and as Formal Intuition: On the Note to B160 in Kant’s Critique of Pure Reason’. Philosophical Review 124 (1): 1–58. O’Shea, James R. 1997. ‘The Needs of Understanding: Kant on Empirical Laws and Regulative Ideals’. International Journal of Philosophical Studies 5 (2): 216–54. Pasternack, Lawrence. 2011. ‘The Development and Scope of Kantian Belief: The Highest Good, the Practical Postulates and the Fact of Reason’. ­Kant-Studien 102 (3): 290–315. Pasternack, Lawrence. 2014. ‘Kant on Opinion: Assent, Hypothesis, and the Norms of General Applied Logic’. Kant-Studien 105 (1): 41–82. Pereboom, Derk. 2019. ‘Kant’s Transcendental Arguments’. In The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta, Spring 2019. https://plato.stanford.edu/archives/spr2019/entries/kant-transcendental/. Pollok, Konstantin. 2001. Kants metaphysische Anfangsgründe der Naturwissenschaft: Ein kritischer Kommentar. Hamburg: Meiner. Pollok, Konstantin. 2008. ‘“An Almost Single Inference” – Kant’s Deduction of the Categories Reconsidered’. Archiv für Geschichte der Philosophie 90 (3): 323–45. Proops, Ian. 2003. ‘Kant’s Legal Metaphor and the Nature of a Deduction’. Journal of the History of Philosophy 41 (2): 209–29. Rauscher, Frederick. 2010. ‘The Appendix to the Dialectic and the Canon of Pure Reason: The Positive Role of Reason’. In The Cambridge Companion to Kant’s Critique of Pure Reason, edited by Paul Guyer, 290–309. Cambridge: Cambridge University Press. Reath, Andrews. 1988. ‘Two Conceptions of the Highest Good in Kant’. Journal of the History of Philosophy 26 (4): 593–619. Reich, Klaus. 1986. Die Vollständigkeit der kantischen Urteilstafel. 3rd ed. Hamburg: Meiner.

h

d i

0 0

9 8 009

2 2 0 6

bli h d

li

b C

b id

i

i

278

Bibliography

Reinhold, Karl Leonhard. 1791. Über das Fundament des philosophischen Wissens. Jena. Renaut, Alain. 1998. ‘Transzendentale Dialektik, Einleitung und Buch I (A293/ B349–A338/B396)’. In Immanuel Kant: Kritik der reinen Vernunft, edited by Marcus Willaschek and Georg Mohr, 353–70. Berlin: Akademie Verlag. Rescher, Nicholas. 1979. Cognitive Systematization: A Systems-Theoretic Approach to a Coherentist Theory of Knowledge. Oxford: Backwell. Rohlf, Michael. 2010. ‘The Ideas of Pure Reason’. In The Cambridge Companion to Kant’s Critique of Pure Reason, edited by Paul Guyer, 190–209. Cambridge: Cambridge University Press. Rosefeldt, Tobias. 2016. ‘Closing the Gap: A New Answer to an Old Objection against Kant’s Argument for Transcendental Idealism’. Proceedings of the Aristotelian Society 116 (2): 181–203. Rosefeldt, Tobias. 2021. ‘Kant on the Epistemic Role of the Imagination’. Synthese 198: 3171–92. Schmucker, Josef. 1990. Das Weltproblem in Kants Kritik der reinen Vernunft. Bonn: Bouvier. Schönecker, Dieter. 2005. Kants Begriff transzendentaler und praktischer Freiheit: Eine entwicklungsgeschichtliche Studie. Berlin: De Gruyter. Schulting, Dennis. 2018a. Kant’s Deduction from Apperception: An Essay on the Transcendental Deduction of the Categories. 2nd ed. Berlin, Boston: De Gruyter. Schulting, Dennis. 2018b. ‘The Current Status of Research on Kant’s Transcendental Deduction’. Revista de Estudios Kantianos 3 (1): 69–88. Schultz, Johann. 2000 [1785]. ‘Institutiones Logicae et Metaphysicae by Johann August Heinrich Ulrich’. In Kant’s Early Critics: The Empiricist Critique of the Theoretical Philosophy, edited by Brigitte Sassen, 210–14. Cambridge: Cambridge University Press. Schulze, Gottlob Ernst. 1792. Aenesidemus: oder, Über die Fundamente der von dem Herrn Prof. Reinhold in Jena gelieferten Elementar-Philopsophie. Helmstedt. Serck-Hanssen, Camilla. 2013. ‘The Significance of Infinite Judgment’. In Kant und die Philosophie in weltbürgerlicher Absicht: Akten des XI. Internationalen Kant-Kongresses 2010, 2:409–20. Berlin: De Gruyter. Shabel, Lisa. 2006. ‘Kant’s Philosophy of Mathematics’. In The Cambridge Companion to Kant and Modern Philosophy, edited by Paul Guyer, 94–128. Cambridge: Cambridge University Press. Smit, Houston. 1999. ‘The Role of Reflection in Kant’s Critique of Pure Reason’. Pacific Philosophical Quarterly 80 (2): 203–23. Specht, Andrew. 2014. ‘F. A. Trendelenburg and the Neglected Alternative’. British Journal for the History of Philosophy 22 (3): 514–34. Stan, Marius, and Eric Watkins. 2014. ‘Kant’s Philosophy of Science’. In The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta, Fall 2014. https://plato.stanford.edu/archives/fall2014/entries/kant-science/. Stang, Nicholas F. 2012. ‘Kant on Complete Determination and Infinite Judgement’. British Journal for the History of Philosophy 20 (6): 1117–39.

h

d i

0 0

9 8 009

2 2 0 6

bli h d

li

b C

b id

i

i

Bibliography

279

Stapleford, Scott. 2008. Kant’s Transcendental Arguments: Disciplining Pure Reason. New York: Continuum. Stern, Robert. 2004. ‘Does “Ought” Imply “Can”? And Did Kant Think It Does?’ Utilitas 16 (1): 42–61. Stern, Robert. 2006. ‘Metaphysical Dogmatism, Humean Scepticism, Kantian Criticism’. Kantian Review 11: 102–16. Stern, Robert. 2019. ‘Transcendental Arguments’. In The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta, Summer 2019. https://plato.stanford .edu/archives/sum2019/entriesranscendental-arguments/. Stevenson, Leslie. 2003. ‘Opinion, Belief or Faith, and Knowledge’. Kantian Review 7: 72–101. Strawson, Galen. 2014. The Secret Connexion: Causation, Realism, and David Hume. Revised ed. Oxford: Oxford University Press. Strawson, Peter. 1966. The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason. London: Methuen. Sturm, Thomas. 2009. Kant und die Wissenschaften vom Menschen. Paderborn: mentis. Tolley, Clinton. 2012. ‘The Generality of Kant’s Transcendental Logic’. Journal of the History of Philosophy 50 (3): 417–46. Tonelli, Giorgio. 1994. Kant’s Critique of Pure Reason within the Tradition of Modern Logic. Hildesheim: Olms. Vahid, Hamid. 2004. ‘Varieties of Epistemic Conservatism’. Synthese 141 (1): 97–122. Vaihinger, Hans. 1881–1892. Commentar zu Kants Kritik der reinen Vernunft. 2 vols. Stuttgart: Verlag von W. Spemann and Union Deutsche Verlagsgesellschaft. Vaihinger, Hans. 1911. Die Philosophie des Als Ob. Berlin: Reuther & Reichard. Van Cleve, James. 1999. Problems from Kant. New York: Oxford University Press. van den Berg, Hein. 2011. ‘Kant’s Conception of Proper Science’. Synthese 183 (1): 7–26. Vinci, Thomas C. 2015. Space, Geometry, and Kant’s Transcendental Deduction of the Categories. Oxford: Oxford University Press. Walden, Kenneth. 2019. ‘Reason Unbound: Kant’s Theory of Regulative Principles’. European Journal of Philosophy 27 (3): 575–92. Ward, Andrew. 2001. ‘Kant’s First Analogy of Experience’. Kant Studien 92 (4): 387–406. Wartenberg, Thomas E. 1979. ‘Order through Reason: Kant’s Transcendental Justification of Science’. Kant-Studien 70 (1–4): 409–24. Wartenberg, Thomas E. 1992. ‘Reason and the Practice of Science’. In The Cambridge Companion to Kant, edited by Paul Guyer, 228–48. Cambridge: Cambridge University Press. Watkins, Eric. 2005. Kant and the Metaphysics of Causality. Cambridge: Cambridge University Press. Waxman, Wayne. 2005. Kant and the Empiricists: Understanding Understanding. Oxford: Oxford University Press.

h

d i

0 0

9 8 009

2 2 0 6

bli h d

li

b C

b id

i

i

280

Bibliography

Willaschek, Marcus. 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 51 (4): 537–64. Willaschek, Marcus. 1998. ‘Phaenomena/Noumena und die Amphibolie der Reflexionsbegriffe (A235/B294–A292/B349)’. In Immanuel Kant: Kritik der reinen Vernunft, edited by Georg Mohr and Marcus Willaschek, 325–51. Berlin: Akademie Verlag. Willaschek, Marcus. 2010. ‘The Primacy of Pure Practical Reason and the Very Idea of a Postulate’. In Kant’s Critique of Practical Reason: A Critical Guide, edited by Andrews Reath and Jens Timmermann, 168–96. Cambridge: Cambridge University Press. Willaschek, Marcus. 2016. ‘Must We Believe in the Realizability of Our Ends? On a Premise of Kant’s Argument for the Postulates of Pure Practical Reason’. In The Highest Good in Kant’s Philosophy, edited by Thomas Höwing, 223–44. Berlin: De Gruyter. Willaschek, Marcus. 2018. Kant on the Sources of Metaphysics: The Dialectic of Pure Reason. Cambridge: Cambridge University Press. Willaschek, Marcus, and Eric Watkins. 2020. ‘Kant on Cognition and Knowledge’. Synthese 197: 3195–213. Williams, Jessica J. 2018. ‘Kant on the Original Synthesis of Understanding and Sensibility’. British Journal for the History of Philosophy 26 (1): 66–86. Wolff, Christian. 1965 [1754]. Vernünftige Gedanken von den Kräften des menschlichen Verstandes und ihrem richtigen Gebrauche in Erkenntnis der Wahrheit. Hildesheim: Olms. Wolff, Michael. 1995. Die Vollständigkeit der kantischen Urteilstafel: Mit einem Essay über Freges Begriffsschrift. Frankfurt am Main: Klostermann. Wolff, Michael. 2018. ‘Freiheit und Natur: Zu Kants arkitetonischem Programm von Philosophie’. In Natur und Freiheit: Akten des XII. Internationalen KantKongresses, edited by Violetta Waibel, Margit Ruffing, and David Wagner, 1:133–56. Berlin: De Gruyter. Wolff, Robert Paul. 1963. Kant’s Theory of Mental Activity: A Commentary on the Transcendental Analytic of the Critique of Pure Reason. Cambridge, MA: Harvard University Press. Wolff-Metternich, Brigitta-Sophie von. 1995. Die Überwindung des mathematischen Erkenntnisideals: Kants Grenzbestimmung von Mathematik und Philosophie. Berlin: De Gruyter. Wood, Allen W. 2020. Kant and Religion. Cambridge: Cambridge University Press. Wundt, Max. 1924. Kant als Metaphysiker: Ein Beitrag zur Geschichte der deutschen Philosophie im 18. Jahrhundert. Stuttgart: F. Enke. Young, J. Michael. 1992. ‘Functions of Thought and the Synthesis of Intuitions’. In The Cambridge Companion to Kant, edited by Paul Guyer, 101–22. Cambridge: Cambridge University Press. Yovel, Yirmiyahu. 1980. Kant and the Philosophy of History. Princeton, NJ: Princeton University Press.

h

d i

0 0

9 8 009

2 2 0 6

bli h d

li

b C

b id

i

i

Bibliography

281

Ypi, Lea. 2011. ‘Practical Agency, Teleology and System in Kant’s Architectonic of Pure Reason’. In Politics and Metaphysics in Kant, edited by Sorin Baiasu, Sami Pihlström, and Howard Williams, 134–52. Cardiff: University of Wales Press. Ypi, Lea. 2017. ‘The Transcendental Deduction of Ideas in Kant’s Critique of Pure Reason’. Proceedings of the Aristotelian Society 117 (2): 163–85. Ypi, Lea. 2021. The Architectonic of Reason: Purposiveness and Systematic Unity in Kant’s Critique of Pure Reason. Oxford: Oxford University Press. Zocher, Rudolf. 1958. ‘Zu Kants transzendentaler Deduktion der Ideen der reinen Vernunft’. Zeitschrift für philosophische Forschung 12 (1): 43–58.

h

d i

0 0

9 8 009

2 2 0 6

bli h d

li

b C

b id

i

i

Index

actuality, of God and immortality, 56, 208, 210–11, 213–16, 218–19 aggregate, 18, See also unity, architectonic vs. technical Allais, Lucy, 137, 140, 144, 178, 180, 181, 231 Allison, Henry E., 108, 109, 123, 141, 178, 217 Ameriks, Karl, 10, 65, 67, 123–5, 150, 248 analysis conceptual, 62, 78–80, 82, 83, 86–7, 126, 233–5, 244, 257, See also knowledge, analytic faculty-, 41–2, 60–2, 207, 228 Anderson, R. Lanier, 81, 231–40, 245 antinomy, 170, 187–8, 240, 260–2 first, 186, 187 third, 213, 261 Antinomy of Pure Reason, 4, 99, 107, 113, 114, 118, 163, 170, 181, 184, 190, 235, 240–2 appearances, totality of, 116, 163, 170, 186–90, 193, 241, See also ideas, cosmological apperception, synthetic unity of, 142–3, 148–9 Architectonic of Pure Reason, 3–5, 18, 23–6, 29–31, 33–4, 37, 50, 51, 69, 225 argument moral, 8 practical, 12, 166, 172–3, 208, 209, 211, 215–17, 219, 222–8 transcendental, 10, 150–1, 166–7, 195

Caimi, Mario, 93, 98, 103, 161 Callanan, John J., 10, 65–8 Canon of Pure Reason, 2, 51, 54–6, 153, 159, 172–3, 207–9, 213, 216, 217, 219, 222, 223, 225–6 Capozzi, Mirella, 43, 99, 100 Carboncini, Sonia, 45 Carl, Wolfgang, 10, 251, 254, 258, 261, 262 Cassam, Quassim, 10 categories as fundamental ways in which we order the manifold of intuition, 93, 102–3, 137, 143, 144, 203 as root concepts, See concepts, derivative vs. root immanent and transcendental use of, 174, 184–5 negative argument for the validity of, 12, 195–206 positive argument for the validity of, 12, 138, 152, 153, 194–200, 206 table of, 88–93, 98, 115, 198, See also completeness, of the table of the categories causality, 156, 254–8, 260 category of, 115, 150, 186, 204 Chance, Brian A., 54, 252, 257, 258 Chignell, Andrew, 55, 81, 159, 210, 214–17, 219, 220, 223 cognitions mathematical, 3, 50, 51 philosophical, 3, 51, 53 practical, 211–14 theoretical, 3, 30–2, 38, 40, 211–12, 222, 225–8 coherence, 57, 173, 187, 227 systematic, 6, 12–13, 28, 29, 33, 36–8, 169–73, 223–4, 259 command, moral, 54–5, 217–18, See also imperative; ought completeness of the table of the categories, 88–92, 98, See also categories, table of of the table of the forms of judgement, 98, See also judgements, table of Conant, James, 123, 137

Bacin, Stefano, 32, 45 Barale, Massimo, 1, 41 Beck, Jakub Sigismund, 9 belief (Glaube), 55, 153, 159, 160, 164–6, 172, 173, 207–9, 219–22, 226–8 doctrinal, 159 moral, 207, 221–2, 227 Bird, Alexander, 135 Bird, Graham, 10, 78, 178, 208, 251, 254 Bohle, Johann Gottlieb, 8 Brandt, Reinhard, 96, 261, 262 Bröcker, Walter, 106

282

h

d i

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

i

i

Index conceivability, 150–1 concepts adequate, 233 construction of, 3, 29, 51, 79, 85, 86 derivative vs. root, 9, 70–2, 169 predicative vs. non-predicative use of, 90–2 pure, 70–2, 86, 129, 169, 198, 239, 248 conceptualism/non-conceptualism, 123, 137–41, 148–9 conditioned, 107, 109–11, 114–18, 184, 186–91, 241, See also unconditioned conditions series of, 111, 114–18, 190 totality of, 107, 109–11, 190 conservatism common sense, 65–7, 124, 134, 135, 149–50, 164, 167, 254 epistemic, 67 methodological, 65–9, 73, 124, 134, 149, 164, 165, 167, 173, 208 continuity, principle of, 20–2, 74 cosmology, rational, 3–4, 7, 8, 226, 254 Critique of Practical Reason, 4, 32, 216, 253, 261, 267–9 critique of pure reason empirical side vs. rational side of, 265 as faculty analysis, See analysis, facultynegative utility of, 172, 173 positive utility of, 169, 172–3, 208–9, 222 as propaedeutic, 1, 3–5, 9–11, 41–2, 61 Critique of the Power of Judgement, 216, 267–9

faculty for judging (Vermögen zu urteilen), 89, 92, 96, 104 of concepts, 75, 90, 96 of inference, 18, 75, 90, 92, 96, 106, 108, 114, 268, See also inference, forms of of judgement (Urteilskraft), 90, 96 Falkenstein, Lorne, 75, 76, 79, 83, 86, 124, 125, 131, 178 Ferrarin, Alfredo, 1, 17, 23, 31, 36, 38, 41, 48, 69, 147, 263, 264, 268 Fichte, Johann Gottlieb, 8–9 Finster, Reinhard, 45 focus imaginarius, 18, 22–3, 32, 40 Fonnesu, Luca, 55, 159 Förster, Eckart, 69, 268 Forster, Michael N., 254, 257, 258, 260 freedom, 8, 37, 54, 171–3, 207–10, 213, 226, 243, 268 concept of, 173, 209 of the will, 54 practical, 55 transcendental, 213, 226, 240 Frierson, Patrick R., 208, 248 Fugate, Courtney, 17 function, 90–7, 100, 104, 142–4 of reason in an inference, 108 of unity, 92, 97

de Boer, Karin, 8–10, 41, 258, 260, 262 de Jong, Willem R., 79 deduction metaphysical, See metaphysical deduction transcendental, See transcendental deduction Discipline of Pure Reason, 2, 51–7, 64, 170, 193, 210, 223, 239, 254, 257, 265, 267 doctrine of method of general logic, 11, 42, 45–50 of particular sciences, 11, 42, 47–50, 58 dogmatic procedure of reason, 13, 229, 231, 246–9 Dyck, Corey W., 248 end essential, 18, 30, 31, 33–5, 37–8 final, 31, 37–9, 171 Ertl, Wolfgang, 257, 260 existence/non-existence, category of, 115 explainability, 151 exposition, 79 metaphysical, 73, 76–88, 123–31, 177–8, 182, 192 transcendental, 67, 73, 76, 87, 123–37, 149, 151, 176–82, 191–2

h

d i

0 0

9 8 009

2 2 0

bli h d

li

b C

283

b id

Garve, Christian, 259 Gava, Gabriele, 10, 24, 26, 31, 34, 36, 55, 61, 79, 86, 144, 146, 151, 159, 160, 164, 165, 195, 213, 219, 220, 230, 243, 247, 248 Gawlick, Günter, 260, 261 geometry, 59, 66–8, 85, 127–32, 135, 136, 151, 175–6, 179 Ginsborg, Hannah, 137 God concept of, 7, 72, 165, 209, 225 existence of, 13, 54, 55, 171, 172, 207–10, 213–18, 222–5, 256, 259, 260, 262 idea of, 72, 105, 112, 162–3 Goldberg, Nathaniel Jason, 20, 161 Golob, Sacha, 137 Goy, Ina, 17, 37 Grier, Michelle, 152, 185 Grüne, Stefanie, 137 Guyer, Paul, 111, 123, 208, 215, 236, 254, 257, 258 Hall, Bryan, 236 Hamann, Johann Georg, 260 Hanna, Robert, 137, 231 happiness, 31, 213–18, 221, 223, 226 Hatfield, Gary, 10, 195, 254 Hegel, Georg Wilhelm Friedrich, 89

i

i

284

Index

Heimsoeth, Heinz, 109 Henrich, Dieter, 10, 123, 138, 139, 143 highest good, 31–3, 37–40, 55–6, 171, 172, 213–22 History of Pure Reason, 2, 51, 56, 229, 262 Hoeppner, Till, 91, 93 homogeneity, principle of, 20–2, 74 Horstmann, Rolf-Peter, 161, 165 Höwing, Thomas, 31, 55, 159, 214, 219, 220 Hoyningen-Huene, Paul, 29 Hume, David, 13, 229, 240, 245, 251–66, 270 hypotheses, 52–3, 154, 160, 164

Kreimendahl, Lothar, 257, 260, 261 Kuehn, Thomas, 257, 258, 260, 261

Ideal of Pure Reason, 4, 22, 115, 118, 171 ideas cosmological, 105–7, 113–19, 162, 174, 175, 183–4, 186–93, 226, 260, See also appearances, totality of psychological, 106, 107, 117–18, 162, 170 regulative, 5, 21, 22, 26, 70, 152, 156–62, 190, 212, 223, 224, 250 theological, 106, 107, 115, 117, 162, 170 imperative, 54–5, 211–13, 217, 249, 269, See also command, moral; ought Inaugural Dissertation, 1, 10 inference, forms of, 92, 108, 113, 116, 120, See also faculty, of inference interest of reason, 54, 160, 240–1 practical, 37, 172, 241 Janiak, Andrew, 78 Jauernig, Anja, 8 judgements categorical, 112 disjunctive, 112 fundamental kinds of, 97, 100, 137 hypothetical, 112 infinite, 98–9 particular, 104 singular, 100 table of, 89–92, 97–100, 105, 109, 121, 144 universal, 99, 100, 104, 108 Kemp Smith, Norman, 43, 78, 253 Kleingeld, Pauline, 31 Klemme, Heiner, 248 Klimmek, Nikolai, 72, 105, 106, 108, 111, 115 knowledge analytic, 79, 80–1, 86, 126, See also analysis, conceptual as a form of taking-to-be-true (Wissen), 55, 207, 213, 219, 220, 222, 227 synthetic, 80–2 Kraus, Katharina T., 248

h

d i

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

La Rocca, Claudio, 1, 17, 36, 41, 43, 161 Land, Thomas, 61, 137, 147 law causal, 115, 155–7, 251, 255 moral, 31, 212–17, 221–2, 226 legitimacy of a concept, 124, 126, 127, 130 Leirfall, Anita, 79 Leitner, Heinrich, 10 logic applied, 43, 44 general, 43, 44, 45, 50, 98–100, 120 particular, 57, 60–1 practical, 43–5, 49 theoretical, 43 transcendental, 93, 98–100, 121 Longuenesse, Béatrice, 10, 81, 93, 123, 137 Lu-Adler, Huaping, 44, 72 Manchester, Paula, 17 manifold of intuition, empirical vs. pure, 103, 141, 144–9 Marwede, Florian, 31, 214, 217 mathematics, 1, 24, 26, 27, 29, 33–5, 39, 50–2, 79, 86, 125, 127, 146, 233–4, 243, 254 McCain, Kevin, 67 McGoldrick, P. M., 79 McLear, Colin, 139 McQuillan, J. Colin, 1, 24, 36, 41, 57, 230 Meer, Rudolf, 161 Meier, Georg Friedrich, 42, 43 Melnick, Arthur, 253 Mendelssohn, Moses, 260 Mensch, Jennifer, 17, 23, 37 Merritt, Melissa McBay, 10, 124–7, 130 Messina, James, 78–81, 139 metaphilosophy, 13, 231, 242–6, 248 metaphysical deduction and conceptual analysis, 78–82, 86–7 general characterization of, 5, 12, 62, 73, 75–6, 122, 176 Metaphysical Foundations, 4, 28, 195, 196, 199–201, 206, 225, 247, 248 metaphysics as a natural predisposition, 2, 8 general, 252, 254, 255, 258, 266 idea of, 7, 13, 15, 17, 29, 37, 40, 56, 171, 172, 223, 228 of morals, 3, 4, 6, 7, 30, 31, 50, 269 of nature, 3, 5, 30, 50, 70, 248, 249 special, 4, 6–8, 50, 53, 70, 72, 248–9, 252, 254–6, 258–9, 266 misologist, 38

i

i

Index Møller, Sofie, 123 Moore, A W., 10 Motta, Giuseppe, 231

Proops, Ian, 123 proposition practical vs. theoretical, 211 prosyllogism, 109–12, 117–19, 121, See also syllogism psychology empirical, 4, 24, 51, 225, 248 rational, 3, 4, 72, 225, 248, 249, 254 purposiveness, 268

necessity/contingency, category of, 115, 186 negation, category of, 115 neglected alternative, 178 Niiniluoto, Ilkka, 135 obligation, moral, 55, 216–19, 221, 223, See also command, moral On a Discovery, 237 On the Use of Teleological Principles, 195, 198, 200 Onof, Christian, 139 ontology, 66–8, 254, 258 opinion (Meinung), 52, 55, 164, 219 organon, 45–7 O’Shea, James R., 152–6 ought, 8, 211–13, 219, 223, See also command, moral; imperative ought implies can principle, 212, 213 Paralogisms of Pure Reason, 72, 115, 118, 171 Pasternack, Lawrence, 55, 159, 219 Pereboom, Derk, 10 philodox, 33, 36–8 philosopher, 32, 36–8 philosophy history of, 229, 252, 262–4 school concept of, 17, 29–38 worldly concept of, 11, 18, 29–40, 55, 171 physics, rational, 3, 4 physiology, 3, 4, 30, 225 plurality, category of, 115 Pollok, Konstantin, 17, 24, 195, 200, 248 possibility logical, 208, 209 real, 215 real practical, 215 predicables, 71 principle of contradiction, 237, 242, 249 immanent vs. transcendent, 185 metaphysical vs. transcendental, 248 of pure empiricism, 240, 242 of sufficient reason, 237, 242 regulative, 157, 255 supreme, 22, 114, 116–18, 184–6, 188–93, See also unconditioned Progress, 263 Prolegomena, 17, 26, 27, 34, 71, 125, 127, 236, 237, 242, 245, 251, 256, 257, 259, 260 proof dogmatic, 53, 235 transcendental, 10, 51–3

h

d i

0 0

9 8 009

2 2 0

bli h d

li

b C

285

b id

Rauscher, Frederick, 159 reality, category of, 114, 186 reason logical requirement of, 159, 160, 163 logical use of, 111, 117 in the narrow sense, 18, 75, 90, 92, 95, 96, 119 nature of, 107, 113, 252, 263, 264 real use of, 117 self-knowledge of, 263–4 Reath, Andrews, 31, 214, 270 reflection, transcendental, 10 Reich, Klaus, 89 Reinhold, Karl Leonhard, 9, 199, 253 Renaut, Alain, 105, 106 representation, root, 71, 76, 77, 86, 136 Rescher, Nicholas, 29 Rohlf, Michael, 111, 113 Rosefeldt, Tobias, 147, 179 rule, object- and cognition-dependent, 15, 42–6, 48–53, 56–8, 193 schema of sensibility, 161 of the idea of a science, 25–6 Schmucker, Josef, 112 Schönecker, Dieter, 213 Schulting, Dennis, 98, 123, 137, 139–41, 144, 147 Schultz, Johann, 197 Schulze, Gottlob Ernst, 253 Schütz, Christian Gottfried, 4 science definition of, 25, 27 idea of, 6–7, 15, 18, 22–9, 34, 35, 169, 171, 223 natural, 1, 17, 28, 154, 253–5 Serck-Hanssen, Camilla, 99 Shabel, Lisa, 247 Smit, Houston, 10 soul concept of, 165 idea of, 72, 105, 112, 162–3 immortality of, 13, 37, 54, 55, 171–2, 207, 210, 213–16, 222–4 substantiality and immateriality of, 171 space, singularity of, 78, 80, 82–7 Specht, Andrew, 178 specification, principle of, 20–2, 74

i

i

286

Index

Stan, Marius, 17, 24 Stang, Nicholas F., 99 Stapleford, Scott, 10 Stern, Robert, 151, 213, 257 Stevenson, Leslie, 55, 159, 219, 220 Strawson, Galen, 252 Strawson, Peter, 10, 123, 253 Sturm, Thomas, 17, 25–7, 248 substance, category of, 118, 235–6 syllogism, 19, See also prosyllogism categorical, 105, 108–10, 112, 118–19 disjunctive, 105, 112 hypothetical, 105, 112, 115, 116 synthesis categorical, 111 disjunctive, 111 hypothetical, 111 of the manifold of intuition, 101–3, 141–3, 146–9, 151 systematicity, See unity, architectonic vs. mere systematicity taking-to-be-true (Fürwahrhalten), 52, 159, 208, 219, 220, 223, 227 objective sufficiency of, 220 subjective sufficiency of, 219, 220 theology moral, 225 rational, 3–4, 7, 8, 72, 249, 254 transcendental, 225 time, singularity of, 80, 82, 86, 87, 143 Tolley, Clinton, 44, 50, 72 Tonelli, Giorgio, 1, 36, 41, 43, 45, 57 totality, category of, 100, 109, 114, 186 Transcendental Aesthetic, 6, 12, 37, 57, 71, 73–6, 78, 86, 88, 119, 123, 124, 126, 137, 139–41, 148, 174, 176–8, 181, 182, 192, 198, 202, 203 Transcendental Analytic, 6, 12, 37, 62, 71, 73–5, 88, 107, 119, 120, 125, 174, 177, 184–6, 188, 190–3, 199, 203, 237 transcendental deduction as a transcendental argument, See argument, transcendental as inference to the best explanation, 135, 136, 167 general characterization of, 5, 73, 122–4, 166–7 unavoidable necessity of, 174–7, 182–3, 192, 196, 199–201, 203, 204 Transcendental Dialectic, 6, 7, 12, 19, 20, 22, 26, 36, 55, 71–5, 92, 96, 105–7, 110, 111, 114–17, 119, 120, 124, 152, 153, 156, 159, 166, 174, 183–6, 188, 190–3, 195, 223, 225, 234, 239, 270 truth conceptual, 231, 245 containment account of, 233

h

d i

0 0

9 8 009

2 2 0

bli h d

li

b C

b id

Ulrich, Johann August Heinrich, 197 unconditioned, 105–20, 184–5, 192, 261, See also conditioned; principle, supreme undecidability, theoretical, 208, 222, 224, 227, 228 understanding in the broad sense, 90–2, 96 in the narrow sense, 75, 88, 90, 96, 97, 104 unity architectonic vs. mere systematicity, 6, 11, 15, 17–24, 26, 27 architectonic vs. technical, 24, 33–6, 38, 39, See also aggregate of space and time as wholes, 138–43, 148 of the manifold of intuition, See synthesis, of the manifold of intuition Vahid, Hamid, 67 Vaihinger, Hans, 76, 124, 125, 152, 253 validity indeterminate, 74, 152–3, 160–2 indirect, 136, 153, 156, 158–60, 162–7 objective (definition), 122–3 practical, 153, 159, 160, 164–7, 177 Van Cleve, James, 236 van den Berg, Hein, 17, 24 Vinci, Thomas C., 136 virtue, 31, 40, 269 Walden, Kenneth, 190 Ward, Andrew, 236 Wartenberg, Thomas E., 20, 152–5 Watkins, Eric, 17, 24, 227, 251, 254, 258 Waxman, Wayne, 258 Willaschek, Marcus, 8, 10, 19, 24, 31, 55, 81, 105, 106, 108, 109, 117, 152, 159, 160, 162, 163, 178, 181, 185, 190, 219, 227 Williams, Jessica J., 137, 139, 144 wisdom, 38–9 doctrine of, 31–2, 40 Wolff, Christian, 13, 35, 44, 48, 93, 95, 229, 230–8, 240, 242, 245–50, 263 Wolff, Michael, 89–3, 95, 96, 98, 208, 226 Wolff, Robert Paul, 253 Wolff-Metternich, Brigitta-Sophie von, 79 Wood, Allen W., 215–17 world concept of, 7, 165, 189, 213 idea of, 72, 105, 162–3 Wundt, Max, 106 Young, J. Michael, 17, 102 Yovel, Yirmiyahu, 31, 214, 264 Ypi, Lea, 8, 17, 36, 38, 152, 161, 225 Zocher, Rudolf, 161

i

i