Self, World, and Art: Metaphysical Topics in Kant and Hegel 9783110290813, 9783110290783

Citation preview

Dina Emundts (Ed.) Self, World, and Art

Self, World, and Art Metaphysical Topics in Kant and Hegel

Edited by Dina Emundts

Printed with the financial support of the Thyssen-Stiftung, Cologne.

ISBN 978-3-11-029078-3 e-ISBN 978-3-11-029081-3 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.dnb.de. © 2013 Walter de Gruyter GmbH, Berlin/Boston Typesetting: Anne Mone Sahnwaldt, Konstanz Printing: Hubert & Co. GmbH & Co. KG, Göttingen ♾ Printed on acid-free paper Printed in Germany www.degruyter.com

For Rolf Horstmann

Content Dina Emundts Introduction

9

Dina Emundts Einleitung 13 Rolf-Peter Horstmann Propositional Activity in Kant and Hegel Barry Stroud Judgement, Self-Consciousness, Idealism Dina Emundts Kant über Selbstbewusstsein

17

41

51

Wolfgang Carl The Copernican Turn and Stroud’s Argument from Indispensability Béatrice Longuenesse Kant and Hegel on the Moral Self

93

Hannah Ginsborg The Appearance of Spontaneity Kant on Judgment and Empirical Self-Knowledge

119

Stefanie Grüne Kant and the Spontaneity of the Understanding

145

Ulrich Schlösser Concept Formation, Synthesis and Judgment Kant’s Theory of the Logical and Cognitive Activities of the Mind Eckart Förster Grenzen der Erkenntnis?

207

79

177

8

Content

Tobias Rosefeldt Dinge an sich und der Außenweltskeptizismus Über ein Missverständnis der frühen Kant-Rezeption

221

Andrew Chignell Ogilby, Milton, Canary Wine, and the Red Scorpion Another look at Kant’s Deduction of Taste 261 Paul Guyer The End of Art and the Interpretation of Geist

283

Anton Friedrich Koch Metaphysik bei Hegel oder analytische, synthetische und hermeneutische Philosophie 307 Gary Hat eld Russell’s Progress Spatial Dimensions, the From-Which, and the At-Which Index of Names

345

321

Dina Emundts

Introduction The contributions collected in this book address metaphysical topics in Kant and Classical German Philosophy. They deal with questions regarding what there is and how that-which-there-is is structured. Each of the essays is also concerned (either directly or indirectly) with the question of how metaphysical the respectively considered views of Kant, Jacobi, Fichte, Hegel, and others actually are. Here, we mean the question regarding the metaphysical claims, presuppositions, and implications of a theory, where these claims, presuppositions, or implications are viewed as metaphysical insofar as they represent assumptions about the existence of non-physical entities. Over the past few decades, hardly anyone has discussed the question regarding the metaphysical nature of Kant’s philosophy and that of the German Idealists more than Rolf Horstmann. He both turned it into one of the central questions intended to make an alleged distinction between Kant and his followers¹ and even employed it against certain analytical interpretations of Hegel to which he himself had always felt closely connected.² This volume begins with a contribution from Rolf Horstmann on the question of which role we ought to attribute to “self-consciousness” in the grounding of the Categories in Kant and Hegel. The majority of the contributions which follow make reference to Horstmann’s ideas regarding the metaphysical orientation of Kant and the German Idealists. They do so by closely examining topics related to self-consciousness, idealism, theories of reality, and aesthetic experience. The rst six contributions look at the topic of self-consciousness and knowledge. Kant’s Idealism plays a special role here. Horstmann’s considerations stand in close relation with his theory of self-consciousness, according to which selfconsciousness appears on the scene together with so-called “conscious propositional acts.” Horstmann investigates the question of how exactly Kant on the one hand and Hegel on the other understand these propositional acts. He thus employs a new means of showing which logical structures make up the world according to Kant and Hegel, what these logical structures are like, and what that means for questions of self-consciousness. The contribution by Barry Stroud engages with

1 Cf. Horstmann, Rolf-Peter. 1997. “Hegels Kritik der Kantischen Kategorien”. In: Horstmann, R.-P. (ed.). Bausteine kritischer Philosophie. Berlin (Philo-Verlag), 181–200. 2 Cf. Horstmann, Rolf-Peter. 2004. “Das Endliche und das Unendliche in Hegels Denken”. In: Menegoni, F.; Illetterati, L. (ed.). Das Endliche und das Unendliche in Hegels Denken. Stuttgart (Klett-Cotta), 83–102.

10

Emundts

Horstmann’s claims regarding self-consciousness and develops an alternative theory of self-consciousness. This alternative theory is not bound by Kant’s Idealism. Rather it takes up various points from Kant’s theory that can hold independently of an idealistic conception. My own contribution picks up here, insofar as I discuss what “self-consciousness” means for Kant. For me it is decisive that Kant develops his theory of self-consciousness in connection with his idealistic philosophy. Wolfgang Carl’s essay also connects up here. He discusses Kant’s Copernican Turn. Via an analysis of the concept of form, he explicates Kant’s understanding of philosophy as metaphysics. His interpretation of Kant sets him apart from the interpretation that he sees as grounding Stroud’s critical analysis of Kant’s Idealism. In the essay that follows, Béatrice Longuenesse discusses Kant’s and Hegel’s conception of the moral self. Longuenesse asks which metaphysical presuppositions Kant and Hegel are committed to, in order to be able to hold their respective conceptions of a moral self. The essay by Hannah Ginsborg, nally, discusses a central aspect of Kant’s theory of self-consciousness, namely in what relation transcendental self-consciousness stands to empirical self-consciousness. The four essays that follow deal with the topic of the possibility of knowledge. In various ways the central question of each of these contributions deals with the limits of knowledge. The essay by Stefanie Grüne picks up on the topic of Ginsborg’s paper, as she discusses Kant’s concept of spontaneity and the tension between Kant’s claims that we are, on the one hand, spontaneous thinking beings yet, on the other hand, that we have to be viewed as beings whose actions are causally brought about. The activity-aspect of our thought is also at the forefront of Ulrich Schlösser’s essay. He defends Kant’s idea of an action or activity of reason by distinguishing between the various activities of reason in the construction of concepts and judgments and by shedding light on their function and connection to one another. In this way, he also attempts to address objections to the Kantian view. The contribution by Eckart Förster discusses Kant’s claims regarding the limits of knowledge. What led Kant – as opposed to, e.g., Fichte – to view an intellectual intuition as impossible? Förster provides an answer to this question via an analysis of Kant’s view of incongruent counterparts at di erent points in his life. Tobias Rosefeldt takes up the topic of the limits of knowledge in his contribution by taking into consideration the weighty objections of Kant’s idealist successors against his theory of the thing-in-itself. Rosefeldt connects this history of the early reception of Kant with an interpretation of Kantian philosophy on which these objections ultimately cannot succeed. In their contributions, Andrew Chignell and Paul Guyer dedicate themselves to the aesthetic conceptions of Kant and Hegel. Chignell deals with Kant’s theory of aesthetic judgment by looking closely at its presuppositions and implications.

Introduction

11

Guyer is concerned with what the concept of Geist means for Hegel’s aesthetics and attempts to uncover the role this concept plays in aesthetic contexts. The central question here is what ontological implications the conception of Spirit has for Hegel. The essays by Anton Koch and Gary Hat eld examine the critical development of Kant and Hegel’s ideas into the 20th Century. Koch develops a sketch of Hegel’s system as metaphysics, in order to raise the question in what sense the philosophers that followed him (and especially Heidegger) took up and/or “overcame” this kind of philosophy. Hat eld gives a detailed analysis of Russell’s views on space and discusses their connection to the Kantian understanding of space on the backdrop of a direct realist theory of perception. Here, the reference to empirical knowledge plays a decisive role. In this way the question of the metaphysical content of the philosophy of Kant and his successors is taken up in various respects. This book is dedicated to Rolf Horstmann. The contributions here stem from a conference I organized in summer 2011 to honor Horstmann for his philosophical work and to thank him personally. Among his many accomplishments, Horstmann initiated a colloquium in Berlin that led to an exciting exchange between philosophers from around the world in various areas, who work on topics in Kant and Classical German Philosophy. Hopefully this volume can serve as an example of this exchange. Rolf Horstmann had a signi cant in uence on every author of this volume, as a teacher and/or a friend. I would like to thank everyone who in some way or another contributed to the conference. I would especially like to thank the authors who have contributed papers to this volume. Additionally, I would like to thank my colleagues, Jochen Briesen and Amber Gri oen, for their careful redaction of the essays. Thank you to Anders Landig and Wolfgang Scha arzyk for the formatting of the texts and to Anne Mone Sahnwaldt for supervising the book and putting together the manuscript. I would also like to extend my thanks to the Fritz-Thyssen-Sti ung for their generous support in funding the conference that gave rise to this volume and for covering the printing costs.

Dina Emundts

Einleitung Die in diesem Buch versammelten Beiträge behandeln metaphysische Themen bei Kant und der Klassischen Deutschen Philosophie. Sie beschä igen sich mit den Fragen, was es gibt und wie dasjenige, was es gibt, strukturiert ist. In allen Aufsätzen geht es dabei auch direkt oder indirekt um die Frage, wie metaphysisch die jeweils betrachteten Konzeptionen von Kant, Jacobi, Fichte, Hegel und anderen sind. Damit ist die Frage nach den metaphysischen Behauptungen, Voraussetzungen und Implikationen einer Theorie gemeint, wobei als metaphysisch solche Behauptungen, Voraussetzungen oder Implikationen angesehen werden, die Annahmen über die Existenz von nicht physikalischen Entitäten darstellen. Die Frage nach den metaphysischen Behauptungen, Voraussetzungen und Implikationen der Philosophie Kants und der Deutschen Idealisten hat in den letzten Jahrzehnten kaum jemand so nachdrücklich diskutiert wie Rolf Horstmann. Er hat sie beispielsweise zur Leitfrage gemacht, um einen Unterschied zwischen Kant und seinen Nachfolgern geltend zu machen,¹ und er hat sie gegen bestimmte analytische Interpretationen Hegels ins Spiel gebracht, denen er sich zugleich immer sehr verbunden fühlte.² Mit einem Beitrag von Rolf Horstmann zu der Frage, welche Rolle dem Thema „Selbstbewusstsein“ bei der Begründung der Kategorien durch Kant und Hegel zukommt, beginnt dieser Band. Die meisten der nachfolgenden Beiträge beziehen sich auf Horstmanns Thesen zur metaphysischen Ausrichtung Kants und der Deutschen Idealisten. Sie tun dies, indem sie die Themen Selbstbewusstsein, Idealismus, Au assung der Wirklichkeit und ästhetische Erfahrung behandeln. Die ersten sechs Beiträge beschä igen sich mit dem Thema Selbstbewusstsein und Erkenntnis. Dabei spielt Kants Idealismus eine besondere Rolle. Rolf Horstmanns Überlegungen stehen im Zusammenhang mit seiner Theorie des Selbstbewusstseins, der zufolge Selbstbewusstsein gemeinsam mit sogenannten bewussten propositionalen Akten au ritt. Horstmann geht der Frage nach, wie Kant auf der einen und Hegel auf der anderen Seite diese propositionalen Akte genauer verstehen. Es wird so auf eine neue Weise gezeigt, welche logischen Strukturen nach Kant und Hegel die Welt strukturieren, von welcher Art diese logischen Strukturen

1 Vgl. Horstmann, Rolf-Peter. 1997. „Hegels Kritik der Kantischen Kategorien“. In: ders. Bausteine kritischer Philosophie. Berlin (Philo-Verlag), 181–200. 2 Vgl. Horstmann, Rolf-Peter. 2004. „Das Endliche und das Unendliche in Hegels Denken“. In: Menegoni, F., Illetterati, L. (Hrsg.). Das Endliche und das Unendliche in Hegels Denken. Stuttgart (Klett-Cotta), 83–102.

14

Emundts

sind und was das für das Thema Selbstbewusstsein bedeutet. Der Beitrag von Barry Stroud entwickelt in Auseinandersetzung mit Horstmanns Thesen zum Selbstbewusstsein eine alternative Theorie des Selbstbewusstseins. Diese alternative Theorie ist nicht auf Kants Idealismus verp ichtet. Sie nimmt vielmehr die Punkte von Kants Theorie auf, die auch unabhängig von einer idealistischen Konzeption bestehen können. Mein Beitrag schließt hier an, indem ich diskutiere, was nach Kant „Selbstbewusstsein“ heißt. Entscheidend ist dabei für mich, dass Kant seine Theorie des Selbstbewusstseins im Zusammenhang seiner idealistischen Philosophie entwickelt. Auch Wolfgang Carls Aufsatz schließt hier an. Er behandelt Kants Kopernikanische Wende. Anhand einer Analyse des Formbegri s wird Kants Au assung von Philosophie als Metaphysik expliziert. Seine Interpretation Kants grenzt Carl von der Interpretation ab, die Strouds kritischer Auseinandersetzung mit Kants Idealismus seines Erachtens zugrunde liegt. Im darauf folgenden Beitrag behandelt Béatrice Longuenesse Kants und Hegels Konzeption eines moralischen Selbst. Longuenesse fragt, welche metaphysischen Voraussetzungen Kant und Hegel in Anspruch nehmen müssen, um ihre Konzeptionen eines moralischen Selbst vertreten zu können. Der Aufsatz von Hannah Ginsborg diskutiert einen zentralen Aspekt von Kants Theorie des Selbstbewusstseins, nämlich die Frage, in welchem Verhältnis das transzendentale Selbstbewusstsein zum empirischen Selbstbewusstsein steht. Die vier folgenden Beiträge befassen sich mit dem Thema der Möglichkeit von Erkenntnis. In unterschiedlicher Weise steht hier die Frage nach den Grenzen der Erkenntnis im Mittelpunkt. Der Aufsatz von Stefanie Grüne schließt thematisch an den Beitrag von Ginsborg an, denn es geht bei Grüne um Kants Begri von Spontaneität und um das spannungsreiche Verhältnis zwischen Kants Thesen, dass wir auf der einen Seite spontan denkende Wesen sind, auf der anderen Seite aber auch als Wesen angesehen werden müssen, deren Handlungen kausal verursacht sind. Der Tätigkeitsaspekt unseres Denkens steht auch bei Ulrich Schlösser im Vordergrund. Er verteidigt Kants Idee einer Handlung oder Aktivität des Verstandes, indem er die verschiedenen Aktivitäten des Verstandes in der Begri sund Urteilsbildung voneinander unterscheidet, ihre Funktion und ihren Zusammenhang beleuchtet und auf diese Weise auch versucht, Einwänden gegen eine Kantische Au assung zu begegnen. Der Beitrag von Eckart Förster behandelt Kants Thesen zur Grenze der Erkenntnis. Was hat Kant dazu veranlasst, anders als beispielsweise Fichte, eine intellektuelle Anschauung für unmöglich zu halten? Die Antwort auf diese Frage wird durch eine Analyse von Kants Au assung inkongruenter Gegenstücke zu verschiedenen Zeiten seines Lebens gegeben. Das Thema der Grenzen der Erkenntnis wird im Beitrag von Tobias Rosefeldt abermals aufgenommen, indem Kants Theorie der Dinge an sich aus der Perspektive behandelt wird, die gewichtige Einwände

Einleitung

15

der idealistischen Nachfolger Kants gegen das Ding an sich berücksichtigt. Diese Geschichte der früheren Rezeption von Kants Theorie des Dings an sich verbindet Rosefeldt mit einer Interpretation dieser Kantischen Lehre, bei der diese Einwände nicht geltend gemacht werden können. Andrew Chignell und Paul Guyer widmen sich in ihren Beiträgen den ästhetischen Konzeptionen von Kant und Hegel. Chignell beschä igt sich mit Kants Theorie des ästhetischen Urteils und analysiert es bezüglich seiner Voraussetzungen und Implikationen. Guyer bestimmt, welche Bedeutung der Begri Geist für Hegels Ästhetik hat, und versucht aufzudecken, welche Rolle dieser Begri in ästhetischen Kontexten spielt. Die leitende Frage ist, welche ontologischen Implikationen die Konzeption des Geistes bei Hegel hat. Die Aufsätze von Anton Koch und Gary Hat eld untersuchen die kritische Fortsetzung von Kant und Hegel bis ins 20. Jahrhundert. Koch entwir eine Skizze von Hegels System als Metaphysik, um von hier aus zu fragen, in welchem Sinne diese Philosophie von den nachfolgenden Philosophen, insbesondere von Heidegger, fortgesetzt bzw. überwunden worden ist. Hat eld gibt eine detaillierte Analyse von Russells Raumau assung und diskutiert deren Beziehung zur Kantischen Raumvorstellung vor dem Hintergrund einer direkt realistischen Au assung von Wahrnehmung. Hierbei spielt unter anderem der Bezug auf empirisches Wissen eine entscheidende Rolle. Auf diese Weise ist insgesamt in verschiedenen Hinsichten die Frage nach dem metaphysischen Gehalt der Philosophie von Kant und den verschiedenen Kant-Nachfolgern thematisiert worden. Das vorliegende Buch ist Rolf Horstmann gewidmet. Die hier versammelten Beiträge gehen auf eine Tagung zurück, die ich im Sommer 2011 veranstaltet habe, um Rolf Horstmann für seine philosophische Arbeit zu ehren und ihm zu danken. Zu Rolf Horstmanns vielen Verdiensten gehört, dass er in Berlin ein Kolloquium initiiert hat, das zu einem regen Austausch zwischen Philosophen aller Richtungen aus aller Welt, die zu Themen Kants und der Klassischen Deutschen Philosophie arbeiten, geführt hat. Von diesem Austausch kann der vorliegende Band ho entlich auch einen Eindruck vermitteln. Als Freund und/oder Lehrer hat Rolf Horstmann einen großen Ein uss auf alle hier versammelten Autoren. Mein Dank gilt allen, die an der Tagung in der einen oder anderen Weise mitgewirkt haben. Besonders danken möchte ich den in diesem Buch vertretenen Autoren. Danken möchte ich außerdem meinen Mitarbeitern Jochen Briesen und Amber Gri oen für das sorgfältige Redigieren der Aufsätze. Anders Landig und Wolfgang Scha arzyk danke ich für die Formatierung der Texte und Anne Mone Sahnwaldt für die Betreuung des Buches und das Herstellen des Manuskripts. Der FritzThyssen-Sti ung danke ich für die großzügige Unterstützung der Tagung und für den Druckkostenzuschuss zum vorliegenden Band.

Rolf-Peter Horstmann

Propositional Activity in Kant and Hegel* The view put forward here is meant to be an attempt to think about self-consciousness or ‘the I’ in the spirit of the tradition of classical German philosophy (Kant and German idealism) and continental Phenomenology (Husserl to Merleau-Ponty). In this tradition self-consciousness is addressed not just as a characteristic of a subject of conscious mental states in virtue of which it (the subject) can become an object of awareness but also as a radically or irreversibly subjective, non-objecti able phenomenon that is intimately linked to an activity which is in charge of providing propositionally-structured content for mental states. According to this approach one has to think of the radically subjective self-conscious I not as an isolated and self-contained phenomenon but as an essential element within the special context of conscious propositional states: one has to take the self-conscious I to be the result of an activity of a conscious subject that is constitutive of a constellation in which both the self-conscious I and a propositional content is established. Basically following this approach, the view presented here is quite simple, and its major claims can be summarized thus: a conscious subject, i.e., a subject that is in the position to be aware of its mental states, happens to experience consciously all sorts of mental states, some of which are propositional states called ‘thoughts’. It is because of these propositional states or ‘thoughts’ that we have to acknowledge the existence of an irreversibly subjective, non-objecti able self-conscious I. Both conscious propositional states and the non-objecti able (irreversibly subjective) self-conscious I are claimed to be essential manifestations of a (synthetic) activity in which self-conscious subjects engage and, which exhausts itself in making available both propositionally-structured content and the self-conscious I to the experiencing subject simultaneously. Because this view places so much weight on propositional contexts let us call it ‘the propositional view’, and let us name this activity ‘the propositional activity’,¹ the states it gives rise to ‘propositional states’ or ‘thoughts’, and the selfconscious I connected with these states ‘the propositional I’ or ‘propositional self-

* This text is a sequel to a paper published a while ago (“The Limited Signi cance of Self-Consciousness”. In: Revue de Metaphysique et de Morale 4, 2010, 435–454). That paper introduces and tries to justify the distinctions whose meaning and implications are elaborated here. Before dismissing the following remarks as based on unsubstantiated assumptions the reader might like to check the earlier text rst. 1 These terms have a tradition, see the article “Proposition”. In: P. Edwards (Ed.): The Encyclopedia of Philosophy. New York (Macmillan) 1967.

18

Horstmann

consciousness’ in order to distinguish it from the I as an object of awareness. If one is to give some credibility to such a propositional view one has to pursue the model underlying it from quite a number of di erent directions. Among them, the most obvious is to clarify the concept of a propositional activity and to elaborate its speci c characteristics. Because this activity is supposed to provide conscious propositional content, one also has to give an account of how this is done. And because this activity is also credited with bringing about the propositional I, one has to focus on questions concerning its features and function and to address some problems connected with it. While all these aspects have to be attended to, in what follows what I want to deal with is primarily the question of how to conceive of what has been called here ‘propositional activity’ within this propositional approach to self-consciousness because this activity is crucial for the occurrence of conscious propositional states. To start with the obvious: if there is such an activity at all there must be a subject to which it can be attributed. What are the conditions a subject has to satisfy in order to qualify for being a subject of a propositional activity? (1) The subject has to be a conscious being because this activity is supposed to account for conscious propositional states of the subject. But consciousness is not enough for this activity to occur because consciousness in itself as conceived of here lacks propositional structure: consciousness is just the basic and unarticulated mode of awareness a subject has of its (inner and outer) environment.² (2) The conscious subject of this propositional activity must also have the ability to deal with the content it is aware of in such a way that it can transform it into propositionallystructured content and at the same time to form its own irreversibly subjective, i.e., propositional I.³ A er all, propositional structure and, at the same time, propositional self-consciousness are supposed to be the characteristic achievements of this activity, and thus the subject has to be able to accommodate them. Although these two requirements sound rather trivial because they just rephrase what is meant by ‘the subject’ in terms of the objectives of the propositional

2 The conception of consciousness relied on here is very well expressed by J. Searle in his article on “Biological Naturalism”. In: S. Schneider, et.al. (Eds.): The Blackwell Companion to Consciousness. Malden, MA (Blackwell Pub.) 2007, though I believe that even in dreamless sleep one is conscious in a rudimentary way. 3 A small note for the purpose of clari cation: One might get the impression that this means that one has to distinguish between two subjects which are numerically distinct, i.e., between the conscious subject on the one hand and the propositional I on the other. This impression is misleading. There is just one single subject. Rather one should think of the propositional I as the result of the (self-)positing act of the conscious subject as soon as it is propositionally active. (See the remarks on identity at the end of this text.)

Propositional Activity in Kant and Hegel

19

activity, they emphasize the important point that subjects of a propositional activity have to be self-conscious entities. This is so because a subject engaged in propositional activity has to be an entity that is capable of having conscious propositional states. These states in turn presuppose, according to the conception supported here, a subject that can relate to propositionally-structured content. This relating again is dependent on the subject’s ability to self-ascribe these states because otherwise the subject could not recognize these states as its own. Now, these states necessarily involve a propositional I (in the radically subjective sense explained above) because in the end propositional self-consciousness is nothing but a relating condition. This line of reasoning leads to a surprising defense of the view deeply rooted in our everyday conceptions of the world that non-human animals or computers do not meet the requirements for counting as subjects of propositional activity: this is so not primarily because of their lack of self-consciousness but because of their inability to be conscious of propositionally-structured content and thus to have propositional states. This result may have some bene cial implications, two of which should be mentioned here: the rst is that it might give rise to the hope that all discussions as to whether non-human animals and machines can have self-consciousness turn out to be pointless as long as they focus on other criteria than susceptibility to propositional content. The second is that it could give some plausibility to the claim that although there might be many creatures that have consciousness, not all of them have to have self-consciousness: it is self-consciousness (and not consciousness alone) which makes a creature an ‘animal propositionale’, and not every conscious creature has to be such an animal. It also might be worth mentioning that the approach outlined here is completely compatible with a naturalist or a realist view with respect to mental (in this case: propositional) states because it relies on consciousness as a necessary condition of everything mental and does not exclude the conviction that in the end consciousness is a natural phenomenon. But to know what a subject of a propositional activity must look like does not help us much to nd out how one has to conceive of that activity itself. To be sure, it has to be taken to be a mental activity of a conscious subject. But how does it operate? Is it a creative activity like imagining or rather an assimilating activity like translating? In order to nd answers to such questions one is well advised to take a look at the history of philosophy because it is here where suggestions have been discussed abundantly. Within the period I am most familiar with there are at least two suggestions which come to mind almost immediately, one of which can be vaguely traced to Kant (and in a certain way to Fichte), the other to Hegel. Although I am neither willing nor in the position to give an exhaustive account of their respective suggestions, I will use their ideas as a guideline for exploring the

20

Horstmann

characteristics and peculiarities of propositional activity more closely, especially as regards the speci c task the propositional activity is supposed to perform. As far as Kant is concerned, he seems to have thought of the propositional activity as consisting in judging.⁴ According to him it is by judging that what he calls ‘the understanding’ actively brings propositional structure into whatever manifold of spatio-temporal data the subject is consciously confronted with by ordering these data, which are provided by the senses according to special concepts in speci c, ‘objective’ ways. Backing away somewhat from Kant’s terminology it looks as if he endorses an argument along the following lines: in order for us to think of the world as a world of experience, and that means as a world of which empirical knowledge is possible, one has to conceive of the subject of experience as a conscious subject that is in the position to form judgments about spatio-temporal objects. This makes the availability of the concept of an object (Objekt überhaupt, cf. CpR 126)⁵ a necessary condition of experience. Now, to form a judgment or to judge means to employ the ability to actively give some conceptual structure to a representational content, i.e., a content the subject is aware (conscious) of, which makes it possible to think of this content as an object. This is done by uniting given representations in such a way that they conform to the rules which are constitutive of the very concept of an object. These rules become known to us by looking at the di erent ways in which judgments provide unity to conceptual content. Because for Kant there are exactly nine (or twelve) of these ways the number of the objectconstituting rules is also strictly limited to nine (or twelve). It is this rule-governed operation of judging which according to Kant not just gives rise to the judgment but also and more importantly to the concept of something that can count as an object (cf. B 138). Therefore the act of judging is both responsible for the resulting judgment and for the fact that we take it (the judgment) to be about an object.⁶

4 The most illuminating analysis of Kant’s theory of judgment is the book by B. Longuenesse: Kant and the Capacity to Judge. Princeton (Princeton University Press) 1998. 5 The Critique of Pure Reason (CpR) is quoted according to the original pagination of the rst (A) and the second (B) edition. Other writings by Kant are quoted following the Akademie edition (Kant’s gesammelte Schri en. Hg. von der Koeniglich Preussischen Akademie der Wissenscha en. Berlin (G. Reimer) 1900 .) by volume and page number. 6 Because of this object-constituting function of judgmental acts Kant thinks it necessary to supplement what he calls ‘general logic’, which only deals with conceptual structure, with his ‘transcendental logic’, which rst and foremost provides us with the concept of an object. Note that the object the judgment is about does not have to be an object of knowledge but could as well be an object of thinking. In order to become an object of knowledge representational content not only has to conform to rules based on the rules of judgment but has also to obey the rules of sensibility. Objects of thinking are not subject to the latter rules. Concerning the distinction between objects of knowledge and objects of thinking, cf. CpR, B 166 n.

Propositional Activity in Kant and Hegel

21

However, this object-constituting function of the propositional activity of judging can, according to Kant, only be accounted for if one accepts the idea of a radically subjective self-conscious I, i.e., a propositional I, as necessarily involved in judging. To focus merely on the object-constituting function is to concentrate on only one aspect which is of importance with respect to the activity of judging. Another aspect has to do with the conditions under which this activity can be taken to result in a judgment about an object. It is here where the concept of the propositional I starts to play a role. This is so because judging means to connect actively or to synthesize representations into the unity of an object. Such an ‘objective’ unity does not come about by merely employing the categorical rules necessary to constitute the representation of an object.⁷ The representation of unity required to get us to the unity of an object has to be provided by a special instance which makes the very concept of unity available to us. Thus the very act of object-constitution by means of judging depends on the existence of such a unity-providing instance. Because this instance cannot be the object itself – a er all, it is constituted by the activity of judging – it has to be something the judging activity brings about in order to endow the object with this characteristic. In other words: Synthesizing a given representational manifold into the unity of an object presupposes an instance which can function as the center of uni cation or can be seen as the provider of unity. This instance is part of the act of judging. The subject of this act has to be able to be conscious of itself as being this center of uni cation since uni cation is a subject-dependant phenomenon: the concept of unity would be empty if there were no subject for whom a manifold is united, and if this subject were not self-conscious there would be no unity for it. If judging consists in actively bringing together divergent representational content into a propositional unity of which the subject is conscious then this subject must also have the unifying resources of a propositional I, for otherwise a conscious unifying center would be missing. This is the reason Kant claims that the propositional activity of judging not only is constitutive of the concept of an object but also has the function of supplying to what he calls “the objective unity of apperception” with content by providing it with a propositional I as the radically subjective center (a subject that can never become an object) of the propositional content (B 135, 141).

7 In § 15 of the second version of the Transcendental Deduction (CpR, B 131), Kant makes it quite clear that the unity we need in order to account for the unity of an object cannot be identi ed with the categorical rule of unity.

22

Horstmann

Thus in following a Kantian approach and determining the propositional activity as judging one is led to a conception according to which the individual act of judging has a double result: on the one hand it produces the judgment, i.e., a propositionally-structured item which is about something, and on the other it gives rise at the very same instant to the propositional I as an accompanying representation (Begleitvorstellung) that provides unity. Both of these results depend on one another: there is no judgment without the propositional I and the other way round. This Kantian version of how to make sense of the idea of propositional activity is quite compelling because it permits us to explain the occurrence of propositional states which are about objects in space and time in a quasi-genealogical manner. This explanation is based on merely three presuppositions: (1) that there indeed is an initial awareness situation in which a subject is conscious of a propositionally-unstructured content (Kant’s given manifold of sensibility), (2) that this content is structurable if it has spatio-temporal determinations, i.e., is subject to the conditions of sensibility, and (3) that the subject has the ability to transform structurable but as yet unstructured content into propositionally-structured content by performing the act of judging or, in other words, that the subject is in the possession of the propositional activity of judging. The story then to be told can be quite short and straightforward if one restricts propositional activity to judging: Whenever a subject is embedded in an environment it is aware (conscious) of, it can transform (parts of) this environment into propositional structures of which it is possible to maintain that they are the case or not the case. This is done by the act of judging. As soon as a subject judges, it establishes both a propositionallystructured content that functions as the content of the judgment and a propositional I that plays the role of the unifying center (Kant’s ‘logical subject’) of that judgment. Thus the occurrence of a conscious propositional state which is about something or which has objective signi cance can only be accounted for by paying attention to the twofold performance of the propositional activity at work in judgmental acts.⁸ This rather sketchy version of Kant’s story of what is necessary if one wants to account for propositional states can be expanded in di erent directions, though only one of these directions is pursued by Kant himself. This has to do with his

8 This conviction is the reason for Kant’s claims in the well-known footnote in the preface to the Metaphysical Foundations of Natural Science that almost solely “from the precisely determined de nition of a judgment in general (an action through which given representations rst become cognitions of an object)” (AA IV, 475 f.) one can infer how experience is possible. Cf. also Prolegomena, § 39: judging is “the act of the understanding that contains all the rest…” (AA IV, 323).

Propositional Activity in Kant and Hegel

23

being mainly interested not in the occurrence of propositional states as such but (at least in his rst Critique) in criticizing traditional metaphysics from an epistemological point of view. For him his theory of judgment is one of the means (the other is his theory of sensible intuition) of showing that knowledge of the alleged objects of traditional metaphysics is not possible – not because, according to this theory, these objects do not meet the conditions necessary for a judgment, but because they do not meet the sensibility conditions (cf. note 7 above), an assessment which amounts to the claim that they are not (cognizable) objects at all. However, it is not this critical claim which is the most remarkable outcome of Kant’s conception of judging. What is much more important is the picture that emerges as to how to conceive of the possibility of an epistemically accessible world, i.e., a world we can think of as an object of knowledge. This picture makes it understandable why and in which sense Kant thought of himself as an idealist both with respect to the world as an object of knowledge and (as is quite o en disregarded) with respect to the propositional I. According to this picture it is the employment of the propositional activity in the act of judging which transforms whatever can be taken to be ‘given’ to a conscious subject into a structured unity of which it is possible to have empirical knowledge, i.e., experience. Because this structure is supposed to be rooted in and governed by valid forms of judgment and because these forms are the result of operations of the ‘mind’, the world as an object of knowledge turns out to be mind-dependent, consisting of ‘ideal’ constructions which constitute what is objectively real for us. And because the act of judging cannot take place without bringing into being a propositional I as its focal point which provides the unity necessary for the representation of an object, this I also has the status of an ‘ideal’ entity, a thoroughly mental product intimately connected with this propositional activity. However, Kant’s ingenious and daring attempt to explain the propositional activity in terms of judging has its price. This has not that much to do with his initial idea to make judging the source of both the objective world and the irreversibly subjective propositional I but with the way Kant executes this conception within his broader epistemological framework. Two of the most challenging obstacles should be mentioned here. The rst is connected with Kant’s implementation of his conception of judging in terms of judgmental forms. According to Kant, not every combination of concepts (even if grammatically correct) yields a judgment. In order to meet the criteria for being a judgment an arrangement of concepts must ful ll certain formal requirements. These formal requirements are codi ed in the list of what Kant regards as the legitimate forms of judgment. Because these forms are based on unifying acts (Funktionen, or ‘functions’) of the propositional activity (of the understanding) which result in judgments, and because these acts are considered to establish the concept of an object, what counts as an object is at least

24

Horstmann

indirectly determined by these forms. From early on quite a number of complaints have been raised against Kant’s list of these forms. They range from the charge of arbitrariness (starting with Reinhold) to the suspicion of incompleteness (going back to Hegel). And although it is still under discussion whether these complaints are indeed justi ed (cf. K. Reich, M. Wol , B. Longuenesse, P. Guyer, H. Allison etc.), they are serious enough to cast doubt on the viability of Kant’s conception of judging. However, the most disturbing implication of his list of judgmental forms comes to the fore when brought together with what Kant takes to be universal conditions of sensibility: this list together with these conditions seems to commit Kant to the claim that the realm of objects which can count as objectively real and thus as genuine elements of our epistemic environment is restricted to the domain of physical objects. This commitment becomes most obvious as soon as one looks at Kant’s treatment of causality. According to his view causality is an object-constituting rule based on the unifying function of the understanding in hypothetical judgments. This rule makes it obligatory that everything that quali es for the status of an objective epistemic item has to have a causal history. In order to meet this requirement something has to t into the general causal chain as speci ed by the causal laws of nature, which in turn are restricted to physical objects. This restriction, so the objection goes, not only seems to be intuitively questionable but even from a Kantian point of view is not without problems as documented, e.g., by Kant’s moral philosophy. This is so because the restriction to physical objects seems to reduce the world of which experience (empirical knowledge) is possible in a much too radical way and to diminish in a most questionable manner the domain of what can count as real in a full-blooded sense. A er all, this restriction excludes from our epistemically accessible world quite a lot of entities (Sachverhalte) that we are inclined to believe we are objectively acquainted with and of which we trust to have some sort of empirical knowledge. Examples might be social entities (e.g., the family), economic constructions (e.g., money) or cultural products (e.g., movies), not to mention aesthetic items (e.g., works of art) or political institutions (e.g., the state). According to Kant, it seems, it is just their physical manifestations that we are entitled to think of as ‘real’ objects, whereas with respect to their non-physical characteristics they have the status of subjective ctions.⁹

9 Kant is well aware of these limitations with respect to the domain of epistemically accessible objects. His way of dealing with natural ends, i.e., organisms, and aesthetic phenomena in the third Critique indicates how he wants to overcome the problems connected with these limitations. Cf. R.-P. Horstmann: “Kant and the Problem of Purposiveness, or how to deal with Organisms (and Empirical Laws and Beauty) in an Idealistic Framework”. (Unpublished Ms.).

Propositional Activity in Kant and Hegel

25

(2) Another considerable problem may be seen in Kant’s rather frugal model of the radically subjective propositional I, a model apparently designed with the sole purpose of matching the equally frugal characterization of an object as a unied collection of spatio-temporal elements. Kant seems to have been content with thinking that what he calls the necessary correlate of a judgment, i.e., a propositional I, is su ciently determined by what he takes to be an inevitable condition in order to account for his very elementary formal concept of an object as a spatio-temporal unity established in the act of judging. Because judging is inseparably connected with the occurrence of a propositional I as a focal point (which makes it a necessary condition for the very possibility of the concept of an object in general), and because both the ideas of spatial unity (togetherness) and temporal unity (identity) presuppose a propositional I as their provider (cf. B 136 n.), it seems to be su cient to characterize the propositional I by means of its unityproviding function and by nothing else. However, if it turns out that Kant’s formal concept of an object as a physical unity is not su cient to cover all sorts of objects established in acts of judging then his characterization of the propositional I as restricted to the role of unity provider could be considered to be too poor to do justice to its role in performing propositional acts (of judging).¹⁰ Problems like these have already been raised by some of Kant’s contemporaries. They may indeed provide a motive not to follow Kant’s ‘judgmental’ interpretation of the propositional activity as constitutive of a propositional I and an epistemically accessible world. However, not to agree with Kant’s interpreta-

10 Actually, Hegel is of the opinion that ultimately such a restricted conception of the propositional I cannot make sense even of physical objects. (Cf. D. Emundts: “Hegel’s Criticism of Kant’s Concept of Physical Laws”. Forthcoming.) A third and perhaps the most challenging problem for Kant’s conception of the propositional activity understood as judging has been seen in the very counter-intuitive consequence that his approach forces him to draw a distinction between the irreversibly subjective propositional I that is necessarily involved in the process of judgment formation and the empirical self-consciousness that consists in my awareness of my inner and outer states or in my awareness of myself as an object. Kant explicitly endorses this distinction and, as is well known, leaves no doubt that is it crucial for his conception of discursive knowledge. However, it is less clear how he thinks of the relation between the two. The easiest way to conceive of this relation is to interpret it in terms of the subject-object distinction. According to such an interpretation the propositional I would be that the subject that has its appearance as an object in the shape of the empirical I. Although this reading has the advantage that it can avoid the impression that Kant has no means to bridge the gap between the propositional and the empirical I, it also has some disadvantages, e.g., it cannot do justice to Kant’s famous explicit denial that the propositional I can ever become an object. It is interesting in its own right to nd out how Hegel understood Kant’s remarks with regard to this question. This cannot be done here. However, I suspect that Hegel also favors a reading that wants to somehow unify the propositional and the empirical I. The following outline of Hegel’s criticism of Kant is based on this assumption.

26

Horstmann

tion does not automatically mean to be prepared to present a viable alternative. Rather, a super cial survey of the last 200 years of philosophy might lead to the impression that there were no alternatives seriously explored at all or at least no very radical ones.¹¹ The only distinguished exception in my eyes is Hegel’s “onetime experiment”, in which he attempts to explain the constitutive achievements of the propositional activity thought to be fundamental to the possibility of an epistemic environment – not (like Kant) by judging but by what could be named rather awkwardly ‘conceptual grasping (begri ich erfassen) under the guidance of experience’. In order to avoid this awkward phrase I will use the term ‘conceptual grasping’ when discussing Hegel’s account of the workings of the propositional activity in what follows. I call his interpretation a “one-time experiment” because his theory of conceptual grasping is documented just once in his writings, namely in his Phenomenology of Spirit, and even there it is not that easily to identify because it is mixed up with considerations which are due to other concerns Hegel is occupied with in the Phenomenology. It might even be doubted that conceptual grasping is understood by Hegel primarily as an epistemological notion. Although I think that this is the case, it is of no importance here. For the present purpose it is enough if it is conceded that Hegel’s conception of conceptual grasping can have an epistemological function. As far as I know there are no serious doubts about that.¹² Presumably Hegel’s approach to questions concerning self-consciousness and objectivity and his presentation of conceptual grasping as an alternative interpretation of the achievements and the function of the propositional activity is encouraged and motivated by the awareness of what he thought to be shortcomings of Kant’s position. However, he is by no means a radical critic of Kant’s basic epistemological assumptions. On the contrary, he shares quite a number of them. He agrees with Kant that we have to think of the epistemically accessible world as conceptually structured and that there is an activity that is responsible for our being

11 This does not mean that there are no interesting suggestions from the last 200 years. In contemporary philosophy the most ambitious and thoughtful example of an attempt to tackle the question of the constitution of an epistemically accessible world I know of is the attempt of B. Stroud to introduce what he calls ‘propositional perception’ as the basic epistemic activity. (For a recent programmatic statement of his view, cf. his exchange with C. Cassam in European Journal of Philosophy 17, 2009, 559–570, esp. 595 f.). But he too seems to be committed to a broadly Kantian framework in that his propositional perception is intimately linked to Kant’s conception of judgment. 12 An in-depth analysis of Hegel’s phenomenological theory of experience from an epistemological perspective is provided by a comprehensive study by D. Emundts: Erfahren und Erkennen. Hegels Theorie der Wirklichkeit. Frankfurt (Klostermann) 2012. The following remarks owe much to her investigation.

Propositional Activity in Kant and Hegel

27

able to have access to this structure and thus to have access to propositional content. He also agrees with Kant that this propositional activity is necessary in order to have both an epistemic object, i.e., an object the propositional content is about, and a self-conscious I¹³ that functions as the epistemic subject of a propositional state. And, most importantly, he even agrees with Kant that there is a relation of mutual dependence between the epistemic object and the self-conscious I: there is no epistemic object without a self-conscious I and vice versa. However, Hegel disagrees with Kant on two essential points: (1) he does not believe in unstructured or non-propositional content, i.e., in Kant’s ‘given sensible manifold’, and hence he does not believe that in order to account for knowledge one has to introduce processes of transformation of non-propositional into propositional content. (2) He does not think of the propositional activity as a creative capacity of a subject with respect both to the constitution of epistemic objects according to categorical rules derived from judging and of the self-conscious I as a provider of unity within the boundaries of these rules. He rather takes this activity to be one of grasping a propositionally-structured content (begri ich erfassen). And he takes the activity of conceptual grasping to be intimately connected with the presence of a self-conscious I which not only provides unity for physical (and mathematical) objects but can also cope with other forms of reality. As far as epistemic objects are concerned Hegel complains that in Kant’s theory of knowledge the rules governing the propositional activity responsible for our having any epistemic objects at all are restricted in a very unconvincing way, in that they allow only physical entities which are causally connected to be objective parts of an epistemically accessible world. He is also rather critical of Kant’s view that the categorical rules which constitute epistemic objects are purely subjective rules. And as far as the I is concerned Hegel challenges the austere, judgment-oriented conception Kant presents because he is convinced that the I is much richer than Kant wants it to be and plays a unifying role within sets of rules which cannot be reduced to judging. Reasons for such a disagreement with Kant with regard to objects and the I are not that di cult to nd. Even from an everyday point of view one might

13 In the case of Hegel I will use the somewhat awkward term ‘self-conscious I’ as a substitute for what I have called the ‘propositional I’. Though I take it that Hegel’s self-conscious I has within his conception of knowledge the same function as Kant’s propositional I has for his theory of judging, i.e., the function of designating the unifying center of propositional content, I believe at the same time that Hegel does not want to restrict this unifying center to the activity of judging as Kant does but rather that he wishes it to be the subject of all activities (among them most notably acting) involved in what for him are epistemic processes. The terminological switch from ‘propositional I’ to ‘self-conscious I’ is meant to indicate that Hegel might rely on a concept of a unifying center which could turn out to be incompatible with Kant’s.

28

Horstmann

feel uncomfortable with a position which claims (1) that from an epistemic perspective the world in which we live is ‘objectively’ nothing but a bunch of physical objects (and events) constructed by us out of subjective rules and (2) that the selfconscious I is nothing but an empty provider of unity. This feeling might become even stronger if one has to acknowledge that (at least for Kant) the main reason for these claims is that only these objects and such an I can be accounted for by the mechanisms of our propositional activity if this activity is to be identi ed with judging. Hegel’s disagreement with Kant on these topics seems to be based on the conviction that one has to think of the epistemically accessible world and the features of the self-conscious I in a much more sophisticated way. Puzzled by Kant’s restriction of objectivity to the world of (Kantian, i.e., causally-mechanically explainable) physical objects and events, he wishes to proceed on the assumption that the world looked at as the totality of epistemic objects, i.e., of objects of which we can have knowledge, contains a lot more ‘objectively real’ items than those Kant tolerates. He wants, e.g., organisms and persons (self-conscious beings) as well as social institutions (corporations, societies, states) and cultural phenomena (works of art, religion) to be objective elements of the world, not just in virtue of their being physical objects but also because of their intrinsic constitution. And he wants a conception of the self-conscious I that can deal with those objects in such a way that their claim to objectivity is justi ed. In short, what Hegel is a er is what could be called a defense of common-sense (though perhaps not in G. E. Moore’s spirit). But – and here things start to become a bit complicated – this defense is meant to be carried out within a broadly Kantian approach to objectivity and self-consciousness. What is required for such a task? Perhaps the easiest way to answer this question is to start with an outlook on what I take to be Hegel’s basic view with regard to our epistemic situation if it is to be described in allusion to something like a Kantian framework of activities and categorical structures. Here a rather sketchy version of such a description. According to Hegel the activity responsible for our being able to know things is not judging interpreted in a Kantian fashion but rather a grasping of propositional content or a conceptual grasping. This activity of grasping depends formally on its having available a non-objective, selfconscious I as its subject and materially on its having a propositionally-structured content as its object. The sum total of what a subject can grasp as conceptually structured in a rule-governed fashion delineates the epistemically accessible world of that subject. However, contrary to Kant, Hegel thinks neither of the I nor of the object as being constituted by this activity but rather thinks of both as conditions for executing it. This means (1) that, according to Hegel, propositional structure cannot be seen as the result of some subjective operation or other of a judging

Propositional Activity in Kant and Hegel

29

mind along Kantian lines but has to be taken to be founded in the very nature of the objects of knowledge themselves. And it means (2) that the self-conscious I can be seen as being grounded in something other than the activity itself (e.g., in an embodied person), thus avoiding the Kantian fate of having no real connection to anything else than to judging. If it is right that Hegel wants to give an account of knowledge and the role the self-conscious I plays based on these two claims, then at least two questions immediately arise. First, what can be said about the conceptual structure as an objective feature, i.e., an essential characteristic of the world of objects? And second, how does a subject realize that in order to know it has to grasp conceptual structure? Hegel is well aware of these questions. According to him the rst one calls for a metaphysical answer whereas the second requires an epistemological one. Up to this point Hegel’s project can be outlined in a fashion which appears to be pretty much in line with Kant’s way of thinking. However, Kantian guidelines are of no help when it comes to the speci c features of Hegel’s account of the I, of knowledge and of the object. This is especially the case when it comes to the rst question just mentioned, namely the question as to what can function as categorical, or object-constituting, conceptual structures and how they come about. According to Hegel it is not the logico-linguistic practice of judging which gives rise to conceptual structure. The problems of such an assumption are amply con rmed by the shortcomings of Kant’s theory mentioned earlier (i.e., the restriction of objects to physical and mathematical objects). Neither are these object-constituting conceptual structures founded in social practices (other than judging), though, as can be mentioned here only in passing, social practices will play an important role in accounting for our ability to grasp these structures.¹⁴ Conceptual and propositional structure is rather the result of what Hegel somewhat opaquely describes as the self-movement of the Concept (Selbstbewegung des Begri s). The question as to what is meant with this turn of phrase and how exactly Hegel conceives of

14 The idea that according to Hegel conceptual structure is founded in social practices has become quite fashionable in recent years, especially in the English-speaking world. One reason for favoring such a view might be that reliance on social practices makes it easier to integrate next to physical objects and events entities like other persons, political institutions like governments and products like movies into our epistemic environment as ‘objectively real’ (to again use a Kantian term). However, if this were a motive on Hegel’s part to introduce social practices as the basis for conceptual structure, then this motive seems at rst sight to be nothing but the expression of an unfounded conviction that there indeed is more to encounter in the epistemically accessible world that can count as a real object than just the physical. This conviction could be easily dismissed as a case of wishful thinking if there were no compelling reasons in its favor. For Hegel these reasons are to be found in his general metaphysical theory concerning the constitution of reality.

30

Horstmann

the constitution of objective conceptual structure leads directly to what could be called ‘the metaphysical basis’ of his philosophy which consists in the conviction that reality has to be seen as a self-organizing conceptual whole. This metaphysical basis is spelled out extensively in what he entitles Science of Logic. Fortunately there is no need to go into the details of his controversial metaphysics here because what is at stake right now are not the mechanisms of conceptual self-organization but Hegel’s suggestions as to why the object-constituting conceptual structures give rise to epistemic objects that can be correlated to an I or can be grasped by a subject. These suggestions, which concern epistemology, can be treated as largely independent of his metaphysics (at least as far as they are put forward within the context of the Phenomenology of Spirit). When it comes to the epistemological side of Hegel’s project, the idea of grasping conceptual or propositional structure¹⁵ comes to center stage. Because it is mainly the function of the self-conscious subject that is of interest here, the following remarks will concentrate rst on this subjective factor. Now, to grasp – that is, to know something or other – presupposes access to propositional structure by an I that is performing the act of grasping. Because there is no propositional content and thus no epistemic object for a subject without a self-conscious subject, and because the only role this subject can play in epistemic contexts is to function as the subjective correlate to an epistemic object (or, as Kant would say, to accompany it), having epistemic access to an object depends on the ability of the subject to ful ll its correlating function. This is to say that only those objects are epistemically accessible or can be objects of knowledge which exhibit a conceptual structure that the subject is in a position to grasp. If this dependence relation between a subject and an object holds, then according to Hegel what can count as an object in an epistemically accessible world is limited to accessible conceptual or propositional structure. Or, to put in Kantian terms, if it is the task of the propositional I to provide objective unity, and if conceptual structure is essential to an object of knowledge, then in order to grasp an object, i.e., to think of it as an objective unity, it must be possible for the propositional I to grasp those structural features which are characteristic of an object. This is so because Hegel rmly believes that without conceptual structure there is nothing to grasp.

15 The legitimacy of the use of the terms ‘conceptual structure’ and ‘propositional structure’ as almost interchangeable would be in need of an explanation if one were to give an exhaustive account of Hegel’s epistemological views. This cannot be done here. It must be enough just to state that for Hegel conceptual structure has to show up in the form of propositional structure if it is to be known.

Propositional Activity in Kant and Hegel

31

All this means that for Hegel a subject must be in the possession of the conceptual tools that it takes to grasp an object correctly. Now what can be said about the way in which a subject acquires these tools, which would give us access to the conceptual features characteristic or essential for an object? According to Hegel the subject has to undergo a learning process which is guided by experience. This process leads the subject to an understanding of the conceptual elements required to know the object. As is well known, Hegel presents this learning process in the Phenomenology in the form of an investigation into the experiences a subject makes when approaching an object with the aim of grasping it.¹⁶ These experiences are meant to be the result of an analysis of the expectations a subject has concerning the means su cient to grasp the essence of the object (or, in Hegel’s language, what the object really or truthfully – in Wahrheit – is). Without going into details here, some remarks on how Hegel proceeds cannot be avoided. According to Hegel, the subject starts with the assumption that in order to grasp what an object really is one has to take it as a bare particular that is just given immediately, i.e., without any conceptual ingredients involved. However, the subject experiences that in the attempt to grasp the object non-conceptually the object simply vanishes because there is nothing at hand that has a stable structure (and so does the subject that is supposed to do the non-conceptual grasping). This result is a quite compelling reason to give up this assumption and to replace it with the belief that what is needed to grasp an object truthfully is a set of basic conceptual elements which can be taken from the object itself by paying close attention to its main characteristics. This belief gives rise to a conception of the object according to which it is in the end nothing but a spatio-temporal physical thing, a substantial being endowed with qualities standing in causal and other law-like relations to other objects. Although this conception makes it possible to deal with physical objects quite successfully, the subject has to nd out that such a conception is not really satisfying because it has the experience that it cannot adequately grasp a lot of objects it intends to grasp truthfully. In being confronted with what might be called ‘epistemic resistance’ or ‘epistemic unruliness’ of the epistemic object, the subject experiences that there are many objects around whose essential features cannot be grasped exhaustively by just relying on categories su cient to determine something as a physical object. Among these objects are most notably living organisms and self-conscious beings. The subject – by now somewhat frustrated (verzweifelt) by having to once again revise his conceptual apparatus – learns that in order to grasp these di erent objects in an adequate way it has to be open to new

16 It is worth mentioning that the original title of what then becomes the Phenomenology of Spirit has been Science of the Experience of Consciousness.

32

Horstmann

conceptual resources that lie in the nature of the object and are a matter of experience. According to Hegel then, the very fact that it is via experience that a subject grasps conceptual structure and thus propositional content is already an indication that the activity responsible for being related to objects of knowledge – and this means to propositional content – is ill conceived if it is taken to be judging in Kant’s sense because Kantian judging underestimates the impact the way we experience things and situations has on propositional content-formation. According to Hegel, it is a fact of experience that we distinguish between (at least) three di erent kinds of objects in the process of propositional content-formation. These are (a) physical bodies, (b) living organisms, and (c) self-conscious beings. This di erentiation is due to the experience that we cannot account for each of them adequately in terms of the characteristics of the others. Thus a selfconscious being cannot be characterized adequately either as a physical body or a living organism – experience tells us that in each of these characterizations there is something missing. This makes experience a negative criterion for an adequate characterization. If a characterization succeeds in integrating what is essential for a speci c object in order to explain its peculiar behavior without violating the way it is experienced, then this would count as a positive criterion for an adequate description of an object. Now, if one agrees with Hegel that there are indeed different types of objects which belong in our epistemic universe, and if one accepts the general idea of the modus operandi of the propositional activity understood as grasping a conceptually-structured content, then this activity of grasping must somehow have access to those concepts which codify the categories necessary for adequate descriptions of di erent types of objects. These categories have to be fundamental concepts that capture what is essential for each type of object. For Hegel, these categories are the rule of causality, or more accurately, the rule of reciprocity (Wechselwirkung) with respect to physical objects, inner purposiveness (innere Zweckmäßigkeit) with respect to living organisms, and freedom or the end in itself (Zweck an sich selbst) with respect to self-conscious beings. According to the procedure in the Phenomenology outlined above, these categories will be acquired in the following way. It sometimes happens that when dealing with objects of knowledge in a propositional mode (i.e., when grasping an object’s conceptual structure) a subject experiences something like resistance on the part of the object to conforming to the conceptual tools available to the subject. The subject experiences that certain concepts it takes to be characteristic of the essence of the object do not succeed in capturing what it takes to be the real or true conceptual structure of what is supposed to be the object to be grasped. Thus – to use one of Hegel’s examples – in a situation where living organisms play a role as the content of the epistemic environment of a subject, the attempt to t such content into a conceptual framework that takes causal reciprocity to be

Propositional Activity in Kant and Hegel

33

the essential characteristic of an object gives rise to the experience of dissatisfaction or of disappointment because of the obstinacy of the content to be grasped to fully obey the demands placed on objects which are essentially determined by the categorical rule of causal reciprocity. This shows that causal reciprocity cannot do (full) justice to what is essential to a living organism as it is immediately present as the content of experience. In the case of living organisms this (partial) disobedience to the rule of causal reciprocity is experienced by their behavior in that they exhibit attitudes like self-movement or self-regeneration which the subject cannot get hold of by relying on causal reciprocity alone. This experience of disappointment in the face of unruly behavior of speci c objective content has to be overcome by allowing other categories to enter the conceptual framework of a subject which have their origin in the respective objective content and which can accommodate the speci c character of this content. Concerning organisms it is, as already mentioned, the concept of inner purposiveness which the object is supposed to reveal to the subject as its essence-constituting category. According to Hegel, this practice of gaining access to essence-constituting concepts in the course of having experiences with objects is not restricted to natural phenomena. Social and cultural objects enter our epistemic universe in pretty much the same way: Here, too, it is the experience of some discrepancy between the conceptual means we have at our disposal in order to grasp an object and the way that object shows itself to be which gives rise to new essence- xing concepts. These concepts might include inter-subjective relations (like recognition), social practices (like marriage), moral facts (like freedom) and legal conditions (like contracts) as well as political circumstances (like shape and organization of the government) or cultural a airs (like religious attitudes).¹⁷ All these concepts nd their justi cation as essence- xing categories in what we experience, and what we experience is dependent on what the object exhibits in terms of its conceptual structure. Thus the process of grasping in the course of having experiences turns out to be the ultimate basis not only for determining what can be epistemically real for us but also for getting insight into the true constitution of reality.¹⁸

17 It is here where social practices come into play. According to Hegel, almost all non-natural objects come into view as objects of grasping only if there is a social environment present. For a more detailed discussion of Hegel’s phenomenological views concerning objects and their conceptual constitution, cf. R.-P. Horstmann: “The Phenomenology of Spirit as a ‘Transcendentalistic’ Argument for a Monistic Ontology”. In: D. Moyar, M. Quante (Eds.): Hegel’s Phenomenology of Spirit. A Critical Guide. Cambridge (CUP) 2008, 43–62. 18 It might be of interest to note in passing that this Hegelian model of how through experience we gain insight into the constitution of epistemic objects – a model which is sketched out by Hegel in terms of a progression from the expectation of immediacy via the experience of resistance to ac-

34

Horstmann

For the purpose at hand it plays no role why Hegel was convinced of the fundamental distinctness of the three types of objects mentioned above (physical objects, living organisms, self-conscious beings). It even plays no role whether this tripartite ‘list of objects’ fully does justice to Hegel’s view as to what can count as a fundamental type of object. (Some might like to include as a fourth type socialcultural entities.) It is also of no importance here to nd out why Hegel thought of the categories mentioned (reciprocity, inner purposiveness, freedom) as appropriate tools of categorizing (though it is easy to notice and interesting in its own right that all three of these categories can be related directly to a central concept in each of Kant’s three Critiques). In the present context, the main objective of these sketchy allusions to his Phenomenology has been to bring to our attention Hegel’s attempt to show (a) why we need more conceptual resources than available from Kant’s table of judgments in order to grasp propositional content, with the aim of determining an object in such a way that it can be considered to be a real or ‘truthful’ element of the epistemically accessible world, and (b) that these resources cannot be brought into being by conceptual activities of the conscious subject alone but have to have a foundation in the essence of the object (hence his preference for the term ‘grasping conceptually = begri ich erfassen’ and his dislike of Kant’s ‘judging’). These conceptual resources, according to Hegel, are necessary in order to arrive at propositional content that can su ciently capture what we experience. If one is prepared to give some credit to what has been put forward so far regarding Hegel’s reasoning as to what grasping – considered as an activity of a subject – is all about, the next question to ask has to do with the object: Why is it that in (Hegelian) conceptual grasping we are bound to nd exactly the particular categories he is suggesting as relevant concepts for the essence of speci c types of objects? How do the objects provide us with these categories? What is their origin and their justi cation? Here again Hegel refers us to his metaphysical teachings in the Science of Logic. It is almost impossible to give a short account of the peculiarities of the so-called ‘logical’ procedure of establishing categories. Thus I will refrain from even trying to do so. But, again, such an account is not really necessary here.¹⁹ This is so because the phenomenological procedure of acquir-

quiring new conceptual items – was very much appreciated by C. S. Peirce. He not only recognizes the value of Hegel’s phenomenological model explicitly he also makes use of it and transforms it (up to a certain point) into his own theory of the universal categories of Firstness, Secondness, and Thirdness, in that he identi es Firstness with immediacy (presence), Secondness with resistence (struggle), and Thirdness with concept (representation). Cf. Lectures on Pragmatism, especially Lectures II to IV. 19 For an exposition of some of the leading ideas underlying Hegel’s ‘logical’ treatment of con-

Propositional Activity in Kant and Hegel

35

ing categories based on experience is more directly related to the epistemological task of explaining conditions of grasping. These remarks will have to su ce to give us an idea of how to conceive of the object-grasping performance of the propositional activity from within a Hegelian framework. What remains to be done is to look at the subject-constituting operation of this activity in Hegel’s phenomenological setting. Here Hegel’s central objective might be described most conveniently again in terms of Kant’s model of self-consciousness. As was pointed out before, the Kantian conception of the propositional I as a necessary element in the process of the constitution of epistemic objects represents self-consciousness as an objective unity of apperception whose sole function consists in providing unity (even with respect to the possibility of spatio-temporal determination in the case of objects of experience, cf. the three footnotes in the B-version of the transcendental deduction in the CpR, B 136, B 144, B 160) to what is immediately given. According to Hegel this view de nitely points in the right direction but has to be modi ed and expanded in such a way that it can explain the possibility of a relation of the propositional I to more items than just physical objects. This modi cation and expansion is required if one is to hold fast to Kant’s basic intuition (shared by Hegel) that the idea of a propositional or, in Hegel’s case, a self-conscious I is irresolvably connected with the propositional activity in that it is taken to be a necessary correlate to propositional content and thus also to epistemic objects. That is to say: if the rules of conceptualizing attributed to the propositional activity include not just Kantian categories but also Hegel’s preferred concepts and thus give rise to the possibility of grasping particular types of objects, and if the I is to be thought of as an integral element of such an activity which can accompany all propositional content, then the propositional activity must be conceived as having the resources to supply not only epistemic objects via Kantian categories but also to provide an I which is in the position to come to terms with the special characteristics of di erent types of epistemic objects. This is so because the conceptual rules (categories) available for determining an epistemic object determine also with respect to what content the self-conscious I can function as the radically subjective center of uni cation. Thus in a Kantian epistemic world the (Kantian) I can be correlated only to an objective propositional content if this content is xed in accordance with Kantian conceptual rules, for otherwise the content would lack objectivity. What is not subject to these rules cannot be conceived as propositional content at all and hence fails to leave room for the self-conscious I as the center for which uni ca-

cepts, cf. R.-P. Horstmann: “Substance, Subject and In nity: a Case Study of the Role of Logic in Hegel’s System”. In: K. Deligiorgi (Ed.): Hegel. New Directions. Chesham (Acumen) 2006, 65–84.

36

Horstmann

tion into an object takes place. (This is the reason that for Kant even the I as an epistemic object can only appear to us in the guise of a physical object, cf. CpR, B 157 n.) For Hegel, all this implies that as soon as there are object-revealing rules in play which bring into focus non-Kantian (physical) objects, one has to make sure that there is an I around which has more resources or which is ‘richer’ than a Kantian I in order to be able to deal with these di erently-constituted objects. In other words: if it is the task of the propositional activity to make knowledge of objects possible, and if there are di erent kinds of objects, then this activity also has to provide an I which can relate to what is speci c about these objects or (to again use Kantian terminology) which can bring under its “objective unity” the essential characteristics of these objects. (In a certain sense what is happening here is the revitalization of the old philosophical saying going back via Aristotle to Empedocles that the same is cognized only by the same – hê gnôsis tou homoiou tô homoiô.) If one shares Hegel’s epistemic universe according to which there are (at least) three di erent types of objects, this means that one has to establish a conception of the I which is responsive to those conceptual tools that are decisive in grasping objects of these types. Thus in the case of living organisms, which are taken to be characterized essentially by inner purposiveness, the self-conscious I, in order to connect to content structured in accordance with the Hegelian category of inner purposiveness, has to be able to function as the supplier of “objective unity” to items revealing this essential characteristic. Roughly the same has to be true of the other object types. This sketchy presentation of Hegel’s model of conceptual grasping is by no means a paradigm of an illuminative and convincing contribution to the question as to how the self-conscious I is supposed to function in propositional contexts. In part this might be due to obscurities in Hegel’s way of expressing his views. But mainly to blame is my inability to give a more tting exposition of his views on this topic. It also is quite obvious that many questions are le open which are of crucial importance for any account of the way the self-conscious I functions in propositional contexts if this account is meant to be attractive. The most pressing of these questions might be seen in the issue of the identity of the self-conscious I as an objectifying center, an issue dominating much of the literature on selfconsciousness. This is not the right place to deal with this issue exhaustively. It will be done at another place. However, at least two of the problems concerning self-identity connected with a conception of self-consciousness based on the operations of propositional activity as put forward here should be mentioned. The rst is best brought into view by taking up what was presented here as Kant’s perspective on judging; the second is best exposed in relation to Hegel’s model of grasping.

Propositional Activity in Kant and Hegel

37

(1) Concerning Kant’s theory of judging, a problem might be seen in the following: If the unity of an object is constituted by the act of judging and if this act is such that it has to give rise to a (radically subjective) propositional I in order to bring about this objective unity, then it seems that each object-constituting act of judging brings into being its own propositional I which is inseparably connected with the individual act. Thus the impression is hard to avoid that there are as many propositional I’s as there are acts of judging. How, then, is one to make sense of the identity of the thinking subject in di erent propositional acts? To answer this question in a satisfactory manner is beyond what can be done here. I just want to hint at Kant’s approach concerning this question in a rather super cial manner. The basic idea is not that the conditions of the identity of the epistemic subject in di erent propositional acts have to be sought in our practice of forming concepts of objects in the act of judging, i.e., (in Kant’s picture) to transform nonpropositional into propositional content, but rather that the identity of the epistemic subject is rooted in the objects constituted by exercising the propositional activity. These objects have to be such that they can form a uniform whole, which he calls ‘nature’, governed by natural laws and law-like regularities. Only with respect to what can be part of this nature can the concept of identity be applied because it is only in nature that we nd spatio-temporal objects, and without relating to space-time determinations the very concept of identity is empty. Strictly speaking, according to Kant, talk about the identity of the propositional I as the radically subjective center of conscious propositional states makes no sense because the concept of identity is not applicable to this I, be it only for the reason that the propositional I is supposed to be such that it can never become an object. Thus what is identical cannot be the propositional I.²⁰ However, the concept of identity can be used with respect to the natural bearer of the propositional activity, i.e., the living conscious subject that performs propositional acts, because this living subject always has the status of an “existence of an appearance (Dasein einer Erscheinung)” (CpR, B 157 n.), i.e., of a spatio-temporal object. This quasimetaphysical line of thought is hard to gure out exactly and might have a number of rather strange consequences which, as is well known, have had damaging results for Kant’s philosophical reputation. (2) A similar problem concerning identity of the self-conscious I can be seen when it comes to Hegel’s model of grasping with respect to di erent types or species of objects: If it is the case that each Hegelian object-type – because of its pe-

20 This reasoning is analogous to what Kant writes about the substantiality of the propositional I in an interesting note to Proposition 2 in the chapter on Mechanics in the Metaphysical Foundations of Natural Science (AA IV, 542 f.).

38

Horstmann

culiar conceptual structure – calls for a particular self-conscious I in order to be grasped adequately, and if it is the case that one and the same object can belong to di erent object-types, how is it possible to think of the self-conscious I as the same in each of the di erent conceptual relations it can have to the same object? The following example might highlight the problem: Even in Hegel’s epistemic universe a self-conscious being is at the same time a living organism and a physical object. This means, according to Hegel, that in the case of a self-conscious being the conscious subject can grasp the very same content in one of three ways. Each of these three ways is determined by a speci c set of categories and thus gives rise to three di erent self-conscious I’s. Why should one think of these I’s as identical? Here, I guess Hegel’s answer would be somewhat in line with Kant’s stance: Like Kant, he would insinuate that identity has little to do with the self-conscious I but starts to be signi cant when it comes to the constitution of the conscious subject as the bearer of propositional states. And as to the question of how to deal with categorically di erent descriptions or with an object belonging to di erent object-types, he would argue that the categories needed to conceptualize an object one way or another are interconnected in a hierarchical order in virtue of which the identity of the object under di erent object-type descriptions is guaranteed. All this points to the conclusion that both Kant and Hegel, though committed to very di erent conceptions of the procedure of the propositional activity (i.e., judging vs. grasping), do share quite a number of views when asked how to deal with the questions concerning the identity of the propositional and the self-conscious I respectively. An interesting endeavor in its own right might be to answer the question as to whether one should favor a Kantian or a Hegelian position concerning the modes of operation of the propositional activity. It is important to notice that this question does not make much sense if one takes their respective approaches to be mutually exclusive alternatives. This is so because obviously both are explanations of this activity within a shared idealistic framework, in that they both start from the assumption that the epistemically accessible world is conceptually structured and both share the belief that some sort of conceptual structure is a necessary condition for objectivity. Thus it cannot be idealism per se that gives reasons to prefer one model over the other. If idealism matters at all then it will be the kind of idealism that is brought into play. The question is rather: Is Kant’s subjective and constructive idealism more convincing than Hegel’s objective and re-constructive idealism? How to decide this question depends in the end on how one is to assess the respective methods of generating object-constituting concepts. And here, it seems to me, one can be of the opinion that Hegel’s approach has certain advantages over Kant’s program. This is not only so because Hegel – if his method works – can integrate many more types of objects into our world of knowledge but more-

Propositional Activity in Kant and Hegel

39

over because he can give within his Science of Logic an account of the way in which object-constituting concepts are generated that does not rely on the dubious basis of judgmental forms. What has been said so far about the Kantian and the Hegelian account of how to think of what was called here ‘propositional activity’, both with respect to the subjective as well as the objective side of this activity, is not meant to be exhaustive. The foregoing remarks can even less pretend to give a historically correct and systematically adequate presentation of the epistemological views of Kant and Hegel. At best they might succeed in pointing out some aspects which, in my eyes, have been at work in their respective approaches to questions concerning knowledge and objectivity. These limitations have their source in the main objectives of the paper: Starting from some observations about the epistemic role of self-consciousness the leading intention here has been to have a closer look at di erent attempts in the German idealistic tradition to come to terms with the origin and the constitution of an epistemic world, i.e., a world of which we can have knowledge, from within a framework that is based on the idea of a propositional activity as a subject- as well as object- xing activity. Obviously, the issues addressed here have been under discussion in a lot of di erent ways for over (at least) two centuries. Even though it is easy to notice that contemporary approaches to these issues, especially in the Anglo-American philosophical community, have the tendency to start from di erent assumptions than those characteristic of the positions presented here, it is not that easy to see why this is so. Given the fact that there are as yet no uncontroversial solutions around to the problems of self-consciousness and object-awareness it is justi ed to revisit a tradition which emphasizes the dynamic aspect connected with self-consciousness and object-awareness by thinking of it in terms of acts and activities.²¹

21 Many thanks to all the participants in the Berlin conference on Kant and Hegel, especially to Dina Emundts for making it all happen. Thanks also to Amber Gri oen for improving grammar and style of this paper considerably.

Barry Stroud

Judgement, Self-Consciousness, Idealism I am increasingly struck in di erent ways by the indispensability of a capacity for judgement for the possibility of the kind of thought and experience we all enjoy. The idea itself is central to the Critique of Pure Reason and stands in no further need of justi cation. Thought and experience of something or other require some way of thinking of or being aware of the thing. That requires concepts, and for Kant concepts are predicates of possible judgements. That means that only someone capable of putting something forward as true – or at least of entertaining or considering something as true or false – can have concepts and so be capable of thought and experience of the fullest human kind. I draw attention to the indispensability of ‘propositional’ thought and experience not because I think anyone would explicitly deny the distinctive character of predicational thought or the centrality of the idea of truth. But I think the necessary conditions of our thinking in those ways have not always been su ciently explored, or even appreciated, in much recent philosophy. An admirable exception is Rolf-Peter Horstmann on the necessary connection between a capacity for judgement or propositional thought and the possession of self-consciousness or self-awareness on the part of the thinking or judging subject.¹ I nd what he says illuminating and promising, and I would like to explore it further. I want to see how far one can go in making clear sense of these matters without being committed to any form of idealism. It is my impression, although he does not explicitly say so, that Horstmann thinks that is not really possible. Just as a creature can be conscious and aware of an object without having a propositional thought about it, thinkers can be aware or conscious of something without being aware of themselves as the subject who is aware of it. Horstmann agrees that consciousness of something does not in general imply self-consciousness on the part of the subject, but he thinks there must be self-consciousness on the part of a subject of ‘propositional’ states of consciousness. “A self-conscious I”, he says, seems to be a “necessary element” of such states (444). “The self-conscious I understood as the self-conscious subject is a necessary condition for being in a conscious state whose content is a proposition” (447).

1 I know of this work primarily from a long paper of his, ‘Self-Consciousness and Propositional Acts’, part of which has been published in revised form as ‘The Limited Signi cance of SelfConsciousness’ in Revue de Métaphysique and de Morale 2010 (4), 435–454 (page numbers alone in parentheses in the text refer to this publication).

42

Stroud

This necessary condition carries with it further necessary conditions. “A conscious propositional state”, Horstmann says, “consists in a mental attitude of a subject towards a propositionally structured content” (444). This means the “subject must have access to or must be able to employ conceptual resources” ordered in such a way as to form a proposition (450). “Thus to be in a conscious propositional state [as in the everyday judgement ‘This house has a blue door’] means for a subject to be conscious in one or the other of the appropriate modes (believing, wishing, etc.) that such and such (e.g., this house) is so and so (e.g., has a blue door)” (450). Furthermore, “this subject has to be conscious of itself as the subject of that state because otherwise it could not be its conscious state” (451). Horstmann explains why he thinks this further necessary condition must be ful lled. To have a propositional thought I must somehow recognize a di erence between my thought and the subject ‘I’ who thinks the thought. “In thinking, I am not identical with my thought, i.e. I am not my thought but the subject of my thought” (451). I can distinguish between me as thinker of this thought and the thought that I now think. And for me to understand and acknowledge that di erence I must be able to think about my thought and about the person who has it. And that requires both that I have some way of thinking of my thought of something’s being so and that I have some way of thinking of myself. Thinking of myself as a thinking subject or ‘I’ is what I think Horstmann means by “the selfconscious I”. He goes on to ask how this self-conscious I “comes about and what is its ontological status” (450). He thinks it “comes into being in virtue of there being a propositionally structured item which can function as its object, and that at the same time the propositionally structured item can occur as an object only if there is also a self-conscious I” (453). Now, Horstmann argues, if you accept the idea that “susceptibility to propositional structure is a result of an activity, an achievement, of the subject,” and you accept the idea that there can be such a propositional object of thought only if there is also a self-conscious I, then “the most natural move”, he says, is “to attribute the formation of the self-conscious I to the very same activity which is responsible for our responsiveness to propositional structure” (453). This “most natural move” – explaining both propositional thought and the self-conscious I who has that thought as generated or constituted by the very same activity – is what Horstmann explores further in what I believe to be still-unpublished work. In what I have seen of that, he explains that “move” as it occurs in Kant, and then in an apparently more tolerant form in Hegel. Horstmann nds serious di culties in both attempts, and gives good reasons to doubt that either attempt as it stands can provide as robust and rich a conception of the thinking subject as we need to make sense of everything human subjects can think and

Judgement, Self-Consciousness, Idealism

43

experience and do. What is more, those earlier e orts to explain human access to a world that a self-conscious I can think about as an object of knowledge apparently led at best to a “mind-dependent” world – a world that is constructed or constituted as it is only in accord with our own forms of representation. This “most natural move” appears to lead to one form or another of idealism. I would like to see to what extent we can understand self-consciousness and its importance for propositional thought while stopping short of idealism. So I want to try resisting what Horstmann thinks is “the most natural move” for explaining self-consciousness. I must say at the outset that it is not easy to understand the kind of “activity” Horstmann thinks generates self-consciousness. He thinks of it as activity that constitutes both the subject of a thought and the propositional item that is the object of that thought. It is apparently an activity of synthesizing something or other into a kind of unity, but there is a question what that activity actually is, and what it operates on. As in Kant, it is said to operate on a “manifold” of something or other, but presumably not on a “manifold” of experiences. Experiences are presumably themselves constituted or generated by just such an activity. When the activity is operating successfully, something new is said to be “constituted” by it, not simply as the result of a blind process, but apparently by some kind of agency. Somebody, or something, does it. But then, what could that agent or agency be? These are familiar, and di cult, questions. I would like to put them to one side and ask about only one apparent aspect of the “synthesizing” that is said to be essential for propositional thought and self-consciousness. It seems at the very least that if this “synthesizing” activity is something that must go on whenever anyone has a thought of something or other’s being so, then each of the alleged products of that activity must be constituted at that time too – both the propositionally-structured thought and the self-conscious I that has it. I am not sure about this. Perhaps this “synthesizing” is not to be understood as an activity or process that takes place in time. Maybe talk of “synthesis” is really only another way of saying that a person’s thought or experience must be uni ed in a certain way in order for it to be the thought or experience it is. That would say nothing about an activity, and so nothing about generating or constituting a self-conscious I either. Horstmann does call what he has in mind an activity. That is the assumption from which the “natural move” starts: that the same activity generates or constitutes both the uni ed structure of a propositional thought and the self-conscious I that has that thought. And some of the things he says suggest that he does think of this object- and subject- “constituting activity” as something that goes on, something that happens. He says that when he is walking along the street and sees the house with the blue door, “this judgment ‘This house has a blue door’[…] cannot

44

Stroud

occur without my being conscious of me as its subject” (447). That is presumably because at that very time the activity of “constituting” the propositional thought also generates consciousness of himself as the subject or thinker of that thought. The activity is both object- and subject-constituting. But this seems to imply something that is not true. It is not true that whenever a thinker has a propositional thought or is in some propositional state that thinker is also aware or conscious of himself as the subject of that state. He is the subject of the propositional state, but whether he thinks or is aware of himself as the subject of the state he is in is a separate question. And in most cases, I would say, a propositional thinker does not usually think of himself at that time as the subject of his thought. It is true that a thinker is, so to speak, present whenever there is a thought of his, but no “self-conscious I” need be present, or present to the thinker, each time. Horstmann walks along and thinks ‘This house has a blue door’. At that moment he does not, at least not typically, also think ‘I think this house has a blue door’. He is not conscious of himself as the subject of the state of thinking that he is actually in. But it seems as if he would have to be aware of himself as subject each time he has a propositional thought if the activity of combining the ingredients of a propositional thought into a unity also and necessarily generates the presence of a self-conscious I that is conscious of itself as the subject of that thought. In denying that there is always, or even very o en, some self-awareness on the part of a thinker with a propositional thought I do not mean to deny that a self-conscious I is a “necessary element” in someone’s being in propositional states. I think it is a necessary element in the sense that any subject or agent of propositional thought is and must be a self-conscious being. He must be capable of thinking of himself as the subject or agent of his thoughts. If he did not have that capacity he would not be capable of propositional thought, for reasons like those Horstmann has given. So for propositional thought I think we must accept something like Kant’s weaker condition that it must at least be possible for the ‘I think’ to accompany each of my propositionally-structured thoughts. I must be capable of attaching ‘I think’ to thoughts that are mine, but no self-awareness or self-ascription of thought is necessary on each occasion in order for a judgement or a propositional thought to occur. This suggests that what is required for the presence of a propositionally-structured thought, and so for the possibility of the required self-consciousness that goes along with it, is something other than an activity that somehow generates or constitutes those elements. Without necessarily thinking there must be some such activity, we can ask what it takes for someone to have a propositionally-structured thought. What does it take for what someone does on a particular occasion to be an expression of such a thought? It would not be enough for the person to utter, even to intentionally utter, what is in fact a full

Judgement, Self-Consciousness, Idealism

45

sentence. Nor would it be enough for the person to respond in some distinctive way to such a sentence. Someone might utter or respond to what is in fact a whole sentence simply as a single unstructured signal or symbol, without understanding or thinking of it as a sentence with an internal structure at all. A dog or a small child could be trained to run to the door every time he hears what is in fact a certain sentence from the Critique of Pure Reason. The child could even be trained to say the sentence out loud. He would not then have grasped the internal structure of that sentence or understood it as expressing something that is either true or false. I think whether someone grasps the internal structure of a particular sentence and understands it and has a thought that says something either true or false depends in part on what other sentences or thoughts of that kind the person can understand and can himself assert in the appropriate conditions. A competent thinker brings to each particular sentence he hears or speaks a general ability or capacity to understand and respond in appropriate ways to a great many di erent sentences of similar or related forms. He understands what the sentence ‘This house has blue door’ says because he knows how it di ers from other sentences he understands such as ‘This house has a black door’, ‘This car has a blue door’, ‘This house has a blue window’, and countless other sentences about objects and their parts and their colours and shapes and so on. It is because of the internal complexity and the wide generality of a speaker’s capacity to understand sentences and thoughts of many di erent kinds that he can discern the relevant structure of a particular sentence, and so have a particular thought with that structure. This ability or capacity is something we all acquire in learning to speak. It can be thought of as a kind of practical knowledge, or know-how; a certain expertise that all speakers and thinkers have. It is a general competence in speaking, thinking, responding, and acting whose intricate structure is at least as complicated as all the di erences among the great variety of sentences we can understand and discriminate from one another. This understanding does not always involve judgement or assertion of the thought. A thinker can wonder whether a certain house has a blue door, or know that if the house has a blue door then soand-so, without judging or believing that any house actually has a blue door. But if the speaker endorses, or judges, or puts the sentence forward as true, he can be said to think or believe that the house has a blue door. Horstmann apparently came to believe that the house has a blue door because he saw that the house has blue door. What he saw to be so is the very thing he thinks to be so in the thought ‘This house has a blue door’. But he is able to see that that is so, and to think or believe it, only because there are a great many other thoughts he could also have, and many other things he could also see to be so (if they were so). Those other thoughts he is capable of are what we might call

46

Stroud

“possible judgements” for him. It is because such possible judgements are within his repertoire that he can be said to possess the concepts he employs in the particular thought he actually has. Concepts are predicates of possible judgements. So what you think, or whether you think such-and-such on a particular occasion, depends on what else you can understand and think and judge. It is a matter of what capacities you possess, and what you express in what you say or do, not a question of what processes or activities go on at the time of thinking the thought, even so-called object- or subject-constituting activities. I have said that when someone sees and so has the thought ‘This house has a blue door’, the person does not also have to have the thought ‘I think that this house has a blue door’; no self-conscious I or self-awareness on the part of the thinker has to be present at all. But, as Horstmann rightly points out, the thought the person then has is distinct from, is not the same as, the person who has that thought. And to understand and acknowledge that di erence the thinker must have some way of thinking of himself. It is also true that what the person thinks when he thinks ‘This house has a blue door’ is something that is true or false independently of whether that thinker thinks it. And any thinker of propositional thoughts could understand that that is true of most of his propositional thoughts. That is something he knows simply by understanding the thoughts he has. Someone who thinks that this house has a blue door understands what it is for a house to have a door and for the door to be blue. In understanding that sentence or thought he knows the conditions under which it would be true, so he knows some of the things it implies and some of the things it does not imply. And he can see that it does not imply that he or anyone thinks that the house has a blue door. For a thinker to recognize or acknowledge that something he thinks to be so could be true (or false) independently of whether he thinks it, the person must be capable of thinking of himself as a thinker of thoughts. Having a capacity to think of himself as subject or bearer of those thoughts does not require that one think of oneself each time one has a thought of an objective state of a airs. But without a capacity to think of myself as a thinker, and to attribute thoughts to myself, I could have no conception of something’s being so (or not) independently of whether I think it is so. The objectivity I ascribe to the things I think to be so is intelligible to me only because I have a special way of thinking of myself. This way of thinking of myself is special in that it is not simply a thought of myself as a subject or thinker of thoughts. I do have such a concept. It is a concept I can apply to other people and so think of them as having certain thoughts and beliefs. With other people’s beliefs it is easy for me to acknowledge that the truth of what they think is independent of its being believed to be so by those thinkers. I can sometimes know or see right before my eyes that what those other thinkers think is not so. I know they believe it, but I can see it is not true. I can recog-

Judgement, Self-Consciousness, Idealism

47

nize their error and know that they are wrong only because I myself at that time believe or know something to be so that is incompatible with what I take those others to believe. And in that case I must then have a thought that I take to be objectively true about what is so. That is, I must understand myself to believe something to be true or false independently of whether I think it. And my having that thought is what requires what I am calling this special way of thinking of myself. I do not think of myself simply as one among many other agents or thinkers of thoughts. That is what I am, of course – one among many other agents or thinkers of thoughts. But when I think of myself as the subject or thinker of the very thought I now have, I must employ this special way I have of thinking of or referring to the particular subject or thinker that I am. My being able to think of myself in this special way does not mean that I cannot also think of other agents or thinkers as capable of thinking of themselves in this same special way. For agents to think of themselves as having thoughts that are true or false independently of whether they think them, they too must have some special way of thinking of themselves. They must be capable of thinking of themselves not simply as one of the agents or thinkers there are. Each of us has this special way of thinking of ourselves in the rst person, typically with a rst-person pronoun. I can think of other thinkers, and also of myself, in the third person. But in so far as I think of others as entertaining and understanding propositional thoughts that are true or false independently of their being thought, I think of those thinkers as also capable of thinking of or referring to themselves in this rst-person way. I can think of them as having the thoughts I attribute to them in the third person only because I also think of them as having a capacity to attribute such thoughts to themselves in the rst person. That is a condition of their having thoughts they can understand to be objectively true or false. And other thinkers must think of me as having the same capacity in so far as they can think of me as a thinker of propositional thoughts. This special way of thinking of or referring to ourselves in the rst person must be within the capabilities of any thinker of propositional thoughts. This seems to me the sense in which the possibility of self-consciousness is required for any thinker of propositional thoughts. Is this the same as the “selfconscious I” that Horstmann thinks is a “necessary element” in any “conscious state whose content is a proposition” (447)? Well, I think we have seen that there does not have to be any self-conscious thought of oneself as the thinker of the thought each time one has a propositional thought. But if “having self-consciousness” or being a “self-conscious being” means only having a general capacity for thinking of oneself and one’s thoughts in this special rst-personal way, then I think it is essential for propositional thought. Only beings capable of this way of thinking of themselves could have propositional thoughts they understand to be

48

Stroud

objectively true or false. So every thinker of such propositional thoughts must be capable of self-consciousness in this sense. The very words ‘self-conscious’ or ‘conscious of myself’ suggest that when I am in such a state there is something I am then conscious of or aware of. This leads to familiar questions about what that something could be. Is it a “self”, or “my self”; is it an “Ich”, or “my Ich”? And if so, what sort of thing is that? When I try to look more closely, whatever it is seems strangely elusive, as Horstmann points out. But how can the thing be so elusive if it is something I am conscious or aware of every time I say or think something about myself in the rst person? I think the self, or myself, considered as an object of awareness, is worse than elusive. On the question of what exactly I am aware of when I think about or turn my attention to myself, I think Hume’s answer was closest to the truth. “For my part, when I enter most intimately into what I call myself,” Hume said, “I always stumble on some particular perception or other […] I never can catch myself at any time without a perception, and never can observe anything but the perception”.² As a point simply about awareness, or what we nd on looking ‘inwards’, what Hume says seems to me more or less right. I don’t mean Hume’s own positive account of the self, or the idea of the self. That seems to me hopeless, as it seemed to Hume. But he seems right about what he says we are aware of when we try to “observe” ourselves. I think it is not just that when I try to do that I never manage to nd the thing I am looking for. I think nothing I notice or nd or am aware of in such a search could be something I can depend on to pick myself out. For any such property or feature that I could notice and think about, I could always doubt or deny the alleged connection between that property and me, with the thing I am, without thereby failing to think about myself. I might have wildly false beliefs about myself, or doubt or deny that whatever I had found was really me. I could even wonder, as apparently some people sometimes do, whether I am actually the thinker of this very thought. Of course, in having the thought ‘I wonder whether I am the thinker of this thought’, I refer with my rst use of ‘I’ to me, the subject, the one who wonders. So I am the thinker of that very thought, whatever I might believe or wonder about myself. But in having the thought that I am (or even am not) the thinker of this thought I do not secure reference to myself by way of the de nite description ‘the thinker of this thought’. I don’t have to apply any concept or description to myself to pick myself out. Reference to myself is secured simply by its being made by me in the rst person, no matter what I might believe or doubt or wonder is true of me. Com-

2 D. Hume, A Treatise of Human Nature (ed. Selby-Bigge), Clarendon Press, Oxford 1958, p. 252.

Judgement, Self-Consciousness, Idealism

49

petent use of ‘I’ is all that is needed for self-reference, and so for an expression of self-consciousness, if self-consciousness is rst-person attribution of psychological states or attitudes to oneself. That is a capacity I think is required for the possibility of propositional thought. But it does not require awareness or experience of an object of some recognizable kind, especially not of an elusive object. The self-consciousness I exhibit in attributing states and attitudes to myself in the rst person is itself a form of propositional thought. It involves saying or thinking something about myself that is expressed in a full sentence. I say ‘I think that house has a blue door’, and that is something that is either true or false, not only about the house and the door, but about me. I can make and understand sentences like that because I am a master of the rst person. I use rst-person forms in sentences to say and think things about myself. My use of the rst-person pronoun, like everyone else’s, refers to whoever utters or thinks it in a sentence. So there is no possibility of failure of reference, or failure to ‘get hold’ of myself when I use it. The rst-person statements I make are true or false depending on what is true of the person they refer to as subject. This suggests that not only is the possibility of self-consciousness essential to propositional thought, as Horstmann and I agree it is, but also that the possibility of propositional thought is essential to self-consciousness. Only in my having rstpersonal propositional thoughts about myself do I possess self-consciousness. If that is so, it further con rms the idea that the self-consciousness we human beings possess is not to be understood as a kind of awareness or consciousness of something that a creature who is incapable of propositional thought could have. A being incapable of propositional thought can be conscious, and conscious of many di erent things – aware of pains and tickles and other sensations, aware of a hawk in the sky overhead, or aware of a swi movement in the grass underfoot – and it could respond accordingly. But whatever such a being was aware of, and whatever its psychological life was like, such a creature could not think of itself as having that psychological life, or as being or having been in any of those psychological states. That requires a capacity for propositional thought. So such a creature could have no self-consciousness of the kind we human beings possess. That both requires and is required by a capacity for propositional thought. The intricately articulated capacity I possess as a propositional thinker is what enables me to understand and, on the right occasion, to assert the thought ‘This house has a blue door’. Walking down the street with that complex capacity, and with the concepts I have acquired in exercising it, it takes only the briefest glance at the house for me to see that the house has a blue door. I acquired the capacities I have as a propositional thinker, and they have taken the speci c forms they have taken, partly because of the way things are in the world in which I acquired them. I learned some things to be so in the world around me at the same time as I was

50

Stroud

learning how to say and think that things are thus and so in the world. I don’t think that what is so in the world is “constituted” as being so by my having the capacities I have to say or think that certain things are so. Nor is it “constituted” by anyone else’s having such capacities, or even by everyone’s having them. Given our natural talents, we all acquire the capacity for propositional thought in a world in which certain things are so. But in the very idea of propositional thought, and in our having a capacity to engage in it, and in the fact that we must be capable of self-consciousness if we can think of an objective world at all, I do not see any encouragement to idealism. The indispensability of propositional thought, and our ful lling all its necessary conditions, give us no reason to conclude that what we thereby think and know about is a “mind-dependent” world.

Dina Emundts

Kant über Selbstbewusstsein Für Kant ist Selbstbewusstsein eine Bedingung für Erkenntnis. Erkenntnis hat etwas damit zu tun, dass man urteilt und verschiedene Urteile in Beziehung zueinander setzt und miteinander verbindet. Um dies in der für Erkenntnis erforderlichen Weise zu tun, muss man sich bewusst sein, dass man urteilt oder dass man der Denker seiner Gedanken ist.¹ Denn man muss in der Lage sein, sich vom Gehalt seiner Gedanken zu distanzieren,² und man muss sich re exiv auf seine Gedanken, Urteile oder Meinungen beziehen können.³ Andernfalls könnte man nicht verschiedene Verbindungen zwischen Gedanken erwägen, man könnte seine Meinungen nicht evaluieren und man könnte nicht hinreichend zwischen wahren und falschen Meinungen unterscheiden, weil zu dieser Unterscheidung der Gedanke gehört, dass etwas unabhängig davon der Fall ist, dass ich es denke. Aufgrund dieser Überlegungen scheint es richtig zu sein, Selbstbewusstsein als eine Bedingung für Erkenntnis anzusehen. In der Kantliteratur werden Kants Ausführungen zum Selbstbewusstsein allerdings verschieden interpretiert. Es ist o enbar nicht eindeutig, was Kant mit Selbstbewusstsein genau meint. Es gibt Unklarheiten in verschiedenen Hinsichten. Eine Unklarheit betri die Funktion von Selbstbewusstsein: Selbstbewusstsein hat für Kant zwar zweifellos die Funktion, eine Bedingung für Erkenntnis zu sein. Aber diese Funktion kann man nur spezi zieren, wenn man die Frage beantwortet, was hier „Erkenntnis“ heißt. Ich habe „Erkenntnis“ bisher im Hinblick auf Urteile ausgeführt. Wenn man jedoch an die Transzendentale Deduktion der Kritik

1 Vgl. Horstmann und Stroud in diesem Band. Der vorliegende Aufsatz ist in Auseinandersetzung mit Horstmanns Text „Self-Consciousness and Propositional Acts. Revisiting Epistemological Issues in Phenomenology and German Idealism“ entstanden. Dieser Text hat (bisher) vier Teile. Von diesen vier Teilen liegen die beiden letzten – „The Self-Conscious I and the Self“ und „Sameness and Identity“ – nur als Manuskript vor. Die beiden ersten Texte sind verö entlicht: (2010): „The Limited Signi cance of Self-Consciousness“, in: Revue de Métaphysique et de Morale 68, 434–454 und „Propositional Activity in Kant and Hegel“ in diesem Band. Ich werde auf diesen Text immer wieder in Fußnoten verweisen. Horstmann gibt dem Thema Selbstbewusstsein bei Kant allerdings von Anfang an eine andere Pointe, als ich es tue, indem er die These in den Vordergrund stellt, dass das „radically subjective consciousness of oneself“ immer zusammen mit propositionalem Gehalt au ritt, als eine Art dem Objekt entsprechender Ich-Pol. 2 Mit „distanzieren“ ist gemeint, dass man sich unter anderem dessen bewusst ist, dass man etwas behauptet, das wahr oder falsch sein kann 3 Carl spricht davon, dass man zu einer re exiven Selbstbewertung in der Lage sein muss. Carl, Wolfgang (1998): „Ich und Spontaneität“, in: Philosophie in synthetischer Absicht, hg. v. M. Stamm. Stuttgart, 105–122; hier: 119.

52

Emundts

der reinen Vernun denkt, scheint für Kant Selbstbewusstsein primär eine Bedingung dafür zu sein, dass ich etwas als Objekt behandeln kann und nicht nur als bloß subjektive Vorstellung ansehen muss. Warum dies so sein soll und was hier „Objekt“ heißt, ist zu klären. Ich werde hierauf im dritten und vierten Abschnitt zurückkommen. Hauptsächlich will ich mich in diesem Aufsatz nicht mit der Funktion von Selbstbewusstsein, sondern direkt mit der Frage beschä igen, was „Selbstbewusstsein“ heißt. Zur Beantwortung dieser Frage, möchte ich zunächst zwei Bedeutungen von Selbstbewusstsein unterscheiden, die den Formulierungen nach bei Kant und den Interpreten o miteinander vermischt werden. Diese Unterscheidung ist meines Erachtens hilfreich, um herauszu nden, was Kant mit Selbstbewusstsein genau meint. Daran anschließend, im zweiten Abschnitt dieses Aufsatzes, werde ich mich mit einigen einschlägigen Textpassagen der Kritik der reinen Vernun beschä igen. Im dritten Abschnitt werde ich über Selbstwahrnehmung reden. Ich will zeigen, dass Selbstwahrnehmung mit Blick auf verschiedene Zustände, in denen wir sind, möglich ist. Die Pointe von Kants Position besteht darin, dass er den Selbstbezug beim Denken oder vielmehr bei bestimmten grundlegenden Tätigkeiten des Denkens auszeichnen will und als „Selbstbewusstsein“ bezeichnet. Mit dieser bestimmten Tätigkeit des Denkens will ich mich dann im vierten Abschnitt dieses Aufsatzes genauer beschä igen.

1 Was ist Selbstbewusstsein? Selbstbewusstsein kann entweder als eine Art von faktisch vorliegendem Bewusstsein seiner selbst oder als eine Fähigkeit, sich beispielsweise als Denker seiner Gedanken zu denken, verstanden werden. In dem, was ich bisher zu Selbstbewusstsein als Bedingung für Erkenntnis ausgeführt habe, könnte man dementsprechend zwei Thesen angelegt sehen: (1) Wenn man etwas erkennt, so nur unter der Bedingung von Selbstbewusstsein als einem Bewusstsein, dass man der Denker seiner Gedanken ist. Dieses Bewusstsein verleiht einem die Fähigkeit, seine Gedanken (in einer objektiven Weise) zu verbinden. (2) Selbstbewusstsein ist die Fähigkeit, sich bewusst zu machen, dass man der Denker seiner Gedanken ist, und diese Fähigkeit ist die Voraussetzung für eine objektive Verbindung von Gedanken und damit für Erkenntnis. Als Vertreter der ersten Lesart sehe ich unter anderem Horstmann und Henrich. Als Vertreter der zweiten Lesart zum Beispiel Carl.⁴

4 Diese Einordnung stammt von mir, die Autoren selbst nehmen diese Unterscheidung nicht vor. Beide Lesarten sind m.E. im vorliegenden Band vertreten, vgl. die Beiträge von Horstmann und Carl. Zu nennen ist auch der Beitrag von Stroud. Auch Stroud meint, wenn ich ihn richtig verstehe,

Kant über Selbstbewusstsein

53

Zunächst scheint vielleicht klar, dass es sinnvoller ist, von Selbstbewusstsein als einer Fähigkeit, sich seiner selbst bewusst zu werden, zu sprechen. Denn erstens scheint die Rede von einer Fähigkeit weniger mysteriös zu sein. Und zweitens scheint auch nur bei dieser Bedeutung von Selbstbewusstsein die Überlegung zu Selbstbewusstsein als Bedingung für Erkenntnis Sinn zu machen. Die Annahme, dass Selbstbewusstsein das Bewusstsein seiner selbst ist, kann man nämlich mit dem Argument zurückweisen, dass sich Selbstbewusstsein aufgrund folgender Überlegung so gar nicht als Bedingung für Erkenntnis verstehen lässt: Wir müssen uns als Denker unserer Gedanken denken können, wenn wir objektive Urteile fällen wollen. Dies lässt sich auch so formulieren: Wir müssen uns bewusst werden können, dass wir Denker unserer Gedanken sind. Also muss Selbstbewusstsein als Bedingung für Erkenntnis als die Fähigkeit verstanden werden, sich bewusst zu werden, Denker seiner Gedanken zu sein. Man kann dagegen unter Selbstbewusstsein nicht ein von dieser Fähigkeit verschiedenes Bewusstsein verstehen. Denn das würde heißen, dass man für Erkenntnis Selbstbewusstsein als ein Bewusstsein, dass man der Denker seiner Gedanken ist, annimmt. Ein solches Bewusstsein liegt beim Denken aber nur hin und wieder vor. Wir sind uns nämlich nicht immer, wenn wir denken, bewusst, Denker unserer Gedanken zu sein. Ein von der Fähigkeit verschiedenes Bewusstsein kann daher keine Bedingung für Erkenntnis sein. Daher, so die Überlegung weiter, kann Kant (wenn man ihm keine absurde These unterstellen will) nur behaupten, dass die Fähigkeit zu Selbstbewusstsein eine Bedingung für Erkenntnis ist. Wenn man sagt, dass Selbstbewusstsein eine Bedingung für Erkenntnis ist, muss man also in Kants Sinn eigentlich meinen, dass Selbstbewusstsein eine Fähigkeit ist.⁵

dass Kant Selbstbewusstsein als Bewusstsein versteht, er will aber gerade in diesem Punkt Kant (und Horstmann) widersprechen. Zur ersten Lesart zähle ich unter anderem auch Dieter Henrich und Patricia Kitcher. Vgl. Henrich, Dieter (1987): „Die Identität des Subjekts in der transzendentalen Deduktion“, in: Kant. Analysen – Probleme – Kritik, hg. von H. Oberer und G. Seel, Würzburg 39–70, besonders 59; Kitcher, Patricia (2011): Kant’s Thinker, Oxford, bes. 175. Zur zweiten Lesart: Carl „Ich und Spontaneität“, 114, und Rosefeldt, Tobias (2000): Das logische Ich. Kant über den Gehalt des Begri es von sich selbst. Berlin, sowie Rosefeldt (2006): „Kants Ich als Gegenstand“, in: Deutsche Zeitschri für Philosophie 54, 277–293. Bei vielen Autoren ist eine Zuordnung aufgrund der Komplexität des Themas nicht leicht. Bei Longuenesse sind einige Formulierung (etwa 153) so, dass ich sie zur ersten Lesart zählen würde. Aber einer ihrer Grundgedanken (154), dass es bei Kant um die Idee der Einheit des Bewusstseins gehe, die noch nicht dessen Realität verbürge, scheint mir in eine andere Richtung zu gehen. Longuenesse, Béatrice (2007): „Kant on the Identity of Persons“, in: Proceedings of the Aristotelian Society, Vol. CVII, Part 2, 149–157. 5 Ähnlich argumentiert, wenn ich ihn richtig verstehe, Stroud in seinem Beitrag in diesem Band.

54

Emundts

Obwohl diese Überlegung überzeugend zu sein scheint, spricht in meinen Augen einiges gegen die These, dass Kant Selbstbewusstsein als die Fähigkeit verstanden wissen wollte, sich seiner selbst bewusst zu werden. Selbstbewusstsein ist für Kant, so will ich argumentieren, das Bewusstsein seiner selbst als eines denkenden Wesens. Wenn Selbstbewusstsein eine Fähigkeit wäre, würde es eine Fähigkeit sein, die andere Fähigkeiten bedingen würde. Im Sinne Kants muss man aber (wie ich zeigen möchte) eher sagen, dass wir als selbstbewusste Wesen bestimmte Fähigkeiten haben. Anders als in der soeben vorgestellten Überlegung, die dafür sprechen sollte, Selbstbewusstsein als eine Fähigkeit anzunehmen, nimmt Kant nicht an, dass „sich seiner selbst als Denker seiner Gedanken bewusst sein“ bedeutet „sich als Denker seiner Gedanken zu denken“. „Bewusstsein“ meint stattdessen, dass mir etwas gegenwärtig ist. Dies soll auch dann der Fall sein können, wenn ich es nicht explizit denke. Auf diese Weise ergibt sich die Möglichkeit, unter Selbstbewusstsein nicht eine Fähigkeit zu verstehen, sondern eine besondere Art von Bewusstsein.⁶ Denn das Bewusstsein meiner selbst als Denker kann dann als etwas angenommen werden, das auch dann vorliegt, wenn ich nicht explizit denke, dass ich der Denker meiner Gedanken bin. Während ich nur hin und wieder denke, dass ich der Denker meiner Gedanken bin, kann mir dies auch sonst bewusst sein (wenn ich denke). Ein wichtiger sachlicher Grund, Selbstbewusstsein nicht als eine Fähigkeit, sondern als ein Bewusstsein anzusehen, besteht in Folgendem: Wenn Selbstbewusstsein nur die Fähigkeit ist, sich seiner selbst bewusst zu werden, dann wird eine Klu zwischen zwei Zuständen behauptet, zwischen denen keine Klu besteht. Wenn ich sage, dass die Sonne den Stein erwärmt, und mich jemand fragt, ob ich denke, dass die Sonne den Stein erwärmt, muss ich in der Regel nicht erst Anstrengungen unternehmen oder Fähigkeiten ausüben, um diese Frage zu beantworten. Dies gilt auch für den Fall, wenn mich jemand fragt, ob ich wirklich meine, dass die Sonne den Stein erwärmt. Auch wenn es zweifellos zutri , dass ich nicht immer, wenn ich etwas denke, auch denke, dass ich es denke, scheint es doch so zu sein, dass ich mir nicht nur in bestimmten Situationen bewusst werde, dass ich denke. Denn sonst würden sich diese Situationen viel mehr von meiner sonstigen Erfahrung abheben. Es ist daher phänomenologisch naheliegender, dass mir irgendwie vorher schon mein Denken gegenwärtig war; dass ich in solchen Si-

6 Als eine besondere Art von Bewusstsein fasst z.B. Horstmann („Limited Signi cance“) Selbstbewusstsein auf: „[…] this subject has to be conscious of itself as the subject of that state“ (451). Vgl. auch seinen Beitrag in diesem Band.

Kant über Selbstbewusstsein

55

tuationen nur etwas explizit mache, dessen ich mir schon bewusst gewesen bin: nämlich, dass ich denke, dass die Sonne den Stein erwärmt. Mir scheint es also eine attraktive Option zu sein, Selbstbewusstsein nicht als eine Fähigkeit, sondern als ein Bewusstsein anzusehen, wobei unter Bewusstsein hier verstanden wird, dass mir etwas – in diesem Fall die Tatsache, dass ich der Denker meiner Gedanken bin – mental gegenwärtig ist. Dann könnte man also sinnvoll davon sprechen, dass man sich als Denker seiner Gedanken bewusst ist, ohne dass man explizit denkt, dass man der Denker seiner Gedanken ist. So etwas hat Kant in meinen Augen angenommen. Dies ist es, was Interpreten wie Dieter Henrich als eine Art implizites Wissen bezeichnet haben.⁷ Das Problem dieses Vorschlags liegt zweifellos darin, dass unklar ist, was ein solches Bewusstsein sein soll. Wenn man sagt, ich weiß implizit, dass ich der Denker dieses oder jenes Gedankens bin, muss man erläutern, was das heißt. Vielleicht kann man über dieses implizite Wissen nicht mehr sagen, als dass es bedeutet, dass man „ich denke“ denken kann oder sich Gedanken zuschreiben kann. Dann stellt sich allerdings die Frage, ob dies von einer Fähigkeit oder Disposition überhaupt zu unterscheiden ist. Will man es unterscheiden, muss man irgendwie angeben können, wodurch sich das Bewusstsein seiner selbst als Denker des Gedankens in Situationen auszeichnen soll, in denen man nicht explizit auch an sich denkt und beispielsweise explizit denkt: „Ich denke, dass die Sonne den Stein erwärmt.“ Wie bin ich mir meiner denkend gegenwärtig, wenn nicht, indem ich denke, dass ich denke? Ist Bewusstsein eine Art Wahrnehmung meiner selbst als denkend? Um diese Fragen zu beantworten, muss man etwas zum Begri des Bewusstseins sagen. Dies werde ich vor allem im dritten Abschnitt tun. An dieser Stelle sieht man, dass von der Frage, ob Selbstbewusstsein eine bloße Fähigkeit oder ein noch näher zu spezi zierendes Bewusstsein ist, einiges abhängt. Die beiden Alternativen führen zu unterschiedlichen Theorien von Selbstbewusstsein als Bedingung für Erkenntnis. Sie unterscheiden sich unter anderem darin, dass es bei der Fähigkeit keine Selbstwahrnehmung geben muss, während so etwas wie Selbstwahrnehmung mit der These, dass Selbstbewusstsein ein anders zu quali zierendes Bewusstsein sei, untrennbar zusammenhän-

7 Unter „Wissen“ kann hier bei meiner Lesart „Gegenwärtighaben“ oder „sich dessen bewusst sein“ verstanden werden. Nach Henrich („Identität des Subjekts“) müssen wir annehmen, dass es hier Wissen gibt, weil wir nur so die Identität des Ichs gegenwärtig haben können, welche wir wiederum zum objektiven Verbinden brauchen. Horstmann („Sameness and Identity“) vertritt dagegen hinsichtlich der Frage nach der Identität des Ichs die These, dass wir das Ich durchaus punktuell denken müssen (auch in diesem Sinn ist es evasiv). Ich kann auf das Thema der Identität hier nicht eigens eingehen.

56

Emundts

gen könnte.⁸ Die Alternativen unterscheiden sich außerdem darin, auf was für ontologische Behauptungen man sich mit ihnen festlegt. Wenn man annimmt, dass wir uns beim Denken immer unserer selbst bewusst sind, dann entsteht vielleicht der Eindruck, es gebe hier ein mysteriöses Selbstverhältnis oder ein rein denkendes Selbst – ein Eindruck, der bei der Annahme der Fähigkeit, sich als Denker seiner Gedanken zu denken, nicht entsteht. Die Theorie ontologisch möglichst wenig aufzuladen, ist sicherlich ein Motiv dafür, Kant die Au assung von Selbstbewusstsein als Fähigkeit zuzuschreiben. Allerdings ist es nicht klar, wie ontologisch harmlos Kant seine Theorie selbst aufgefasst hat. Bedenkend, wie Kant sich mit den auf der Hand liegenden Fehlschlüssen seiner Theorie des Selbstbewusstseins beschä igt, kann man auf jeden Fall sagen, dass er sie nicht so verstanden wissen wollte, dass das Ich oder Selbst etwas ontologisch Eigenständiges ist. Aber die Tatsache, dass Kant sich wiederholt und ausführlich mit möglichen Fehlschlüssen beschä igt, zeigt auch, dass er seine eigene Theorie nicht für ontologisch völlig harmlos gehalten hat. Wenn man Selbstbewusstsein als eine Fähigkeit au asst, ist Kants Theorie zweifellos anschlussfähiger für heutige Theorien⁹ – aber dies allein ist natürlich kein Grund anzunehmen, dass Kant selbst sie so aufgefasst hat. Ich möchte im Folgenden jedenfalls fortfahren zu zeigen, dass Kants Au assung von Selbstbewusstsein insgesamt dagegen spricht, Selbstbewusstsein als eine Fähigkeit aufzufassen. Hierfür werde ich mich im dritten Abschnitt dieses Aufsatzes mit dem Thema der Selbstwahrnehmung beschä igen. Ich werde hier die These vertreten, dass Kant eine eigentümliche Form des vom Denken verschiedenen Bewusstseins seiner selbst beim Denken annimmt. Anschließend werde ich Kants Begründung für Selbstbewusstsein als Bedingung für Erkenntnis näher untersuchen und hierdurch aufzuklären versuchen, was Kant dazu bewogen hat, diese eigentümliche Form von Bewusstsein anzunehmen. Im folgenden zweiten Abschnitt werde ich mich zunächst einigen Passagen aus der Kritik der reinen Vernun zuwenden.

8 Ob das tatsächlich so sein muss, hängt, wie schon angedeutet, vom Bewusstseinsbegri ab. 9 Die Annahme einer Fähigkeit entspricht den meisten heutigen Au assungen zu Selbstbewusstsein als Bedingung für Erkenntnis, vgl. Stroud in diesem Band, S. 46: „But without a capacity to think of myself as a thinker, and to attribute thoughts to myself, I could have no conception of something’s being so (or not) independently of whether I think it is so. The objectivity I ascribe to the things I think to be so is intelligible to me only because I have a way of thinking of their independence from their being thought or believed by me. And that requires that I have a special way of thinking of me.“ Wenn man behauptet, Kant ginge es mit Selbstbewusstsein um eine Fähigkeit, wird es entschieden leichter, von Kant eine Verbindung zu aktuellen Theorien zum Thema Selbstbewusstsein zu ziehen wie beispielsweise zu der von John Perry.

Kant über Selbstbewusstsein

57

2 Die Begleitvorstellung „Ich denke“ und die Einheit des Bewusstseins Die Frage, ob Kant Selbstbewusstsein als eine Fähigkeit oder als ein genuines Bewusstsein verstanden wissen will, lässt sich schon deshalb nicht ohne weiteres anhand von Kants Schri en beantworten, weil es viele verschieden lautende Textstellen gibt, auf die man sich bei der einen oder der anderen Lesart beziehen kann. Die Interpreten, die Kant so verstehen, als ginge es ihm um eine bloße Fähigkeit, berufen sich gerne auf den berühmten Satz: Das: Ich denke, muß alle meine Vorstellungen begleiten können; denn sonst würde etwas in mir vorgestellt werden, was gar nicht gedacht werden könnte, welches ebensoviel heißt, als die Vorstellung würde entweder unmöglich, oder wenigstens für mich nichts sein (B 131 f).¹⁰

Dieser Satz wird zumeist so verstanden, dass Gedanken nur dann meine Gedanken sind, wenn ich sie mir auch explizit als meine Gedanken zuschreiben kann. Betont wird daher das „können“ in Kants Satz, weil dies darauf hinzudeuten scheint, dass man in der Lage sein muss, sich einen Gedanken als seinen eigenen zuzuschreiben.¹¹ Damit scheint Selbstbewusstsein die Fähigkeit zu meinen, sich Gedanken als seine zuschreiben zu können. Dies ist jedoch nicht das, was Kant mit diesem Satz sagt. Mit dem Satz wird nicht gesagt, dass ich, wenn ich denke, diese Gedanken mit dem „Ich denke“ begleiten kann. Gesagt wird, dass alle meine Vorstellungen vom „Ich denke“ begleitet werden können müssen, weil ich nur die Vorstellungen miteinander verbinden kann, die durch das „Ich denke“ begleitet werden können. Also müssen alle meine Vorstellungen vom „Ich denke“ begleitet werden können, damit sie miteinander verbindbar sind – denn sind sie das nicht, so sind sie, wie Kant sagt, für mich nichts. Dies legt folgende Annahme zumindest nahe: Wann immer ich Vorstellungen wirklich miteinander verbinde – also objektiv urteile – muss das „Ich denke“ diese Vorstellungen begleiten – und nicht nur

10 Ich zitiere Kant im Text unter Angabe von Band- und Seitenzahl der Akademieausgabe (AA) von Kant’s gesammelte Schri en. Berlin und Leipzig: de Gruyter 1900 ., die Kritik der reinen Vernun dagegen wie üblich nach der ersten (A) bzw. zweiten (B) Au age. Es ist zu bemerken, dass Kants Aussage, dass diese Eindrücke für mich nichts sind, eine ziemlich starke These zur Beziehung Bewusstsein–Selbstbewusstsein impliziert. Denn auf viele Eindrücke, die zunächst nicht geordnet scheinen, kann man im Zusammenhang anderer Gedanken manchmal doch wieder zugreifen, auch sie würden demnach Selbstbewusstsein voraussetzen. Hier kann man vielleicht sagen: Diese Eindrücke sind für mich zunächst nichts und sie können sich nur unwillkürlich wieder einstellen. Es stellt sich das Problem, wie die Vorstellungen, denen Selbstbewusstsein vorausgehen muss, mit denen zusammenhängen, bei denen das nicht der Fall ist. 11 Rosefeldt Das logische Ich, 101.

58

Emundts

begleiten können. Mit anderen Worten: Dass Vorstellungen überhaupt meine sind, bedeutet, dass ich sie mit anderen Vorstellungen von mir verbinden kann. Diese Verbindung von Vorstellungen ist eine aktive und selbstbewusste Handlung von mir als denkendem Subjekt. Kants Text scheint mir daher eher dafür zu sprechen, dass Kant Selbstbewusstsein als ein Bewusstsein meiner selbst versteht und nicht als eine Fähigkeit, mich als Denker meiner Gedanken zu denken. Es gibt andere Stellen, die man gegen meine Interpretation anführen könnte; zum Beispiel „Diese Vorstellung [ich] mag nun klar (empirisches Bewusstsein) oder dunkel sein, daran liegt hier nichts, ja nicht einmal die Wirklichkeit desselben; sondern die Möglichkeit der logischen Form alles Erkenntnisses beruht notwendig auf dem Verhältnis zu dieser Apperzeption als einem Vermögen“ (A 118). Wenn dieser Satz bedeuten soll, dass es auch ausreicht, das Vermögen zu haben, „Ich denke“ zu denken, so spricht er gegen meine These. Man kann ihn aber auch so lesen, dass eine Vorstellung nur dann Teil der Erkenntnis ist, wenn sie vom „Ich denke“ begleitet wird – daher kann jede Vorstellung vom „Ich denke“ begleitet werden –, aber wenn sie Erkenntnis ist, gibt es auch Bewusstsein meiner selbst (und nicht nur die Fähigkeit, mich als Denker zu denken). Die Frage, ob Kant eine Fähigkeit oder ein Bewusstsein im bisher erläuterten Sinn meint, ist auch deshalb schwer anhand der Kantischen Schri en zu beantworten, weil die Antwort auf diese Frage mit der Interpretation des mit dem Thema Selbstbewusstsein verbundenen Projekts insgesamt zusammenhängt. Bei dem, was ich letztlich hier als Kants Begri des Selbstbewusstseins vorstellen werde, handelt es sich um ein Bewusstsein meiner selbst oder auch: ein Bewusstsein einer von mir vollzogenen Tätigkeit.¹² Diese Interpretation verp ichtet Kant auf eine Theorie, der zufolge mit Selbstbewusstsein ein genuiner Bewusstseinszustand verbunden ist, ein Zustand, bei dem man sich tatsächlich seiner selbst oder seiner Tätigkeit direkt und über die Zeit hinweg bewusst ist. Diese Theorie von Selbstbewusstsein passt meines Erachtens zu dem, was Kant insgesamt für ein Projekt verfolgt. Dies kann ich hier aber nicht ausführen, sondern nur andeuten. Die Grundidee zu Selbstbewusstsein in der Deduktion ist meines Erachtens, dass ich mich auf mich als tätiges Wesen beziehe und dies eine Voraussetzung für die objektive Verbindung von Mannigfaltigen ist. Umgekehrt ist die Verbindung

12 Die Rede davon, dass man sich der Tätigkeit des Denkens bewusst wird, ist für meine Interpretation wichtig. Auch hierin könnte ein Unterschied zu der These von Selbstbewusstsein als Fähigkeit liegen. Zwar schreibt man sich auch zufolge dieser Interpretation Urteilen als Tätigkeit zu, aber nicht in dem Sinn, dass es wichtig ist, dass man sich seiner auch als tätig bewusst ist. Dass Kant (besonders in der B-Au age) mit Selbstbewusstsein eine Tätigkeit meint, hat Horstmann herausgearbeitet; Horstmann, Rolf-Peter (1993): „Kants Paralogismen“, in: Kant-Studien 83, 408–425.

Kant über Selbstbewusstsein

59

von etwas zu einer objektiven Einheit die Voraussetzung für eine bestimmte Form von Selbstbezug, nämlich für einen Selbstbezug, bei dem ich meine Identität objektiv bestimmen kann. Dies kann ich nämlich, indem ich mich auf objektiv verbundene Ereignisse beziehe, also etwa feststelle, dass ein Gegenstand durch ein Ereignis in der Vergangenheit in die Lage gebracht wurde, in der er jetzt ist. In diesem Sinn der beiden möglichen Selbstbezüge lese ich zum Beispiel die folgende Stelle aus der Deduktion: Ich bin mir also des identischen Selbst bewußt, in Ansehung des Mannigfaltigen der mir in einer Anschauung gegebenen Vorstellungen, weil ich sie insgesamt meine Vorstellungen nenne, die eine ausmachen. Das ist aber so viel, als, daß ich mir einer notwendigen Synthesis derselben a priori bewußt bin, welche die ursprüngliche synthetische Einheit der Apperzeption heißt, unter der alle mir gegebene Vorstellungen stehen, aber unter die sie auch durch eine Synthesis gebracht werden müssen (B 135 f).

Es kommt Kant meines Erachtens in der Deduktion nicht nur darauf an, dass ich für eine Verbindung von Mannigfaltigen, die diese objektive Weise des Selbstbezugs ermöglicht, Regeln – d.i. die Kategorien – brauche. Es kommt ihm auch darauf an, dass dieser Selbstbezug über solche Regeln nur möglich ist, wenn die Verbindung durch uns, also aktiv und spontan zustande kommt und nicht an sich gegeben ist. Für diesen letzten Punkt bedarf es meines Erachtens das Bewusstsein meiner selbst als denkend tätig. Denn nur wenn ich ein Bewusstsein meiner Tätigkeit des Verbindens habe, kann ich das regelgeleitete objektive Verbinden von einem bloß subjektiven unterscheiden. Das mit einem Bewusstsein meiner Tätigkeit verbundene objektive Verbinden kann re ektiert, korrigiert und ausgewertet werden. In diesem Sinn eines begleitenden Bewusstseins meiner Tätigkeit lese ich beispielsweise folgende Passage: Ich nenne sie [die Vorstellung „Ich denke“, die als Akt der Spontaneität aufgefasst werden muss; D. E.] die reine Apperzeption, um sie von der empirischen zu unterscheiden, oder auch die ursprüngliche Apperzeption, weil sie dasjenige Selbstbewußtsein ist, was, indem es die Vorstellung Ich denke hervorbringt, die alle andere muß begleiten können, und in allem Bewußtsein ein und dasselbe ist, von keiner weiter begleitet werden kann (B 132).

Obwohl ich diese Stellen als Beleg für meine Au assung von Selbstbewusstsein au asse, schätze ich die Textlage nicht als eindeutig ein. Meines Erachtens sind Kants Äußerungen zum Selbstbewusstsein in der Kritik der reinen Vernun insgesamt so, dass sie nicht nur mit Blick auf die bisher diskutierte Frage, ob Selbstbewusstsein als ein Bewusstsein oder als eine Fähigkeit verstanden werden muss, uneindeutig sind. Es gibt auch mit Blick auf den Zusammenhang von Selbstbezug und Gegenstandserkenntnis zwei mögliche, sich aber ausschließende Interpretationen. Die eine lautet, kurz gesagt, dass Selbstbewusstsein als eine Bedingung

60

Emundts

für Erkenntnis ein Bewusstsein seiner selbst meint, das mit Gegenstandserkenntnis eng zusammenhängt, von diesem aber auch verschieden ist. Die andere lautet, dass Selbstbewusstsein gar nichts anderes als Gegenstandserkenntnis ist. Beide Thesen muss man auf das Thema Einheit bzw. Identität beziehen, weil dies beim Zusammenhang von Selbstbewusstsein und Gegenstandserkenntnis für Kant der entscheidende Aspekt ist. Die erste These besagt dann, dass wir ein Bewusstsein der Identität unserer selbst brauchen, um Gegenstandserkenntnis zu erlangen, wobei wir dieses Bewusstsein nur dann erfolgreich erlangen können, wenn wir uns tatsächlich auf Gegenstände beziehen. Die zweite These besagt, dass das Bewusstsein unserer Identität die Gegenstandserkenntnis ist. Das heißt, ich weiß, dass ich über die Zeit identisch bin, indem ich Urteile in einer bestimmten Weise miteinander verbinde, weil dies in einem Bewusstsein statt nden muss; aber abgesehen von diesem Wissen, dass das so sein muss, gibt es nicht noch irgendein Bewusstsein einer Tätigkeit oder einen genuinen Selbstbezug. Diesen gibt es aber zufolge der ersten These, weil hier das Bewusstsein der Tätigkeit des Verbindens von Mannigfaltigem oder von Urteilen dasjenige ist, was verbürgt, dass die Verbindung, die ich vornehme, objektiv werden kann. Diese Unterscheidung von möglichen Interpretationen zum Selbstbewusstsein ist mit der Alternative zwischen Selbstbewusstsein als Bewusstsein oder als Fähigkeit, mit der ich bisher operiert und die ich im ersten Abschnitt erläutert habe, nicht identisch. Dies liegt daran, dass die Vorschläge Antworten auf verschiedene Fragen sind. Man kann fragen, ob Selbstbewusstsein eine Fähigkeit oder ein Bewusstsein ist, und man kann fragen, inwiefern Selbstbewusstsein von objektiver Erkenntnis unterschieden ist. Den Zusammenhang der beiden Themen sehe ich folgendermaßen:¹³ Wenn man die These vertritt, dass Selbstbewusstsein eine Fähigkeit ist, dann lässt sich das mit den beiden hier vorgestellten Thesen verbinden: Man kann die Fähigkeit meinen, sich direkt auf seine Tätigkeit (des Urteilens) zu beziehen, oder die Fähigkeit, durch objektive Urteile auf die eigene Identität Bezug zu nehmen (indem man darauf schließt, dass es diese Identität geben muss). Wenn man die These vertritt, das Selbstbewusstsein ein spezi sches Bewusstsein ist, so lässt sich das hingegen nur mit der These eines von Gegenstandserkenntnis verschiedenen Selbstbewusstseins tun. Daher muss ich diese erste These vertre-

13 Dies scheint mir generell ein interessanter Punkt. Man muss sich überlegen, ob und wie die These, dass das „Ich denke“ eine Idee (Longuenesse „Kant on the Identity of Persons“) oder ein Gedankending (Rosefeldt „Kants Ich als Gegenstand“) ist, mit der These vereinbar ist, dass wir uns (wirklich) unserer selbst bewusst sind. Mir scheint, dass man beides zusammen nicht behaupten kann, sondern dass man dann, wenn man Selbstbewusstsein als Denknotwendigkeit, Idee oder Gedankending au asst, Selbstbewusstsein auch als eine Fähigkeit (diese Idee zu bilden) verstehen muss.

Kant über Selbstbewusstsein

61

ten. Sie scheint mir durch die Textstellen, die ich oben bereits aus Paragraph 16 der Deduktion angeführt habe, auch belegt zu werden. Für die zweite These sprechen Stellen wie die folgende: Also nur dadurch, daß ich ein Mannigfaltiges gegebener Vorstellungen in einem Bewußtsein verbinden kann, ist es möglich, daß ich mir die Identität des Bewußtseins in diesen Vorstellungen selbst vorstelle, d.i. die analytische Einheit der Apperzeption ist nur unter der Voraussetzung irgend einer synthetischen möglich (B 133).

Diese Stelle kann man so lesen, dass die analytische Einheit der Apperzeption meint, dass ich klarerweise ein Bewusstsein bilden muss, wenn die Vorstellungen miteinander verbunden werden, und dass ich diese analytische Einheit aufgrund der Synthese des Mannigfaltigen annehmen kann. Bei dieser Lesart muss kein besonderes Bewusstsein meiner selbst angenommen werden. Meine Identität wäre nur dadurch gegeben, dass ich mir die Einheit des Mannigfaltigen als durch mich bewirkt denken muss.¹⁴ Ich lese die Stelle dagegen so, dass das „synthetisch“ sich auch auf das Selbstbewusstsein bezieht, so dass das bedeutet: Ich kann mich selbst in der objektiven Einheit nur als identisches Wesen vorstellen, wenn ich auch ein Bewusstsein meiner synthetisierenden Tätigkeit habe. Auch diese Textstelle scheint mir daher für meine Lesart zu sprechen oder zumindest mit ihr kompatibel zu sein. Bisher habe ich allerdings noch wenig dazu gesagt, was bei meiner Interpretation unter „Bewusstsein“ positiv verstanden werden kann. Dies soll der nächste Abschnitt erläutern.

3 Selbstwahrnehmung Wenn für Kant Selbstbewusstsein ein Bewusstsein seiner selbst ist, das nicht nur vorliegt, wenn man sich explizit als Denker der Gedanken denkt, dann muss Kant etwas darüber sagen, was dieses Bewusstsein ist.¹⁵ Die These, dass „Bewusstsein“

14 Viele Stellen bei Kant lassen sich meines Erachtens für beide Interpretationen in Anspruch nehmen. Dies liegt daran, dass Kant die Abhängigkeit von Gegenstandserkenntnis in jedem Fall betont. Die hier entscheidende Frage ist daher: Wie weit geht diese Abhängigkeit? Bei der Antwort auf diese Frage kommt es darauf an, wie man Wendungen wie „Das ist aber soviel, als“ versteht (vgl. etwa B 135). Eindeutig ist die Textlage daher nicht. Zu der alternativen Interpretation vgl. z. B. Wunderlich, Falk (2005): Kant und die Bewußtseinstheorien des 18. Jahrhunderts. Berlin, 178. 15 Für Horstmann („Limited Signi cance“) ist Selbstbewusstsein das Korrelat für Objektbewusstsein. Aber da auch er annimmt, dass wir uns unseres Denkens bewusst sein müssen, beantwortet das nicht hinreichend die Frage, worin dieses Bewusstsein besteht. An dieser Stelle verstehe ich meine Überlegungen als Ergänzung.

62

Emundts

heißt, dass ich etwas denke, fällt hier, wie bereits deutlich geworden sein sollte, aus. Dies wäre eine unplausible These, denn dann müsste man annehmen, immer wenn wir irgendetwas denken, denken wir auch zugleich, dass wir es denken. Das scheint in der Tat absurd. Allerdings gibt es Stellen in der Kritik der reinen Vernun , die diese These durchaus nahe legen. So beispielsweise wenn Kant sagt, die Vorstellung des „Ich denke“ sei „ein Denken, nicht ein Anschauen“ (B 157).¹⁶ Man kann versuchen, diese Stellen so zu interpretieren, dass Kant hier nicht Selbstbewusstsein als Bedingung für Erkenntnis meint, sondern den besonderen Denkakt, mit dem wir uns explizit auf uns beziehen. Mir scheint es allerdings nahe liegender diese Stellen als Ausdruck einer gewissen Verlegenheit zu verstehen, in der sich Kant bei diesem Thema befunden hat. Denn aus Gründen, die ich noch ausführen werde, wollte Kant nicht die These vertreten, dass Selbstbewusstsein eine bestimmte Form von Selbstanschauung ist. Zugleich wollte er aber ein Bewusstsein unserer selbst annehmen, das auch dann besteht, wenn wir nicht explizit denken, dass wir denken. Wenn wir Kants Verlegenheit in diesem Punkt zur Kenntnis nehmen, fällt es leichter, Passagen aus Kants Texten zu verstehen, die sonst eher Befremden auslösen. Im Anschluss an das, was ich im ersten Abschnitt ausgeführt habe, will ich dies etwas genauer erklären. Was kann es überhaupt heißen, wenn ich sage, ich bin mir meiner selbst als Denker meiner Gedanken bewusst, wenn damit nicht gemeint ist, dass ich explizit denke, dass ich der Denker meiner Gedanken bin, und nicht, dass ich die Fähigkeit habe, mich als Denker meiner Gedanken zu denken? Vielleicht könnte es heißen, dass ich auf das Wissen, dass ich der Denker meiner Gedanken bin, jederzeit zugreifen kann. Diesem Vorschlag folgend ist mir etwas dann bewusst, wenn ich jederzeit auf es zugreifen kann.¹⁷ Tatsächlich sagt man damit etwas anderes als, dass Selbstbewusstsein eine Fähigkeit sei. Denn gemäß des jetzt diskutierten Vorschlags eines möglichen Zugri s nimmt man nicht an, dass eine Fähigkeit ausgeübt werden muss, um Bewusstsein davon zu haben, dass ich denke, dass p. Vielmehr wird behauptet, dass ich mir jederzeit dessen bewusst bin und es mir daher jederzeit explizit machen kann. Dennoch ist der Vorschlag nicht befriedigend. Das liegt vor allem an Folgendem: Dass man denkt, dass p, ist einem sicherlich nicht in der Weise präsent, in der einem präsent ist, dass zwei und zwei vier ist. Letzteres ist aber ein Wissen, für das gilt, dass man

16 Ausführlicher lautet die Stelle: „Dagegen bin ich mir meiner selbst in der transzendentalen Synthesis des Mannigfaltigen der Vorstellungen überhaupt […] bewußt, nicht wie ich mir erscheine, noch wie ich an mir selbst bin, sondern nur daß ich bin. Diese Vorstellung ist ein Denken, nicht ein Anschauen.“ 17 Diese Bedeutung von Bewusstsein grei Ned Blocks Vorschlag zum Bewusstsein (im Anschluss an Freud) auf.

Kant über Selbstbewusstsein

63

auf es (wenn man es weiß) jederzeit zugreifen kann. Das Problem bei diesem Vorschlag ist, dass man die Rede von etwas, auf das man zugreifen kann, vielleicht bei Überzeugungen oder Propositionen verstehen kann. Aber dass ich denke – also die Denktätigkeit und das mit ihr verbundene Selbstbewusstsein – ist nicht in dieser Weise abru ar. Es ist nicht falsch wie bei dem gerade diskutierten Vorschlag zu sagen, dass ich mir meiner selbst als Denker insofern bewusst bin, als ich mir dies jederzeit explizit machen kann. Aber der Fall verlangt eine andere Erklärung als die eines möglichen Zugri s auf Sachverhaltswissen.¹⁸ Dies ist sicherlich der Grund für die Interpretation von Selbstbewusstsein als einer Fähigkeit, weil man durch sie nicht darauf verp ichtet ist, dass man das Bewusstsein davon, Denker der eigenen Gedanken zu sein, so abrufen kann wie Sachverhaltswissen.¹⁹ Will man Selbstbewusstsein weder als möglichen Zugri auf ein Wissen noch als Fähigkeit verstehen, so muss man behaupten, dass wir, wenn wir denken, auch irgendwie faktisch gegenwärtig haben, dass wir denken. Diese Behauptung versetzt Kant jedoch auch wiederum in eine gewisse Verlegenheit. Eine Möglichkeit wäre zu sagen, dass wir wahrnehmen, dass wir denken. Diesen Vorschlag werde ich hier im Folgenden diskutieren. Es scheint, als würde Kant diesem Vorschlag ambivalent gegenüberstehen: Einerseits scheint er ihn ablehnen und andererseits dann doch in einer bestimmten Weise vertreten zu wollen. Die Gründe, die ihn zu dem einen und dem anderen bewogen haben, lohnen eine eingehende Untersuchung. Hierfür werde ich Kants Thesen zu Selbstwahrnehmung skizzieren. Dies werde ich tun, indem ich auf den Unterschied von Kants Au assung und heutigen Theorien der Selbstwahrnehmung eingehe. Auf diese Weise möchte ich zeigen, welche Probleme sich für Kant stellen, wenn er das Selbstbewusstsein, um das es ihm in seiner Transzendentalphilosophie geht, als eine Form der Selbstwahrnehmung au assen will. Meine These lautet, dass Kant „Bewusstsein“ und somit auch „Selbstbewusstsein“ als etwas verstanden wissen wollte, bei dem etwas jederzeit gegenwärtig ist, dass er dies aber von verschiedenen anderen Formen der Selbstwahrnehmung abgrenzen musste. Dies zu zeigen ist Aufgabe der folgenden Absätze. Selbstbewusstsein als Bedingung für Erkenntnis ist für Kant Selbstbezug bei der Tätigkeit des Denkens oder, vielleicht besser,²⁰ bei grundlegenden, mit unserem Intellekt zusammenhängenden Tätigkeiten. Mit anderen Worten: Im Selbst-

18 Man könnte hier vielleicht sagen: Anders als bei Sachverhaltswissen kann ich darauf, dass ich denke, nur zugreifen, wenn ich denke. Aber bei der These der Fähigkeit ist dies nicht klar. Die Interpretationen können sich daher auch darin unterscheiden, was unter „Denken“ verstanden wird. Hierauf werde ich zumindest kurz im vierten Abschnitt noch einmal zurückkommen. 19 Dies scheint mir bei John Perry besonders deutlich zu werden. 20 Dies werde ich im vierten Abschnitt genauer erläutern.

64

Emundts

bewusstsein als Bedingung für Erkenntnis sind wir uns unserer als denkend bewusst. Dies ist nicht die einzige Weise, wie wir auf uns selbst Bezug nehmen können. Wir können beispielsweise auch auf uns selbst Bezug nehmen, wenn wir uns bewusst werden, dass wir Schmerzen haben oder spazieren gehen. Und wir können auf uns selbst als jemanden Bezug nehmen, der eine Person ist, die bestimmte Eigenscha en und eine Geschichte hat. All dies kann man auch nach Kant als Formen von Selbstbewusstsein ansehen. Fälle von Selbstbezügen bei mentalen Zuständen generell, also beispielsweise Schmerzemp ndungen, werden häu g – besonders in den philosophischen Diskussionen über Selbstwahrnehmung in den letzten vierzig Jahren – als eine besondere Art eines Selbstbezugs angesehen und unter Titeln wie Selbstbewusstsein, Selbstwahrnehmung oder Selbstwissen thematisiert.²¹ Das Besondere des Selbstbezugs in solchen Fällen soll unter anderem darin bestehen, dass wir uns auch dann erfolgreich auf uns beziehen könnten, wenn wir alles über uns als bestimmte Individuen oder Personen vergessen hätten.²² Dadurch wird nahegelegt, dass wir uns unmittelbar auf uns beziehen. Weiterhin liegt eine Besonderheit darin, dass die Frage, ob man sich sicher sei, wer Schmerzen habe, unsinnig ist. Dies legt nahe, dass ich keine Identi zierung von mir als körperlicher Person vornehmen muss, um mir erfolgreich Schmerzen zuzuschreiben. Es gibt, so die Überlegung, nicht die Möglichkeit einer Fehlidenti kation des Subjekts der Zuschreibung. Damit scheint auch klar – was vielleicht weniger o ensichtlich, für das Folgende aber wichtig ist –, dass wir uns hier gar nicht als ein kontinuierliches Wesen verstehen müssen, sondern nur im Moment auf uns Bezug nehmen. Von dieser Art des Selbstbezugs lassen sich Fälle unterscheiden, in denen wir uns mehr oder weniger so auf uns beziehen, wie wir uns auch auf andere Personen beziehen. So können wir uns aufgrund unseres Äußeren Merkmale zuschreiben, wie wir es im Prinzip auch bei anderen tun. Und wir nehmen auf uns als bestimmte Personen, die eine gewisse Kontinuität haben, Bezug, wenn wir uns beispielsweise die Handlung des Zuspätkommens zuschreiben. In der gegenwärtigen Diskussion werden (im Anschluss an Wittgenstein) die beiden hier unterschiedenen Arten des Selbstbezugs zumeist durch die Unterscheidung eines Subjekt- und eines Ob-

21 Ich orientiere mich hier (wie üblich) vor allem an Shoemaker, Sydney (1968): „Self-Reference and Self-Awareness“, in: The Journal of Philosophy, Vol. 65, No. 19, 555–567. 22 Damit die gemeinte Besonderheit mentaler Zustände zur Geltung kommt, muss man an extreme Fälle denken, bei denen sogar jegliches Körpergefühl verloren ist (vgl. Anscombe, G. E. M. (1975): „The First Person“, in: Mind and Language, ed. by S. Guttenplan, Oxford, 45–65). Die verschiedenen Fälle körperlichen Bezugs müssen verschieden analysiert werden, aber dies ist hier nicht meine Aufgabe; vgl. Prosser, Simon; Recanati, François (eds.) (2012): Immunity to Error through Misidenti cation: New Essays. Cambridge.

Kant über Selbstbewusstsein

65

jektgebrauchs des Ich-Ausdrucks charakterisiert. Angesichts dessen, dass bei den ersten Fällen eine Art unmittelbarer Selbstwahrnehmung angenommen wird, während wir uns bei den anderen auf uns so wie auf andere empirische, raum-zeitliche Gegenstände beziehen, könnte man ebenso die Möglichkeit einer unmittelbaren Selbstwahrnehmung von Fällen empirischen Selbstbewusstseins unterscheiden, wobei „empirisches Selbstbewusstsein“ heißen soll, dass wir uns auf uns wie auf (andere) empirische Gegenstände beziehen. Es ist wichtig zu sehen, dass Kant den soeben skizzierten Unterschied nicht macht. Er unterscheidet nicht Fälle von Selbstwahrnehmung bei mentalen Zuständen generell und Fälle von empirischem Selbstbewusstsein. Dies jedenfalls nicht in der Weise, dass er zwischen den beiden einen kategorialen und philosophisch besonders interessanten Unterschied gesehen hätte. Kant kennt, wie ich später ausführen werde, nur den Akt des Denkens als Fall eines besonderen Selbstbezugs, der auch philosophisch wirklich interessant ist. Dagegen denkt er bei Schmerzen und anderen Emp ndungen Selbstwahrnehmung als einen Fall von empirischem Selbstbewusstsein. Zwar macht er innerhalb des empirischen Selbstbewusstseins einen Unterschied von innerem und äußerem Sinn. Beides sind aber Bezüge auf sich, die als Fälle des empirischen Selbstbewusstseins zu gelten haben. Dass dies so ist, sieht man daran, dass Kant in der Deduktion der Kritik der reinen Vernun den Selbstbezug von mir als denkendem Ich von allen anderen Formen des Selbstbezugs unterscheidet und hierbei alle anderen Formen als Fälle des empirischen Selbstbewusstseins begrei . Unter die empirischen Fälle von Selbstbewusstsein fallen dann beispielsweise auch Schmerzemp ndungen. Kant nimmt folglich auch an, dass ich im Akt der Selbstwahrnehmung von mir als Schmerzen habend auf mich als ein Wesen Bezug nehme, das in der Zeit kontinuierlich existiert. Wenn ich meine Schmerzen wahrnehme, so nehme ich wahr, dass ich Schmerzen habe, und hierbei nehme ich auf mich als etwas Bezug, das nicht nur punktuell existiert. Hier könnte man im Sinne Kants hinzufügen, dass es nur dadurch, dass ich auf mich als kontinuierliches Wesen Bezug nehme, auch überhaupt möglich ist, meinen Zustand der Schmerzemp ndung mit anderen Zuständen in Verbindung zu setzen.²³ Wenn Kant dies annimmt, dann nimmt er auch an, dass ich in der Selbstwahrnehmung der Schmerzen zumindest indirekt auf mich als Objekt Bezug nehme.

23 In diesem Punkt scheint mir Evans These zum Selbstbewusstsein derjenigen von Kant ähnlich zu sein, da auch bei Evans die Idee zu nden ist, dass wir uns auf uns als kontinuierliche Wesen beziehen müssen, um uns überhaupt etwas als Eigenscha zuzuschreiben; vgl. Evans, Gareth (1982): The Varieties of Reference, Oxford, 205–233.

66

Emundts

Denn „Objekt“ kann man nennen, was über die Zeit hinweg und trotz Veränderungen seiner Eigenscha en dasselbe ist. Kant redet im Zusammenhang von Selbstwahrnehmung bei Emp ndungen auch ausdrücklich vom Ich als „Objekt“ (B 157) und von „Selbstanschauung“. Dies wir allerdings die Frage auf, in welchem Sinn wir in der Selbstwahrnehmung oder Selbstanschauung auf ein Objekt Bezug nehmen. Hier scheint es zwei Möglichkeiten zu geben: Man kann ein inneres Selbst annehmen, das uns durch den inneren Sinn präsent ist. Das wäre ein Selbst bestehend aus Wünschen, Gedanken und Gefühlen. Oder man kann annehmen, dass ich mich auch in der Wahrnehmung meiner mentalen Zustände auf mich als körperliches Wesen beziehe, also auf ein Objekt in Raum und Zeit, das sich in einem bestimmten Zustand be ndet. Kant ist in diesem Punkt nicht immer ganz klar. Meines Erachtens muss er jedoch für die zweite Option eintreten – und hat dies letztlich, nämlich in der 2. Au age der Kritik der reinen Vernun – auch deutlich getan.²⁴ Dass ich Kant dazu verp ichtet sehe, das Selbst als Objekt als ein räumliches Ding aufzufassen, liegt kurz gesagt an Folgendem: Nach Kants sogenannten Analogien der Erfahrung sind Kontinuität und Beharrlichkeit Eigenscha en, die wir nur einem Ding mit räumlicher Ausdehnung zuschreiben können. Dies leuchtet auch mit Blick auf die Vorstellung des Bezugs auf uns selbst ein: Wenn man von einem Selbst reden will, muss man sich Wünsche, Gedanke und Gefühle als Zustände eines einheitlichen, identi zierbaren Wesens vorstellen. Ein solches Wesen ist für uns auch körperlich. Selbstwahrnehmung ist eine besondere Wahrnehmung (mit dem „inneren Sinn“) von inneren Zuständen einer Person, die auch körperlich sein muss. Es gibt also eine Art innerer Wahrnehmung bei mentalen Zuständen, die mentalen Zustände sind aber als Zustände einer körperlichen Person anzusehen und nicht zu isolieren. Außer beim Denken sind für Kant damit alle Formen des Selbstbezugs klarer Weise Fälle empirischen Selbstbewusstseins. Vermeintlich punktuelle Selbstwahrnehmung gibt es für Kant nicht – jedenfalls nicht, wenn man von dem für Kant besonderen Fall des Selbstbezuges beim Denken absieht. Unsere Schmerzen nehmen wir wahr, indem wir auf uns als körperliche Wesen Bezug nehmen. Folgt man meinen Ausführungen, so vertritt Kant also nicht die These, dass es bei mentalen Zuständen generell eine Besonderheit des Ich-Ausdrucks gibt, die von philosophischer Relevanz ist. Er betrachtet vielmehr die meisten Fälle von Selbstbezug auch bei mentalen oder „inneren“ Zuständen als Fälle, in denen wir

24 Dafür habe ich an anderen Stellen argumentiert, u.a. Emundts, Dina (2007): „Kant über innere Erfahrung“, in: Was ist und was sein soll: Natur und Freiheit bei Immanuel Kant, hg. v. U. Kern, Berlin, 189–205.

Kant über Selbstbewusstsein

67

uns – in vielleicht unterschiedlicher Weise – wahrnehmend auf uns als empirische Wesen beziehen. Es gibt einschlägige Kritik an der Vorstellung, dass wir einen inneren Sinn haben oder zu innerer Wahrnehmung in der Lage sind. Zwei Kritikpunkte möchte ich kurz ansprechen und auf Kants Theorie des empirischen Selbstbewusstseins beziehen. (1)Shoemaker kritisiert das Modell einer inneren Wahrnehmung vor allem mit dem Ziel, der Besonderheit des Ich-Ausdrucks bei der Zuschreibung mentaler Prädikate Rechnung zu tragen. Er will bei mentalen Zuständen einen Selbstbezug annehmen, bei dem wir uns nicht identi zieren müssen, der aber gleichzeitig nicht in einen Cartesianismus führt. Wer sich beim Selbstbezug bei mentalen Zuständen am Wahrnehmungsmodell orientiert, der sucht, so Shoemaker, das Objekt, das wir als Träger der Prädikate annehmen können. Wenn man hier ein Objekt sucht, so kommt man zur Annahme eines inneren Selbst, das heißt zu einer Art Cartesianismus. Diese Suche wird, so Shoemaker, dann als unsinnig entlarvt, wenn man bedenkt, dass das Wahrnehmungsmodell im Fall des Bewusstseins meiner mentalen Zustände (zum Beispiel Schmerzen) unzulänglich ist. Anders als bei mentalen Zuständen können wir bei Wahrnehmungen eine Sache aus verschiedenen Perspektiven betrachten. Diese und weitere Unterschiede verbieten die Analogie des Wahrnehmungsmodells für den Fall mentaler Zustände. Die Konsequenz, die Shoemaker aus der Unzulänglichkeit des Wahrnehmungsmodells zieht, ist die, dass die Besonderheit des Selbstbezugs nicht zu einem cartesianischen Selbst führt, weil der Selbstbezug ganz anderer Natur ist. Auf diese Weise stellt Shoemaker alle Philosophen, die eine innere Wahrnehmung oder einen inneren Sinn annehmen, unter den Verdacht, hier mit falschen Vorgaben zu operieren. Dieser Verdacht erweist sich bei Kant als unbegründet. Denn wenn das, was ich oben ausgeführt habe, richtig ist, so nimmt Kant nicht an, dass es das Selbst als wahrnehmbaren inneren Gegenstand gibt. Er nimmt vielmehr an, dass man zu seinen inneren Zuständen als einem körperlichen Gegenstand einen besonderen Zugang hat. Diesen besonderen Zugang bezeichnet Kant als „inneren Sinn“.²⁵

25 Wenn man Kants Thesen zum inneren Sinn so liest wie ich und die Konzeption so zusammenfasst, dass es bei mentalen Zuständen nur einen besonderen Zugang zum körperlichen Objekt gibt, dann könnte man fast sagen, dass Shoemaker und Kant hier gar nicht weit voneinander entfernt sind (außer dass für Kant das Denken als mentaler Zustand vor allen anderen mentalen Zuständen ausgezeichnet wird). Bei Shoemaker ist dieser Punkt nicht ganz leicht zu verstehen, denn er nimmt an, dass der Ich-Ausdruck immer auf eine körperliche Person referiert, beschreibt aber die Besonderheit des Selbstbezuges so, dass man den Eindruck hat, dass der Ich-Ausdruck bei mentalen Zuständen nicht referieren muss. Ich werde hierauf im vierten Abschnitt noch einmal zurückkommen.

68

Emundts

Weiter nimmt Kant an, dass ich mir diese Zustände nur dann als meine zuschreiben kann, wenn ich mich zugleich auch als kontinuierliches, körperliches Wesen ansehe. (2) Gegen die Annahme, dass wir Zustände wie Schmerzen wahrnehmen, ist von Nagel (gegen Armstrong) vorgebracht worden, dass wir diese Zustände direkt erleben und daher Wahrnehmung das falsche Modell ist.²⁶ Auf einen solchen Einwand würde Kant wohl sagen, dass das direkte Erlebnis für uns nur dann überhaupt etwas sein kann, auf das wir uns auch beziehen können, wenn wir es als ein zeitlich zu lokalisierendes – punktuelles oder sich in der Zeit erstreckendes – Ereignis au assen. Dann ist damit laut Kant auch die Vorstellung eines in der Zeit kontinuierlich existierenden Wesens gegeben. In diesem Fall ist aber die Rede von Wahrnehmung oder Anschauung gerechtfertigt. Kant muss also nicht leugnen, dass es direktes Erleben mentaler Zustände gibt, für die Bezugnahme auf diese Erlebnisse muss aber noch etwas hinzukommen, und dies kann man in der bisher erläuterten Weise als eine Form der Wahrnehmung oder Anschauung (ich verwende Wahrnehmung und Anschauung hier synonym) ansehen. Ich kann diese Punkte hier nicht ausdiskutieren. Ich wollte mit ihnen Kants Theorie eines möglichen wahrnehmenden oder anschauenden Selbstbezugs bei mentalen Zuständen nur sehr grob verteidigen. Die gängigen Einwände gegen innere Wahrnehmung widerlegen diese Theorie nicht (oder zumindest nicht ohne weiteres). Bei all dem, was ich zu Kants Theorie der Selbstwahrnehmung bisher ausgeführt habe, ist nun zu berücksichtigen, dass Kant die Selbstzuschreibung von Gedanken elementar von der Selbstzuschreibung anderer mentaler Prädikate unterscheidet. Was ich über den Bezug auf die eigenen mentalen Zustände ausgeführt habe, gilt nach Kant nicht beim Denken oder bei allen (noch näher zu erläuternden) grundlegenden intellektuellen Tätigkeiten.²⁷ Dies sieht man daran, dass Kant in der Deduktion mit Blick auf Selbstbewusstsein als Bedingung für Erkenntnis entschieden verneint, dass hier ein Fall von Selbstanschauung vorliegt (B 158). Warum er meint, dies tun zu müssen, werde ich im letzten Abschnitt meines Auf-

26 Nagel, Thomas (1980): „Armstrong on the Mind“, in: Readings in Philosophy of Psychology, ed. by N. Block, Vol. 1, Cambridge, 200–206. 27 Man kann daher auch sagen: Das Thema Selbstbewusstsein als Bedingung für Erkenntnis wird für Kant nicht virulent als eine Antwort auf die Frage, was das Besondere an unserem Selbstbezug ist, sondern was das Besondere an unserem Bezug auf uns als denkende Wesen ist. Ähnlich in diesem Punkt Carl „Ich und Spontaneität“, 115. Daher ist auch zu notieren, dass Kant (anders als beispielweise Shoemaker am Ende des angeführten Aufsatzes) die Selbstzuschreibung mentaler Prädikate nicht als notwendige Bedingung für Erkenntnis ansieht.

Kant über Selbstbewusstsein

69

satzes noch ausführen. Zunächst soll untersucht werden, was das für die bisher diskutierte These zu Bewusstsein bzw. Selbstbewusstsein heißt. Indem Kant innere Wahrnehmung oder Anschauung im Fall des Denkens ausschließt, schließt er eine Möglichkeit der Explikation von Bewusstsein aus, nämlich die, die für Anschauung charakteristisch ist. Denn bei der Selbstanschauung als Objekt können wir erläutern, was es heißt, dass uns etwas (also zum Beispiel eine Schmerzemp ndung) bewusst ist. Es kann uns etwas in der Wahrnehmung gegenwärtig sein, ohne dass wir explizit an es denken. „Bewusst“ bedeutet demnach, etwas im Modus des Denkens oder des Anschauens oder Wahrnehmens präsent zu haben. Man könnte also annehmen, dass wir unsere Denktätigkeit wahrnehmend präsent haben und uns jederzeit explizit auf sie beziehen können. Aber o ensichtlich ist für Kant die Möglichkeit der Selbstanschauung von uns als denkend keine Option für die Erklärung dafür, wie wir uns unserer selbst denkend bewusst sind. So erklärt sich, wie es bei Kant zu Äußerungen wie der früher erwähnten kommt, mit denen Kant explizit sagt, dass die Vorstellung „Ich denke“ keine Anschauung ist. Es ist aber bemerkenswert, dass Kant sich auch mit der negativen Antwort, dass das „Ich denke“ keine Anschauung ist, nicht zufrieden geben will. Er ringt vielmehr damit, etwas zu nden, durch das wir uns unseres Denkens bewusst sein können, ohne damit zu behaupten, dass wir explizit denken müssen, dass wir denken. Damit sucht er etwas, das keine Selbstanschauung ist, aber dieselbe Funktion einnimmt. Kant thematisiert die Weise des bewussten Selbstbezugs nicht in den Passagen, in denen er Selbstbewusstsein als Bedingung für Erkenntnis analysiert, also in der Deduktion der Kritik der reinen Vernun . Denn hier will er zeigen, dass wir uns unserer selbst bewusst sein müssen. Er betont in diesem Zusammenhang, dass dies nicht heißt, dass wir uns unserer als Objekt – also im Modus der Selbstanschauung – bewusst sind. Das ist für ihn wichtig, weil er die Form des Selbstbezugs beim Denken von anderen Formen des Selbstbezugs (wie der Schmerzempndung) streng unterscheiden will.²⁸ Wie sich das Bewusstsein meiner selbst im Falle des Denkens positiv ausbuchstabieren lässt, thematisiert Kant hingegen an späterer Stelle der Kritik der reinen Vernun und in den Prolegomena. Statt einer Selbstanschauung als Objekt behauptet Kant für das Denken eine besondere Form des Selbstbezugs. Obwohl diese Passagen bei den Lesern Erstaunen hervorrufen, können sie nach dem, was

28 An einigen Stellen klingt es hier sogar (wie schon angeführt) so, als wolle Kant sagen, dass wir das Denken bewusst nur im Modus des Denkens haben können. Dies wäre als Bedingung für Erkenntnis aber, wie ausgeführt, nicht überzeugend.

70

Emundts

ich ausgeführt habe, nicht mehr wirklich überraschen: Nach all dem muss sich etwas darüber sagen lassen, wie ich mir selbst beim Denken meines Denkens bewusst bin, ohne dass ich explizit denke, dass ich denke. Da die Option einer Selbstanschauung als Objekt für Kant o enbar ausfällt, muss er hier nach Alternativen suchen. Die Alternative, die Kant in wiederum verschiedener Weise ins Auge fasst, ist meines Erachtens eine Art der Selbstvergegenwärtigung, bei der es zu keiner Anschauung als Objekt kommt. So sagt Kant in der Kritik der reinen Vernun : Das „Ich denke“ „drückt […] eine unbestimmte empirische Anschauung, d.i. Wahrnehmung“ aus; eine Wahrnehmung, so Kant weiter, die „vor aller Erfahrung vorher (geht) […]“ (B 422) – das heißt für Kant: Eine Wahrnehmung, in der ich mich noch gar nicht auf mich als Objekt beziehen kann – aber dieser Zusatz nimmt bereits auf den Grund für die Besonderheit der Selbstvergegenwärtigung Bezug, den ich erst im nächsten Abschnitt erläutern werde.²⁹ In den Prolegomena spricht Kant vom „Ich denke“ als einer unmittelbaren Anschauung (334). Dort behauptet er weiter, dass das Ich gar kein Begri sei, der einen Gegenstand bezeichne, sondern „nichts mehr als ein Gefühl eines Daseins“. Hier will Kant darauf hinaus, dass wir uns mit der Vorstellung „Ich denke“ nicht nur auf uns als einen Gegenstand beziehen können – dies tun wir mit dem Begri „Ich“ als einer Art Vorstellung, und wir tun es in anderen Fällen des Selbstbezugs als der des Denkens – es gibt vielmehr, so Kant, Weisen des Gegenwärtigseins des Denkens, in der mir kein Objekt gegenwärtig ist. Dass Kant das Gegenwärtigseins der Denktätigkeit in dieser Weise ausführt, hängt o ensichtlich mit dem Grund zusammen, warum er die Selbstanschauung in diesem Fall verneint. Den werde ich im folgenden Abschnitt ausführen.

4 Selbstbewusstsein als Bedingung für Erkenntnis Worin liegt für Kant die Besonderheit des Denkens? Sich seiner in seiner Tätigkeit des Denkens bewusst werden, kann man nur, wenn man denkt.³⁰ Dies kann aber für Kant gerade nicht die Besonderheit sein. Denn das ist bei Schmerzen ähnlich, da man sich nur als Schmerzen habend wahrnehmen kann, wenn man Schmerzen

29 Vgl. zu dieser Stelle auch: Emundts, Dina (2006): „Die Paralogismen und die Widerlegung des Idealismus in Kants Kritik der reinen Vernun “, in: Deutsche Zeitschri für Philosophie 54, 295–309. 30 Ich habe schon einmal darauf aufmerksam gemacht, dass es wichtig ist, auf die Tätigkeit Bezug zu nehmen (vgl. Fn. 12). Durch die Parallele zur Selbstwahrnehmung bei mentalen Zuständen erscheint dies jetzt nicht mehr so seltsam, wie wenn man über das Urteilen spricht.

Kant über Selbstbewusstsein

71

hat.³¹ Die Wahrnehmung von Schmerzen erklärt Kant, wie im letzten Abschnitt deutlich geworden sein sollte, als eine Art der Wahrnehmung meiner inneren Zustände eines zugleich körperlichen Objekts. Aber beim Denken soll diese Form der Selbstanschauung nicht möglich sein. Beim Denken gibt es laut Kant eine Besonderheit, die darin besteht, dass es ein Bewusstsein des Denkens gibt, bei dem man das Denken nicht einem Objekt als Träger der Tätigkeit zuschreibt. Als Grund dafür, im Denken eine besondere Art von mentalen Zuständen zu sehen, bei der keine Selbstanschauung möglich ist, nennt Kant, dass Selbstanschauung eine Ordnung in der Zeit voraussetzt, das Bewusstsein meiner als denkend aber dieser Ordnung in der Zeit vorhergehen muss, weil es diese Ordnung bedingt (B 158). Diese Begründung ndet sich bei Kant in verschiedenen Varianten: So heißt es zum Beispiel auch, dass die Tätigkeit des Denkens das Bestimmende und nicht das Bestimmbare sei und als solches (als das Bestimmende) nicht bestimmt werden könnte (B 158). Man kann das, was Kant meint, mithilfe einer Analogie erklären: Sich selbst als denkend anschauen, wäre so wie ein Film, der zeigt, wie dieser Film selbst als ganzer geschnitten wird. Das ist unmöglich. Den Prozess seiner vollständigen Entstehung kann der Film nicht zum Gegenstand machen.³² Diese Analogie soll unter anderem zeigen, dass die Pointe von Kants Überlegungen in seinem Idealismus liegt. Unter Kants Idealismus verstehe ich hier die These, dass wir das gegebene Mannigfaltige nach objektiven Regeln verbinden müssen, um überhaupt auf eine strukturierte Welt mit Objekten Bezug nehmen zu können. Zu dieser strukturierten Welt gehören auch wir als Objekte der Welt. Wenn wir ohne Denken gar nicht auf uns als Objekt bzw. als eine raum-zeitliche Einheit Bezug nehmen können, dann setzt die Selbstanschauung als Objekt oder als Einheit das Denken voraus. Soll uns dieses Denken gegenwärtig sein, kann es dies nicht in Form einer Selbstanschauung als Objekt. Selbstanschauung ist hier nicht möglich, weil das, was man anschauen würde, erst Produkt des Prozesses ist. Es scheint, als müsse Kant aufgrund dieses Arguments alle Formen der phänomenalen Vergegenwärtigung ausschließen, und in der Deduktion klingt dies auch zuweilen so.³³ Wenn er Selbstbewusstsein als eine Form des Bewusstseins ansieht, scheint er aber auch irgendeine Form der Vergegenwärtigungsmöglichkeit einräumen zu müssen. In diesem Sinn lässt sich die im letzten Abschnitt er-

31 Ein besonderer Fall sind Phantomschmerzen, aber dies kann ich hier nicht diskutieren. Ich denke, auch dann kann man sagen, dass man Schmerzen hat. 32 Für dieses Bild danke ich Martin Otter. 33 Vgl. zum Beispiel die oben angeführte Stelle zum Bestimmen, wo davon die Rede ist, dass ich mir die Spontaneität meines Denkens vorstelle (B 158).

72

Emundts

wähnte Stelle aus den Prolegomena jetzt weiter erläutern: Das Ich sei, so Kant dort, „nichts mehr als Gefühl eines Daseins ohne den mindesten Begri und nur Vorstellung desjenigen, worauf alles Denken in Beziehung steht“ (334).³⁴ Die Pointe liegt hier darin, dass uns das Ich als etwas Reales in der unmittelbaren Anschauung gegenwärtig sein kann, ohne dass wir es als solches (also ohne Bezugnahme auf anderes) erkennen können oder erkennen können, dass und inwiefern es ein Gegenstand ist. Den Ausdruck „Gefühl“ wählt Kant, um die Rede von Selbstanschauung zu vermeiden. Von Selbstanschauung will Kant nicht sprechen, weil Anschauungen einen Gegenstand als bestimmte, einheitliche Entität haben. Bezüglich unseres Denkens soll uns aber etwas präsent sein, das wir nicht als etwas Bestimmtes ansehen oder begreifen können. Der Selbstbezug im Denken ist für Kant aus diesem Grund ein Bezug auf etwas, das evasiv und radikal subjektiv ist.³⁵ Evasiv ist es, insofern es sich trotz seiner Präsenz nicht xieren lässt. Folgt man meiner Darstellung, so liegt dies daran, dass die Tätigkeit, auf die wir uns beziehen, die Bedingung der Fixierung von etwas darstellt. Radikal subjektiv ist das, worauf wir uns beziehen, indem es sich als die jede Form von Objektivierung bedingende Tätigkeit per se nicht zum Objekt machen lässt. Und das Phänomen der Präsenz ist hier o enbar genauso schwer zu beschreiben wie dasjenige, was wir präsent haben. Jedenfalls ist bei diesem Vorschlag schwer genau anzugeben, was mit „Wahrnehmung“ oder „Anschauung“ oder „Gefühl“ gemeint ist. Es ist ein Gegenwärtighaben von etwas, das sich so, wie es gegenwärtig ist, nicht als Objekt bestimmen lässt. Wichtig für meine Fragestellung in diesem Aufsatz ist vor allem das Ergebnis, dass Kant zwar das Modell der Selbstanschauung ablehnt, dass er aber dennoch daran festhält, dass es eine Form der Präsenz gibt. Diese besteht nicht darin, dass ich explizit denke, dass ich denke, sondern sie besteht in einer besonderen Form der unbestimmten oder unmittelbaren Anschauung. Damit ist auch klar, dass Kant Bewusstsein nicht als bloße Fähigkeit verstehen will. Denn in diesem Fall ist die Annahme eines besonderen Modus des Gegenwärtighabens nicht sinnvoll. Es ist zu vermuten, dass es dafür, dass Kant Selbstbewusstsein nicht als Fähigkeit verstehen will, Gründe gibt, die auch mit seiner idealistischen Konzeption zusammenhängen: Man könnte beispielsweise sagen, dass die Annahme von

34 Mohr bezieht diese Stelle der Prolegomena auf die Apperzeption, nicht auf das Ich als Gegenstand einer unmittelbaren Anschauung. Allerdings stimme ich mit ihm überein, dass hier nicht von einer Selbstanschauung die Rede ist, die Identitätsbewusstsein liefern könnte: Durch die unmittelbare Anschauung wird gerade kein Gegenstand gegeben; Mohr, Georg (1991): Das sinnliche Ich. Innerer Sinn und Bewusstsein bei Kant. Würzburg, 65. 35 Die These der Evasivität und Subjektivität in diesem Sinn vertritt Horstmann in allen in Fußnote 1 genannten Texten.

Kant über Selbstbewusstsein

73

Selbstbewusstsein als Fähigkeit nur Sinn macht, wenn wir die Fähigkeit als Fähigkeit einer körperlichen Person bestimmen können. Folgt man dieser Überlegung, so gilt, dass Kant aufgrund seines Idealismus’ die These der Fähigkeit gar nicht vertreten konnte. Bei Kant ist dasjenige, dessen wir uns bewusst sind, etwas, bei dem wir gerade nicht sagen können, was es ontologisch gesehen ist. Die Bestimmung von mir als Objekt, auf das ich mich als etwas beziehen kann, das Fähigkeiten hat, ndet erst durch die bewusste Tätigkeit statt. Dies wird besonders in der oben angeführten Stelle aus den Prolegomena deutlich. Das heißt natürlich nicht, dass Kant ein cartesianisches Selbst annimmt; wir werden uns ja auch nicht einer geistigen Substanz bewusst. Was das „Ich denke“ ist oder wessen Tätigkeit es ist, kann, so ist gerade die Pointe, durch das Bewusstsein dieser Tätigkeit nicht bestimmt werden. Folgt man meiner Darstellung, stehen die Bestimmungen der Selbstrelation als evasiv und subjektiv o ensichtlich mit Kants idealistischer Konzeption in engstem Zusammenhang. Es gilt zu bedenken, dass dieser Idealismus bei Vielem, was auch Bestandteil von Kants Theorie des Selbstbewusstseins ist, nicht o ensichtlich ist bzw. dass Vieles aus Kants Theorie des Selbstbewusstseins nicht per se auf einen Idealismus verp ichtet. Selbstbewusstsein als Bedingung für Erkenntnis, so wie ich es zu Beginn dieses Aufsatzes erläutert habe, setzt zum Beispiel keineswegs einen Idealismus voraus. Ich habe (wie andere zeitgenössische Philosophen) die Argumentation zwar in Anlehnung an Kant gegeben, aber sie nur auf Erkenntnis in einem anspruchsvollen Sinn bezogen und nicht, wie bei Kant selbst, schon auf Gegenstände als raum-zeitlich geordnete Einheiten. Man muss also zwischen den anfangs ausgeführten Überlegungen dazu, warum Selbstbewusstsein eine Bedingung für Erkenntnis ist – nämlich aus dem Grund, dass wir uns als Denker unserer Gedanken denken müssen, um objektiv zu urteilen –, und Kants Überlegungen unterscheiden. Ich will damit nicht behaupten, dass nicht auch Kant sagen würde, dass wir Selbstbewusstsein annehmen müssen, weil wir uns als Denker unserer Gedanken denken müssen, um objektiv zu urteilen. Aber hier muss man bedenken, dass Kant die Herstellung einer raum-zeitlichen Ordnung von Mannigfaltigem mit der Tätigkeit des Urteiles in einen engen, aber bekanntlich auch komplexen Zusammenhang stellt. Dies wir die wichtige Frage auf, was in den vorliegenden Ausführungen eigentlich mit „Denken“ gemeint ist. Aufgrund der idealistischen These entsteht der Eindruck, dass es Kant gar nicht um alle Formen von Denktätigkeiten geht, sondern man vielmehr sagen muss, dass es für Kant eine spezielle Denktätigkeit gibt, für die es keine innere Wahrnehmung des Denkens geben kann, in welcher ich mir zugleich Objekt bin. Diese spezielle Denktätigkeit ist die, mit der ich Mannigfaltiges gemäß den Kategorien synthetisiere. Allerdings steht diese Tätigkeit für Kant mit der des Urteilens in einer so engen Verbindung, dass man den Selbstbezug, bei dem ich mir nicht Objekt werden kann, auch auf

74

Emundts

das objektive Urteilen beziehen muss. Wie dies genau aussehen soll, ist erläuterungsbedür ig. Wenn man die Thesen zum Selbstbewusstsein auch auf das Urteilen bezieht, so kann jedenfalls Kants oben zusammengefasste Begründung dafür, dass hier keine Selbstanschauung möglich ist, nur dann überzeugen, wenn wir ohne Urteilen auf gar kein Objekt Bezug nehmen könnten. Denn sonst könnte man ja annehmen, dass erst die synthetisierende Tätigkeit statt ndet und wir uns dann auf uns als Objekte, die die Tätigkeit des Urteilens ausüben, beziehen können. Es scheint mir daher klar, dass die Theorie von Selbstbewusstsein, die ich Kant unterstelle, die Handlungen des Synthetisierens und die des Urteilens als dieselbe Tätigkeit in einem ziemlich strengen Sinne au assen muss.³⁶ Um die Frage, was hier „Denken“ heißt, ausführlicher zu beantworten, müsste man ausführen, wie man den Zusammenhang der beiden Tätigkeiten und wie man deren Zusammenhang zum Thema Selbstbewusstsein genau versteht. Dies kann ich hier nicht tun. Ich werde daher wie bisher von Denken im Sinne von „grundlegenden intellektuellen Tätigkeiten“ reden. Darunter fällt auf jeden Fall unsere Syntheseleistung von Mannigfaltigem. Alternativ zu meiner These, dass Evasivität und Subjektivität des Selbstzugs primär mit Kants Idealismus zusammenhängen, lässt sich auch behaupten, dass diese vielmehr mit der Eigenart des Selbstbewusstseins zu tun haben. Dies scheint mir die zumeist – zumindest implizit – vertretene These zu sein, wenn versucht wird, an Kants Theorie des Selbstbewusstseins als einer Art impliziten Wissens positiv anzuschließen. Will man sich seiner selbst bewusst werden, so ist man, wie eine berühmte These mit philosophischer Tradition besagt, in Zirkeln gefangen, weil man das, was man sich bewusst machen will, immer wieder voraussetzen muss. Ich behaupte, dass der Grund bei Kant für die These der Evasivität der Idealismus (im oben erläuterten Sinn) ist. Auch bei einigen anderen Philosophen wie beispielsweise Sartre ist dies der Fall.³⁷ Dennoch entsteht o der Eindruck, als hinge die Weise des Selbstbezugs mit einer Eigenart von Selbstbeziehungen zusammen, die unabhängig vom Idealismus besteht. Dies liegt daran, dass es erstens phänomenologisch überzeugend ist, dass es einen Bezug auf uns selbst gibt, bei dem sich dasjenige, was erfasst werden soll, per se als Gegenstand entzieht. Zweitens (damit zusammenhängend) gibt es Eigenarten der Selbstbeziehungen, welche Ähnlichkeit mit dem haben, was Kants Theorie besagt. Auf diese nehmen die neueren Theorien über Selbstbewusstsein Bezug. Und auch sie sprechen o

36 Vgl. hierzu vor allem Horstmann in diesem Band. 37 Ebenso scheint mir klar, dass dies bei Horstmann („Limited Signi cance“ und in diesem Band) der Fall ist, auch wenn es hier so klingt, als ginge es (auch) um die genuine Eigenart des Selbstbezugs.

Kant über Selbstbewusstsein

75

von einer besonderen Subjektivität des Selbstbezugs bei mentalen Zuständen. Es ist so, als habe die Theorie von Kant durch die zeitgenössischen Theorien der Selbstwahrnehmung Unterstützung bekommen. Letzteres scheint mir allerdings tatsächlich nicht der Fall zu sein. Wie ich schon ausgeführt habe, unterscheidet man im Anschluss an Wittgenstein und Shoemaker in der heutigen Debatte zwischen Subjekt- und Objektgebrauch des Ich-Ausdrucks. Der Subjektgebrauch liegt nach Shoemaker in Fällen vor, in denen keine Identi zierung als Objekt nötig ist – also beispielsweise wenn ich mich auf mich als Schmerzen habend beziehe. Man könnte vermuten, dass Kant und Shoemaker dasselbe Phänomen im Blick haben: eine Bezugnahme auf sich, die ohne Identi kation von statten geht. Nach meinen bisherigen Ausführungen ist klar, dass man hier auf jeden Fall einräumen muss, dass Kant – anders als Shoemaker – nur im Fall des Denkens die Selbstanschauung als Objekt ausschließt. Aber denkend werden wir uns nach Kant unserer selbst tatsächlich nicht als Objekt, sondern als Subjekt bewusst. Daher besteht hier allem Anschein nach eine Ähnlichkeit zu dem, was wir bei Shoemaker als den subjektiven Gebrauch des Ich-Ausdrucks kennen.³⁸ Nach der genaueren Betrachtung lässt sich aber auch diese These einer Ähnlichkeit nicht halten: Shoemaker nimmt – wie die meisten Philosophen heute – an, dass der Ich-Ausdruck immer auf die körperliche Person referieren muss.³⁹ Damit stimmt er mit Kant in dessen Überlegungen zur Selbstwahrnehmung, verstanden als ein Fall von empirischem Selbstbewusstsein, überein. Aber wenn Shoemaker annimmt, dass der Ich-Ausdruck immer auf eine körperliche Person referiert, so ist die Rede davon, dass es in bestimmten Fällen zu einem Subjektgebrauch des Ich-Ausdrucks komme, strenggenommen misslich, denn der suggeriert (wie es bei Wittgenstein auch gedacht ist), dass der Ich-Ausdruck auf kein Objekt referiert. Wenn der Ich-Ausdruck auf die körperliche Person referiert und ‚nur‘ die Zugangsweise zu dieser in manchen Fällen eine besondere ist, so besteht kein Grund für

38 Dass hier eine Parallele besteht, macht jüngst vor allem Longuenesse in verschiedenen Arbeiten (z.B. Longuenesse „Kant on the Identity of Persons“) geltend. Auch bei Kitcher spielt diese Parallele argumentativ eine wichtige – aber meines Erachtens aus hier ausgeführten Gründen nicht überzeugende Rolle, vgl. Kitcher Kant’s Thinker, 253 . 39 Die Sache ist bei Shoemaker tatsächlich nicht immer eindeutig. Aber durch seine Abgrenzung von Kant und Wittgenstein lässt sich folgern, dass er dies annimmt. Er stimmt dann weitgehend mit Peter Strawson überein. Allerdings sind beide Theorien in diesem Punkt recht kompliziert. Je nachdem, wie man sie versteht, stellt sich auch die Beziehung zu Kant anders dar.

76

Emundts

die Rede vom Subjektgebrauch des Ich-Ausdrucks.⁴⁰ Aber abgesehen von der Frage, ob Shoemaker hier eine unglückliche Wortwahl hat, besteht sachlich an dieser Stelle ein erhellender Kontrast von Kant zu Shoemaker: Nicht nur nimmt Kant bei den meisten mentalen Zuständen an, dass wir uns als Objekt anschauen können. Für Kant ist vielmehr auch im Fall der speziellen Denktätigkeit der Selbstbezug tatsächlich derart, dass das „Ich denke“ gar nicht auf ein Objekt bezogen werden kann⁴¹ und dass eine Selbstidenti zierung gar nicht möglich ist.⁴² Dies begründet Kant aber in der oben erläuterten Weise, das heißt durch die idealistische Theorie. Kant vertritt diese Thesen, weil er Idealist ist. Als Idealist nimmt er eine besondere Art des rein subjektiven Selbstbezugs an. Worauf ich hier hinaus will, lässt sich so zusammenfassen: Es mag durchaus sein, dass es eine Eigenart des Selbstverhältnisses gibt. Was Kant angeht, so muss man aber festhalten, dass sich die Besonderheit, die die Rede eines subjektiven Gebrauchs des Ich-Ausdrucks rechtfertigt, nur bei einer speziellen Denktätigkeit ndet, die Kant im Rahmen seines Idealismus annimmt. Eine weitere Besonderheit des Selbstbezugs kann Kant auch behaupten. Mit ihr habe ich mich hier nicht ausführlich beschä igt. Wie aber deutlich geworden sein sollte, liegt die (vergleichbar harmlose) Besonderheit bei mentalen Zuständen überhaupt nach Kant sicherlich nicht daran, dass wir hier auf uns ohne Identi zierung Bezug nehmen können (wie unter anderem bei Shoemaker angenommen). Denn, wie deutlich geworden sein dür e, nimmt Kant längst nicht für alle Fälle mentaler Zustände an, dass wir es hier mit einem Selbstbezug zu tun haben, bei dem das Ich grundsätzlich evasiv oder subjektiv wäre. Die Besonderheit liegt darin, dass wir auf uns als räumliches Wesen in solchen Fällen in anderer Weise (nämlich durch den inneren Sinn) Bezug nehmen als in anderen Fällen.

40 Dass Shoemaker diesen Ausdruck wählt, liegt daran, dass er sich auf Wittgenstein bezieht. Auch wenn er Wittgenstein gerade an der hier entscheidenden Stelle beim Subjektgebrauch widerspricht, behält er die Rede bei. 41 Dies vertritt, wie gesagt, sehr deutlich Horstmann, der daher in den in Fußnote 1 genannten Texten auch immer von „radically subjective consciousness of oneself“ und von „the nonobjecti able (irreversibly subjective) self-conscious I“ spricht. Damit hat er mich auf diesen wichtigen Punkt aufmerksam gemacht. 42 Wie ich diese Tätigkeit überhaupt mir als Person und auch einer anderen Person zuschreiben kann, ist erklärungsbedür ig. Natürlich kann man auch nach Kant sich und anderen als körperlichen Personen – also Objekten – das Prädikat „denkt“ oder „ist im Zustand des Denkens“ zuschreiben. Dies muss schon deshalb möglich sein, damit wir uns und anderen Menschen zuschreiben können, objektiv zu urteilen. Aber nehmen wir hier eine Identi kation besonderer Art vor? Kant äußert sich zu diesen Thema nur sporadisch (KrV B 157; Preisschri über die Fortschritte der Metaphysik AA XX, 268). Ausführlich dazu Hannah Ginsborg in diesem Band.

Kant über Selbstbewusstsein

77

Ich will also nicht leugnen, dass es eine Eigenart des Selbstverhältnisses gibt. Was den phänomenologischen Aspekt angeht, so könnte man meines Erachtens mit Blick auf beide Formen der hier thematisierten Selbstverhältnisse eine passende phänomenologische Geschichte erzählen. Es scheint mir richtig, dass der Selbstbezug bei mentalen Zuständen etwas Evasives hat. Es scheint mir sogar auch zutre end, dass wir phänomenologisch Fälle von Selbstbezug auf unsere Denktätigkeit noch besonders herausstellen können. Aber dies zu tun, würde zu einer ganz anderen Form von Theorie führen als derjenigen von Kant. Bei Kant geht es nicht um die Phänomenologie, sondern um Fragen zu Bedingungen von Objektbezug. Daher müssen wir den Selbstbezug als einen transzendentalen Selbstbezug ansehen. Was man daher am Ende der Kantischen Theorie des Selbstbewusstseins unter anderem entnehmen kann ist Folgendes: Ohne Idealismus (im erläuterten Sinn) kann man entweder Selbstanschauung als die Weise ansehen, in der wir uns auf uns selbst beziehen, wenn wir dies nicht explizit denkend tun – denn dagegen spricht bei Kant nur der Idealismus. Oder man kann Selbstbewusstsein als eine Fähigkeit behaupten. Auch dagegen spricht bei Kant, folgt man meiner Darstellung, letztlich der Idealismus. Oder man kann Kant aus anderen Gründen – zum Beispiel aus den erwähnten phänomenologischen Gründen – darin folgen, dass es beim Denken ein besonderes Selbstverhältnis gibt. Als Idealist (im erläuterten Sinn) hat man dagegen auf jeden Fall gute, von Kant gegebene Gründe, an der Idee eines radikal subjektiven Selbstverhältnisses festzuhalten.⁴³

43 Dies ist, wie deutlich geworden sein sollte, in meinen Augen die Position, die Horstmann vertritt. Meine Dank für Anregugungen und Diskussionen gilt vor allem Rolf Horstmann. Außerdem danke ich Andrea Lailach-Hennrich und Anne Mone Sahnwaldt für ihre Anmerkungen zum Text.

Wolfgang Carl

The Copernican Turn and Stroud’s Argument from Indispensability Simple and short formulas may have the disadvantage that their familiarity conceals the real issue. The catchword ‘The Copernican Turn’ is well-known as a short designation of Kant’s transcendental philosophy and is distinguished by being invented by Kant himself for his “altered method of our way of thinking”.¹ There are di erent views about what this alteration amounts to. According to Husserl, the Kantian “change in the ways of thinking”² is nothing else than a “complete turnover of our natural way of thinking”³ which has been achieved in principle by Descartes’ Cogito as the starting point of philosophy.⁴ On the other side, Rorty claims that the “strategy of the Copernican revolution” is “to insure that objects will conform to our knowledge rather than be able to demand conformity from us.”⁵ In a similar way Stroud takes “the Copernican point” to be the claim that “objects must conform to our knowledge”⁶ and explains such a conformity by assuming that “these objective states of a airs are constituted by the condition of thought and experience.⁷ However, if one looks closely at the text to which these philosophers are referring one immediately notices that Kant himself doesn’t talk about a change of our natural attitude toward the world or about a new way of conceiving the relation between knowledge and its object in general, but rather is concerned with a particular kind of knowledge. The “change in the ways of thinking” he proposes is restricted to metaphysics and has the scope of explaining the apriority of metaphysical knowledge. To understand Kant’s Copernican Turn it is essential to take account of its restriction and of its explanatory scope, as Heidegger pointed out a long time ago.⁸ For an assessment of this turn one has to answer two questions. First, one has to clarify what it means that objects must conform to our knowledge. Secondly, one has to show why this view explains the possibil-

1 I. Kant, Kritik der reinen Vernun , B XVIII. 2 B XVI. 3 E. Husserl, Erste Philosophie, Husserliana VII, 1956, 243; cf. 249, 254. 4 Loc. cit., 284. 5 R. Rorty, Philosophy and the Mirror of Nature, Oxford 1980, 153. 6 B. Stroud, The Signi cance of Philosophical Scepticism, Oxford 1984, 195. 7 B. Stroud, “The Allure of Idealism”, in: B. Stroud, Understanding Human Knowledge, Oxford 2005, 88. 8 M. Heidegger, Phänomenologische Interpretation von Kants Kritik der Reinen Vernun , Frankfurt 1977, 55/6.

80

Carl

ity of metaphysical knowledge. These issues are connected with each other. The view that objects must conform to our knowledge must be clari ed in such a way that it explains that possibility. In other words: There is no adequate clari cation of that view without such an explanation. Without taking into account the restriction of the discussion to metaphysics and the scope of its explanation, the claim that the objects must conform to our knowledge sounds rather strange because Kant several times asserts its negation. In the well-known letter to Marcus Herz of February 1772 he writes: “Our understanding is not by means of his representations the cause of the objects nor is the object the cause of the representations of the understanding.”⁹ In the Prolegomena he points out: “My understanding and the conditions which hold for the possible connections of the determinations of things, as far as their existence is concerned, doesn’t prescribe any rule for the objects. They don’t conform to my understanding.”¹⁰ The understanding and our cognitive faculties in general are not the cause of the objects of our knowledge, nor is there any way in which they can “prescribe” something or other. Thus, what is the Copernican Turn all about? In the rst Critique Kant draws a distinction between two ways of conceiving the relation between a representation and its object: First, “the object alone makes the representation possible”. And second, “the representation makes the object possible”.¹¹ The rst claim is supposed to be true of all empirical representations, while the second holds for a priori representations only, but with a certain modication. This modi cation is important. Kant writes: But if it is the second, then since representation in itself (for we are not here talking about its causality by means of the will) does not produce its object as far as its existence is concerned, the representation is still a priori determinant of the object if it is possible through it alone to cognize something as an object.¹²

Whatever the precise content of the idea of an a priori determination may be, it is fairly obvious that Kant does not want to rely upon the alternative between the object making possible the representation and the representation making possible the object. Rather, he attempts to introduce a third way a representation might be related to its object. This way is called the “a priori determination of an object” and is explained by reference to conditions of the possibility of knowledge of an object.

9 Kant’s Briefwechsel I, 2. Au age, Berlin 1922, AA 10, 130. 10 Kants Werke IV, Berlin 1968, AA 4, 294. 11 B 124 f.; cf. Preisschri , AA 20, 274. 12 B 125.

The Copernican Turn and Stroud’s Argument from Indispensability

81

Both of these ideas are missing in the alternative Kant started with. Given this explication, one has to conclude that Kant himself didn’t consider the two alternatives as complete or exhaustive, because there is no place for a priori representations as conditions of knowledge of an object. Thus, if the Copernican Turn is conceived within the framework of that alternative, the real point of Kant’s argument is lost. What does he have in mind? It may be helpful to have a look at the development of his thought. In the letter to Marcus Herz mentioned above Kant argues in a way more or less similar to the considerations to be found almost ten years later in the rst Critique. However, there are two di erences. First, Kant doesn’t mention his proposal to consider a third way a representation may be related to its object. Second, there is an indication of the reason to draw the distinction – an indication missing in the rst Critique. Kant points out that he wants to discover “the key to the secret of metaphysics”. In 1781 he was convinced that he had discovered that key. In the ‘Preface’ to the second edition of the Critique of Pure Reason the key is described as “an altered method of our way of thinking”, better known under the title ‘The Copernican Turn’. Thus, an account of the development of Kant’s thought conrms what is disclosed by an impartial reading of the ‘Preface’ and allows us to state two conditions which have to be satis ed by any interpretation of the Copernican Turn. First, it is a project exclusively concerned with metaphysics. Second, the idea that “objects must conform to our knowledge” has to be explained within the framework of Kant’s epistemological turn in metaphysics. The interpretations mentioned at the beginning of this paper do not satisfy any of these conditions. As we have seen, Kant draws a distinction between two ways a representation may be related to its object. In contemporary terminology, on the one side there are representations having an intentional content with conditions of satisfaction depending on how things are in the world; and, on the other side, there are representations having an intentional content with conditions of satisfaction which are supposed to bring about changes in the world. In the latter case Kant refers to the “causal e ciency by means of our will”. These di erent kinds of representations have di erent “directions of t”, as Searle says.¹³ Perceptions, beliefs, experiences are representations having a content which is supposed to match an independent world (“mind-to-world direction of t”), while intentions, desires, commands have a content that is concerned with bringing about changes in the world (“world-to-mind direction of t”). Kant, who as far as I know was the rst to draw that distinction, was always eager to point out that the representations theoretical philosophy is concerned with have a “mind-to-world direction of t”.

13 J. Searle, Intentionality, Cambridge 1983, 7/9.

82

Carl

In other words: They are representations which “have to conform to the objects”¹⁴ and are not “active with regard to the object”.¹⁵ However, Kant is not satis ed with drawing the distinction between two classes of representations with di erent directions of t, rather he introduces a third way a representation can match its object. This way is described as an “a priori determination of an object” and has something to do with representations being conditions of the knowledge of objects. What he has in mind is explained with regard to intuitions in the following way: “These conditions underlie a priori objects in the mind according to the form”.¹⁶ With regard to concepts as conditions of knowledge he claims that “through them alone experience is possible (as far the form of thinking is concerned)”.¹⁷ Thus, the third way of a representation being related to its object has something to do with its form. The key role of the notion of form is pointed out by Kant in several ways. He emphasizes that his idealism has to be understood as a “formal idealism”.¹⁸ Confronted with Schlosser’s objection that critical philosophy is nothing else than “a pedantocracy under the name ‘Formgebungsmanufaktur’” he points out that “the formal part or element of our knowledge is the main concern of philosophy”.¹⁹ It is quite surprising that Kant justi es his view by referring to a claim of the “Ancients” as found in Aquinas²⁰ as well as in the ‘Summulae Logicales’ of Petrus Hispanus.²¹ The claim states: Forma dat esse rei. It is even more surprising that Kant took interest in this claim even before 1769,²² and, although the claim is not mentioned in the rst Critique, he o en repeats this claim a er 1787 in order to sum up what is peculiar to transcendental philosophy – like a headline: In the lectures on metaphysics, called L 2 from the winter-term 1790/1,²³ in the aforementioned “Von einem neuerdings erhobenen vornehmen Ton in der Philosophie”, published in 1796, and several times in the Opus postumum. Why did he care about this claim of the “Ancients”?

14 R 4633, AA 17, 615. 15 AA 10, 130. 16 B 125. 17 B 126. 18 B 519; Prolegomena, AA 4, 375; R 4953, AA 18, 40. 19 Von einem neuerdings erhobenen vornehmen Ton in der Philosophie, AA 8, 404. 20 Thomas Aquinas, Summa contra Gentiles, II. 58, 1260. 21 P. Hispanus, Summulae Logicales Venezia 1593, 294. 22 R 3850/2, AA 17, 312. 23 AA 28, 575.

The Copernican Turn and Stroud’s Argument from Indispensability

83

The notion of form belongs to a pair of “concepts of re ection,” form and matter – two concepts “inseparably bound up with every use of the understanding”.²⁴ The identi cation of the form with essence – esse rei – is supplemented by a restriction: The essence as far as it is known by reason.²⁵ What this restriction amounts to can be seen from the examples given by Kant: Is the thing [Sache/res] an object of the senses, its essence is the form of things given by intuition (as appearances), and even mathematics is nothing else than a doctrine of the forms of pure intuition, while metaphysics as pure philosophy is based upon forms of thought under which every object (the matter of knowledge) has to be subsumed.²⁶

As one can see from these examples, the restrictive determination of essence amounts to the claim that the essence of an object is whatever one can know a priori about it. Thus, the identi cation of essence with form is not concerned with a substantial or rich notion of essence, as in Aristotle or what today are called ‘metaphysical necessities’, but rather with an epistemological notion of essence. This essence is identi ed with form. Kant quotes the claim forma dat esse rei in order to justify his view that the main concern of philosophy has something to do with whatever belongs to the form (das Formale).²⁷ This is true in the rst place of metaphysics which is the real issue of the Copernican Turn, and for which Kant tries to establish a new conception of a representation being related to its object. This conception supplements the dichotomy of the two di erent “directions of t” by suggesting a third possibility described as an “a priori determination of an object” and restricted to representations of forms.²⁸ Thus, we get a preliminary idea of what the knowledge is about for which Kant proposes his “altered method of our way of thinking”. What this alteration amounts to can be explored by considering his views about the notion of form as developed in the chapter ‘On the amphiboly of concepts of re ection’.²⁹ I am concerned here only with the pair of concepts ‘matter’ and ‘form’ which, as quoted above, is “inseparably bound up with every use of the understanding.”³⁰ These concepts are relations: matter and form are always matter and form of something or other. Secondly, they are correlative concepts: There is no matter without

24 25 26 27 28 29 30

B 322. Vornehmen Ton, op. cit., 404. Loc. cit. Loc. cit. B 125. B 316 . B 322.

84

Carl

form, and there is no form without matter. Finally, these concepts are contextsensitive: What they refer to depends upon the realm of their application and is quite di erent in logic and transcendental philosophy. Kant claims that in the “transcendental sense” ‘matter’ signi es “the determinable” while ‘form’ signies “the determination”, “actus determinandi”.³¹ Thus, the transcendental use of these concepts only requires that something is given, and that there is an actus determinandi.³² What does it mean that form is an actus determinandi? How can we give an account of the intimate connection between form and activity emphasized by Kant as early as the end of the 1760s? Because he applies the distinction between form and matter to intuitions as well as to concepts, the notion of form cannot be explained by the spontaneity of the understanding. Form as an actus determinandi cannot be only an operation of the understanding. A positive view emerges from considering the form of our external intuition. Kant writes: “The perception represents (staying for now only with outer intuitions) something real in space. For rst, perception is the representation of a reality, just as space is the representation of a mere possibility of coexistence”.³³ Perception is a representation of something real, because its matter is a sensation “which designates a reality in space and time”.³⁴ The view that sensations designate something or other is without doubt in need of further clari cation; and the same is true of the distinction between “feelings” and “objective sensations” drawn by Kant in the Critique of Judgement.³⁵ However, I will not discuss this problem here, because for applying the distinction between matter and form to outer empirical intuitions it may be su cient to point out that sensations are indispensable for such intuitions and that because of “the real of sensation” outer intuition has to be considered as a representation of something real.³⁶ That it is a representation of “something real in space” has something to do with the form of outer intuition. Kant refers to its form by pointing out that “space is the representation of a mere possibility of coexistence”.³⁷ What does that mean? In the Transcendental Aesthetic Kant claims that the form of an object of an empirical intuition, called ‘appearance’, is that “which allows that the manifold of appearances can be ordered in certain relations”.³⁸ This explication is to be

31 32 33 34 35 36 37 38

Vornehmen Ton, loc. cit., 404. Metaphysik-Vorlesung L 2, AA 28, 575. A 374. Loc. cit. AA 5, 206. B 207. A 374. B 34.

The Copernican Turn and Stroud’s Argument from Indispensability

85

found in the second edition, while in the rst edition Kant says that the form “allows the manifold of appearance to be intuited as ordered in certain relations”.³⁹ In other words: In ‘A’ the form of an intuition explains the content of a given representation, while in ‘B’ the form explains the possibility of representations with a given content. Kant prefers the latter version, because the form of an object of an empirical intuition has something to do with the possibility of certain relations between appearances.⁴⁰ This version emphasizes the connection between form and the possibility of various relations and explains why space as the form of outer perception is determined as a “mere possibility of coexistence” as pointed out in A 374. Because of its matter such a perception is a representation of “something real”, and because of its form what is represented is represented within the framework of possible spatial relations. By its matter and form an outer perception represents something that connects the real with the possible. The signi cance of such a connection can be appreciated by having a look at the rst argument of the “metaphysical exposition” of the concept of space by which Kant tries to show that the representation of space cannot be an empirical concept, because an empirical external intuition presupposes that representation. Such an intuition refers to something outside me or to several things outside me. What is referred to is the content of the intuition being in another place from that in which I nd myself, or in several such places. All these places belong to one and the same space. Given the way Kant describes the content of an empirical outer representation it follows that the representation of space to which di erent places belong is presupposed by such a representation and cannot derived from it. Kant elaborates this point in R 4673: “The order of things which are side by side is not the space, but the space is what makes possible such an order.”⁴¹ That things are side by side is a relation that has to be established by empirical means. But such a relation entails as well as excludes other spatial relations. There are symmetrical, asymmetrical and transitive spatial relations. To establish empirically that there is a certain spatial order between objects implies statements about the possibilities and impossibilities of spatial relations between objects being in di erent places. These modal statements presuppose the representation of a single, in nite space. As Warren has pointed out, “Kant’s central point is that our recognition of these modal features requires a representation of space”.⁴² Thus, the argument for the aprioricity of our representation of space explains that space is not a relation be-

39 A 20. 40 R 4673, AA 17, 639. 41 Loc. cit. 42 D. Warren, “Kant and the Apriority of Space,” in Philosophical Review 107, 1998, 208.

86

Carl

tween di erent places but rather the very possibility of such a relation. As Kant notes in his copy of the rst edition of the rst Critique, “space is what underlies the possibility of external relations”.⁴³ Leaving aside the issue of the aprioricity of our representation of space and of our representations of forms in general, I want to explore the connection between form and possibility or, more precisely, the connection between the pair of concepts ‘form’ and ‘matter’ and the modal distinction ‘possible’ and ‘real’. The latter is implied by the former, and both distinctions are essential features of the human mind. In the Critique of Judgement Kant points out that “it is indispensably necessary for the human understanding to distinguish between the possibility and the reality of things”.⁴⁴ As we all know, Kant is discussing the issue of what is peculiar to human knowledge and the validity of concepts and distinctions relativized to this kind of knowledge by taking into consideration the idea of a divine knowledge exempli ed by an intuitive understanding or by an intellectual intuition. For an intuitive understanding there is no distinction between what is possible and what is real. That does not mean that what is thought in this way is neither possible nor real, but rather whatever is thought is real: “If our understanding were intuitive, all its objects would be real”.⁴⁵ What is possible would be reduced to what is real; and the idea that something is possible or could be possible doesn’t make sense for such an understanding. Thus, the distinction between “thinking of an object” and “having knowledge of it” would disappear. What about intellectual intuition? In his lectures on Metaphysics, called L1, Kant points out that “such an intuition is not sensorial; the object will be cognized without there being any a ection”.⁴⁶ An intuition having no sensation as its matter cannot have a form either: “It is for this reason that God intuits the world without space and time […]”⁴⁷ . However, intellectual intuition is for human beings “non-sense” (Unding), because our cognitive relations to the world are quite di erent from those of God: While our knowledge doesn’t create its object, God’s knowledge does. His intuition of the world has neither a matter nor a form. Because we are not “the creator of the things” we have empirical intuitions of, these things have to be given to us, and “therefore we cannot cognize them except under a certain condition”.⁴⁸ This condition is the way in which something is given to us, the form of our intuitions. Our intuitions have both a form and a matter,

43 44 45 46 47 48

AA 23, 22. The Critique of Judgement, § 76, AA 5, 401. Loc. cit., 402. AA 28, 179. Loc. cit. Loc. cit.

The Copernican Turn and Stroud’s Argument from Indispensability

87

because our intuitions do not create their own objects. The latter point is made by Wittgenstein in the following way: “Observing does not produce what is observed. (That is a conceptual remark.)”⁴⁹ Parenthetically, I want to point out that the notion of whatever is given to us implies, according to Kant, the notion of a way of being given to us. Nothing can be given to us, except “under a certain condition” or in a certain way. This reminds one of Frege’s claim that there can be no reference without sense. Or, in more general terms: every representation has a mode of representation or perspective. An empirical external intuition represents because of its matter “something real”, and because of its form it is represented “in space”, that is: as a manifold of possible spatial relations. The distinction between form and matter implies the distinction between possible and real. The divine understanding which is not able to apply the latter distinction is a faculty of non-conceptual cognitions. There are no possibilities for him, because whatever he thinks is real. For the intellectual intuition there are no possibilities either, because where there is no matter, there can be no form. On the other hand, for our cognitions the distinction between form and matter is indispensable; and this is true as well of the distinction between possible and real. This distinction is applied to both of our basic cognitive faculties, sensitivity and understanding, but in di erent ways. Because we cannot cognize anything except by concepts, our knowledge is shaped and determined by concepts – “the realm of reason”, as one might say. Because we can only cognize something if it is given to us in some way or other, our knowledge is connected with a realm of possibilities determined by our forms of intuitions: the manifold of possible and spatial relations. There are forms of conceptual cognitions, and there are forms of sensorial intuitions. These forms are connected with di erent kinds of relations between the possible and the real. The form, understood as actus determinandi, articulates these relations in di erent ways according to the distinction between understanding and sensibility. As far as concepts are concerned, Kant draws a distinction between logical and real possibility. The latter has to do with the conditions of the possibility of experience.⁵⁰ Knowing something by concepts establishes a relation between a conceptual possibility and what is real by nding an intuition which “corresponds” to a given concept. The way, so to say, leads from the possible to the real. With regards to empirical intuitions the situation is di erent. The real is given by their matter, their sensations; and by their forms the real is embedded in a manifold of possible spatial and temporal relations. Here the way leads from the real to the possible. Because of the correlation

49 L. Wittgenstein, Philosophische Untersuchungen II, Oxford 1953, IX. 50 B XXVI; B 302; B 624 n.

88

Carl

between form and matter of empirical intuitions the real as represented by them is not an isolated item, but something that is essentially connected by spatial and temporal relations with other real items. It seems to me that in the current debate about non-conceptual content and Kant’s notion of intuition his notion of a form of intuition is rather neglected. What does all this have to do with the Copernican Turn? As we have seen, the two di erent relations between a representation and its object conceived according to the two “directions of t” do not allow us to explain a representational relation to the world which is neither practical nor exclusively empirical. It is for this reason that Kant introduces a third possibility: If a representation gures as a condition of the possibility of knowledge then it is “a priori determinant” of its object. Such a representation concerns the form of our knowledge, either of our sensitivity or of our understanding. Such a representation is a priori and refers to something that is an actus determinandi. The idea of determination has to be understood in a way that it doesn’t t only the spontaneity of the understanding, because otherwise the assumption of forms of intuitions would be in con ict with the view that sensitivity is independent of the understanding.⁵¹ The notion of form which can be applied to both kinds of cognitive faculties has something to do with the relation between possibility and reality – a relation that applies to both of them, but in di erent ways. The form in “the transcendental sense” is nothing else than di erent ways of correlating what is possible and what is real. Given Kant’s notion of form, I will try to make sense of his claim that “objects must conform to our knowledge”. The Copernican Turn is a turn to metaphysics and is supposed to explain the possibility of metaphysical knowledge. Metaphysics is concerned with the form of objects and our knowledge of them. Thus, the metaphysical turn proposed by Kant amounts to the claim that the forms of objects must conform to the forms of our knowledge. The notion of form is, as we have seen, a key-notion of transcendental philosophy and belongs to his theory of concepts of re ection. These concepts are always pairs of correlative concepts. Thus, the notion of form has to be considered in connection with the notion of matter. The distinction of form and matter is “inseparably bound up with every use of the understanding” and entails the distinction between possibility and reality which is indispensable for the understanding of a nite being, for the “human point of view”.⁵² The indispensability of these distinctions, combined with their restricted validity, provides the key, as we shall see, for understanding the relation between an object and its cognition, as proposed by the Copernican Turn.

51 Cf. B 123. 52 B 322.

The Copernican Turn and Stroud’s Argument from Indispensability

89

Kant calls this relation sich richten nach, – a phrase which has in German a variety of meanings. It doesn’t mean einrichten, that is, to arrange something or other, because it has nothing to do with “an arranged arbitrary design” (eine eingerichtete willkürliche Formengebung), but rather with “the diligent and careful, attentive work of the subject to account for his own faculty of reason”, as Kant points out in 1796.⁵³ The really “critical” project – in the Kantian sense of the word – is the issue of the Copernican Turn which claims that the form of objects has to be explained by the forms of our knowledge. How can this claim be justi ed? The notion of form belongs to the pair of concepts of re ection ‘form’ and ‘matter’ which is distinguished by its indispensability for the human mind and its restricted validity. What is true of the pair is true as well of each of its elements. Thus, because of its indispensability the notion of form is a constitutive condition of our knowledge; and this indispensability holds only for our knowledge, because there is no application of the notion of form except to our knowledge, as we have seen from Kant’s views about God’s knowledge. The indispensability of the notion of form is a constitutive as well as a restricted indispensability. Because of the latter the assumption that objects have a form cannot be justi ed and their determination cannot be given without relying upon our forms of knowledge. And because of their constitutive indispensability the forms of our knowledge are the forms of its object. In the lectures on Metaphysics L 1 Kant describes this indispensability in the following way: “The objects must conform to the conditions which hold for the possibility that they are known […]”⁵⁴ Expressed in the terminology of the Copernican Turn this claim states that the forms of objects must conform to the formal conditions, that is, to the forms of our knowledge. The reason for this is, according to Kant, “the nature of the human understanding”. To make the same point in more austere terms: Knowledge implies truth. Thus, the restricted indispensability of the notion of form and its constitutive indispensability, taken together, amounts to nothing else than the Copernican Turn. Stroud has criticized the Copernican Turn several times, most recently in his latest book, Engagement and Metaphysical Dissatisfaction.⁵⁵ There are few contemporary philosophers who have carried out investigations into the indispensability of certain ways of thinking with such depth and brilliance. However, although he is well aware of the fact that his own considerations point in the “direction of an answer that it is at least partly Kantian”,⁵⁶ he always keeps his distance

53 54 55 56

Vornehmen Ton, Loc. cit., 404. Metaphysik-Vorlesung L 1, AA 28, 239. B. Stroud, Engagement and Metaphysical Dissatisfaction, Oxford 2011. Loc. cit., 156.

90

Carl

from anything that might remind one of the Copernican Turn. The idea of a world that must conform in general to “our necessary ways of thinking of and experiencing it” is for him committed to idealism; and that seems to him “too high a price to pay for the prospect of metaphysical satisfaction”.⁵⁷ I don’t want to consider idealism here, neither Kant’s version of idealism nor Stroud’s views about it. By way of concluding, I would instead like to explore some features of Stroud’s notion of indispensability and to distinguish it from the Kantian idea of indispensability. As we have seen, Kant is concerned with concepts “which are inseparably bound up with every use of the understanding”;⁵⁸ and it is one of the many tasks Kant attributes to metaphysics to give an account of them. Stroud claims: “For Kant indispensability was the key to the very possibility of putting metaphysics once and for all on the secure path of science”.⁵⁹ He does not give any argument for this claim which, according to my view, is wrong. But the real problem with the claim is that the indispensability of certain ways of thinking cannot be such a “key”, because it cannot establish any “positive metaphysical verdict about those ways of thinking”.⁶⁰ Thus, the view attributed by Stroud to Kant contains the focal element of what is wrong with the Kantian strategy, i.e., “to draw a conclusion about what the world is like from facts about how thinkers must think the world is”.⁶¹ In order to assess this criticism one has to explore Stroud’s idea of indispensability: Indispensability implies invulnerability, and it is the invulnerability of certain beliefs or ways of thinking he is mainly concerned with: “Beliefs that have to be accepted by anyone who is faced with a metaphysical question about their status would be metaphysically invulnerable”.⁶² That a belief is invulnerable means that you cannot entertain its negation. You cannot do that because the invulnerability of the belief is based upon its indispensability. However, from the fact that you cannot accept its negation you cannot conclude that it is false and, thus, that the invulnerable belief is true: “That would be to draw a conclusion about what the world is like from facts about how thinkers must think the world is”.⁶³ This conclusion is a non sequitur. According to Stroud, Kant drew this conclusion – “the problematic step of the Kantian strategy”. Did Kant make this mistake?

57 Loc. cit., 134/5. 58 B 322. 59 Engagement, Loc. cit., 126. 60 Loc. cit., 128. 61 Loc. cit., 135. 62 Loc. cit., 146. 63 Loc. cit., 135.

The Copernican Turn and Stroud’s Argument from Indispensability

91

How thinkers must think the world is is what indispensable beliefs are about. These beliefs guarantee neither their truth nor the falsity of their negation. We have to acknowledge “a di erence in general between our believing something and its being so”;⁶⁴ and this holds as well for invulnerable beliefs. Thus, Stroud’s strategy of indispensability does not lead to any conclusion about the world; it doesn’t go beyond an exploration of how we must think about the world.⁶⁵ This strategy secures a certain “transcendental invulnerability” of our beliefs about the world,⁶⁶ but no truth about the world can be established in this way. However, Stroud recognizes an application of the strategy of indispensability which establishes the truth of some indispensable beliefs as well. Such an application can be found in Descartes. He is, according to Stroud, the forerunner of Kant in operating with the idea of indispensability and with the Cogito provided the most “compelling” example of its application: Beliefs which are self-verifying and known to be true by the very act of believing them.⁶⁷ Another example is my judgement that I am here now. Although it states a contingent truth I cannot be wrong in making it. Thus, we have to conclude that there are two kinds of invulnerable beliefs: Beliefs we cannot avoid believing, without there being any guarantee of their truth or of the falsity of their negation, and beliefs we know to be true by the very act of having them. Indispensability implies invulnerability: Because certain beliefs are indispensable, they are invulnerable. But does invulnerability imply indispensability? The invulnerability of Cogito-beliefs has something to do with the self-verifying character of indexical thoughts. There is no reason to take for granted – or, at least, no reason mentioned by Stroud – that these beliefs are indispensable; and their invulnerability is not due to their indispensability, if they are supposed to be indispensable at all. Thus, we have to conclude that there are two kinds of invulnerable beliefs: beliefs having a “transcendental invulnerability” because of their indispensability and beliefs of the Cogito-type being invulnerable because one cannot be wrong in believing them. These present-tensed beliefs are essentially indexical, and it is for this reason that they belong to “a very special class” of beliefs a person might have. Stroud claims: “This idea was richly exploited by Kant. He was interested not just in the thinking of an individual thinker, but in the possibility of thinking in general by anyone”.⁶⁸ Because of this interest he

64 Loc. cit. 65 Cf. “Kantian Argument, Conceptual Capacities, and Invulnerability,” in: B. Stroud, Understanding Human Knowledge, Oxford 2005, 168/73. 66 Loc. cit., 173. 67 Cf. T. Burge, “Our Entitlement to Self-Knowledge,” in PAS 1996, 92. 68 Engagement, Loc. cit., 128.

92

Carl

did not and could not care about the invulnerability of certain indexical beliefs; and it gives a distorted view of his metaphysical project if one considers it as an exploitation of the Cartesian project. The idea of indispensability of certain ways of thinking or believing plays an important role in Kant’s philosophy. However, this idea can be used in di erent ways. The scope of Stroud’s strategy of indispensability is to establish a particular class of invulnerable beliefs, not Cogito-beliefs, but beliefs which are invulnerable to sceptical challenges. The kind of invulnerability he is interested in doesn’t imply its truth or the falsity of its negation, but allows for neutralizing sceptical challenges raised against our beliefs about an external world which regard the colours of objects, the values of what we care about, etc. But Kant’s strategy of indispensability was not concerned with the refutation of scepticism, either global or local. He wanted to establish that there are non-empirical conditions of empirical knowledge; and the kind of indispensability invoked by the Copernican Turn has something to do with these conditions. This turn was not designed to answer any sceptical challenge, and it cannot do that. Its basic fault cannot be that Kant tries to obliterate the di erence between our believing or even necessarily believing something and its being so. Stroud’s dissatisfaction with Kant’s way of putting metaphysics on the secure path of science and his diagnosis of the problematic “second and stronger step of Kant’s reasoning” are largely due to his conviction that Kant’s philosophy can and should be considered in the tradition of the Cartesian challenge to scepticism. That is the picture that holds so many contemporary interpreters of his philosophy captive.

Béatrice Longuenesse

Kant and Hegel on the Moral Self* 1 The Problem By “moral self” I mean whatever entity is referred to by the term “I” when “I” is used as the subject of a proposition expressing a moral command: “I ought to X,” or “X is what I ought to do.” For Kant, “I ought to X” is a moral command just in case it is determined under the discriminating principle of the categorical imperative, “I ought never to act except in such a way that I could also will that my maxim should become a universal law” (Grundlegung, 4:402).¹ In section 3 of Groundwork of the Metaphysics of Morals, Kant calls this moral self “the proper self” (das eigentliche Selbst): The human being claims for himself a will which lets nothing be put to his account that belongs merely to his desires and inclinations, and on the contrary thinks as possible by means of it – indeed as necessary – actions that can be done only by disregarding all desires and sensible incitements. The causality of such actions lies in him as intelligence and in the laws of e ects and actions in accordance with principles of an intelligible world […]. Since it is there, as intelligence only, that he is the proper self [das eigentliche Selbst, my emphasis]

* The original dra of this paper was prepared for the conference Dina Emundts organized in honor of Rolf-Peter Horstmann’s 70th birthday in Berlin, June 2011, under the title: “Metaphysical Themes in Kant and Hegel.” My warmest thanks to Dina for organizing what turned out to be one of the very best conferences I ever attended, where intellectual challenge did not preclude a deep sense of friendship and common purpose. Revised versions were presented at Yale University and at the New School in Spring 2012. My gratitude for the helpful comments of participants in all three discussions. Special thanks to Rolf for being an incisive commentator on my presentation at the New School. I am sure I have not answered all his objections to his satisfaction, but I hope the paper is better for trying to address them. 1 Works of Kant are cited by volume and page of Kant’s Gesammelte Schri en, hg. von der königlichen bzw. Preußischen Akademie der Wissenscha en, Berlin. Abbreviations and English Translations used: A, B: Kritik der reinen Vernun , 1. bzw. 2. Au age. Critique of Pure Reason, Transl. P. Guyer and A. Wood, Cambridge University Press, 1998. Grundlegung: Grundlegung der Metaphysik der Sitten. Groundwork of the Metaphysics of Morals, Transl. M. Gregor, Cambridge University Press, 1998. KpV: Kritik der praktischen Vernun . Critique of Practical Reason, Transl. M. Gregor, Cambridge University Press, 1997. KU: Kritik der Urteilskra . Critique of the Power of Judgment, Transl. P. Guyer and E. Matthews, Cambridge University Press 2000. Naturwissenscha : Metaphysische Anfangsgründe der Naturwissenscha . Metaphysical Foundations of Natural Science, Transl. M. Friedman, Cambridge: Cambridge University Press, 2004. Sitten: Metaphysik der Sitten. The Metaphysics of Morals, Transl. M. Gregor, Cambridge University Press, 1996.

94

Longuenesse

(as a human being he is only the appearance of himself) those laws apply to him immediately and categorically, so that what inclinations and impulses (hence the whole nature of the world of sense) incite him to cannot infringe upon the laws of his volition as intelligence; indeed he does not hold himself accountable for the former or ascribe them to his proper self [seinem eigentlichen Selbst, my emphasis], that is, to his will, though he does ascribe to it the indulgence he would show them if he allowed them to in uence his maxims to the detriment of the rational law of his will (Grundlegung, 4:457–8).²

The “proper self” is opposed to the “dear self” (das liebe Selbst), which we can take to be what is represented by the term “I” in “I ought to X” when the “ought” is determined, not under the strict command of the categorical imperative, but under prudential or technical rules dictated by our impulses and inclinations (Grundlegung, 4:407). But since, according to Kant, we are both pathologically a ected beings and intellectual beings or “intelligences,” the ends we set for our actions depend both on the ways we seek or avoid the objects a ecting us with pleasure or pain, and on the command of duty. So in any given case of self-assigned maxim for our actions, the “self” that “I” refers to is what Kant calls, in the Critique of Practical Reason, the “entire” self – both sensible and intellectual. But our “dear self” has a natural propensity to “make its claims primarily and originally valid, as if it constituted our entire self” (KpV, 5:74). The question I would like to examine in this paper is the following: how much metaphysics, and what kind of metaphysics, does Kant need in order to account for the existence and nature of the “moral self” or “proper self”? As is well known, in the Critique of Pure Reason Kant has cut to size what he takes to be the illusory attempt of rationalist metaphysicians to derive from the logical features of our concept “I” in “I think” the real features of an entity, an immaterial soul distinct from the body. However, the rationalist representation of the self as an immaterial substance, criticized in the Paralogisms of Pure Reason in the rst Critique, seems to regain its credentials in the second Critique, as the object of a belief called forth by reason in its moral use.³ Does Kant’s “proper self” call for such a reinstatement of the idea of an immortal soul? Kant makes no mention of it in either Groundwork or the Analytic of the Critique of Practical Reason. Rather, Kant’s postulate of the existence of an immaterial and thus immortal soul is introduced not in connection with the moral self or autonomous will, but rather as part of Kant’s solution to the problem of the real possibility of the Highest Good, namely the synthetic unity of

2 See also Grundlegung, 4:461: “The law interests since it is valid for us as human beings, since it arose from our will as intelligence and so from our proper self [aus unserem eigentlichen Selbst].” 3 See KpV, 5:122–24. Also Kritik der reinen Vernun , Preface to the Second Edition, B 30. As is customary, the Critique of Pure Reason is cited by page numbers in the rst edition (1781) and the second edition (1787), respectively indicated by “A” and “B”.

Kant and Hegel on the Moral Self

95

virtue with happiness. Only in the context of the solution to that problem, o ered in the Dialectic of Pure Practical Reason, does Kant assert that we must postulate, on behalf of pure practical reason, the immortality of the soul and the existence of God. In other words, the solution to the problem of the Highest Good provides the speci c context in which something like the structure of rationalist “special metaphysics” (rational psychology, cosmology, and theology) is provided new legitimacy. However, this structure is not preserved in Kant’s own system of metaphysics. The model for Kant’s system of metaphysics is not that of Wol an “ rst philosophy or ontology” followed by “special metaphysics.” Rather, in his own system Kant returns to the model he praised in the Preface to Groundwork of the Metaphysics of Morals, that of the ancient Stoic system – logic, physics, ethics (Grundlegung, 4:387) – which for Kant become transcendental logic, metaphysics of nature, metaphysics of morals. So it seems worth asking: what kind of metaphysics, and how much metaphysics, is in fact needed to explain Kant’s notion of a moral self? In Part II of this paper, I shall sketch out Kant’s answers to this question and evaluate those answers. In Part III, I shall sketch out Hegel’s answers to this very same question: How much metaphysics, and what kind of metaphysics, is needed to account for the speci cally Kantian notion of a moral self? Unsurprisingly, Kant’s and Hegel’s respective answers are going to be signi cantly di erent. But it is worth noting from the outset that, like Kant, Hegel models his philosophical system on that of the Stoics: Logic (Science of Logic), Physics (Philosophy of Nature), Ethics (Philosophy of Spirit). Within this system, Hegel’s account of the Kantian notion of a moral self is both historical and metaphysical. On the one hand, Hegel gives a historical account of a form of thought of which Kantian morality – and thus the Kantian notion of the moral self or “eigentliches Selbst” – is an especially salient and systematically developed instance. As always for Hegel, this form of thought, namely this way of thinking of oneself under the concept of morality, is also a form of being or existence, namely a moment in the development of what he calls “Spirit.”⁴ But precisely for this reason, the account is also metaphysical. That

4 See Phän., VI, C: “Spirit Certain of Itself. Morality,” S. 3, pp. 441–63. PR, Part II, “Morality,” §§ 105–141, S. 7, pp. 203–291. PG, S. 10, §§ 503–512, “Morality.” References to works of Hegel are cited by volume and page in Werke in Zwanzig Bänden, Theorie Werkausgabe, Frankfurt: Suhrkamp Verlag, 1970, followed by the page number in the English edition for ease of reference. All translations are mine, except for the Philosophy of Right where I have used A. Wood’s translation. Abbreviations used: EG: Enzyklopädie der philosophischen Wissenscha en im Grundrisse, dritter Teil: die Philosophie des Geistes. (Tr. Hegel’s Philosophy of Mind, from the 1830 edition, together with the Zusätze by W. Wallace and A. V. Miller; with revisions and commentary by M. J. Inwood).

96

Longuenesse

there is such a way of thinking as Kantian morality indicates that there is such a way of existing as Kantian morality and the Kantian moral self – a way of existing that calls for metaphysical elucidation: what must reality in general be like, for such a determination of existence as Kantian morality to be even possible? Of course, Hegel’s discussion of morality (Moralität)⁵ goes well beyond a mere discussion of Kant’s view of morality and the moral self. The ambition of this paper, however, is only to compare what Kant has to say about his own notion of the moral self and Hegel’s interpretation and transformation of that notion. A clari cation is in order here: as I hope will be clear from what follows, I do not mean to imply that Hegel endorses Kant’s moral theory or that he shares Kant’s view of moral practical reason as what gives absolute value to human beings. As I have commented on elsewhere, Kant’s moral philosophy and Kant’s notion of moral practical reason represent for Hegel the most extreme development of the “philosophies of re ection” whose limited standpoint Hegel takes to be de nitively disquali ed by his own philosophy.⁶ This being said, Hegel also thinks that morality (Moralität) and the moral view of the world (die moralische Weltanschauung) are an indispensible moment in the development of what he calls Spirit (more on this below). My hope is that comparing Kant’s own notion of the moral self and Hegel’s re-interpretation of it will yield new insights into Kant’s and Hegel’s respective views, and especially into the metaphysical commitments each view depends on. An independent discussion of Hegel’s view of morality in relation to his own notion of Sittlichkeit, ethical life, is beyond the limits of this paper.

LE: Enzyklopädie der philosophischen Wissenscha en im Grundrisse, erster Teil: die Wissenscha der Logik. (Tr. Hegel’s Logic. Being Part One of the Encyclopaedia of the Philosophical Sciences (1830), by W. Wallace, with foreword by J. N. Findlay). Phän.: Phänomenologie des Geistes. (Tr. Hegel’s Phenomenology of Spirit, by A. V. Miller with analysis of the text and foreword by J. N. Findlay, Oxford: Oxford University Press, 1977). PR: Grundlinien der Philosophie des Rechts. (Tr. Elements of the Philosophy of Right, by H. B. Nisbet, ed. by A. Wood, Cambridge: Cambridge University Press, 1991). WL: Wissenscha der Logik. Science of Logic. (Tr. Hegel’s Science of Logic, by A. V. Miller, foreword by J. N. Findlay, Humanity Books, George Allen & Unwin, Ltd, 1969). 5 See the works cited in the previous note. 6 Thanks to Rolf-Peter Horstmann for pressing me to clarify this point. See also Longuenesse, Béatrice (2007), Hegel’s Critique of Metaphysics, Transl. N. J. Simek, Cambridge: Cambridge University Press, Chapter 5, especially pp. 170–171.

Kant and Hegel on the Moral Self

97

2 Kant’s Moral Self and Kant’s Metaphysics I will consider ve di erent contexts in which Kant characterizes what I call “the moral self” and he calls “the proper self.” I will argue that the metaphysical baggage needed is signi cantly di erent from one context to the next. I will try to understand why and ultimately to answer my question above: how much metaphysics, and what kind of metaphysics, is needed to account for Kant’s notion of a moral self? II.1 In the Paralogisms of Pure Reason, in the Critique of Pure Reason, Kant argues that rationalist metaphysicians’ defense of the dualism of mind and body results from a series of implicit and invalid inferences from logical features of the concept “I” in the proposition “I think” to real features of an entity that this “I” is supposed to represent.⁷ However, Kant concludes his criticism of each sophistical inference from logical features of “I” (properties of the concept “I”) in “I think” to real features of a thing (a purported substance) by what may appear to be a concession to rationalist metaphysicians. The relevant rationalist concepts can remain in use, Kant says, as long as one does not lose sight of the fact that they are just that: concepts from which, on their own, no property of a corresponding thing can be derived. Kant makes no mention of a moral use of these concepts until he reaches the Third Paralogism, the Paralogism of Personhood.⁸ In this Paralogism Kant crit-

7 Thus (i) “I” can only have the place of subject, not predicate, in the proposition “I think;” (ii) “I” is simple; (iii) “I” in “I think” expresses the unity and identity of an activity of thinking we take to be our own; (iv) in thinking “I think” we are immediately certain of our own existence. From these merely logical features (features of the representation “I” as a concept embedded in the proposition “I think”) rationalist metaphysicians conclude to the features of a purported entity that can exist only as subject not predicate, that is simple (indivisible), that is conscious of its own identity as a continuing entity, that is immediately certain of its own existence. (See A 348–80, B 406–32. For discussions of Kant’s Paralogisms of Pure Reason, see Horstmann, RolfPeter (1984), “Kants Paralogismen”, in Kant Studien 84 (4), 408–25; Rosefeldt, Tobias (2000), Das logische Ich. Kant über den Gehalt des Begri es von sich selbst, Berlin: PHILO, Verlagsgesellscha GmbH; Emundts, Dina (2006), “Die Paralogismen und die Widerlegung des Idealismus in Kants Kritik der reinen Vernun ”, in Deutsche Zeitschri für Philosophie 54 (2), 295–309; Longuenesse, Béatrice (2007), “Kant on the Identity of Persons”, in Proceedings of the Aristotelian Society, vol. CVII, Part 2, 149–167; and Longuenesse, Béatrice (2008), “Kant’s ‘I think’ versus Descartes’s ‘I am a thing that thinks’”, in D. Garber & B. Longuenesse (eds.), Kant and the Early Moderns, Princeton: Princeton University Press, 9–31.) 8 As Michael Rosen has pointed out to me, in this context “personhood” is a better translation for Persönlichkeit than “personality”. What Kant has in mind is the property of being a person, which he identi es with the property of being conscious of one’s own numerical identity through time, where “being conscious” is a factive notion: being conscious of one’s own identity through

98

Longuenesse

icizes the rationalist inference from the consciousness of numerical identity expressed in the mere concept “I” as used in the proposition “I think,” to the purported consciousness of the numerical identity of an entity, a person. In fact, Kant argues, the consciousness of numerical identity expressed by “I” in “I think” is merely the consciousness of the unity, and thus numerical identity, of a process of thinking I am justi ed, for reasons elucidated elsewhere in the rst Critique, in taking to be my own. In no way does that make it a consciousness of the numerical identity of an entity. So, from the fact that we do have the former kind of consciousness of identity (expressed in “I think”), we cannot infer that we have the latter kind of consciousness of identity and thus are persons in the metaphysical sense cited above, where a person is an entity that remains numerically identical to itself through time and is conscious of that numerical identity through the changes of its own states.⁹ However, immediately a er completing his criticism of the Paralogism of Personhood, Kant adds: Meanwhile the concept of personhood, just like the concepts of substance and of the simple, can remain (insofar as it is merely transcendental, i.e. a unity of the subject which is otherwise unknown to us but in whose determinations there is a thorough connection of apperception), and to this extent this concept is necessary and su cient for practical use (A 366, emphasis mine).

It seems clear that the concept that “can remain” is the concept of a person as an entity of which we are supposed to be immediately conscious in thinking “I” in “I think,” namely the concept Kant has just criticized. This concept nevertheless “can remain” as long as one does not claim to know anything about the object of that concept beyond the fact that it is one (it remains numerically identical through time) and its states are “thoroughly connected by apperception” (A 365). To that extent, namely within these limits of what we can claim to know about it, this concept is “necessary and su cient for practical use.”

time is being conscious of what is, in fact, the identity through time of an entity, oneself as the thinker of the thought “I think.” 9 Note that according to Kant’s argument in the Third Paralogism, the only way we can be conscious of the numerical identity of ourselves as an entity is in being conscious of ourselves as a spatiotemporal, empirical, entity among other spatiotemporal empirical entities (see A 363). But this consciousness does not rest only on the pure consciousness of identity expressed by “I” in “I think.” The consciousness of the identity through time of oneself as a spatiotemporal entity is the consciousness of oneself as a person in the empirical sense. This consciousness will have an important role in the third tenet of Kant’s moral self. See below, II.3, the moral self in the Metaphysics of Morals.

Kant and Hegel on the Moral Self

99

Why “necessary and su cient for practical use”? Kant does not give much indication here. “Practical use” might mean what he calls elsewhere technicalpractical use, namely any use of reason in the determination of action, which calls upon the capacity to adjust means to ends and to keep track of the action as one’s own. Or it might mean moral-practical use, where in addition to the technical capacity for adjusting means to ends and keeping track of one’s actions, the autonomous determination of the will under the moral law would have to be in play.¹⁰ Kant, I submit, has the latter sense of “practical” in mind here. What he means is that taking the mere concept “I” in “I think” to express the consciousness of the numerical identity of an entity is taking this “I” to be the concept of a purely intelligible object, an object to which we have cognitive access by the mere concept “I” and thus independently of the conditions of sensibility. The reason Kant claims that the concept of a person which “can remain” is “necessary and su cient for practical use” would then be that such a concept makes it possible to think of oneself as belonging to a purely intelligible world, free of the conditions of causal determinism proper to the sensible world, and thus metaphysically free. And metaphysical or “transcendental” freedom, for Kant, is a necessary condition of moral responsibility.¹¹ Now, leaving space for the possibility of metaphysical freedom is not asserting the existence of metaphysical freedom. On Kant’s own account, the latter depends on our consciousness of the determination of our actions under laws of our own doing,¹² a concept that is neither contained in nor entailed by the concept of a person as an entity that is conscious, through the mere concept “I,” of its own identity and that has unity of apperception of its states. So the concept that “can remain” a er the criticism of the Third Paralogism, although necessary, does not seem su cient for practical use, least of all for the moral-practical use.¹³

10 See KU, Introduction, 5:172. 11 See A 534/B 562. 12 See for instance the solution to the Third Antinomy of Pure Reason (A 545/B 573–A 558/B 586). See also the general remark to the Paralogisms of Pure Reason in the second edition (B 430–31). 13 One might also object that the concept of rationality is not contained in the concept of person that “can remain” a er the criticism of the Third Paralogism, which makes the latter even further from being “su cient” for any practical use at all. But here Kant has the resources to reply that “unity of apperception” does entail rationality, at least in the theoretical sense. The “connection in apperception” of one’s states is not just their psychological continuity in memory (as Locke, for instance, would have it), but their rational connectedness, a connectedness that, as Kant has argued earlier in the Critique, is a necessary condition for any objective cognition as well as for the instrumental and prudential use of reason. Still, this does not entail practical rationality in the moral sense, which takes us back to the same objection I formulate in the main text: even if we take “person” to be an existing entity that we represent by the mere concept “I” in “I think” and

100

Longuenesse

Nevertheless, we at least have here a rst tenet for Kant’s characterization of a moral self, in the form of a necessary condition for being a moral self or, in Kant’s terms, a “proper self” or a “self properly speaking”: an entity is a moral self only if it is a subject conscious, in referring to itself by the concept “I,” of its own identity through time¹⁴ and of the thorough connection of its determinations in apperception. II.2 As noted above, this rst tenet appears to be a necessary, not a su cient, condition for being a moral self. The additional condition for the moral-practical use of reason is stated in the concluding remarks to the Paralogisms in the second edition of the Critique.¹⁵ The latter is a follow up to section III of Groundwork of the Metaphysics of Morals, which precedes the second edition of the Critique. There Kant maintains that insofar as I think of myself as bound by the moral law, I think of myself under the idea of freedom. And insofar as I think of myself under the idea of freedom, I just am free, namely spontaneously self-determining, laying down the laws of my own existence independently of the conditions of sensibility. To that extent, I am a pure intelligence, belonging in a purely intelligible world. Kant adds, however: The concept of a world of understanding is […] only a standpoint that reason sees itself constrained to take outside appearances in order to think of itself as practical, as would not be possible if the in uences of sensibility were determining for the human being but is nevertheless necessary insofar as he is not to be denied consciousness of himself as an intelligence and consequently as a rational cause active by means of reason, that is, operating freely (Grundlegung, 4:458).

Is this a metaphysical statement? Not really. It does not say anything about the nature of the being that necessarily thinks of itself as free. So the statement belongs neither to a rationalist metaphysics, which would claim knowledge of such a being by a priori conceptual reasoning, nor to a metaphysics in Kant’s sense, which, as we will see in a moment, requires that some minimal amount of empirical information be available for any claim to a priori cognition of an existing entity to get o the ground. This being so, it is tempting to give the talk of “two standpoints”

that is, as such, conscious of its own identity through time, we do not thereby have the concept we need to ground its characterization as freely determining its own actions. 14 Note that “conscious of its own identity through time” leaves it open whether the entity one is conscious of in using the term “I” is an entity known by mere concepts – in particular, the mere concept “I” – or also by its empirical properties as an entity determined in space and time. The concept of a person that “can remain” is the former. But it can also be interpreted as the latter. It is the latter that we will encounter in the third tenet, the moral self in the Metaphysics of Morals. See n. 9 above. 15 See n. 13.

Kant and Hegel on the Moral Self

101

a fairly de ationary meaning. One will then interpret Kant as just saying that to think of oneself by the concept “I” in “I ought to X,” namely to think of oneself as capable of setting for oneself the moral norms for the permissibility or impermissibility of one’s intended actions, just is occupying a standpoint on oneself that is radically distinct from the standpoint of causal explanation. And to think of oneself in that way just is being that way. From a contemporary standpoint this is a familiar thesis. The standpoint of justi catory reasons, as the standpoint of agency, is irreducible to the standpoint of causal explanation. Indeed, substituting the latter for the former amounts to a form of mental pathology. Someone occupying the stance of the spectator with respect to her life and actions is someone in whom agency just comes to a halt. One might even say: even with the best knowledge of the causal explanation of one’s own motivations for producing the e ect one is in the process of producing in the world, the sense that committing oneself to an action is up to oneself is a condition for the very possibility of action.¹⁶ But of course Kant does not reduce the distinction between the intelligible and sensible worlds to a distinction between two standpoints in the way I have just described. The description I have given says no more than this: however complete the causal explanation of a particular action/event, it does not make irrelevant the standpoint of the agent seeking and evaluating justi catory reasons for her actions. The two standpoints in this description are those of third-person theoretical cognition and rst-person agency regarding one and the same event and belonging to one and the same empirical world. In contrast, Kant’s “two-standpoints” view means that there exists, as the ground of the empirical world and its universal causal laws, an intelligible world which we cannot know, but where we are at least justi ed in thinking that our true nature is that of pure intelligences determining their actions under laws of freedom. It’s still true that for Kant, our ignorance of our own purely intelligible nature is no hindrance to our agency or to its justi catory structure. In this sense, at least, no metaphysics is presupposed in Kant’s account of the moral self. To characterize a self as a moral self or a “self properly speaking” (eigentliches Selbst), nothing further is needed than the structure of her motivation (here: being motivated un-

16 For a forceful development of this kind of view, see Moran, Richard (2001), Authority and Estrangement, an Essay on Self-Knowledge, Princeton and Oxford: Princeton University Press, esp. Chapter 3. Note that “up to oneself” here does not mean that the choice of how to act is arbitrary. Rather, it means that a necessary condition for the action to take place at all is that the agent be in a position to endorse it and commit herself to it. The loss of that capacity amounts, in pathological cases, to a form of mental and agential paralysis.

102

Longuenesse

der the primacy of the categorical imperative). That structure alone counts as the way she knows herself as free.¹⁷ In any event, we now have the second tenet in Kant’s characterization of a moral self: to be a moral self is to act under the highest normative principle of the moral law, which for us takes the form of a categorical imperative whose command is not dependent on the conditions of sensibility but rather legislates on those conditions. Now for Kant, one progresses from the mere critique of reason in its theoretical and practical use to metaphysics properly speaking only when one includes in one’s considerations the minimal amount of empirically determinate content required to allow the a priori principles one has established in the Critiques to be applied to actually existing entities. In the case of the metaphysics of nature, the minimal empirical content is that of the concept of matter. In the case of the metaphysics of morals, the minimal empirical content is the concept of a human being. The third context in which Kant’s concept of a moral self is to be considered, then, is that of the Metaphysics of Morals. II.3 Just as the Critique of Pure Reason is supposed to justify the fundamental principles that make possible a metaphysics of nature, so the Groundwork of the Metaphysics of Morals and the Critique of Practical Reason are supposed to justify the fundamental principles that make possible a metaphysics of morals. One progresses from the Critique of Pure Reason to a metaphysics of nature by introducing the minimal amount of empirical information that gives its content to the concept of matter as “object of the outer senses” (Naturwissenscha , 4:481) and “movable in space” (Naturwissenscha , 4:480). Similarly, one progresses from the Critique of Practical Reason to a metaphysics of morals (a system of systematically organized, synthetic a priori laws for actually existing entities) by introducing the minimal empirical concept of a human being as a living being that is both pathologically a ected and rationally self-determining, a person.¹⁸

17 This knowledge of freedom as knowledge of the moral law is most strikingly stated in the Preface to the Critique of Practical Reason: the moral law is the ratio cognoscendi of freedom, freedom is the ratio essendi of the categorical imperative. Still, the fact that freedom is the ratio essendi of the moral law makes it ontologically prior, in the sense of an ontological condition, for the very existence of the moral law as the structure of moral motivation. And this does give a strong metaphysical avor to Kant’s characterization of the moral self: a moral self just is an entity that is metaphysically free (i.e., impervious to natural causal determinism). 18 See Naturwissenscha , 4:473–79; Sitten, 6:216–17. As Paul Guyer has pointed out to me, Kant’s discussion of duties in the Metaphysics of Morals draws on a profusion of particular empirical facts concerning the situation of particular human beings in particular social circumstances. This makes it very di erent from his discussion of the laws of motion in the Metaphysical Foundations of Natural Science, which consists in applying to the empirical concept of matter as “what is mov-

Kant and Hegel on the Moral Self

103

The concept of a person is de ned in Groundwork II: Rational beings are called persons because their nature already makes them out as an end in itself, that is, as something that may not be used merely as a means, and hence so far limits all choice (and is an object of respect) (Grundlegung, 4:428). […] Every rational being, must always take his maxims from the point of view of himself, and likewise every other rational being (who for this reason are also called persons) (Grundlegung, 4:438).

This concept of person is the rational, a priori component in the empirical concept of human being, that is, an empirical entity that has the property of personhood: Just as there must be principles in a metaphysics of nature for applying those higher universal principles of nature in general to objects of experience, a metaphysics of morals cannot dispense with principles of application, and we shall o en have to take as our object the particular nature of human beings, which is cognized only by experience, in order to show in it what can be inferred from universal moral principles. But this will in no way detract from the purity of these principles or cast doubt on their a priori source. This is to say, in e ect, that a metaphysics of morals cannot be based upon anthropology but can still be applied to it (Sitten, 6:216–17).

And now the two aspects of personhood we encountered in the remarks to the Paralogisms of Pure Reason – consciousness of one’s own identity through time and connection of one’s own states in apperception – are explicitly related to the notion of imputability:

able in space” the Principles of Pure Understanding justi ed in the rst Critique. In the Metaphysics of Morals, by contrast, Kant discusses for instance the institution of marriage, the mutual obligations of masters and servants, and so on. The structural correspondent in the domain of nature would seem to be a discussion of particular empirical laws rather than the general laws of motion (e.g., Newton’s three fundamental laws in Principia), and thus of empirical science rather than the metaphysics of nature. I’m inclined to say that what would strictly speaking correspond, in the domain of morality, to the fundamental laws Kant discusses in the Metaphysical Foundations of Natural Science would be the three formulations of the categorical imperative and their variants in section 2 of Groundwork. For these formulations do rely on a minimal concept of a human being as not just rational (a person) but also sensible and pathologically a ected. In fact, section II of Groundwork is entitled: “Transition from popular moral philosophy to metaphysics of morals,” where the “metaphysics of morals” does seem to be the system of formulations of the categorical imperative. If this is correct, contrary to Kant’s description the bulk of the Metaphysics of Morals would then parallel empirical natural science, including, e.g., Newton’s law of universal gravitation, but also a profusion of more particular and contingent empirical laws, rather than metaphysical laws of nature as Kant understands them. I will ignore this complication, which is of no consequence for my discussion since I draw only on what Kant says, even in The Metaphysics of Morals, about moral selves as human beings in general.

104

Longuenesse

A person is a subject whose actions can be imputed to him. Moral personhood is therefore nothing other than the freedom of a rational being under moral laws (whereas psychological personhood is merely the ability to be conscious of one’s identity in di erent conditions of one’s existence). From this it follows that a person is subject to no other laws than those he gives to himself (either alone or at least along with others) (Sitten, 6:223).

With those two minimal empirical components – the empirical yet a priori determined concept of matter and the empirical yet a priori determined concept of a human being, or a person in the empirical sense – we now have the structure of Kant’s system: Logic (the a priori exposition of the structure of thought, both theoretical and practical), metaphysics of nature, metaphysics of morals. Within that system, the entity of which it can be said that the laws of its existence are those of self-determination under the moral law is the human being that has the a priori character of being a person: a being accountable for her own actions and, as such, one who acts under the representation of the moral law. This gives us the third tenet in Kant’s notion of a moral self: a moral self is an empirically given human being that has personhood. Here the concept of personhood contains both the psychological concept – to be a person is to be an entity that is conscious of its own identity through time – and the moral concept – to be a person is to be an agent who is accountable for her own actions. Strikingly, in this context, that of the Metaphysics of Morals, we do know what kind of entity this is: a human being, a “complete” self that is both a “proper” self and a “dear” self.¹⁹ We still have not encountered a moral self as an immortal soul. It only now makes its entrance. This will be the fourth tenet. II.4 As I said earlier, a characterization of the moral self as an immortal soul is presented only in the context of the solution to the Dialectic of Pure Practical

19 To be clear: I do not mean to say that for Kant the moral self has to be an empirical self, or that consciousness of the identity through time of an empirical human being is a necessary condition for moral personhood. What is a necessary condition for moral personhood is the consciousness of one’s own identity through time as an entity. The concept of a person that “can remain” a er the criticism of the Paralogisms is such a concept, where it is accepted, “as necessary [and su cient] for moral use” that such a consciousness is carried by the mere use of the concept “I.” This is the concept of a person as a purely intelligible being (having access to itself via the mere concept “I”). However, in the Metaphysics of Morals, imputability is a property of the empirical person, who is conscious of her own identity as a spatiotemporal entity. This does not make the concept of a person an anthropological concept, since it is still the case that a person is a human being having personhood where the concept of personhood is the a priori concept de ned in Groundwork, and any rational being, whether human or not, would count as a person and be subject to the moral law. Thanks to Rolf-Peter Horstmann for pressing me on this point. On the relation between the a priori and empirical concepts of person, see Longuenesse (2007) “Kant on Identity of Persons”.

Kant and Hegel on the Moral Self

105

Reason. There Kant claims that the highest good, namely not only the supreme good (das oberste Gut), virtue, but the complete good (das vollendete Gut), virtue and happiness proportionate to it, is the complete object of pure practical reason. The rst condition for the realization of the highest good is that virtue, its unconditioned condition, be achieved. But given the radical imperfection of human beings, we can believe in the possibility of such achievement only as the result of an inde nite progress toward virtue. So pure practical reason must postulate the immortality of the soul if the object it necessarily sets itself as its highest goal is even to be a possible object for it. Aiming at the highest good, made necessary by respect for the moral law and the presupposition owing from this of its objective reality, lead through the postulates of practical reason to concepts that speculative reason could indeed present as a problem but could never solve. Thus it leads to 1) the problem in the solution of which speculative reason could do nothing but commit paralogisms (namely, the problem of immortality) because it lacked the mark of permanence by which to supplement the psychological concept of an ultimate subject, necessarily ascribed to the soul in self-consciousness, so as to make it the real representation of a substance; this mark practical reason furnishes by the postulate of a duration required for conformity with the moral law in the highest good as the whole end of practical reason (KpV, 5:134).

This takes me back to the point I was making at the beginning of this paper. Absent the theory of the highest good, Kant’s conception of the moral self could have been framed within a metaphysical model that had de nitively set aside the Wolfan structure of metaphysics. Just as the metaphysics of nature is the application to the concept of matter of the a priori principles laid out and justi ed in the Critique of Pure Reason, the metaphysics of morals is the application to empirical human beings of the a priori principles laid out and justi ed in the Groundwork of the Metaphysics of Morals and the Critique of Practical Reason. On this metaphysical picture, empirical reality would be partitioned into those entities for which natural causal laws are the only relevant laws on the one hand; and those entities for which normative, and especially morally normative laws are relevant.²⁰ But now, with the postulate of the immortality of the soul introduced in the context of the resolution of the problem of the highest good, we are back to the Wol an picture of special metaphysics: rational cosmology (with the idea of freedom), psychology (with the idea of the immortality of the soul), theology (with the idea of God,

20 I am leaving out of consideration the added level of complexity introduced by Kant’s Third Critique, the Critique of the Power of Judgment and the consideration of living beings. Let me just note that they do not, according to Kant, constitute a new domain endowed with its own kind of legislation.

106

Longuenesse

which is the objet of the second postulate of pure practical reason, necessary for thinking the possibility of the highest good as the complete good: that virtue be rewarded with happiness). What happened? A detailed assessment of the reasons for this return to the Wol an model is beyond the scope of this paper. Let me only note that the idea of the moral self as an immortal soul rests on distinctly di erent grounds from those that I have called the rst three tenets of Kant’s notion of the moral self. I brie y mentioned one ground at the beginning of this paper. In making the Highest Good, as the complete good (the unity of virtue and happiness), the necessary object of pure practical reason, Kant attempts to re-unify what he has rigorously divided: the end of moral-practical reason, virtue, and the (albeit indeterminate) end of our sensible nature, happiness. Even more importantly, the idea of the Highest Good as the complete good rests on an idea of an unconditioned totality, which our nite mind can represent only as the unattainable goal of an inde nite progress.²¹ The idea of the moral self as an immortal soul is thus based on grounds that are signi cantly di erent from those that I have called the rst three tenets of Kant’s notion of a moral self. In any event, we now have a fourth tenet for Kant’s notion of a moral self: a moral self is an entity that necessarily postulates its own nature as an immortal soul. II.5 Kant adds, however, that the idea of an inde nite progress is itself only an approximation of what for an intuitive intellect is represented as the atemporal standard of holiness presented to our nite will: The eternal being, to whom the temporal condition is nothing, sees in what is to us an endless series, the whole of conformity with the moral law. And the holiness that his command in exibly requires in order to be commensurable with his justice in the share he determines for each in the highest good, is to be found whole in a single intellectual intuition of the existence of rational beings (KpV, 5:123).

This gives us a surprising h tenet in Kant’s notion of a moral self: a moral self is an atemporal component in the eternally existing whole of virtue that is the condition of “the share of each” in the highest good.

21 Thanks to Ralf Bader for reminding me that the argument for the Postulate of the immortality of the soul does not even appeal to the unity of virtue and happiness: the latter constitutes the complete good that, as a necessary object of pure practical reason, grounds the postulation of a benevolent God in the second Postulate. What is common to the arguments for both postulates is the idea of an unconditioned totality. This concept is what gives rise to a seemingly insuperable con ict of reason, in its practical as well as in its theoretical use. This con rms again that the idea of the immortality of the soul rests on signi cantly di erent grounds than what I have called the rst three tenets in the characterization of a moral self, where what is at stake is just the structure of motivation of a moral self, independently of any notion of a totality of the Good (see KpV, 5:124–132). Thanks to Kelley Schi mann for pressing me on this point.

Kant and Hegel on the Moral Self

107

So now we have ve stages and ve tenets in Kant’s construction of his metaphysical account of the moral self – namely his answer to the question: what features if any, should we assign a priori to the entity referred to by “I” in the moral “I ought to”? Let me list them again: (i) A moral self or “proper self” is a subject who connects its own determinations in apperceptions and is thus conscious of its own identity through time in representing itself as “I.” (ii) A moral self is an entity that acts under the highest normative principle of the moral law. As such it has two standpoints on itself: one according to which it belongs to a purely intelligible world, the other according to which it belongs to an empirical world, determined according to empirical causal laws. (iii) A moral self is an empirically given human being that has personhood both in the psychological sense (consciousness of its own numerical identity through time) and in the moral sense (moral accountability). (iv) A moral self is an immortal soul. (v) For an intuitive intellect, each moral self is an atemporal component in the eternally existing whole of virtue (holiness) as the unconditioned condition of the highest good, the unity of virtue and happiness proportionate to it. There are, of course, connections between these characterizations. But they are also strikingly heterogeneous. Given (i), (ii), and (iii), and supposing we gave the de ationary account I suggested above to (ii) – and thus also to (i) – Kant could have characterized the moral self as a human being having the property of personhood in both the psychological and the moral sense, as de ned in the Metaphysics of Morals. His system would then have been complete on the old Greek model of logic, physics, ethics: transcendental system of principles, metaphysics of nature, metaphysics of morals. But he does not do that. Instead, the need to solve the problem of the highest good leads him to return to the structure of the rationalist metaphysics he criticized in the rst Critique, as he expressly indicates in the solution to the Dialectic of the Critique of Practical Reason: rational cosmology, rational psychology, rational theology (KpV 5:132 f.). Now Hegel, in contrast to Kant, does hold on to the old Greek model of logic, physics, ethics, which become for Hegel Science of Logic, Philosophy of Nature, Philosophy of Spirit, and which he takes to restore in its old glory the metaphysical project that Kant had reduced to insigni cance by promoting the motto: We have no knowledge of things as they are in themselves. How signi cant is this model for Hegel’s account of Kant’s moral self? And does this account shed retrospective light on Kant’s notion of a moral self?

108

Longuenesse

3 Hegel’s Metaphysics and Kant’s Moral Self In the Preface to the Science of Logic in its rst edition, Hegel laments the contempt into which metaphysics has fallen.²² Responsible for this downfall is the at-footed popular version of Kant’s philosophy that reduces the critical system to the easy motto: no knowledge beyond the limits of experience (WL, 5:13–14; 25). Kant is not without responsibility in this asco, since his critical philosophy is the most systematically developed form of the modern “philosophy of re ection” according to which “thinking is mere thinking, and depends for its content on sensible intuition” (WL, 5:38; 45), with the result that we have no knowledge of things as they are in themselves. Any truth available to us is only subjective truth, truth for us. It is, says Hegel, as if one granted a man correct insight, but added that he nevertheless has insight into nothing true, but only into what is untrue. As inconsistent as this would be, no less inconsistent is a true cognition that supposedly doesn’t cognize the object as it is in itself (WL, 5:39; 46).

Hegel repeats his charge against Kant in the Introduction to the Subjective Logic and adds that Kant should have known better. He had, with his concept of an intuitive intellect, what was needed to overcome the separation between thinking and its object or content (WL, 6:266; 593). A few pages earlier Hegel also credited Kant with the groundbreaking discovery of the relation between “I” and concepts. I don’t have concepts. I just is [sic] the concept (WL, 6:253–4; 583). I cannot discuss these two points with the detail they would call for: Hegel’s endorsement of Kant’s intuitive intellect²³ and Hegel’s characterization of Kant’s “I” as the concept. I will at least try to clarify them as much as needed to make some progress in answering the following question: in the context of Hegel’s new metaphysics, what becomes of Kant’s metaphysical account of the moral self in the ve stages I have identi ed in Part II? Kant’s most developed account of what he means by “intuitive intellect” is given in §§ 76–77 of the Critique of the Power of Judgment. An intuitive intellect is an intellect for which thinking and intuiting would be one and the same. Such an intellect would not be dependent on sensibility for providing its concepts with the intuition that gives them their content. There would thus be no distinction,

22 Note that this is strangely reminiscent of Kant’s own deploration of the downfall of metaphysics in the Preface to the rst edition of the Critique of Pure Reason. 23 As should be clear from Hegel’s complaints about Kant’s responsibility in the downfall of metaphysics, this is a very quali ed endorsement, to say the least. See also III.1, below, p. 112. And Longuenesse (2007) Hegel’s Critique of Metaphysics, pp. 190–91.

Kant and Hegel on the Moral Self

109

for such an intellect, between possibility and actuality. This distinction makes sense for us because the mere concept of an object is not su cient to attest to its existence. But an intuitive intellect, in thinking the object, would present it to itself in its very existence (KU, § 76, 5:402). Most importantly, the representations of such an intellect would have the structure by which Kant, in the rst Critique, has characterized intuitions rather than concepts. They would have the part/whole structure of intuitions – in which, moreover, the whole of intuition is the precondition for the determination of its parts – rather than the structure of concepts in a discursive understanding such as ours, in which lower or less general concepts are thought under more general or higher concepts, and individual objects of intuitions are never completely determined under concepts. Given its structure, an intuitive intellect is an intellect for which, unlike for us, there is no obstacle to conceiving each part of an organism as determined in its reality, not just in its position or shape, by the whole. For such an understanding, conceiving the whole just is conceiving each part of the whole, in its reality, as dependent on the whole in both its individual constitution and its combination with all other parts (KU § 77, 5:407). In contrast, our “merely discursive” understanding can represent such a dependence of the parts upon the whole only as if it was the dependence of the parts on a concept of the whole, namely an end, “by a remote analogy” with the dependence of parts on the concept of the whole in the artifacts of our own human agency. Now, Hegel’s Concept, which he also calls “I” or “pure self-consciousness,” has a structure similar to the one Kant assigns to an intuitive intellect’s representation of the whole in §§ 76–77 of the third Critique – for the structure of Hegel’s Concept is not just that of a genus under which species are subordinated, as in discursive concepts. Rather, its structure is that of a totality that is present in, or inhabits, each of its instances.²⁴ Life is such a totality: it is the totality of living beings, each of which, individually, is an instance of life, and all of which, collectively, are just what life is. Spirit is such a totality: it is the totality of individual human beings each of which is an instance of Spirit and all of which, collectively re ecting and endorsing, or as the case may be, entering into con ict with, their own collective reality, are just what Spirit is. So both Life and Spirit have the structure of what Hegel calls “the Concept.” Here’s what Hegel writes in the section on “the Universal Concept” in the Subjective Logic:

24 It is appropriate to talk of the relation of the totality to its instances rather than the relation of the whole to its parts. This does constitute a di erence between Hegel’s Concept and the representation of the totality of the world by an intuitive intellect according to Kant. I cannot further expand on this point. All I can do is o er the two examples of “Life” and “Spirit” (see below).

110

Longuenesse

Life, I, Spirit, Absolute Concept, are not Universals merely [nur] as higher genera [Gattungen]. Rather, they are concrete entities [Konkrete], whose determinations are not species [Arten] or lower genera. Rather, they [those Universals] are absolutely in themselves in their reality, and they are ful lled [erfüllt] in that reality. (WL, S. 6, 279; 605)

The list that opens this passage is somewhat awkward. “Life, I, Spirit, Absolute Concept” do not belong in the same place in the system, and yet Hegel cites all of them as examples of what he calls “the Universal Concept” and as examples of the way “the Universal Concept” relates to its particular instantiations. I’ll rst focus on “Absolute Concept” and “I,” and then return to “Life” and “Spirit.” Here’s how I understand Hegel’s point about “I” and “Concept” or “Absolute Concept.” Kant should have seen, and was on the verge of seeing, that his formal “I” in each “I think,” individuated in each particular thinker, is itself just the expression, in each individual thinker, of an activity of thinking to which each individual thinker belongs or in which each is participating. That activity is an I²⁵ in the sense in which Kant’s thinker is an I or, rather, re ects itself in the concept “I.” For thought has a normative structure for which the collective thinker takes itself to be accountable just as, and in virtue of the fact that, individual thinkers take themselves to be accountable for the validity of theoretical and practical reasoning. The same is true of the collective responsibility of the I as an agent setting for itself the norms of its actions: the universal I is normative for action in virtue of being instantiated in individuals who set the norms for their actions in the course of their interaction with other individuals who do just the same (namely set the norms for their actions in the course of their interaction with other individuals), so that all of them set the norms of their individual and collective actions in the context of their interaction with a given whole of ethical life. That’s how the Concept is both universal and singular or, as Hegel says in the Introduction to the Subjective Logic, “individual personhood (individuelle Persönlichkeit),” (WL, 6:253; 583). Moreover, the Concept, as a uni ed, ongoing thinking activity that is realized in individual thinkers is a structure “both of nature and of spirit” (WL, 6:257; 586). In nature, it is Life. In the Science of Logic, Life is the rst “moment,” or gure, of the Idea, namely the Concept knowing itself in its own reality or objectivity (WL, 6:462–63; 755). Life is the Concept as immediate, unre ected experience of its own reality, or of itself in its own reality, where its reality is that of a totality of living

25 Here I am writing I, without scare quotes and in italics, to indicate that I am talking about Hegel’s ontological I, not the representation, concept or term “I” as it appears in Kant’s “I think.” As I explain in the main text, Hegel’s I or Concept is neither an individual self (the referent of the Kantian “I”) nor just a concept or representation.

Kant and Hegel on the Moral Self

111

beings, each of which is an instance of that immediate self-relatedness of the Concept, as a whole. As Spirit, the Concept is also knowledge of its own reality, or knowledge of itself in its reality, that is to say, in Hegel’s vocabulary “Idea.” But in this case, it is consciousness of its own reality not just as Life but as having emerged from Life as a process of knowing and a process of willing that generates is own reality. And again, knowing and willing are processes that exist only as a totality, albeit as a totality that is realized in each individual thinker and agent. What does any of this have to do with the notion of an intuitive intellect to which Hegel claims Kant should have done better justice? I have already pointed out the structural similarity between the two: in the Concept, just as in the representations of the Intuitive Intellect according to Kant, thinking the whole just is thinking the parts, indeed thinking the whole is thinking the determinate existence of the parts (in Hegel it would be truer to say: the particular elements, rather than the parts).²⁶ To this structural similarity corresponds an ontological similarity: Hegel’s Concept, like Kant’s intuitive intellect, is pure activity. There is, however, a fundamental dissimilarity between Hegel’s Concept and Kant’s intuitive intellect. As Hegel says, the Concept just is the self-re ection of the whole, buried in the life of the whole as “mere” Concept, emerging from the whole as knowing and willing, realizing itself in the whole as Spirit. There is no indication, in Kant’s very minimal characterization of the intuitive intellect, that it should be conceived as one with what it thinks: Kant’s intuitive intellect is more like a creationist God than like Spinoza’s Substance. But Hegel’s Concept inherits the monism of Spinoza’s substance. It is Spinoza’s substance plus personhood (WL, 6:195; 536), where re ection and willing have been given their full due, and where individuals are not just modes of the in nite substance. Rather, they are those entities by which and in virtue of which the substance is a subject in Hegel’s sense, inherited from Kant’s subject as represented by the concept “I”²⁷ – selfre ective, self-normative. Now we can attempt to answer our question: what becomes of Kant’s moral self in Hegel’s new metaphysics? Or in other words: what are the consequences, for Kant’s notion of a moral self, of the fact that Hegel starts where Kant ended? As we saw, Kant ended with a vague gesturing to, rather than a systematic elaboration of, what both the world and moral selves would look like for an intuitive intellect. Hegel starts with a drastically revised version of an intuitive intellect, the

26 Cf. above, n. 24. 27 See Phän., S. 3:28; 14.

112

Longuenesse

standpoint of the Concept. What are the consequences of this shi of perspective for Kant’s notion of a moral self? My answer will be all too schematic. What I propose to do is to take one by one, but in reverse order, the ve tenets of Kant’s moral self I outlined in the second part of this essay. III.1 Hegel on Kant’s Fi h Tenet, the intuitive intellect. As we saw, according to Kant an intuitive intellect would grasp at one glance the atemporal whole of reality in which, from all eternity, we are ascribed our share of the highest Good, namely the a priori synthetic unity of virtue and happiness. In contrast, given the limitations of our cognitive capacities (an understanding that is “only” discursive, a sensibility whose intuition has temporal form) we can represent the highest good only as the horizon of the inde nite progress of our imperfect souls toward virtue. For Hegel, however, the standpoint of an intuitive intellect is our own, insofar as we have reached the standpoint of the Concept. And the Concept – namely we, as instantiations of the Concept – has immediate access to its own reality in life, and mediate access to it in Spirit, where the Concept re ects upon its own reality and further develops it, according to its own self-set norms. One Idea of the Good, whatever the internal tensions and contradictions it may be riddled with insofar as the universal and particular are not recognized as one, sets the immanent norms for one and the same activity, that of the Concept.²⁸ Whereas for Kant we do not and cannot have any notion of the way an intuitive intellect presents to itself the notion of the Highest Good, for Hegel we do know how it does so: indeed, we are the intuitive intellect insofar as we know ourselves, theoretically and practically, to be that in which the activity of the Concept dwells, immersed in life on the one hand, and generating its own world of objective Spirit on the other hand – or rather doing both at once and inseparably. Having this knowledge makes irrelevant the supposition of an immortal soul, which brings me to the next point. III.2 Hegel on Kant’s Fourth Tenet, the moral self as an immortal soul. The postulation that each individual moral self is an immortal soul and thus apt to inde nite progress toward virtue and its corresponding degree of happiness is a representation that belongs to what Hegel calls in the Phenomenology of Spirit the “moral view of the world.” As explained in the corresponding chapter of the Phenomenology of Spirit,²⁹ the moral view of the world is the expression in individual consciousness of a stage in the development of the Concept in which thought de-

28 See, for example, WL, S. 6:241–48; 818–23; LE § 234 and Addition, S. 8:286–87; 290–91; PR §§ 130–131, S. 8:243–44; 157–58. 29 See Phän., S. 3:442–464; 365–74.

Kant and Hegel on the Moral Self

113

terminations are re ected as being in opposition to the reality from which they come and to which they return. Subjects that postulate themselves as immortal souls and depend on their individual immortality for instantiating the realization of the highest good do not know themselves for what they really are: an imperfect, one-sided moment in the development of Spirit. Exit the immortal soul as a true account of even Kant’s notion of the moral self. For Kant’s notion of a moral self, however limited, has more resources than this sad recourse to mere belief or faith would have us think: However essential it may be to emphasize the pure and unconditional self-determination of the will as the root of duty – for knowledge of the will rst gained a rm foundation and point of departure in the philosophy of Kant, through the thought of its in nite autonomy […] – to cling on to a merely moral point of view without making the transition to the concept of ethical life [Sittlichkeit] reduces this gain to an empty formalism, and moral science to an empty rhetoric of duty for duty’s sake. […] If a duty is to be willed merely as a duty and not because of its content, it is a formal identity which necessarily excludes every content and determination. The antinomies and shapes assumed by this perennial obligation, among which the merely moral point of view simply dri s to and fro without being able to resolve them and to get beyond obligation are developed in my Phenomenology of Spirit. (PR, § 135, S. 7:252–53; 162–63)

The “merely moral point of view” that does not make “the transition to the concept of ethical life” is prevalent all the way to Kant’s Metaphysics of Morals. At least there, however, both the universality of personhood and the particular traits of empirical human beings are taken into account. Which brings us to Kant’s third tenet. III.3 Hegel on Kant’s Third Tenet, the moral self as an empirical human being that has personhood. This is a characterization of the moral self that Hegel, I submit, accepts: this is what a moral self truly is, with all the limitations involved in the separation between the property of personhood and its “in nite autonomy”³⁰ and the particular needs of empirical human beings that give their content to moral obligation. But Hegel’s account of the moral self, or the self of morality, is a historical account. According to the chapter on Morality in the Principles of the Philosophy of Right and in the Philosophy of Spirit (Part 3 of the Encyclopedia of Philosophical Sciences), the emergence of the moral self occurs in connection with the emergence of an objective con guration of Spirit: abstract right. Abstract right is the system of contractarian laws regulating the institution of private ownership. A condition for the institution of such laws is that individuals think of each other and treat each other as persons in the juridical sense, namely as individuals hav-

30 See the text cited above.

114

Longuenesse

ing the capacity to enter into freely-endorsed relations of contract (PR § 36, 7:96; 69; cf. also EG § 486, S. 10:304; 218). So the existence of the institution of property rights depends on the fact that Spirit, namely a determinate totality of nite minds, has reached the point where it can think the notion of “abstract freedom,” freedom as the pure capacity to enter, by one’s own choice, into determinate, selfprescribed contractarian rules for ownership over external things. But the existence of such systems of rules is what, in turn, breeds the emergence of morality, that is to say, in Hegel’s terms, the existence of individuals as not only persons (a “person” is, for Hegel, a strictly juridical notion, inherited from Roman law), but also moral subjects: self-determining wills whose freedom consists in making their own both the norm of abstract universality in the form of the moral law and the goal of happiness or well-being (PR § 130, S. 7:243; 157) that has prevailed in the more immediate system of mutual relations human beings enter into, that of the family and its life-preserving and life-perpetuating function. The ground is thus prepared for generating a world in which the norms of living, in all their aspects, are the result of a will whose universality is that of the social and cultural life in which individuals enter into multiple confrontations and negotiations about what it is appropriate to will. This means that Hegel’s elaboration of the concept of a “subject” in the chapter on morality in the Philosophy of Right, which replaces the “person” of abstract right, contains an implicit or explicit criticism of Kant’s account of what a moral subject or moral self is. But it is also an account of what its truth is, namely what Kant truly grasped with this notion. Pace Kant, according to Hegel “abstract universality,” willing the law just because it is the law, belongs in the sphere of abstract right, not of morality. The will of morality, for Hegel, is already beyond such merely “formal” freedom. What it wills is not only freedom but also happiness. Morality, as a phase in the development of objective Spirit, is the expression, via the individual will of each particular subject, of the life of the universal subject, Spirit, which emerges out of the immediacy of life and generates a life mediated by re ection: the ethical life. But the moral subject mired in the forms of Kantian morality is not aware that it belongs in this process. Hegel o ers an explanation for this lack of awareness that is both historical and metaphysical: there is an actual history behind the emergence of the Kantian gure of morality, but the necessity of that history, or the unavoidability of this moment in history, has for Hegel a metaphysical explanation: Kantian morality in all its guises is a moment in the development of the Concept, emerging from life and returning via the travails of re ection and separation, to a life-world it has shaped: Spirit. So again, this is what becomes of the third tenet in Kant’s characterization of the moral subject, the empirically determined person of the Metaphysics of Morals. In Hegel’s Philosophy of Objective Spirit, Kant’s empirical person becomes the

Kant and Hegel on the Moral Self

115

moral subject that emerges from the person of abstract right and is not yet aware that it belongs in a world of particular needs, in which the universal will can nd its satisfaction by shaping that world according to the norms of ethical life. III.4 Hegel on Kant’s second tenet, the moral self as belonging to an intelligible world, not the phenomenal world, and as such as determined under the moral law, not causal determinism. When I presented Kant’s second tenet, I noted that its formulation in terms of “two standpoints” in section 3 of Groundwork (Grundlegung, 4:458) could be read as metaphysically de ationary. I added, however, that Kant did not mean it to be read in this way. Rather, he meant to defend the metaphysically robust thesis that as morally self-determining beings, we belong in an intelligible world in which the moral law just is the law of determination of what, in the world of appearances, appears as an event determined under universal causal laws. Now according to Hegel, Kant’s characterization of the “proper self” or moral subject is in fact the expression of a limited standpoint Spirit has on its own development. But the duality of standpoints or stances, and the fact that the agent’s stance, the stance of self-determination, is what moves the world, is a view Hegel shares with Kant. He gives it his own slant: the duality of standpoints Kant de nes in the text of Groundwork cited above is not a static duality. Causal necessity, rather than being “incompatible” with self-determination, is revealed as the way a self-re ecting, self-determining whole necessarily re ects itself at a stage of its development that must eventually be proved incomplete. But at bottom, Hegel shares with Kant the idea that goal-directed, normatively-structured agency is both the true nature of reality and what makes individual human beings “selves” – the referent of “I” in “I think” and “I ought to.” Contrary to Kant, however, Hegel claims that we do have knowledge of that true nature of reality: it is substance, in Spinoza’s sense, plus personhood. Which leads me to Kant’s rst tenet. III.5 Hegel on Kant’s First Tenet, the moral self as a person “in the psychological sense,” an entity conscious of its own numerical identity through time in referring to itself as “I,” and of the connection of its determinations in apperception. In the solution to the Third Paralogism, this concept of a person was “merely transcendental,” meaning that we could not claim any knowledge at all of the entity referred to by the term “I” in “I think.” In the Metaphysics of Morals, in contrast, the person under consideration is the empirical human being, who has the property of personhood, both psychological and moral. Now I said earlier that Hegel’s Concept is both universal and particular. Let me now cite the relevant texts in full. The Concept is First this pure unity relating to itself and this not immediately, but insofar as it abstracts from all determination and content and returns to the unlimited identity with itself. In this

116

Longuenesse

way it is universality. […] Second, I is also singularity, absolute determination, which distinguishes itself from some other entity [welches sich Anderem gegenüber stellt] and excludes it: individual personhood (WL, S. 6:253; 583).

Here’s one way to read the rst point. Kant’s pure “I” of “I think,” this “pure unity relating to itself” is just one way “I” is present in thought. Thought in this way, “I” refers to whoever is currently thinking “I” or referring to herself via the concept “I.” It thus holds true of, or is instantiated in, any human being capable of thinking in the rst person. Human beings acquire this capacity when entering into mutually opposing, exclusive, and thereby mutually conditioning relations, the system of which constitutes what Hegel calls Spirit. In Hegel’s complicated vocabulary, the identity that is thought by the most abstract (because mostly devoid of empirical determinations) concept “I” is “the identity of identity and nonidentity”: the identity (one-and-the-sameness) of the universal, self-referential activity of thinking “I” (Kant’s “pure I”) and the activity of thinking of oneself as singular, as one individual entity in a world of mutually excluding, mutually conditioning, self-determining entities: human beings belonging to a world of Spirit. In other words, for Hegel, the notion of identity on which the true understanding of the moral self depends is not the Kantian consciousness of one’s own identity through time, whether via the “pure” concept “I” or as consciousness of an empirical entity tracking its own agency through space and time. Rather, it is the consciousness of the identity of identity (the “pure” I) and non-identity (the empirical, historically determinate, individual persons of abstract right, subjects of morality and individuals of ethical life, mutually conditioning for one another in a system of Spirit).

4 Concluding Remarks Let me quickly take stock of the questions I raised at the beginning of this paper: how much metaphysics, and what kind of metaphysics, is needed to account for Kant’s moral self (the referent of “I” in the moral “I ought to”)? Why does Hegel claim, against Kant, to reestablish metaphysics to its past glory, and how much metaphysics does he need to account for Kant’s notion of a moral self? In Part II of this essay, I suggested that it is possible to give an account of Kant’s moral self with very little metaphysical baggage, by understanding Kant’s language of “two standpoints” as stressing the opposition and equal indispensability of the observer’s standpoint and the rational agent’s standpoint on one and the same temporal sequence, read as an event (to be causally explained) or as an action (to be morally evaluated). Note that even on this metaphysically economi-

Kant and Hegel on the Moral Self

117

cal reading, Kant’s language of “two standpoints” would still partition reality into (1) the kinds of entities for the understanding of which the standpoint of causal explanation is su cient and (2) the kinds of entities for which that standpoint has to be supplemented by the irreducibly distinct standpoint of moral justi cation and evaluation. The division of Kant’s metaphysics into a metaphysics of nature and a metaphysics of morals is the natural outcome of this partition, although metaphysics so understood is limited to partitioning reality as it appears from a human standpoint, and makes no claim to any knowledge of things as they are in themselves. Supposing we remained within that standpoint, Hegel’s historical account would o er the added resource of explaining the Kantian moral agent’s structure of motivation as an internalization of the structure of “abstract right,” a system of juridical laws with respect to which each individual has equal claim to being acknowledged as a freely contracting person. And the division of Hegel’s philosophical system into a philosophy of nature and a philosophy of Spirit could be read as a transformation of the very same partition of reality o ered by the Kantian system, with a historical slant: both the system of a priori concepts and principles Kant took to be necessary conditions for any metaphysics of nature and the system of a priori concepts structuring morality or the moral view of the world, and thus the notion of a moral self, would stand to be re-evaluated in light of Hegel’s history of Spirit understood as the collective history of humankind. Clearly, however, this minimalist reading is one neither Kant nor Hegel would accept. For Kant, the causal determination of events in human lives is the mere appearance of the free self-determination of pure intelligences belonging in a purely intelligible world of things as they are in themselves. For Hegel, the subject of morality as Kant understands it nds its place in a history of Spirit that has as its motivating force the Concept, a process of thinking emerging from the world of Life and producing reality according to self-set norms. In other words, Kant and Hegel share the conviction that the fundamental fact about reality, in light of which its ultimate nature is to be understood, is that it thinks and is self-determining. For Kant, this is a fact about our reality, so that we are for ourselves, qua moral beings, the only access to the intelligible world of things as they are in themselves. For Hegel, the fact that the ultimate nature of reality is such that it is thinking and self-determining, is a fact about reality as a whole. But if the analyses I have o ered in this essay are correct, neither Kant’s nor Hegel’s version of the strong metaphysical thesis concerning the ultimate nature of reality are necessary to preserve tenets i), ii) and iii) in Kant’s account of the moral self or to acknowledge the force of Hegel’s account of the historical emergence, value and demise of Kantian morality and the Kantian moral self.

Hannah Ginsborg

The Appearance of Spontaneity Kant on Judgment and Empirical Self-Knowledge This paper aims to draw a connection between two issues in the interpretation of Kant, both of which have gured signi cantly in the work of Rolf Horstmann. The rst issue concerns the status of the thinking subject as discussed by Kant in the Critique of Pure Reason, in particular the Transcendental Deduction and the Paralogisms. What, if anything, is the transcendental subject of apperception, the “I that thinks”? More speci cally, how are we to make sense of Kant’s contrast between the transcendental self-consciousness expressed by the formula “I think” and the empirical self-knowledge which allows us to cognize ourselves and our mental states as part of the spatio-temporal world? The second issue concerns the relation of the Critique of Judgment to the rest of the Kantian system, and in particular to the views of the rst Critique. The Critique of Judgment introduces what appears to be a new faculty with its own a priori principle, distinct from the faculties of understanding and reason, and it seems to do so in order to address a problem about cognition le unanswered in the rst Critique. But what exactly is that problem and how is the principle of judgment supposed to address it? I have learned a great deal from Horstmann’s discussions of both of these issues. But as will emerge in what follows, I disagree with him on a number of points, the most general of which bears on the relation between the two issues. To the best of my knowledge, Horstmann does not see the faculty of judgment in the third Critique as playing any systematic role in Kant’s philosophy with respect to the notion of the thinking subject in the Deduction and Paralogisms, and in his own work he treats the two issues independently. But the hypothesis motivating this paper is that there is a deep connection between the two. In particular, as I shall go on to suggest, a philosophically satisfactory interpretation of Kant’s views on the thinking subject in the rst Critique requires that we appeal to the resources of the third. The starting point for the interpretive proposal I am o ering is a claim with which Horstmann, at least o cially, disagrees: that the I that thinks – the referent of the “I” introduced at § 16 of the B deduction – must be identi ed with a particular human being in space and time, something which can be the object of empirical self-knowledge. Or to be a little more precise: that each of us, in entertaining the thought expressed by “I think” in the relevant sense, is referring to a particular human being, a human being of which we can acquire empirical

120

Ginsborg

knowledge through both outer and inner sense.¹ Even though the transcendental self-consciousness invoked in the Deduction and the Paralogisms is not the consciousness of oneself as a human being or under any other empirical description, we must still, on this assumption, understand the “self” of which one is conscious as, in fact, a human being. For example, when the human being Rolf Horstmann says the words “I think”, intending to express what Kant calls transcendental self-consciousness, we must understand the expression “I” as referring to Rolf Horstmann, even though Rolf himself is not conceiving himself under that description. Failure to accept this point, it seems to me, makes the I that thinks mysterious, or worse, leads us to think of it as some kind of disembodied universal mind outside of space and time, distinct from the particular minds belonging to particular spatio-temporal human beings. Each of us, in going through the reasoning of the Deduction and Paralogisms, has to understand the “I” of the “I think” as referring to him- or herself. But there is a di culty standing in the way of this identi cation. This is that my knowledge of myself as a human being is knowledge of myself as an object, as embedded in a spatio-temporal causal order governed by natural laws. This knowledge seems on the face of it to be incompatible with understanding myself (the human being) as a thinking subject, endowed with the spontaneity characteristic of the I that thinks. It is in addressing this di culty, I shall go on to argue, that the Critique of Judgment contributes to Kant’s account of self-knowledge. If we understand human beings as endowed with a faculty of judgment – which, I shall argue, is compatible with understanding ourselves also as part of empirically determined nature – then we can understand how the “I” of the “I think” can, for each of us, refer to a particular human being. For judgment, as I shall argue, is, in a sense, the appearance of spontaneity. The spontaneous subject can, in Kant’s words, “ nd itself in nature”² because nature includes human beings who are ca-

1 Note that I want to remain neutral on whether or not Kant is an empirical dualist, leaving open that a human being might be composed of an object of outer sense (a human body) in some kind of association with an object of inner sense (a human mind). 2 Critique of Judgment, Introduction VIII, 5:193. (Except in the case of the Critique of Pure Reason, which I cite according to the usual A and B page numbering, references to Kant’s works cite Kant’s Gesammelte Schri en, edited by the Prussian (now German) Academy of Sciences (1900–). All translations, including quotations from the German-language articles I go on to cite, are my own.) My use of Kant’s words in this context might seem like a stretch, since, on the face of it, the passage cited is concerned with the possibility of orienting ourselves in nature’s overwhelming diversity by bringing its objects under a system of empirical concepts. But as I will indicate in section IV, I take the ascription of judgment to human beings to be equivalent to the claim that human beings are capable of conceptualizing nature, or, equivalently, that nature is conceptualizable by human beings. So if the argument of this paper is correct, there is at least some philosophical warrant for

The Appearance of Spontaneity

121

pable of judgment, and who can thus manifest, within the world of appearances in space and time, the spontaneity in virtue of which cognition of nature is possible for us. The paper is organized in six sections. Section I o ers a brief explanation and defence of the claim that the I that thinks must be identi ed with a human being. Section II considers how Kant’s view of cognition might allow for this identi cation, drawing on a suggestive line of thought o ered by Béatrice Longuenesse. Sections III and IV raise an objection to Longuenesse’s view, illustrating the more general di culty I see with the identi cation. Section V aims to show how appeal to the faculty of judgment addresses the di culty, and section VI considers an objection to my approach.

1 In his critique of rational psychology in the Paralogisms, Kant argues that the mere representation “I think” does not a ord any cognition of what it is that thinks. This representation does a ord consciousness of a certain unity among the representations which it accompanies, and it allows me to think of these representations as belonging to a single subject. But because the representation contains no intuition it does not constitute cognition of an object, nor, a fortiori, of myself as an object. Descartes was wrong, then, to suppose that I have a priori cognition of myself as a thinking substance in which my representations inhere. Even though I can think of myself as numerically identical from one representation to another, this identity pertains to me only as a subject of thought. That I can think the manifold of representations as all belonging to the same thinking subject does not allow me to infer the existence of a single enduring substance in which those representations inhere. Does this leave any room for an answer to the question “What is the I that thinks”? That is, even if this representation itself a ords nothing which could qualify as knowledge of a particular self, could it still be that the expression “I” picks out or refers to some entity, perhaps an item which could be cognized otherwise than through the mere consciousness of thinking? In an article from 1993, Rolf Horstmann argues that the answers depend on whether we appeal to the rst or to the second edition of the Critique. In the rst edition, according to Horstmann, the representation of the I, for all its emptiness, is still referred to what

interpreting Kant’s talk of the understanding’s “ nding itself” in nature in the terms suggested here.

122

Ginsborg

Kant calls a “substrate”.³ The point of the Paralogisms, as far as the rst edition goes, is not to deny that the representation is related to an object (gegenstandsbezogen) but simply to argue that this object is not knowable for us. In the B edition Paralogisms, by contrast, Kant o ers an understanding of the I as simply an activity or action (“dass das Ich als etwas gedacht werden muss, das als Aktivität, als Handlung zu beschreiben ist”). Here Kant’s account, according to Horstmann, is of a piece with his characterization of the representation “I think” in the B deduction as an “act of spontaneity” (§ 16, B 132). On this account it is a mistake to suppose that there is any entity picked out by the I of apperception: to represent the I is simply to represent the spontaneous activity of thought. Horstmann’s view of the Paralogisms has been challenged, in di erent ways, by Tobias Rosefeldt and by Michael Wol . For Rosefeldt, we must understand the “I” in both editions as picking out an object, but where this object is, in Kant’s terms, a kind of “nonreal” object, a Gedankending, something of which we lack intuition but which still quali es as an object.⁴ Wol is less clear about whether the “I” of the “I think” unambiguously refers to an object, but in contrast to both Horstmann and Rosefeldt, he holds that if it refers to an object, then that object is nothing other than a human being in space and time. “[If] the little word ‘I’ in the sentence ‘I think’ […] can be referred at all to an intuitable object (distinct from my state of thinking), this object can be nothing other for Kant than me as this human being here, which I can intuit both through my inner and my outer sense”.⁵ Following Wol ’s suggestion, although perhaps taking it further than he intended, I want to claim that the “I” should in fact be understood, at least if we are aiming for a sympathetic reading of Kant, as referring to a human being. I agree with Horstmann that the consciousness expressed by the “I think” is in the rst instance consciousness of the activity of thinking and that, through this consciousness alone, we learn neither that there is anything which performs this activity or, if there is anything, what it is. But this does not rule out that the activity of which we are conscious is performed by, or in some more general sense is attributable to, an entity which is cognizable through other means, in particular through empirical means: namely by a human being of whom we can come to be aware both through inner and outer sense. And if it does indeed belong to a human being, then there is a sense in which the little word “I” refers to that hu-

3 Rolf-Peter Horstmann, “Kants Paralogismen,” Kant-Studien 83 (1993): 408–425, p. 418. 4 Tobias Rosefeldt, “Kants Ich als Gegenstand,” Deutsche Zeitschri für Philosophie 54 (2006): 277–293. 5 Michael Wol , “Empirischer und transzendentaler Dualismus,” Deutsche Zeitschri für Philosophie 54 (2006): 265–275, p. 268.

The Appearance of Spontaneity

123

man being, even though there is, so to speak, no descriptive content associated with that word which would allow me, through entertaining that content alone, to represent myself as a human being. In other words, the mere fact that I cannot claim a priori that the “I who thinks” is a human being does not mean that it is not a human being: as Henry Allison puts it, “one can always nd a perfectly good empirical answer to the question [“what is the I that thinks?”]; for example, Henry Allison”.⁶ Moreover, if we do not make this identi cation, it is hard to see how Kant’s account of the conditions of cognition in the rst Critique can have any bearing on cognition that is “human” in the ordinary sense, namely such that we can ascribe it to individual members of the human species. This might explain why many commentators in fact tacitly make this identi cation, for example Horstmann himself in a recent attempt to explain the activity of thinking as it gures in Kant’s account of the unity of apperception. While walking absentmindedly to the station I have […] visual, acoustical, olfactory and tactile impressions […] I am aware of all of them without noticing any one of them in particular […] Suddenly I realize that on the le side of the street in which I am walking, there is a house with a blue door […] what has happened is that in focusing for reasons normally beyond my grasp on a particular collection of my impressions, I end up being in a propositionally structured state (i.e. that this house has a blue door).⁷

This state, Horstmann goes on to explain, is “the outcome of an activity of mine, of my bringing together into a speci c relation certain elements I have not been aware of in a determined manner until then”.⁸ In this explanation “I” clearly refers to Rolf Horstmann, since it is unquestionably Rolf Horstmann who walks to the station, who has visual impressions, who realises that there is a house with a blue door, and who ends up in a propositionally structured state. But since there is no indication of a change in referent, the passage suggests that it is also Rolf Horstmann who performs the activity responsible for this propositionally structured state, that is who brings the elements of the sensory manifold into the unity of apperception. So it seems that Horstmann himself cannot accept without quali cation his own o cial claim, in the same article, that “the self-conscious I is not a self-standing entity”:⁹ there is a perfectly straightforward sense in which Rolf Horstmann is the entity Rolf Horstmann refers to with “I” when he entertains the thought “I think.”

6 Henry Allison, Kant’s Transcendental Idealism (Yale: Yale University Press, 1983), p. 290. 7 Rolf-Peter Horstmann, “The Limited Signi cance of Self-Consciousness,” Revue de Metaphysique et de Morale 68 (2010): 434–454, pp. 446–447. 8 ibid., p. 447. 9 ibid., p. 453.

124

Ginsborg

2 I have indicated the desirability of identifying the I that thinks or the I of apperception with a human being, and I have also cited some commentators who seem to endorse – explicitly or implicitly – that identi cation. Now I want to ask whether this identi cation is consistent with the overall view of cognition o ered by Kant in the rst Critique. It might seem that it is simply ruled out by Kant’s transcendental idealism, especially when that idealism is understood along strongly phenomenalist lines. For example, on a common – if not always clearly articulated – reading of the Transcendental Aesthetic and the Transcendental Deduction, the “I” of apperception expresses the standpoint of a being in some sense “outside of” space and time which is responsible for the constitution of spatio-temporal objects – appearances, or phenomena – through an activity of unifying representations given from a source which is also outside of space and time. Both the processes through which the representations are given and the activity of synthesis through which they are uni ed must, on this interpretation be understood as non-spatial and non-temporal and in a sense prior to space and time. If this is how we understand Kant, then the human being Rolf Horstmann with its representations of a blue door is, along with the blue door and all other objects in space and time, something constructed by the I of apperception, albeit out of materials possibly given from elsewhere. So we cannot possibly identify the I of apperception with, say, Rolf himself. However, some commentators seem to adopt a less phenomenalist reading of Kant’s idealism, on which the activity of the I does not literally constitute spatio-temporal objects out of representations given to it, but rather confers objectivity on those representations, bringing it about that they are intentionally directed towards an objective spatio-temporal world. The activity of the I on such a reading serves to generate, not the spatio-temporal world as such, but a representation which has such a world as its intentional object. A reading of this kind is suggested by Béatrice Longuenesse, for whom spatio-temporal objects are made possible by the activity of the I in the sense that they could not be “represented objects” without that activity, although the I is not responsible for their existence, which depends on the in-itself. They are thus in a sense “internal” to our representations – as she puts it, albeit tentatively, “mental intentional correlates of our representational capacities and activities”¹⁰ – although there re-

10 Béatrice Longuenesse, Kant and the Capacity to Judge (Princeton: Princeton University Press, 1998).

The Appearance of Spontaneity

125

mains a sense in which they exist independently of those activities.¹¹ This seems to leave room for the possibility that the I might identify itself with one of those spatio-temporal objects, speci cally one of the human beings in the world whose representation it generates. The I would thus correspond to the standpoint of a human being whose existence does not, per impossibile, depend on its own activity of thinking but which can, however, be represented to itself as a spatio-temporal object only in virtue of this activity. That activity, in other words, would make possible the representation of a world of objects, including human beings, one of which the I can recognize as itself, that is, as the very I whose activity makes possible the representation of a world of objects in the rst place. Longuenesse suggests something like this conception of self-knowledge in her discussion of the Analogies of Experience. On her reading, the Third Analogy in particular shows how, by drawing on the forms of judgment under the heading of relation, we – and here the “we” is in the rst instance the plural of the “I” of apperception – generate the representation of a uni ed space and time in which all empirical objects may be cognized through their relations of interaction with one another. Kant’s argument has, she says, a “particularly striking aspect”: namely the claim that, in cognizing these causal interactions, “we also situate ourselves, as empirical unities of consciousness associated to a body we represent as our own, in the uni ed empirical space and time whose representation we thereby generate”.¹² This is because, as she puts it in her discussion of the Anticipations of Perception, “our awareness of the universal ‘community or interaction of the objects we perceive in space […] includes an awareness of their interaction with our own body, and thus of their causal determination of the ‘matter of our perceptions, [namely] sensations”.¹³ We synthesize a given sensory manifold so that it comes to represent – to be intentionally directed towards – a spatio-temporal world of interacting substances. And among the substances we thus represent are our own bodies, which we recognize both as causally interacting with other bodies, and as causally related to our mental states, in particular our sensations, and, more speci cally, the sensations which serve as the material of our synthesis. So, for Longuenesse, the very sensations from which we generate the representation of an objective world, and which thus come to be intentionally directed towards objects

11 See especially the section on the “internalization within representation of the relation between representation and its object” at pp. 20–26 of her Kant and the Capacity to Judge. I have some doubts, independent of the worry I go on to raise for her account of self-knowledge, about the coherence of the account Longuenesse presents in that section. But I leave these aside for the purposes of the present paper. 12 ibid., p. 378. 13 ibid., p. 322.

126

Ginsborg

in that world, are also cognized by us as caused by those objects. She sums up the view towards the end of her discussion of the Third Analogy: The ‘objects of sensation,’ that is, to which sensations are intentionally related are the substrates of changing determinations [that is, substances] in relations of community with all other substances in space. We cognize these objects only through their rule-governed relations, among which are their relations to our own body. The objects are the ‘causes of the sensations we cognitively relate to them in experience, insofar as between the objects and the sensations we empirically cognize the temporal relation that is the schema of causality: for instance, the regularly repeated succession between my body’s carrying another body and my feeling of weight […]. The subjective manifold of our empirical representations is itself temporally determined in causal relation to the universal temporal order of the objects of experience […].¹⁴

As she puts it in conclusion: the astonishing edi ce of Kant’s Analogies of Experience comes to completion […] by the location of ‘us’ in the empirically given world […] an ‘us’ […] consisting of unities of empirical consciousness associated with a phenomenal body of our own, unities of consciousness both passive (receptive, capable of conscious sensation and associative imagination) and active (spontaneous, intellectual, capable of judgment and synthesis speciosa). And, as such, the authors of the representation of the very world in which ‘we’ locate ‘ourselves’: transcendental subjects.¹⁵

Longuenesse does not speak explicitly here of the identi cation of the I of apperception with a particular human being in space and time. But I take this identi cation to be implicit in her claim that, through its spontaneous activity the I locates itself as a unity of empirical consciousness associated with a phenomenal body.¹⁶ So I take it to be a consequence of her reading of the Analogies that the word “I” in the expression “I think” refers to, or picks out, a human being which can thus be understood both as a transcendental subject, responsible for generating the representation of the spatio-temporal world, and itself an object located in that world.¹⁷ And as I understand her view, this is possible because the sensations

14 ibid., p. 393. 15 ibid. 16 See note 1. 17 This identi cation is also implicit in the example of seeing a tower in Béatrice Longuenesse, “Kant’s ‘I think’ versus Descartes’ ‘I am a thing that thinks’” (in D. Garber and B. Longuenesse (eds.) Kant and the Early Moderns [Princeton: Princeton University Press, 2008], 9–31), p. 15, and the examples of seeing a tree and carrying out a mathematical proof in Béatrice Longuenesse, “Two Uses of ‘I’ as Subject?” (in S. Prosser and F. Recanati (eds.) Immunity to Error through Misidenti cation: New Essays, [Cambridge: Cambridge University Press, 2012] 81–103), pp. 89–93.

The Appearance of Spontaneity

127

from which the transcendental subject generates its representation of the world, and which thus come to have the spatio-temporal world as their intentional object, are also themselves part of that spatio-temporal world.¹⁸ The transcendental subject receives sensations, including, for example, sensations of weight, which it synthesizes to form the representation of a spatio-temporal world containing such things as bodies endowed with properties such as weight. But the spatio-temporal world thus represented includes a human body, which sometimes li s other bodies, and a temporal series of sensations and other mental states which are correlated with changes in that human body. For example, it includes sensations of weight whose occurrence is correlated with occurrences of the human body li ing another body. Because these sensations are the very same sensations out of which the transcendental subject synthesizes its representation of the spatio-temporal world, the transcendental subject can identify itself with the human being – the object comprising the human body and its correlated states of mind – thus locating itself in the world whose representation is made possible by its own activity of synthesis. I shall try to elaborate Longuenesse’s suggestion in terms of an example. How do I – in the sense of “I” corresponding to the apperceptive “I think” – identify myself with the particular human being now forming the ink marks which are responsible for the words you are now reading? In order not to beg the question, we need to think of “I” here as referring in the rst instance only to the transcendental subject, leaving open whether it also picks out the (or for that matter any) human being. So while we agree that there is a spatio-temporal human being – call her HG – writing with a pen and sometimes uttering words like “I think”, and a transcendental self-consciousness aware of itself as generating the representation of a world which includes pens, human beings and utterances of “I think”, we are leaving unsettled whether the “I” of the “I think” which expresses the transcendental self-consciousness refers to anything spatio-temporal, and in particular whether it refers to HG. From the point of view of the transcendental subject – that is, from “my” point of view, on the usage being employed in the example – HG is just another object of cognition, along with the pen, the table at which HG is writing, a cup on the table, and other human beings sitting at tables in HG’s vicinity. I cognize HG through synthesizing a manifold of sensations, for example

In each of these examples Longuenesse intends us to understand the pronoun “I” both as the “I” of the Kantian “I think” and as referring to a speci c human being in a determinate empirical context. 18 She does not say this explicitly, but I think it is implicit in her discussion, at (1998) Capacity to Judge, 322, of the “shi in point of view” regarding the relation of sensation to the real in appearance.

128

Ginsborg

visual sensations of the colour and shape of a hand, an arm, and part of a torso; in synthesizing these sensations I represent the hand, arm and torso as integrated into a larger scene which incorporates the pen which the hand is holding, the tables, and the other human beings. And it is part of the upshot of this synthesis that I cognize correlations between observed changes in HG’s body and other observed changes: for example, between the movement of HG’s hand when it is in contact with the pen and the movement of the pen. But HG plays a special role in my experience, because I also cognize correlations between changes in HG, observed through outer sense, and other occurrences of which I am aware only through inner sense: correlations which distinguish HG from other objects of my experience, including other human beings. For example I cognize that HG’s li ing a cup is correlated with the occurrence of a sensation of weight, whereas I nd no such correlation between sensations of weight (at least as perceived through inner sense) and cups’ being li ed by other human beings. More generally, it is only in the case of HG, as opposed to other objects observed through outer sense, that I come to discover predictable regularities in the occurrence of sensations and other mental states relative to changes in the observed object. The close association between the occurrence of mental states, observed through inner sense, and changes in HG, observed through outer sense, allows me to label these mental states as belonging speci cally to HG as opposed to any other object of my outer experience, and thus to regard HG as an object of both inner and outer sense (whereas other human beings, the pen, the cup and so on are objects only of outer sense).¹⁹ I can regard my outer perception of this hand, and my inner perception of a feeling of weight, as relating to one and the same object: a “unity of empirical consciousness associated with a phenomenal body,” as Longuenesse puts it, or, as it might also be put, a human being. But – and this is the crucial element in the present discussion – I can also, in Longue-

19 This explanation of why HG (in the example) plays a special role in the experience of the transcendental subject represents my own attempt to esh out Longuenesse’s account, since Longuenesse herself does not o er an explanation of why it is HG, rather than some other human being, which the I identi es with itself. I suspect that the explanation I have o ered on her behalf is inadequate, because I doubt that we can make sense of the contrast between inner and outer sense without smuggling in the assumption that the “I” of the “I think” already refers to a human being. Without this contrast, however, it is hard to see why HG’s mental states have a privileged role over the mental states of the other human beings which I represent as a result of my synthesis of the given manifold. For the intentional objects of my experience are not restricted to objects of immediate perception but include, for example, mental states whose occurrence can be inferred from my perception of other people’s bodies. This might constitute a reason, independent of the di culty I go on to raise in sections III and IV, to question Longuenesse’s account, but I will not pursue it further here.

The Appearance of Spontaneity

129

nesse’s terms, “locate” or “situate” myself in the empirical world by identifying HG’s body as my own, that is, as belonging to me qua subject of apperception and not just qua object of inner sense. For I can recognize, on philosophical re ection, that the very sensations of which inner sense makes me aware, and which I can consequently label as HG’s sensations, are the material out of which I generate, through my synthetic activity, my representation of the empirical spatio-temporal world. The sensation of weight caused in HG by her hand li ing the cup is identical with the sensation of weight which provides the material for my objective experience of the body as heavy. And this allows me – the “me that thinks” – to identify myself with the HG who is an object of the experience made possible through my activity of thinking.

3 I have elaborated Longuenesse’s view because it represents an unusually explicit attempt to confront the question how a transcendental subject, conceived initially as independent of the spatio-temporal world, can be “located” or “situated” within that world. In this section I will raise what I think is a di culty for Longuenesse’s answer. Let us suppose that Longuenesse’s view is correct at least up to this point: the transcendental subject synthesizes, out of a given manifold of sensation, and in accordance with the categories, the representation of a spatiotemporal world containing human beings with mental states caused by the impingement of other bodies on their bodies. So, to go back to our example, let us suppose that I – here, again, the I that thinks, conceived initially in a way which leaves open the question of its identi cation with an object in the spatio-temporal world – come, through the synthesis of a given manifold of representation, to have objective cognition of a human being HG, of a cup in HG’s hand, and of a feeling of weight which belongs to HG and is caused by HG’s li ing the cup. Is Longuenesse right to claim that I – the I that thinks – am in a position to identify myself with HG? It does indeed seems plausible that if I am in a position to identify myself with any object of my experience, that object can only be HG. For, as noted, HG plays a special role in my experience: it is only the mental states associated with HG’s body, and not those associated with the body of any other human being, that I cognize directly through inner sense. However the question is not whether I can identify myself with HG as opposed to some other object of my experience, but whether I can identify myself with HG as opposed to being unable to locate myself in the empirical world at all. And there seems to be an obstacle in the way of that identi cation: namely that there is a di erence between my own relation to the objects which I cognize, and the relation which I represent HG as having to

130

Ginsborg

those same objects. My own relation to the cup is a cognitive one: the cup is the intentional object of my experience, and, in experiencing it, I ascribe properties to it: for example, I cognize it to be heavy. By contrast, the relation I represent HG as having to that same cup is not, at least on the face of it, cognitive, but rather merely causal. Through my synthesis of the sensory manifold in accordance with the categories I come to represent the cup as standing in a causal relation to HG, in particular as causing her to have a sensation of weight. But that does not amount to my representing the cup as an intentional object of HG’s mental states, and in particular, to my representing her as cognizing the cup to be heavy. It would seem, then, that I cannot identify myself with HG, or more generally “locate” or “situate” myself at HG’s position in the spatio-temporal world, since HG’s point of view on the cup is di erent from mine. Better put, perhaps, she has no point of view on the cup at all. The cup indeed causes sensations in her, but it does not present itself to her, the way it presents itself to me, as a determinate object endowed with speci c properties: it is not “in her view” the way it is in mine. It might be objected that this underestimates my resources for representing HG’s relation to the cup. For one thing, it might be pointed out, I recognize that the cup’s causing the feeling of weight in HG is not an isolated occurrence. Rather, I recognize it as happening in accordance with an empirical natural law on which one kind of event – a body li ing another body – is correlated with another kind of event – a sensation belonging to the empirical unity of consciousness associated with the rst body. So I represent HG’s sensation at the very least as indicating the presence of weight in the cup, and this might be thought to be su cient for my taking HG to represent the cup as heavy. For another thing, I am not restricted to ascribing sensations to HG, but can also ascribe to her an imaginative activity through which she organizes and processes those sensations, to arrive at more complex mental states which might more plausibly be thought of as intentionally directed towards objects like the cup. I can represent HG, that is, as carrying out a kind of cognitive processing, and this might again be thought to support the idea that HG is not merely passively a ected by the cup, but also actively synthesizes her representations to arrive at cognition of it. But these resources are insu cient to allow for the representation of HG as cognizing the cup, at least given Kant’s understanding of cognition. A quick way to see this is to note that the same resources could be used to represent a non-human animal as cognizing the cup, since the sensations of animals are no less nomologically correlated with the corresponding properties than in the case of humans, and animals no less than humans can be understood as engaging in the kind of imaginative processing through which sensations can be incorporated into more complex representational states. Since, for Kant animals are not capable of cognition in the relevant sense, whatever psychological activity I ascribe to HG cannot be su cient to establish that her relation

The Appearance of Spontaneity

131

to the cup is cognitive. The fact that the state of mind caused in HG is nomologically correlated with her li ing heavy things, and that it might be the outcome of a psychologically complex operation, does not take away from the fundamental di culty, that the relation between HG’s state of mind and the cup is conceived of as governed by causal laws, and hence as falling short of the spontaneity which Kant takes to be required for cognition. To state the point brie y in Kantian terms: I – the I of the “I think” – conceive HG as passive with respect to the cup, as capable only of combining the sensations it causes in her according to empirical laws of association. But if I am to identify myself with HG, I must be able to conceive HG as also engaged in some kind of genuine activity with respect to the cup: not only responding to it but also making judgments about it. And there is nothing so far in my resources for thinking about HG which allows me to do this. The di culty identi ed here is in the territory of familiar, and more general worries about what Quassim Cassam describes as the “elusiveness” of the self: in Kant’s case the worry that, as Horstmann puts it, “the self-conscious I can […] never be an item of which I can be conscious as an object”.²⁰ Dina Emundts raises a version of this di culty when she says that what we perceive as spontaneous, we cannot see as object-like: “if we wanted to determine the I as being present [daseiend] […] the I would be determined as appearance. But then it is represented not as determining [das Bestimmende] but as determined [das Bestimmte]”.²¹ The

20 Horstmann “Limited Signi cance”, p. 449. See for example Quassim Cassam, Self and World (Oxford: Oxford University Press, 1997), Chapter 1. My understanding of the issues concerning self-knowledge for Kant is very much indebted to Cassam’s illuminating treatment in this book. However, whereas Cassam sees Kant as committed to the elusiveness of the self and the related “exclusion thesis” according to which the self is not an item in the phenomenal world (ibid., p. 10), this paper starts from the assumption that a sympathetic reading of Kant must nd a way to avoid committing Kant to the exclusion thesis. The aim of this paper is to show how the Kantian thinking subject can, pace Cassam’s reading of Kant, “make itself the object of its own thought and intuition” (ibid., p. 22). Indeed, I see Kant in the Critique of Judgment as holding something like the position which Cassam identi es with the “robust response” to the exclusion thesis, namely that subjects of thought and experience are human beings, and as such can certainly think of themselves as objects among others. It should be noted, though, that whereas Cassam’s own development of the robust response appeals in part to a conception of awareness of one’s self qua subject in terms of the notion of immunity to error through misidenti cation, I do not think that this does justice to Kant’s own notion of the subject’s self-awareness, which requires that one be aware of oneself as a thinker and judger, and not just (for example) as passively undergoing experiences. Here I am in partial agreement with Béatrice Longuenesse, “Self-Consciousness and Consciousness of One’s Own Body: Variations on a Kantian Theme,” Philosophical Topics 34, (2006): 283–309, and Longuenesse “Two Uses”. 21 Dina Emundts, “Die Paralogismen und die Widerlegung des Idealismus in Kants ‘Kritik der reinen Vernun ’,” Deutsche Zeitschri für Philosophie 54 (2006): 295–309, p. 306.

132

Ginsborg

version we have arrived at emerges as a consequence of our attempting to identify the I with a particular appearance, namely with a human being. If we think the candidate human being – HG in my example – solely as an element in causally determined nature, then we would seem to be debarred from thinking it as spontaneous, and hence from being identical with the self-conscious I.²² A related version of the di culty is raised by John McDowell when he asks how the referent of the ‘I’ in the ‘I think’ can also be “a third person, something whose career is a substantial continuity in the objective world”,²³ and more speci cally how something which “starts out conceiving itself as a merely formal referent for ‘I’” could come to “identify itself with a particular living thing”.²⁴ McDowell describes this identi cation as a matter of the subject’s “appropriating a body” in that the subject might “register a special role played by a particular body in determining the course of its experience”,²⁵ and this sounds like Longuenesse’s answer, at least on the sketch I have given so far. But, McDowell says, we can only “pretend to make sense of” the idea that the subject could, in this way, identify itself with a living thing, for the idea “would not provide for it to conceive itself, the subject of its experience, as a bodily element in objective reality – as a bodily presence in the world”.²⁶ For McDowell, the di culty is insoluble. Kant is unable to “accommodate the fact that a thinking and intending subject is a living animal” because of “his rm conviction that conceptual powers are non-natural, in a sense which equates nature with the realm of law”.²⁷ According to McDowell it is because Kant does not recognize what McDowell calls “second nature,” a responsiveness to reasons arrived at as a result of inculcation into human practices of language-use, but which still quali es as natural, that he is blocked from identifying the standpoint of the I with that of a living organism.

22 Cf. also Allison’s suggestion that the act of thinking is incapable of grasping itself as object because “the conceptual activity through which the mind represents an object, including itself as object, cannot itself be given to it as an object. Insofar as one objecti es thinking, that is, treats it merely as a psychological occurrence, one eo ipso destroys its character as thinking” (Allison Transcendental Idealism, p. 278). 23 John McDowell, Mind and World (Cambridge, Mass.: Harvard University Press, 1994), p. 102. 24 ibid., p. 102–103. 25 ibid., p. 103 26 ibid. 27 ibid., p. 104.

The Appearance of Spontaneity

133

4 A simple way of putting the problem I have identi ed, one suggested by McDowell’s formulation of it, is just to say that the “I” of the “I think” cannot pick out a human being because a human being is part of the natural causal order and the I is spontaneous. But once it is put in this way, the problem might seem to admit of an equally simple solution. Can’t it just be insisted that being spontaneous in the sense of the rst Critique is compatible with being part of the natural causal order, so that HG is not only passively a ected by the cup, but also capable of a spontaneous activity which allows her to cognize it? And indeed perhaps this is just what Longuenesse has in mind when she describes the “us” that we locate in the empirical world as “unities of consciousness both passive (receptive, capable of conscious sensation and associative imagination) and active (spontaneous, intellectual, capable of judgment and synthesis speciosa)”.²⁸ I – the I of the “I think” – can identify myself with HG, on this suggestion, because I conceive HG as endowed not just with sensibility and the related capacity of empirical imagination, but also with understanding and the related capacity of transcendental imagination. Admittedly, I have no empirical warrant for ascribing understanding to HG: I as a thinker do not have empirical cognition of HG’s spontaneity, since I as a thinker am not presented to myself as an object. But I do have empirical warrant for judging that if any one of the intentional objects of my cognition is identical with me, that object can only be HG. HG’s special role as an object of my inner as well as my outer experience makes her the only thing in the spatiotemporal world which could be identi ed with me, or more generally, understood as occupying my own standpoint on the world. And if it can be argued that a full account of the possibility of cognition requires that I be capable of locating myself in the spatio-temporal world, then that would constitute a transcendental justi cation for ascribing understanding to HG. Once it is allowed, then, that the spatio-temporal character of an object does not debar it from being capable of spontaneous thought, then there is nothing to stop me representing HG as having just the same capacity that I do. I can represent her as identical with me simply by, so to speak, projecting my own spontaneity on to her: by ascribing to her my own spontaneous capacity for synthesizing a given manifold of sensible intuition so as to generate a representation of a world of objects in space and time. Now, I do think that this suggestion is on the right track, insofar as it recognizes that HG’s spatio-temporal character need not rule out the ascription to her of spontaneity. However, in its present form it is inadequate. As it stands,

28 Longuenesse Kant and the Capacity to Judge, p. 392.

134

Ginsborg

the suggestion is that I – the I that thinks – simply ascribe to HG the same spontaneous understanding which (on transcendental re ection) I ascribe to myself. That is to say, I ascribe to HG a capacity not only of being a ected by things so as to undergo sensations, but also of spontaneously synthesizing those sensations, in accordance with the categories and forms of judgment, so that they come to form a uni ed representation of a spatio-temporal world of causally interacting substances. I conceive of her as coming to cognize objects, such as the cup she is li ing, by generating, out of the material provided by the cup’s a ecting her senses, a representation with that cup as its intentional object. And I conceive of the activity through which the representation is generated as amounting to, or resulting in, her cognition of the cup as having determinate properties. For example, I represent her as combining the sensation of weight caused by her li ing the cup with her other representations in a way which is governed by the categories: she relates her sensation of weight to her other representations in what Kant calls an objective unity, something which can be expressed by her saying not just “if I li a body, I feel an impression of weight” but “it, the body, is heavy” (Critique of Pure Reason § 19, B 142). At rst sight this might seem to solve the problem. HG, as I represent her, is not merely responding causally to the cup but cognizing it: she is judging it to be heavy. And if that is what HG is doing, then I can identify myself with her, for she, like me, stands in a cognitive relation to the cup. But the solution is illusory. The problem is that HG on this conception, does not in fact cognize the cup in the example: more speci cally, she doesn’t cognize the very same cup which I cognize. The cup which I cognize plays, for HG, the role that the unknowable thing-in-itself plays for me: it is not the intentional object of her representations, but rather their unknowable ground. HG does indeed cognize a cup: she synthesizes her representations, including her sensation of weight, so that they come collectively to stand in an objective unity, which she can express by saying (among other things) “the cup is heavy.” But I cannot represent her as referring to the very same cup which I cognize, that is to say, to the cup which I represent as a ecting her senses so as to produce the impressions which she synthesizes. That cup is no more the intentional object of her experience than the unknowable thing in itself is the intentional object of my experience. So while my ascription of a spontaneous synthesizing understanding to HG allows me to represent her as a judging and cognizing subject rather than a mere animal, it does not allow me to represent her as the same judging subject which I am. In order to do that, I need to represent her not just as cognizing objects as a result of having her senses a ected, but as cognizing the very objects which a ect her senses, which is to say, the very same objects which I cognize and which are the intentional objects of my experience. It is not enough for her to respond to the cup by making a judgment with a content expressible by “the cup is heavy”: she must judge, of

The Appearance of Spontaneity

135

the very cup which causes her impression of weight, that it is heavy. Otherwise I cannot conceive her as cognizing the same spatio-temporal world which I cognize, namely the spatio-temporal world in which she is located.

5 The argument so far is that I – the I that thinks, as conceived, roughly, on Longuenesse’s model – can identify myself with a particular human being in the world only by conceiving that human being as endowed with spontaneity. But I cannot do that by what we might call “doubling the understanding,”²⁹ that is, by regarding the candidate human being as standing in the same spontaneous relation to her sensory manifold as I take myself, in transcendental re ection, to stand to my own. What I need to ascribe to HG is not my own spontaneous activity, conceived of as transcendental, but rather its phenomenal correlate, something which belongs to HG qua spatio-temporal appearance but which at the same time allows her to have cognition of, and not merely to be a ected by, other spatio-temporal appearances in her environment. In short, what HG needs, in order to be me, is the appearance of spontaneity. The main proposal of this paper is that this role is lled by the faculty of judgment as characterized by Kant in the third Critique. The faculty of judgment can be regarded as the empirical or phenomenal correlate of the spontaneous activity of the I as invoked in the Critique of Pure Reason. It is – to speak metaphorically – what this activity “looks like” when viewed from the empirical perspective, in which the I is conceived of not as a transcendental subject in some sense prior to, or outside of, space and time, but rather as itself located in space and time and, moreover, subject to empirical causal laws. Alternatively – and perhaps preferably – to talk about the spontaneous activity of the I is simply to talk about human beings’ exercise of judgment, but in a way which abstracts from our empirically determined, and more generally, our spatio-temporal character. On this way of thinking, the spontaneity of the I just is the faculty of judgment, viewed from the transcendental rather than the empirical perspective. Either way, though, appeal to the faculty of judgment makes possible the identication of the I of apperception with a particular human being. In terms of our example, I can identify myself with HG by regarding her not merely as responding psychologically to the cup which a ects her senses, but as doing so in a way which involves the exercise of judgment.

29 I owe this helpful formulation to Paul Guyer.

136

Ginsborg

This proposal depends on a controversial understanding of the faculty of judgment, but I shall introduce it by making some relatively uncontroversial points. Kant de nes the faculty of judgment as the capacity for “thinking the particular as contained under the universal” (Critique of Judgment, Introduction IV, 5:179), and distinguishes two ways in which it can be exercised, namely as determining and as re ective. Re ective judgment, which is exercised where “merely the particular is given, for which judgment is to nd the universal” (ibid.) is the more fundamental in the context of the Critique of Judgment, for it is only in its re ective aspect that judgment shows itself as an autonomous faculty, distinct from understanding. So, in what follows, my references to “judgment” will be to the faculty of judgment in its re ective aspect. Kant describes re ective judgment as making various contributions to empirical cognition, in particular the systematization of particular empirical cognitions into overarching scienti c theories,³⁰ but also, at a more fundamental level, the conceptualization of particular objects: for example, conceptualizing a stone as an instance of the general kind granite.³¹ And he also understands it as the faculty responsible for judging things to be beautiful, where this in turn is a matter of responding to them with a feeling of pleasure and, in so doing, taking it that all other perceivers of the object ought to respond the same way, or, equivalently, that one’s response is appropriate to, or normatively called for by, the object.³² A more general feature of the faculty of judgment which is rarely remarked on, but which is obvious enough that it should not be controversial either, is that it is ascribed in a context which does not abstract from the empirical spatio-temporal character of human beings. The beings who judge are human beings in an environment which includes other human beings, nonhuman animals, other natural objects and artefacts. So they are regarded not only as attempting to conceptualize and to understand the natural world, but also as themselves a part of nature: that is, as themselves belonging to the same empirical natural world which they attempt to conceptualize. A further point, also rarely remarked on, but again not likely to be controversial, is that the possession of judgment distinguishes human beings from other animals in the natural world. Human beings and non-human animals share such psychological capacities as sensation, imagination (including memory) and de-

30 See sections IV and V of the Introduction and sections II, IV and V of the First Introduction. 31 See especially section V of the First Introduction, where Kant de nes re ective judgment as a “capacity to re ect on a given representation for the sake of a concept which is thereby made possible” (20:211) and describes it as “working with given appearances so as to bring them under empirical concepts of determinate natural things” (20:213). The example of granite is in the footnote at 20:215–216. 32 See in particular the Second and Fourth Moments of the Analytic of the Beautiful.

The Appearance of Spontaneity

137

sire. But while there is a sense in which animals’ psychological capacities allow them to classify objects in their environment – they can discriminate objects of di erent kinds through responding psychologically to them in ways that are sensitive to, or register, the properties of objects in their environment – they do not conceptualize those objects: they do not represent them as having the properties registered, or as being of the kinds discriminated. A fortiori, animals do not construct systematic theories in an attempt to understand their environment. They are certainly capable of quite sophisticated patterns of response to their environment which re ect their having registered connections between the holding of different states of a airs (the connection, say, between the rustling of leaves in a tree and another animal’s hiding in the tree). But because animals do not conceptualize the relevant objects, they do not represent the states of a airs as holding, so they cannot be said to understand, for example, why the leaves are rustling, or to recognize that there are regular correlations between rustling leaves and other animals in trees. And while animals certainly feel pleasure in objects, they do not take it that other creatures should share the pleasure they feel, nor do they regard their own pleasure in normative terms, as appropriate to the object which causes it. So animals do not judge objects to be beautiful. Now I come to something more controversial. I think that these various exercises of judgment can be uni ed by thinking of judgment, at the most fundamental level, as a capacity human beings have to respond to the objects around in us in ways that incorporate a claim to the normativity of those very responses with respect to those objects. In other words: for human beings to possess a faculty of judgment is for it to be possible for objects outside us to a ect us in such a way that the resulting states of mind involve a legitimate claim to their own appropriateness with respect to those objects.³³ Animals lack judgment, not because

33 “Legitimate” is important here. To regard a human being as capable of judgment is not just to make an empirical psychological claim about her – that she in fact takes (some of) her responses to objects to be appropriate to those objects – but to regard her, at least by default, as entitled to take those responses to be appropriate. The ascription of judgment itself makes a normative claim, and in particular implies that, at least by default, one endorses the judger’s own normative attitudes. The point is important in the context of my overall argument, since without it, (to return to the example) my ascription of judgment to HG would be just another empirical claim about HG’s psychological responses to her surroundings and would not put me in a position to recognize HG as having the same cognition of her environment as I do. In what follows the point is marked by reference to human beings as “aware of” the appropriateness of their responses – where “awareness” is meant to have a factive connotation. To say that I – the thinking subject – represent HG’s psychological response as involving “awareness of” its appropriateness to her circumstances is to say that I not only ascribe the attitude of appropriateness but endorse it. This endorsement is crucial to the possibility of my identifying myself with HG.

138

Ginsborg

they have inferior sensory and imaginative capacities, but because they exercise these capacities without any awareness of a normative dimension to what they are doing, and, more speci cally, without the awareness of what they are doing as appropriate. I take judgment, so understood, to be manifested paradigmatically in the case of judgments of beauty. For a judgment of beauty, as I understand Kant’s account, is in the rst instance a state of mind in which a human being is immediately conscious of her state of mind – the very state of mind which constitutes that judgment – as one which is appropriate to the object, or (equivalently) one which all perceivers of the object should, or ought to, share. But I also take judgment to be operative in cases where we conceptualize, and in turn come to devise theories which systematize, the particular objects which a ect our senses. What makes it the case that human beings do not merely react to objects in our environment in a way which re ects sensitivity to their properties, but also conceptualize them, is that our reactions incorporate awareness of their own appropriateness with respect to those objects. It is in virtue of this awareness, on my understanding of judgment, that these reactions do not merely register the presence of the corresponding property, as in the case of an animal, but ascribe the property, or represent the object as having it. I will not try to elaborate this account of judgment in detail here, since I have presented it elsewhere,³⁴ but I will try to o er enough clari cation to indicate how I take it to address the di culty under discussion. The account has as background the assumption – to which both Kant and other eighteenth-century philosophers, notably Hume, are committed – that our cognitive responses to objects are not just sensory but imaginative, and involve in particular both the integration of current sensations with one another to form uni ed perceptual images (as when current visual and tactile sensations of colour, shape and texture are uni ed in the perception of an individual physical object), and the imaginative recall of previous sensations and images (as when the perceptual image of a presented object calls to mind previous images of other objects of the same kind, or with similar properties). For both Kant and Hume, this kind of imaginative processing characterizes both human and animal responses to the world; but Kant, unlike Hume, holds that there is a di erence in principle between its character in animals and its character in human beings. According to Kant, the activity is not merely imaginative but also conceptual: human beings do not merely associate their representations according to natural psychological laws, but combine them in accordance with con-

34 In various places, but see especially Hannah Ginsborg, “Thinking the Particular as Contained Under the Universal” (in R. Kukla (ed.) Aesthetics and Cognition in Kant’s Critical Philosophy [Cambridge: Cambridge University Press, 2006], 35–60).

The Appearance of Spontaneity

139

cepts, both pure and empirical. But, against this relatively uncontroversial background, di cult questions emerge about what the conceptuality of human experience amounts to. Does combination of sensations “in accordance with” concepts presuppose a grasp of the relevant concepts antecedent to the activity of combining? If so, how are we to explain our possession of these concepts, in particular the empirical concepts? And if not, what content can we give to the idea that the activity is governed by concepts? My account of judgment in the third Critique aims to provide a particular understanding of the conceptual character of experience, one which does not require that the human being grasp concepts in advance of the imaginative activity which accords with them. As I see it, the di erence between conceptualizing objects, as humans do, and merely discriminating among them, as animals do, can be made out without supposing that human beings must possess concepts prior to their conceptualizing activity. What makes our imaginative activity one of conceptualizing, as opposed to the merely associative use of imagination ascribed to human beings by Hume, is simply that it is carried out with an awareness of its appropriateness to the circumstances, an awareness which does not in turn depend on the recognition that this activity conforms to a concept.³⁵ To see this in the context of a simple example, consider an animal which picks up a heavy object in its mouth. An animal in these circumstances might not only experience a sensation characteristic of li ing heavy objects – the kind of sensation I have been labelling a “sensation of weight”³⁶ – but also, in some not necessarily conscious way, recall previous experiences which involved sensations of the same kind. This recall of prior experiences would account for the animal’s behaving in certain ways a er it has picked up the object, for example for the fact that it avoids dropping the object on its own foot. This behaviour would be explained in terms of the animal’s registering the connection between its present sensation of weight and the sensation of pain involved in a previous experience where it did drop something heavy on its foot. Now the animal’s tendency, on picking up a heavy thing, to recall previous experiences in which it had picked up

35 The awareness is correspondingly, as I’ve put it elsewhere, “primitive”: in particular, it is not the awareness that one is representing one’s environment veridically. I discuss this primitive character in my “Thinking the Particular”, pp. 54–58, and in more depth in my “Aesthetic Judgment and Perceptual Normativity,” Inquiry 49 (2006), 403–437, pp. 415–417 and pp. 419–427. 36 To say that it has a “sensation of weight” is in fact misleading, since the sensation as such lacks intentional content. It is “of” weight only in the sense that sensations of that phenomenological type are characteristically the e ect of an object’s weight. In the case of a human being, we can think of it proleptically as a sensation “of weight” in the intentional sense since, as a result of imaginative processing, carried out with awareness of its appropriateness to the circumstances, the human comes, in having the sensation, to represent the object as heavy.

140

Ginsborg

a heavy thing (as opposed to any of its other experiences, say of picking up blue things or warm things), amounts in a sense to its classifying the thing as heavy. That is to say: the animal shows sensitivity to, or registers, the property of weight which the thing shares with the heavy things it previously perceived. But it does not represent or conceptualize the object as heavy, and on my view this is because it recalls its previous experiences – and as a result anticipates pain if the object is dropped – without any awareness of this imaginative response as appropriate to its present sensation. Compare now a human being who, like the animal, li s an object and feels the same, or a similar “sensation of weight”: that is, a sensation of a phenomenological kind typically associated with li ing heavy things. For her too the sensation will trigger an imaginative process involving the recall of previous sensations of the same phenomenological kind, and this might, as in the case of the animal, lead her to anticipate pain if the object is dropped on her foot. However, in contrast to the animal, that imaginative response involves the awareness of its appropriateness to her present circumstances. So, if she happens to recall the experience of some speci c heavy thing she has li ed in the past, she will take this recollection to be appropriate to her present experiential state, in a way that recalling a previously perceived blue or warm thing is not.³⁷ And if she happens, as a result, to anticipate pain if the thing is dropped, she will take this imaginative association too to be appropriate to the present sensation of weight. This makes it the case, on my view, that what she does is not merely to classify the present object with other heavy things, and to anticipate pain if the object is dropped: it is also to bring the object under the concept heavy, and to make the inference, from her recognition that the object is heavy, to the judgment that, if dropped on her foot, it will cause pain. But this account of what makes her activity conceptual does not presuppose the idea of her grasping the concept heavy; nor, relatedly, does it imply that she needs to grasp the concept heavy in order to recall her previous experiences of

37 The idea that the present experience of a heavy object leads us not only to call to mind previous experiences of heavy objects, but also, in so doing, to take the previous experiences to be appropriate to the present experience, might seem strained. It can come to seem more plausible, though, when we re ect on language-use. A less strained example would invoke the fact that a human being li ing a heavy object might express her experience by saying “it’s heavy,” where the word “heavy” both comes to mind naturally (in preference to other words like “blue” or “warm”) and (again in preference to “blue” or “warm”) is regarded by her as appropriate to her present circumstances. This can be seen as a manifestation of her tendency both to recall previous experiences of heavy things (since it was in connection with those previous experiences that she learned the use of the word “heavy”) and to do so with a sense of the appropriateness of what she is doing to her current situation.

The Appearance of Spontaneity

141

heavy things. Rather, it portrays her as grasping the concept heavy, and as capable of using it in inferences, in virtue of the same kind of naturally determined exercise of imagination that we might ascribe to animals. The only di erence³⁸ is that, unlike an animal, she carries out the activity with the awareness of its appropriateness to the circumstances, an awareness which does not in turn depend on a prior recognition that the object is, in fact, heavy. That is, if my interpretation is granted, her activity is an exercise of judgment, speci cally in its re ective aspect. How does this address the di culty described in section III? It does so by offering a way in which a human being can be conceived as both passive and spontaneous, as part of nature, but also as standing in a cognitive relation to nature. To return to the example, it allows me to represent HG not just as a ected by the cup she li s, but as cognizing it – that very cup – to be heavy. So it removes the obstacle identi ed to my identifying myself with HG. Now, it was argued in section IV that this could not be achieved simply by my regarding HG as spontaneous in the same sense of spontaneity that emerges from transcendental re ection on my own cognitive activity. I could not conceive of HG as identical to me by simply gra ing my own conception of myself as spontaneous on to a conception of HG as an empirically determined object. For that would have required me to conceive the objects of my own cognition – the objects a ecting HG’s senses – as epistemically inaccessible to her. The cup which I judge to be heavy would not have been the object of HG’s cognition, but only the unknowable ground of her sensations. But the ascription of judgment to HG allows me to integrate a conception of HG as empirically determined with a conception of her as spontaneous, by understanding her empirical synthesis of the impressions she receives as incorporating a normative dimension through which it quali es as cognitive, and more speci cally as a ording cognition of the very objects which a ect her senses.³⁹ It allows me to

38 “Only” here is something of an overstatement, because I take this di erence to have farreaching implications for the variety and sophistication of human imaginative responses in contrast to those of animals. Most importantly, the fact that human imaginative responses can involve this normative awareness makes possible the meaningful use of language, which in turn enables humans to re ne their psychological responses to their environment in ways which are not available to animals (e.g. we can develop dispositions to respond not just to presented objects, but to other people’s utterances and also to our own linguistic tokens, so that we become capable of inferential and symbolic thought). But I take this contrast between the range of human and animal capacities, striking though it is, to derive from the more fundamental di erence I’ve characterized here. 39 In terms of the passage from Allison’s Transcendental Idealism quoted in note 22, the ascription of judgment as I have described it o ers a way of “objectifying” thinking, that is, “treating it as a psychological occurrence”, without at the same time “destroy[ing] its character as thinking.” It does so by allowing us to consider something initially conceived of as a psychological occur-

142

Ginsborg

conceive of HG both as responding psychologically to the cup’s weight and as – in so doing – judging the cup to be heavy. And this means that I can represent her as making the very same judgment which I myself make when I judge that the cup is heavy. More generally – and to summarize the core thesis of this paper – the ascription of judgment in this sense to HG allows me to conceive of her not only as an object of my cognition, nor even only as an object of my cognition which is itself a subject of cognition; but as the subject of my cognition. It thus removes the obstacle to explaining how there can be such a thing, for Kant, as empirical self-knowledge.

6 Following Horstmann, Rosefeldt, Longuenesse and many other commentators, I have framed my discussion of empirical self-knowledge for Kant by assuming at the outset that the “I” of the “I think” cannot be taken unproblematically as referring to a human being, but must be understood in the rst instance as playing some other role: for example, picking out a non-spatio-temporal activity of thinking, or a non-spatio-temporal subject or agent which carries out that activity. My discussion has been organized around the question of how to reconcile this understanding of the “I” with our ordinary understanding of “I” as referring to a human being, a reconciliation which, following Longuenesse more speci cally, seems to require that the thinking subject be able to “locate” or “situate” itself in the spatiotemporal world by identifying itself with a particular human being in that world. I have argued that there is a di culty standing in the way of this identi cation which can be overcome only by appeal to the faculty of judgment as introduced by Kant in the third Critique (and as understood on the interpretation sketched here). However, it might be objected that this way of framing the discussion distorts the issues, and, further, that it is this distortion which is responsible for the difculty I have described. The di culty, on this objection, is nothing but an artefact of my initial, mistaken conception of the transcendental subject as something other than a human being. It disappears if we understand the “I” from the outset as referring to a human being, so that the “activity” of thought is understood not as taking place in some sense outside of space and time, but as the temporally successive psychological process through which a human being, a er being causally

rence – the empirical synthesis of the manifold according to natural psychological laws – as, at the same time, an exercise of thought.

The Appearance of Spontaneity

143

a ected by spatio-temporal objects impinging on her sense-organs, comes to form cognitive representations of those very objects. Given this understanding of the thinking being and its activity, the question of how the thinking subject comes to pick out some human being to identify with itself – a question whose coherence is in any case open to doubt – might be thought simply not to arise. And in that case, so the objection goes, there is no need to appeal to the Critique of Judgment to make sense of the possibility of empirical self-knowledge. I am sympathetic to the worry that my framing of the question represents a distortion. My aim in adopting it has been to make clear how the di culty arises for the commentators I have mentioned, and I do not want to commit myself one way or another on whether it represents an accurate, or even a coherent, model of Kant’s view. But I do not think that the di culty disappears if we reject the model, and instead take the “I” of the “I think” to pick out a human being from the outset. For in that case we are confronted with the same di culty in a di erent form: how can the empirically determined cognitive processing which takes place in a human being a ected by spatio-temporal objects be thought of as involving the kind of spontaneity Kant means to capture with the “I think”? How are we to reconcile a view of human beings as part of a spatio-temporal world governed by causal laws, and in that sense passive, with a view of those same human beings as engaged in genuinely active thought and judgment with respect to the sensations caused in us by external objects? In short, the question which I dramatized above by asking how the transcendental subject in the example can identify itself with HG, can be asked less dramatically by asking simply how a given human being can be both empirically determined on the one hand, and endowed with spontaneity on the other. I want to suggest in conclusion that the answer to this question is the same as the answer I gave to the question in its previous form. For a human being to respond psychologically to the objects a ecting her senses in a way which is not only naturally determined, but a manifestation of spontaneity, is for her response to involve a legitimate claim to its own normativity, or equivalently – on the interpretation I have been suggesting – for it to be an exercise of the faculty of judgment. In the previous section I embedded this conception of human spontaneity within an account of how a thinking being, conceived of initially as a transcendental subject of experience, must conceive of a human being in order to identify itself with her. But we can dispense with this framing device, and think of judgment not as the appearance of spontaneity, but as spontaneity itself: that is, spontaneity in so far as it can be ascribed to human beings.⁴⁰ On this way of looking at things,

40 This is the approach which I took in my “Kant and the Problem of Experience,” Philosophical

144

Ginsborg

the representation “I think” is simply what a human being is le with when she abstracts from her own spatio-temporal character and that of the objects in her environment. But that there is something for her to be le with, a representation of pure spontaneity, is a re ection of her status – from this empirical perspective, more fundamental than her capacity to entertain the thought “I think” – as not merely a sentient part of nature, but a sentient part of nature which is endowed with judgment.⁴¹

Topics 34 (2006): 59–106), where I argued that what it is for experience to involve understanding or spontaneity is for the empirical synthesis of the sensory manifold to involve awareness of its own appropriateness, that is (in the terms of the present paper) to amount to an exercise of the faculty of judgment as characterized in the third Critique. 41 I am grateful to the other participants at the conference where this paper was initially presented, and to the audience at a subsequent presentation at the University of Potsdam, for helpful comments and discussion. Thanks in particular to Dina Emundts, Stefanie Grüne, Paul Guyer, Béatrice Longuenesse, Tobias Rosefeldt, Ralf Stoecker, Bernhard Thöle, and Daniel Warren. I would also like to express my gratitude to Rolf Horstmann for many years of friendship, hospitality and intellectual stimulation.

Stefanie Grüne

Kant and the Spontaneity of the Understanding In the Critique of Pure Reason there are two chapters in which Kant talks about spontaneity. The rst is the Transcendental Analytic, especially the transcendental deduction, where Kant characterizes the understanding as being spontaneous. The second chapter is the Resolution of the Third Antinomy, in which Kant tries to show that determinism concerning human actions is compatible with those actions being spontaneous. In the Third Antinomy Kant makes clear that by ‘spontaneity’ he means absolute spontaneity. Absolute spontaneity is the “faculty of beginning a state from itself” (CpR A 533/B 561), that is, the faculty of causing oneself or another substance to be in a state without being caused to do so. From what Kant says in the Critique of Practical Reason, it follows that according to him absolute spontaneity is not the only kind of spontaneity, but can be contrasted with something one might call comparative or relative spontaneity. Relative spontaneity is a property of human actions that are not caused by something external to the agent but rather by her own prior states.¹ Unlike the antinomy chapter, the text of the Transcendental Analytic is not very explicit about what kind of spontaneity is meant when the understanding is characterized as being spontaneous. For that reason there is an ongoing discussion among Kant scholars whether Kant conceives of the spontaneity of the understanding as absolute or as relative spontaneity.² The reason why Kant scholars are interested in this question usually is that they want to know whether acts of the understanding are caused by prior events or states, i.e., whether we can describe human cognition as a causally determined process. Since the assumption that such acts are absolutely spontaneous seems to rule out that they are caused by prior events or states, some Kant scholars argue

1 I will give a more detailed analysis of absolute and relative spontaneity in section one of my paper. 2 Allison, Henry (1996), “On naturalizing Kant’s transcendental philosophy”, in: H. Allison: Idealism and Freedom, Cambridge: Cambridge University Press, 53–66; Pippin, Robert (1987), “Kant on the Spontaneity of the Mind”, in: Canadian Journal of Philosophy, Vol. 17, No. 2, 449–476, and Willaschek, Marcus (2009), “Die ‚Spontaneität des Erkenntnisses‘. Über die Abhängigkeit der ‚Transzendentalen Analytik‘ von der Au ösung der Dritten Antinomie”, in: J. Chotas, J. Karásek, J. Stolzenberg (ed.): Metaphysik und Kritik. Interpretationen zur ‚Transzendentalen Dialektik‘ der Kritik der reinen Vernun , Würzburg: Königshausen und Neumann, 165–183, for example, argue that the understanding is absolutely spontaneous; Kitcher, Patricia (1990), Kant’s Transcendental Psychology, Oxford/New York: Oxford University Press, argues that it is relatively spontaneous.

146

Grüne

against a causal account of the acts of the understanding by maintaining that the understanding is absolutely spontaneous. Whereas scholars take very di erent positions on whether the understanding is absolutely spontaneous or not, they all seem to agree on the assumption that, if it is absolutely spontaneous, then acts of the understanding cannot be a caused by prior events or states. Allison, for example, argues that the understanding is absolutely spontaneous and that therefore Kitcher’s causal conception of sensible synthesis, which is one of the activities of the understanding, cannot be right.³ By contrast, Kitcher defends her interpretation by pointing out that the understanding is only relatively spontaneous.⁴ In this paper, I will argue for two claims: My rst claim is that Kant’s characterization of the spontaneity of the understanding in the Transcendental Analytic leaves it open whether or not the understanding has relative or absolute spontaneity. My second claim is that the aforementioned assumption, which is shared by Kitcher, Allison, Pippin and Willaschek, is wrong: Even if the understanding were absolutely spontaneous, the acts of the understanding could be caused by prior events, and hence an absolute spontaneity of the understanding would be compatible with a description of cognition as a causal process. In order to argue for these two claims, I will rst examine what exactly Kant means by relative and absolute spontaneity. In the second part of this paper, I will make a suggestion as to how to understand Kant’s conception of the spontaneity of the understanding and then argue for the claim that this conception is neutral with respect to the question of whether the understanding is relatively or absolutely spontaneous. Since this result does not rule out that the understanding is in fact absolutely spontaneous, the third part of the paper will be concerned with the question of whether an absolute spontaneity of the understanding would rule out the possibility that acts of the understanding are causally determined by prior events.

1 Relative and Absolute Spontaneity In this section I will rst examine what exactly Kant means by ‘relative spontaneity’ and a er this what he means by ‘absolute spontaneity’. Even though, as far as I know, apart from one passage in the Opus postumum Kant nowhere uses the expression ‘relative spontaneity’,⁵ in the secondary literature it is common to contrast absolute spontaneity with relative spontaneity. The

3 Allison, Henry (1996), “On naturalizing Kant’s transcendental philosophy”, p. 62 . 4 Kitcher, Patricia (1990), Kant’s Transcendental Psychology, p. 253 n. 5. 5 Opus postumum, AA XXII 492.

Kant and the Spontaneity of the Understanding

147

two passages which are central for interpreters who ascribe the concept of relative spontaneity to Kant are a passage from the Pölitz version of the Lectures on Metaphysics, in which Kant distinguishes between absolute spontaneity or spontaneitas simpliciter talis and spontaneitas secundum quid, and a passage from the Critique of Practical Reason, in which Kant distinguishes between transcendental freedom and comparative freedom. Since Kant uses the expressions ‘transcendental freedom’ and ‘absolute spontaneity’ synonymously,⁶ and since the concept of spontaneitas secundum quid from the Pölitz version of the Lectures on Metaphysicsand the concept of comparative freedom are closely related,⁷ it is usual to translate what Kant says about comparative freedom into statements about a spontaneity that is not absolute but only comparative or relative. As the question I want to answer in this section is the question of how the concept of the spontaneity of the understanding relates to the concepts of relative and absolute spontaneity, and since the concept of the spontaneity of the understanding that I will analyze stems from Kant’s critical period, in examining Kant’s concept of relative spontaneity my main focus will be on the relevant passages from the Critique of Practical Reason. Kant introduces the concept of comparative freedom in the context of discussing the problem of how it is possible to make human beings morally responsible for their actions even though such actions are caused by prior events. According to Kant, in order to solve this problem it is not enough to point out that actions of human beings are comparatively free. In the Critique of Practical Reason, he characterizes the concept of comparative freedom in the following way: […] according to [the concept of comparative freedom] that is sometimes called a free e ect, the determining natural ground of which lies within the acting being, e.g., that which a projectile accomplishes when it is in free motion, in which case one uses the word ‘freedom’ because while it is in ight it is not impelled from without; or as we also call the motion of a clock a free motion because it moves the hands itself, which therefore do not need to be pushed externally; in the same way the actions of the human being, although they are necessary by their determining grounds which preceded them in time, are yet called free because the actions are caused from within, by representations produced by our own powers, whereby desires are evoked on occasion of circumstances and hence actions are produced at our own discretion. (AA V 96)

One page later Kant writes: […] if the freedom of our will were nothing but […] psychological and comparative but not also transcendental, i.e. absolute, then it would at bottom be nothing better than the free-

6 Cf. CpR A 446/B 474; A 533/B 561. 7 See n. 10.

148

Grüne

dom of a turnspit, which, when once it is wound up, also accomplishes its movements of itself (AA V 97).

In the two passages Kant alternates between ascribing comparative freedom to movements, to actions, to the will and to substances. In order not to make things too complicated I will only talk of relative spontaneity or comparative freedom as being a property of actions. In the rst quotation Kant claims that an action is comparatively free if and only if the natural determining ground of the action is internal to the acting being. Since it is reasonable to suppose that at least in this context natural determining grounds are causes, an action turns out to be comparatively free if and only if its cause is internal to the acting being. How Kant conceives of these internal causes becomes more perspicuous in the following passage, in which Kant explains why comparative freedom is not su cient for moral responsibility: In the question about that freedom which must be put at the basis of all moral laws and the imputation appropriate to them, it does not matter whether the causality determined in accordance with a natural law is necessary through determining grounds lying within the subject or outside him […]; if these determining representations have the ground of their existence in time and indeed in the antecedent state, and this in turn in a preceding state, and so forth, these determinations may be internal and they may have psychological instead of mechanical causality, that is, produce actions by means of representations and not by bodily movements; they are always determining grounds of the causality of a being insofar as its existence is determinable in time and therefore under the necessitating conditions of past time, which are thus, when the subject is to act, no longer within his control and which may therefore bring with them psychological freedom (if one wants to use this term for a merely internal chain of representations in the soul) but nevertheless natural necessity; and they therefore leave no transcendental freedom […] (AA V 96 f.).

According to this passage, determining grounds or causes of actions or states are states of the acting beings, which are temporally prior to the actions or states they cause. Thus, the concept of comparative freedom or relative spontaneity can be explained in the following way: Def.: relative spontaneity/comparative freedom An action is relatively spontaneous or comparatively free if and only if it is caused by a state of the acting being itself which is temporally prior to the action. In the passage from AA V 97 Kant characterizes relative spontaneity as ‘freedom of a turnspit’. According to the de nition just given, the movements of a turnspit are relatively spontaneous because the cause of its movement is a prior state of

Kant and the Spontaneity of the Understanding

149

the turnspit itself, more exactly a prior state of the weights, which drive the turnspit. The contrast between a relatively spontaneous action and an action that is not spontaneous at all can be illustrated for example by the di erence between a mechanical turnspit that is driven by a weight and a turnspit that is moved by a so-called turnspit dog running on a wheel. (This kind of turnspit was popular in England until the 19th century.)⁸ Whereas the movement of the mechanical turnspit is relatively spontaneous, because the cause of its movement is a prior state of the turnspit itself, or rather of the weight which is part of the turnspit, the movement of the second kind of turnspit is not relatively spontaneous (it is not spontaneous at all), because at every moment the cause of its movement is an external force,⁹ namely the movement of the turnspit dog.¹⁰

8 Cf. the Wikipedia entry on turnspit dogs: http://en.wikipedia.org/wiki/Turnspit_Dog. 9 If you have the inclination to conceive of the dog as a part of the turnspit, which for that reason has relative spontaneity, you could think instead of a mill wheel as an example for a thing the movement of which is not spontaneous at all: At every moment the cause of the movement of the mill is an external force, namely the water owing over it. 10 In the Pölitz version of the Lectures on Metaphysics the concept of spontaneitas secundum quid is de ned in a similar way to the concept of comparative freedom: But now the transcendental concept of freedom follows; this means absolute spontaneity, and is selfactivity from an inner principle according to the power of free choice. Spontaneity is either absolute or without quali cation , or quali ed in some respect . – Spontaneity in some respect is when something acts spontaneously under a condition. So, e.g., a body which is shot o moves spontaneously, but in some respect . This spontaneity is also called automatic spontaneity , namely when a machine moves itself according to an inner principle, e.g. a watch, a turnspit. But the spontaneity is not without a quali cation because here the inner principle was determined by an external principle . The internal principle with the watch is the spring, with the turnspit the weight, but the external principle is the artist who determines the internal principle (ML₁ AA XXVIII 267).

Since Kant illustrates spontaneitas secundum quid and comparative freedom by using exactly the same examples both phenomena must at least be closely related. Furthermore, also Kant’s definitions of the two concepts sound similar. Whereas an action displays spontaneitas secundum quid if and only if it takes place because of a principle that is internal to the acting being and that in turn is determined by an outer principle, it is comparatively free if and only if the natural or determining ground of the action is internal to the acting being. Thus, according to both de nitions, an action is comparatively free or displays spontaneitas secundum quid only if it is caused by something that is internal to the acting being. Still, there are also two di erences between the concept of spontaneitas secundum quid and the concept of comparative freedom. The rst is that an action has to ful ll only one condition in order to fall under the concept of a comparatively free action, but has to ful ll two conditions in order to fall under the concept of an action that displays spontaneitas secundum quid. As I have already pointed out, in order to fall under both concepts an action has to be caused by something that is internal to the acting being. For an action to display spontaneitas secundum quid, it furthermore

150

Grüne

Before I turn to the topic of absolute spontaneity let me make one remark concerning a conception of relative spontaneity that is o en ascribed to Kant but, I think, falsely. In the secondary literature, it is common to illustrate the idea that the understanding has relative spontaneity by means of the spontaneity which is displayed by the behavior of computers.¹¹ The main idea, which stems from Sellars, seems to be the following: The state a computer is in a er having computed for a while is one which is not caused by an external or foreign cause. Instead two causal factors are relevant for bringing it about that the computer is in the relevant state: (i) the state the computer is in before being in the relevant state and (ii) a set of dispositions of the computer (its program). Accordingly, on the computer model of relative spontaneity an action is relatively spontaneous if and only if it is brought about by a cause consisting of (i) a state of the acting being prior to the action and (ii) a set of dispositions of the acting being. It is certainly true that the movement of the mechanical turnspit and the movement of a clock’s hand are not due only to the prior states of the turnspit and the

has to be the case that this internal cause has to be causally determined by a cause that is external to it. The second di erence concerns the way in which Kant characterizes the internal cause: Whereas in the Critique of Practical Reason the cause is characterized in temporal terms as a state of the acting being that is prior to the action it causes, in the Pölitz version of the Lectures on Metaphysics no temporal terms are used in order to describe the cause. Kant is only reported as saying that the inner principle of the clock is the spring and the inner principle of the turnspit is the weight. As I see it, the two di erences between the concept of spontaneitas secundum quid and the concept of comparative freedom are related in the following way: The reason why in the Critique of Practical Reason Kant drops the second condition contained in the de nition of the concept of spontaneitas secundum quid is that there he characterizes the cause of a comparatively free action in temporal terms. Since there are no temporal beings that have an in nite duration and since we know from the principles of pure understanding that every alteration has a cause, it follows that it cannot be the case that the chain of the causes of the action of a temporal being only comprises states of this being. Thus, even though the fact that an action is caused by a prior state of the acting being does not imply that this state itself is caused by some outer cause, it does imply that in the chain of the causes of this state at some point there appears an outer cause. 11 Cf. for example Pippin, Robert (1987), “Kant on the Spontaneity of the Mind”, 449–476; Sellars, Wilfrid (1970/71), “’…this I or he or it (the thing) which thinks…’. Immanuel Kant, Critique of Pure Reason (A 346/B 404)”, in: Proceedings and Addresses of the American Philosophical Association 44, 5–31, and Willaschek, Marcus (2009), “Die ‚Spontaneität des Erkenntnisses‘”, 165–183.

Kant and the Spontaneity of the Understanding

151

clock but also to the turnspit and the clock having certain dispositions. Still, in characterizing relative spontaneity Kant does not mention any dispositions or capacities. I assume that the reason why he does not mention them is the following: It is equally true for relatively spontaneous acts and for acts that are not spontaneous at all that they only take place if the acting beings have certain dispositions. For example, just as the mechanical turnspit only moves because it has the disposition to rotate if the weight pulls downwards, the turnspit driven by a turnspit dog running on a wheel only moves because it has the disposition to rotate if the wheel on which the dog is running turns around.¹² Thus, even though it is true of every relatively spontaneous act that dispositions of the acting being are causally relevant for the act’s taking place, this is not characteristic for relatively spontaneous actions but rather a common feature of any action whatsoever. Therefore, the second causal factor from the computer model of relative spontaneity can be dropped, and if it is dropped this model turns out to be the same model as the one I have proposed. As I see it, the problem with the analogy between the spontaneity of a computer and relative spontaneity is that some interpreters seem to think that contrary to what I have just pointed out it is only the second causal factor which is important for the analogy. Therefore, they assume that an action is relatively spontaneous if and only if it is an action for which it holds true that external force on its own is not su cient for the taking place of the action but leads to the performance of the action only in combination with some disposition of the acting being.¹³ This clearly is not the right way to interpret Kant’s concept of relative spontaneity, since according to this analysis the movement of the turnspit driven by a turnspit dog

12 Furthermore, when Kant in the Critique of Practical Reason characterizes a relatively spontaneous act as an act that is caused by an inner state of the acting being, this inner state cannot be a disposition, because according to the Critique of Practical Reason the causing inner state is a state that is prior to the act, whereas as the disposition is not a causing prior state. 13 According to Willaschek for example characterizing the spontaneity of the understanding as relative spontaneity is to characterize it in the following way: “Die [relative; S. G.] Spontaneität des Verstandes besteht darin, dass die Weise, wie wir die Eindrücke unserer Sinne konzeptualisieren und diese Konzeptualisierungen (Urteile) zu einem einheitlichen Bewusstsein verbinden, durch die gegebenen Eindrücke selbst nicht, oder zumindest nicht vollständig, determiniert ist. Das allein schließt aber nicht aus, dass es weitere kausale Faktoren gibt, die, eventuell in Verbindung mit unseren Sinnesemp ndungen, unsere Begri sbildung und Begri sverwendung kausal vollständig festlegen. Man könnte sich das Kantische ‚Gemüt‘ zum Beispiel wie einen Computer vorstellen, in dem das Programm (Anschauungsformen und Kategorien) ‚a priori bereitliegt‘ (vgl. A 20/B 34). Dann wäre das Output des Computers (die Urteile) durch das sinnliche Input und das Programm vollständig festgelegt.” (Willaschek, Marcus (2009), “Die ‚Spontaneität des Erkenntnisses‘”, p. 171).

152

Grüne

as well as the movement of the mechanical turnspit turn out to be relatively spontaneous. Let us now turn to the question of how Kant conceives of absolute spontaneity. As I have already pointed out at the very beginning of the paper, in the Transcendental Dialectic of the Critique of Pure Reason Kant characterizes absolute spontaneity or transcendental freedom as the “faculty of beginning a state from itself” (CpR A 533/B 561), that is, as a faculty of causing a state without being caused to do so by a prior event. Furthermore, he o en talks of absolute spontaneity as a property of actions. It is rather complicated to determine what exactly this property consists in because, as it turns out, Kant has two di erent concepts of absolute spontaneity as a property of actions. Most of the passages in which Kant talks of actions as being transcendentally free are located in the Resolution of the Third Antinomy, where he tries to show that contrary to the rst impression the assumption that some empirical actions ¹⁴ are transcendentally free (and therefore morally imputable) is compatible with the assumption that all actions are causally determined by a prior state of the world and the laws of nature. According to Kant, these two assumptions can be shown to be compatible only if one accepts transcendental idealism, i.e., his claim that things have spatio-temporal properties only insofar as they are appearances, but that as things in themselves they do not exist in space and time. Kant’s fundamental idea for solving the apparent inconsistency between the two assumptions is the following: Because all empirical actions are performed by appearances, and appearances “must have grounds that are not appearances” (CpR A 537/B 565), they have two di erent kinds of causes: On the one hand they are caused by appearances, on the other hand they are caused by intelligible causes or things in themselves. In the rst respect, actions are causally determined by prior events;¹⁵ in the second respect, they can be transcendentally free. In A 537/B 565, for example, Kant writes: The e ect can […] be regarded as free in regard to its intelligible cause, and yet simultaneously, in regard to appearances, as their result according to the necessity of nature (CpR A 357/B 565).

And some pages later he claims: Thus freedom and nature, each in its full signi cance, would both be found in the same actions, simultaneously and without any contradiction, according to whether one compares them with their intelligible or their sensible cause. (CpR A 541/B 569)

14 Empirical actions are actions that take place in time and/or space. 15 I use the expressions ‘to be caused’ and ‘to be causally determined’ synonymously.

Kant and the Spontaneity of the Understanding

153

Clearly, in both of these passages, freedom is ascribed to empirical actions. A er all, Kant claims of those actions, to which freedom is ascribed, that they are caused by appearances. And it is only empirical actions or events that can be caused by appearances.¹⁶ Furthermore, even though in these two passages empirical actions are only explicitly characterized as being free, but not as being transcendentally free, by ‘freedom’ Kant has to mean transcendental freedom. As I have already pointed out, in the Resolution of the Third Antinomy Kant wants to prove that the claim that there is transcendental freedom is compatible with the claim that “everything in the world happens solely in accordance with laws of nature” (CpR A 445/B 473) by showing that one and the same action can both be free and happen in accordance with laws of nature. Obviously, showing this only amounts to proving that the two claims are compatible, if freedom here is transcendental freedom. A much more detailed explanation of this double causation of empirical actions is given in the following passage: But every e ective cause must have a character, i.e. a law of its causality, without which it would not be a cause at all. And then for a subject of the world of sense we would have rst an empirical character, through which its actions, as appearances, would stand through and through in connection with other appearances in accordance with constant natural laws, from which, as their conditions, they could be derived; and thus, in combination with these other appearances, they would constitute members of a single series of the natural order. Yet second, one would also have to allow this subject an intelligible character, through which it is indeed the cause of those actions as appearances, but which does not stand under any conditions of sensibility and is not itself appearance. The rst one could call the character of such a thing in appearance, the second its character as a thing in itself. Now this acting subject, in its intelligible character, would not stand under any conditions of time, for time is only the condition of appearances but not of things in themselves. […] In its empirical character, this subject, as appearance, would thus be subject to the causal connection, in accordance with all the laws of determination; and to that extent it would be nothing but a part of the world of sense, whose e ects, like those of any other appearance, would ow inevitably from nature. […] But in its intelligible character […], this subject would nevertheless have to be declared free of all in uences of sensibility and determination by appearances; and since, in it, insofar as it is a noumenon, nothing happens, thus no alteration requiring a dynamical time determination is demanded, and hence no connection with appearances as causes is encountered in its actions, this active being would to this extent be independent and free of all natural necessity present solely in the world of sense. […] Thus freedom and nature, each in its full signi cance, would both be found in the same actions, simultaneously and without any contradiction, according to whether one compares them with their intelligible or their sensible cause (CpR A 539–41/B 567–9).

16 Further passages in which freedom is ascribed to empirical actions are CpR A 551/B 579 and A 557/B 585.

154

Grüne

According to this passage, every empirical action has two di erent causes: First, the acting being as an appearance in virtue of having an empirical character causes the action to take place; secondly, the acting being as a thing in itself in virtue of having an intelligible character causes the action to take place. To use Kant’s own example from the end of the Resolution of the Third Antinomy, consider the action of telling a lie at a particular time. Let us assume that the empirical character of the liar consists amongst other things in his disposition to perform such actions that magnify his wealth. By having this disposition the liar as an appearance causes the lie to take place. Furthermore, as Kant points out in CpR A 554/B 582 the empirical character itself can be causally explained. For example, the liar’s disposition to perform actions that magnify his wealth might be caused by the way he is raised by his parents and the way he is in uenced by his friends. Thus, as Kant points out in the rst quotation, insofar as an empirical action is caused by the empirical character of the acting being it stands “through and through in connection with other appearances in accordance with constant natural laws” and therefore is part of the natural order. The reason why, in addition to claiming that an empirical action is caused by the acting being in virtue of having an empirical character, Kant can also claim that it is caused by the acting being in virtue of having an intelligible character is that the empirical character depends on the intelligible character. For Kant, the empirical character is “determined in the intelligible character (in the mode of thought [der Denkungsart])” (CpR A 551/B 579). He characterizes the intelligible character as the “transcendental cause” of the empirical character, the empirical character as “the sensible sign” (CpR A 546/B 574) of the intelligible character, and he claims: […] another intelligible character would have given another empirical one […] (CpR A 556/B 584).

As Kant points out in the passage from CpR A 539–41/B 567–9, the acting subject as a thing in itself does “not stand under any conditions of time” and therefore “would not be subject to the law […] that everything that happens must nd its cause in the appearance (of the previous state).” This is exactly the reason why an empirical action insofar as it is caused by the acting subject in virtue having an intelligible character at least might be transcendentally free. If the acting subject as a thing in itself in causing the empirical action is not causally determined by a prior event, it might be the case that it is not causally determined at all.¹⁷ And

17 Two remarks: (1) Here the concept of a cause obviously cannot be the schematized category

Kant and the Spontaneity of the Understanding

155

in this case, the empirical action would be transcendentally free. Thus, transcendental freedom or absolute spontaneity of empirical actions can be characterized in the following way: Def.: absolute spontaneity/transcendental freedom of empirical actions An empirical action of an acting subject is transcendentally free or absolutely spontaneous, if and only if (i) it is caused by the acting subject as a thing in itself and (ii) this causing is not causally determined by anything else. Something that might seem puzzling about this de nition is that whereas so far I have always talked about actions being caused by prior states or events, according to the proposed de nition an action is caused by an acting subject, that is, by a substance. In the secondary literature, there is no agreement concerning the question of the way in which Kant conceives of causes. Whereas it is usually thought that Kant like Hume believes that causes are events, Eric Watkins has argued that, for Kant, what gures as causes are substances or, more precisely, substances in virtue of their causal powers.¹⁸ I do not want to take a stand on this debate. In the passages that are relevant for this paper, Kant sometimes talks of events or states as being causes, and sometimes he talks of substances (in virtue of having a character) as being causes. As I see it, both ways of talking are compatible with each

of a cause, but has to be the unschematized category. (2) In the Resolution of the Third Antinomy there are several passages which convey the impression that for an empirical action to be transcendentally free it is su cient that (i*) it is caused by the acting subject as a thing in itself and that (ii*) the acting being in causing the empirical action is not caused by a prior event (cf. for example A 551 f./B 579 f). Still, this impression has to be misleading. If Kant really believed that every empirical action that ful lls (i*) and (ii*) is transcendentally free, then the Resolution of the Third Antinomy would su er from the following two problems: First, since transcendental idealism implies that all appearances “have grounds that are not appearances” (CpR A 537/B 565), that is, have grounds that are things in themselves, and since no things in themselves are causally determined by prior events, a direct implication of transcendental idealism would be that empirical actions indeed are transcendentally free. Yet, as Kant makes clear at the end of the Resolution of the Third Antinomy, he takes himself only to have shown that empirical actions might be transcendentally free, not that they are transcendentally free (cf. A 557 f./B 585 f.). Secondly, since for Kant every acting being (not just human beings, but also animals, plants and inanimate substances) has an empirical and an intelligible character and thus is the cause of its empirical actions not just as an appearance, but also as a thing in itself, and since no thing in itself is causally determined by prior events, accepting (i*) and (ii*) as a de nition of the concept of absolute spontaneity would have the consequence that all empirical actions are transcendentally free. 18 Watkins, Eric (2005), Kant and the Metaphysics of Causality, Cambridge: Cambridge University Press, pp. 230–297.

156

Grüne

other. If substances cause actions in virtue of having a certain character, i.e., in virtue of having a set of dispositions, then this implies that prior to the action the subject is in some state which triggers the activation of the disposition and thus is a further causal factor in the explanation of the action. For example, if the liar causes himself to lie by having the disposition to perform such actions that magnify his wealth then, prior to the action, he has to be in the state of believing that lying in the relevant situation will magnify his wealth. This is the reason why in the Resolution of the Third Antinomy Kant writes that “all the actions of the human being in appearance are determined in accord with the order of nature by his empirical character and the other cooperating causes” (CpR A 549/B 577; italics mine). Similarly, there is also an implication in the other direction: For Kant, the state of believing that lying in the relevant situation will magnify the subject’s wealth only causes the lie to take place, if the acting subject furthermore has some disposition, for example the disposition to perform such actions that magnify her wealth. Thus, both ways of talking about causality complement each other.¹⁹ Whether one of the two ways is more fundamental is something that can be le open in this paper. For Kant empirical actions are not the only kind of actions that can be transcendentally free. In addition, there are also intelligible actions, which can be transcendentally free. Whereas in the Resolution of the Third Antinomy Kant does not distinguish between these two kinds of transcendentally free actions in a very explicit way,²⁰ he does so in his later work Religion within the Boundaries of Mere Reason in the context of answering the question of how it is possible to make human beings morally responsible for evil deeds. In Book One of Religion, Kant explains that the intelligible character of human beings consists in having a highest maxim in which the incentive [Triebfeder] of morality is subordinated to the incen-

19 This way of explaining why characterizing events or states as causes is compatible with characterizing substances as being causes presupposes that for Kant, contrary to what one rst might think, causes are not su cient conditions. Rosefeldt, Tobias (forthcoming), “Kants Kompatibilismus”, sec. 5, argues that Kant’s conception of causes is very close to that of Mackie who characterizes a cause as an “insu cient but necessary part of a condition which is itself unnecessary but su cient for the result” (Mackie, J. L. (1965), “Causes and Conditions”, in: American Philosophical Quarterly, Vol. 2/4, p. 245). What I say about causes is based on the assumption that this is the right analysis of Kant’s conception of causality. 20 Whereas in most passages of the Resolution of the Third Antinomy in which Kant talks of free actions, these actions clearly are empirical actions, there is at least one passage, in which he describes an intelligible action as being free (cf. CpR A 544/B 572). Furthermore, it follows from the way in which he characterizes transcendental freedom as a faculty that there are intelligible actions that might be transcendentally free. See p. 157 f.

Kant and the Spontaneity of the Understanding

157

tive of self-love.²¹ Furthermore, he claims that “nothing is, however, morally (i.e. imputable) evil but that which is our own deed” (AA VI 31). In Kant’s terminology a deed is a special kind of action, namely a free action.²² Thus, in order to be able to make a human being morally responsible for what she has done it is not enough that the action has been caused by her intelligible character. In addition, the human being has to adopt her intelligible character by acting freely. Clearly, the free action by which a human being acquires its intelligible character cannot be an empirical action but has to be an intelligible action, that is, an action performed by the human being as a thing in itself. Accordingly, in Religion Kant states that one has to distinguish between two kinds of (transcendentally) free actions: Now the term ‘deed’ can in general apply just as well to the use of freedom through which the supreme maxim (either in favor of, or against, the law) is adopted in the power of choice, as to the use by which the actions themselves (materially considered, i.e. as regards the objects of the power of choice) are performed in accordance with that maxim. […] The former is an intelligible deed, cognizable through reason alone apart from any temporal condition; the latter is sensible, empirical, given in time (factum phaenomenon) (AA VI 31).

In this passage, Kant only states that there are two di erent kinds of free actions, namely the intelligible action of adopting the supreme maxim, that is, of adopting an intelligible character, and the empirical action of doing something in accordance with the highest maxim, for example lying to somebody. He does not say anything about how he understands the expression ‘freedom’ in this context. I have already explained how I interpret the concept of transcendental freedom or of absolute spontaneity of empirical actions. How Kant understands transcendental freedom or absolute spontaneity of intelligible action can be shown by analyzing two passages from the Third Antinomy and its Resolution in which Kant talks about transcendental freedom as a faculty: Suppose there were a freedom in the transcendental sense […], namely a faculty of absolutely beginning a state […]; then not only will a series begin absolutely through this spontaneity, but the determination of this spontaneity itself to produce the series, i.e., its causality, will begin absolutely, so that nothing precedes it through which this occurring action is determined in accordance with constant laws (CpR A 445/B 473; italics mine). [R]eason creates the idea of a spontaneity, which could start to act from itself, without needing to be preceded by any other cause that in turn determines it to action according to the law of causal connection (CpR A 533/B 561).

21 Cf. Religion, AA VI 21 f., 36 f., 47. 22 Cf. Pölitz version of the Metaphysical Lectures AA XXVIII 564.

158

Grüne

In both passages transcendental freedom or absolute spontaneity is characterized as the faculty to act without being caused to do so. Clearly, in this context, these uncaused actions can only be intelligible not empirical actions, since empirical actions are causally determined, namely by both an intelligible and an empirical or sensible cause. Thus, the passages I have quoted suggest that we understand absolute spontaneity or transcendental freedom of intelligible actions in the following way: Def.: absolute spontaneity/transcendental freedom of intelligible actions An intelligible action is transcendentally free or absolutely spontaneous if and only if it is not caused by anything.

2 The Spontaneity of the Understanding in the Transcendental Analytic In this section, I will analyze several passages from the Transcendental Analytic in which Kant describes the understanding as being spontaneous. In this way I will present what I take to be Kant’s conception of the spontaneity of the understanding. Furthermore, I will argue for the claim that this conception of the spontaneity of the understanding leaves open whether the understanding is relatively or absolutely spontaneous. Before I start let me make two preliminary remarks: First of all, in the Transcendental Analytic, Kant uses the expression ‘spontaneity’ in di erent ways. Sometimes he uses it as a term for a faculty. In this case he employs it as a synonym for ‘understanding’.²³ Sometimes he uses ‘spontaneity’ as a term for a property of thinking, i.e., for a property of acts of the understanding.²⁴ Furthermore, there is at least one passage in which he talks of the spontaneity of concepts,²⁵ and outside the Transcendental Analytic he also talks of the spontaneity of the understanding.²⁶ Depending on the context, I will use the expressions ‘spontaneity’ and ‘spontaneity of the understanding’ either as a term for a faculty or as a term for a property of acts of the understanding. Secondly, I want to stress that, for Kant, the understanding is the capacity to perform three di erent kinds of acts: It

23 Cf. CpR A 51/B 75, CpR B 129 f. 24 Cf. CpR A 68/B 93, A 77/B 102, B 157 f. 25 Cf. CpR A 50/B 74. 26 Cf. What real progress has metaphysics made in Germany since the time of Leibniz and Wol AA XX 275 f., and Letter to Beck from 20. 1. 1792 (AA XI 316).

Kant and the Spontaneity of the Understanding

159

is the capacity to form concepts,²⁷ the capacity to judge,²⁸ and the capacity to synthesize or process sensations in such a way as to form intuitions, an activity I will call ‘sensible synthesis’.²⁹ Thus, to determine what exactly Kant’s conception of the spontaneity of the understanding is means to nd out what the spontaneity of concept formation, judging and sensible synthesis consists in. So, let us proceed to answering this question. At the very beginning of the Transcendental Logic, Kant characterizes sensibility and understanding in the following way: If we will call the receptivity of our mind to receive representations insofar as it is a ected in some way sensibility, then on the contrary the faculty for bringing forth representations itself, or the spontaneity of cognition, is the understanding (CpR A 51/B 75).

According to this passage, sensibility and understanding both are capacities to supply the mind with representations. The di erence between sensibility and understanding consists in the way in which they do this. The sensibility is the capacity of the mind to react in a certain way, if acted upon, namely to react by being in a representational state. The understanding or the ‘spontaneity of cognition’, by contrast, is the capacity to produce representations, that is, it is the capacity to be in a representational state as the result of having acted in a certain way.³⁰ At least in this passage we do not learn anything about the question of whether this act of producing representations is an activity that is caused by prior states or one that is not.

27 Cf. CpR A 126 f. In this passage Kant does not explicitly describe the understanding as the faculty to form concepts. Instead he characterizes it as “a faculty for making rules through the comparison of the appearances.” But since for Kant rules are concepts, it follows from this passage that the understanding is the faculty to form concepts. 28 Cf. CpR A 69/B 94 and A 126. 29 Kant nowhere explicitly states that one of the functions of the understanding is to synthesize sensations in such a way as to produce intuitions. Still among interpreters there is a very wide agreement that by far the most plausible way to interpret many passages from the transcendental deduction is by attributing this position to Kant. (Two interpreters who doubt that the understanding plays a role in the formation of intuitions are Allais, Lucy (2009), “Kant, NonConceptual Content and the Representation of Space”, in: The Journal of the History of Philosophy 47(3), 383–413, and Hanna, Robert (2001), Kant and the Foundations of Analytic Philosophy, Oxford: Clarendon Press, ch. 1.) One of the passages of the deduction which in my view speaks very much in favor of the assumption that Kant attributes the function to perform the activity of sensible synthesis to the understanding is CpR B 130. 30 As will turn out further below, the concept of action that is relevant here, is a very broad concept of action, much broader than the concept of a conscious or intentional action. See p. 161 f.

160

Grüne

The assumption that spontaneity is the capacity to act in such a way as to produce representations is also con rmed by other passages. In § 15 of the Bdeduction Kant writes: The manifold of representations can be given in an intuition that is merely sensible, i.e., nothing but receptivity, and the form of this intuition can lie a priori in our faculty of representation without being anything other than the way in which the subject is a ected. Yet the combination (conjunctio) of a manifold in general can never come to us through the senses, and therefore cannot already be contained in the pure form of sensible intuition; for it is an act of the spontaneity of the power of representation, and, since one must call the latter understanding, in distinction from sensibility, all combination […] is an action of the understanding, which we would designate with the general title synthesis in order at the same time to draw attention to the fact that we can represent nothing as combined in the object without having previously combined it ourselves, and that among all representations combination is the only one that is not given through objects but can be executed only by the subject itself, since it is an act of its self-activity (CpR B 129 f.).

Two other very similar sounding passages (which do not stem from the Transcendental Analytic) are the following: Since composition, either through the object or through its representation in intuition, cannot be given but must be produced, it must rest on the pure spontaneity of the understanding in concepts of objects in general (of the composition of the given manifold) (Letter to Beck from 20. 1. 1792, AA XI 316). All representations which constitute an experience can be assigned to sensibility, with one solitary exception, namely that of the composite as such. Since compounding cannot fall under the senses, but has to be performed by ourselves, it belongs, not to the receptive nature of sensibility, but to the spontaneity of the understanding (What real progress has metaphysics made in Germany since the time of Leibniz and Wol AA XX 275 f.; italics mine).

In all three passages, Kant’s reasoning is more or less the same. He claims that every combination of representations is the result of something we execute [verrichten], we perform [machen], we produce [machen], of an action or an act. From this he concludes that combination can never be brought about by receptivity, but only by spontaneity. Clearly, in order for the reasoning to be valid Kant has to accept that receptivity is not a capacity to act in a certain way, but that only spontaneity is such a capacity. At rst glance, a di erence between the last three passages and the passage from the very beginning of the Transcendental Analytic seems to be that, according to the last three passages, what is produced by spontaneity is not a representation, but rather the combination of representations. In this case, these passages would only show that spontaneity is a capacity to act or to produce something, but they would not show that it is a capacity to produce representations. Still,

Kant and the Spontaneity of the Understanding

161

this impression is misleading. The only way for the understanding to combine representations is by representing their content as being combined or as forming a unity. For Kant, combining representations and representing their content as forming a unity is one and the same thing. That is why in B 129 f. and in Real Progress Kant switches so easily from claiming that only spontaneity can combine representations to claiming that the representation which can only be produced by spontaneity is the representation of combination. Thus, these passages conrm that the spontaneity of the understanding is the capacity to act in such a way as to produce representations.³¹ And again, in none of the passages does Kant say anything concerning the question of whether these acts are caused by prior events or not. My claim that Kant’s characterization of the spontaneity of the understanding as a capacity to act in such a way as to produce representations leaves open whether acts of the understanding are caused by prior events can be further substantiated by examining Kant’s concept of an action.³² Kant adopts this concept from the metaphysical tradition of his time, especially from Baumgarten’s Metaphysica. In the Pölitz version of the Lectures on Metaphysics, the concept of an action is characterized in the following way: A substance acts if and only if in virtue of its own power it causes itself or another substance to be in some state.³³ This is a very broad concept of an action, according to which not only human beings, but also animals and inanimate substances have the capacity to act.³⁴ In the characterization of the concept of an action nothing is said concerning the question of whether a substance, when it causes itself or another substance to be in some state, is itself caused to do so or not. But only if this were not the case – if the causing was itself uncaused – would the action be absolutely spontaneous. Thus, actions for Kant can either be absolutely spontaneous or not. Accordingly, in the Pölitz version of the Lectures on Metaphysics he points out that only a subclass of acting is ‘Thun’ (facere), which means acting from freedom.³⁵ Since actions do

31 Since, as I have pointed out in the beginning of this part of my paper, the understanding is the capacity to perform three di erent kinds of actions, there are three di erent kinds of representations that are produced by the understanding, namely intuitions, concepts and judgments. 32 I use the expressions ‘act’ and ‘action’ synonymously. 33 ML₁ AA XXVIII 564. That Kant here is not only reporting on Baumgarten’s opinion concerning the concept of action, but is expressing his own view is shown by the fact that in the Critique of Pure Reason for somebody who wants to inform himself about the predicables of the understanding, which comprise amongst others the concept of action, Kant recommends reading the ontological textbooks, the most important of which is Baumgarten’s Metaphysica (cf. CpR A 82/B 108.). 34 Obviously this concept of an action is not the concept of an intentional or conscious action. 35 ML₁ AA XXVIII 564.

162

Grüne

not have to be absolutely spontaneous, from Kant’s characterization of the spontaneity of the understanding as the capacity to act in a certain way it does not follow that the spontaneity of the understanding is a case of absolute spontaneity. Still, if one considers the further development of the deduction, it turns out that Kant’s notion of the spontaneity of the understanding is narrower than the notion of a capacity to act in such a way as to produce representations. This follows from the fact that in the deduction Kant distinguishes between two di erent kinds of synthesis, namely between transcendental or productive and empirical or reproductive synthesis. In § 24 of the B-deduction for example Kant writes: […] but insofar as this synthesis [i.e., the synthesis of imagination; S. G.] is still an exercise of spontaneity […], the imagination is to this extent a faculty for determining the sensibility a priori, and its synthesis of intuitions, in accordance with the categories, must be the transcendental synthesis of the imagination, which is an e ect of the understanding on sensibility […]. Now insofar as the imagination is spontaneity, I also occasionally call it the productive imagination, and thereby distinguish it from the reproductive imagination whose synthesis is subject solely to empirical laws, namely those of association […] (CpR B 151 f.).

The di erence between transcendental or productive and empirical or reproductive synthesis is the di erence between a synthesis of representations according to the categories and a synthesis according to empirical laws. Only the rst kind of synthesis is an activity that Kant is willing to ascribe to the understanding (he says that the transcendental or productive synthesis of the imagination is an effect of the understanding) and is willing to describe as being spontaneous. So, here Kant takes back his claim from § 15 that every combination or synthesis is a spontaneous act of the understanding.³⁶ But if according to Kant, there is a nonspontaneous activity of synthesis, by which the representation of combination is produced, then spontaneity cannot be the general capacity to produce representations by acting in a certain way.³⁷

36 The apparent con ict between B 129 f. and § 24 can be resolved in the following way: In the transcendental deduction Kant argues that with respect to self-conscious beings reproductive synthesis cannot take place without transcendental synthesis (see, for example, CpR A 119–123, B 139 f.). Therefore Kant is justi ed in claiming that (in fact) every synthesis that takes place in self-conscious beings is an act of the understanding, even though reproductive synthesis, were it to occur independently of transcendental synthesis, would not be an act of the understanding. 37 As should be clear from n. 36, I do not want to claim that with respect to self-conscious beings reproductive synthesis can take place apart from transcendental synthesis. I only want to claim that Kant distinguishes between two concepts of synthesis, and that the concept of reproductive synthesis is not the concept of a spontaneous act.

Kant and the Spontaneity of the Understanding

163

If one wants to nd out why Kant only quali es transcendental, but not reproductive synthesis as being spontaneous, one has to nd out more about the di erence between these two kinds of synthesis. Kant’s explanation of reproductive synthesis runs as follows: When causal impact on sensibility regularly brings it about that the mind receives the same sensations, this causes the mind, or more speci cally, the imagination to acquire the capacity to associate, combine or synthesize these representations. If, a er the imagination has acquired this capacity, the same sensations again are presented by sensibility, then the imagination actualizes its capacity, i.e., it actually combines the sensations. Reproductive synthesis of imagination just consists in this actualization of an acquired capacity to combine speci c sensations and thereby to produce the representation of combination. (Thus, contrary to what is suggested by the names of these two kinds of synthesis, the di erence between productive and reproductive synthesis is not that only by the rst, but not by the second representations are produced. Rather, the di erence consists in the way in which representations are produced and in the kind of representations that are produced).³⁸ In addition to this capacity human beings also have the capacity to combine sensations according to the categories. If this capacity to combine sensations according to the categories is actualized, transcendental synthesis takes place. Just to give one example for such a synthesis: To synthesize sensations according to the category of quantity means to synthesize sensations which form a homogeneous manifold, that is, sensations whose contents fall under the same concept. The result of such a synthesis is the formation of the representation of a quantum, that is, the representation of an entity that consists of a multitude of homogeneous parts and therefore can be determined quantitatively. The fact that reproductive synthesis is the actualization of a capacity that is acquired in the way described above is the reason why in the A-Deduction Kant writes: [O]nly the productive synthesis of the imagination can take place a priori; for the reproductive synthesis rests on conditions of experience. (CpR A 118)

I assume that in this context the expression ‘conditions of experience’ refers to sensible input. That sensibility supplies sensations is clearly a condition of experience. Still, the quotation cannot be taken to mean that reproductive, in contrast to productive, synthesis only takes place if sensibility presents sensations. In this respect, productive and reproductive syntheses do not di er from each other, but

38 This is the way, at least, that I understand what Kant says in CpR A 100 and A 120 f.

164

Grüne

absolutely look alike. Both only take place if the sensibility supplies sensations.³⁹ As long as there are no sensations there is nothing that can be synthesized in the rst place. Therefore, I suggest we understand the quotation in the following way: To say that reproductive synthesis rests on conditions of experience is to say that it is the actualization of a capacity whose existence is causally dependent on sensible input. In this respect there really is a di erence between reproductive and productive synthesis: Productive synthesis is the actualization of a capacity whose existence is not causally dependent on sensible input. In the case of productive synthesis even though the actualization of the relevant capacity is dependent on sensible input, the existence of the capacity itself is not; instead this capacity is innate. This di erence, I want to suggest, is the reason why Kant quali es only productive, but not reproductive synthesis as being spontaneous. If this is right, then we have found out what the spontaneity of the understanding consists in: Def.: spontaneity of the understanding An act is an act of the understanding or a spontaneous act if and only if (i) it is the act of producing a representation and (ii) this act of producing a representation is the actualization of a an innate capacity.⁴⁰ Before answering the question of whether this conception of the spontaneity of the understanding implies that the understanding is relatively or absolutely spontaneous, I want to discuss one possible objection to my interpretation of Kant’s concept of spontaneity. This objection runs as follows: Productive synthesis is a synthesis in accordance with the categories. This means that in productive synthesis the categories are used in some way.⁴¹ Therefore, productive synthesis only takes

39 Depending on how one conceives of the synthesis that leads to the formation of the pure intuitions of space and time, this might not be true for such a synthesis. But clearly it is true for the synthesis that leads to the formation of empirical intuitions. 40 As I will show below, in the case of sensible synthesis this innate capacity consists in the capacity to synthesize sensations in a lawful way. In the case of judging, I would suggest that the innate capacity consists in the capacity to form judgments according to the di erent logical forms of judging. Since Kant says so little on the topic of concept formation, it is di cult to determine what the relevant innate capacity might be. Concerning this question, chapter 6 of Longuenesse’s book Kant and the Capacity to Judge might be helpful. In this chapter she argues that the logical forms of judgment guide the generation of empirical concepts (Longuenesse, Béatrice (1998), Kant and the Capacity to Judge, Princeton: Princeton University Press, pp. 131–166). 41 As I see it, concepts for Kant can be used in di erent ways. In judgments concepts are used in a di erent way than in sensible synthesis. Thus, saying that the categories are used in sensible synthesis does not entail that sensible synthesis is a kind of judgment.

Kant and the Spontaneity of the Understanding

165

place if the synthesizing subject possesses the categories. Yet, Kant explicitly denies that the categories are innate concepts. According to him, all concepts, not just empirical concepts, are “made according to their form” (Jäsche-Logic, AA IX 93).⁴² And to make the categories according to their form presupposes sensible input. Thus, productive synthesis cannot be the actualization of an innate capacity. This objection can be based on passages in which Kant explains how the categories are formed. In his Inaugural Dissertation he writes: Since, then, empirical principles are not found in metaphysics, the concepts met with in metaphysics are not to be sought in the senses but in the very nature of the pure understanding, and that not as innate concepts but as concepts abstracted from the laws inherent in the mind (by attending to its actions on the occasion of experience), and therefore are acquired concepts. To this genus belong possibility, existence, necessity, substance, cause etc. […] (AA II 395).

It is unclear what exactly the “actions” to which one attends “on the occasion of experience” are supposed to be. Such acts could either be acts of forming judgments or they could be acts of sensible synthesis. In the rst case we would acquire the pure concepts of the understanding by paying attention to the laws according to which we form judgments. According to the second model we would acquire them by paying attention to the laws according to which sensations are synthesized.⁴³ Both models, however, entail that the formation of the categories presupposes sensible input. We can pay attention to the way in which we form judgments only if we already have formed judgments. As long as we have not acquired the pure concepts of the understanding, such judgments can only be composed of empirical concepts, and the possession of empirical concepts presupposes the presence of intuitions. The presence of intuitions in turn presupposes sensible input. Accordingly, on the rst model the acquisition of categories requires sensible input. The same holds for the second model. One can pay attention to the activity of sensible synthesis only if it takes place, and in order for sensible synthesis to take place,

42 See Jäsche-Logic, AA IX 93 f., CpR A 56/B 80, Re ection 2851 (AA XVI 546), Re ection 2855 (AA XVI 547), Re ection 2856 (AA XVI 548) and 2859 (AA XVI 549). 43 Of course, it is highly unlikely that according to the Dissertation we form the categories by paying attention to the act of sensible synthesis. A er all, in the Dissertation Kant has not yet developed his theory of sensible synthesis. I just want to call attention to the fact that the Dissertation’s model for forming the categories by re ecting on acts of the understanding allows being transformed in such a way that the acts of the understanding on which the subject re ects are conceived of as acts of sensible synthesis instead of as acts of judging. As I am about to show, this is exactly what happens in CpR A 85 f./B 118.

166

Grüne

sensibility has to supply sensations as the material that is synthesized. Therefore, according to both models, the formation of the categories presupposes sensible input. But then productive synthesis cannot be the actualization of an innate capacity, and, therefore, the proposed characterization of productive imagination and, accordingly, the interpretation of the spontaneity of the understanding must be false. Or so the objection goes. One can react to this objection by pointing out that even though Kant clearly says that the categories are not innate concepts he also claims that there is something which is closely connected to the categories and which is innate. For example, in the passage from De Mundi Sensibilis quoted above he writes that the categories are “abstracted from the laws inherent in the mind [aus den der Erkenntniskra eingep anzten Gesetzen abgezogen]”. Thus, there are laws that are ‘inherent’ or ‘implanted’ in the power of cognition or the understanding. These laws are laws concerning the actions to which one can attend – and which, therefore, also take place – “on the occasion of experience.” To say that these laws are implanted in the understanding is to say that the understanding has the innate capacity or disposition to act in a certain lawful way when having an experience. Since the only two acts that the understanding performs “on the occasion of experience” are sensible synthesis and judging, this capacity is the capacity to synthesize or judge in a certain lawful way. Furthermore, this innate capacity is very closely related to the categories, since the categories are formed by re ecting on its actualization. In the Critique of Pure Reason Kant still holds a very similar view: Now, we already have two sorts of concepts of an entirely di erent kind, which yet agree with each other in that they both relate to objects completely a priori, namely the concepts of space and time, as forms of sensibility, and the categories, as concepts of the understanding. […] in the case of these concepts, as in the case of all cognition, we can search in experience, if not for the principle of their possibility, then for the occasional causes of their generation, where the impressions of the senses provide the rst occasion for opening the entire power of cognition to them and for bringing about experience, which contains two very heterogeneous elements, namely a matter for cognition from the senses and a certain form for ordering it from the inner source of pure intuiting and thinking, which, on the occasion of the former, are rst brought into use and bring forth concepts. (CpR A 85 f./B 118)

According to this passage, thinking or understanding has a certain form for ordering the matter of cognition. Since the matter of cognition consists in impressions of the senses or sensations, the form of thinking is the form for ordering or synthesizing sensations. This is con rmed by Kant’s claim that the understanding’s form is brought into use or actualized when impressions of the senses, that is, sensations, are present. Now, a form for synthesizing sensations seems to be nothing else than the capacity to synthesize sensations in a regular or lawful way. Even though Kant does not explicitly say that this capacity is implanted or innate, he

Kant and the Spontaneity of the Understanding

167

characterizes it as stemming from the inner source of thinking, which suggests that it is not acquired but rather innate.⁴⁴ And he claims that the actualization of this capacity of the understanding leads to the formation of the categories. Thus for Kant the understanding has the innate capacity to synthesize representations in a regular or lawful way and the categories are formed by re ecting on the actualization of this capacity. Now, it is very plausible to assume that this innate capacity is exactly the capacity that is relevant for productive synthesis. And if this is right, then it turns out that contrary to the objection I have discussed, Kant indeed conceives of productive synthesis as the actualization of an innate capacity to synthesize sensible representations. Still, in order to refute the objection completely one furthermore has to explain how to understand Kant’s claim from § 24 of the B-deduction that productive synthesis takes place according to the categories. If productive synthesis is the actualization of an innate capacity and if the categories are formed by paying attention to the actualization of this capacity, then in productive synthesis the categories cannot already be deployed. Here one has two options. The rst assumes that characterizing productive synthesis as taking place in accordance with the categories does not entail that in productive synthesis the categories are used.⁴⁵ As I see it, Kant’s formulations in § 24, indeed, do not prove that the categories are used in productive synthesis. Still, I think that there are many other reasons which speak in favor of believing that, for Kant, categories do play this role.⁴⁶ Therefore,

44 Besides, in the Eberhard Controversy which also stems from the critical period, Kant characterizes the categories as being originally acquired and writes: “These likewise [the transcendental concepts of the understanding] are acquired and not innate, but their acquisition, like that of space, is originaria and presupposes nothing innate except the subjective conditions of the spontaneity of thought (in accordance with the unity of apperception)” (Entdeckung, AA VIII 223). Here Kant explicitly claims that the formation of the categories presupposes something innate, namely innate conditions for the spontaneity of thought. This ts very well with my account of the spontaneity of the understanding, for according to this account it is a necessary condition for an act to be a spontaneous act that the understanding is an innate capacity for producing representations. Thus on my view there are indeed innate conditions for the spontaneity of thought. 45 Cf. Waxman, Wayne (1991), Kant’s Model of the Mind. A new Interpretation of Transcendental Idealism, New York/Oxford: Oxford University Press. 46 The most important reason is that the argumentation from the transcendental deduction would not go through if possession of the categories was not a necessary condition for sensible synthesis to take place. In the transcendental deduction Kant wants to prove that the categories apply to objects of experience by showing that they are necessary conditions for experience. And he tries to show that the categories are necessary conditions of experience by arguing that they are necessary conditions for the presence of intuitions, which he does in turn by establishing that sensible synthesis would not lead to the formation of intuitions if the categories were not used in sensible synthesis.

168

Grüne

I prefer the second option, which is to assume that Kant uses the expression ‘concept’ in two di erent ways. Concepts as clear concepts are applied in judgments and are always made according to their form. Concepts as obscure concepts or as schemata are used in sensible synthesis and do not have to be made according to their form.⁴⁷ According to this view, in productive synthesis the categories are used as obscure concepts or schemata, whereas the categories as clear concepts are formed by paying attention to productive synthesis.⁴⁸ Let us now turn to the question of whether it follows from Kant’s conception of the spontaneity of the understanding that the understanding is relatively or absolutely spontaneous. In the last part of this section I have suggested that an act is an act of the understanding or a spontaneous act if and only if (i) it is the act of producing a representation and (ii) this act of producing a representation is the actualization of an innate capacity. And in the rst part of my paper I have de ned relative and absolute spontaneity in the following way: Def.: relative spontaneity An action is relatively spontaneous if and only if it is caused by a state of the acting being which is temporally prior to the action. Def.: absolute spontaneity of empirical actions An empirical action of an acting subject is absolutely spontaneous, if and only if (i) it is caused by the acting subject as a thing in itself and (ii) this causing is not causally determined by anything else. Def.: absolute spontaneity of intelligible actions An intelligible action is absolutely spontaneous if and only if it is not caused by anything.

47 Longuenesse, Béatrice (1998), Kant and the Capacity to Judge, argues that the categories as rules for sensible synthesis are schemata. In Grüne, Stefanie (2009), Blinde Anschauung, Frankfurt: Klostermann, I argue that the categories as such rules are obscure concepts. 48 If, by contrast, one believes that in productive synthesis the categories as clear concepts are used, then one cannot accept my account of productive synthesis and my interpretation of the spontaneity of the understanding.

Kant and the Spontaneity of the Understanding

169

Obviously, that an act falls under the concept of a spontaneous act of the understanding leaves open whether the act is relatively spontaneous or not. A er all, an act of producing a representation that is the actualization of an innate capacity can be caused by a state of the acting being, but can also be caused by an outer cause (or it may not be caused at all). Similarly, the de nition of the spontaneity of the understanding does not imply that the understanding and its acts are absolutely spontaneous. First of all, an act of producing a representation that is the actualization of an innate capacity can be either causally determined or not causally determined. Thus, being a spontaneous act of the understanding leaves open whether this act is an absolutely spontaneous intelligible act or not. Secondly, being a spontaneous act of the understanding has also no implications concerning the question of whether the act is an absolutely spontaneous empirical action. For even if acts of producing representations that are actualizations of an innate capacity meet the rst part of the de nition of absolutely spontaneous empirical actions, because they are caused by the subject as a thing in itself, this causing may still either be caused by something else or not, that is, spontaneous acts of the understanding in the sense of the above de nition may either meet the second part of the de nition of absolute spontaneity of empirical acts or not. Hence, Kant’s conception of the spontaneity of the understanding has no implications concerning the question of whether acts of the understanding are absolutely or relatively spontaneous. Before I come to the last part of my paper I would like to mention brie y one further argument against the assumption that the spontaneity of the understanding is absolute. Kant repeatedly emphasizes that we do not have any knowledge of ourselves as things in themselves.⁴⁹ Yet, he claims that acts of the understanding are spontaneous. Regardless of whether by acts of the understanding he means intelligible or empirical acts, if the spontaneity of the understanding were absolute spontaneity, this would imply that we would have knowledge of ourselves as things in themselves: If acts of the understanding were intelligible acts, we would know that we as things in themselves perform acts of synthesizing, judging or concept formation that are not caused by anything. If acts of the understanding were empirical acts, we would know that we as things in themselves are causally undetermined causes of empirical acts of synthesizing, judging and concept formation. Since for Kant we do not have knowledge of ourselves as things in themselves the spontaneity of the understanding cannot be absolute spontaneity. In the secondary literature, there is a huge debate about the question of until when Kant still believed that we can prove theoretically and therefore can know that we

49 Cf. CpR B 67 ., B 155–159; B 409, B 429 .

170

Grüne

are absolutely spontaneous or transcendentally free.⁵⁰ Thus, whether one thinks that the aforementioned argument is a good one depends on how one falls out on this debate, which is something I will not do in this paper. What I nd interesting, however, is that Allison and Willaschek, who argue that the spontaneity of the understanding is absolute spontaneity, both believe that according to the Critique of Pure Reason we do not know whether in thinking we are absolutely spontaneous.⁵¹ They claim that “qua engaged in cognition”⁵² we have to conceive of ourselves as being absolutely spontaneous, but that this is compatible with ourselves in fact not being absolutely spontaneous. Thus for them, Kant only gives a practical but no theoretical proof for the absolute spontaneity of the understanding. I do not see how one can grant that we might not be absolutely spontaneous and still claim that for Kant the spontaneity of the understanding is absolute spontaneity. A er all, in the Transcendental Analytic Kant does not say that in order for us to be able to conceive of ourselves as cognizing or thinking beings, we have to conceive of the understanding as being spontaneous, or that we cannot but think under the idea of spontaneity. Instead he says that the understanding is spontaneous. It would be very strange if Kant would ground his conception of the spontaneity of the understanding in the Critique of Pure Reason on a practical proof without ever mentioning it, whereas in the Groundwork of the Metaphysics of Morals he explicitly emphasizes that his proof of freedom is only a practical proof which leaves open whether human beings really are free or not.⁵³

3 Acts of the Understanding as Absolute Spontaneous Acts? So far I have argued that Kant’s conception of the spontaneity of the understanding is neutral with respect to the question of whether the understanding is relatively or absolutely spontaneous. But even if this is true, it might still in fact be the case that acts of the understanding are absolutely spontaneous. The claim that

50 Cf. Ameriks, Karl (1982), Kant’s Theory of Mind. An Analysis of the Paralogisms of Pure Reason, Oxford/New York: Oxford University Press, ch. 6; Ludwig, Bernd (2010), “Die ‚consequente Denkungsart der speculativen Kritik‘. Kants radikale Umgestaltung seiner Freiheitslehre im Jahre 1786 und die Folgen für die Kritische Philosophie als Ganze”, in: Deutsche Zeitschri für Philosophie, 58, 595–628; Rosefeldt, Tobias (2000), Das logische Ich. Kant über den Gehalt des Begri s von sich selbst, Berlin/Wien: Philo Verlag, ch. 7. 51 Cf. Allison, Henry (1996), “On naturalizing Kant’s transcendental philosophy”, p. 64, Willaschek, Marcus (2009), “Die ‚Spontaneität des Erkenntnisses‘”, sect. 4. 52 Allison, Henry (1996), “On naturalizing Kant’s transcendental philosophy”, p. 64. 53 Cf. Groundwork AA IV 448.

Kant and the Spontaneity of the Understanding

171

acts of the understanding are absolutely spontaneous would then not be an analytic but a synthetic judgment. This would be the case, for example, if Willaschek was right in claiming that it follows from Kant’s considerations concerning the conditions of giving reasons for judgments in his review of Schulz that acts of the understanding are absolutely spontaneous.⁵⁴ In this part of my paper I will argue that even if acts of the understanding were absolutely spontaneous, this would not imply that they are not caused by prior events. My own personal interest in this matter is that I have proposed a reconstruction of Kant’s theory of sensible synthesis which relies heavily on the idea that sensible synthesis is a process that can be described as a causally determined process.⁵⁵ As I have already pointed out, Kant distinguishes between two di erent kinds of actions, which are absolutely spontaneous in di erent ways: (i) Intelligible actions are absolutely spontaneous, if and only if they are not caused by anything. (ii) Empirical actions are absolutely spontaneous, if and only if they are caused by the acting being as a thing in itself and this causing is not causally determined by anything else. Now, let us assume that acts of the understanding are indeed absolutely spontaneous. We then have to nd out whether they are absolutely spontaneous in the sense of (i) or (ii). Since acts of the understanding are either acts of sensible synthesis or acts of judging or acts of concept formation, we have to determine with respect to all three kinds of acts whether they are intelligible or empirical actions. I will start with sensible synthesis. From the way Kant characterizes the activity of sensible synthesis in the transcendental deduction, it follows that sensible synthesis is an activity that takes

54 Since according to Willaschek in Kant’s review of Schulz he only gives a practical proof of the absolute spontaneity of the understanding, as I have pointed out at the end of the second part of this paper, I do not see why Kant’s review of Schulz should have implications for the question of whether acts of the understanding are absolutely spontaneous or not (cf. Willaschek, Marcus (2009), “Die ‚Spontaneität des Erkenntnisses‘”, sect. 3 and 4). Furthermore, in the review Kant only argues that as judging beings we have to conceive of ourselves as absolutely spontaneous. Thus, Kant’s review of Schulz does not even amount to a general practical proof that insofar as we perform acts of the understanding we have to conceive of ourselves as absolutely spontaneous beings. In other words: If Willaschek was right and it really followed from Kant’s review of Schulz that acts of the understanding were absolutely spontaneous, then it would only follow that some acts, namely acts of judging, were absolutely spontaneous. It would not also follow that acts of sensible synthesis and concept formation would be absolutely spontaneous. 55 Cf. Grüne, Stefanie (2009), Blinde Anschauung.

172

Grüne

place in time. In the A-deduction, for example, he explains why there has to be a synthesis of reproduction in the following way: Now it is obvious that if I draw a line in thought, or think of the time from one noon to the next, or even want to represent a certain number to myself, I must necessarily rst grasp one of these manifold representations a er another in my thoughts. But if I were always to lose the preceding representations (the rst parts of the line, the preceding parts of time, or the successively represented units) from my thoughts and not reproduce them when I proceed to the following ones, then no whole representation […] could ever arise. (CpR A 102; italics mine)

And in the paragraph on the synthesis of recognition he writes: Without consciousness that that which we think is the very same as what we thought a moment before, all reproduction in the series of representations would be in vain. For it would be a new representation in our current state, which would not belong at all to the act through which it had been gradually [nach und nach] generated […] (CpR A 103; italics mine)

Similar passages can also be found in the B-deduction, for example in § 24 where Kant says: We cannot think of a line without drawing it in thought, we cannot think of a circle without describing it, we cannot represent the three dimensions of space at all without placing three lines perpendicular to each other at the same point, and we cannot even represent time without, in drawing a straight line […], attending merely to the action of the synthesis of the manifold through which we successively determine the inner sense, and thereby attending to the succession of this determination in inner sense. (CpR B 154; italics mine)

According to all these passages, sensible synthesis is an activity that takes place in time. Since all temporal properties are properties things only have as they appear to us, sensible synthesis is not an intelligible but an empirical action. Therefore, if sensible synthesis is absolutely spontaneous, it is absolutely spontaneous in the sense of (ii) and not of (i). The reason why sensible synthesis in the transcendental deduction is characterized as an act that takes place in time is, I think, the following: In order for sensible synthesis to take place, the human being has to be causally a ected by other things. Sensibility reacts on this causal a ection by presenting sensations. Since time and space are the forms of sensibility the content of these sensations is temporally and spatially structured. In the case of inner sense things get more complicated because time as the form of inner sense is not only responsible for the content of representations of inner sense being temporally structured but also for the fact that these representations themselves appear in temporal form, namely

Kant and the Spontaneity of the Understanding

173

as following one another. It is these inner representations, appearing to the subject as following each other, that are synthesized by the understanding. And it is because the representations appear to the subject in temporal form that the act of synthesizing them takes place in time as well, i.e., is an empirical action. This does of course not mean that this temporal act of synthesis is something that has its temporal structure in itself – this would be atly inconsistent with transcendental idealism. Rather one should suppose that the empirical act of synthesis, which takes place in time, is grounded in some kind of activity or state which pertains to the subject in itself and therefore is atemporal. This activity or state, however, is clearly not what Kant is talking about when he introduces sensible synthesis and characterizes it in temporal terms. The same is true of acts of judging and concept formation. Acts of judging, insofar as we are conscious of them, are given to us in inner sense⁵⁶ and thus appear to us in temporal form. Thus, all acts of judging we know of are empirical actions and can only be absolutely spontaneous in the sense of (ii), not in the sense of (i). Kant says only very little about concept formation, but it is plausible to assume that if acts of judgments are only given to us in inner sense, the same is true of acts of concept formation. If this is true, then acts of concept formation appear to us as taking place in time, therefore are empirical acts and can be absolutely spontaneous only in the sense of (ii). Now, it is important to note that, if an act is absolutely spontaneous in the sense of (ii), this does not imply that it is not caused by a prior event. Kant’s fundamental idea for resolving the Third Antinomy is that if one accepts Kant’s transcendental idealism, then one and the same empirical action can be absolutely spontaneous and also be caused by prior empirical events or states. Thus, from the fact that an empirical action is absolutely spontaneous one cannot infer that it is not caused by a prior event or state. The inference from the absolute spontaneity of an action to its being not caused by a prior event is only justi ed if the action is an intelligible action, but it is not justi ed, if the action is an empirical action. This is why the reasoning of those philosophers who use the assumption that the understanding is absolutely spontaneous in order to argue against a causal conception of the acts of the understanding is fallacious: From the assumption that acts of the understanding are absolutely spontaneous they infer that they are not causally determined, even though they conceive of the acts of the understanding as empirical actions taking place in time.⁵⁷

56 Cf. CpR A 358 and Rosefeldt, Tobias (2000), Das logische Ich, pp. 214–219. 57 Allison, for example, tries to show that the spontaneity of the understanding is absolute spontaneity by arguing that for Kant acts of judging presuppose that the judging subject recognizes

174

Grüne

As I have already pointed out, Kant nowhere claims that the understanding is absolutely spontaneous, and, given his contention that we know ourselves only as appearances but not as things in themselves, he also should not make this claim. What he could say (though as far as I know nowhere does) is that acts of the understanding might be absolutely spontaneous in the sense of (ii). In this case, Kant’s conception of sensible synthesis would be parallel to his conception of empirical actions in the Resolution of the Third Antinomy. I will end my paper with a short and very speculative remark about how Kant could conceive of the act of sensible synthesis if he believed that sensible synthesis might be absolutely spontaneous in the sense of (ii). As we already know, according to the Resolution of the Third Antinomy, human beings as practical beings are causes of their empirical actions in virtue of having an empirical character. This character consists in dispositions to carry out certain actions under certain circumstances, and it is possible to give a causal explana-

reasons for the judgment as reasons and that such recognition is something the subject cannot take itself as being caused to do. He writes: “[U]nderstanding that p requires not merely arriving at one’s belief that p by the correct causal route, that is, having this belief produced by the appropriate set of prior beliefs; it also requires taking or recognizing these prior beliefs as warranting the belief that p […]. [T]his recognition cannot, in turn, be analyzed simply as the possession of a further, second order belief […]. Now, it is precisely because recognition cannot be understood simply as having another belief […], that it is not something that one can coherently take oneself as caused to do, not even by the internal state of the ‘system’ (as the conception of relative spontaneity would have it). […] Since this [conceptual recognition] is an act that the subject must perform for itself (self-consciously) rather than a cognitive state in which it nds itself, it follows that we must assume an absolute spontaneity and not merely a relative spontaneity in order to conceive of its possibility” (Allison, Henry (1996), “On naturalizing Kant’s transcendental philosophy”, in: H. Allison: Idealism and Freedom, Cambridge: Cambridge University Press, p. 63 f.; italics mine). Here, judgments or beliefs clearly are conceived of as states or acts of human beings as appearances and not as things in themselves. Otherwise Allison could not talk of our recognizing prior beliefs as being reasons. (Even though in the passage I have quoted Allison only uses the expression ‘belief’, it follows from the context of the passage that he is talking about judging (cf. p. 61 f.). He does not seem to distinguish between beliefs and judgments.) Furthermore, for Allison judging or believing consists partly in performing an action which one has to conceive of as something one is not caused to do, namely the act of recognizing reasons as reasons. This is exactly the reason why according to him one has to assume an absolute spontaneity in order to conceive of the possibility of believing or judging. Thus, absolute spontaneity is rstly ascribed to empirical states or actions, and it secondly consists in the property of not being caused by a prior event. Clearly, if the understanding were spontaneous in this sense, then acts of judging could not be caused by prior events. But as I have already pointed out this is not Kant’s conception of absolute spontaneity from the Resolution of the Third Antinomy. The same objection applies to Willaschek’s way of showing that the spontaneity of the understanding is absolutely spontaneous (cf. Willaschek, Marcus (2009), “Die ‚Spontaneität des Erkenntnisses‘”, 165–183).

Kant and the Spontaneity of the Understanding

175

tion of why a human being has the empirical character it has. Still, we can also regard the empirical character as being determined by the human being’s intelligible character. This is the reason why for Kant it is also true that human beings as things in themselves are causes of their empirical actions. If, in causing an action (in virtue of having a certain intelligible character) the human being as a thing in itself is not causally determined, the empirical action is transcendentally free or absolutely spontaneous. Now, one might apply this model to the case of sensible synthesis in the following way: The empirical character of the understanding consists in our having the disposition to synthesize sensations according to the categories. Here it is important to note that this disposition is the disposition to engage in an activity that takes place in time. Therefore, it cannot constitute the intelligible character of the understanding but has to be its empirical character. Human beings, in virtue of having this disposition together with other causal factors, namely the reception of certain sensations, cause sensible synthesis to take place. As in the case of the empirical character of human beings as practical beings, it might be possible to give a causal explanation of why human beings have this disposition. For example, one might give an evolutionary explanation. Still, we can also regard the empirical character of ourselves as thinking beings as being determined by the intelligible character of ourselves as such beings.⁵⁸ I assume that this intelligible character consists in the constitution of our understanding. This constitution is something our understanding shares with the understanding of all cognizing subjects regardless of what kind of sensibility these subjects have. In the rst part of the B-deduction, Kant argues that all cognizing subjects, regardless of what kind of sensibility they have, as long as their understanding is not intuitive, structure or synthesize their representations according to the categories (§ 21). So all cognizing subjects have the corresponding disposition, namely the disposition to structure or synthesize representations according to the categories. This disposition has to be the disposition to engage in a nontemporal activity, because it is a disposition that could be had by subjects whose sensibility does not have the form of time. This can be shown in the following way: Because of Kant’s transcendental idealism, no activity can in itself have a temporal form. Instead the only way for an activity to have temporal form is to appear as having temporal form. Furthermore, thinking, or rather acts of the understanding, appear only in inner, but not in outer sense. Thus, the only way for acts of the understanding to have temporal form is to appear to subjects in an inner sense, the form of which is time. If this is right, the intelligible character of human beings as thinking beings

58 Here I use ‘thinking’ in a very broad sense, not as a synonym to ‘judging’ but as a generic term for all activities of the understanding.

176

Grüne

consists in the disposition to synthesize sensible representations according to the categories in a non-temporal way. Because our understanding has the intelligible character it has, we organize sensible representations according to the categories. Because time is the form of inner sense, this organization or synthesis takes place in time. Thus, as in the case of ourselves as practical beings, our empirical character depends on our intelligible character, the empirical character of our understanding depends on the intelligible character of our understanding. Therefore, Kant would be justi ed in claiming that in virtue of having the disposition to synthesize sensible representations in a non-temporal way human beings as things in themselves cause sensible synthesis to take place. If this causing itself is not causally determined by anything else, sensible synthesis is transcendentally free or absolutely spontaneous. If, by contrast, it is causally determined, then sensible synthesis is only relatively, but not absolutely, spontaneous. In both cases, it would be true that sensible synthesis is caused by a prior state, namely by the state of receiving certain sensations. Thus, regardless of whether sensible synthesis is only relatively or also absolutely spontaneous, it is an activity that is causally determined by prior states or events.

Ulrich Schlösser

Concept Formation, Synthesis and Judgment Kant’s Theory of the Logical and Cognitive Activities of the Mind 1 Kant’s Logic as a Theory of the ‘Actions of the Understanding’ In the 20th century, Kant’s way of interrelating theory of cognition with metaphysics was considered a particularly promising part of his philosophy. According to the Kantian approach, we take up a re ective attitude and refer to our own cognitive apparatus – we refer in particular to space and time as forms of intuition and to the categories. On this ground, one can argue for the following: No available object of reference can ever have features that contradict the constraints of this very apparatus. The considerations on logic and semantics underlying this project, in sharp contrast, were generally considered to be of little more than historical interest. There seemed to be no way of modernizing logic and semantics based on Kant’s own presentation of them. The reason for this widespread assessment is not only the fact that the traditional logic we nd in Kant is considered to be a less powerful tool than the system of mathematical logic a er Frege. For at least parts of Kant’s presentation seem to attribute a di erent status to the logical sphere than we do nowadays – one that is entirely unfamiliar to us: According to Kant, the central topic of logic is the ‘actions’ of the understanding and their principles.¹ Since we are dealing with actions, we can highlight the unfamiliar status of this logical sphere by drawing on a parallel to ethics. The actions governed by the principles of pure logic correspond to the actions of the pure “holy” will Kant discusses in the second chapter of the Groundwork of the Metaphysics of Morals.² The principles of logic spell out what the understanding does under ideal

1 On the historical background of this conception of logic, see Reimarus, Hermann Samuel (1756): Vernun lehre (Nachdruck 1979 München: Hanser), and also Wol , Michael (1995): Die Vollständigkeit der kantischen Urteilstafel (Frankfurt: Klostermann), 22 . On the Kantian term ‘action’ (Handlung) more generally, see Gerhardt, Volker (1986): “Handlung als ein Verhältnis von Ursache und Wirkung”, in: Prauss, G., ed. (1986): Handlungstheorie und Transzendentalphilosophie (Frankfurt: Klostermann). 2 AA 4, 414. Apart from the Critique of Pure Reason, references to Kant are to the volume and page of Kants gesammelte Schri en, ed. by the Deutsche Akademie der Wissenscha en, 1900 . I follow mostly, though not always, the translations in the Cambridge Edition of the Works of Immanuel Kant. In particular I refer to: Immanuel Kant (1998): Critique of Pure Reason, transl. by P. Guyer and A. Wood (Cambridge: Cambridge UP); Immanuel Kant (2000): Critique of the Power

178

Schlösser

circumstances, that is, if nothing interferes with its activity. Adding empirically given, psychological limitations such as the limits of our memory or the in uence of mere imagination corresponds to adding merely subjective inclinations in the domain of ethics. In the case of logic, one thereby derives what Kant calls “applied logic”. In relation to applied logic, pure logic acquires a normative status.³ NeoKantians have argued on the grounds of the pure and normative status of logic that Kant’s theory does not fall prey to psychologism.⁴ Nevertheless, it is still true that Kant’s logic makes, at least as a hypothesis, an existence claim, namely the claim that there is an understanding the operations of which are the subject of logic. In contrast, the 20th -century, post-psychologistic approach to logic denies that logic makes any existential presuppositions.⁵ Furthermore, the activities Kantian logic is dealing with are probably to be understood as mental activities. Examples are the acts of subordinating or combining representations, and also the act of positing. However, this may imply that mental acts (such as subordination) and logical relations (such as the subject and predicate positions) are not distinguished clearly enough. In addition, the following holds: If the operations in question are mental operations, then ‘inner’ representations – Vorstellungen in German – will be the vehicles of logic. As Frege has argued, Vorstellungen are private.⁶ It is not possible to compare these representations if they belong to di erent persons; nor is it possible that di erent persons exchange these representations. On the other hand, Kant does not introduce anything like Fregean “senses”⁷ which would not be identical with either psychological states or physical objects. How can the philosophically minded reader of Kant respond to these concerns? Is it recommendable to cut Kant’s considerations on logic and semantics out of his project of combining metaphysics and the theory of cognition? Should

of Judgment, transl. by P. Guyer and E. Matthews (Cambridge: Cambridge UP); Immanuel Kant (1992): Lectures on Logic, transl. by M. Young, (Cambridge: Cambridge UP); Immanuel Kant (1997): Lectures on Metaphysics, transl. by K. Ameriks and S. Naragon (Cambridge: Cambridge UP); Immanuel Kant (1997): Prolegomena to Any Future Metaphysics, transl. by G. Hat eld (Cambridge: Cambridge UP) 3 See B 78/A 54 4 On the arguments of Windelband, Rickert and Kroner, see Kusch, Martin (1995): Psychologism (London: Routledge), 64 .; Husserl o ers a critical discussion of this claim in Husserl, Edmund (1900/1901): Logische Untersuchungen (Halle: Niemeyer), A 26 . 5 An early version of this claim can be found in Husserl 1900/1901 Logische Untersuchungen: A 69. 6 Frege, Gottlob (1918): “Der Gedanke”, in: Frege, G. (1993): Logische Untersuchungen (Göttingen: Vandenhoeck), 43 . 7 See Frege, Gottlob (1892): “Über Sinn und Bedeutung”, in: Frege, G. (1962): Funktion, Begri , Bedeutung (Göttingen: Vandenhoeck).

Concept Formation, Synthesis and Judgment

179

we replace these considerations with more viable contemporary theorems on logic and semantics? A brief example drawn from Kant’s justi cation of the category concepts illustrates why this cannot be done easily. Most interpreters⁸ agree that the central step of the justi cation is the following claim: If it is possible for me to ascribe thoughts to myself, then these thoughts must be such that they can (and regularly do) relate to objects in the sphere of appearances (and not just to mere intuitions). However, this step considered in isolation does not tell us anything about the categories. Kant introduces the categories in another, separate step called the “metaphysical deduction”, as opposed to the “transcendental deduction”. Only the latter deduction deals with the relation between self-consciousness and object reference. In the metaphysical deduction Kant is tying the categories, understood as principles of transforming intuitions into representations of objects, to the principles of logic: The same understanding, therefore, and indeed by means of the very same actions through which it brings the logical form of a judgment into concepts by means of the analytic unity, also brings by means of the synthetic unity of the manifold in intuition in general a transcendental content into its representations, on account of which they are called pure concepts of the understanding […] (A 79/B 105).⁹

It is important to note that starting from the empty logical form of judgment no content can be derived. Thus, from the mere logical form one cannot derive the category concepts which determine the structure of the objects we experience. However, this means that the types of actions mentioned in the above quotation and the capacity underlying both cases of action must play a crucial role in the overall argument. If this is true, then Kant’s deduction depends on his account of logic understood as a theory of the actions of the understanding. Even Kant’s argument in the deduction can’t be spelled out without his unfamiliar account of logic. Instead of bypassing Kant’s account of the ‘actions of the understanding’ I will try in this paper to assign a constructive role to it. In a preliminary consideration in the second part of the paper I will o er an outline of how self-directed transcendental arguments can help justify Kant’s use of his mentalistic vocabulary. In the third part of the paper I want to show that concept formation, judgment and the synthetic integration of perceptions are based on di erent actions of the

8 See, for example, Henrich, Dieter (1989): “The Identity of the Subject in the Transcendental Deduction”, in: Schaper, E./Vossenkuhl, W., eds. (1989): Reading Kant (Oxford: Blackwell), and Guyer, Paul (1987): Kant and the Claims of Knowledge (Cambridge: Cambridge UP), 133 . 9 Translation changed by U. S.

180

Schlösser

understanding even though Kant seems to suggest that this is not the case. In the fourth part of the paper I will outline a few objections to Kant’s account. It has been claimed that Kant’s account of concept formation is circular. In addition, it seems as if both concept formation and judgment and judgment and the synthesis of intuition mutually presuppose each other. Starting from concept formation, I try to show in parts ve, six and seven of the paper how Kant’s emphasis on the active nature of the understanding contributes to solving these problems.

2 Self-directed Transcendental Arguments and Kant’s Philosophy of Mind The approach I want to sketch relates to Quassim Cassam’s description of selfdirected transcendental arguments. Just as world-directed arguments tell us something about the nature of the world in which our thinking and experience takes place, so self-directed arguments tell us something about the cognitive faculties of the thinking or knowing self. If it is a necessary condition of the possibility of a certain cognitive achievement that our faculties are thus and so, then, given the assumption that the achievement is actual, it follows that our cognitive faculties are thus and so.¹⁰

This starting point needs to be modi ed for the purpose of the paper: Instead of talking about cognitive faculties, I would like to talk about types of cognitive activities. These are characterized in terms of their functions. Let us suppose that the cognitive achievements consist in the fact that objects are accessible to us in a speci c way. The following two examples may illustrate what I have in mind. Both examples yield di erent cognitive activities: a) Several coloured surfaces are attached to a wall opposite to a spectator. At face value, they have no aspects in common; even their colour is not entirely identical. Nevertheless, the surfaces can be given to the cognizing subject as

10 Cassam, Quassim (1999): “Self-Directed Transcendental Arguments”, in: Stern, R., ed. (1999): Transcendental Arguments (Oxford: OUP), 85; Kitcher (1990) follows a similar strategy: Instead of talking of ‘cognitive achievements’ she uses the term ‘cognitive tasks’: “[…] I argue that transcendental psychology analyses cognitive tasks to determine the general speci cations for a mind capable of performing those tasks.” Capacities can then be spelled out “by showing that any faculty that can perform the task at all must meet certain speci cations.” (Kitcher, Patricia (1990): Kant’s Transcendental Psychology (Oxford: Oxford UP), 13).

Concept Formation, Synthesis and Judgment

181

being coloured in shades of blue. Similarly, my environment may present itself as structured according to basic colours even though no two colours given are literally identical. b) The objects are accessible to me as going beyond the mere form of individual spatial extents of quality given right now. If the objects were just given in this form, we could only communicate using utterances such as “now, here: white”. Instead, we are dealing with objects as the underlying entities to which we ascribe properties. We take for granted that an object can be reidenti ed later and also that it is possible to pick out the object from another perspective. Objects are accessible as the bearers of a plurality of properties. They are three-dimensional and temporally extended. The fact that we humans access what is present to us in the ways just described can be contrasted with a minimalistic conception of sense impressions. In the 20th century such a conception has been illustrated by using the model of rays of light hitting the retina.¹¹ 18th -century theories of sense data talk of atomistic (and nominalistically conceived) sensory impressions present in a sequence of what we nowadays would call a mental lm. Let us suppose we start with a minimalistic conception of this type simply because it is the least demanding conception. Then the following holds: The richer and more developed access to the sphere of objects is possible only if certain functions of the mind other than those of the minimalistic sensibility are ful lled. According to Kant, these functions are ful lled (mostly)¹² by the actions of the understanding. In Kant’s own presentation of this reasoning one can nd another element I would like to mention: In addition to the function and the individual case of an activity ful lling this function, there is a rule. The activity has to be informed by this rule; otherwise the activities ful lling the function would just be arbitrary and could break down at any moment. The activities would not ful l the function in a systematic and continuous way. Overall, Kant claims: If we humans possess a cognitive achievement, then a certain function of the mind must be ful lled by our cognitive activities in a well-ordered way; thereby, these activities are speci ed necessarily by certain rules.

11 See, for example, Quine, Willard V. O. (1960): Word and Object (Cambridge, Mass.: Harvard UP), 31. 12 This formulation presupposes that Kant’s complex theory of space and time can be considered as spelling out conditions for minimal sensibility as described above. In addition to the understanding and its activities, Kant also relies on the power of imagination which acts in accordance with the understanding (see, for example, B 151).

182

Schlösser

This Kantian strategy can easily be applied to our examples: If the achievement is the capacity to grasp what two surfaces with di erent shades of colour (thus, prima facie, with two di erent colours) have in common, then there must be an activity which ful ls the corresponding function, that is, the function of providing the general. This is the very same function which also underlies the acquisition and the capacity to master concepts. Thereby, the activity is necessarily directed by the rules of logic that apply to this task. Similarly, suppose the cognitive achievement is to pick out and refer to stable objects in thought. Then, there must be a synthetic activity ful lling the function of integrating the prior visual information in the relevant way, thus providing unity among representations. At the same time, this activity is speci ed by the rules of transcendental logic, because they are the rules “without which no object can be thought at all” (B 87/A 62). “No cognition can contradict it [that is, the transcendental logic as logic of truth; U. S] without at the same time losing all content” (B 87/A 62–63). The last claim is stronger than my argument in favour of a systematic procedure (instead of arbitrariness) suggests. This is the case because the relevant achievement – thinking about stable objects – already demands that the activity cannot ful l the function other than systematically. Thus, Kant does not address the issue of the mental and potentially private character of our cognitive activities by re-arranging the relationship between processes in the inner realm and external processes. Neither is it crucial to emphasize that some of the activities also have an external use; this is the case with concept application in linguistic communication. Instead, Kant relies on a modal concept, in particular on the concept of a conditional necessity. If we humans have certain cognitive achievements, then certain functional roles must be ful lled. This implies that speci c rules are e ective in relation to our mind. The cognitive activities can then be characterized as the very processes that instantiate these necessary conditions. This way of justifying claims about our cognitive apparatus is still distinct from the core arguments of the rst Critique, such as the deduction argument. In the reasoning I have suggested here, object reference is taken for granted. The assumption that we refer to objects is understood as a contingent premise without further justi cation. In contrast, the deduction argument tries to show that object reference is implied in the possibility of ascribing thoughts to oneself.¹³

13 This di erence depends on the claim that a strong, anti-sceptical reading of the deduction can be given. If one gives up this reading, the di erence between the self-directed arguments suggested here and the deduction itself is less obvious. (See Guyer’s way of defending the deduction against the charge of psychologism in Guyer, Paul (1989): “Psychology and the Transcendental Deduction”, in: Förster, E., ed. (1989): Kant’s Transcendental Deductions (Stanford: Stanford UP). In this context it is important to note that by tracing back psychologism to Hume

Concept Formation, Synthesis and Judgment

183

Finally, it is worth mentioning that Kant’s reference to logical and cognitive activities is not based on introspection – nor does his account merely put features forward which also would be available by introspection and thus independently of Kant’s own methodological procedure. For, (1) inner sense, too, provides us only with appearances; in addition, he calls the consciousness of our activities frequently weak or dark.¹⁴ (2) Conditions for the possibility of representing need not be part of what is represented. (3) Introspection cannot ground the conditional necessity Kant is looking for. And (4) Kant does not describe the activities as we experience them but by characterizing their functional role.

3 Do All Acts of the Understanding Belong to the Same Type? Let us suppose that Kant can justify taking up logical and cognitive activities in the rst place; thus let us suppose that the strategy using self-directed transcendental arguments is successful. The obvious next step is to analyze how Kant spells out his model of these activities. The Kantian classi cation of logic is derived from the 18th -century tradition. The classi cation distinguishes between the acts of concept formation, judging and inferring. Here, I will rely mostly on his account of concept formation and judging. In addition, I will refer to the activity of synthetically integrating the manifold of representations on the level of intuitions. This activity is not a topic of the Kantian logic, yet it is at the centre of his theoretical philosophy. The activity of synthetic integration is meant to allow for the fact that mental representations can represent complex, three-dimensional and temporally extended objects. Since these objects are also the point of relation of our logical states, the activity of synthesizing is tied to our logical activities as well. The rst question I would like to address is the following: Are we really entitled to speak of a plurality of logical and cognitive activities? Or is Kant’s claim rather that we are ultimately just dealing with one uniform type of activity of the understanding? There are di erent ways in which the three processes emphasized could all belong to one and the same type of activity. The rst option is that the same activity of the understanding is processed on di erent cognitive levels. This could be the case with the activities of synthesis and of judging. The rst one operates at the level of intuition whereas the second one relates to concepts, that is, to gen-

and Tetens, Guyer identi es psychologism with empiricism, thus ignoring the issues of privacy and individualism.) 14 See, e.g., A 103–104.

184

Schlösser

eral representations. By claiming that there is just one uniform type of activity Kant would claim the following: The activity of synthetically integrating the visual information I receive in a sequence of perceiving my environment would be the same as the activity involved in judging. Another option to interpret the claim that there is just one uniform type of activity is the following: The very same activity could be described from two different angles (both related to the same cognitive level). This could be the case with the activities of concept formation and of judging. From one angle, the activity could be described as the process of judging, which consists in attributing a general mark to an object. From the other angle, the process would constitute a representation as a general representation in the rst instance. Then, the activity of judging would be identi ed with the act of forming the concept that partakes in the judgment. This way of spelling out the di erent options shows that they amount to two di erent claims. There is evidence for both claims in Kant. I have already mentioned Kant’s claim that “the very same actions through which it [the understanding; U. S).] brings the logical form of judgment into concepts” are also e ective in establishing “the synthetic unity of the manifold in intuition in general” (B 105/A 79). In addition, Kant writes in the chapter on “The Logical Use of the Understanding in General”: “We can […] trace all actions of the understanding back to judgment, so that the understanding in general can be represented as a faculty for judging” (B 94/A 69). This suggests that the other logical actions – namely concept formation and inference – can be spelled out in terms of judgment, which corresponds to the second of the two claims mentioned above. Thus, overall, we would be dealing with one uni ed type of action: judgment. This would be the key to Kant’s philosophy of mind. However, there are objections to both claims. First, I would like to consider the claim that the very same activity that establishes the logical form of judgment also synthesizes the manifold at the level of intuitions. In this context, it is important to note that the logical form of judgment is primarily based on the activity of subsuming. Suppose I judge on the ground of the mark (Merkmal in German) that this small, red, insect-like animal has eight legs that it is an arachnid. In this judgment, the concept of a mite (the animal we are dealing with) and the concept of arachnid are related by subordination; the generality of the concept ‘arachnid’ is potentially larger than the generality of ‘mite’. This claim can also be formulated in a corresponding way in terms of the marks the judgment represents the object as having. The marks of being insect-like and having eight legs are subordinated to the mark of being an arachnid; according to Kant, marks are subordinated “insofar as one mark is represented in the thing only by means of the other” (AA 9, 59).

Concept Formation, Synthesis and Judgment

185

When we run through the given material and thereby synthesize our representations, the underlying relations are di erent. In this case, the representations are at the same level of generality. If one concedes that the term ‘mark’ can be applied at the level of intuition at all, then we are dealing with coordinated marks. The front and the rear side of the animal would be an example for features of this type.¹⁵ Given the di erence one may wonder whether it is indeed the same activity which establishes both relations.¹⁶ Identifying the activity of concept formation with judging is similarly problematic. This becomes obvious once Kant introduces the distinction between reection and determination.¹⁷ Re ection is a logical activity that ascends from the individual to the general. Determination proceeds in the opposite direction, thus applying the general representation either to an object or to the intuition corresponding to it. Like our aesthetic response to the object, concept formation centres on re ection alone. In judging, in turn, the issue of how I want to determine the object is dominant – on this topic, see also section 5, subitem b. (p. 192). Thus, both claims underlying Kant’s thesis that there is just one uniform type of the activity of the understanding turned out to be controversial. In the following I will rely on the assumption that there is a plurality of correlated activities, since this is the weaker claim.

15 I assume here that the operation of synthesis as represented in sections 1–3 of the A-deduction (A 98–110) is primarily concerned with representing individual objects and the task of tracking their identity through time. For a di erent reading, see Longuenesse, Béatrice (1998): Kant and the Capacity to Judge (Princeton: Princeton UP), 49. 16 A standard reply to this objection is the following: When we synthesize our representations and thereby relate coordinated marks, we also subsume the bundle we have produced under a concept (e.g., the concept of a substance). However, this claim does not establish that the activities in question are identical. That they still may be di erent can be shown by the following possibility: There may be an empirical judgment which establishes the relationship of subsumption between two concepts without thereby synthesizing the coordinated marks of the underlying object – apart from those represented in the subject concept. Performing the synthesis would be di erent from this particular act of judging. Therefore, both activities can still be distinct even if they turn out to be correlated. A negative reply to this question obviously a ects the success of the “metaphysical deduction”: Is the claim that merely the same faculty and its norms (and not also the same actions) operate both at the intuitive level and the judgmental level still su cient to introduce the transcendental content and thus the categories? 17 AA 20, 211, and AA 5, 179.

186

Schlösser

4 Five Challenges to Kant’s Model of Concept Formation, Judgment and Synthesis In the next step I want to look more closely at the Kantian descriptions of the activities in question. I want to point out that there are internal problems in how, according to Kant, the activities are supposed to accomplish their task. In addition, there are tensions in Kant’s description of the relationship between di erent activities. Overall, I want to point out ve problems for Kant’s theory. Since Kant’s account of concepts plays a crucial role in formulating these problems I will begin by analysing this account. The Jäsche-Logic explains the activity of concept formation in the following way: To make concepts out of representations one must be able to compare, to re ect and to abstract, for these three logical operations of the understanding are essential and universal conditions for generation of every concept whatsoever. I see, e.g., a spruce, a willow and a linden. By rst comparing these objects with one another I note that they are di erent from another in regard to the trunk, the branches and the leaves, etc.; but next I re ect on that which they have in common among themselves, trunk, branches and leaves themselves, and I abstract from the quantity, the gure, etc., of these; thus I acquire a concept of a tree. (AA 9, 94)

Hannah Ginsborg¹⁸ has outlined two internal problems in Kant’s account of concept formation. The rst (1) objection deals with what is o en considered to be the standard case of concept formation: We start with a given plurality of internally complex objects of everyday life. The willow, the spruce and the linden mentioned in the Jäsche-Logic fall under this category. An e ort to explain how concept formation proceeds from this starting point faces the following crucial question: What accounts for the fact that we take up this particular set of marks in order to form the concept rather than another set of marks the objects also have in common? The issue of which marks we take up a ects the application of the concept in the future. If we choose a di erent set of marks, di erent applications in the future may turn out to be correct, even though both sets of marks are present in the examples given right now. Why do we refer to leaves, branches and the trunk, but ignore for example the size, which could potentially exclude small Japanese Bonsai trees?

18 Ginsborg, Hannah (2006): “Thinking the Particular as Contained under the Universal”, in: Kukla, R., ed. (2006): Aesthetics and Cognition in Kant’s Critical Philosophy (Cambridge: Cambridge UP), 39 . See also Pippin, Robert (1982): Kant’s Theory of Form (New Haven und London: Yale UP), 112 .

Concept Formation, Synthesis and Judgment

187

Why do we abstract from the wooden material that trees have in common with other objects, including artefacts made of plants? Why don’t we include the ivy growing on the tree or the surrounding earth it is growing in? Apparently a given object does not x the relevant concept that allows the classi cation of other objects as being of the same type. If one object cannot establish what the relevant concept is, a plurality of objects cannot establish this either. A straightforward response to this di culty is to claim that in seeing the linden, the spruce and the willow I see them as trees. However, this response seems to imply that the concept of a tree is already in some way available to me. According to the second (2) objection the following holds: I can form the concept of a tree on the basis of the marks having leaves, having branches and having a trunk only if I have the capacity to recognize leaves, branches and trunks as common features even though they di er considerably in each case in their shape, colour and so forth. Thus I have to grasp these diverging features as leaves, as branches and as the trunk. In order to do so, I must already be in possession of the concept of these features. If this is true, the Kantian model could only explain how I form a new concept if I already possess some other relevant concepts. The model could not explain how we manage to generate (empirical) concepts in the rst place.¹⁹ In addition to these internal problems in Kant’s theory of concept formation, there is also a considerable tension between this theory, on the one hand and Kant’s description of judgments and the role concepts are meant to play in judgments, on the other. The following question (3) arises: How can the theory of concept formation I have sketched be related to Kant’s more modern claim in the Critique of Pure Reason that concepts are always “predicates of possible judgments” (B 94/A 69)? This claim suggests that Kant could be read as a forerunner of Frege and the primacy of judgments over concepts Frege argues for. Kant’s identi cation of concepts with predicates of possible judgments seems to suggest that a representation quali es as a concept only insofar as it does play this role, that is, only if it enters into a combination of representations which includes a relationship

19 Ginsborg’s own approach to solve these problems relates to Hume. According to Hume, concepts are based on our having the capacity to call objects to mind which are similar to those that I represent right now (Ginsborg “Thinking the Particular”, 44 .). Unlike Hume, Ginsborg suggests that we take our modes of association as having normative signi cance (Ginsborg “Thinking the Particular”, 51). In this context, she introduces the idea of a primitive normativity. Her approach seems to imply that all the rules governing concepts acquire the status of primitive normativity. In the following, I will suggest a reading which attributes primitive normativity only to the rules of rationality. Discussing Ginsborg’s highly innovative approach in detail is beyond the scope of this paper. Doing so would demand a sketch of her reading of the Critique of Judgment as well.

188

Schlösser

of subordination. If this claim implies that the concept in its role as a predicate always already relates to another concept (namely the concept in the subject position), then a genuine ascent to the sphere of thought does not exist. The ‘modernist’ strand in Kant’s theory could be paraphrased as follows: Insofar as we are dealing with a cognitively responsible subject, the propositional level has always already been realized. It is impossible to step behind this point of view even in hypothetical considerations. A possible weaker position is to claim that there is an ascent to the sphere of thought, but the ascent only takes place in judgment. This claim still implies that there is no separate theory of concept formation as the Jäsche-Logic suggests. (4) The question arises of how exactly concepts are correlated to the synthesis underlying intuition. According to Kant, this synthesis plays a crucial role in the explanation of how intuitions are intentionally directed at objects. The question does not primarily concern the well-known relation between category concepts and synthesis. Rather, the focus is on empirical concepts such as the concept of a tree in Kant’s example. On the one hand, Kant’s description of how we generate an empirical concept seems to presuppose that we already refer to objects given in a well-ordered way – for example to the spruce, the linden and the willow, or to three arachnids. If the description presupposes reference to these objects, then it also presupposes the synthesis involved in representing these objects. On the other hand, one may wonder whether the synthesis can yield the representation of objects which are both stable and have clearly delineated empirical boundaries if one does not take for granted that the empirical concepts are already available; it seems as if the empirical concepts already have to be applied in the synthesis. Once again, considering the role of the concepts in the genetic process reveals a circularity problem. Finally, there is a h (5) critical question. This question refers to the relationship between judgment and synthesis. According to Kant, judgment and synthesis are also interlocking. In order to understand the h question one has to bear in mind that an essential feature of judgment is articulating a relation to objectivity. Judgment, therefore, is distinct from a way of combining representations which merely articulates a subjective impression of how something appears to me individually: “That is the aim of the copula is in them: to distinguish the objective unity of given representations from the subjective” (B 141–142). Kant’s standard example is the sentence, The body is heavy. As is well known to readers of Kant, the reference to objectivity articulated in a judgment is grounded on the synthesis of intuition, insofar as this synthesis itself is tied to the unity of apperception as a source of objectivity. Judgments are then “nothing other than the way to bring given cognitions to the objective unity of apperception” (B 141). But is it conceivable that the relation to the objective sphere provided by the synthesis has to be

Concept Formation, Synthesis and Judgment

189

established in each case of judgment? Do I perform the synthesis by judging? This question is particularly important if one takes the presentation of judgment in the Jäsche-Logic into consideration as well. According to the Jäsche-Logic, a judgment primarily maps a speci c combination of concepts – it is a “representation of a relation among them” (AA 9, 101). However, we know already that the concepts which are supposed to be combined in a judgment presuppose an underlying synthesis. Thus it is impossible that the synthesis is established by the judgment in the rst place, as Kant seems to claim. These ve critical worries provide reasons for focusing on Kant’s theory of concepts despite Kant’s own emphasis on the primacy of judgment among the logical and cognitive activities. For all the problems mentioned relate primarily to the role of concepts and to their generation. Thus in the following, I will pursue the unusual strategy of reconstructing Kant’s account of the activities of the mind starting from the issue of concept formation.

5 What xes which Marks we must select in order to form the relevant Concept? A few E orts to solve the rst Problem a. Longuenesse’s reference to schemata as mediating between concepts and perception I will begin by discussing two prominent e orts to solve the rst problem. Béatrice Longuenesse’s e ort is the most well known.²⁰ It has been frequently discussed. According to Longuenesse, the comparison at the beginning of concept formation cannot be a comparison of perceptions, that is, a comparison of representations which represent individuals. Of course, a comparison of perceptions is possible. However, Longuenesse argues that it cannot lead us to the level of determinate concepts. The comparison in question cannot be a comparison of concepts either, even though this procedure is also familiar to Kant. In the process of thought, a comparison of concepts precedes the endorsement of a judgment and thus the formation of a belief. In order to understand concept formation, an intermediate level, a level between mere perception and concepts, is needed. Within the Kantian apparatus, Longuenesse identi es this level with the level of schemata. This approach has been criticized from di erent directions. One important problem has been pointed out by Hannah Ginsborg.²¹ It seems as if Longuenesse

20 Longuenesse Capacity to Judge, 107 . 21 Ginsborg “Thinking the Particular”, 41.

190

Schlösser

ultimately has to accept the claim that schemata can only be schemata insofar as we compare them, whereby the very objects of the comparison are nothing other than the schemata themselves. Therefore, Longuenesse’s solution to the rst problem is itself circular. I want to focus on a more basic question. I want to nd out what the reference to schemata involves and in what sense schemata can be objects of comparison at all. First, a preliminary remark: Compared to Kant’s own presentation of the topic, Longuenesse has to reverse the role schemata play as the middle ground between concepts and the perceptual access to an individual. Kant takes for granted that we already possess the concept. Given this starting point, schemata are conceived of as a representation of the procedure whereby an intuition is assigned to the merely intellectual concept.²² Longuenesse, in contrast, has to assume that schemata can generally be available even without the relevant concept. The function they are meant to ful l points precisely in the opposite direction. Schemata are supposed to contribute to forming a concept in the rst place, given that we have information available at the level of intuition. At the same time Longuenesse agrees with the Kantian claim that schemata cannot be located at the level of mere sensibility. One may wonder whether the intermediate position of schemata is even possible if we do not already possess the relevant concept. In addition, schemata cannot reasonably be conceived of as entities which are general in just the same way as concepts are general. According to this approach, the only di erence between schemata and concepts would be that schemata are immediately involved in perception. By focusing on the role Kant attributes to the schemata one can see why this approach must be wrong. Suppose, we follow the top-down direction as Kant himself does. Then, a line of thought becomes salient that is similar to Kant’s reasoning about the power of judgment in the passage preceding the schematism chapter.²³ Suppose applying a concept or providing an image in intuition for it would again demand a mental item of a similar type as the concept. Then the gap between a general representation and its image would open up again; this time as a gap between the schema and the instance that it is supposed to provide. The heterogeneity of the intellectual and its exhibition would still be present. The heterogeneity causes the problem, not the schemata’s status of being either implicit or explicit. In Kant there is a tendency to cover up this problem (as opposed to solving it). This is the case when Kant calls a schema “a representation of a general procedure of the imagination for providing a concept with its image” (B 179–180/A 140). Certainly, the relation between the rep-

22 See B 179–180/A 140. 23 B 171–172/A 132–133.

Concept Formation, Synthesis and Judgment

191

resented procedure in its generality and an individual case of its application is an issue here.²⁴ On the other hand, schemata cannot be conceived of as sample images either. If we go back to the case of the tree, a sample image would be a standardized image of a tree in which all the speci c aspects of willows, spruces and lindens would have been dropped. This, however, would still be an image of an individual tree – an image of a boring tree which probably does not exist in reality. Due to its twofold di erence to both the concept and the image, the schema is a procedure (or a representation of a procedure) which cannot be reduced to any particular content of either thought or perception. The implicit character follows as a conclusion from this reasoning. This is why Kant calls the “schematism of our understanding […] a hidden art in the depth of the human soul, whose operations we will hardly ever divine from nature and lay before our eyes” (B 180–181/A 141).²⁵ Once again Kant is drawing on the parallel to the power of judgment for the exercise of which no general rule can be given. Understood this way, it is doubtful whether schemata can reasonably be considered as objects of comparison at all as Longuenesse suggests. This is particularly doubtful given the fact that the concepts corresponding to these schemata are not yet supposed to be available.²⁶

24 The same criticism applies to Johannes Haag’s interpretation of schemata. According to Haag, schemata of empirical concepts are again concepts. See Haag, Johannes (2007): Erfahrung und Gegenstand (Frankfurt: Klostermann), 287. 25 Translation changed by U. S. 26 I read the central evidence in favour of Longuenesse’s approach in a di erent way than she does. It is Re ection 2880: “We compare only what is universal in the rule of our apprehension. For example, one sees a sapling, so one has the representation of a tree; an elongated rectangle makes one think of a square [gibt Anlaß zum Quadrat]. A unicorn is a horse, where the horn has been taken from other animals” (AA 16, 557, transl. by Longuenesse). It is helpful to relate this passage to other re ections nearby in order to understand its purpose better. In an amendment to Re ection 2876 (AA 16, 556), Kant raises the following question: “Is it possible to possess a concept as repraesentationem communem prior to and without comparison with others?” His answer to the question is, “Comparison is not always necessary in order to get a general concept, but instead the consciousness that the representation is possible in various ways” (transl. by U. S.). A straightforward way to account for this possibility is the following: Given one representation, we can access another representation nearby in the realm of possibilities by means of association. This association can provide us with a starting point for concept formation if the association proceeds in such a way that a mark present in the given representation is also taken to be present in the representation we associate with it: “For example, one sees a sapling, so one has the representation of a tree; an elongated rectangle makes on think of a square. A unicorn is a horse, where the horn has been taken from other animals”. The “rule” Kant is talking about is the rule governing the process of association; what is “general in the rule” is the very feature the associa-

192

Schlösser

b. Are we always already at the conceptual level? Sebastian Rödl’s interpretation of concept formation as an example of a propositionalist reading Our reasoning about the issue of concept formation started from the following observation: Suppose three individual objects, e.g., a spruce, a linden and a willow are given to us. In this situation, it is not at all obvious how it happens that we generate this particular concept (tree) rather than another concept – a concept that would structure the present and future input in a di erent way. In addition, the previous line of reasoning has revealed that introducing schemata as mediating between intuitive and conceptual levels does not solve the problem. Claiming that we cannot step behind the conceptual level at all becomes more appealing given the negative results so far. This claim can be based on Kant’s well-known thesis that cognition always has two sources: it is always also conceptual. As Kant famously puts it: Intuitions without concepts are blind.²⁷ It is important to note that the Kantian slogan can be related to the problem of concept formation in general only if the slogan is meant to include empirical concepts as well. In consequence, one has to accept a very ambitious reading of Kant’s thesis: According to this reading, we can access objects via intuitions only if we already possess and apply the relevant empirical concepts (and not just category concepts). Accepting this approach means accepting that there is no genuine ascent from mere intuitions to the conceptual sphere. The three activities of comparing, re ecting and abstracting thus have to be tied to what Kant calls the use of concepts. The use of concepts takes place in judgments. Therefore the interpreter’s task is to read judging in such a way that the actions of concept formation can be identi ed within the process of judging. Sebastian Rödl is among those interpreters who endorse the strategy just sketched. I take up his reading of re ection as a crucial example: According to Kant “general representation” means “re ected representation”. Thus, I start with re ecting. In re ecting I look at a spruce in respect of what it has in common with a linden, a willow and other trees. This means I take the spruce to be a tree. I can make this explicit in judging “The spruce is a tree”.²⁸

However, it is doubtful whether the judgment mentioned by Rödl can be read as primarily expressing re ection in the Kantian sense. One should bear in mind

tion transmits from one representation to the other. The procedure of association is merely part of the psychology involved in the theory of concept formation. However, this psychological process should not be confused with the necessary operations of synthesis, as Longuenesse does. 27 B 75/A 51. 28 Rödl, Sebastian (2001): “Handlungen, die einen Begri ausmachen”, in: Gerhardt, V./ Horstmann, R.-P./Schumacher, R. (2001): Kant und die Berliner Au lärung. Akten des IX. Internationalen Kant-Kongress (Berlin/New York: De Gruyter), 433.

Concept Formation, Synthesis and Judgment

193

that the concept of re ection as a part of Kant’s theory of concept formation ought to be compatible with Kant’s comments on re ection in the Critique of Judgment. In the Critique of Judgment Kant distinguishes the activity of re ection from the opposing movement of determination. Kant characterizes re ecting by using the following formulations: “we re ect on a given representation, according to a certain principle, in relation to a concept which is thereby made possible” (AA 20, 211, emphasis added). According to Kant, the principle of re ection is the claim that “for all things in nature empirical concepts can be found […]” (AA 20, 211, emphasis added).²⁹ Determination, in contrast, starts from an “underlying concept” (AA 20, 211). According to the most prominent descriptions³⁰ it proceeds top-down, that is, in the opposite direction, and aims at the object or the intuition corresponding to it.³¹ Kant emphasizes that it is only in the case of the concept of nature that re ection already implies determination.³² This holds because the concept of nature constitutes the corresponding area of investigation; therefore, it has always already been applied. In all other cases, the activity of re ection which is directed towards a “possible” concept, or a concept that can be “found”, does not yet include the performance of determination. The sample judgment in the quotation from Sebastian Rödl stands primarily for the application of a concept and thus for the process of determination: The concept tree is applied to a spruce. The activity of re ection seems to be both presupposed and by-passed. The judgment “the spruce is a tree” is certainly not a good starting point for making explicit what re ection is.³³

29 Translation changed by U. S. 30 E.g., AA 5, 179. 31 On the ambivalences in Kant’s use of the term ‘determination’, see Allison, Henry (2001): Kant’s Theory of Taste (Cambridge: Cambridge UP), 18–19. 32 AA 20, 212. 33 For similar reasons I also hesitate to endorse Rödl’s account of abstraction and comparison. Since Rödl wants to take into account that in concept formation we perform a transition, he presents abstraction and comparison as syllogisms. His example for abstraction is the following: “(1) The spruce is a tree. (2) Trees have a trunk. (C) The spruce has a trunk” (Rödl “Handlungen”, 433). Once again, the dominant focus seems to be on the determination, in particular on the rule, “What belongs to or contradicts a higher concept also belongs to or contradicts all lower concepts that are contained under those higher ones” (AA 9, 98). Again, this does not seem to be abstraction, because abstraction is concerned with the question of which marks the concept ‘tree’ includes. This seems to be presupposed in the syllogism. Comparison is nally rendered as “(1) The linden has a trunk with such and such characteristics; (2) The spruce has a trunk with such and such characteristics; (3) The willow has a trunk with such and such characteristics; (C) Trees have a trunk” (Rödl “Handlungen”, 434). However, it is hard to see how comparison as focusing on both what objects have in common and on their di erence could be rendered as a syllogism.

194

Schlösser

c. Kant’s solution to the rst and the fourth problems: Synthesis prepares concept formation So far I have reached the following conclusions: Taking up schemata as a means of preparing the formation of general concepts does not solve the rst problem on our list. Nor can I claim (as Sebastian Rödl does) that the three operations of comparison, re ection and abstraction are literally the same as our use of concepts in propositions. Thus, I have to go back to the claim that for Kant there is a genuine ascent from the level of mere sensibility to the level of concepts. This means, however, that the original problem is present again – namely the problem of how we can explain our selecting the relevant marks without circularity. This problem can be addressed by taking up the Kantian claim that a cognitive activity is already e ective at the underlying level of perception. With the support of the power of imagination, the activity is performed by the very same capacity that also grounds the conceptual sphere. The activity, however, can only contribute to solving the puzzle of concept formation if it precedes concept formation. This condition can easily be formulated in Kant’s own words: synthesis precedes analysis. Kant states this assumption in a well-known passage from § 10 of the Critique of Pure Reason: Prior to all analysis of our representations these must rst be given, and no concept can arise analytically as far as the content is concerned. The synthesis of the manifold, however, (whether it be given empirically or a priori) rst 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 uni es them into a certain content […] (B 103/A 77).

Thus, concepts can arise analytically as far as their form, that is, insofar as their generality is concerned, whereas it is impossible that they can arise analytically with respect to their content. The synthesis is the very procedure which “collects the elements for cognitions”. The Metaphysics Mrongrovius adds to this observation: “From pure sensations one cannot make any concepts or communicate them to others […] But one can make concepts from the synthesis of perception” (AA 29, 794).³⁴

Rödl’s second syllogism rather seems to stand for abstraction (if it can be related to one of the three activities at all). 34 In this context one has to bear in mind that the synthesis is closely related to the category concepts. In consequence, my approach implies that category concepts di er from empirical concepts both in status and in the role they play in the cognitive process. I do not think that this implication causes problems. In particular, I do not think that a claim from the Jäsche-Logic, according to which the form of a concept is always produced (AA 9, 93), undermines my approach.

Concept Formation, Synthesis and Judgment

195

This claim leads directly to the fourth problem mentioned on the list: Can the synthetic activity establish su ciently stable and clearly delineated boundaries in the realm of objects without presupposing empirical concepts? Is an activity (implicitly) directed solely by the category concepts in a position to determine which empirical marks belong to each other? In other words: Does the synthetic activity have su cient selective power to allow for a representation of distinct empirical objects? I will outline why an a rmative answer to these questions is plausible. First one has to bear in mind that we are dealing with a synthetic activity guided by the concept of an object in its relation to gurations in space and time. If the cognitions in question represent the object in this way, the following must be true: “insofar as they are to relate to an object our cognitions must also necessarily agree with each other in relation to it, i.e., they must have that unity that constitutes the concept of an object” (A 104–105). Thus, cognition will represent objects as having marks which – once we have conceptualized the marks – cannot yield contradictory judgments. However, this is just a minimal condition. In addition, the demand of uni cation relates particularly to space and spatial features. Therefore, the understanding is directed towards objects which are uni ed in space. These are objects which occupy an enclosed spatial area. They can be composed as having sides which connect to each other. In our access to the world we will not give priority to the referents of what we nowadays call mass terms³⁵ nor to objects which are discontinuous in space (or in time) and therefore disparate. Finally, we have to take into account that the understanding will follow a pattern in its procedure of establishing unity that ful ls basic norms of (economic) rationality.³⁶ The understanding is de ned as the rational capacity. This aspect of the meaning of ‘understanding’ is obvious in the corresponding German term Verstand. The constraints of rationality in question do not primarily apply to one isolated instance of perceiving one’s environment. The constraints of rationality relate to a sequence of perceptual situations and the repetitions of similarities and correspondences they exhibit. The

Even in the case of synthesis, the activity itself comes rst. We have to re ect on this activity in order to form the category concepts as explicitly present representations. On this topic, see B 103104/A 78. In addition, the Logic Philippi claims that the understanding acquires pure concepts “by paying attention to his own procedure on the occasion of experience” (AA 24, 452, transl. by U. S). 35 A standard example is ‘water’. On mass terms, see Quine Word and Object, 91 . 36 Economic rationality is understood here as aiming at a mini-max relationship balancing unity and diversity. Its most explicit treatment already relates to a higher level: Economic rationality underlies the possibility of having a network of concepts.

196

Schlösser

fact that the constraints of rationality³⁷ already shape our perception is the very feature of perception that lays the ground for concept formation. This is particularly true for the act of comparison involved in concept formation and for our capacity to see what two appearances have in common.³⁸ These minimal assumptions taken together imply that integrating certain speci c unities is more likely than others, which in turn informs the process of concept formation. I would like to illustrate this implication by going back to the tree example and the potential discrepancies between ways of integrating. The example was used before in Ginsborg’s rst objection.³⁹ Focussing on the wooden material would unite what is spatio-temporally discontinuous. We would not be dealing with an object having a unity in space and time, but with an object that consists of an assembly of pieces of wooden material on di erent locations taken together all at once. However, given our reasoning, this cannot be what I primarily consider to be a unity in perception. A very small tree with its leaves, branches and trunk, on the other hand, would qualify as such a unity. The criterion of economy matters for this example as well. In the case of wooden material, highly diverse items would be conceived as a unity. But not integrating our representation of a small tree in the same way as we integrate other perceptions of trees in a sequence of perceiving them would render our perception too disparate despite a high degree of similarity in the input. Both cases violate the principles of (economic) rationality. One can similarly explain why we do not include the soil the tree is growing in or the ivy surrounding it in what we perceive as the relevant unity when we face a tree. The con guration of trees is not always connected to these features which again concerns the issue of economy. In addition, the soil is not an enclosed body. These examples show the following: Aiming at a uni ed object, or – formulated in Kant’s own terminology – being guided by the category of substance in relation to space and time does provide the synthetic activity with selective power. The synthetic activity understood this way contributes to our structuring of empir-

37 The constraints of rationality are meant to respond to a worry about a potential gap that might open up if the reading of the deduction would rely only on category concepts. Stefanie Grüne lls this gap by introducing dark empirical concepts as sharing important features with mechanisms of association (Grüne, Stefanie (2009): Blinde Anschauungen. Die Rolle von Begri en in Kants Theorie sinnlicher Synthesis, (Frankfurt: Klostermann), 232 .). The advantage of the constraints of rationality is to be found in the fact that they can be traced back directly to the understanding without introducing additional empirical assumptions. Like Grüne, I assume that the mechanism operates recursively in a sequence of applications. 38 Once we have concepts, the principles of rationality guide us by developing the architecture of our conceptual systems. See Kant’s comments on the ‘principle of re ection’ (AA 20, 211 .). 39 Ginsborg “Thinking the Particular”, 38 .

Concept Formation, Synthesis and Judgment

197

ical objects in perception. The activity allows us to explain why we can perceive something as an enclosed object even if the corresponding empirical concept is not yet available for us – as is the case with the savage who sees a house for the rst time.⁴⁰ If one concedes that the synthesis does play this role, then it follows that all humans operate under the same side constraints. These side constraints are involved in guiding the focus of our attention towards a speci c set of marks in a situation of concept formation. There is, of course, still some room for variation. However, this scope is not bigger than the scope a good theory of concept formation wants to preserve – a er all, di erent people do to a certain extend generate di erent concepts in corresponding situations.⁴¹

6 Ascending from the Individual to the General Let us suppose that synthetic activity can provide representations of enclosed objects by proceeding in a way that conforms to the category concepts. Within the process of representing, the synthetic activities provide boundaries by drawing these boundaries. This makes plausible why it is likely that in concept formation we all take up this set of properties or distinctions rather than others. However, this reasoning does not contribute to solving the second internal problem of concept formation. Using the standard example again, the problem can be spelled out in the following way: Kant’s presentation of the concept ‘tree’ presupposes that we already possess the sub-concepts ‘trunk’, ‘branches’ and ‘leaves’. I don’t think we can address this problem by emphasizing (1) that only complex concepts are tied to sub-concepts and (2) that there must be a basic level of simple concepts which have to be explained in a di erent way. For accepting this suggestion implies endorsing a speci c model of how we humans actually generate concepts. One would have to claim that starting from childhood we would, as a matter of fact, proceed from concepts of what is simple to concepts of what is composed. However, this is an implausible empirical claim. It seems more convincing to assume that humans deal from the very beginning primarily with the complex objects of everyday life. They immediately form concepts of these objects. These concepts in turn are themselves internally complex. Kant’s focus on synthesis conforms well to this common-sense approach.

40 See AA 9, 33. 41 It is important to note that so far I have only explained why certain features are uni ed as belonging to one object. Thereby, I have not yet explained the generality of the concept.

198

Schlösser

Thus, the second problem can’t be solved by using the concepts ‘complex’ and ‘simple’. Instead, one has to refer directly to the concepts ‘individual’ and ‘general’. In order to do so I would like to make a background assumption explicit: I do not assume that in perception the characteristics of objects are always already given as general.⁴² We do not in each case perceive the characteristics in objects as being general ones, that is, from the point of view that they allow for multiple instantiation. Given this background assumption, Kant’s position can be summarized in the following way: The situation is not such that I already have to possess the (empirical) concept in order to be able to see what di erent objects have in common. It is precisely the other way around: While seeing – or better: while focusing on what several instances have in common – I have thereby acquired the concept. Kant emphasizes this claim in a bold formulation found in the Metaphysics Mrongrovius: “A concept is the consciousness that the (same) is contained in one representation as in another […]” (AA 29, 888, emphasis added).⁴³ Once I possess the concept, the mark of the object presents itself to me as general. (However, it is important to note that my gaze is not always focused on correspondences among objects.) Thus, Kant tries to o er a very simple solution to the second problem of concept formation. One may wonder whether the solution is actually not too simple. A er all, there are two di erent paradigmatic situations even if we are just dealing with simple features such as colours. In the rst setting, there is a plurality of surfaces in my visual eld, e.g., on a wall opposite to me that I am looking at; they are all coloured in the same shade of red. The surfaces have a di erent location and may also di er in shape. In the second setting we see the same surfaces in terms of their spatial properties, but all of them are coloured in di erent shades of red. In the rst case, it is easy to tell what the di erent instances have in common – and thus, what the representation of this common aspect will be like and how we generate it. While we are comparing, we have to attend to the colour and the shade and neglect the spatial positions and the shapes of the surface. Each of the di erent surfaces has individual, spatio-temporal marks. Despite the individuality of the marks, there can nevertheless be an identity in terms of shade. In the

42 (1) This implies that Kant’s theory allows for intuitive marks. On this matter, see Re ection 2286 (AA 16, 299) and the logic lectures Dohna-Wundtlacken (AA 24, 725) and Busolt (AA 24, 634). See also Stuhlmann-Laeisz, Rainer (1976): Kants Logik (Berlin/New York: De Gruyter), 73, 89, 93; Wol 1995:66; Smit, Houston (2000): “Kant on Marks and the Immediacy of Intuitions”, in: The Philosophical Review 9, 102; and Grüne Blinde Anschauungen, 68. (2) My claim implies that the intuitively given marks are not again general or discursive. In this respect, I follow Grüne Blinde Anschauungen, 66 against Haag Erfahrung und Gegenstand, 166. 43 Translation by U. S.

Concept Formation, Synthesis and Judgment

199

second case, however, it is not obvious how one would proceed, since prima facie we are dealing with surfaces which do not literally correspond to each other from any perspective, not even in their precise colour. This problem can also be spelled out by referring to the process of abstraction: It is insu cient just to know what it is we are abstracting from. We also need to know towards what the abstraction is leading us. In the second case, the latter is not present anywhere in the surfaces that are given to us. But does this not mean that we already have to presuppose the concept, because the individual instances – the di erent shades on the surfaces – do not x where the process of abstraction is heading? Kant’s approach to addressing this issue is similar to his e ort at solving the rst problem of concept formation by referring to the process of synthesis. Once again, the active determination of the understanding is the key to the solution. What it is we pay attention to in comparing and re ecting and what it is that we thereby neglect (such as the di erences in the shades of colour) creates a common feature from our point of view. This common feature would not be present in the merely given material independently from our access to it.⁴⁴ Unlike synthesis, the activity involved in our seeing common features does not proceed in a way that is governed all the way down by necessity.⁴⁵ This would not be desirable anyway, since the conception should allow for variations in how we carve up our concepts. But the activity isn’t unregulated or arbitrary either. As an activity of re ection, it is guided by the corresponding principle of re ection: “that for all things in nature empirically determinate concepts can be found”, as Kant says (AA 20, 211). In addition, the principles of economy always direct us in building a system of concepts. Furthermore, the concepts of identity and similarity must be present in the activity as well – these concepts must be present in re ection in a way that corresponds to the presence of the category concepts in the process of synthesizing intuitions. Kant does not tell us much about this last issue. The Kantian distinction between concepts that are objects of re ection (re ektierte Begri e) and concepts

44 It is important to note that this approach does not imply that Kant endorses nominalism: Once we have the concept we can recognize universal properties in the object. See also Longuenesse Capacity to Judge, 120: Universals “are revealed in things only by the acts of comparison, re ection and abstraction.” 45 This claim does not contradict part II of the paper. What is necessary even here is what the general rules of logic demand, that is, that the procedure moves from singular representations to a representation of what di erent objects can have in common on the ground of economic rationality. However, from this starting point it does not follow that it is necessary which particular concepts humans generate in a given situation.

200

Schlösser

involved in the process of re ecting (re ektierende Begri e) may indicate that he had concepts such as identity and similarity in mind.⁴⁶ However, we still have not reached a satisfactory interpretation of Kant’s theory of concept formation. Several interrelated concerns still need to be addressed. The rst concern relates to Kant’s use of the term ‘abstract’.⁴⁷ According to Kant, we do not abstract a representation – the concept. Rather, we abstract from certain marks of the object and keep other marks. Therefore, one representation can be more or less abstract than other representations. However, this comparative use of the term ‘abstract’ seems to miss an absolute distinction between abstract and concrete which is present in Frege and post-Fregean philosophy. According to the post-Fregean tradition, thoughts understood as the content of sentences, and the concepts that contribute to them are abstract in the absolute sense, whereas spatio-temporal objects and their features are concrete. The charge of not being able to account for the right meaning of ‘abstract’ is closely related to other critical questions concerning Kant’s theory of concept formation: Why is the product of abstraction also not again just singular? How can it be more than the result of a function based on just these very speci c sample cases (that is, this particular spruce, willow and linden)? If so, how is it possible that this product puts reasonable constraints on the application of the concept in the future? I will brie y sketch three features of the Kantian account which contribute to answering these questions. 1. Let us begin with a background remark: Kant’s conception of concepts should not be read in a one-sided way. The content of the representation produced by means of comparison, re ection and abstraction is from the very beginning also de ned by its role in our logical and cognitive activities. Considering the content in isolation, that is, treating it merely as a partly indeterminate product of abstraction, is misleading. The content has been produced as a vehicle that allows for multiple applications. The activity of concept formation is tied to the process of applying the concept in judgments. This relationship can be reformulated using the terminology of the Critique of Judgment: Re ection and determination are correlated in such a way that re ection already anticipates the possibility of determination.⁴⁸ On the other hand, the inverse relation holds as well: I restrict the process of possible determination by binding it to the particular content re ection has produced for this purpose. The content provides me with a guideline that

46 On this distinction, see Re ection 2865 (AA 16, 552). 47 See, for example, AA 9, 95. 48 This possibility is expressed in the Kantian claim that re ection provides a concept as a “ground for cognizing the object” (Re ection 2854, AA 16, 547).

Concept Formation, Synthesis and Judgment

201

makes the process of determination possible. The key feature of Kant’s theory of concepts I try to highlight here can also be spelled out in a di erent way: Starting from my sensory experience I may derive some content, but I do not derive the concept of a concept. This concept is based on the structure of our interrelated logical and cognitive activities. The content I have derived ful ls a function within these activities. 2. Consider again the example of the coloured surfaces which di er in the shade of their colour. Consider also that it is part of the use of the representation in question that it could be applied to these di erent surfaces. Now suppose the mental item I use in my mind nevertheless still has the singular character of an intuition, that is: it is an image. In this case, the image in my mind, which would unavoidably have a speci c colour, could not t precisely to the objects to which it is applied. The representation (the image I use) would not be taken literally anymore. Thereby, we come to understand the Kantian claim that the use of concepts is always symbolic. Unlike geometrical drawings, the symbolic realisation of a concept does not mirror the characteristics of the represented object.⁴⁹ It is this feature of the Kantian theory that allows us to account for theoretical intuitions about the abstractness of concepts. 3. As far as the future applications of concepts are concerned, we nd at least some clues as to how Kant could respond to the problems outlined by Wittgenstein.⁵⁰ Wittgenstein uses the case of a person who, guided by our example, performs the mathematical operation of adding: She successively adds two units given a speci c starting point of the operation. Once she has reached the number 1000, she moves on by adding four units. The crucial question is the following: On what grounds can we blame her for having made a mistake? She may simply have been following a di erent concept. This problem does not just relate to mathematical concepts, but similarly to empirical concepts as discussed in this paper. Prima facie, there is a striking di erence between the thought experiment and the Kantian approach: In Wittgenstein’s thought experiment we look at the person from the outside, whereas Kant always focuses on the rst-person point of view. From my point of view it cannot be entirely indeterminate which marks belong to

49 This distinction is already present in AA 2, 291–292. 50 Wittgenstein, Ludwig (1984/1952): Philosophische Untersuchungen (Frankfurt: Suhrkamp), 336 . (that is, §§ 185 . of the Philosophical Investigations). On this problem, see also Ginsborg, Hannah (2011): “Primitive Normativity and Scepticism about Rules”, in: The Journal of Philosophy, 108 (5).

202

Schlösser

a concept insofar as I myself have made the concept.⁵¹ A change in procedure in applying the concept would have to be grounded in the marks that constitute the concept. In Kripke’s presentation of Wittgenstein’s problem, however, the thought experiment is no longer formulated from an external point of view and still the problem seems to arise. Kripke, too, starts from the rst-person perspective⁵² and reasons in the following way: Even in my own case, there are no facts of the matter in the past which determine which way of continuing corresponds to the previous procedure. The previous procedure of applying the concept would be compatible with di erent ways of continuing since. Previous uses do not x which procedure is correct in the future. For Kant, the issue of which features I paid attention to⁵³ while forming the concept and of which marks I thereby included in the concept are themselves facts of the matter which are present right now. They inform my applications of the concept because I meant them to do so. In order to understand the role of the concept Kant has in mind I would like to draw a double contrast – a contrast that is more striking if we focus again on empirical concepts rather than mathematical ones: According to Kant, the future application of a concept cannot just proceed mechanically. But it is also not true that in applying the concept I retrospectively constitute a determination of the concept which had been le indeterminate before. For each application of the concept we need both a general content (characterized by marks) and the genuine power of judgment. Our power of judgment relates to the content as its rule. However, since the content is general, applying it to various cases always demands an additional, substantial contribution. Here, the power of judgment has its genuine role; a role which cannot be su ciently explained by binding it to another concept or rule. Given the marks and the power

51 Here, some subtle distinctions are necessary. In the case of an empirical concept, I made the concept in the following sense: I provided its general form and I was involved in selecting the relevant marks (I share the second assumption with interpreters such as Hannah Ginsborg). The origin of the matter of the concept is nevertheless empirical; in order to form the concept, I have to be responsive to my sensory environment. Mathematical concepts are made all the way down, because they can be de ned by stipulation. Going back to concept formation introduces a di erent context than Kant’s statements at B 755/A 727 . In that passage, Kant refers to a concept “as it is given” (B 756/A 728, transl. U. S). Thus, he is concerned merely with analyzing the current usage of a concept as it is given in a society. 52 See Kripke, Saul (1982): Wittgenstein on Rules and Private Language (Oxford: Oxford UP), 8 . 53 The role of attention becomes more prominent in the contemporary philosophy of mind, too: In his Simon Lectures, John Campbell, for example, classi es attention as being part of the ‘hard facts of semantics’.

Concept Formation, Synthesis and Judgment

203

of judgment, the question how to proceed is not entirely open as the sceptical reasoning suggests. There is another side constraint which is relevant for how Kant could address the Wittgenstein-Kripke problem. As I have mentioned before, both the idea of synthesis and the principle of re ection introduce demands of economic rationality. These demands are e ective both in forming and in applying concepts. This issue leads back to the intersubjective case. Only if the demands of rationality are e ective in the background are mutual expectations justi ed. If I, as the teacher, would have meant that the other person should change her behaviour above 1000 (a threshold I haven’t reached), then this would not have been communicable; nor would the person have any ground to rationally grasp my expectation.

7 Concept, Judgment and Synthesis In a nal, seventh step I would like to address brie y the relationship between concepts and judgments – as is well known, judgments are at the centre of Kant’s own presentation of his theory. The previous step has already revealed that successful concept formation demands the subsequent use of concepts in judgments. I now want to argue that the opposite relation holds, too: The possibility of judgment also depends on the process of concept formation as described in the logic. Against this claim two passages from the chapter on the “Clue for the Discovery of All Pure Concepts of the Understanding” in the Critique of Pure Reason can be put forward: “All intuitions, as sensible, rest on a ections, concepts therefore on functions. By a function, however, I understand the unity of the action of ordering di erent representations under a common one […]” (B 93/A 68). In the same paragraph, Kant adds: “We can, however, trace all actions of the understanding back to judgments, so that the understanding in general can be represented as a faculty for judging” (B 94/A 69). Adding the second passage to the rst may lead the reader (1) to claim that the activity of subsuming described in the rst quotation already quali es as a rudimentary form of judging. (2) Since the second quotations states that all actions of the understanding can be traced back to judgment, it looks as if a genuine account of concept formation has been given up. Both assumptions together suggest that a representation becomes a concept in the rst place if another representation is subsumed under it as the rst quotation says. However, this line of reasoning is unconvincing. It is important to note that a representation B can be subsumed under a representation A only if A is already a general representation. Subsuming would be impossible if A were originally just a sensory representation of an individual. Sensory representations of individuals can only be co-ordinated, they cannot be subordinated. The re ection at the centre of

204

Schlösser

concept formation is therefore still a necessary starting point since it can provide the general representation that is needed. So far, I have addressed the rst, second and fourth objection against Kant’s overall model of the actions of the understanding. In the previous paragraph, I have also responded to the third concern. Thus, in a nal section, I would like to comment on the relation between judgment and synthesis. The potential tension in this relationship I have outlined in the fourth part of the paper can be summarized as follows: On the one hand, Kant says that a judgment “is nothing other than the way to bring given cognitions to the objective unity of apperception” (B 141). Thereby Kant relates judgment also to the performance of the synthesis of intuitions. This may suggest that the synthetic connections are actualized by our judging. On the other hand, it seems as if the very possibility of judgment already presupposes that the synthesis of intuition is e ective. The latter seems to be the case for several reasons. One of them is that judgments presuppose concepts. The formation of concepts in turn depends on the performance of a synthesis on the intuitive level as we have seen in part V. The alleged problem can be solved by giving up the rst claim: Unity and order in the given sensory material are not established by means of our judging. The synthetic connections are necessary and therefore always already e ective in relation to all potential content. I would like to suggest a di erent reading of the Kantian passage: A logical combination of originally merely subjective states only quali es as a judgment, that is: as something that transcends the individual person and her subjectivity, if it is bound to the objective unity of apperception and the synthesis underlying it.⁵⁴ Thus the relation to objectivity established by the synthesis explains why judgments and the concepts that partake in them are not merely private even though they are grounded on mental representations. Kant can therefore respond to the Fregean challenge. The claim that our cognitive activities establish a relation to

54 According to the terminology of the Prolegomena, this merely subjective combination of representations is called ‘judgment of perception’. “Empirical judgments, insofar as they have objective validity, are judgments of experience; those, however, that are only subjectively valid I call mere judgments of perception. The latter do not require a pure concept of the understanding, but only the logical connection of perceptions in a thinking subject” (AA 4, 298). Insofar as the judgments of experience are related to the concepts of the understanding they are also tied to the synthesis. Thereby, the judgments of experience are more than just my subjective states. This universal aspect is captured in the following formulation: “Now this judgment can be of two types: rst, when I merely compare the perceptions and conjoin them in a consciousness of my state, or, second, when I conjoin them in a consciousness in general” (AA 4, 300). In the given context, one should bear in mind that it is controversial whether Kant is entitled to call the judgments of perceptions ‘judgments’ at all. Doing so seems to contradict the account of judgments in §§ 18–19 of the B-deduction.

Concept Formation, Synthesis and Judgment

205

objectivity and thus allow for the communicability of the content of our mental representations is only an apparent paradox. The complex discussion can be summarized as follows: (1) Kant’s emphasis on the active role of the understanding is not itself a problem; for Kant, it is rather a tool to solve several problems surrounding concept formation and judgment. (2) These problems can be solved precisely because we are not dealing with just one type of action, but with a plurality of di erent activities which interact in an organic way. They contribute to the Kantian image of the purposiveness of our mind. As the cognizing being in question, we can’t help assuming that our mind is organized in a purposive way. This assumption, hardly ever spelled out explicitly, is the keystone of Kant’s reasoning.⁵⁵

55 I am grateful to the audience in Berlin and also to Tyson Go on for his comments on this paper.

Eckart Förster

Grenzen der Erkenntnis? Rolf-Peter Horstmann hat in seinem Buch Die Grenzen der Vernun auf beeindruckende Weise vorgeführt, wie sich die an Kant anschließenden Philosophen mit der Kantischen Behauptung von den notwendigen Grenzen unserer Erkenntnis auseinandergesetzt haben und wie weit sich Ziele und Motive des deutschen Idealismus aus dieser Auseinandersetzung verstehen lassen. „Sie alle“, schreibt Horstmann, „von J. G. Fichte über F. W. J. Schelling bis zu G. W. F. Hegel, waren sich einig darin, daß Kant zu kurz grei und daß es gilt, mit ganz anderen Mitteln und auf ganz anderen Wegen gegen Kant die Einsicht in die Grenzenlosigkeit der Vernun durchzusetzen“.¹ Ich möchte im Folgenden Horstmanns Thema von der anderen Seite her aufgreifen und fragen, was Kant denn überhaupt motiviert hat, unserer Erkenntnis Grenzen zu setzen bzw. ob man auch von seinem eigenen Standpunkt aus sagen kann, er habe mit dieser Grenzziehung zu kurz gegri en.

1 Fichte war bekanntlich der erste, der die Kantische Grenze überschritten hat. Dazu ging er davon aus, dass das Ich sein Sein nur als seine eigene Tat erkennen kann – eine Einheit von Denken und Sein, die keine rezeptive, sondern eine produktive oder intellektuelle Anschauung ist. Das impliziert zweierlei: Da das Ich nur durch sich selbst sein kann, was es ist, muss es sich selbst „setzen“; da es, um ein Ich zu sein, sich seiner selbst bewusst sein muss, muss es auch die Schritte, die es notwendig tut, um sich zu setzen, ins Bewusstsein heben können. Die Wissenscha slehre, die Fichte ab 1794 in Jena vortrug, beschrieb er selbst einmal so: „eben in der Erforschung der für Kant unerforschlichen Wurzel, in welcher die sinnliche und die übersinnliche Welt zusammenhängt, dann in der wirklichen und begrei ichen Ableitung beider Welten aus Einem Princip, besteht ihr Wesen“.² Dieser ers-

1 Rolf-Peter Horstmann, Die Grenzen der Vernun . Eine Untersuchung zu Zielen und Motiven des Deutschen Idealismus. Frankfurt: Anton Hain 1991. 2 „Die Wissenscha slehre“, II. Vortrag 1804. In: J. G. Fichte – Gesamtausgabe der Bayrischen Akademie der Wissenscha en. Nachgelassene Schri en, Bd. 8. Hg. von R. Lauth und H. Gliwitzky. Stuttgart-Bad Cannstatt: Frommann 1985, 32.

208

Förster

te Schritt über die Kantische Erkenntnisgrenze ist insofern besonders interessant, als Kant selbst ursprünglich eine intellektuelle Anschauung des Ich annahm. So notierte er sich z.B. in seinem Exemplar von Baumgartens Metaphysica: „[W]ir haben … intellectuelles inneres Anschauen (nicht den innern Sinn) unserer Thatigkeit“ (Re . 4336, 17:509; vgl. Re . 4228, 17:467).³ An anderer Stelle, in einer Vorlesungsnachschri , heißt es: „Es ist dieses Ich aber ein absolutes Subject […]. Dieses ist der einzige Fall, wo wir die Substanz unmittelbar anschauen können. Wir können von keinem Dinge das Substratum und das erste Subject anschauen; aber in mir schaue ich die Substanz unmittelbar an. Es drückt also das Ich nicht allein die Substanz, sondern auch das substantiale selbst aus“ (28.1:225–226). Das wird bekanntlich in der kritischen Phase von Grund auf verworfen. Was hat dazu geführt? Wie ist Kant zu der Au assung gekommen, dass eine intellektuelle Anschauung unmöglich ist und dass unserer Erkenntnis notwendige Grenzen gezogen sind?

2 Zu den Schri en, die Kant bereits während seiner Studienzeit nachhaltig geprägt haben, gehört neben Newtons Werken zweifellos der berühmte Briefwechsel zwischen Leibniz und dem Newtonianer Samuel Clarke. Er dür e dazu beigetragen haben, dass Kant sich trotz seiner Bewunderung für Newton nicht mit dessen Lehre vom absoluten Raum anfreunden konnte und stattdessen wie Leibniz eine relationale Position vertrat.⁴ In seinem vierten Brief an Clarke hatte Leibniz geschrieben: „Wenn der Raum eine absolute Wirklichkeit ist, so wird er, gegenüber der Substanz weit mehr als eine Eigenscha oder ein zufälliges Merkmal, realer sein als alle Substanzen, und Gott könnte ihn nicht zerstören, noch auch im geringsten verändern. Er ist dann nicht nur als ganzer unermeßlich, sondern auch in jedem

3 Ich zitiere Kant im Text unter Angabe von Band- und Seitenzahl der Akademieausgabe (AA) von Kant’s gesammelte Schri en. Berlin und Leipzig: de Gruyter 1900 ., die Kritik der reinen Vernun dagegen wie üblich nach der ersten (A) bzw. zweiten (B) Au age. 4 Die von David Walford vertretene These, wonach Kant in seinen frühen Schri en „a form of compatibilism […] a synthesis of Newtonian absolutism and Leibnizian relationalism“ vertreten habe, lässt sich m. E. nicht halten und wird, soweit ich sehe, auch durch die von Walford angeführten Stellen nicht belegt; vgl. David Waldorf, „Towards an Interpretation of Kant’s 1768 Gegenden im Raume Essay“. In: Kant-Studien 92 (2001), 407–439, 433.

Grenzen der Erkenntnis?

209

seiner Teile unwandelbar und ewig. Es gibt dann eine unendliche Anzahl ewiger Dinge außer Gott.“⁵ Für Leibniz selbst ist der Raum nichts anderes als die Relationen zwischen Dingen, so dass es gar keinen Raum gibt, wenn keine Dinge existieren. Kant akzeptierte diese Au assung mit Modi kationen bis 1768, als er deren Unhaltbarkeit in einem Gedankenexperiment glaubte demonstrieren zu können. In seinem Aufsatz Von dem ersten Grunde des Unterschiedes der Gegenden im Raum unterscheidet er zunächst Lage und Gegend voneinander, um dann zu zeigen, dass die Lage der Teile des Raumes untereinander die Gegend bereits voraus setzen – und damit einen absoluten Raum. („Gegend“ hat hier noch die alte Bedeutung von „Richtung“).⁶ [I]m abgezogensten Verstande besteht die Gegend nicht in der Beziehung eines Dinges im Raume auf das andere, welches eigentlich der Begri der Lage ist, sondern in dem Verhältnisse des Systems dieser Lagen zu dem absoluten Weltraume. Bei allem Ausgedehnten ist die Lage seiner Teil gegeneinander aus ihm selbst hinreichend zu erkennen, die Gegend aber, wohin diese Ordnung der Theile gerichtet ist, bezieht sich auf den Raum außer demselben und zwar nicht auf dessen Orter, weil dieses nichts anders sein würde, als die Lage derselben Theile in einem äußeren Verhältniß, sondern auf den allgemeinen Raum als eine Einheit, wovon jede Ausdehnung wie ein Theil angesehen werden muß (2:377–8).

Dass sich Gegend nicht auf Lage reduzieren lässt, wie Leibniz Theorie des Raumes es verlangen würde, glaubt Kant 1768 anhand von inkongruenten Gegenstücken zeigen zu können, d.h. solchen Gegenständen, die hinsichtlich Größe, Proportion sowie Lage der Teile untereinander völlig gleich sind, aber keine gemeinsamen Grenzen haben können, also verschieden sind. Das bekannteste Beispiel solch inkongruenter Gegenstücke sind die Hände des Menschen:⁷ „Die rechte Hand ist der linken ähnlich und gleich, und wenn man bloß auf eine derselben sieht, auf die

5 Samuel Clarke, Der Briefwechsel mit G. W. Leibniz von 1715/1716, hg. von E. Dellian. Hamburg: Meiner 1990, 43. 6 Vgl. z.B. Christian Wol , Mathematisches Lexicon. Leipzig 1716, 659: „Die Gegend (Plaga): Ist eine gerade Linie, die aus dem Auge gegen den Himmel mit den Horizont Parallel gezogen wird. Mehr mathematisch redet man, wenn man den Durchschnitt des Vertical=Circuls und des Horizonts nennt.“ So auch noch in Grimms Deutsches Wörterbuch. Band 5. Leipzig 1897, 2230: „Gegend: ‚der [Begri ] einer Richtung, von mir aus bestimmt‘“. Kant nennt die Beispiele: oben und unten, rechts und links, vorne und hinten (2:379). 7 Auf dieses Beispiel werde ich mich im Folgenden der Einfachheit halber beschränken. Auch werde ich nur immanent die Bedeutung dieses Beispiels für die Kantische Au assung von notwendigen Erkenntnisgrenzen betrachten und von allen physikalischen und mathematischen Fragen, die inkongruente Gegenstücke aufgeben, abstrahieren. Vgl. dazu die Aufsätze in J. van Cleve and R. E. Frederick (eds.), The Philosophy of Right and Le : Incongruent Counterparts and the Nature of Space, Dordrecht 1991; Felix Mühlhölzer, „Das Phänomen der inkongruenten Gegenstücke

210

Förster

Proportion und Lage der Theile untereinander und auf die Größe des Ganzen, so muß eine vollständige Beschreibung der einen in allen Stücken auch von der andern gelten“ (2:381). Zwar können wir nur im Vergleich zweier Gegenstände deren Inkongruenz erkennen; das bedeutet aber nicht, dass sie selbst ihre Bestimmtheit (Gerichtetheit) erst in einem solchen Vergleich erlangen würden. Um das zu verdeutlichen, schlägt Kant folgendes Gedankenexperiment vor. Angenommen, es sei noch keine Welt gescha en und das erste Schöpfungsstück sei eine menschliche Hand. Sie muss entweder eine rechte oder linke sein: würde als nächster Gegenstand ein menschlicher Torso gescha en, dann passte sie nur auf einen der Arme, nicht auf den anderen. Dass sie eine linke oder eine rechte ist, kann aber in diesem Fall nicht durch äußere Relationen, d.h. durch die Beziehung auf andere Gegenstände bestimmt werden, denn per Annahme ist die Hand der bisher einzige Gegenstand: ihre Bestimmtheit – ihre Gerichtetheit – muss also, wie Kant hier sagt, auf einem „inneren Grunde“ (2:382) beruhen. Das ist das erste wichtige Ergebnis. Was heißt das aber genau? Kant de niert den Begri des „inneren Grunds“ zwar nicht, identi ziert ihn aber mit dem „erste[n] Grund der Möglichkeit ihrer [der Materie] Zusammensetzung“, und diesen wiederum mit dem „allgemeinen Raum als einer Einheit, wovon jede Ausdehnung wie ein Theil angesehen werden muß“ (2:378). Der innere Grund ist also dasjenige Ganze, welches die Teile bzw. deren Zusammensetzung erst möglich macht. Im Falle inkongruenter Gegenstücke kann dies nicht deren physisches Ganze sein, da dieses die Summe ihrer jeweiligen Teile und deren Lagen ist, aber auch nicht Leibniz’ Raum, der nichts anderes ist als die Lagen der Teile – folglich bleibt nur die Beziehung des Gegenstands (der Teile) zum absoluten Raum. So schreibt Kant: Es ist hieraus klar: daß nicht die Bestimmungen des Raumes Folgen von den Lagen der Theile der Materie gegen einander, sondern diese Folgen von jenen sind, und dass also in der Bescha enheit der Körper Unterschiede angetro en werden können, und zwar wahre Unterschiede, die sich lediglich auf den absoluten und ursprünglichen Raum beziehen, weil nur durch ihn das Verhältnis körperlicher Dinge möglich ist (2:383).

Wir haben hier gewissermaßen ein Kantisches Pendant zu Newtons berühmten Eimerversuch⁸ für die Existenz des absoluten Raumes: Die empirische Tatsache inkongruenter Gegenstücke bezeugt Kant zufolge dessen „eigene Realität“. Ein

aus Kantischer und heutiger Sicht“, in: Kant-Studien 83 (1992), 436–453; P. Rusnock, R. George, „A Last Shot at Kant and Incongruent Counterparts“, in: Kant-Studien 86 (1995), 257–278. Zur Bedeutung inkongruenter Gegenstücke für die Entstehung des transzendentalen Idealismus vgl. J. Vance Buroker, Space and Incongruence. The Origin of Kant’s Idealism, Dordrecht: Reidel 1981. 8 Vgl. Newtons „scholium“ zu seinen De nitionen, in Isaac Newton, Mathematischen Prinzipien

Grenzen der Erkenntnis?

211

solcher Raum muss folglich als wirklich angenommen werden und „nicht für ein bloßes Gedankending“ gehalten werden (2:383). Trotzdem hat Kant schon kurz darauf in seiner Inauguraldissertation De mundi sensibilis atque intelligibilis forma et principiis die Annahme des absoluten Raumes wieder verworfen und diesen sogar „ins Märchenland [ad mundum fabulosum]“ relegiert, da er doch „unendliche wahrha e Relationen ohne irgendwelche aufeinander bezogene Dinge erdichtet“ (2:404, § 15 D).⁹ Über die Gründe für diese erstaunliche Kehrtwendung sagt Kant uns nichts. Er sagt nur, dass das folgende Jahr (1769) ihm „großes Licht“ (Re . 5037, 18:69) gab. Wieder ein Jahr später (1770) präsentierte er dann in seiner Inauguraldissertation seine eigene revolutionäre Theorie von Raum und Zeit als Anschauungsformen. Warum ist er nicht bei Newtons Position geblieben? Zunächst ist festzuhalten, dass Kants Gedankenexperiment lediglich eine reductio ad absurdum der Leibnizschen Raumtheorie darstellt. Eine positive Lösung des Problems inkongruenter Gegenstücke skizziert er in dem Aufsatz von 1768 nicht. Oder besser gesagt: Kant behauptet nur, erstens, dass es einen inneren Grund für deren Inkongruenz geben muss, und er identi ziert, zweitens, diesen Grund mit Newtons absolutem Raum, weil „der allgemeine Raum eine Einheit ist, wovon jede Ausdehnung wie ein Theil angesehen werden muß“ (2:378). Er sagt aber nicht, wie aus der Einheit des absoluten Raumes die Gerichtetheit z.B. einer Hand erklärt werden kann, d.h. wie sich die Tatsache, dass es entweder eine rechte oder eine linke ist, aus dem Ganzen des Raumes limitando verstehen lässt. Nun macht, wenn ich recht sehe, Newtons absoluter Raum eine solche Erklärung in der Tat unmöglich. In diesem Raum gibt es keine realen Unterscheidungsgründe, keine Topologie, wie wir heute sagen würden, derart, dass sie die in diesem Raum seienden Dinge beeinträchtigen könnte. Wenn Leibniz im Briefwechsel mit Clarke einwendet, unter Voraussetzung des absoluten Raumes hätte Gott keinen Grund gehabt, die Welt an ihrem jetzigen Ort statt an einem anderen zu scha en, was dem Prinzip des zureichenden Grunds widerspricht, gesteht Clarke zu, „all places are originally alike […] [and there is an] indi erency of all places“.¹⁰ Er kann Leibniz darum nur entgegen halten, dass der zureichende Grund für die Erscha ung der Welt an einem bestimmten Ort in Gottes Wille, nicht in der Natur des Raumes zu suchen sei.

der Naturlehre, Hg. von J. Ph. Wolfers. Darmstadt: Wissenscha liche Buchgesellscha 1963, 29. Für die Unterscheidung von „Gegend“ und „Lage“ bei Newton, siehe ibid., 31. 9 Ich folge Klaus Reichs Übersetzung von Kants Inauguraldissertation in I. Kant, Über die Form und die Prinzipien der Sinnen- und Geisteswelt. Hamburg: Meiner 1958. 10 „Dr. Clarke’s Third Reply“ in: The Leibniz-Clarke-Correspondence, hg. von H. G. Alexander. Manchester: Manchester University Press 1956, 30.

212

Förster

Kant scheint nun geschlossen zu haben: Die Annahme einer einzelnen Hand als erstem Schöpfungsstück ist eine widerspruchsfreie Annahme. Ihre notwendige Gerichtetheit kann nur auf einem inneren Grund beruhen. Wenn weder Leibniz’ noch Newtons Raum die Möglichkeit dieser Gerichtetheit verständlich machen können, dann muss es noch eine Alternative zu den beiden Möglichkeiten absoluter oder relationaler Raum geben. Aber wie könnte die aussehen? Um Kants Gedankengang nachzuvollziehen, gehen wir am besten von dem aus, was sich bisher ergeben hat, nämlich dass die unterschiedliche Gerichtetheit z. B. von linker und rechter Hand intuitiv deutlich, begri ich aber gar nicht grei ar ist.¹¹ Die erste Lektion, die Kant daraus ziehen musste, ist, dass Anschauung und Verstand nicht nur graduell unterschieden sein können, wie Empiristen und Rationalisten gleichermaßen annahmen, sondern zwei grundsätzlich verschiedene Erkenntnisquellen sein müssen mit je eigenen Prinzipien und Formen. Worin besteht dann aber ihr wesentlicher Unterschied? Mittels der Sinnlichkeit wird uns etwas gegeben, mittels des Verstandes aber nicht, oder wie Kant in der Inauguraldissertation schreibt: Sinnlichkeit ist die Empfänglichkeit eines Subjekts, durch die es möglich ist, daß sein Vorstellungszustand durch die Gegenwart irgendeines Objekts auf bestimmte Weise betro en wird. Der Verstand oder die Vernun ist das Vermögen eines Subjekts, kra dessen es das, was in seine Sinne wegen seiner eigentümlichen Bescha enheit nicht fallen kann, vorzustellen vermag (§ 3, 2:392; Reich 19).

Die Sinnlichkeit als Möglichkeit, a ziert zu werden, ist somit immer passiv, eine reine Rezeptivität. Der Verstand dagegen ist spontan und aktiv. Ist das richtig, dann ist damit für Kant zugleich ausgeschlossen, dass wir eine intellektuelle Anschauung haben können. Sie könnte, so scheint er geschlossen zu haben, nur ein Oxymoron sein, eine spontane Passivität: „Bei den intellektuellen Erkenntnissen gibt es (für den Menschen) keine Anschauung, sondern nur eine symbolische Erkenntnis“ (2:396, § 10; Reich 31). Sind Anschauung und Verstand derart zu unterscheiden, dann lässt sich Kants Gedankengang vielleicht folgendermaßen rekonstruieren: 1. Die Existenz inkongruenter Gegenstücke wie linke und rechte Hand ist eine Tatsache. 2. Inkongruente Gegenstücke sind intuitiv verschieden, begri ich aber identisch und ununterscheidbar.

11 „Der Unterschied kann durch keine Schärfe des Verstandes diskursiv beschrieben oder auf intellektuelle Merkmale zurückgeführt werden“ (2:403, § 15 C, Reich 51).

Grenzen der Erkenntnis?

213

3. Inkongruente Gegenstücke sind somit keine Verstandesgegenstände (=Dinge an sich), da diese „sich notwendig auf objektive Begri e müßten bringen lassen“ (4:484, vgl. 4:286). 4. Inkongruente Gegenstücke sind folglich Sinnesgegenstände, d.h. Dinge, die „von der Natur des Subjekts, insofern es von [einem] Objekt modi zierbar ist“ (2:293, § 4, Reich 21), abhängig sind – also Erscheinungen. 5. In der Natur des Subjekts, genauer: in seiner Sinnlichkeit, muss der innere Grund der Inkongruenz solcher Gegenstände gesucht werden. 6. Liegt der innere Grund in der Natur der Sinnlichkeit, dann kann er nur darin bestehen, dass jeder Raum ein bestimmter (in diesem Fall: gerichteter) Raum dadurch wird, dass er eine Begrenzung des umfassenden Anschauungsraums (Form der Sinnlichkeit) ist.¹² Dass damit der entscheidende Grund für die Einführung des subjektiven Raumbegri s bezeichnet ist, sagt Kant 1770 in der Inauguraldissertation ausdrücklich: Der Raum ist also ein absolut erstes formales Prinzip der Sinnenwelt, nicht bloß deswegen, weil nur mit Hilfe seiner Vorstellung die Gegenstände des Weltalls Erscheinungen sein können, sondern vor allem deswegen, weil er wesentlich einzig ist und absolut alle äußeren Sinnengegenstände in sich befaßt und daher ein Prinzip der Allheit, nämlich eines Ganzen, das kein Teil eines anderen sein kann, abgibt (III § 15 E, Reich 57).

Damit sind zunächst einmal der Sinnlichkeit Grenzen gesetzt: die sinnlichen Erkenntnisse sind nur Vorstellungen der Dinge, wie sie erscheinen, die Verstandeserkenntnis liefert 1770 noch Vorstellungen der Dinge, wie sie an sich sind. Dabei konnte Kant allerdings nicht lange stehen bleiben. Denn ist der Verstand im Gegensatz zur Sinnlichkeit als aktives, spontanes Vermögen bestimmt, der aber die Gegenstände seiner Erkenntnis nicht selbst hervorbringt, muss sich über kurz oder lang die Frage stellen, die Kant 1772 im Brief an Marcus Herz so formulierte: „wenn solche intellectuale Vorstellungen auf unsrer innern Thätigkeit beruhen, woher komt die Übereinstimmung die sie mit Gegenständen haben sollen?“ (10:131). Das führt dann dazu, auch dem Verstand und der Vernun Grenzen zu setzen und damit zur Position der Kritik, die ich hier als bekannt voraussetze.

12 D.h. „die innere Bestimmung eines jeden Raumes ist nur durch die Bestimmung des äußeren Verhältnisses zu dem ganzen Raume, davon ein jener ein Theil ist (dem Verhältnisse zum äußeren Sinne), d.i. der Theil ist nur durchs Ganze möglich“ (4:286).

214

Förster

3 Nun ist der Einwand naheliegend, dass auch bei Kants kritischer Theorie des Raums als Form der äußeren Anschauung unklar ist, wie dieser ein innerer Grund für die Inkongruenz empirischer Gegenstände sein kann. Dazu scheint der Gedanke, dass der Raum als Anschauungsform ein Ganzes ist, von dem jeder empirische Raum eine Einschränkung darstellt, genauso wenig hinreichend zu sein wie Newtons absoluter Raum. Kant hat in diesem Zusammenhang gelegentlich den Vergleich gemacht mit einem „Marmorlager zu Bildsäulen von unendlicher Mannigfaltigkeit, welche insgesamt nur durch Einschränkung (Absonderung des Übrigen von einem gewissen Teil des Ganzen […]) möglich“ sind (20:302). Dabei scheint er an bekannte Äußerungen Michelangelos gedacht zu haben. Aber obwohl z. B. Michelangelos David nur durch Einschränkung eines Marmorblocks zustande kam, ist der Marmor kaum der „innere Grund“ der Statue, sonst hätten Agostino di Duccio oder Antonio Rossellino, die sich vorher bereits vergeblich an dem Stein versucht hatten, die Statue ja ebenfalls aus dem Block herausholen können. Der Marmor, so wird man sagen müssen, ist nur das Medium für die Statue, aber nicht der innerer Grund ihrer Form, der vielmehr in der Idee des Künstlers von dem zu scha enden Werk zu sehen ist. Es bliebe meines Erachtens auch tatsächlich unklar, warum Kant glauben konnte, seine Theorie des Raumes könne besser mit inkongruenten Gegenstücken zurechtkommen als Newtons absoluter Raum, würde die Identi kation der Sinnlichkeit mit Passivität nicht noch eine entscheidende Implikation haben. Daraus folgt nämlich, dass uns überhaupt keine Gegenstände gegeben werden können, sondern dass Eindrücke immer unverbunden in der Sinnlichkeit au reten und erst von uns zu Gegenständen der Erfahrung konstituiert werden müssen. In einem passiven Medium als solchem kann es grundsätzlich nur Unverbundenes geben. Ist dem so, dann muss der innere Grund inkongruenter Gegenstände mit der Möglichkeit der Zusammensetzung ihrer Vorstellungen zusammenhängen; und zwar so, dass die Zusammensetzung in einer bestimmten, durch die Form der Sinnlichkeit bedingten, aber für den diskursiven Verstand nicht fassbaren Weise geschieht. Was heißt das aber genau? Damit verschiedene, in meinem Bewusstsein einzeln au retende Vorstellungen überhaupt als Vorstellungen von einem Gegenstand betrachtet werden können, muss ich denken können, dass sie notwendig zusammengehören und so einen Gegenstand ausmachen. Ich muss sie also von mir unterscheiden können. Nun habe ich aber keine andere Möglichkeit, etwas von mir zu unterscheiden, als dass ich es an einem anderen Ort vorstelle. Der Raum macht somit zwar nicht die Vorstellungen, wohl aber deren Zusammensetzung zu einem von mir unterschiedenen Gegenstand möglich: Der Raum mit dem

Grenzen der Erkenntnis?

215

darin enthaltenen Mannigfaltigen, könnte man sagen, ist wie das Marmorlager, und das Erkenntnissubjekt ist wie der Künstler, der daraus den Gegenstand der Erfahrung konstituiert. So wie die Materialität des Marmor dem Künstler, so legt die Form der Sinnlichkeit der objektivierenden Handlung des Erkenntnissubjekts gewisse Beschränkungen auf. Damit hat sich das Problem aber nur verschoben. Das ursprüngliche Gedankenexperiment, von dem Kant ausging, ist jetzt gar nicht mehr durchführbar. Der Gedanke, dass es nur eine einzige Hand gibt, lässt sich innerhalb der kritischen Konstitutionstheorie nicht mehr sinnvoll formulieren. Einerseits kann es nicht mehr irrelevant sein, ob ich erkennen kann, ob es eine linke oder rechte Hand ist. Andererseits gilt für linke und rechte Hand gleichermaßen, dass ich deren Mannigfaltigkeit von mir unterscheide und an einem anderen Ort vorstelle – das kann also nicht der innere Grund von deren Inkongruenz sein. Zu ihrem Verständnis ist Kants subjektiver Raumbegri genauso unzureichend wie der absolute und der relationale. In Abwandlung eines bekannten Satzes Jacobis könnte man sagen: Ohne die Voraussetzung inkongruenter Gegenstücke konnte ich nicht in die kritische Raumtheorie des transzendentalen Idealismus hineinkommen, und mit jener Voraussetzung kann ich darin nicht bleiben.

4 Wie konnte es dazu kommen? Kant hat recht, scheint mir, wenn er darauf insistiert, dass (1), um Inkongruenz zu verstehen, wir Richtungen unterscheiden können müssen. Und er hat auch recht, dass (2) Inkongruenz auf einem inneren Grund beruht, genauer, auf der Abhängigkeit der Teile vom Ganzen. Das ist aber Zweierlei, das von Kant zunächst zusammengeworfen und wie eines betrachtet wird. (Ad 1) Zwei Dinge können nur unterschieden werden, insofern es ein Drittes gibt, hinsichtlich dessen sie verschieden sind. So sind zwei Richtungen ununterscheidbar, wenn es nicht einen Punkt oder eine Grenze gibt, auf den hin oder von dem weg die Richtung weist. Vorher kann man gar nicht von unterschiedenen Richtungen, und damit letztlich auch nicht einmal von Richtungen reden. (Ad 2) Das Ganze (der innere Grund), der die Teile z.B. einer Hand möglich macht, hat aber mit einem solchen Punkt unmittelbar nichts zu tun, und auch nichts mit der Natur des Raumes. Dasjenige Ganze, das z.B. eine Hand möglich macht, ist der Organismus, dessen Teil sie ist. Wie Kant später selbst in der Kritik der Urteilskra argumentierte, haben bei einem Organismus dessen Teile gar keine vom organischen Ganzen unabhängige Existenz. Ein für sich seiender Teil des Organismus, z.B. eine einzelne Hand wie in Kants ursprünglichem Gedankenexperiment, ist eine Unmöglichkeit. Empirisch zeigt sich das auch darin, dass eine

216

Förster

Hand, die vom Körper abgetrennt ist, keine Hand mehr ist (worauf schon Aristoteles hinwies) und sich in seine chemischen Bestandteile au öst.¹³

5 Damit scheint das Thema inkongruenter Gegenstücke aus Kantischer Sicht abgeschlossen zu sein. Das ist es aber nicht. Allerdings ging seine Weiterentwicklung zunächst nicht von diesen selbst aus, sondern von Kants fortgesetzter Beschä igung mit dem Problem des Raumes. Nach der Publikation der Kritik der reinen Vernun musste Kant sich mit der Behauptung auseinandersetzen, dass sein Idealismus derselbe wäre wie derjenige Berkeleys. Seine Erwiderungen konzentrierten sich zunächst auf den Nachweis, dass der Raum nicht empirischen Ursprungs sein könne, wie Berkeley annahm. Denn damit ich aus sinnlich Gegebenem die Vorstellung des Raumes abstrahieren könnte, müsste ich ein Mannigfaltiges der Anschauung als Bleibendes oder Zugleichseiendes vorstellen können.¹⁴ Das ist aber im inneren Sinn allein nicht möglich, denn dort ist alles sukzessiv und damit jede neue Wahrnehmung später als ihr Vorgänger. Es ist also unter dieser Voraussetzung gar nicht möglich, ein Mannigfaltiges der Anschauung als gleichzeitig vorzustellen und damit auf etwas von mir Unterschiedenes zu beziehen. Das bedeutet dann aber auch, dass ohne den Raum keine Zeitbestimmung möglich ist. In seiner ursprünglichen Abgrenzung von Newton und Leibniz hatte sich Kant ganz darauf konzentriert, dass Raum und Zeit nicht vom Subjekt unabhängige reale Größen sein können, sondern nur Formen der Anschauung. Jetzt fällt das Licht auf einen anderen Aspekt: Sie können nur zusammen Erfahrung möglich machen. Keine Raumbestimmung ohne Zeit; keine Zeitbestimmung (Schematismus) ohne Raum. Der Raum selbst kann aber nicht wahrgenommen werden; vielmehr muss er für uns durch das Zugleichsein der Gegenstände in ihm

13 Auch für Kants andere Beispiele ließe sich m.E. zeigen, dass ein jeweiliges ‚Ganzes‘ der innere Grund der Inkongruenz ist, nur muss von Fall zu Fall entschieden werden, was das Ganze im einzelnen Fall ist. So gibt es z.B. keine sich windende Bohne ohne den Zusammenhang von u.a. Sonne, Gravitation, Erdrotation; kein sphärisches Dreieck ohne Kugel bzw. ohne den Mathematiker, der es in der Anschauung konstruiert; keine Schrauben ohne den Fräser, der die Gewinderichtung festlegt. 14 „Zugleich“ sind zwei Dinge, der dritten Analogie zufolge, wenn die Wahrnehmung des einen (A) auf die Wahrnehmung des anderen (B) wechselseitig folgen kann, die Apprehension (nicht nur das Denken) also sowohl von A nach B fortschreiten, als auch von B wieder zu A zurückschreiten kann.

Grenzen der Erkenntnis?

217

repräsentiert werden. Das hat Konsequenzen für die Transzendentalphilosophie. In ihr muss dann nämlich auch noch gezeigt werden, wie sich die Kategorien a priori auf Gegenstände des äußeren Sinns anwenden lassen.¹⁵ Der erste Schritt muss darin bestehen anzugeben, wodurch sich „[e]twas, das ein Gegenstand äußerer Sinne sein soll“ (4:476) und damit den Raum wahrnehmbar macht, a priori vom Raum unterscheiden lässt. Kants Antwort: Dadurch, dass es einen Raum erfüllt und anderes hindert, denselben Raum einzunehmen. „[Es] ist der Widerstand, den eine Materie in dem Raum, den sie erfüllt, allem Eindringen anderer leistet, eine Ursache der Bewegung der letzteren in entgegengesetzter Richtung. Die Ursache einer Bewegung heißt aber bewegende Kra . Also erfüllt die Materie ihren Raum durch bewegende Kra und nicht durch ihre bloße Existenz“ (4:497). Damit ein Gegenstand des äußern Sinns als solcher erfahren werden kann, muss das Subjekt der Erfahrung folglich nicht nur an den Sinneseindrücken die dreifache Synthese der Apprehension, Reproduktion und Rekognition ausüben. Nicht dadurch wird Raumerfüllung schon wahrnehmbar, dass ich Vorstellungen von mir unterscheide und auf einen Punkt im Raum beziehe, sondern dadurch, dass ich dessen Erfüllung als Wirkung auf mich erfahre (und nicht bloß denke). Das kann ich nur dadurch, dass ich selbst Bewegungen ausübe – und zwar Bewegungen körperlich-leiblicher Art, die über die transzendentalen Bewegungen der drei genannten Synthesen hinausgehen: „Die bewegenden Krä e der Materie sind das was das bewegende Subject selbst thut mit seinem Körper an Körpern“ (22:326). Damit ist auch die Passivität der Sinnlichkeit nur noch als relative zu deuten: Etwas kann nur gegeben werden, wenn es in einer entsprechenden Bewegung entgegen genommen wird: „Die A ectibilität des Subjects als Erscheinung ist mit der Incitabilität der correspondirenden bewegenden Krä e als Correlat in der Wahrnehmung verbunden“ (22:396). Wie ist dies aber transzendentalphilosophisch zu verstehen?

6 Damit ich bewegende Krä e ausüben kann, muss ich selbst einen Körper haben. Nun ist es, wie wir gesehen haben, ein Axiom der Kantischen Theorie der Sinnlichkeit, dass kein sinnlicher Gegenstand gegeben ist, sondern konstituiert werden muss, also auch mein eigener Körper. Interessanterweise hat Kant gerade dieses

15 Vgl. 4:478 und ausführlicher in Eckart Förster, Die 25 Jahre der Philosophie, Frankfurt: Klostermann 2011, 64–7, 77–85.

218

Förster

Thema in einem vor kurzem von Werner Stark aufgefundenen Losen Blatt, das auf die erste Häl e von 1799 zu datieren ist,¹⁶ aufgegri en. Auf die Frage: „Wie ist der Begri eines organischen Körpers möglich?“ antwortet Kant nun: Als Mensch organisiere ich „ursprünglich mich selbst“, indem ich das Mannigfaltige der Anschauung meiner selbst im äußeren Sinn „empirisch und physisch zu einem Körper bilde“, dessen Teile nicht nur „auf Zwecke hinweisen“, sondern deren „Zwekbestimmung“ sich der Verstand, der das so Zusammengefasste auf Begri e bringt, „durch die That bewust wird“ (LB Bodmer 3, S. 2). Das ist das Entscheidende! In dem, was ich selbst tue, in der eigenen Tat, erweist sich mir die reale Zweckmäßigkeit des physikalischen Organismus, zu dem ich mich selbst konstituiere, weil ich kann, was ich will: Ich erkenne durch die eigene Tat, dass ich – als körperliches Wesen – ein „Organismus der Natur“ und der eigenen Willkür zugleich bin und eigene Krä e in die Erscheinungen hineinlegen kann, um im Wechselspiel von actio und reactio die auf meinen Körper wirkenden Krä e der Dinge zu erfahren. In der Kritik der Urteilskra von 1790 hatte Kant es noch für eine „Eigentümlichkeit“ (§ 77) unseres diskursiven Verstandes gehalten, dass wir alle Organismen betrachten müssen, als ob sie Naturzwecke wären, „ohne doch zu behaupten, daß der Grund eines solchen Urtheils im Objecte liege“ (5:401).¹⁷ Denn: „A priori ist es sogar für uns unmöglich, einen solchen Begri [der Zweckmäßigkeit] seiner objektiven Realität nach als annehmungsfähig zu rechtfertigen“ (5:399). Dies hat sich 1799 grundlegend geändert. Eine bloße „als ob“ Einstellung kann meiner eigenen Organisation gegenüber gar nicht erst greifen. Sie könnte sich im Laufe weiterer Erfahrung gar nicht als falsch erweisen, da sie selbst Bedingung der Möglichkeit äußerer Erfahrung ist: „[D]er Begri eines organischen Körpers ist also a priori d.i. vor der Erfahrung aber zum Behuf der Moglichkeit der Erfahrung an meinem eigenen Subjecte gegeben“ (LB Bodmer 3, S. 2). Nun können wir allerdings noch einen Schritt weitergehen. Denn was ist das für ein Organismus, den ich in der Tat erfahre und der zugleich Bedingung der Möglichkeit der Erfahrung aller von mir unterschiedenen (und von mir zu unterscheidenden) Dinge ist? Es ist vor allem ein Körper, der symmetrische, aber nicht identische Teile hat, die ich relativ unabhängig voneinander betätigen kann. Ich werde überhaupt erst zu einem Erkunder im Raum, indem ich eine aufrechte Hal-

16 Kant, „Loses Blatt Bodmer 3“. In: Kant-Studien 95 (2004), 17–19. Zur Datierungsfrage vgl. Eckart Förster, „Zwei neu aufgefundene Lose Blätter zum Opus postumum“, ibid., 28. 17 „[S]o wird der Begri der Zweckmäßigkeit der Natur in ihren Producten ein für die menschliche Urtheilskra in Ansehung der Natur nothwendiger, aber nicht die Bestimmung der Objecte selbst angehender Begri sein, also ein subjectives Princip der Vernun für die Urtheilskra , welches als regulativ (nicht constitutiv) für unsere menschliche Urtheilskra eben so nothwendig gilt, als ob es ein objectives Princip wäre“ (§ 76).

Grenzen der Erkenntnis?

219

tung erwerbe und zu Laufen lerne; indem ich lerne, meine Arme und Hände zu bewegen und Dinge zu greifen; indem ich lerne, die Achsen meiner Augen zu kreuzen und damit den Blick auf etwas vor mir zu richten, und so weiter. Dies sind die Bedingungen dafür, dass ich zwischen oben/unten, links/rechts, vorn/hinten unterscheiden kann, was es wiederum ermöglicht, dass ich drei Dimensionen auch unabhängig von ihrem körperlichen Ursprung denken kann. Aber hören wir, wie Kant dies selbst ausdrückt: der Raum selbst [muß] als Erfahrungsgegenstand (spatium perceptibile) vorgestellt werden wenn gleich nur indirect, durch einen Zwischenbegrif, der Betastung seines eigenen Körpers nach seinen drey Abmessungen imgleichen der Händebewegung Linien zu ziehen sie durch Puncte zu begrenzen und so sich von Flächen als Grenzen endlich auch von einem körperlichen Raume empirisch eine Vorstellung zu machen, und sagen zu können es existirt etwas Räumliches und ist als das Gantze zur Einheit notwendig verbundener Warnehmungen ein Gegenstand möglicher Erfahrung (21:590).

Gegen Ende seiner langen Karriere ist Kant also zu der Einsicht gekommen, dass es die Inkongruenz meines eigenen Körpers ist, die die Gegenden im Raum bestimmt, und nicht anders herum, wie er ursprünglich angenommen hatte.

7 Was folgt daraus für die Kantische Behauptung der Erkenntnisgrenzen? Abschließend möchte ich nur noch darauf hinwiesen, dass Kant in der Konsequenz dieser Überlegungen sich auch dem eingangs erwähnten Grundgedanken Fichtes ein Stück weit angenähert hat. Kein Selbstbewusstsein ohne Gegenstandsbewusstsein, soviel hatte bereits die transzendentale Deduktion der Kategorien ergeben. Nun muss aber auch gelten: Kein Gegenstandsbewusstsein ohne organische Leiblichkeit. D.h., ich muss mich zu einem Gegenstand des äußeren Sinnes machen, als Bedingung des eigenen Selbstbewusstseins. Auch für Kant ist das Ich, was es ist, nur durch sich selbst, d.h. es setzt sein eigenes Sein. Und es muss sich dessen, was es ist, bewusst werden können. Es erkennt sich als Subjekt, indem es sich zum Objekt macht, und es macht sich zu einem bestimmten Objekt, indem es sich von anderem unterscheidet. D.h. es schreitet zur Erkenntnis seines eigenen Seins als eines bestimmten durch Hineinlegen bewegender Krä e in die „unbestimmte doch bestimmbare Anschauung“ (22:414). Die dazu notwendigen Schritte rekonstruiert die Transzendentalphilosophie und hebt damit ins Bewusstsein, was das Ich seinem Wesen nach ist. Damit hat auch Kant, ähnlich wie Fichte, die „Wurzel, in welcher die sinnliche und die übersinnliche Welt zusammenhängt“, sowie die „Ableitung beider Welten aus Einem Princip“ darzustellen begonnen:

220

Förster

Der erste Act des Erkenntnisses ist das Verbum: Ich bin das Selbstbewustseyn da Ich mir selbst Object bin: – Hierin liegt nun schon ein Verhältnis was vor aller Bestimmung des Subjects vorhergeht namlich das der Anschauung zu dem des Begri s wo das Ich doppelt d.i. in zwiefacher Bedeutung genommen wird indem ich mich selbst setze … Das Subject ist hier das Ding an sich weil es Spontaneität enthält. Die Erscheinung ist Receptivität. Jenes ist nicht ein anderes Object sondern eine andere Art sich selbst zum Object zu machen (22:413–5).

Tobias Rosefeldt

Dinge an sich und der Außenweltskeptizismus Über ein Missverständnis der frühen Kant-Rezeption Es ist in der Geschichte der Philosophie wahrscheinlich einmalig, dass sich unter ihren Hauptvertretern innerhalb nur weniger Jahre das Verständnis davon, auf welche Weise man Philosophie betreiben sollte, d.h. welche Fragen es in ihr zu beantworten gilt und mit welcher Methode man dies tun sollte, so radikal geändert hat, dass einer von ihnen – Kant – ein nur zwei Jahre nach seinem Tod erschienenes Hauptwerk eines anderen von ihnen – Hegels Phänomenologie des Geistes – wohl vollständig unverständlich gefunden hätte, wenn er es denn hätte lesen können. Eine der vielen sympathischen Eigenscha en von Rolf-Peter Horstmanns Beschä igung mit dem Deutschen Idealismus liegt darin, dass er das Phänomen der latenten Unverständlichkeit der Texte dieser philosophischen Strömung nie kaschiert hat, sondern es immer als Aufgabe der exegetischen Auseinandersetzung mit ihnen gesehen hat, diesem Phänomen plausible Erklärungen entgegenzusetzen, wie es zu den radikalen Änderungen im Philosophieverständnis und der philosophischen Methode in der nachkantischen Philosophie gekommen ist.¹ Nicht wenige dieser Erklärungen handeln von der unter Kants Nachfolgern allgemein verbreiteten Unzufriedenheit mit Kants Philosophie, d.h. von ihrer Meinung, dass Kant zwar Überragendes geleistet habe und der Philosophie eine völlig neue und erstmals die richtige Richtung gegeben habe, dass er aber andererseits die Voraussetzungen und Potentiale seines philosophischen Systems selbst nicht richtig durchschaut habe und man deswegen vom Wortlaut und manchmal auch der Methode der Kantischen Philosophie abweichen müsse, um ihrer eigentlichen Idee treu zu bleiben. Der folgende Beitrag wird in gewissem Sinne in der Tradition solcher Erklärungen stehen und einen Strang der ersten 15 Jahre der Entwicklungsgeschichte von Kant bis Hegel nachzeichnen. Er wird dabei von dem vielleicht wichtigsten Vorwurf gegen Kants System handeln. Dieser Vorwurf lautet, dass Kants Rede vom Ding an sich, genauer seine These, dass wir nie wissen können, wie die Dinge an sich selbst bescha en sind, sondern immer nur ihre Erscheinungen kennen, zu Inkonsistenzen in seinem System führt. Er hat viele von Kants unmittelbaren Nach-

1 Vgl. z.B. Horstmann, Rolf-Peter (1991). Die Grenzen der Vernun . Eine Untersuchung zu Zielen und Motiven des Deutschen Idealismus. Frankfurt: Verlag Anton Hain.

222

Rosefeldt

folgern davon überzeugt, dass man Kants transzendentalen Idealismus entweder ganz aufgeben sollte oder aber, um daraus ein konsistentes System zu machen, so radikalisieren muss, wie das etwa Fichte in seiner Version des Idealismus getan hat. Der Beitrag soll allerdings nicht ausschließlich davon handeln, anhand von Inkonsistenzen in Kants Lehre vom Ding an sich den Weg der Radikalisierung von Kants Idealismus zu demjenigen Fichtes verständlich zu machen. Ich möchte vielmehr dafür argumentieren, dass der Inkonsistenzvorwurf auf einem gewissen Missverständnis von Kants Lehre vom Ding an sich basiert. Ferner möchte ich zeigen, dass originellerweise der Erste, der diesem Missverständnis aufgesessen ist, Kant selbst ist. Als er das Missverständnis dann erkannt und revidiert hatte, hatte der philosophische Weltgeist sein Urteil über das Schicksal von Kants System bereits gesprochen, und den entscheidenden Figuren der nachkantischen Philosophie ist die Änderung in der Darstellung der Lehre vom Ding an sich gar nicht aufgefallen – so jedenfalls meine These. Ich werde im ersten Teil meines Beitrags diejenigen Textpassagen aus der ersten Au age der Kritik der reinen Vernun vorstellen, die meines Erachtens das Verständnis, das Kants Nachfolger von seiner Lehre vom Ding an sich hatten, maßgeblich bestimmt hat. Es handelt sich dabei um den Abschnitt über den sogenannten vierten Paralogismus der reinen Vernun , eine Passage, die Kant bekanntlich nicht in die zweite Au age der Kritik übernommen hat. Im zweiten Teil werde ich dann einen Schnelldurchlauf durch die Geschichte der Rezeption von Kants Lehre vom Ding an sich von 1782 bis 1797 unternehmen und die beiden wichtigsten Einwände vorstellen, die in dieser Zeit gegen diese Lehre vorgebracht worden sind. Ich werde ferner dafür argumentieren, dass diese Einwände vollkommen berechtigt sind und gegen mögliche Repliken aufrechterhalten werden können, wenn man Kants im ersten Teil referierte Selbstdarstellung zu Grunde legt, und dass Fichtes System vor diesem Hintergrund tatsächlich als konsequente Fortentwicklung der eigentlichen Kantischen Idee erscheinen muss. Im dritten, vierten und fün en Teil meines Beitrags werde ich dann zu zeigen versuchen, dass Kant spätestens in der zweiten Au age der Kritik der reinen Vernun seine Aussagen über Dinge an sich so revidiert hat, dass diese Aussagen sich einerseits gegen den Vorwurf der Inkonsistenz verteidigen lassen und nun andererseits Kants philosophisches Projekt von demjenigen Fichtes fundamental abgrenzen.

1 Transzendentaler und empirischer Idealismus In der Kritik der reinen Vernun argumentiert Kant bekanntlich dafür, dass Raum und Zeit bloße Formen unserer sinnlichen Anschauung sind und raum-zeitliche Eigenscha en folglich nicht den Dingen an sich, sondern nur ihren Erscheinun-

Dinge an sich und der Außenweltskeptizismus

223

gen zukommen. Dieser These gibt er den Titel „transzendentaler Idealismus“. Aus dem transzendentalen Idealismus folgt für Kant die Unerkennbarkeit der Dinge an sich, weil er annimmt, dass wir ohne Beteiligung unserer Sinne gar nichts erkennen können. Für den transzendentalen Idealismus führt Kant im wesentlichen zwei Argumente an. Das erste geht von der Annahme aus, dass wir von der raumzeitlichen Struktur der Welt a priori, d.h. unabhängig von der Erfahrung, Vorstellungen und Wissen haben, und versucht dann zu zeigen, dass dieser Sachverhalt nur dadurch erklärt werden kann, dass sich die Raum-Zeitlichkeit der von uns wahrgenommenen Welt nicht dieser Welt selbst, sondern der Struktur unseres Wahrnehmungsapparats verdankt. Das zweite Argument für den transzendentalen Idealismus soll zeigen, dass man nur mit Hilfe der These, dass die Dinge an sich nicht in Raum und Zeit existieren, bestimmte Probleme aus der philosophischen Tradition lösen kann, so z.B. das Problem der Vereinbarkeit von menschlicher Freiheit und kausaler Determiniertheit der Natur oder die sogenannten Antinomien des Weltanfangs oder der Teilung der Materie. In der ersten Au age der Kritik der reinen Vernun verfolgt Kant nun die Idee, den transzendentalen Idealismus noch durch ein weiteres, drittes Argument zu stützen. Das Argument ndet sich im Kapitel über die sogenannten „Paralogismen der reinen Vernun “ und basiert auf der These, dass nur der transzendentale Idealismus eine plausible Antwort auf den Außenweltskeptizismus geben kann. Der Außenweltskeptizismus behauptet, dass wir nie wirklich wissen können, ob es außer uns und unseren Vorstellungen auch irgendwelche materiellen, im Raum existierenden Gegenstände gibt. Kant verwendet für diesen Außenweltskeptizismus den Terminus „empirischer Idealismus“ und für dessen Gegenteil – die These, dass wir von der Existenz äußerer Gegenstände wissen können – den Terminus „empirischer Realismus“ (KrV A 369 f.). Diese Wortwahl erklärt sich sicher zum Teil daraus, dass er seinem Argument dadurch einen gewissen terminologischen P geben wollte. Das Argument für den transzendentalen Idealismus lautet nämlich, dass dessen Gegenteil – der transzendentale Realismus – unweigerlich zu einem empirischen Idealismus führt, der transzendentale Idealismus aber mit dem empirischen Realismus vereinbar ist (KrV A 370). In etwas leichter zugänglichen Worten: Wer annimmt, dass raum-zeitliche Eigenscha en den Dingen an sich zukommen, der kann sich nicht gegen den Außenweltskeptizismus wehren, d.h. muss akzeptieren, dass wir niemals wissen können, ob es materielle Körper im Raum gibt. Wer aber annimmt, dass räumliche Eigenscha en gar nicht den Dingen an sich, sondern nur Erscheinungen zukommen, der kann das Wissen von Körpern im Raum gegen den Skeptizismus verteidigen. Sehen wir uns Kants Argumentation etwas genauer an. Sie basiert darauf, dass Kant dem transzendentalen Realisten ein bestimmtes Modell davon unterstellt, wie wir von der Existenz von an sich existierenden Körpern wissen können.

224

Rosefeldt

Dieses Modell kann man in folgendem Schaubild zusammenfassen:

Subjekt

Vorstellungen nimmt unmittelbar wahr

Körper verursachen

Abb. 1: Transzendentaler Realismus

Diesem Modell liegt die Annahme zu Grunde, dass es sich bei den Dingen, die ein Subjekt unmittelbar wahrnehmen kann, immer nur um dessen eigene Vorstellungen handelt. Von der Existenz der von den durch diese Vorstellungen repräsentierten materiellen Gegenständen wissen wir nur deswegen, weil wir annehmen dürfen, dass unsere Vorstellungen in der Regel von Gegenständen verursacht werden, die ihnen entsprechen (vgl. KrV A 372). Kant meint nun, dass dieses Modell des Wissens von materiellen Körpern unweigerlich in den Außenweltskeptizismus führt. Er begründet das so: Wenn wir tatsächlich immer nur unsere eigenen Vorstellungen wahrnehmen, dann ist der Schluss auf materielle Körper als die Ursache dieser Vorstellungen unzulässig. Dieser Schluss wäre nur dann legitim, wenn wir zumindest manchmal einen Blick hinter unsere Vorstellungen werfen könnten, um zu sehen, was diese verursacht. Da alles Blicken im Haben neuer Vorstellungen besteht und uns dieser Blick hinter unsere Vorstellungen deswegen versagt ist, können wir nicht ausschließen, dass unsere Körpervorstellungen von etwas ganz anderem als von Körpern verursacht werden. Wir können nicht einmal ausschließen, dass wir selbst diese Ursache sind. In Kants eigenen Worten: […] der Schluss von einer gegebenen Wirkung auf eine bestimmte Ursache [ist] jederzeit unsicher, weil die Wirkung aus mehr als einer Ursache entsprungen sein kann. Demnach bleibt es in der Beziehung der Wahrnehmung auf ihre Ursache jederzeit zweifelha , ob diese innerlich oder äußerlich sei, ob also alle sogenannte äußere Wahrnehmungen nicht ein bloßes Spiel unseres innern Sinnes sein, oder ob sie sich auf äußere wirkliche Gegenstände als ihre Ursache beziehen (KrV A 368).

Kant argumentiert nun dafür, dass wir dem Außenweltskeptizismus entkommen, wenn wir den transzendentalen Realismus aufgeben und wie er selbst transzendentale Idealisten werden. Er begründet dies folgendermaßen: Der transscendentale Idealist kann hingegen ein empirischer Realist [sein], mithin […] das Dasein der Materie eben so auf das Zeugniß unseres bloßen Selbstbewußtseins [annehmen] und dadurch für bewiesen […] erklären, wie das Dasein [s]einer selbst als eines denkenden

Dinge an sich und der Außenweltskeptizismus

225

Wesens. Denn ich bin mir doch meiner Vorstellungen bewußt; also existieren diese und ich selbst, der ich diese Vorstellungen habe. Nun sind aber äußere Gegenstände (die Körper) blos Erscheinungen, mithin auch nichts anderes, als eine Art meiner Vorstellungen […]. Also existiren eben sowohl äußere Dinge, als ich selbst existiere, und zwar beide auf das unmittelbare Zeugniß meines Selbstbewußtseins […] (KrV A 370 f.).²

Was Kant hier sagt ist unter anderem deswegen interessant, weil er dabei eine sehr eindeutige Selbstinterpretation seines transzendentalen Idealismus liefert, insbesondere der These, dass raum-zeitliche Eigenscha en nur Erscheinungen zukommen. Dieser Passage zufolge sind mit „Erscheinungen“ schlicht unsere eigenen Vorstellungen gemeint, und die These besagt also, dass in Raum und Zeit ausgedehnte materielle Körper nichts anderes als unsere eigenen Vorstellungen sind. Es ist dann kaum verwunderlich, dass uns der transzendentale Idealismus ermöglicht, die skeptischen Bedenken hinsichtlich der Existenz materieller Körper auszuräumen. Von deren Existenz können wir nun nämlich ohne den problematischen Schluss von unseren Vorstellungen auf deren Ursachen wissen, und zwar ohne die Annahme aufgeben zu müssen, dass wir unmittelbar immer nur unsere eigenen Vorstellungen wahrnehmen. Denn da Körper ja schlicht nichts anderes sind als bestimmte unserer eigenen Vorstellungen, können wir auch diese unmittelbar wahrnehmen. So gradlinig diese Widerlegung des Außenweltskeptizismus ist, so sehr hat man doch das ungute Gefühl, dass sie etwas billig ist – oder vielleicht sollte man eher sagen: zu teuer erkau . Angesichts von Kants Versicherung, dass wir sehr wohl von der Existenz materieller Gegenstände wissen können, weil solche materiellen Gegenstände in Wirklichkeit nichts als Vorstellungen sind, fühlt man sich ein bisschen so wie ein Verdurstender in der Wüste, dem man erst sagt: „ja, die Oase am Horizont existiert tatsächlich“, um dann nach einer kurzen und gemeinen Pause die Begründung hinterher zu schieben „die Oase ist schließlich nichts anderes als eine Fata Morgana und Fata Morganen sind reale optische Phänomene“. Der Verdurstende interessiert sich für echte Oasen mit echtem Wasser, und der Außenweltskeptiker dafür, ob es außer einem selbst und den eigenen Vorstellungen noch etwas anderes in der Welt gibt. Es ist deswegen interessant zu fragen, ob Kant zugesteht, dass es außer den Körpern, die nichts als Vorstellungen sind, auch vorstellungstranszendente Gegenstände gibt, die diese Vorstellungen in uns hervorrufen. Schauen wir uns zur Beantwortung dieser Frage wieder ein Zitat an:

2 Vgl. auch KrV A 370: „Der transscendentale Idealist kann hingegen ein empirischer Realist, mithin, wie man ihn nennt, ein Dualist sein, d.i. die Existenz der Materie einräumen, ohne aus dem bloßen Selbstbewußtsein hinauszugehen und etwas mehr als die Gewißheit der Vorstellungen in mir, mithin das cogito, ergo sum anzunehmen.“

226

Rosefeldt

Nun kann man zwar einräumen, daß von unseren äußeren Anschauungen etwas, was im transscendentalen Verstande außer uns sein mag, die Ursache sei; aber dieses ist nicht der Gegenstand, den wir unter den Vorstellungen der Materie und körperlicher Dinge verstehen; denn diese sind lediglich Erscheinungen, d.i. bloße Vorstellungsarten, die sich jederzeit nur in uns be nden, und deren Wirklichkeit auf dem unmittelbaren Bewußtsein eben so, wie das Bewußtsein meiner eigenen Gedanken beruht. Der transscendentale Gegenstand ist sowohl in Ansehung der inneren als äußeren Anschauung gleich unbekannt (A 372).

Diese Passage erweitert das bisherige Erkenntnismodell des transzendentalen Idealismus um die Idee, dass unsere Vorstellungen von Dingen verursacht werden, die „im transzendentalen Sinne außer uns sind“, d.h. die unabhängig von uns und unseren Vorstellungen – oder auch: an sich selbst – existieren. Wieder kann man das Modell mit einem Schaubild veranschaulichen:

Subjekt

Körper = Vorstellungen

nimmt unmittelbar wahr

Dinge an sich verursachen

Abb. 2: Transzendentaler Idealismus 1

Die entscheidende Frage, die sich angesichts der eben zitierten Textstelle stellt, ist nun, welchen Status Kant dem Sachverhalt zuschreibt, der durch den grau unterlegten Teil dieses Schaubilds repräsentiert wird. Ist dies ein Sachverhalt, dessen Bestehen er explizit behaupten will? Oder einer, dessen Bestehen er bloß in Erwägung zieht und nicht ausschließen will, ohne es allerdings selbst zu bekrä igen? Ich denke, dass die Textpassage beide Interpretationen zulässt, je nachdem wie man den Satz versteht: „Nun kann man zwar einräumen, dass von unseren äußeren Anschauungen etwas, was im transscendentalen Verstande außer uns sein mag, die Ursache sei.“ Dass man die Existenz von vorstellungstranszendenten Dingen an sich als Ursache unserer Vorstellungen einräumen mag, könnte nach einer ersten Interpretation so viel heißen, dass Kant nicht bestreiten will, dass es außer uns und unseren Vorstellungen auch noch etwas anderes in der Welt gibt, und nur bestreitet, dass wir wissen können, dass dies tatsächlich der Fall ist. Nach einer zweiten Interpretation würde Kant ein bisschen mehr zugestehen und selbst behaupten, dass es von uns und unseren Vorstellungen verschiedene Dinge an sich gibt und dass diese Dinge Ursache für unsere Vorstellungen sind; bestreiten würde er lediglich, dass diese Dinge an sich Körper in Raum und Zeit sind. Es wird sich im Folgenden zeigen, dass es nicht nur die Unklarheit des Kantischen Textes hinsichtlich dieser Interpretationsfrage war, die Kants Nachfolger

Dinge an sich und der Außenweltskeptizismus

227

an seiner Rede vom Ding an sich zu beanstanden hatten. Vielmehr war man bald mehrheitlich der Meinung, dass Kant im Rahmen seines Systems der These über Dinge an sich in keiner der beiden Lesarten zustimmen dür e.

2 Die Kritik an Kants Idealismus Die erste Rezension von Kants Kritik der reinen Vernun erschien 1782 in den Göttinger gelehrten Anzeigen.³ Sie stammt ursprünglich von Christian Garve, wurde aber vor der Verö entlichung vom Herausgeber der Zeitschri , dem Göttinger Philosophieprofessor Johann Georg Heinrich Feder, massiv redigiert, wobei Feder vor allem das bei Garve durchaus noch zu ndende Lob des Buches abgeschwächt und dessen Kritik noch einmal forciert hat. Der für unseren Zusammenhang wichtigste Vorwurf der Rezensenten lautet: Kants transzendentaler Idealismus besagt letztlich genau dasselbe wie der Idealismus von George Berkeley und das bedeutet, dass seine Theorie ebenso absurd ist, wie es von einem Idealismus à la Berkeley gemeinhin angenommen wurde.⁴ Nennen wir diesen Vorwurf den Absurditätsvorwurf. Wie berechtigt ist der Vergleich mit Berkeley und der damit einhergehende Absurditätsvorwurf? Zwar gibt es sicher gravierende Unterschiede zwischen Kant und Berkeley, insbesondere was die Motivation ihres jeweiligen Idealismus betri . Die zuvor dargestellte Kantische Überlegung – wir können von der Existenz von Körpern wissen, weil wir diese unmittelbar wahrnehmen, und wir können Körper unmittelbar wahrnehmen, weil Körper nichts anderes als unsere eigenen Vorstellungen sind – ndet sich aber genau so auch bei Berkeley. Auch die These, dass ein solcher Idealismus absurd ist, ist nicht von der Hand zu weisen. Feder hat dies 1787, also fünf Jahre nach Erscheinen der Rezension, in seinem Buch Über Raum und Kausalität noch einmal auf den Punkt gebracht. Er schreibt dort:

3 Vgl. Feder, Johann Georg Heinrich; Garve, Christian (1991 [1782]). „Critik der reinen Vernun von Immanuel Kant (Rezension)“. In: A. Landau (Hg.): Rezensionen zur Kantischen Philosophie 1781–87. Bebra: Albert Landau Verlag, 10–17. Ich kann hier nicht auf den Ein uss anderer Rezensionen Kantischer Werke für die Entwicklung von Kants Idealismus eingehen, obwohl dies ohne Zweifel lohnend wäre. Eine ausführliche Auseinandersetzung mit diesem Ein uss ndet sich in Sassen, Brigitte (1997). „Critical Idealism in the Eyes of Kant’s Contemporaries“. In: Journal of the History of Philosophy 35, 421–455. 4 Feder, J. G. H.; Garve, C. (1991 [1782]). Critik der reinen Vernun , 13. Der Vergleich mit Berkeley lag für Garve und Feder, die beide stark vom Britischen Empirismus, insbesondere von Locke beein usst waren, nahe.

228

Rosefeldt

Ich habe nie das Glück gehabt, Königsberg und seinen berühmten Philosophen zu sehen; ob ich gleich eine lebha e und in vielen Stücken richtige Vorstellung von ihm habe. Viele Personen, die ihn von Angesicht kennen, lieben und verehren, habe ich gesprochen. Aber wer darf Anstand nehmen, ihm ein von den Vorstellungen derer die ihn gesehen haben, so wohl als von der meinigen, unabhängiges Daseyn zuzugestehen?⁵

Das ist natürlich eine rhetorische Frage und man kann sich kaum vorstellen, dass Kant selbst – und angesichts des von Feder gewählten Beispiels möchte man sagen: gerade er – sie anders beantwortet hätte als mit: „Niemand nimmt daran Anstand.“ Und in der Tat hat Kant bereits in den Prolegomena eine Passage eingefügt, in der er deutlich zu machen versucht, dass er mit seinem transzendentalen Idealismus keinen Idealismus à la Berkeley vertreten wollte. Er schreibt dort: Der Idealismus besteht in der Behauptung, daß es keine andere als denkende Wesen gebe, die übrige Dinge, die wir in der Anschauung wahrzunehmen glauben, wären nur Vorstellungen in den denkenden Wesen, denen in der That kein außerhalb diesen be ndlicher Gegenstand correspondirte. Ich dagegen sage: es sind uns Dinge als außer uns be ndliche Gegenstände unserer Sinne gegeben, allein von dem, was sie an sich selbst sein mögen, wissen wir nichts, sondern kennen nur ihre Erscheinungen, d.i. die Vorstellungen, die sie in uns wirken, indem sie unsere Sinne a ciren. […] Kann man dieses wohl Idealismus nennen? Es ist gerade das Gegentheil davon (AA IV 288 f.).

Ich werde später noch einmal ausführlich auf diese Passage zurückkommen. Im Moment interessiert mich nur Kants Behauptung, dass seine Position „gerade das Gegentheil“ vom Idealismus ist, weil sie beinhalte, dass unseren Vorstellungen wirkliche Gegenstände korrespondieren, die von diesen Vorstellungen und von uns selbst unterschieden sind, und nur bestreiten will, dass wir wissen, wie genau diese Gegenstände an sich selbst bescha en sind. Projiziert man diese Aussage zurück auf das Erkenntnismodell, das Kant in der Skeptizismuswiderlegung in der Kritik der reinen Vernun vertreten hatte, dann muss man folgern, dass Kant seine Behauptung über den fraglichen Sachverhalt eindeutig im Sinne der zweiten Lesart verstanden wissen will: Er will nicht nur nicht ausschließen, sondern behaupten, dass unsere Körpervorstellungen von vorstellungstranszendenten Dingen an sich verursacht werden. Diese Selbstauskun wir allerdings die Frage auf, ob Kant diese These im Rahmen seines Systems überhaupt vertreten darf. Gegen eben diese Annahme richtet sich der zweite berühmte Einwand gegen Kants Rede vom Ding an sich, zu dem ich nun komme. Diesen Einwand kann man als Inkonsistenzvorwurf bezeichnen. Er hat die folgende Struktur: Erst wird gezeigt, dass Kant die Behauptung der A ektion durch

5 Feder, Johann Georg Heinrich (1787). Ueber Raum und Causalität: zur Prüfung der Kantischen Philosophie. Göttingen: Dietrich, 78.

Dinge an sich und der Außenweltskeptizismus

229

subjektunabhängige Gegenstände nicht nur machen muss, um dem Absurditätsvorwurf zu entgehen, sondern dass diese Behauptung ein integraler Bestandteil seiner Theorie menschlicher Erkenntnis ist. Nennen wir diesen Teil des Vorwurfs die Unverzichtbarkeitsbehauptung. Dann wird dafür argumentiert, dass Kant die genannte Behauptung aus verschiedenen Gründen nicht machen dür e, wenn seine Theorie menschlicher Erkenntnis wahr wäre. Dieser Teil des Einwands soll im folgenden Unmöglichkeitsbehauptung genannt werden. Der Inkonsistenzvorwurf ist erstmals von Friedrich Heinrich Jacobi im berühmten Anhang „Über den transzendentalen Idealismus“ zu seinem Buch David Hume über den Glauben, oder Idealismus und Realismus von 1787 erhoben worden. 1792 wird er in Gottlob Ernst Schulzes anonym erschienener Schri Aenesidemus oder über die Fundamente der von dem Herrn Professor Reinhold in Jena gelieferten Elementar-Philosophie wiederholt, zum Teil terminologisch präzisiert und zudem gegen die These verteidigt, dass man dem Inkonsistenzvorwurf durch Kants Revisionen in der zweiten Au age der Kritik oder durch Reinholds Reaktionen auf Jacobi entgehen könnte. Da ich der Meinung bin, dass Jacobi und Schulze Kant im wesentlichen auf dieselbe Weise kritisieren wollen, werde ich den Inkonsistenzvorwurf parallel anhand von Aussagen beider Philosophen darstellen.⁶ Kommen wir zunächst zur Unverzichtbarkeitsbehauptung. Jacobi stellt fest, dass Kant tatsächlich annehmen muss, dass es außer den Körpern, die wir erkennen können, die aber nichts als Vorstellungen sind, auch vorstellungsunabhängige Dinge an sich als Ursachen unserer Vorstellungen gibt. Diese Annahme folgt schlicht aus dem Ausgangssetting, von dem aus Kant in der transzendentalen Ästhetik seine Argumente für den transzendentalen Idealismus entwickelt. Dieses Setting beinhaltet nämlich, dass wir erstens keine Erkenntnis ohne Anschauung haben können, diese Anschauung zweitens bei uns nur sinnlich sein kann und dass drittens sinnliche Vorstellungen als solche de niert sind, die nur durch Einwirkung von Gegenständen auf das Subjekt – die sogenannte ‚A ektion‘ – hervorgebracht werden können (vgl. KrV A 19/B 33). Auf diesen Begri von Sinnlichkeit bezieht sich Jacobi, wenn er schreibt: […] das Wort Sinnlichkeit ist ohne alle Bedeutung, wenn nicht ein distinctes reales Medium zwischen Realem und Realem, ein würkliches Mittel von Etwas zu Etwas darunter verstan-

6 Birgit Sandkaulen hat vor kurzem auf sehr subtile Weise dafür argumentiert, dass Jacobis Kritik sich von derjenigen Schulzes fundamental unterscheidet (vgl. Sandkaulen, Birgit (2007). „Das ‚leidige Ding an sich‘. Kant – Jacobi – Fichte“. In: J. Stolzenberg (Hg.). Kant und der Frühidealismus. Hamburg: Felix Meiner Verlag, 175–201.). Ich möchte durch die Parallelbehandlung von Jacobis und Schulzes Kritik zeigen, dass die Unterschiede zwischen den beiden Kritiken weniger dramatisch sind, als Sandkaulen behauptet, und werde in zwei weiteren Fußnoten auch noch etwas konkretere Einwände gegen Sandkaulens Interpretation formulieren.

230

Rosefeldt

den werden, und in seinem Begri e, die Begri e von ausseinander- und verknüpt seyn, von Thun und Leiden, von Causalität und Dependenz, als realer und objektiver Bestimmungen schon enthalten seyn sollen […].⁷

Man kann diese Passage als Stellungnahme zu der Frage lesen, welche der in Abschnitt 1 erwähnten Gegenstände – die von uns laut Kant erkennbaren Körper, die selber nichts als Vorstellungen sind, oder die Gegenstände, die „im transzendentalen Verstande außer uns“, d.h. von uns und unseren Vorstellungen unterschieden sind – als zweites Relatum der A ektionsrelation in Frage kommen. Laut Jacobi können das nur letztere sein. Sehr plausibel an dieser Behauptung ist die These, dass es sich bei den a zierenden Gegenständen nicht um die empirisch erkennbaren Körper handeln kann, wenn diese gemäß der Konzeption im vierten Paralogismus verstanden werden; denn die Rede von dem Hervorbingen von Vorstellungen durch Gegenstände scheint keinen rechten Sinn zu ergeben, sollte es sich bei diesen Gegenständen selbst um Vorstellungen handeln. Jacobis Rede von einem „distincten realem Medium zwischen Realem und Realem“, in dessen Begri „die Begri e von auseinander- und verknüp seyn“ enthalten sind, legt zudem nahe, dass er nicht nur behaupten will, dass der a zierende Gegenstand für Kant nicht selbst eine Vorstellung sein kann, sondern dass er im Fall äußerer Anschauung auch numerisch verschieden vom vorstellenden Subjekt sein muss. Obwohl man diese These nicht mit einer Analyse des Begri s der A ektion begründen kann (der schließlich Selbsta ektion nicht ausschließt), gibt es gute Gründe dafür, dass sie Kants Konzeption sinnlicher Erkenntnis korrekt beschreibt. Erstens legt die Weise, auf die Kant in der transzendentalen Ästhetik von Sinnlichkeit spricht, ein solches nicht-solipsistisches Verständnis von A ektion nahe. Zweitens charakterisiert Kant in der oben zitierten Passage aus A 372 den vermeintlich vorstellungstranszendenten Gegenstand äußerer Anschauungen als einen, der „im transzendentalen Verstande außer uns“ ist. Es macht aber wenig Sinn, davon zu sprechen, dass das erkennende Subjekt selbst im transzendentalen Verstand außer sich ist. Und drittens deckt sich Jacobis Interpretation sehr gut mit Kants zitierter Selbstinterpretation in der Prolegomena, in der Kant sich auf die These festlegt, dass es nicht bloß denkende Wesen und ihre Vorstellungen gibt. Es gibt also sehr gute Gründe, Kant mit Jacobi die Annahme zu unterstellen, dass es von uns und unseren Vorstellungen unterschiedene, an sich selbst existierende Gegenstände sind, die uns a zieren.

7 Jacobi, Friedrich Heinrich (2004 [1785]). „Ueber den Transscendentalen Idealismus“. In: Werke – Gesamtausgabe. Hg. von K. Hammacher und W. Jaeschke. Band 2,1. Schri en zum transzendentalen Idealismus. Hg. von W. Jaeschke und I.-M. Piske. Hamburg: Felix Meiner Verlag, 109.

Dinge an sich und der Außenweltskeptizismus

231

Dass es ein unverzichtbarer Bestandteil von Kants Lehre sinnlicher Erkenntnis ist, als zweites Relatum der A ektionsbeziehung nicht eine Kantische Erscheinung, sondern einen subjektunabhängigen Gegenstand anzunehmen, meint auch Schulze. Er schreibt: Der Gegenstand außer unsern Vorstellungen, (das Ding an sich) der nach der Vernun kritik durch Ein uß auf unsere Sinnlichkeit die Materialien der Anschauung geliefert haben soll, ist nun aber nicht selbst wieder eine Anschauung oder sinnliche Vorstellung, sondern es soll etwas von denselben realiter Verschiedenes und Unabhängiges sein […].⁸

Dass die A ektionsbeziehung also – in Jacobis Worten – „ein distinktes reales Medium zwischen Realem und Realem“ ist, führt Schulze vor allem gegen Reinholds Versuch, Kants Theorie gegen den Inkonsistenzvorwurf zu verteidigen, an. Darauf werde ich unten noch einmal zurückkommen. Wenden wir uns zuerst dem zweiten Teil dieses Vorwurfs zu, der Unmöglichkeitsbehauptung. Jacobi argumentiert dafür, dass die Annahme der Existenz von subjektunabhängigen Dingen an sich als Ursachen unserer sinnlichen Vorstellungen gegen die Erkenntnisbeschränkungen verstößt, die ein wesentlicher Bestand-

8 Schulze, Gottlob Ernst (1996 [1792]): Aenesidemus oder über die Fundamente der von Herrn Professor Reinhold in Jena gelieferten Elementar-Philosophie. Nebst einer Verteidigung des Skeptizismus gegen die Anmaßungen der Vernun kritik. Hg. von M. Frank. Hamburg: Felix Meiner, 184 (Originalpaginierung 264). Sandkaulen schreibt zum Unterschied zwischen Jacobis und Schulzes Kant-Kritik Folgendes: „Kant eine missbräuchliche Anwendung der Kategorien über das Gebiet der Erscheinungen hinaus zu unterstellen [gemeint ist: wie Schulze das tut; T. R.], bedeutet nämlich, den Theorierahmen der transzendentalen Re exion vorauszusetzen und zugleich zu behaupten, daß Kant sich mittels einer Verobjektivierung eben dieses Rahmens eine Dimension erschlossen habe, in die er unter den Restriktionsbedingungen seiner Theorie nicht hätte vordringen dürfen. Jacobi hingegen kommt es auf etwas ganz anderes an. Unverzichtbar und zugleich unhaltbar ist die Voraussetzung a zierender Gegenstände hier nicht deshalb, weil sie sich einer fälschlichen Überdehnung des Theorierahmens verdankte, sondern deshalb, weil sie in einen theoriefremden Hohlraum fällt, den Kant sich mit den Mitteln des transzendentalen Idealismus in gar keiner Weise aneignen kann“ (Sandkaulen, Birgit (2007). „Das ‚leidige Ding an sich‘“, 188 f.). Das klingt so, als würde Schulze – anders als Jacobi – behaupten, dass Kant zu der Annahme afzierender Dinge an sich nur dadurch gelangen kann, dass er die Kausalitätskategorie über den Bereich der Erscheinungen hinaus ausweitet und schlussfolgert, dass es a zierende Gegenstände geben muss, wenn wir Vorstellungen in uns haben. Wie die zitierte Passage deutlich macht, behandelt Schulze Kants Annahme a zierender Dinge an sich aber genauso wie Jacobi als eine nicht weiter begründete Prämisse von Kants System. Die Rede von einem unzulässigen Kategoriengebrauch (der „fälschlichen Überdehnung des Theorierahmens“, wie Sandkaulen es nennt) bringt er nicht im Zuge der Begründung der hier so genannten Unverzichtbarkeitsbehauptung ins Spiel, sondern erst im Rahmen seiner Kritik der im Folgenden dargestellten Unmöglichkeitsbehauptung, der These also, dass die vorausgesetzte Rede von a zierenden Gegenständen gegen Kants eigene Beschränkung des Kategoriengebrauchs verstößt.

232

Rosefeldt

teil von Kants kritischer Philosophie sind. Um seine Kritik zu verstehen, ist es hilfreich, sich diejenige Passage anzusehen, in der er explizit auf Kants oben zitierte Aussage aus dem Paralogismenkapitel eingeht, dass man die Existenz eines transzendentalen Gegenstands „einräumen mag“ (A 372): Denn wenn nach ihr [der Kantischen Philosophie; T. R.] auch eingeräumt werden kann, daß diesen bloß subjektiven Wesen, die nur Bestimmungen unseres eigenen Wesens sind, ein transscendentales Etwas als Ursache entsprechen mag:so bleibt doch in der tiefsten Dunkelheit verborgen, wo diese Ursache und von was Art ihre Beziehung auf die Würkung sei.⁹

Wie ich im Folgenden darlegen möchte, führt Jacobi in dieser Passage genau genommen drei Gründe dafür an, dass Kant im Rahmen seines Systems eine Affektion durch subjektunabhängige Gegenstände nicht behaupten darf. Diese drei Gründe nden sich auch in anderen Passagen seines Textes und zudem in ähnlicher Form bei Schulze. 1. Grund: Kant ist nicht berechtigt anzunehmen, dass die vorstellungstranszendente Ursache unserer Vorstellungen ein von uns selbst verschiedener Gegenstand ist. Dieser Grund kommt in Jacobis Bemerkung zum Ausdruck, dass Kant selbst dann, wenn er zu Recht eine vorstellungstranszendente Ursache unserer Vorstellungen annimmt, keine Behauptungen darüber machen dürfe, „wo“ diese Ursache ist, d.h. – das darf man wohl ergänzen – ob sie in uns selbst oder in einem von uns verschiedenen Gegenstand liegt. Noch deutlicher ist dieser Punkt bei Schulze formuliert, der schreibt, dass Kants Theorie der sinnlichen Erkenntnis nicht nur impliziert, dass „wir uns überhaupt einen Grund unserer Erfahrungskenntnisse denken müssen“, sondern auch, dass „dieses Etwas für ein von dem Gemüte verschiedenes Ding an sich gehalten werden müsse“, und dann einwendet, dass „aber auch das Gemüt als der alleinige Grund aller unserer Erkenntnis gedacht werden“ kann und Kant nicht berechtigt ist, diese Möglichkeit auszuschließen.¹⁰ Angesichts von Kants Kritik am Erkenntnismodell des transzendentalen Realismus im vierten Paralogismus scheint dieser Einwand sehr berechtigt. Kants Vorwurf dort war, dass in diesem Modell ein unzulässiger Schluss von einer gegebenen Vorstellung (einer Körpervorstellung) auf eine bestimmte Ursache (einen Körper) vorliegt. Nun scheint aber Kant einen ebensolchen Schluss von einer gegebenen Vorstellung auf

9 Jacobi, F. H. (2004 [1785]). „Ueber den Transscendentalen Idealismus“, 110. 10 Schulze, G. E. (1996 [1792]): Aenesidemus oder über die Fundamente der von Herrn Professor Reinhold in Jena gelieferten Elementar-Philosophie, 185 (OP 265).

Dinge an sich und der Außenweltskeptizismus

233

eine bestimmte Ursache vorzunehmen, wenn er auf die Existenz von subjektunabhängigen Dingen an sich als Ursachen für unsere Vorstellungen schließt, denn er schließt mit dieser Annahme ja zumindest aus, dass die Ursache für unsere Vorstellungen wir selbst sind. 2. Grund: Kant ist nicht berechtigt anzunehmen, dass es überhaupt vorstellungstranszendente Ursachen unserer Vorstellungen gibt. Durch die Hervorhebung des Wortes „mag“ scheint Jacobi zu betonen, dass Kant nicht behaupten, sondern nur nicht ausschließen darf, dass es überhaupt vorstellungstranszendente Ursachen von Vorstellungen gibt. Das heißt, dass auch der Schluss von einem gegebenen Sachverhalt auf irgendeine Ursache dieses Sachverhalts in Rahmen von Kants System unzulässig ist, wenn es sich nicht um eine Ursache aus dem Bereich der Erscheinungen handelt. Jacobi kann sich mit dieser These auf Kants für die Kritik der reinen Vernun zentrale Behauptung berufen, dass die reinen Verstandesbegri e – wie die von Ursache und Wirkung – sowie die damit verbundenen Grundsätze des reinen Verstandes – wie der Grundsatz, dass es zu jedem Ereignis eine Ursache gibt (vgl. A 189/B 231), – nur für den Bereich der Erscheinungen, nicht aber für Dinge an sich Geltung haben. Jacobi schreibt: Unsere allgemeinen Vorstellungen, Begri e und Grundsätze [d.h. die reinen Verstandesbegri e und die Grundsätze des reinen Verstandes; T. R.] drücken nur die wesentliche Form aus, in welche jede besondere Vorstellung und jedes besondere Urteil zufolge der Beschaffenheit unserer Natur sich fügen muß, um in Einem allgemeinen oder transscendentalen Bewußtseyn aufgenommen und verknüp werden zu können und dergestalt relative Wahrheit, oder relativ objective Gültigkeit zu erhalten. Aber diese Gesetze unseres Anschauens und Denkens sind, wenn man von der menschlichen Form abstrahiert, ohne alle Bedeutung und Gültigkeit, und geben über die Gesetze der Natur an sich nicht die entfernteste Weisung. Weder der Satz des zureichenden Grundes, noch selbst der Satz, daß aus Nichts Nichts werden kann, geht die Dinge an sich an.¹¹

Die Gültigkeit der zuletzt genannten Prinzipien über den Bereich der Erscheinungen hinaus, müsste Kant aber annehmen, wenn er darauf schließen will, dass es eine vorstellungstranszendente Ursache unserer Vorstellungen gibt. Schulze hat dieses Argument Jacobis terminologisch auf den Punkt derjenigen Formulierung gebracht, unter der es seitdem diskutiert wird, wenn er schreibt, dass Kants Annahme von a zierenden Dingen an sich eine im Rahmen seiner

11 Jacobi, F. H. (2004 [1785]). „Ueber den Transscendentalen Idealismus“, 110 f.

234

Rosefeldt

Theorie unzulässige Anwendung der Kategorie der Kausalität impliziere.¹² Wie die folgenden Zitate zeigen, versteckt sich hinter dieser Formulierung – zumindest in einer Lesart¹³ – aber genau dasselbe Argument wie bei Jacobi: Nach der transzendentalen Deduktion der reinen Verstandesbegri e […] sollen […] die Kategorien Ursache und Wirklichkeit nur auf empirische Anschauungen, nur auf etwas, so in der Zeit wahrgenommen worden ist, angewendet werden dürfen, und außer dieser Anwendung sollen die Kategorien weder Sinn noch Bedeutung haben. […] [A]lso darf auf [das Ding an sich] nach den eigenen Resultaten der Vernun kritik weder der Begri Ursache, noch auch der Begri Wirklichkeit angewendet werden […].¹⁴

Dass Schulze so wie Jacobi der Meinung ist, dass die Einschränkung des Kategoriengebrauchs deswegen problematisch ist, weil damit eine Einschränkung der mit den Kategorien jeweils verbundenen Grundsätze einhergeht und man diese Grundsätze in uneingeschränkter Form voraussetzen muss, um auf Dinge an sich als Ursachen unserer Vorstellungen zu schließen, wird an der folgenden Stelle deutlich: Die kritische Philosophie behauptet nun allerdings wohl, daß es […] Dinge an sich objektiv gäbe, und daß sie der Real-Grund des Inhalts unserer Erfahrungskenntnisse seien: Allein sie behauptet dies ohne allen Grund, und hat durch ihre Lehren über die Natur und Bestimmung der Grundsätze des reinen Verstandes und der reinen Vernun alle Möglichkeit, jene Behauptung zu erweisen, gänzlich zerstört.¹⁵

3. Grund: Kant ist nicht einmal berechtigt, es für möglich zu halten, dass es vorstellungstranszendente Ursachen unserer Vorstellungen gibt.

12 Sandkaulen ist der Meinung, dass „Jacobi das Argument fälschlich in Anwendung gebrachter Kategorien und insbesondere der Kausalkategorie nicht gebraucht“ (Sandkaulen, Birgit (2007). „Das ‚leidige Ding an sich‘“, 180). Angesichts der zuletzt zitierten Passage aus dem Jacobi-Text fällt es mir schwer, dem zuzustimmen. Zwar wird der Term „Kategorie“ dort nicht verwendet, aber es ist aus dem Kontext klar, dass mit dem Ausdruck „unsere allgemeinen Vorstellungen, Begri e und Grundsätze“ die reinen Verstandesbegri e und die Grundsätze des reinen Verstandes gemeint sind. Diese dürfen laut Jacobi nur dazu dienen, Vorstellungen in ein allgemeines transzendentales Bewusstsein zu vereinigen, und verlieren „alle Bedeutung und Gültigkeit“, wenn man sie auf die „Natur an sich“ anwendet. Schulze hat mit seiner Rede von Kants unzulässigem Gebrauch der Kategorien also nur einen Vorwurf terminologisch auf den Punkt gebracht, der sich auch bei Jacobi ndet. 13 Zur anderen Lesart gleich mehr. 14 Schulze, G. E. (1996 [1792]): Aenesidemus oder über die Fundamente der von Herrn Professor Reinhold in Jena gelieferten Elementar-Philosophie, 184 (OP 263 f.). 15 Schulze, G. E. (1996 [1792]): Aenesidemus oder über die Fundamente der von Herrn Professor Reinhold in Jena gelieferten Elementar-Philosophie, 206 (OP 296 f.).

Dinge an sich und der Außenweltskeptizismus

235

Es gibt einige Passagen bei Schulze und Jacobi, die so klingen, als behaupteten diese, dass Thesen über vorstellungstranszendente Ursachen innerhalb von Kants System nicht bloß unbegründet, sondern sogar sinnlos seien. In dem vorletzten Zitat aus dem Aenesidemus kommt diese Kritik zum Beispiel dann zum Ausdruck, wenn Schulze – in direkter Anlehnung an eine Passage aus der Kritik der reinen Vernun (vgl. A 240/B 298 f.) – schreibt, dass die Kategorien für Kant „weder Sinn noch Bedeutung haben“, wenn wir sie auf Dinge außerhalb der Zeit (wie Dinge an sich) anwenden. Einen ähnlichen Vorwurf könnte man hinter Jacobis obiger Formulierung vermuten, dass dann, wenn wir von vorstellungstranszendenten Ursachen sprechen, „in der tiefsten Dunkelheit verborgen [bleibt], […] von welcher Art ihre Beziehung auf die Wirkung sei“.¹⁶ Die genannten drei Gründe für die Unmöglichkeitsbehauptung generieren zusammen mit der ebenfalls gut begründeten Unverzichtbarkeitsbehauptung eine scheinbar unüberwindliche Spannung in Kants System. Jacobi bringt sie mit seinen berühmten Worten über Kants Voraussetzung a zierender Dinge an sich zum Ausdruck: Ich muß gestehen, daß dieser Anstand mich bey dem Studio der Kantischen Philosophie nicht wenig aufgehalten hat, so daß ich verschiedene Jahre hintereinander die Critik der reinen Vernun immer wieder von vorne anfangen mußte, weil ich unau örlich darüber irre wurde, daß ich ohne jene Voraussetzung in das System nicht hineinkommen, und mit jener Voraussetzung darinn nicht bleiben konnte.¹⁷

Wie kann man auf die Spannung zwischen Unverzichtbarkeitsbehauptung und Unmöglichkeitsbehauptung reagieren, wenn man grundsätzlich Sympathien für Kants Projekt hat? Zweifellos am besten dadurch, dass man zeigt, dass eine dieser beiden Behauptungen falsch ist. Im Falle der Unmöglichkeitsbehauptung müsste man dazu alle drei von Jacobi und Schulze angeführten Gründe für diese Behauptung ausräumen. Eine Möglichkeit auf den dritten Grund zu reagieren, besteht darin, Kants Aussagen über die Sinnlosigkeit von Aussagen über Dinge, die nicht in Raum und Zeit sind, weniger stark zu lesen, als der Einwand dies voraussetzt. Und

16 Jacobi, F. H. (2004 [1785]). „Ueber den Transscendentalen Idealismus“, 110. Sollte dieser dritte Grund tatsächlich eine Rolle für Jacobis Unmöglichkeitsbehauptung gespielt haben, würde das erklären, weshalb er schreibt, Kants transzendentaler Idealismus sei „mit sich selbst in wahrha unaussprechliche Widersprüche geraten“ (ebd. 112). Die These, dass die Inkonsistenz in Kants System „unaussprechlich“ ist – Jacobi betont diese Charakterisierung durch Hervorhebung im Druck –, würde dann bedeuten, dass sie nicht so sehr darin besteht, dass Kant der These der Affektion durch Dinge an sich sowohl zustimmen als auch sich dieser Zustimmung enthalten muss, als vielmehr darin, dass er ihr zustimmen und sie zugleich für sinnlos halten muss. 17 Jacobi, F. H. (2004 [1785]). „Ueber den Transscendentalen Idealismus“, 109.

236

Rosefeldt

in der Tat hat Kant selbst in seinem eigenen Handexemplar der ersten Au age der Kritik der reinen Vernun Änderungsvorschläge in genau diesem Sinne notiert. Er versieht dort Behauptungen der Form, dass wir die Kategorie der Kausalität sinnvoll gar nicht auf Dinge an sich anwenden können, systematisch mit dem Zusatz „[…] wenn sie Erkenntnis verscha en sollen“ (vgl. z.B. A 246). Nimmt man diese Zusätze ernst, dann lässt Kant es also zu, dass wir zumindest sinnvolle Gedanken über Dinge an sich haben können, wenn wir auch nie erkennen können, ob diese Gedanken wahr sind.¹⁸ Kann man auf der Grundlage der Unterscheidung zwischen Denken und Erkennen auch die ersten beiden Gründe für die Unmöglichkeitsbehauptung entkrä en? Laut Schulze kann man Reinholds Auseinandersetzung mit Jacobis Kritik als den Versuch verstehen, eben dies zu tun. Er unterstellt Reinhold die These, dass man als Kantianer nur annehmen müsse, dass wir uns Dinge an sich als Ursachen unserer Vorstellungen denken müssen, nicht aber dass wir diese erkennen können.¹⁹ Und er zitiert eine Passage, in der Reinhold den ontologischen Status des Dinges an sich auf die folgende Weise charakterisiert, um die These, dass es solche Dinge gibt, mit dem Kantischen System kompatibel zu machen. Die Passage lautet: „Das Ding an sich ist […], wenn wir es von den Erscheinungen, denen es zum Grunde liegt, unterschieden, nichts als ein Produkt der Vernun und als ein logisches Ding“.²⁰ Die Rede vom Ding an sich als „logischem Ding“ beschreibt tatsächlich sehr gut die Stellung, die Gegenständen, die nicht solche unserer sinnlichen Anschauung sind, in Kants System zukommen kann, ohne diesem System zu widersprechen. Wie Kants Metaphysikvorlesungen deutlich machen, verwendet er die Ausdrücke „Objekt im logischen Verstande“ und „logisches Etwas“ zur Rede über Dinge, die unter denjenigen Gegenstandsbegri fallen, den er am Ende der transzendentalen Analytik als höchsten Begri der Transzendentalphilosophie bezeichnet und als „Begri von einem Gegenstande überhaupt (problematisch genommen und unausgemacht, ob er Etwas oder Nichts sei)“ (A 290/B 346)

18 Im Rahmen von Kants Gesamtphilosophie ist diese Einschränkung der Sinnlosigkeitsthese sehr nachvollziehbar, schließlich macht Kant im Rahmen seiner Moralphilosophie an zentraler Stelle Aussagen über unseren freien Willen als die nicht-sinnliche Ursache unserer Handlungen, und es wäre sehr merkwürdig, wenn er all diese Aussagen für sinnlos halten müsste. 19 Vgl. Schulze, G. E. (1996 [1792]): Aenesidemus oder über die Fundamente der von Herrn Professor Reinhold in Jena gelieferten Elementar-Philosophie, 207 (OP 299 .). 20 Schulze, G. E. (1996 [1792]): Aenesidemus oder über die Fundamente der von Herrn Professor Reinhold in Jena gelieferten Elementar-Philosophie, 212 Anm. 34 (OP 307).

Dinge an sich und der Außenweltskeptizismus

237

charakterisiert.²¹ Zu den Gegenständen, die „nicht Etwas sondern Nichts“ sind, gehört für Kant das sogenannte „Gedankending“ bzw. „ens rationis“, das er als den „Gegenstand eines Begri s, dem gar keine anzugebende Anschauung correspondirt“ (A 290/B 347), de niert. Da Kant deutlich macht, dass man auch solche Gedankendinge durch die Kategorien denkt, scheint er in seinem System Platz zu haben für Gegenstände, auf die die Kategorie der Kausalität angewandt werden kann, obwohl sie nicht Gegenstände der Anschauung sind. Schulze führt im Aenesidemus zwei Argumente dafür an, dass die Unterscheidung zwischen Denken und Erkennen und ein Verständnis von Dingen an sich als Gedankendingen nicht dazu geeignet ist, Kant gegen eine Kritik wie die von Jacobi zu verteidigen. Sein erstes Argument lautet folgendermaßen: Nun ist ja aber in dem vorliegenden Streit zwischen der Elementar-Philosophie und ihren Gegnern gar nicht davon die Rede, ob wir nicht vielleicht vermöge der Einrichtung unsers Erkenntnisvermögens gezwungen sind, die Anschauungen der Gegenstände im Raume als abhängig von einem Dinge an sich zu denken […]; sondern vielmehr davon, ob es außer unsern Vorstellungen und unabhängig von denselben, so ein Etwas realiter gäbe, als wir unter einem Dinge an sich verstehen, und ob dieses Etwas mit den Anschauungen wirklich in Kausal-Verbindung stehe. Behauptet man nun, das Ding an sich existiere nur dem Begri e davon nach in unserer Vernun […], so muß man auch eingestehen, daß es eigentlich gar keine Dinge an sich realiter gäbe, und daß also unsern Vorstellungen gar keine wirkliche Abhängigkeit von diesen Dingen an sich zukomme.²²

Man kann Schulze hier so verstehen, dass die Unterscheidung zwischen Denken und Erkennen nichts gegen den oben genannten zweiten Grund für die Unmöglichkeitsbehauptung auszurichten vermag (und also a fortiori auch nichts gegen den ersten). Selbst wenn man mit Kant die Möglichkeit inhaltsvoller Gedanken über Dinge an sich als Ursachen unserer Vorstellungen zugestehen kann, scheint Kant keinerlei Rechtfertigung für die von ihm gemachte Annahme zu haben, dass es solche Dinge gibt, wenn er zugesteht, dass wir dies nicht erkennen können. Schulzes zweites Argument macht deutlich, dass sich die Charakterisierung von Dingen an sich als Gedankendingen gerade nicht mit der Rolle verträgt, die diese in Kants Theorie sinnlicher Erkenntnis spielen: Ist nun das Ding an sich in seinem Unterschiede von den sinnlichen Vorstellungen bloß eine Idee oder ein logisches Ding, so muß man, um dartun zu können, daß der Sto der sinnlichen

21 Vgl. Vorlesungen über Metaphysik (L₂ nach Pölitz), AA XXVIII 544 und 555; vgl. dazu Rosefeldt, Tobias (2000). Das logische Ich. Kant über den Gehalt des Begri es von sich selbst. Berlin: Philo Verlag, 71 . 22 Schulze, G. E. (1996 [1792]): Aenesidemus oder über die Fundamente der von Herrn Professor Reinhold in Jena gelieferten Elementar-Philosophie, 211 (OP 305 f.); vgl. auch die Fußnote ebd. 212 (OP 306 .).

238

Rosefeldt

Vorstellungen etwas von demselben Gegebenes sei, voraussetzen, daß eine Idee und ein bloß logisches Ding etwas geben und die Rezeptivität a zieren könne, welches doch o enbar absurd ist.²³

Schulze hat völlig Recht: Zwar kann Kant im Rahmen seines Systems zugestehen, dass es außer Erscheinungen auch Gedankendinge gibt, aber er weist diesen Gedankendingen einen denkbar bescheidenen ontologischen Status zu. Gedankendinge sind so etwas wie die rein intentionalen Gegenstände von Begri en, denen keine Anschauung entspricht, und es gibt sie nur, weil und solange jemand einen Begri von ihnen hat. Sie sind gerade nicht vorstellungstranszendente Gegenstände wie Dinge an sich es ja sein sollen and angesichts von Kants A ektionstheorie auch sein müssen. A ektion soll – um Jacobis Worte zu gebrauchen – ein „reales Medium zwischen Realem und Realem, ein wirkliches Mittel von Etwas zu Etwas“ sein. Ein Gedankending ist nach Kants eigener Theorie aber kein ‚Etwas‘, sondern ein ‚Nichts‘ (vgl. KrV A 290 ./B 346 .). Bis auf weiteres muss der Versuch, die Spannung zwischen Unverzichtbarkeitsbehauptung und Unmöglichkeitsbehauptung dadurch aufzuheben, dass man die Unmöglichkeitsbehauptung widerlegt, also als gescheitert angesehen werden. Jacobi kommt zu dem Schluss, dass jemand, der dem Kantischen Projekt mit Sympathie gegenübersteht, letztlich keine andere Wahl hat, als bei der Unverzichtbarkeitsbehauptung anzusetzen. Wenn Kant ein konsistentes System vertreten möchte, müsse er auf die Annahme vorstellungstranszendenter Gegenstände und also auch auf die Abgrenzung von einem radikalen Außenweltskeptizismus verzichten. Er schreibt: „Der transscendentale Idealist muß also den Muth haben, den krä igsten Idealismus, der je gelehrt worden ist, zu behaupten, und selbst vor dem Vorwurfe des spekulativen Egoismus [gemeint ist der Solipsismus; T. R.] sich nicht zu fürchten, weil er sich unmöglich in seinem System behaupten kann, wenn er auch nur diesen letzten Vorwurf von sich abtreiben will.“²⁴ Jacobi selbst hält diese Konsequenz für so absurd, dass sie für ihn einer Widerlegung der Kantischen Philosophie nahekommt. Erst Fichte hat versucht, Kants Grundideen im Sinne des von Jacobi geforderten „krä igen Idealismus“ a rmativ weiterzuentwickeln. In seinem Versuch einer neuen Darstellung der Wissenscha slehre von 1797/98 sagt er über die Idee, im Rahmen des Kantischen Systems subjektunabhängige Gegenstände für unsere Vorstellungen und deren Gehalte verantwortlich zu machen: „Diese Absurdität irgend einem Menschen, der

23 Schulze, G. E. (1996 [1792]): Aenesidemus oder über die Fundamente der von Herrn Professor Reinhold in Jena gelieferten Elementar-Philosophie, 212 f. Anm. (OP 308). 24 Jacobi, F. H. (2004 [1785]). „Ueber den Transscendentalen Idealismus“, 112.

Dinge an sich und der Außenweltskeptizismus

239

seiner Vernun noch mächtig ist, zuzutrauen, ist mir wenigstens unmöglich“,²⁵ und er will sie deswegen auch nicht Kant selbst zutrauen. Was wir als gute Kantianer annehmen dürfen ist laut Fichte nichts als die Existenz unserer Vorstellungen und das mit manchen von ihnen verbundene Bewusstsein, dass wir über ihren Gehalt nicht willkürlich bestimmen können und insofern beschränkt sind (ebd. 242 f.). Die Vorstellungen, bei denen dies der Fall ist, sind diejenigen, die Kant Emp ndungen genannt hätte, die Fichte aber lieber als Gefühle bezeichnet: als „das Gefühl des Süßen, Roten, Kalten u. d. gl.“ (ebd. 243) oder allgemein als das Gefühl „so und so bestimmt“ zu sein (ebd.). „Dieses ursprüngliche Gefühl aus der Wirksamkeit eines Etwas weiter erklären zu wollen“, so schreibt Fichte weiter, „ist der Dogmatismus der Kantianer, den ich soeben gezeigt habe, und den sie gern Kant au ürden möchten. Dieses ihr Etwas ist nothwendig das leidige Ding an sich“ (ebd.). Als nicht-dogmatische Kantianer dürfen wir laut Fichte zwar annehmen, dass sich unsere Vorstellungen auf Gegenstände beziehen, diese Gegenstände müssen selbst aber als Produkt des menschlichen Verstandes verstanden werden, so dass unter dem Gegenstandsbezug von Vorstellungen nicht mehr verstanden werden darf als die Tatsache, dass unsere Vorstellungen auf eine gewisse nicht unserem Belieben anheimgestellte Weise miteinander verbunden sind. Wenn dies der einzig sinnvolle Begri eines unabhängigen Gegenstands ist, ist es unzulässig, die Tatsache, dass Vorstellungen mit einem bestimmten empirischen Gehalt in uns sind, dadurch zu erklären, dass ihr Gegenstand auf bestimmte Weise bescha en ist. Schließlich ist dieser Gegenstand ja selbst durch die Vorstellungen diesen Gehalts konstituiert. Die genannte Erklärung ist laut Fichte im Rahmen des Kantischen Systems aber nicht nur unmöglich, sondern auch unnötig. Wir sollten zugestehen, so Fichte, dass „bei dem unmittelbaren Gefühle […] alle transscendentale Erklärung ein Ende“ hat (ebd.), d.h. dass wir – anders als die Unverzichtbarkeitsbehauptung sagt – die Tatsache, dass wir Vorstellungen mit einem bestimmten empirischen Gehalt haben, im Rahmen der Transzendentalphilosophie gar nicht weiter erklären müssen, sondern als unhintergehbares Faktum anerkennen sollten. So kursorisch diese Darstellung der ersten 15 Jahre Rezeptionsgeschichte von Kants Lehre vom Ding an sich gewesen ist, so hat sie doch ho entlich eines deutlich gemacht: Wenn man mit dem transzendentalen Idealismus diejenige Strategie zur Widerlegung des Außenweltskeptizismus verbindet, die Kant im vierten Paralogismus der A-Au age verfolgt, und wenn man die Unterscheidung zwi-

25 Fichte, Johann Gottlieb (1970 [1797/98]). Versuch einer neuen Darstellung der Wissenscha slehre. In: J. G. Fichte, Gesamtausgabe der Bayerischen Akademie der Wissenscha en. Hg. von R. Lauth und H. Gliwitzky. Stuttgart-Bad Cannstatt: Friedrich Frommann Verlag, Band I, 4., 239.

240

Rosefeldt

schen erkennbaren Erscheinungen und unerkennbaren Dingen an sich nach dem Modell versteht, das dieser Strategie zu Grunde liegt, dann führt der Versuch, Kants System konsequent zu Ende zu denken, zwangsläu g zu einem Idealismus à la Fichte. Das kann leicht den Eindruck erwecken, als sei die Entwicklung von Kant zu Fichte insgesamt systematisch alternativlos, d.h. artikuliere nur Spannungen, die in Kants Philosophie objektiv vorhanden sind und überwinde diese schließlich. Im Folgenden möchte ich zeigen, dass dieser Eindruck täuscht. Bereits 1783 in den Prolegomena, d.h. zwei Jahre nach Erscheinen der ersten Au age der Kritik der reinen Vernun , besonders aber 1787 in deren zweiter Au age hat Kant seine These über die Unterscheidung zwischen Dingen an sich und Erscheinungen und seine Ansicht über das Verhältnis dieser Unterscheidung zum Problem des Außenweltskeptizismus so radikal gegenüber seiner Darstellung im Paralogismenkapitel aus der ersten Au age geändert, dass seine Rede vom Ding an sich gegen die zuletzt dargestellte Kritik verteidigt werden kann. Leider ist das keinem seiner Kritiker aufgefallen.

3 Die Dinge, die uns erscheinen Als die Garve/Feder-Rezension erschien, schrieb Kant gerade an seinen Prolegomena. Er war über die Rezension so erbost, dass er diesem Werk einen Anhang hinzufügte, in dem er dem Rezensenten vorwir , praktisch keines der zentralen Argumente, die er für seinen transzendentalen Idealismus liefert, erwähnt zu haben (vgl. AA V 372 .). Mit dieser Kritik hat Kant völlig Recht. Dennoch hat ihn der oben dargestellte Vorwurf, sein transzendentaler Idealismus liefe auf einen Idealismus à la Berkeley hinaus, so sehr verunsichert, dass er sich genötigt sah, sowohl in den Prolegomena als auch später in der zweiten Au age der Kritik der reinen Vernun darauf durch einige Klarstellungen zu reagieren. Ich zitiere noch einmal aus der bereits erwähnten Passage aus den Prolegomena, in der er den Idealismusvorwurf zurückweist: […] es sind uns Dinge als außer uns be ndliche Gegenstände unserer Sinne gegeben, allein von dem, was sie an sich selbst sein mögen, wissen wir nichts […]. Demnach gestehe ich allerdings, daß es außer uns Körper gebe, d.i. Dinge, die, obzwar nach dem, was sie an sich selbst sein mögen, uns gänzlich unbekannt, wir durch die Vorstellungen kennen, welche ihr Ein uß auf unsre Sinnlichkeit uns verscha , und denen wir die Benennung eines Körpers geben […] (AA IV 289).

Sehen wir uns genauer an, was Kant hier über Körper sagt. Im Paralogismenkapitel der ersten Au age war seine Position klar: Körper sind letztlich selbst nichts

Dinge an sich und der Außenweltskeptizismus

241

als Vorstellungen, d.h. mentale Zustände des Subjekts. Hier lautet die Charakterisierung nun anders: Körper sind die „außer uns be ndlichen Gegenstände unserer Sinne“, also diejenigen Dinge, die wir mit Hilfe unserer Sinne wahrnehmen. Sie sind nicht selbst Vorstellungen, sondern Dinge, die wir „durch die Vorstellungen kennen, welche ihr Ein uß auf unsre Sinnlichkeit uns verscha “. Anders als zuvor fungieren Vorstellungen hier also nicht mehr als die Gegenstände unserer Wahrnehmung, sondern als Wahrnehmungsvehikel. Wir haben Vorstellungen, wenn wir Körper wahrnehmen, aber wir nehmen nicht Vorstellungen wahr, wenn wir Körper wahrnehmen. Das bedeutet, dass Kant eine Annahme fallen lässt, die sowohl dem Erkenntnismodell des transzendentalen Realismus als auch dem des transzendentalen Idealismus unhinterfragt zu Grunde lag (vgl. Abb. 1 und 2), die Annahme nämlich, dass die unmittelbaren Gegenstände unserer Wahrnehmung immer unsere eigenen Vorstellungen sind. Er ersetzt sie nun durch die Annahme, dass wir diejenigen vorstellungstranszendenten Gegenstände wahrnehmen, die die Vorstellungen in uns hervorrufen.²⁶ Diese Änderung wir natürlich umgehend die folgende Frage auf: Wenn wir vorstellungstranszendente Gegenstände wahrnehmen können, was bleibt dann von Kants These, dass wir immer nur Erscheinungen, nicht aber Dinge an sich erkennen können? In der zitierten Passage schreibt Kant über die wahrgenommenen Körper, dass sie „Dinge [seien, die] nach dem, was sie an sich selbst sein mögen, uns gänzlich unbekannt“ sind. Dinge an sich tauchen in dieser Formulierung also

26 In dem zuvor genannten Anhang zu den Prolegomena reagiert Kant auf den Berkeley-Vergleich mit dem Hinweis, dass Berkeley anders als er selbst keine Theorie synthetischer Erkenntnis a priori entwickle, deswegen über kein Wahrheitskriterium verfüge und also nicht zwischen Schein und Wirklichkeit unterscheiden könne (vgl. AA V 375). Diese Replik unterscheidet sich stark von derjenigen im Haupttext der Prolegomena, die ich hier dargestellt habe, und ist mit einer Konzeption der Unterscheidung zwischen Dingen an sich und Erscheinungen wie der aus dem vierten Paralogismus vereinbar. Die Tatsache, dass Kant in den Prolegomena zwischen zwei so verschiedene Strategien zur Erwiderung auf den Idealismus-Vorwurf schwankt, ist meines Erachtens ein Beleg dafür, wie unvorbereitet ihn dieser Vorwurf gegen seine Theorie getro en hat. Wie ich im Folgenden deutlich machen möchte, hat er sich spätestens bis zur zweiten Au age der Kritik der reinen Vernun dann für die Strategie aus dem Haupttext entschieden. Dass diese Strategie besser zu Kants eigentlichen Intentionen passt, ist in der Literatur natürlich umstritten. Dina Emundts zum Beispiel argumentiert dafür, dass eine Abgrenzung von Kant und Berkeley im Sinne des Anhangs zu den Prolegomena erfolgversprechender ist als eine, die Kant starke Annahmen über subjektunabhängige Gegenstände zuschreibt, weil sie der Meinung ist, dass Kant diese Annahmen im Rahmen seines Systems nicht machen darf (vgl. Emundts, Dina (2008). „Kant and Berkeley on Objectivity“. In: B. Longuenesse, D. Garber (ed.). Kant and the Early Moderns. Princeton: Princeton University Press, 117–142.). Ich werde im folgenden Abschnitt 4 zu zeigen versuchen, dass man diese letzte Annahme aufgeben kann, wenn man die hier vorgeschlagene Interpretation von Kants Rede von Dingen an sich zu Grunde legt.

242

Rosefeldt

nicht als eine weitere von den sinnlich wahrnehmbaren Körpern unterschiedene Entität auf. Kant sagt vielmehr von ein und denselben Entitäten, den Körpern, dass wir sie einerseits durch die Vorstellungen kennen, die sie in uns hervorrufen, dadurch aber andererseits nicht erkennen, wie sie an sich selbst bescha en sind. Noch unmissverständlicher macht er diesen Punkt in einer Passage aus der zweiten Au age der Kritik der reinen Vernun deutlich, in der er sich wiederum gegen den Vorwurf wehrt, dass er einen Idealismus vertrete, der äußere Gegenstände in Vorstellungen, d.h. in bloßen Schein verwandelt. Er schreibt dort: Wenn ich sage: im Raum und der Zeit stellt die Anschauung [die] äußeren Objecte […] vor, so wie es unsere Sinne a cirt, d.i. wie es erscheint, so will das nicht sagen, daß diese Gegenstände ein bloßer Schein wären. Denn in der Erscheinung werden jederzeit die Objekte, ja selbst die Bescha enheiten, die wir ihnen beilegen, als etwas wirklich Gegebenes angesehen, nur daß, so fern diese Bescha enheit nur von der Anschauungsart des Subjects in der Relation des gegebenen Gegenstandes zu ihm abhängt, dieser Gegenstand als Erscheinung von ihm selber als Object an sich unterschieden wird (KrV B 69).

Kant ist hier so explizit wie man es sich nur wünschen kann: Dinge an sich und Erscheinungen sind nicht zwei verschiedene Arten von Gegenständen (wahrnehmbare, mit Vorstellungen identische Körper einerseits und ihre vorstellungstranszendenten Ursachen andererseits); vielmehr wird ein und derselbe Gegenstand als Erscheinung von ihm selber als Ding an sich unterschieden. In anderen Worten: Von ein und demselben Gegenstand erkennen wir einerseits, wie er uns erscheint, andererseits bleibt uns verborgen, wie er an sich selbst bescha en ist.²⁷ Die Passage gibt zudem einen Hinweis darauf, was Kant eigentlich damit meint, wenn er sagt, dass wir nicht erkennen können, wie die Dinge an sich selbst bescha en sind. Er schreibt, dass alle Bescha enheiten, die wir an den von uns

27 In diesem Sinn äußert sich Kant auch in B XXVII, wo er von der „Unterscheidung der Dinge als Gegenstände der Erfahrung von eben denselben als Dingen an sich selbst“ spricht, und in B 306, wo er „Gegenstände als Erscheinungen Sinnenwesen (Phaenomena)“ und „eben dieselbe nach [ihrer Bescha enheit an sich selbst], wenn wir sie gleich in derselben nicht anschauen, […] Verstandeswesen (Noumena)“ nennt (Hervorhebungen von mir; vgl. auch Opus postumum, AA XXII 26 und 43). Die Idee, dass die Unterscheidung zwischen Erscheinungen und Dingen an sich keine zwischen zwei verschiedenen Arten von Gegenständen, sondern eine zwischen zwei verschiedenen Aspekten dieser Gegenstände ist, geht auf Prauss und Allison zurück (vgl. Prauss, Gerold (1974). Kant und das Problem der Dinge an sich. Bonn: Bouvier Verlag; Allison, Henry (1983). Kant’s Transcendental Idealism. An Interpretation and Defense. New Haven, London: Yale University Press). Meine Interpretation der Unterscheidung unterscheidet sich von derjenigen von Prauss und Allison dadurch, dass ich die Rede von Aspekten nicht methodologisch, sondern ontologisch interpretiere. Es geht Kant meiner Meinung nach nicht darum, zwei verschiedene Weisen zu unterscheiden, auf die man ein und dieselben Dinge thematisieren kann, sondern um zwei verschiedene Arten von Eigenscha en, die ein und dieselben Gegenstände haben.

Dinge an sich und der Außenweltskeptizismus

243

verschiedenen Gegenständen wahrnehmen können (d.h. alle seine raum-zeitlichen Eigenscha en), „nur von der Anschauungsart des Subjects in der Relation des gegebenen Gegenstandes zu ihm abhäng[en]“. Bei den Eigenscha en, die wir an einem Ding erkennen können, handelt es sich also um Eigenscha en, die in einer gewissen Weise von uns Subjekten abhängen, genauer von unserer Anschauungsart, d.h. von der Weise, wie wir kognitiv bescha en sind und auf das Ding reagieren. In einer Fußnote zu der eben zitierten Passage erläutert Kant diese Form von Subjektabhängigkeit durch eine Analogie zu traditionellen sekundären Qualitäten: So wie Farben außergeistigen Gegenstände nicht an sich selbst, d.h. unabhängig von uns zukommen, sondern nur deswegen, weil wir mit bestimmten Emp ndungen auf diese Gegenstände reagieren, so sollen alle raum-zeitlichen Eigenscha en den außergeistigen Gegenständen nicht an sich selbst zukommen, sondern nur deswegen, weil unsere Anschauungen gemäß unserer Anschauungsformen Raum und Zeit strukturiert sind (vgl. B 69 f., Anm.).²⁸ Alle erkennbaren raum-zeitlichen Eigenscha en außergeistiger Gegenstände sind demnach Eigenscha en, die unsere kognitive Reaktion auf diese Gegenstände betre en. In B 59 schreibt Kant: Wir haben also sagen wollen: […] daß die Dinge, die wir anschauen, nicht das an sich selbst sind, wofür wir sie anschauen, noch ihre Verhältnisse so an sich selbst bescha en sind, als sie uns erscheinen […]. Was es für eine Bewandtniß mit den Gegenständen an sich und abgesondert von aller dieser Receptivität unserer Sinnlichkeit haben möge, bleibt uns gänzlich unbekannt. Wir kennen nichts als unsere Art, sie wahrzunehmen, die uns eigenthümlich ist, die auch nicht nothwendig jedem Wesen, ob zwar jedem Menschen, zukommen muß.

Statt davon zu sprechen, dass wir nicht wissen, wie die Dinge an sich selbst bescha en sind, sondern lediglich „unsere Art, sie wahrzunehmen“ kennen, spricht Kant an verschiedenen anderen Stellen seines Werks auch davon, dass wir lediglich die Weise erkennen, auf die uns die Dinge, deren An-sich-Sein uns unbekannt ist, a zieren. An diesen Stellen wird zudem deutlich, inwiefern die Rede von Dingen als Erscheinungen auf die Annahme verp ichtet ist, dass diese Dinge auch irgendwie an sich selbst bescha en sind. Ich zitiere einige Stellen:

28 Lucy Allais stellt diese Analogie zwischen Kantischen Erscheinungseigenscha en und sekundären Qualitäten in das Zentrum ihrer Interpretation von Kants transzendentalen Idealismus (vgl. Allais, Lucy (2007). „Kant’s Idealism and the Secondary Quality Analogy“. In: Journal of the History of Philosophy 45, 459–484), ohne allerdings wie ich auf der Grundlage der Analogie Erscheinungseigenscha en als dispositionale Eigenscha en zu interpretieren (vgl. dazu neben der folgenden Darstellung auch Rosefeldt, Tobias (2007). „Dinge an sich und sekundäre Qualitäten“. In: J. Stolzenberg (Hg.). Kant in der Gegenwart. Berlin, New York: De Gruyter, 167–209).

244

Rosefeldt

[W]enn wir die Gegenstände der Sinne wie billig als bloße Erscheinungen ansehen, so gestehen wir hiedurch doch zugleich, daß ihnen ein Ding an sich selbst zum Grunde liege, ob wir dasselbe gleich nicht, wie es an sich bescha en sei, sondern nur seine Erscheinung, d.i. die Art, wie unsre Sinnen von diesem unbekannten Etwas a cirt werden, kennen. Der Verstand also, eben dadurch daß er Erscheinungen annimmt, gesteht auch das Dasein von Dingen an sich selbst zu, und so fern können wir sagen, daß die Vorstellung solcher Wesen, die den Erscheinungen zum Grunde liegen, mithin bloßer Verstandeswesen nicht allein zulässig, sondern auch unvermeidlich sei (Prolegomena, AA IV 314 f.). [A]lle Vorstellungen, die uns ohne unsere Willkür kommen (wie die der Sinne), [geben] uns die Gegenstände nicht anders zu erkennen […], als sie uns a ciren, wobei, was sie an sich sein mögen, uns unbekannt bleibt […], [so daß wir] bloß zur Erkenntniß der Erscheinungen, niemals der Dinge an sich selbst gelangen können. Sobald dieser Unterschied […] einmal gemacht ist, so folgt von selbst, daß man hinter den Erscheinungen doch noch etwas anderes, was nicht Erscheinung ist, nämlich die Dinge an sich, einräumen und annehmen müsse, ob wir gleich uns von selbst bescheiden, daß, da sie uns niemals bekannt werden können, sondern immer nur, wie sie uns a ciren, wir ihnen nicht näher treten und, was sie an sich sind, niemals wissen können (Grundlegung zur Metaphysik der Sitten, AA IV 451). Die Sinnlichkeit und ihr Feld, nämlich das der Erscheinungen, wird selbst durch den Verstand dahin eingeschränkt: daß sie nicht auf Dinge an sich selbst, sondern nur auf die Art gehe, wie uns vermöge unserer subjectiven Bescha enheit Dinge erscheinen. Dies war das Resultat der ganzen transscendentalen Ästhetik, und es folgt auch natürlicher Weise aus dem Begri e einer Erscheinung überhaupt: daß ihr etwas entsprechen müsse, was an sich nicht Erscheinung ist, weil Erscheinung nichts für sich selbst und außer unserer Vorstellungsart sein kann, mithin, wo nicht ein beständiger Cirkel herauskommen soll, das Wort Erscheinung schon eine Beziehung auf Etwas anzeigt, dessen unmittelbare Vorstellung zwar sinnlich ist, was aber an sich selbst, auch ohne diese Bescha enheit unserer Sinnlichkeit (worauf sich die Form unserer Anschauung gründet), Etwas, d.i. ein von der Sinnlichkeit unabhängiger Gegenstand, sein muß (KrV A 251 f.).²⁹

Dass wir die Dinge nicht so kennen, wie sie an sich selbst sind, sondern nur so, wie sie uns a zieren, heißt für Kant also, dass wir von Gegenständen nur wissen können, dass sie so bescha en sind, dass sie bestimmte sinnliche Vorstellungen in uns hervorrufen. Da die Bescha enheit dieser Vorstellungen wiederum von unserer kognitiven Verfasstheit abhängt, können wir ausschließlich subjektabhängige Eigenscha en von Gegenständen erkennen.³⁰ Wie die Zitate zeigen, ist Kant ferner der Meinung, dass wir von einem Gegenstand allein dadurch, dass wir ihn als Erscheinung in seinem Sinne charakterisieren, ipso facto annehmen, dass er irgendwie an sich selbst bescha en ist, d.h. irgendwelche subjektunabhängigen

29 Vgl. auch B 59–61, A 44/ B 61 und A 358. 30 Dass die Bescha enheit der Vorstellungen zum Teil von Anschauungsformen abhängt, die unserem Geist wesentlich sind, erklärt dann, dass wir Wissen a priori davon haben können, wie uns Gegenstände erscheinen.

Dinge an sich und der Außenweltskeptizismus

245

Eigenscha en haben muss, auch wenn wir nicht wissen, welche subjektunabhängigen Eigenscha en dies sind. Es ist der Begri der Erscheinung, der festlegt, dass es nur dann Erscheinungen geben kann, wenn es auch Dinge an sich gibt.³¹ Ich habe an anderer Stelle dafür argumentiert, dass man sowohl die Subjektabhängigkeit sinnlich erkennbarer raum-zeitlicher Eigenscha en als auch die Tatsache, dass die Existenz solcher Eigenscha en diejenige subjektunabhängiger Eigenscha en impliziert, am plausibelsten dadurch rekonstruieren kann, dass man raum-zeitliche Eigenscha en als Dispositionen dazu versteht, in Wesen mit unseren Anschauungsformen eine bestimmte Art von Vorstellung hervorzurufen, und eine solche Disposition wiederum als die höherstu ge Eigenscha en, irgendeine subjektunabhängige Eigenscha zu haben, durch deren Vorliegen unter geeigneten Umständen die jeweilige Vorstellung hervorgerufen wird.³² An dieser Stelle möchte ich mich damit begnügen, Kants Behauptungen anhand einer Analogie verständlich zu machen. Ein philosophisch unkontroverses Beispiel für eine Eigenscha von Gegenständen, die von unserer Bescha enheit und unserer Reaktion auf diese Gegenstände abhängt, ist die Eigenscha gi ig zu sein. Gi ig zu sein ist einerseits eine Eigenscha , die von uns verschiedenen Dingen zukommt

31 Vgl. KrV B 306 „Gleichwohl liegt es doch schon in unserm Begri e, wenn wir gewisse Gegenstände als Erscheinungen Sinnenwesen (Phaenomena) nennen, indem wir die Art, wie wir sie anschauen, von ihrer Bescha enheit an sich selbst unterscheiden: daß wir […] eben dieselbe nach dieser letzteren Bescha enheit, wenn wir sie gleich in derselben nicht anschauen, […] als Gegenstände, bloß durch den Verstand gedacht, jenen gleichsam gegenüber stellen.“ und KrV B XXVII f.: „Gleichwohl wird, welches wohl gemerkt werden muß, doch dabei immer vorbehalten, daß wir eben dieselben Gegenstände auch als Dinge an sich selbst, wenn gleich nicht erkennen, doch wenigstens müssen denken können. Denn sonst würde der ungereimte Satz daraus folgen, daß Erscheinung ohne etwas wäre, was da erscheint.“ 32 Für eine ausführliche Darstellung der Interpretation vgl. Rosefeldt, Tobias (2007). „Dinge an sich und sekundäre Qualitäten“, 167–209. Ich ho e, mit meiner Lesart Horstmanns Anforderung an eine aus heutiger Sicht akzeptable Interpretation von Kants transzendentalem Idealismus zu liefern, die dieser folgendermaßen formuliert: „Die gegenwärtigen Probleme bestehen […] darin, eine Deutung der Kantischen Lehre von den Dingen an sich und von Raum und Zeit zu nden, die eine bescheidene Lesart dessen zulassen, was der transzendentale Idealismus will, eine Lesart, die es erlaubt, Kants Position so zu präsentieren, daß sie sich wenigstens als kompatibel mit der einen oder anderen Version des Empirismus oder des wissenscha lichen Realismus erweisen läßt“ (Horstmann, Rolf-Peter (1997). „Was bedeutet Kants Lehre vom Ding an sich für seine transzendentale Ästhetik?“. In: Ders. Bausteine kritischer Philosophie. Arbeiten zu Kant. Bodenheim: Philo Verlag, 36). Ich ho e, mit meiner Interpretation auch Horstmanns These widerlegen zu können, dass Kants Theorie es ausschließt, im Rahmen einer solchen Realismus-kompatiblen Interpretation „Kant mit irgendwelchen Dingen an sich liebäugeln zu lassen, bei denen sich die Frage nach dem Vorliegen oder Nichtvorliegen von irgendwelchen Eigenscha en sinnvoll stellen läßt“ (ebd. 45).

246

Rosefeldt

– z.B. Fliegenpilzen –, andererseits eine Eigenscha , die diese von uns unabhängigen Gegenstände nur in Relation zu uns haben, Fliegenpilze z.B. nur deswegen, weil wir auf eine bestimmte Weise, nämlich mit Vergi ungssymptomen, auf sie reagieren. Wären wir biologisch anders gebaut, wären Fliegenpilze auch nicht gi ig. Trotzdem hängt das Gi igsein der Fliegenpilze natürlich nicht ausschließlich von uns ab, sondern liegt ebenso an einer von uns unabhängigen Eigenscha en des Pilzes – einer Eigenscha , die ihm an sich selbst zukommt. In diesem Fall ist das seine Eigenscha , den chemischen Sto Muscimol zu enthalten, der kausal für die Vergi ungssymptome verantwortlich ist. In analoger Weise ist die Tatsache, dass raum-zeitliche Eigenscha en den wahrgenommenen Dingen nicht an sich selbst zukommen, sondern subjektabhängig sind, damit verträglich, dass diese Dinge auch an sich selbst irgendwie bescha en sind, d.h. subjektunabhängige Eigenscha en haben, ja sogar haben müssen, um subjektabhängige Eigenscha en zu haben. So wie der Pilz nur deswegen gi ig sein kann, weil er eine bestimmte chemische Bescha enheit hat, die für unsere Reaktion auf ihn verantwortlich ist, so müssen die Gegenstände unserer Sinne auch irgendwelche subjektunabhängigen Eigenscha en haben, um uns erscheinen, d.h. Vorstellungen in uns erregen zu können. Im Falle des Pilzes ist dies deswegen eine begri iche Wahrheit, weil Gi igsein nichts anderes ist als die Eigenscha , irgendeine reaktionsunabhängige Eigenscha zu haben, die unter geeigneten Umständen bei uns Vergi ungssymptome hervorru . Im Falle Kantischer Erscheinungen erhält man eine analoge begri iche Wahrheit, wenn man annimmt, dass eine raum-zeitliche Eigenscha nichts anderes ist als die Eigenscha , irgendeine reaktionsunabhängige Eigenscha zu haben, auf Grund deren uns der Gegenstand a ziert, d.h. bestimmte Vorstellungen in uns hervorru . Nach diesem Modell ist die Weise, wie die Dinge an sich selbst bescha en sind, also mitverantwortlich dafür, wie sie uns erscheinen: Dass uns überhaupt etwas als raum-zeitlich strukturiert erscheinen kann, liegt an uns und unseren Anschauungsformen. Aber dass uns dieser Gegenstand rund und jener eckig erscheint, liegt darüber hinaus auch an diesen Gegenständen. So wie es an uns liegt, dass wir überhaupt mit Vergi ungssymptomen auf etwas reagieren können, aber nicht allein an uns liegt, dass der Fliegenpilz gi ig, der Steinpilz aber essbar ist. Dies liegt auch an der jeweiligen chemischen Bescha enheit dieser beiden Pilze. Anders als in der Pilzanalogie können wir laut Kant allerdings überhaupt keine der subjektunabhängigen Eigenscha en der von uns wahrgenommen Gegenstände erkennen und also auch von keiner solchen Eigenscha wissen, dass sie der

Dinge an sich und der Außenweltskeptizismus

247

Grund dafür ist, dass uns das Ding auf die-und-die Weise erscheint. Und genau dies besagt seine These von der Unerkennbarkeit der Dinge an sich.³³ Ohne Zweifel wir Kants neues Modell unserer Erkenntnis von äußeren Gegenständen und seine Explikation des Verhältnisses von Erscheinungen und Dingen an sich viele sowohl exegetische als auch systematische Fragen auf, deren Beantwortung den Rahmen dieses Beitrags allerdings sprengen würde. Mir reicht es hier, darauf hinzuweisen, wie verschieden dieses Modell von demjenigen ist, das Kants Argument gegen den Außenweltskeptizismus im vierten Paralogismus zu Grunde lag. Dies kann man anhand einer weiteren Graphik deutlich machen:

verursachen Vorstellungen in

subjektunabhängige Eigenscha en von außergeistigen Dingen

Subjekt nimmt unmittelbar wahr (hat Vorstellungen von)

subjektabhängige Eigenscha en

Abb. 3: Transzendentaler Idealismus 2

Im Rahmen dieses Beitrags stellen sich nun vor allem zwei Fragen zu diesem Modell. Erstens gilt es zu klären, ob Kants neue Explikation des Verhältnisses von Erscheinungen und Dingen an sich ebenfalls von den Einwänden betro en ist, die Garve, Feder, Jacobi und Schulze gegen Kant formuliert haben. Zweitens wüsste man gerne, welche Konsequenzen die Änderung für Kants Verhältnis zur Herausforderung des Außenweltskeptizismus hat. Ich werde die erste dieser beiden Fragen im nächsten und die zweite im übernächsten und letzten Teil meines Beitrags zu beantworten versuchen.

33 Genau in diesem Sinne ist die Bemerkung in dem obigen Zitat aus der Grundlegung zu verstehen, dass man „hinter den Erscheinungen doch noch etwas anderes, was nicht Erscheinung ist, nämlich die Dinge an sich, einräumen und annehmen müsse, ob wir gleich uns von selbst bescheiden, daß, da sie uns niemals bekannt werden können, sondern immer nur, wie sie uns a ciren, wir ihnen nicht näher treten und, was sie an sich sind, niemals wissen können“ (AA IV 451).

248

Rosefeldt

4 Kant antwortet seinen Kritikern Dem Idealismus-Einwand von Garve und Feder kann Kant klarerweise entgehen. Er gesteht ja, anders als Berkeley, zu, dass wir von uns und unseren Vorstellungen verschiedene Gegenstände wahrnehmen, und sagt lediglich, dass diese Gegenstände nur für uns Körper sind, d.h. im Raum existieren. Dem neuen Modell zufolge müsste Feder also auch nicht die absurde Annahme machen, dass der Königsberger Philosoph, von dem ihm schon so viele Menschen erzählt haben, nichts als eine Vorstellung in seinem Geist ist, sondern er kann weiter annehmen, dass dieser eine von seinen Vorstellungen verschiedene Person ist. Was Feder schlucken muss ist lediglich die Annahme, dass dem Philosophen all die raum-zeitlichen Eigenscha en, die man an ihm wahrnehmen kann – seine Größe und Gestalt zum Beispiel – nicht unabhängig vom menschlichen Geist zukommen, sondern dass er diese nur für Wesen hat, denen die Welt in Raum und Zeit erscheint. Wie steht es um die Einwände von Jacobi und Schulze? Es ist klar, dass es im Rahmen des neuen Modells bei einer Verteidigung gegen den Inkonsistenzvorwurf darum gehen muss, die Unmöglichkeitsbehauptung zu widerlegen, denn das Modell nimmt ja an, dass es außergeistige Gegenstände sind, die in uns Vorstellungen hervorrufen. Dabei wurde bereits gezeigt, wie man auf den genannten dritten Grund für die Unmöglichkeitsbehauptung reagieren kann. Wie oben dargestellt, muss Kant nicht annehmen, dass Gedanken über nicht raum-zeitliche Ursachen sinnlos sind. Er kann also zugestehen, dass wir sinnvolle Gedanken über Dinge an sich als Ursachen unserer Vorstellungen haben können. Um zu sehen, ob der zweite Grund für die Unmöglichkeitsbehauptung stichhaltig ist, sollte man sich zwei Fragen stellen: erstens die Frage, ob Kant nach dem neuen Modell zu der Annahme berechtigt ist, dass Dinge an sich – d.h. Träger subjektunabhängiger Eigenscha en – wirklich existieren, d.h. in Jacobis Sinne etwas „Reales“ und nicht bloße Gedankendinge sind; zweitens die Frage, ob er berechtigt ist, diesen Dingen eine kausale Rolle bei der Genese unserer Vorstellungen zuzusprechen. Ein Problem mit der ersten Frage könnte sein, dass Kant nach seinen eigenen Maßgaben nur die Existenz von solchen Gegenständen annehmen darf, die wir sinnlich wahrnehmen, und dass er deswegen die Existenz von Dingen an sich nicht behaupten darf. Der Einwand wäre zweifellos berechtigt, wenn das Modell des Paralogismenkapitels korrekt wäre und es sich bei Dingen an sich um von den wahrgenommenen Erscheinungen unterschiedene Entitäten handelte. Denn dann wüssten wir von der Existenz von Dingen an sich nur durch einen Schluss von einer wahrnehmbaren Wirkung auf eine nicht-wahrnehmbare Ursache und müssten, wie Jacobi moniert, den Satz des zureichenden Grundes über den Bereich der uns erscheinenden Gegenstände hinaus anwenden. Dem neuen Modell zufolge ist dies aber nicht der Fall. Die Dinge, von denen Kant annimmt, dass sie

Dinge an sich und der Außenweltskeptizismus

249

irgendwie an sich selbst bescha en sind, nehmen wir ja wahr. Es sind schließlich genau diejenigen Dinge, von denen wir erkennen können, wie sie uns erscheinen. Man kann auch sagen: Wir nehmen die Dinge wahr, die irgendwie an sich selbst bescha en sind, obwohl wir nicht wahrnehmen, wie sie an sich selbst beschaffen sind, d.h. obwohl wir ihre subjektunabhängigen Eigenscha en nicht kennen. Wenn wir sagen, dass diese Dinge existieren, wenden wir den Begri der Existenz also gar nicht auf Gegenstände an, die wir nicht wahrnehmen. Dies würden wir nur tun, wenn wir die Existenz von etwas behaupten würden, das uns gar nicht sinnlich erscheint, wie z.B. diejenige Gottes. Aber stellt sich die erste Frage nun nicht in neuem Gewande? Müssen wir nun nicht fragen, wie wir wissen können, dass die Gegenstände, die wir wahrnehmen, auch irgendwelche subjektunabhängigen Eigenscha en haben, d.h. überhaupt auch an sich selbst existieren? Dies ist letztlich die Frage des klassischen Außenweltskeptizismus, der wissen will, was uns zu der Annahme berechtigt, dass ein vermeintlich wahrgenommener Gegenstand nicht denselben ontologischen Status hat wie die Dinge, die uns in unseren Träumen erscheinen und die keinerlei Eigenscha en haben außer denjenigen, die wir an ihnen wahrzunehmen meinen. Ich werde im nächsten Abschnitt dazu Stellung nehmen, welche Auswirkungen Kants neues Verständnis des transzendentalen Idealismus für seine Haltung zum Außenweltskeptizismus hat. Im Moment möchte ich nur festhalten, dass Kant nach dem neuen Modell in dem Moment, in dem er berechtigterweise behaupten darf, dass Erscheinungen existieren, auch zu der Annahme berechtigt ist, dass Dinge an sich existieren. Dies hat den folgenden Grund: Erscheinungen sind nach dem neuen Modell Gegenstände, die subjektabhängige Eigenscha en haben, also Eigenscha en, die damit zu tun haben, wie uns die Gegenstände erscheinen, d.h. auf welche Weise sie kausal auf unseren Geist einwirken. Nach der oben skizzierten Analyse ist eine solche Erscheinungseigenscha E₁ nichts anderes als die Eigenscha , irgendeine subjektunabhängige Eigenscha E₂ zu haben, die bestimmte Vorstellungen in uns hervorru . E₁ selbst ist dabei eine subjektabhängige Eigenscha , denn der Gegenstand hat sie nur deswegen, weil wir auf bestimmte Weise auf ihn reagieren. E₁ ist ihrer De nition nach aber zugleich eine Eigenscha , die ein Gegenstand nur dann haben kann, wenn er auch irgendeine subjektunabhängige Eigenscha E₂ hat – welche auch immer das sein mag.³⁴ Das heißt: Der Schluss von der Existenz von Erscheinungen (als Trägern subjektabhängiger Eigenscha en) auf die Existenz von Dingen an sich (als Trägern subjektunabhängiger Eigenscha en) ist durch einen begri ichen Zusammenhang gerechtfertigt,

34 Vgl. für eine genauere Darstellung dieses Punktes Rosefeldt, Tobias (2007). „Dinge an sich und sekundäre Qualitäten“, 167–209, Abschnitt 3.

250

Rosefeldt

also einen, den man durch ein analytisches Urteil zum Ausdruck bringt. Und um ein solches Urteil berechtigterweise vorauszusetzen, müssen wir nicht die Gültigkeit eines synthetischen Grundsatzes wie des Satzes des zureichenden Grundes für einen Bereich von Gegenständen annehmen, für den Kant ihn explizit nicht behaupten will. Auf den genannten begri ichen Zusammenhang kann man sich auch bei der Beantwortung der zweiten Frage berufen, der Frage danach, ob Kant berechtigt ist, Dingen an sich eine kausale Rolle bei der Genese unserer Vorstellungen zuzusprechen. Wieder ergibt sich die Annahme einer solchen Rolle schlicht aus dem Begri der Erscheinung. Eine Erscheinungseigenscha ist schließlich gerade deswegen subjektabhängig, weil sie etwas damit zu tun hat, „wie uns die Gegenstände a zieren“, d.h. welche Wirkungen ein Gegenstand auf Grund seiner subjektunabhängigen Eigenscha en einerseits und der Bescha enheit unseres Geistes andererseits auf uns hat. Der erste Grund für die Unmöglichkeitsbehauptung hatte mit dem Vorwurf zu tun, dass Kant selbst dann, wenn er gerechtfertigt wäre, eine vorstellungstranszendente Ursache unserer Vorstellungen anzunehmen, nicht behaupten dür e, dass diese Ursache im Fällen von Vorstellungen des äußeren Sinns ein vom denkenden Subjekt verschiedener Gegenstand ist. Auch dieser Vorwurf lässt sich im Rahmen von Kants neuem Modell des Verhältnisses von Erscheinungen und Dingen an sich ausräumen. Wenn es ein und dieselben Dinge sind, von denen Kant sagt, dass wir wissen, wie sie uns erscheinen, aber nicht wissen, wie sie an sich selbst bescha en sind, dann ist jeder Grund dafür, dass ein äußerer Gegenstand als Erscheinung numerisch verschieden vom denkenden Subjekt als Erscheinung ist, auch ein Grund dafür, dass dieser Gegenstand an sich selbst, d.h. verstanden als Träger subjektunabhängiger Eigenscha en, verschieden vom Subjekt an sich selbst ist. Dann aber folgt die numerische Verschiedenheit des a zierenden Gegenstands vom a zierten Subjekt einfach daraus, dass es sich bei dem a zierenden Gegenstand um einen handelt, der als Erscheinung einen anderen Ort im Raum einnimmt, als es das denkende Subjekt tut, denn räumliche Verschiedenheit ist im Falle von Erscheinungen hinreichend für numerische Verschiedenheit (vgl. KrV A 263/B 319).³⁵

35 Dass die These von einer A ektion durch Dinge an sich mit Kants Erkenntnisbeschränkungen kompatibel ist, wenn man die Unterscheidung zwischen Dingen an sich und Erscheinungen nicht als eine zwischen zwei Arten von Gegenständen, sondern zwischen zwei Aspekten derselben Gegenstände interpretiert, hat bereits Marcus Willaschek gezeigt – ebenfalls in einem RolfPeter Horstmann gewidmeten Band (vgl. Willaschek, Marcus (2001). „A ektion und Kontingenz in Kants transzendentalem Idealismus“. In: R. Schumacher (Hg.). Idealismus als Theorie der Repräsentation? Paderborn: Mentis, 211–231). Die Zwei-Aspekte-Interpretation, die Willaschek dis-

Dinge an sich und der Außenweltskeptizismus

251

Für jemanden, der mit Jacobis und Schulzes Kritik an Kant sympathisiert, ist es naheliegend, gegen die eben vorgetragene Verteidigung einzuwenden, dass Kants vermeintlich neue Beschreibung des Verhältnisses von Erscheinungen und Dingen an sich genauso wie die alte dem Vorwurf ausgesetzt ist, dass darin die Kategorien von Ursache und Wirkung auf unzulässige Weise, d.h. über den Bereich der Erscheinungen hinaus, angewendet werden. Schließlich impliziert die neue Beschreibung, dass Gegenstände auf Grund von Eigenscha en, die ihnen an sich zukommen, auf uns einwirken. Dass diese Implikation nun sogar analytisch aus dem Begri der Erscheinung folgt, scheint die Chancen für eine Rettung des Kantischen Projekts eher zu verringern als zu verbessern. Ich denke, man kann zeigen, dass das neue Modell keine Aussagen über Dinge an sich als Ursachen unserer Vorstellungen beinhaltet, die Kant selbst bedenklich gefunden hätte oder die sich mit den wesentlichen Inhalten seines philosophischen Systems nicht vertragen. Die entscheidende Frage dabei ist, wie Kant seine These, dass die Kategorien nicht auf Dinge an sich angewendet werden dürfen, verstanden wissen wollte oder aus Gründen der Konsistenz seines Systems hätte verstehen müssen. Eine extreme Lesart dieser These wäre: Jede mit behauptender Kra gemachte Aussage,³⁶ die impliziert, dass Dinge auf Grund ihrer subjektunabhängigen Bescha enheit Wirkungen auf uns ausüben, ist unzulässig. Wäre diese Lesart richtig, dann wäre Kants Theorie zweifellos auch in der neuen Lesart widersprüchlich. Doch ist die Lesart zwingend? Sehen wir uns noch einmal eine Passage aus den Prolegomena an, die ich vorher in Ausschnitten bereits zitiert hatte: In der That, wenn wir die Gegenstände der Sinne wie billig als bloße Erscheinungen ansehen, so gestehen wir hiedurch doch zugleich, daß ihnen ein Ding an sich selbst zum Grunde liege, ob wir dasselbe gleich nicht, wie es an sich bescha en sei, sondern nur seine Erscheinung, d.i. die Art, wie unsre Sinnen von diesem unbekannten Etwas a cirt werden, kennen. Der Verstand also, eben dadurch daß er Erscheinungen annimmt, gesteht auch das Dasein

kutiert, versteht die Unterscheidung zwischen den beiden Aspekten allerdings methodologisch im Sinne von Prauss und Allison und nicht wie ich ontologisch (vgl. dazu oben Anm. 27). Willaschek weist zu Recht darauf hin, dass die methodologische Zwei-Aspekte-Interpretation scheitert, wenn sie erklären soll, wie Kant das An-sich-Sein der a zierenden Gegenstände für den empirischen Gehalt unserer Vorstellungen verantwortlich machen kann, und dass eben dies nur unter der Voraussetzung möglich wäre, dass der Unterschied zwischen den beiden Aspekten ontologischer Natur ist. Eine ontologische Zwei-Aspekte-Theorie ist laut Willaschek allerdings genauso wenig mit Kants Erkenntnisbeschränkung vereinbar wie eine Zwei-Welten-Interpretation (ebd. 223 ., besonders Anm. 26). Man kann die folgenden Ausführungen als den Versuch lesen, diese letzte Annahme zu widerlegen. 36 oder in Kants Terminologie: „Jedes assertorische und nicht bloß problematische Urteil […]“.

252

Rosefeldt

von Dingen an sich selbst zu, und so fern können wir sagen, daß die Vorstellung solcher Wesen, die den Erscheinungen zum Grunde liegen, mithin bloßer Verstandeswesen nicht allein zulässig, sondern auch unvermeidlich sei. Unsere kritische Deduction schließt dergleichen Dinge (noumena) auch keinesweges aus, sondern schränkt vielmehr die Grundsätze der Ästhetik dahin ein, daß sie sich ja nicht auf alle Dinge erstrecken sollen, wodurch alles in bloße Erscheinung verwandelt werden würde, sondern daß sie nur von Gegenständen einer möglichen Erfahrung gelten sollen. Also werden hiedurch Verstandeswesen zugelassen, nur mit Einschärfung dieser Regel, die gar keine Ausnahme leidet: daß wir von diesen reinen Verstandeswesen ganz und gar nichts Bestimmtes wissen, noch wissen können, weil unsere reine Verstandesbegri e sowohl als reine Anschauungen auf nichts als Gegenstände möglicher Erfahrung, mithin auf bloße Sinnenwesen gehen, und, so bald man von diesen abgeht, jenen Begri en nicht die mindeste Bedeutung mehr übrig bleibt (AA IV 314 f.).

Natürlich kann man diese Passage so lesen, dass Kant hier behauptet, dass sein System die Existenz von Dingen an sich als Ursachen unserer Vorstellungen sowohl „unvermeidlich“ beinhaltet als auch impliziert, dass jede Aussage über Dinge an sich als Ursachen unzulässig ist. (Die Geschichte der nachkantischen Philosophie hat gezeigt, dass diese Lesart möglich ist!) Man muss Kant dann entweder unterstellen, dass er kurzzeitig vergisst, dass der Begri der A ektion ein kausaler Begri ist oder dass der Begri der Ursache ein reiner Verstandesbegri ist, oder man muss annehmen, dass Kant innerhalb von nur einer Textseite einen Widerspruch behauptet, bei dem es – um Fichtes oben zitierte Worte zu verwenden – schwer fällt, ihn „irgend einem Menschen, der seiner Vernun noch mächtig ist, zuzutrauen“. Nach allen Regeln des wohlwollenden Interpretierens wäre es allerdings geboten, Kants Formulierung, dass wir von Dingen an sich „ganz und gar nichts Bestimmtes wissen“ und von reinen Verstandesbegri en wie dem der Ursache bei der Anwendung auf Dinge an sich „nicht die mindeste Bedeutung mehr übrig bleibt“, so zu interpretieren, dass sie nicht das Wissen davon ausschließt, dass Dinge an sich auf uns einwirken. Die Frage ist dann natürlich: Wie soll man diese Formulierungen dann verstehen? Gibt es überhaupt eine plausible Alternative zu der eben genannten extremen Lesart von Kants Formulierung? Ich denke, dass die zitierte Passage und zahlreiche andere Stellen, in denen Kant die Einschränkung des Kategoriengebrauchs auf den Bereich der Erscheinungen behauptet, nahelegen, dass es insbesondere drei Dinge sind, die er im Zusammenhang der Verwendung von Kategorien für die Rede über Dinge an sich als unzulässig ausschließen will. Wie ich im Folgenden zeigen werde, schließt keine der drei Einschränkungen das Wissen davon aus, dass Dinge auf Grund ihrer subjektunabhängigen Bescha enheit Wirkungen auf uns ausüben.

Dinge an sich und der Außenweltskeptizismus

253

Einschränkung 1: Man darf die Kategorien nicht auf Gegenstände anwenden, die uns nicht erscheinen. Kants Hauptargument für die These, dass wir durch den Gebrauch von Kategorien nur dann etwas erkennen können, wenn wir diese auf Erscheinungen anwenden, lautet: Um Gegenstände zu erkennen, reicht es nicht aus, sie zu denken, sondern sie müssen uns auch in der Anschauung gegeben sein. Unsere Anschauung ist aber sinnlich und durch sinnliche Anschauung können uns nur Erscheinungen gegeben werden.³⁷ Ganz in diesem Sinne warnt Kant in den Prolegomena davor, dass „sich der Verstand unvermerkt an das Haus der Erfahrung noch ein viel weitläu igeres Nebengebäude an[baut], welches er mit lauter Gedankenwesen anfüllt, ohne es einmal zu merken, daß er sich mit seinen sonst richtigen Begri en über die Grenzen ihres Gebrauchs verstiegen habe“ (AA IV 316). Das heißt: Der Gebrauch einer Kategorie wie derjenigen der Ursache ist dann problematisch, wenn wir dabei Gegenstände zu erkennen meinen, die mit keinem der uns sinnlichen erscheinenden Gegenstände identisch sind.³⁸ Tatsächlich ist die transzendentale Dialektik, der Teil der Kritik der reinen Vernun , in dem Kant seine Kritik an einem unzulässigen Kategoriengebrauch im Detail ausführt, voll von Beispielen für Gegenstände, die uns nicht sinnlich erscheinen, von denen die traditionelle Metaphysik aber Wissen zu haben meinte: erste Ursachen, unteilbare kleinste Teile der Materie, einfache und den menschlichen Tod überdauernde Seelen, Gott.³⁹ Gegen Einschränkung 1 verstößt Kant nicht, wenn er behauptet, dass Dinge auf Grund ihrer subjektunabhängigen Eigenscha en die Ursachen für unsere Vorstellungen sind. Denn die Dinge, von denen dabei die Rede ist, sind ja laut Kants

37 Vgl. z.B. § 22 der transzendentalen Deduktion (KrV B 146 .), A 246 f./B 302 f., Prolegomena AA IV 316 f. 38 Vgl. auch A 252, wo Kant nur ausschließt, dass „ein Noumenon einen wahren, von allen Phänomenen zu unterscheidenden Gegenstand bedeute“. 39 Dass Kant mit seiner Kritik am Gebrauch der Kategorien über den Bereich der Erscheinungen hinaus diejenigen philosophischen Verfahren im Sinn hat, die er in der transzendentalen Dialektik kritisiert, wird auch an einer Stelle der Prolegomena deutlich, wo er den unzulässigen Kategoriengebrauch folgendermaßen charakterisiert: „[…] so wird der Verstand aus seinem Kreise getrieben, um theils Gegenstände der Erfahrung in einer so weit erstreckten Reihe vorzustellen, dergleichen gar keine Erfahrung fassen kann, theils sogar (um sie zu vollenden) gänzlich außerhalb derselben Noumena zu suchen, an welche sie jene Kette knüpfen und dadurch von Erfahrungsbedingungen endlich einmal unabhängig, ihre Haltung gleichwohl vollständig machen könne. Das sind nun die transscendentalen Ideen, welche […] einen unvermeidlichen Schein dem Verstande einen transscendenten Gebrauch ablocken […]“ (AA IV 332 f.). Man bedenke, wie verschieden das Vorgehen, das Kant hier kritisiert, von der harmlosen Behauptung ist, dass die Dinge, die uns erscheinen, auch irgendwie an sich selbst bescha en sind.

254

Rosefeldt

neuer Version des transzendentalen Idealismus keine anderen Dinge als diejenigen, die uns erscheinen. Anders formuliert: Kant muss von keinem einzigen Ding, das uns nicht erscheint, behaupten, dass es subjektunabhängige Eigenscha en hat und auf Grund dieser Eigenscha en Ursache für etwas anderes ist. Alle an sich existierenden und wirkenden Gegenstände, auf die seine Theorie verp ichtet ist, sind gleichzeitig solche, die uns durch unsere sinnliche Anschauung gegeben sind. Einschränkung 2: Wir können davon, wie die Dinge an sich selbst bescha en sind, nie etwas „Bestimmtes“ wissen. Kants in der zitierten Passage aus den Prolegomena gemachte Behauptung, dass wir von dem a zierenden Ding an sich „ganz und gar nichts Bestimmtes wissen, noch wissen können“,⁴⁰ kann in einer wohlwollenden Lesart wiederum nicht bedeuten, dass wir von dem a zierenden Gegenstand überhaupt nichts wissen können, denn nach Kants eigener Theorie wissen wir von ihm ja, dass er irgendwelche subjektunabhängigen Eigenscha en hat, auf Grund deren er in uns bestimmte Vorstellungen hervorru . Es liegt nahe, Kants Bemerkung so zu verstehen, dass wir nur dann etwas Bestimmtes von dem a zierenden Gegenstand wissen würden, wenn wir wissen würden, welche seiner An-sich-Eigenscha en kausal dafür verantwortlich sind, dass er uns so-und-so erscheint. Auf die obige Pilz-Analogie übertragen hieße das: Solange wir von dem Pilz nur wissen, dass er gi ig ist, haben wir keinen bestimmten Begri davon, wie er an sich selbst (chemisch) bescha en ist, und zwar selbst dann nicht, wenn wir wissen, dass es irgendeine seiner ihm unabhängig von uns zukommenden chemischen Eigenscha en ist, die für unsere Vergi ungssymptome verantwortlich ist. Die vorgeschlagene Interpretation kann sich auf eine Passage aus den Prolegomena stützen, in der Kant sagt, dass „Erscheinungen doch jederzeit eine Sache an sich selbst voraussetzen und also darauf Anzeige thun, man mag sie nun näher erkennen, oder nicht.“ Er fährt fort: „Da wir nun aber diese Verstandeswesen nach dem, was sie an sich selbst sein mögen, d.i. bestimmt, niemals erkennen können, gleichwohl aber solche im Verhältniß auf die Sinnenwelt dennoch annehmen und durch die Vernun damit verknüpfen müssen: so werden wir doch

40 Vgl. auch KrV B 307, wo Kant davon spricht, dass wir nur über einen „ganz unbestimmten Begri von einem Verstandeswesen als einem Etwas überhaupt außer unserer Sinnlichkeit“ verfügen, der kein „bestimmte[r] Begri von einem Wesen [ist], welches wir durch den Verstand auf einige Art erkennen könnten“, und A 252, wo Kant sagt, dass unser Begri vom Ding an sich keine „bestimmte Erkenntniß von irgend einem Dinge, sondern nur das Denken von Etwas überhaupt bedeutet“.

Dinge an sich und der Außenweltskeptizismus

255

wenigstens diese Verknüpfung vermittelst solcher Begri e denken können, die ihr Verhältniß zur Sinnenwelt ausdrücken. Denn denken wir das Verstandeswesen durch nichts als reine Verstandesbegri e, so denken wir uns dadurch wirklich nichts Bestimmtes […]“ (AA IV 355). Wenn Kant hier davon spricht, dass man Dinge an sich voraussetzen muss, egal ob man sie „näher erkennen“ kann „oder nicht“, meint er wohl, dass wir Dinge an sich nicht „näher“ erkannt haben müssen, um sagen zu können, dass es sie gibt und dass sie uns irgendwie a zieren. Nur der Anspruch eines solches „näheren“ Erkennens, bei dem wir etwas Bestimmteres über das An-sich-Sein des Gegenstands herausbekommen, wird von ihm zurückgewiesen. Die Interpretation passt zudem sehr gut zu Kants Rede davon, dass wir im Bereich der Erscheinungen durch die Anwendung der Kategorien von Ursache und Wirkung die objektive Zeitfolge von wahrgenommenen Zuständen „bestimmen“ (vgl. KrV 232 .). Solches Bestimmen kann nur dann statt nden, wenn man den Zustand kennt, über dessen Verhältnis zu anderen Zuständen man mit Hilfe des Begri s der Ursache urteilt. Es ndet also auch bei Kants Aussagen über a zierende Dinge an sich nicht statt, denn diese Aussagen implizieren, wie gesagt, keine Angaben darüber, auf Grund welcher subjektunabhängigen Eigenscha en die Dinge Vorstellungen in uns hervorrufen. Als letzter Beleg kann Kants De nition von „Bestimmen“ aus der Fortschrittsschri dienen: „Bestimmen […] heißt synthetisch urtheilen“ (AA XX 268). Jedes Urteil darüber, welche subjektunabhängige Eigenscha eines Gegenstands unsere jeweiligen Vorstellungen verursacht, müsste synthetisch sein. Das Urteil aber, dass ein uns erscheinender Gegenstand irgendeine subjektunabhängige Eigenscha hat, auf Grund deren er uns auf bestimmte Weise erscheint, ist nicht synthetisch, sondern ergibt sich, wie gezeigt, aus dem Begri der Erscheinung. Einschränkung 3: Daraus, dass man unter Verwendung der Kategorien „unbestimmte“ Aussagen darüber macht, wie Dinge an sich selbst bescha en sind, darf man nicht folgern, dass die normalerweise mit diesen Kategorien verbundenen Grundsätze für Dinge an sich gültig sind. Wenn Kant die Anwendung der Kategorien auf den Bereich der Erscheinungen einschränkt, verbindet er das fast immer mit einem Zusatz wie dem folgenden: „[…] Alle synthetische Grundsätze a priori sind nichts weiter als Principien möglicher Erfahrung […] und können niemals auf Dinge an sich selbst, sondern nur auf Erscheinungen als Gegenstände der Erfahrung bezogen werden“ (AA IV 313; vgl. auch A 238/ B 297). Die These, dass wir nicht zu der Annahme berechtigt sind, Grundsätze wie der der Kausalität würden für Dinge an sich gelten, ist zentral für Kants Projekt einer Kritik der vorkritischen Metaphysik, die viele ihrer Argumente auf der Grundlage solcher Grundsätze entwickelt. Auch Kants Versuch, durch

256

Rosefeldt

den transzendentalen Idealismus Freiheit und Naturnotwendigkeit zu vereinbaren, kann nur dann erfolgreich sein, wenn der Geltungsbereich der Grundsätze eingeschränkt ist. In der Vorrede zur zweiten Au age der Kritik der reinen Vernun begründet Kant die Wichtigkeit der Unterscheidung zwischen Erscheinungen und Dingen an sich folgendermaßen: „Nun wollen wir annehmen, die durch unsere Kritik nothwendiggemachte Unterscheidung der Dinge als Gegenstände der Erfahrung von eben denselben als Dingen an sich selbst wäre gar nicht gemacht, so müßte der Grundsatz der Causalität und mithin der Naturmechanism in Bestimmung derselben durchaus von allen Dingen überhaupt als wirkenden Ursachen gelten. Von eben demselben Wesen also, z.B. der menschlichen Seele, würde ich nicht sagen können, ihr Wille sei frei, und er sei doch zugleich der Naturnothwendigkeit unterworfen, d.i. nicht frei, ohne in einen o enbaren Widerspruch zu gerathen, weil ich die Seele in beiden Sätzen in eben derselben Bedeutung, nämlich als Ding überhaupt (als Sache an sich selbst), genommen habe und ohne vorhergehende Kritik auch nicht anders nehmen konnte“ (KrV B XXVII). Wenn man bedenkt, wie wichtig für Kants Philosophie die beiden philosophischen Projekte sind, für die die Einschränkung der Grundsätze auf den Bereich der Erscheinungen notwendig ist, liegt es nahe anzunehmen, dass die dritte Einschränkung die für ihn wichtigste ist. Wieder verstößt Kant gegen diese Einschränkung nicht dadurch, dass er behauptet, dass die Gegenstände, die uns erscheinen, dies dadurch tun, dass sie uns auf Grund von irgendwelchen subjektunabhängigen Eigenscha en auf bestimmte Weise a zieren. Diese Behauptung impliziert nicht, dass der Grundsatz der Kausalität für diese Gegenstände als Dinge an sich gültig ist, dass es also für jeden ihrer subjektunabhängigen Zustände eine Ursache gibt. Und wie oben erläutert muss man auch nicht die Gültigkeit dieses Grundsatzes für Dinge an sich annehmen, um auf der Basis einer gegebenen Wirkung (unserer Vorstellungen) auf deren Ursache (das Ding an sich) zu schließen – anders als dies im Modell des vierten Paralogismus der Fall war. Die drei genannten Einschränkungen bringen meiner Ansicht nach genau das zum Ausdruck, was Kant mit der These gemeint hat, dass der Gebrauch der Kategorien auf den Bereich der Erscheinungen eingeschränkt ist. Da Kant seit den Prolegomena gegen keine dieser drei Einschränkungen mehr verstößt, tri der Inkonsistenzvorwurf spätestens seit diesem Punkt der Entwicklung seines Systems nicht mehr zu, und Kant hat sich also der Kritik von Jacobi und Schulze entzogen, bevor diese überhaupt formuliert war.

Dinge an sich und der Außenweltskeptizismus

257

5 Transzendentaler Idealismus und Außenweltskeptizismus Ein Vorwurf bleibt. Die Rettung von Kants Aussagen über Erscheinungen vor dem Absurditätsvorwurf und diejenige seiner Behauptungen über Dinge an sich vor dem Inkonsistenzvorwurf gelingt nur, weil Kant dabei einen gegenüber dem vierten Paralogismus wesentlich anspruchsvolleren, d.h. realistischeren Erscheinungsbegri ins Spiel bringt. Erkenntnis von einer Erscheinung haben wir jetzt nur noch dann, wenn der Gegenstand unserer Wahrnehmung und unserer Urteile auch unabhängig von uns existiert und unabhängig von uns irgendwie bescha en ist. Das war im Modell des vierten Paralogismus anders. Die dortigen Erscheinungen sind nichts als Vorstellungen, und es ist begri ich nicht ausgeschlossen, dass solche Erscheinungen existieren, ohne dass irgendetwas von uns Unabhängiges existiert. Die Einführung eines realistischeren Erscheinungsbegri s hat philosophische Kosten. Es stellt sich nun nämlich die Frage, was Kant berechtigt anzunehmen, dass wir jemals Erkenntnis von Erscheinungen haben: Wenn die Gegenstände, die uns im äußeren Sinn erscheinen, also Körper, unabhängig von uns existieren, d.h. nicht selbst Vorstellungen sind oder zumindest keine Dinge, die vollständig durch unsere Vorstellungen von ihnen konstituiert sind, wie können wir dann sicher sein, dass wir jemals Körper wahrnehmen? Anders formuliert: Welche Antwort hat der transzendentale Idealismus in der neuen Interpretation noch auf den Außenweltskeptizismus, den Kant im Paralogismenkapitel widerlegen wollte? Diese Frage ist sehr berechtigt. Durch die Modi kation seiner Darstellung des transzendentalen Idealismus allein hat Kant kein Argument gegen den Außenweltskeptizismus geliefert. Vielmehr hat er sich damit die Möglichkeit des „billigen“ Argumentes aus dem Paralogismenkapitel – Körper existieren, denn Körper sind bloße Vorstellungen – verbaut. Kant hat diesen Punkt sehr klar gesehen und er hat deswegen in der zweiten Au age der Kritik der reinen Vernun im Abschnitt über die „Widerlegung des Idealismus“ einen ganz neuen Beweis der Realität der Außenwelt vorgelegt. Der Beweis soll „darthun, daß wir von äußeren Dingen auch Erfahrung und nicht bloß Einbildung haben“ (KrV B 275), d.h. – nach der hier vertretenen Interpretation – dass es sich bei den Gegenständen unseres äußeren Sinns um wirkliche, von uns unabhängige Dinge handelt, und nicht bloß um Geschöpfe unserer Einbildung, die bloß in uns existieren. Ich kann mich im Rahmen dieses Beitrags nicht mit den Details von Kants ‚Widerlegung‘ auseinandersetzen. Ganz grob läu Kants Argumentation darauf hinaus, zu zeigen, dass wir nur dann Wissen von der objektiven zeitlichen Abfolge unserer eigenen geistigen Zustände haben können – ein Wissen, das der Außenweltskeptiker anerkennen muss – wenn wir zumindest manchmal wirkliche, das heißt von unseren Vorstel-

258

Rosefeldt

lungen verschiedene Gegenstände wahrnehmen.⁴¹ Das Bewusstsein der eigenen Existenz in der Zeit, schreibt Kant, sei „nur durch ein Ding außer mir und nicht durch die bloße Vorstellung eines Dinges außer mir möglich“. Bemerkenswert an dieser These ist, dass auch an ihr deutlich wird, wie sehr sich Kants Position gegenüber der im Paralogismenkapitel geändert hat. Das Insistieren auf dem Unterschied zwischen „einem Ding außer mir“ und „der Vorstellungen eines Dinges außer mir“ wäre dort undenkbar gewesen, denn das Argument im Paralogismenkapitel beruhte ja gerade auf der Annahme, dass Dinge außer mir selbst nichts als Vorstellungen sind. Ich möchte nicht behaupten, dass Kants Versuch einer Widerlegung des Außenweltskeptizismus erfolgreich oder auch nur besonders plausibel ist. Nehmen wir einmal an, er scheitert. Wie schlimm wäre das? Die Antwort auf diese Frage hängt natürlich davon ab, für wie wichtig man die Lösung des Problems des Außenweltskeptizismus hält. Jacobi, Schulze und Fichte hielten sie o ensichtlich für sehr wichtig, und das ist natürlich auch ihr gutes Recht. Auch Kant hielt die Argumente, die in der philosophischen Tradition für den Außenweltskeptizismus entwickelt wurden, o ensichtlich für bedenkenswert genug, um auch der zweiten Au age seiner Kritik der reinen Vernun eine Widerlegung des problematischen Idealismus hinzuzufügen. Aber er stellt diese Widerlegung nun klugerweise nicht mehr als weiteres indirektes Argument für die Richtigkeit seines transzendentalen Idealismus dar. Und das zeigt, dass er nun selbst der Meinung war, dass es keinen direkten Zusammenhang zwischen seinem transzendentalen Idealismus und der Möglichkeit einer Widerlegung des Außenweltskeptizismus gibt.⁴² In unserem Zusammenhang muss die Frage aber vor allem lauten, ob die Widerlegung des Außenweltskeptizismus wesentlich für Kants philosophisches Projekt ist, denn der Vorwurf, um den es bei Jacobi und Schulze ging, war ja nicht, dass dieses Projekt unvollständig ist, sondern dass es inkonsistent ist. Ich denke, dass die Antwort auf diese Frage „nein“ lauten muss. Wesentlich für Kants Projekt ist allenfalls die Widerlegung eines Humeschen Kausalitätsskeptizismus. Der

41 Diese Darstellung ist in der Literatur selbstverständlich umstritten, nicht zuletzt deswegen, weil natürlich nicht alle Kant-Interpreten Kant einen so realistischen Begri eines Gegenstandes des äußeren Sinnes unterstellen, wie ich ihn hier propagiert habe. Zu der Problematik, wie stark die Konklusion ist, für die Kant mit der Widerlegung argumentieren will, vgl. Guyer, Paul (1987). Kant and the Claims of Knowledge. Cambridge: Cambridge University Press, Kap. 12. 42 Zudem gesteht Kant zu, dass sein Argument in der „Widerlegung des Idealismus“ nur zeigt, dass wir überhaupt manchmal echte äußere Erfahrung haben, und dass dies allein uns nicht zwischen echter äußerer Erfahrung und nur scheinbarer äußerer Erfahrung unterscheiden lässt (vgl. KrV B 278 f.). Wichtige Argumente für den Außenweltskeptizismus wie etwa Descartes’ Traumargument sind davon also gar nicht betro en.

Dinge an sich und der Außenweltskeptizismus

259

transzendentale Idealismus wird von Kant eingeführt, um die Möglichkeit synthetischer Urteile a priori verständlich zu machen, und möglicherweise auch dazu, ihre Existenz gegen skeptische Zweifel zu verteidigen. Das Urteil, dass die Zustände der von uns wahrgenommenen Gegenstände in Kausalbeziehungen zueinander stehen, ist ein solches synthetisches Urteil a priori. Die Annahme, dass wir uns die wahrgenommene Welt nicht ausdenken, sondern von uns unabhängige Gegenstände dafür verantwortlich sind, was wir wahrnehmen, ist es nicht. Wenn meine Darstellung der ersten 15 Jahre der Rezeptionsgeschichte von Kants Lehre vom Ding an sich korrekt ist, ist diese Geschichte die Geschichte zweier Missverständnisse. Das erste Missverständnis besteht darin, dass keinem der von mir referierten Kritiker Kants die grundlegende Änderung zwischen dem Paralogismenkapitel der ersten Au age der Kritik der reinen Vernun und den Aussagen über Erscheinungen und Dinge an sich in deren zweiten Au age aufgefallen ist. Das zweite Missverständnis liegt in der Annahme, dass die Überwindung jeder Art von Skeptizismus wesentlich für Kants transzendentalphilosophisches Projekt ist. Wie die Geschichte der nachkantischen Philosophie zeigt, hat dieses Missverstehen von Kant eine enorm produktive philosophische Kra entfaltet. Wir müssen den beteiligten Philosophen dafür also eigentlich dankbar sein.

Andrew Chignell

Ogilby, Milton, Canary Wine, and the Red Scorpion Another look at Kant’s Deduction of Taste* 1 The Ogilby Problem In “Of the Standard of Taste” (1757), Hume introduces a classic puzzle regarding what he calls the “two species of common sense regarding taste.” On the one hand, Hume says, when we theorize about taste at an abstract level we naturally think that beauty is “no quality in the things themselves […] but exists only in the mind which contemplates them.” As a result, we conclude that “every individual ought to acquiesce in his own sentiment, without pretending to regulate those of others.”¹ At a suitably abstract level, it seems, we take it to be just obvious that de gustibus non disputandum est: you have your taste and I have mine, just as we do in genuinely gustatory matters. On the other hand, Hume says, when we go to the level of particulars, there is something in common sense that resists this permissive attitude with respect to aesthetic taste. I don’t mind that you adore cucumbers while I think they taste like dirty socks, but I do mind, says Hume, if our aesthetic reactions di er, at least with respect to paradigm cases. Here he cites our comparative evaluations of, on the one hand, John Milton, famous 17th century poet and author of the epic Paradise

* I am happy to dedicate this piece to Rolf-Peter Horstmann upon his retirement from teaching. Rolf has kindly hosted me (along with numerous other Ausländern) at the Humboldt University nearly every Sommersemester for the past twelve years or so. His generosity and warmth – combined with a healthy skepticism and complete intolerance for obscurantism – is a large part of what made turn-of-the-millennium Berlin such a vibrant place for those working on classical German philosophy. I am grateful to have been a part of it. Quotations from Kant are translated from Kant, Immanuel. 1900–. Immanuel Kants Schri en. Berlin: Ausgabe der koeniglich preussischen Akademie der Wissenscha en (now de Gruyter). I will follow standard practice and cite the pagination from this edition as (volume:page), unless the rst Critique is being referred to, in which case I cite it as (A-edition pagination/B-edition pagination). I have o en consulted or used the English translations in Kant, Immanuel. 2000 (1790). Critique of the Power of Judgement. Trans. and eds. P. Guyer and E. Matthews. New York: Cambridge University Press, and Kant, Immanuel. 1987 (1790). Critique of Judgment. Trans. and ed. W. Pluhar. Indianapolis: Hackett Publishing Company. 1 Hume, David. 1898 (1757). “Of the Standard of Taste,” in: Essays Moral, Political and Literary, ed. by T. H. Green and T. H. Grose, London: Longmans, Green and Co., 269.

262

Chignell

Lost and, on the other hand, John Ogilby, obscure 17th century author of the notso-epic Frog, or, the Netherland Nightingale, Sweet Singer of Amsterdam. Ogilby was an important atlas-maker, but his poetry and translation work had already been derided by the likes of Dryden and Pope, as Hume would have known. At the level of particular objects or works such as these, says Hume, we don’t readily tolerate disagreement: “Whoever would assert an equality of genius and elegance between O and M ,” he writes, would be thought to defend no less an extravagance, than if he had maintained a molehill to be as high as T , or a pond as extensive as the ocean. Though there may be found persons, who give the preference to [Ogilby]; no one pays attention to such a taste; and we pronounce without scruple the sentiment of these pretended critics to be absurd and ridiculous (269).

In the end, Hume appears to resolve the puzzle by simply jettisoning the rst common sense thought about the indisputability of taste and holding onto the second one – i.e. the one according to which we must take someone who thinks Ogilby’s poetry is more valuable than Milton’s to be making a mistake akin to thinking a mole-hill is bigger than the largest volcano in the Canary Islands. Hume still thinks that beauty is a mere sentiment of the mind, but he claims that there has to be something about particular objects – including Milton’s poems – that is disposed to produce that sentiment in all properly-functioning and properlycultivated readers. He thus grounds the normativity of our judgments of taste – even if not their intersubjectivity – in dispositions such as these. In the early parts of Critique of the Power of Judgment (1790) Kant goes through what he calls an “Analytic of the Beautiful” – an analysis of the main moments in a pure judgment of taste – and notices, in the process, the same puzzle that Hume did. On the one hand, a pure judgment of taste is a subjective and singular state that is generated by an individual beholder on the basis of sensory experience: as a result, it seems destined to be an expression of mere personal preference. On the other hand, there is a kind of universality or even normativity that attaches to our paradigmatic aesthetic judgments – those that are made in abstraction from any practical and economic interests we may have. They speak, as he puts it, with a “universal voice.” This obviously raises a problem, which according to Kant can be represented thus: How is a judgment possible which, merely from one’s own feeling of pleasure in an object, independent of its concept, judges (beurtheilte) this pleasure as attached to the representation of that same object in all other subjects, and does so a priori, i.e., without having to wait for the assent of others? (5:288, my italics; cf. 5:290n)

We do judge that much of Milton’s poetry is beautiful, and that much of Ogilby’s is not, and without any consultation we expect, assume, or even demand (Kant

Ogilby, Milton, Canary Wine, and the Red Scorpion

263

uses words like “fordern,” “ansinnen,” and “zumuten” in this context) that others will do so as well. When someone judges to the contrary, we are surprised and unimpressed; we typically ask him to take another look at Paradise Lost alongside Frog, or, the Netherland Nightingale, Sweet Singer of Amsterdam. If he persists in his judgment, we regard him askance – as somehow kidding us or kidding himself, as having some agenda or even an economic interest in Ogilby’s literary success, as missing the point or, worse, as maladjusted or malfunctioning. If we care about him or his reputation, we may seek to draw his attention to the features of Paradise Lost that we take to contribute to our feeling of disinterested pleasure in the object, or to the features of Frog, or, the Netherland Nightingale that we take to underwrite our feeling of distaste. But we do not simply accept his judgment as, in the undergraduate phrase, a “valid opinion,” and then remind ourselves that taste, a er all, is not disputable. Kant considered this problem in lectures dating back to the middle 1770s, but he only seriously tries to deal with it in the third Critique. In what follows I will sketch a solution to the problem that I think we nd in Kant’s texts (even if it is not the only one that we nd in Kant’s texts). The solution crucially involves what Kant calls “aesthetic ideas” and their capacity to symbolize rational ideas. Here I am joining commentators like Hermann Cohen, Anthony Savile, Heiner Bielefeldt, and Kenneth Rogerson in holding that the deduction of taste is only really completed in § 42 and following, when Kant begins the discussion of aesthetic ideas.² A er sketching my own version of the proposal, I’ll go on to discuss how this kind of view can handle a further problem related to Hume’s question about Ogilby. Finally, I’ll note that Rolf-Peter Horstmann’s discussions of Kant’s “principle of purposiveness” in general o ers a clue as to how the account that I sketch with respect to art objects might be extended to natural objects as well.

2 The Gap in the Deduction of Taste To understand Kant’s approach to this problem, we rst need to know more about the proximate basis of an aesthetic judgment for him, given that it isn’t any physical feature of the object itself and also isn’t any sort of interest or desire on the

2 Cohen, Hermann. 1889. Kants Begründung der Ästhetik, Berlin: Dümmler; Bielefeldt, Heiner. 2001. Kants Symbolik: Ein Schlüssel zur kritischen Freiheitsphilosophie, Freiburg: Karl Alber; Savile, Anthony. 1987. Aesthetic Reconstructions: The Seminal Writings of Lessing, Kant, and Schiller, Oxford: Basil Blackwell; Rogerson, Kenneth. 2008. The Problem of Free Harmony in Kant’s Aesthetics, Albany, NY: SUNY Press.

264

Chignell

part of the subject. Here is just a brief sketch of what I take to be his view:³ In ordinary experience, the faculty of understanding (Verstand) employs concepts to synthesize and sort the material presented by the senses. But, says Kant, in the course of some exceptional experiences, the world comes to us in such a way that it seems especially intelligible or signi cant, even apart from its susceptibility to the usual conceptualization. The object or vista we’re (disinterestedly) contemplating seems to be unusually deep, to have a supra-conceptual signi cance, to point beyond itself to something of ultimate importance. Another way to put this is to say that the object, as well as the episode that it occasions, strike us as extraordinarily “purposive” (Zweckmässig) even apart from our use of a concept to determine the object’s actual nature or “purpose” (Zweck) (5:188 .). There seems to be a kind of cognitive coyness about the object: we nd that we are not able to apply concepts to it in such a way that we feel we’ve fully analyzed its signi cance for us. We can of course bring it under some concepts – concepts of a painting or a sculpture or a work of music, of representing an Italian woman or the painter himself or an ancient Egyptian queen. But, again, there is something special about this object that seems to resist normal cognitive techniques – a kind of inarticulable depth or richness or import that brings the understanding up short with respect to its drive to categorize, conceptualize, analyze, and dissect. A surprising feature of this cognitive coyness in beautiful objects is that we respond to it not with annoyance or enervation, but rather with an exquisite kind of fascination.⁴ The object presents itself to us as somehow full of purport, and so we keep looking, trying out di erent analyses, taking new critical angles, trying to pluck out the heart of its mystery even while enjoying, in a unique way, the fact that we do not succeed. We also reject critical e orts to spell out in some nalized way the true and complete signi cance of the object. This state or series of states, I think, is a large part of what Kant dubs the “free play” or “unexpected harmony” of the faculties of cognition. It is a free play, because the information o ered by sense/imagination is not fully captured by determinate concepts in the usual way. It is unexpectedly harmonious because, despite our self-conscious lack of full understanding, we still feel that what is presented is somehow purposive or signi cant for us – that it promises a kind of happiness (to use Alexander Ne-

3 See Chignell, Andrew. 2007. “Kant on the Normativity of Taste: The Role of Aesthetic Ideas,” in: Australasian Journal of Philosophy 85 (3), 415–33, for more. I have drawn on this paper for some of the exposition of my view here. The discussion of the Ogilby problem, however, is entirely new and, indeed, formulated in response to feedback on this earlier piece. 4 On the cognitive “re ection without frustration” involved here, see Joseph Cannon’s very illuminating 2008. “Intentionality of Judgments of Taste in Kant’s Critique of Judgment,” in: Journal of Aesthetics and Art Criticism 66 (1), 53–65.

Ogilby, Milton, Canary Wine, and the Red Scorpion

265

hamas’ phrase), or seeks to tell us something important, even if we can’t place our nger on precisely what.⁵ It is on the basis of – or in virtue of having – the pleasure occasioned by such harmonious free play that an authentic judgment of taste is made (5:186–8).⁶ It should be clear that, from the point of view of grounding the universality of particular judgments of taste, there is still a signi cant slip between cup and lip. The fact that we all possess faculties that make us capable of this experience of free play does not entail that we all will have that experience in the presence of the same objects. This is a problem that Paul Guyer brought out in his early work on Kant’s aesthetics, starting in the late 1970’s, though it was implicit in previous commentators such as Donald Crawford.⁷ The problem points to a gap in the Deduction of Taste, one that subsequent commentators have responded to either by trying to ll it using resources from other parts of the third Critique, or by denying that the gap needs lling at all. My own preferred approach is to view Kant as going along with Hume in response to the puzzle with which we started, at least for some distance. That is, I think Kant ultimately locates the cause of the free play in something about the object of the judgment. But he does so in a unique way, or so I want to suggest – one that is not fundamentally at odds with the “subjective” and “singular” characteristics of aesthetic judgment that he highlights in the Analytic.

3 Filling the Gap: Rational Ideas, Aesthetic Ideas, Aesthetic Attributes We need a few more conceptual building blocks in order to see how (I submit) the gap in the Deduction can be lled. They are the concepts of (1) a rational idea; (2) an aesthetic idea; and (3) an aesthetic attribute. By rational or transcendental idea Kant means a concept to which we are led “in an entirely necessary way by reason according to its original laws,” but which refers to something beyond our cognitive

5 Nehamas, Alexander. 2007. Only the Promise of Happiness, Princeton, NJ: Princeton University Press. 6 I articulate the view in this way in order to stay neutral between competing positions on whether the judgment or the pleasure is more fundamental. See § 9 of the Critique as well as the classic discussion of this issue in Ginsborg, Hannah. 1991. “On the Key to Kant’s Critique of Taste,” in: Paci c Philosophical Quarterly 72 (4), 290–313. 7 Crawford, Donald. 1974. Kant’s Aesthetic Theory. Madison, WI: University of Wisconsin Press; Guyer, Paul. 1997 (1979). Kant and the Claims of Taste, 2nd edition, New York: Cambridge University Press.

266

Chignell

ken (KrV A 339/B 397). Included among these ideas are not just representations of the familiar supersensibilia of metaphysics and theology, but also mathematical ideas of in nity, the in nitesimal, and “the maximum” generally; various cosmological ideas of absolute motions, temperatures and distances, the cosmic whole, and the rst moment of time; exemplary ideas of a perfect human body or a perfect example of a species; and also moral ideas of perfect virtue or perfect love or freedom unconstrained by natural laws. Finally, there is the important idea of full and complete natural systematicity. This idea is really the synthesis of two other ideas: that of the in nitely expansive empirical cosmos or world-whole, and that of a perfect epistemological system in which concepts of particulars fall under concepts of contingent empirical laws which are in turn arranged hierarchically such that minds like ours can aim at full scienti c comprehension. As recent commentators (including Horstmann) have pointed out, the principle of natural systematicity grows in signi cance for Kant over the course of his career.⁸ Initially in the Critique of Pure Reason he took it to be a merely “logical” maxim of reason – a subjective heuristic device that we can use in order to make provisional claims about the goal-directed character of certain parts of nature. Later, and most clearly in the introductions to the third Critique, he seems to view it as a necessary presupposition of empirical concept-formation, and thus of scienti c inquiry about the empirical world. Unlike determinate empirical concepts such as ower or fork, rational ideas cannot be adequately ‘exempli ed’ by any empirical experience [A 327/B 383]. Moreover, speculative reason cannot prove a priori that these ideas have actual instances – Kant famously rejects the attempts of his scholastic/rationalist predecessors to use speculative considerations to prove that, for instance, the First Cause of the world, or the Most Real Being, or the free will, or the bounded worldwhole, or the immaterial soul exist. Although we can’t demonstrate that there are objects of these rational ideas via either pure reason or experience, Kant notes that we have a strong natural and rational propensity to generate these ideas and then just presume that they have actual objects. Reason has the drive to go to a maximum, to have a complete explanation, to nd rest in a systematic whole, to locate “the therefore to every wherefore” (zu allem Warum das Darum) (A 585/B 613). This is one of Kant’s great innovations: using a kind of erotic apostrophe, he ascribes

8 See Ginsborg, Hannah. 2006. “Aesthetic Judgment and Perceptual Normativity,” in: Inquiry 49 (5), 403–37; Zuckert, Rachel. 2007. Kant on Beauty and Biology: An Interpretation of the “Critique of Judgment”. Cambridge: Cambridge University Press; Ostaric, Lara. 2009. “Kant’s Account of Nature’s Systematicity and the Unity of Theoretical and Practical Reason,” in: Inquiry 52 (2), 155–78; and the Horstmann papers discussed below.

Ogilby, Milton, Canary Wine, and the Red Scorpion

267

to reason itself various “needs,” “inclinations,” “drives,” “propensities” that ultimately lead us to accept uncritical theory-building principles like the rationalist’s Principle of Su cient Reason. In this way Kant rejects (or at least considerably complicates) the Platonic/Christian view of reason as the faculty that whips the horses of the passions into shape and keeps the chariot of the self moving in the right direction, as well as the Humean view of reason as the slave of the passions, meekly brought along to nd rationalizations for and means to the ends set by the non-rational passions that truly rule us. Rather, reason’s own inherent and natural inclinations towards fully articulate explanations are o en what get us into speculative trouble. When faced with these natural but illegitimate aspirations of reason, Kant says that we must take up our Critiques and resist: “Our age is an age of critique, and to critique all must be subjected” (KrV A xi). But since it is reason itself that generates these ideas, such critique is also an exercise in self-morti cation, and through it one can be spared many di cult and nevertheless fruitless e orts, since [we] would not attribute to reason anything which obviously exceeds its capacity, but would rather subject reason, which does not gladly su er constraint in the paroxysms of its lust for speculative expansion, to the discipline of abstinence (A 786/B 814).

Thus the second half of the rst Critique is an extended meditation on the various ways in which reason seduces us into the forbidden realm of things-in-themselves, o en via the Principle of Su cient Reason, and a therapeutic attempt to convince us that these illicit inclinations in reason must be identi ed and, as far as possible, extirpated or suppressed. Despite the fact that we must discipline ourselves to the fact that we cannot have theoretical cognition (Erkenntnis) or knowledge (Wissen) that there are objects of any rational ideas we still naturally nd these ideas fascinating and important. Many of them have heuristic or pedagogical uses, others provide moral ideals that we try to approximate, and some provide the content of the Belief (Glaube) for which Kant says he has to deny knowledge in the rst place. So we’re certain to be fascinated by any attempts to represent these ideas or even any hints that their objects are really possible. (2) This brings us to our second piece of terminology: An aesthetic idea is characterized by Kant as a phenomenon to which “no determinate thought, i.e., concept, can be adequate, so that no language can fully attain to it or make it understandable” (5:314). The reason that an aesthetic idea cannot be captured by a determinate concept is exactly because it is the sensible expression or “counterpart (pendant) of a rational idea.” An aesthetic idea deserves the name “idea” just insofar as it is a sensible representation which “strives toward something that lies

268

Chignell

beyond the bounds of experience, and hence seeks to approximate an exhibition of rational ideas” (ibid.). Although Kant sometimes calls it a “representation” (Vorstellung) in the singular, I have argued in other work that an important and indeed essential aspect of an aesthetic idea is that it involves a plurality of representations linked together by a certain theme.⁹ Kant seems to a rm this when he remarks that the “supplementary representations of the imagination […] which let one think more than one can express in a concept determined by words […] yield (geben) an aesthetic idea” (5:315). This suggests that, on the subject side of the equation anyway, an aesthetic idea is constituted by an “inexhaustible” and “non-exponible” series or multitude of representations. The series is uni ed by a certain theme, perhaps, but is too rich to be fully comprehended (5:342). Put another way, an aesthetic idea is a “coherent whole,” but a “coherent whole of an unspeakable fullness of thought” (5:329). It is this un nalizable series of associations, rather than any particular state, that “aspires” to exhibit a rational idea – an idea that, strictly speaking, cannot be sensibly exhibited at all. A psychological corollary of this view about the structure of an aesthetic idea is that the mental episode of having or undergoing such an idea will have certain essential features. Someone having an aesthetic idea will experience a “quickening” of her cognitive faculties, for instance, as her associative imagination brings to mind this “wealth of sensations and supplementary representations for which no expression is found” (5:316). This quickening, I submit, is just another way of characterizing the “animated feeling” or “pleasure” that accompanies or constitutes the harmonious free play of the faculties discussed earlier: the imagination runs through a series of representations that are associated with the object somehow, a series that yet seems so inexhaustible and un nalizeable as to elude capture by determinate concepts. The main point for our purposes, however, is that it is not the content of these representations that is of primary importance, but rather the formal manner in which they are strung together by the mind into a “coherent whole” that has the phenomenological feel of both unity and inexhaustibility. It is on the basis of having an experience with this formal structure (and that is thus accompanied by a feeling of pleasure) that we can judge the object that occasioned it to be beautiful.¹⁰

9 Again, see Chignell “Kant Normativity of Taste” for more details. 10 Hannah Ginsborg asks, in discussion, why it is that we wouldn’t just stop the account of the Deduction at a claim about the harmonious free play of the faculties itself. In other words, why not just say that any object that can lead to that mental phenomenon will count as beautiful for Kant? I agree that we might leave the account there, but nd an advantage in the present interpretation’s ability to further explain (though still only partially) what in the object generates the relevant

Ogilby, Milton, Canary Wine, and the Red Scorpion

269

(3) A nal piece of terminology: Kant distinguishes in § 49 between two di erent kinds of attributes that objects can have – logical attributes and aesthetic attributes. The logical attributes of a beautiful object are what we commonly think of as its attributes – being a painting, being a meadow, representing an Italian woman or an Egyptian queen, depicting the Roman god Jupiter. Aesthetic attributes, by contrast, “accompany the logical ones,” says Kant, and yet perform a distinct function (5:315–6). Of the Jupiter painting, for example, he says that the content of the painting possesses logical attributes such as “having long, owing hair” or “sitting atop a large throne.” Under the right circumstances, our mind does not just present these logical attributes of the object of a beautiful work of art, however: it also “calls to mind” a series of “supplementary representations […] expressing the concept’s implications (Folgen) and its a nity (Verwandtscha ) with other concepts.” These representations in the mind are the aesthetic attributes of the object. The main example that Kant o ers in connection with Jupiter is that of an “eagle with lightning in its claws” – an image traditionally associated with Jupiter in Roman mythology. Presumably, however, the imagination, together with re ective judgment, also conjures up countless other “related representations” that it associates, in some loose and free fashion, with the rational idea of God and the “sublimity and majesty of creation” that Jupiter symbolizes. Because Jupiter is an object whose concept is a rational idea – i.e., the rational idea of a creative deity – it will possess a richness such that the set of aesthetic attributes which “animate it” by way of these mental associations will seem inexhaustible to the properly-situated subject (5:315–6). Another example that Kant provides in the Critique itself comes from contemporaneous literature. Kant didn’t encounter much visual art, having never le the area around Königsberg, and presumably didn’t hear music very o en (apart from the prisoners singing hymns in the castle near his house). But he certainly read a great deal, and he explicitly claims that poetry or letters is the best art form for expressing rational ideas. The literary example he provides is a poem composed not by Milton (though Kant is far more in uenced by Milton than the average Prussian philosopher, as a recent book by Sanford Budick makes clear),¹¹ but rather by Friedrich der Große. In the poem, which was written in French, the proper attitude toward death is compared to the resigned and digni ed passing of the sun over the horizon at sunset:

feeling of “animation.” I will suggest below that it is because an object symbolically presents a rational idea (in a particularly e ective way) that it leads to the generation of an aesthetic idea, and thus to the characteristic feeling of mental harmony that Kant calls “free play”. 11 See Budick, Sanford. 2010. Kant and Milton. Cambridge, MA: Harvard University Press.

270

Chignell

Oui, nissons sans trouble, et mourons sans regrets, En laissant l’Univers comblé de nos bienfaits. Ainsi l’Astre du jour, au bout de sa carrière, Répand sur l’horizon une douce lumière, Et les derniers rayons qu’il darde dans les airs Sont ses derniers soupirs qu’il donne à l’Univers. (Quoted by Kant at 5:315–6)

Let’s set possible political motives aside and take Kant at his word when he says that the poem is beautiful and thus demands a positive judgment from all properly disinterested readers. Why does he say this? Clearly there are formal and material attributes of the object (subject-matter) of the poem: being a sunset is the major one; being compared to death, and being characterized as having gentle light are some others. But the imagination of the disinterested reader, in considering these attributes and in “remembering all the pleasures of a completed beautiful summer day,” will also (says Kant) conjure up a rich series of thoughts which it associates with its theme – the rational idea of “cosmopolitan virtue.” The series of representations brought to mind by the free play of the faculties in the contemplation of such an object will not exhaust its content (there is always more to say about great art!), but each thought in that series will, “ to be sure, pertain to the concept of the object” (5:315). And so the reader , in re ecting on this whole experience, will connect the having of the aesthetic idea to the object itself. Now it is always possible in principle for the imagination to call to mind a string of associations in connection with experience of any object and any subjectmatter. But Kant thinks that such an attempt with respect to a non-beautiful object will be neither easy nor pleasurable. Contemplation of such things “leaves nothing behind as an idea and makes the spirit dull, the object gradually disgusting, and the mind dissatis ed with itself and moody because it is conscious that in reason’s judgment its disposition is contrapurposive”. Signi cant for our discussion is that he concludes this comment by saying that “if the beautiful arts are not combined, whether closely or remotely with […] ideas – which alone carry with them a self-su cient satisfaction – then the latter is their ultimate fate” (5:326). This may or may not provide the basis for an account of the ugly, but it certainly suggests that Kant wants to leave room for the aesthetically neutral. The account just sketched lies at the heart of Kant’s solution to Hume’s problem: aesthetic ideas – with their exhilarating, provocative, and yet pleasurable phenomenology – will only be reliably occasioned in us by objects that “sensibly render” or, as Kant puts it elsewhere, “symbolize” a rational idea. As a result, “taste is basically a faculty for judging (Beurteilungsvermögen) the sensible rendering (Versinnlichung) of (rational) ideas” (5:356). Or, in the somewhat overstated phrase of a lecture, “the entire utility of the beautiful arts is that they set

Ogilby, Milton, Canary Wine, and the Red Scorpion

271

[…] propositions of reason in their full glory and powerfully support them” (25:33). This also hints at an explanation of why we keep coming back to particular works: not for the sensible renderings of rational ideas per se, but rather for the pleasure they occasion in us by sensibly rendering the ideas in the particular way that they do.¹²

4 The Lingering Ogilby Problem By now it should be obvious why Ogilby presents such a problem for this reading of Kant. I have been arguing that the way to make sense of Kant’s claims about the universality of our aesthetic judgments while retaining a subjective basis of these judgments is to focus on the features of particular art-objects that can be regarded as symbolically exhibiting a rational idea. But Ogilby’s poetry, too, (not to mention Bunyan’s allegories, which Hume also denigrates from an aesthetic point of view) can be regarded as taking various moral virtues or mathematical and dynamical “maxima,” supersensibilia, and so forth as its themes. Thus the account so far cannot rule out Ogilby’s or Bunyan’s work as a plausible occasion for the production of aesthetic ideas. Indeed, reference to rational ideas is pretty easy to come by, and given a suitably powerful capacity for associative imagination it seems that almost anything could provoke the rich series of seeming endless representations that constitutes an aesthetic idea and involves the free play of the faculties. So how, if at all, are we able justi ably to judge (with a universal voice) that Milton’s sonnets and Friedrich’s stanzas are beautiful, while Ogilby’s doggerel is not? Apart from simply conceding that anything that can be in some way connected with rational ideas is beautiful, there seem to be two main options here. First, we might say that only some of the objects that symbolically exhibit rational ideas are productive of genuine aesthetic ideas in properly situated beholders. The symbolic exhibition of rational ideas thus becomes something like a necessary but not a su cient condition of beauty on this view, and we have to say there is something more to the successful art-object – the way it symbolizes ideas – that makes it a suitable object of positive aesthetic judgments.

12 But why doesn’t Kant talk about the role of aesthetic ideas in judgments of beauty early on in the Analytic? My guess is that he only realizes that he needs to bring together his discussions of the beauty of nature and of the role of aesthetic ideas later on when he is thinking through the argument of the Deduction. Thanks to Franz Knappik for discussion here.

272

Chignell

Alternatively, we could say that only the objects that occasion aesthetic pleasure in properly-situated beholders succeed in truly or fully symbolizing rational ideas. There might be other ways in which objects refer or allude to a rational idea, but only beautiful objects succeed in genuinely symbolizing or symbolically exhibiting such a topic or theme (correlatively, a genius is an artist who is able to produce or perform such exhibitions). The distinction between these two alternatives may be largely verbal. Both say that a successful work will symbolically exhibit a rational idea by way of its content and/or form, and thereby occasion an aesthetic idea and the concomitant pleasure in properly-situated beholders. Both agree that a rational idea can be in some way associated with or referred to by an object, even if that object is not aesthetically successful. And both agree that this is an idealized sort of criterion that tells us what a successful work will do, even though (as Kant o en points out) we may not always have the ability to tell, in a given case, whether we are suitably disposed and su ciently disinterested to apply this criterion e ectively. The alternatives di er merely over whether unsuccessful works like (say) Ogilby’s should count as genuinely symbolizing rational ideas or not. But since it seems that Kantian “symbolization” is something that admits of rules or formulae (as Kant himself says when telling us what an analogical “symbol” is in § 59), perhaps the rst alternative is preferable – the one according to which lots of objects manage to symbolize rational ideas in their content or their form, but only those that do so in a particularly e ective way are proper occasions for a positive judgment of taste. The proposed solution to the Ogilby problem also ensures that it is not easy to create successful art: one cannot just at-footedly symbolize the content or structure of a rational idea and trust that it is the adequate basis for the production of aesthetic ideas. Rather, the speci c way that a great artwork ts a certain content to a particular form becomes an essential ingredient of aesthetic success. That said, perhaps we can allow that beauty is partial or comes in degrees: by focusing intently on almost any object we may be able to make some associations to rational ideas, and then get some short-lived free play going in such a way as to cause a few murmurs of aesthetic pleasure. (Even Ogilby had his fans.) But paradigmatic beauties will powerfully induce us to such activity as a result of the unique and compelling “way” that their content/form symbolically exhibits rational ideas. Note that this suggests that there is room for degrees of beauty in the account, or at least for judgments of degrees of beauty. Putting all of this together, we now have the resources to formulate a Kantian account of artistic beauty:

Ogilby, Milton, Canary Wine, and the Red Scorpion

273

An artwork is beautiful, for Kant, if its content and/or form symbolizes a rational idea in such a way that it occasions in properly-situated subjects an aesthetic idea (the having of which involves the “free” and “harmonious” play of the faculties, and is thus uniquely pleasurable).

The “way” in which a particular artwork symbolizes a rational idea will be exceedingly complex and di cult to analyze, and so the account remains inarticulate and uninformative at just this juncture. But this inarticulacy is salutary, I think, since any simple rules or formulae regarding how rational ideas must be symbolized in successful art would be inconsistent with Kant’s overall opposition to a science of beauty, and his endorsement of the “singular” or rst-person character of aesthetic judgments. Indeed, the artist herself will o en be unable to work out or even be aware of all the complex ways in which her work symbolizes rational ideas, and we may as a result be willing to credit her success to a kind of “spirit” (Geist) working through her rather than to a self-conscious and calculated e ort.¹³ An interesting implication of this proposal is that a Kantian critic will want to focus his attentions not merely on the rational ideas symbolized in the work, but on the complex “ways” in which the artist has achieved this symbolic exhibition in a particular physical or literary medium, and on the new aesthetic ideas that these ways occasions in him. With respect to the very greatest of works, however, even the most acute understanding and the most penetrating critic will nd the e ort to pluck out its mystery frustrated. That mystery will consist in the work’s unique way of symbolically exhibiting or picturing rational ideas – a content that can’t really be sensibly exhibited at all. It is part of the unanalyzable je ne sais quoi of a great work – the aspect that, again, makes a science of the beautiful based on explicit principles or rules impossible.

5 Systematicity and the Deduction of Taste It would be worth lling out the general picture I am sketching here and comparing it to some of the other broadly expressionistic interpretations of Kant. It would also be worth saying more about the role of genius or spirit in the production of beautiful art, and about the role of the Kantian art critic. Here, however, I want to focus on how judgments of beauty in nature t into this expressionistic account. This is important because Kant famously says in the third Critique that beauty in

13 It should be clear, then, that the view I am sketching here provides an idealized account of the normativity of a true judgment of beauty, rather than a criterion that allows us, in a particular cases, to know who is getting it right, and who is getting it wrong.

274

Chignell

nature as well as beauty in art has to do with aesthetic ideas (5:320). It will prove useful to do this in the context of discussing Rolf-Peter Horstmann’s illuminating work on the deduction of the principle of purposiveness generally. Horstmann has two main pieces on the issue: one in a volume on transcendental deductions edited by Eckart Förster and commented on by Reinhard Brandt, the second in a forthcoming volume.¹⁴ Again, the main focus of both is not the deduction of taste in particular but the deduction of the overarching principle of purposiveness. According to Horstmann, the purposiveness that the natural world presents consists (at least in part) in its susceptibility to being known by us, especially in its contingent aspects – the empirical particulars that fall under empirical laws. Horstmann thinks of this as a claim about the “material of sensation” rather than about the things-in-themselves: the claim is that this material could have been so chaotic or recalcitrant to our understanding of it that, even if we were able to categorize it in some very general way, we wouldn’t have been able to bring it under speci c empirical laws or concepts. Fortunately for us, however, it isn’t recalcitrant in that way: there is a systematic purposiveness in empirical things that allows us to know them, to predict events reliably, to study natural objects and systems, to construct scienti c theories, and to live in the natural world without fear of unlawful chaos. Horstmann claims that in the third Critique this principle of purposiveness becomes, for Kant, a necessary condition “for the unity of knowledge in view of the multitude of empirical laws” (Horstmann 1989, 169). Indeed, Kant’s recognition of its importance leads him to assign it a new status: it is no longer a merely “logical” maxim, as it was in the rst Critique, but rather a new kind of “subjective transcendental principle,” one that Horstmann is willing to call “constitutive” in his second paper on the topic. The content of the principle, again, is that we encounter in experience a systematic natural world: a world whose features can be known not just a priori but also empirically because it is structured in terms of part and wholes, kinds, empirical laws, a nities, and hierarchies.

14 Horstmann, Rolf-Peter. 1989. “Why must there be a Transcendental Deduction in the Critique of Judgment?” in: E. Förster (ed.). Kant’s Transcendental Deductions. The Three Critiques and the Opus postumum. Stanford, CA: Stanford University Press; and Horstmann, Rolf-Peter. forthcoming 2013. “Kant and the Problem of Purposiveness, or How to deal with Organisms (and Empirical Laws and Beauty) in an Idealistic Framework” (this essay is forthcoming in a volume in honor of Paul Guyer’s 60th birthday). Here I cite from a dra , and so the page numbers will presumably be di erent in the published version.

Ogilby, Milton, Canary Wine, and the Red Scorpion

275

But what is the connection between all this and the critique of taste? Horstmann argues that aesthetic experience of beauty simply is a pleasing sense of the present sensory manifold as especially cognitively purposive: It is at the very center of [Kant’s] argument that if, when we contemplate an object, all conditions are ful lled that would have to be ful lled in order to put us in the position of acquiring a concept of that object, then this object produces in us the feeling of pleasure. This feeling in turn indicates that the object is purposive for our faculty of re ective judgment, and it is because of this that we call the object beautiful. Now these conditions of concept acquisition presuppose the transcendental principle of purposiveness, and this implies that the very possibility of an aesthetic judgment of re ection is based on that principle (ibid., 174).

A worry we might have about this way of construing Kant’s point is that the apparent purposiveness of the world looks like it will be su cient on its own for a positive judgment of taste. But if the transcendental principle of purposiveness is applied universally, as Horstmann indicates that it is, then we are threatened with an account on which everything becomes beautiful, or at least capable of being the object of positive aesthetic judgment, in much too easy a way. This makes it puzzling that Horstmann goes on to follow Guyer in holding that there is a gap between an aesthetic judgment of this sort, grounded in the principle of purposiveness, and the claim that it is intersubjectively valid (ibid., 173). For presumably the very deduction of that principle that Horstmann discusses is, like all Kantian deductions, supposed to be intersubjectively valid, and so it is hard to see where the gap would be with respect to any subject and any experience of an object. Horstmann appears to recognize this problem in the second paper, and simply leaves it as an “interesting question” whether there can be “aesthetically neutral objects […] or (whether) every possible object of knowledge has an aesthetic value” (ibid., 17). Setting that issue aside, however, I now want to sketch a slightly di erent account of how the connection between natural systematicity and a judgment of taste might be made. The account here is implicit in something else that Horstmann says in the more recent paper: that beauty in nature “points to or hints at a purposefully organized nature” (ibid., 16). His own explanation of this is, once again, a direct and transcendental one: beauty could not be “explained as belonging to the objective elements of constituted nature without relying on the notion of purposiveness” (ibid.). I want to suggest, by contrast, that beauty in nature and art “points to or hints at a purposefully organized nature” in a less direct and more roundabout way – namely, by symbolically exhibiting the rational idea of it. Just as an art object that symbolizes the idea of a mathematical maximum or a creative deity in a partic-

276

Chignell

ularly rich way induces in us an aesthetic idea that in turn provides the basis for our judgment that it is beautiful, so too an object in nature can symbolize the idea of natural systematicity and in that way produce in us the right kind of aesthetic idea. Before describing how this symbolization might go, let me note rst a wellknown passage in which I take Kant to be making precisely this point while discussing natural beauty in particular: Although our concept of a subjective purposiveness of nature in its forms, in accordance with empirical laws, is not a concept of the object at all, but only a principle of the power of judgement for providing concepts in the face of this excessive multiplicity in nature (in order to be able to be oriented in it), we nevertheless hereby ascribe to nature as it were (gleichsam) a regard to our faculty of cognition, in accordance with the analogy of an end; and thus we can view natural beauty as the exhibition (Darstellung) of the concept of formal (merely subjective) purposiveness and natural ends as the exhibition of the concept of a real (objective) purposiveness, one of which we judge through taste (aesthetically, by means of the feeling of pleasure), the other through understanding and reason (logically, in accordance with concepts) (5:193).

The rst part of this passage says that natural systematicity isn’t something that we rst run across in the world and then somehow “logically” grasp with concepts. Instead we presuppose in an a priori albeit subjectively justi ed way that the world is systematically ordered under hierarchies of laws such that it has, in Kant’s words, “a regard to our faculty of cognition.” In the unpublished Introduction, Kant suggests that the presupposition is downright required for us rationally to engage in scienti c inquiry, and even to form any empirical concepts whatsoever.¹⁵ For present purposes, the more interesting part of the passage just quoted from 5:193 comes next: “natural beauty” counts as an “exhibition of the concept of formal (merely subjective) purposiveness […] which we judge through taste (aesthetically, by means of the feeling of pleasure).” Formal subjective purposiveness here refers to the structure, exhibited by systems, that allows them to be cognized and comprehended by our minds and methods. Kant is thus saying that there is something in natural beauty that “exhibits” this structure somehow. But how? Later in the Critique he explains that To exhibit (dartun) the reality of our concepts, intuitions are always required. If they are empirical concepts, then the latter are called examples. If they are pure concepts of the understanding, then the latter are called schemata (5:351).

15 Cf. 20:203.

Ogilby, Milton, Canary Wine, and the Red Scorpion

277

So some sort of intuitional content is required for exhibition. Kant goes on to say, however, that transcendental ideas cannot have their reality exhibited in either of these ways, but we can at least exhibit some of them symbolically: All hypotyposis (presentation, subjecto sub adspectum), as making something sensible, is of one of two kinds: either schematic, where to a concept grasped by the understanding the corresponding intuition is given a priori; or symbolic, where to a concept which only reason can think, and to which no sensible intuition can be adequate, an intuition is attributed with which the power of judgement proceeds in a way merely analogous to that which it observes in schematization (ibid.).

Kant is asserting here that non-empirical concepts may acquire intuitional content either through the process of schematism or through symbolization. Ideas, unlike categories, can’t be schematized, and so if they are to have any positive content at all it must be symbolic content. Return now to the quotation from 5:193: many commentators, Horstmann included, assume that the exhibition of formal subjective purposiveness here goes by way of a straightforward “example”: we nd ourselves confronting, in an experience of a beautiful object, something that seems especially suited to our understanding. But this is pretty vague, and again it leaves us with the question about whether there can even be aesthetically neutral objects, since we are already in the context of the transcendental presupposition that everything we encounter in nature is suited for our understanding. My own suggestion is that Kant thinks of natural beauty as providing a kind of content to the idea of subjective purposiveness – itself a component of the complex idea of natural systematicity – via symbolic exhibition rather than straightforward example. Natural structures that seem complete in their internal relations, whose parts bear various organic value relations to the whole, and in which the diversity of forms and colors are brought together into a discrete articulate object, present us with an exhibition by analogy of the rational idea of full and complete systematicity. This in turn leads, in the ways described earlier, to the production of an aesthetic idea in us, and ultimately to the pleasure that is the basis of a judgment of taste. Let me emphasize again that I’m not saying that certain bits of nature exhibit natural systematicity, full-stop, and that this is the basis of their aesthetic appeal. I’m saying rather that objects in nature (and art) – in particular those with the kinds of features that the rationalist tradition in aesthetics focused on – can symbolize (rather than provide an example of) the relations of unity and harmony amid diversity that is characteristic of systematicity. I am also not claiming that this is what Kant is arguing for in § 59 of the third Critique, according to which “beauty is a symbol of morality.” The point there has to do with analogies

278

Chignell

between the way our faculties operate in aesthetic contexts and the way they operate in moral contexts. Rather, my claim here is that the qualities of unity amidst diversity, internal relations that seem somehow complete, and perfect part-whole interconnections allow an object to symbolically serve as a kind of analogical exhibition of the rational idea of natural systematicity, even apart from any connection to morality.¹⁶ In order to make this clearer, let me try to give an example from the world of art. I’m wary in the context of an aesthetics paper of trying to do justice to wellknown works. So since this volume is in honor of Rolf-Peter Horstmann, I’ll take the example of a more obscure artistic creation that he has clearly relished over the years in the context of our post-colloquium discussions – something called “The Red Scorpion.” A quick search online reveals a couple of items that go by the title “The Red Scorpion.” Actually on google.de the rst thing that comes up is a strip tease artist who, for a certain price, will come to your party dressed in a very elaborate and very red scorpion costume – described on the website as “exotic” – and proceed to remove the costume to the thrill of your assembled guests. Fortunately, however, this is not the Red Scorpion of which Professor Horstmann is a devotee. The other main item that the internet delivers, ttingly enough in the context of a discussion of the third Critique, is a biological organism – namely, a scorpion with a red spot on its back. According to the information I could nd, this kind of scorpion comes out only at night, and is one of the most dangerous predators in the desert ecosystem. People who know Kant’s works extremely well might know that Kant himself talks about scorpions in his lectures on physical geography, and I think we can assume for the sake of argument here that they were the red kind: The scorpion is, in Italy, no bigger than a little nger, has a crab-like shape and wounds its enemy with its tail, which contains a hook. One has to avail oneself of the crushed scorpion, in order to place it on the wound and pull the poison back out. In emergencies involving such a poisonous bite, the Indians further take to burning the place that was bitten. In India they are much bigger. It is said that when one places a scorpion under a glass and then blows tobacco smoke under it, it will kill itself with its own tail (9:352).

But obviously this has nothing to do with aesthetics, and Kant’s report about the tobacco smoke would be an un tting theme in a paper devoted to Horstmann,

16 See Chignell, Andrew. 2006. “Beauty as a Symbol of Natural Systematicity,” in: British Journal of Aesthetics 46; and Chignell, Andrew. 2010. “Real Repugnance and Belief about Things-inThemselves,” in: Kant’s Moral Metaphysics, B. Lipscomb and J. Krueger (eds.), Berlin: de Gruyter, 177–209; for more expansive versions of this argument and of the claims about symbolization generally.

Ogilby, Milton, Canary Wine, and the Red Scorpion

279

from whom many of us have received in gratitude the occasional tobacco item of a summer’s evening in Berlin. Somewhat closer to aesthetics is Kant’s intriguing remark, earlier in the same lecture material, that someone who is bitten by tarantulas will alternately cry, laugh, dance, and be sad. Such a person cannot tolerate [the colors] black or blue. One cures him through music, primarily of the cither, oboe, trumpet, and violin, via which, when one achieves the right tone and the most tting melody, he is brought to dancing, sweating, and ultimately to health. Those who are stung by scorpions also love music, chie y the sackpipe and drums (9:350, my emphasis).

The connection Kant draws here between the purpose of curing someone who has been bitten by a scorpion and the purposive aesthetic qualities of danceable sackpipe music is, to say the least, somewhat strained. Fortunately, this too is not the kind of Red Scorpion we are concerned with here, although the biological kind may be the ultimate origin of the term. The Red Scorpion I’m referring to, rather, is a product of the cocktail artist at a bistro called Via Nova, just a block away from our colloquium’s meeting place at the Humboldt Universität zu Berlin. Like its namesake, this Red Scorpion tends to be seen only at night, and it has a bright red color – a result of an artful mixture of brown rum, white rum, maracuja syrup, triple sec, lemon juice and orange juice.¹⁷ In place of a stinging tail, the Red Scorpion is served with one of those wooden toothpick umbrellas that can, if you’re not careful while sipping from the glass, e ect a minor injury to the face or eye. Now Kant himself would not have approved of the thought that a drink could be aesthetically appealing: he cites “Canary wine” as something that is entirely agreeable but not pleasurable in the aesthetic way, even if it comes from Teneri e. But let’s suppose, in honor of Rolf, that a cocktail like the Red Scorpion really is a kind of artwork … the way it appeals to the eye with a scarlet that is deeper than cinnabar, the way the bright umbrella seems rise, e ortlessly, from the carefully arrayed Eiswürfel (so hard to nd in Germany!), the way its coolness and he feels in the hand as one brings it to the lips, the internal relations between the taste of rums and marajuca and orange and lemon creating a whole that is much more than the sum of its parts. It is this kind of articulated and yet somehow sealed-o and complete structure in the Red Scorpion that allows it perfectly to complement

17 Citation for this recipe is from the Via Nova menu, Summer 2011 edition, which lists their cocktails alphabetically. Red Scorpion has an auspicious place just a er a drink called “Melon Kiss” and just before the familiar drink called “Sex on the Beach” and the somewhat less familiar variation “Sex on the Swimming Pool.”

280

Chignell

a slice of Via Nova’s Pizza Napoli and provide us with grounds to abandon our Stammkneipe of many years – the trusty Deponie on Georgenstraße – for the sake of Via Nova, despite their awful table service. It is this kind of (apparently!) effortless combination of aesthetic qualities, indicative of real genius on the part of the bartender, that makes it reasonable for the devotee of the Red Scorpion (a er urging everyone else in the colloquium to try it and encountering responses ranging from feigned appreciation to neutrality to open repulsion from those who did) to follow the example of Kant’s young poet in the third Critique and maintain that only he is engaging in a truly disinterested and properly aesthetic evaluation of it (5:282–3). But, getting back to our serious inquiry here, it is also precisely these qualities of unity amidst diversity, internal relations that seem somehow complete, and perfect part-whole interconnections that allow it symbolically to serve as an analogical exhibition of the rational idea of systematicity. And this would be true not only of a brilliant Gesamtkunstwerk like the Red Scorpion, but also of natural objects that displayed these kinds of properties.¹⁸ Signi cantly, then, formal relations as well as form-content relations in the objects – either natural or arti cial – can symbolically exhibit a key rational idea: that of systematicity in nature. When they do so in a particularly rich or e ective way, we are led, via the richness of this ideational content, naturally to re ect in a kind of associative fashion on what is being presented, to call up the object’s various aesthetic attributes and other associations, and ultimately to undergo the experience of an aesthetic idea. The pleasure in the subjective form of this experience is essential to the aesthetic judgment, but the connection to a particular object goes by way of the latter’s ability to symbolize –in a particularly rich and e ective way – the object of a rational idea, and thus to cause an aesthetic idea in us. An objector might worry that Kant’s theory of taste on this interpretation threatens to be too narrow – it ascribes beauty only to those objects, works, or vistas that can be associated somehow with rational ideas. Like Plato on a common reading of the Symposium, Kant on the present interpretation is so xated on the ideas of reason that he denigrates important this-worldly aspects of art and nature, aspects which clearly contribute to their aesthetic success. But note, as an initial response to this worry, that although this would be a problem for readings that insist that an appeal to moral ideas is the only way to ll the gap in the Deduction, Kant’s own appeal to the example of Friedrich’s poem

18 For example, a proportionate and well-formed egg (which, it is worth noting, has traditionally been thought preferable to a scorpion of any sort – see, e.g., the rhetorical question at Luke 11:12: “Or if he asks you for an egg, would you give him a scorpion?”).

Ogilby, Milton, Canary Wine, and the Red Scorpion

281

highlights the fact that the domain of ideas, for him, is quite vast. It is not that all great art points narrowly to the One or the Good; rather, there is a huge array of cosmological, metaphysical, mathematical, and (yes) moral ideas that can be symbolized in art – even in purely formalistic art, it turns out – and all such symbolizations (when they are wrought in a certain “way”) can serve as the occasion for aesthetic response. By way of further response, however, it is worth admitting that there is an unmistakable Platonic avor to the theory. Kant is suggesting that one of the main goals of art- and nature-appreciation is to help us catch sight of the transcendent objects of ideas. His language is that of both aesthetic appreciation and Platonic eros when he asks, in the rst Critique: Why has Providence set many objects, although they are intimately connected with our highest interest, so high that it is barely granted to us to encounter them in an indistinct perception, doubted even by ourselves, through which our searching glance is more enticed than satis ed?” (A 743–4/B 771–2).

According to the picture presented here, one answer would be: So that we would make beautiful things, and learn to appreciate beautiful nature. That kind of beauty entices us by giving us symbols – indistinct perceptions, doubted even by ourselves – of rational ideas. The account as I’ve begun to outline it here meets the conditions set out by the Analytic: aesthetic pleasure arises out of the form and feeling of the subjective experience and is not based directly in any intellectual or empirical interests. It just happens that, for Kant, rational ideas are (or o er) the only themes rich enough to evoke aesthetic ideas in us. So the metaphysical, mathematical, cosmological, or moral content of the artwork will be indirectly –Kant says “remotely” – connected to the judgment that the object is beautiful (5:326). However, this content itself – whatever other interests it may satisfy or engender – is not the direct basis for a judgment of taste. Rather, the disinterested pleasure involved in experiencing aesthetic ideas is. Thus we can say – a priori, as Kant puts it – that only those objects that occasion aesthetic ideas by way of symbolically exhibiting rational ideas in a particularly fruitful way will be beautiful for every beholder. The clause about the “way,” of course, is what allows us to rule out Ogilby on Kant’s behalf. We have also seen that this condition provides a sense of what sort of art critic a true Kantian would be. She would seek to be fastidiously disinterested herself, and would point out the possible prejudices and interests that lead people to make misleading judgments in a given context. The claim in this paper is that the ideal Kantian critic would also seek to draw our attention to the subtle, creative, and

282

Chignell

hard-to-describe ways in which ideas are symbolically expressed in successful artworks and beautiful nature, and thus to the complex manner in which metaphysical, moral, cosmological and mathematical content impinges on the domain of aesthetic taste.¹⁹

19 My gratitude to audiences at the Institute for Advanced Study, Princeton, and LudwigMaximilian-Universität, München, and to Bradley Murray, Rachel Zuckert, Paul Guyer, Omri Boehm, Joseph Cannon, and Franz Knappik for helpful discussions of these ideas (usually framed without the Red Scorpion example). Thanks, too, to Dina Emundts for organizing the conference at the Humboldt-Universität, Berlin, at which this paper was rst presented in something like the current form, and for editing the volume in which it now appears.

Paul Guyer

The End of Art and the Interpretation of Geist 1 Interpreting Geist For decades, the central problem of Hegel scholarship has been the interpretation of the concept of Geist. On the one side, there is the traditional position that by Geist Hegel means “cosmic spirit,” something larger than although including humankind, or at least the mental life of human beings. On the other side, there is what may be considered the revisionist position that by Geist Hegel means nothing other than human mental life or (as some now like to say) “mindedness.” We might call the traditional interpretation a metaphysical or transcendent interpretation of Geist, the more recent revisionist kind an immanent and humanistic interpretation. A leading proponent of the transcendent approach has been Charles Taylor, from whom comes the term “cosmic spirit” and who equates Geist with a God who has a “larger rational plan” for the world than that of human reason, although such “Geist or God cannot exist separately from the universe which he sustains and in which he manifests himself” (and from this we “can already see why Hegel had to su er accusations of Spinozism or pantheism”).¹ Rolf-Peter Horstmann has also taken a transcendent approach to Geist, or rather “reason,” which he regards as the central concept of Hegel’s philosophy: for him, “‘reason’ is not merely the name for a human faculty which contributes in a speci c manner to our gaining knowledge; he also uses ‘reason’ to describe that which is ultimately and eminently real,” a “primary structure,” also called the “Absolute,” “which forms the basis for all facts which are real in [a] speci c sense.”² The immanent and humanistic approach to Hegel’s conception of Geist goes back at least to Klaus Hartmann,³ but among its most recent proponents have certainly been Terry Pinkard and Robert Pippin. For Pinkard, Geist is “self-relation as ‘subjectivity’ […] or […] to be an agent,” and Hegel’s philosophy of Geist “aims at providing a comprehensive account […] of what it means to be an agent – or more

1 Charles Taylor, Hegel (Cambridge: Cambridge University Press, 1975), p. 87. 2 Rolf-Peter Horstmann, “Hegel,” in E. Craig, ed., Routledge Encyclopedia of Philosophy (London: Routledge, 1998), vol. 4, p. 264; cp. Dina Emundts and Rolf-Peter Horstmann, G. W. F. Hegel: Eine Einführung (Stuttgart: Reclam, 2002), p. 32. 3 Klaus Hartmann, “Hegel: A Non-Metaphysical View,” in A. MacIntyre, ed., Hegel: A Collection of Critical Essays (Garden City: Anchor Doubleday, 1972), pp. 101–24.

284

Guyer

generally, what it means to be human.” Even more speci cally, on his approach the philosophy of Geist concerns “a certain sociality of agency, of our being and becoming agents only in and through reciprocal recognition.”⁴ Pippin adds that Geist is “collective subjectivity,”⁵ though like Pinkard he also equates human subjectivity with human agency, and thereby with freedom: according to Pippin, for Hegel “the highest value or aspiration is freedom, […] a form of rational agency, the actualization of reason.”⁶ On these accounts, the philosophy of Geist is thus the account of the realization of collective human freedom and of the human selfunderstanding of such freedom, and Geist as collective human reason, agency, and freedom does not have to be understood as part of some larger cosmic spirit that includes but is not exhausted by human spirit. Hegel’s aesthetics, more precisely its central theme of the “end of art,” its claim that “art, considered in its highest vocation, is and remains for us a thing of the past” that “no longer a ords that satisfaction of spiritual needs which earlier ages and nations sought in it,”⁷ gives us a way to settle this dispute. For Hegel, as a form of “absolute knowing,” art is interpreted as a form of knowledge of Geist, but as essentially involving sensuous media, art is a speci cally human form of knowledge of Geist. Yet art is a form of knowledge of Geist that cannot have the same signi cance it had at earlier, less philosophically enlightened stages of human history just because it is a sensuous form of knowledge, and Geist is ultimately not cognizable by even partially sensuous means. I will argue that this is precisely because Geist is supposed to be entirely non-sensuous, purely intellectual, while human being, even human subjectivity, seems to be essentially sensuous as well as intellectual. The signi cance of art is thus for us something of the past, which must give way to religion and ultimately to philosophy, because Geist is not reducible to human subjectivity. Or more precisely, Hegel’s thesis of the end of art can be reconciled with the premise that Geist as the subject of art is human agency only on a conception of human agency that undermines the humanistic approach to Geist by e acing any normal distinction between the human and the divine. Pippin himself actually concedes this point when he describes

4 Terry Pinkard, “Symbolic, Classical, and Romantic Art,” in S. Houlgate, ed., Hegel and the Arts (Evanston: Northwestern University Press, 2007), pp. 3–28, at pp. 4–5, 7. 5 Robert B. Pippin, “What Was Abstract Art? (From the Point of View of Hegel),” in S. Houlgate, ed., Hegel and the Arts, pp. 244–70, at p. 249. 6 Robert B. Pippin, “The Absence of Aesthetics in Hegel’s Aesthetics,” in F. C. Beiser, ed., The Cambridge Companion to Hegel and Nineteenth-Century Philosophy (Cambridge: Cambridge University Press, 2008), pp. 394–418, at p. 397. 7 G. W. F. Hegel, Aesthetics: Lectures on Fine Art, transl. by T. M. Knox, 2 volumes (Oxford: Clarendon Press, 1975), vol. I, pp. 10–11.

The End of Art and the Interpretation of Geist

285

Hegel’s aesthetics as resting on a “quite radical humanism (or divinization of the human).”⁸ To be sure, the dispute between the immanent and transcendent interpretations of Geist cannot be settled simply by citing some decisive text that can only be interpreted in one way. Hegel famously says that ne art is only “truly art, and it only ful ls its supreme task when it has placed itself in the same sphere as religion and philosophy, and when it is simply one way of bringing to our minds and expressing the Divine, the deepest interests of mankind, and the most comprehensive truths of the spirit”⁹ – this is the premise that leads to the end of art thesis, because it turns out that although art is only truly art when it tries to do this, in the end it cannot successfully do it. The occurrence of the word “divine” here might seem to settle the interpretative debate in favor of the transcendent rather than immanent interpretation of Geist. But the humanists are able scholars, and do not ignore this well-known statement but instead interpret it in a way consistent with their approach. Thus in two articles Pippin interprets away the threat to the humanistic approach that the passage might appear to o er by claiming that Hegel follows the word “divine” (das Göttliche) “with a number of appositives and quali cations that strip it of much traditional religious association,”¹⁰ a “set of appositives [that] appears to gloss the divine as ‘the deepest interests of mankind and the most comprehensive truths of spirit’ rather than vice versa”¹¹ – though since the entire debate about “spirit” has been whether it means something strictly human or something more than just human, perhaps something divine as well as human, what he actually needs to claim is that the phrase “the deepest interests of mankind” is “appositive” for both “the divine” and “the most comprehensive truths of the spirit.” However, such an assertion is not an argument, and Pippin owes us an explanation for why the three terms of the quotation should be taken as appositives, the quotation thus saying that art concerns the divine and the most comprehensive truths of the spirit which equal the deepest interests of mankind, rather than as a conjunction, saying that art concerns the deepest interests of mankind plus the divine and the most comprehensive truths of the spirit. Such an argument might begin with the citation of other passages in which Hegel identi es the divine with the human, but passages in which he does not use exactly the same grammatical construction as in the present one, and which could therefore provide evidence for the construal of the present one.

8 Pippin, “What Was Abstract Art?”, p. 250. 9 Hegel, Aesthetics, vol. I, p. 7. 10 Pippin, “The Absence of Aesthetics,” p. 397. 11 Pippin, “What Was Abstract Art?”, p. 250.

286

Guyer

Such passages can be found. But then again, passages can also be found that appear to recognize the di erence between the human and the divine, and that suggest that the subject of art is the human relation to the divine rather than the identi cation of the divine with the human. So at that level of argumentation, we will quickly reach a stand-o . My proposal is that in order to resolve such a stand-o between interpretations of Hegel’s words we need to look more closely at the actual argument of the lectures on aesthetics, and that when we do we will realize that Hegel’s argument that art cannot have the signi cance for us that it had for earlier times and must be superseded by philosophy makes sense only because he assumes that art must provide sensuous representation of something more than the merely human, or more precisely the divine in the human, but that this cannot in the end be sensuously represented. It is only because art tries and must try to represent the divine as well as the human as the divine that once the divine and the human are properly understood art is ultimately doomed to failure. Hegel’s argument is that when the divine itself is understood in merely human terms, it can be adequately represented in art – that is the triumph of classical art in its paradigmatic medium, sculpture, but in every other medium as well. But when the divine, or the human as divine, is adequately understood, then it can no longer be adequately represented by art, even in its most dematerialized medium, namely poetry – that is the ultimate failure of both romantic art and its paradigmatic medium, poetry. The key to my argument will thus be that Hegel’s diagnoses of the ultimate limitations of romantic art and its paradigmatic medium turn precisely on the incapacity of poetry to represent the divine and the human as divine. While romantic art might seem to have an unlimited potential to represent the human, Hegel ultimately argues that this is so only on a conception of the human that fails to recognize its divinity, or the real nature of Geist. That we need to appeal to the denouement of Hegel’s diagnosis of the end of art rather than to his initial programmatic statements alone must be evident from the ambiguity of the latter, or the fact that it is easy for each approach to Geist to nd programmatic statements that seem to support it. Thus this passage seems to support the humanistic interpretation of Geist: [A] work of art is such only because, originating from the spirit, it now belongs to the territory of the spirit; it has received the baptism of the spiritual and sets forth only what has been formed in harmony with the spirit. Human interest, the spiritual value possessed by an event, an individual character, an action in its complexity and outcome, is grasped in the work of art and blazoned more purely and more transparently than is possible on the ground of other non-artistic things. Therefore the work of art stands higher than any natural product which has not made this journey through the spirit. For example, owing to the feeling

The End of Art and the Interpretation of Geist

287

and insight whereby a landscape has been represented in a painting, this work of the spirit acquires a higher rank than the mere natural landscape.¹²

This passage straightforwardly seems to say that art is in the business of representing human interests and actions, and that even when art super cially represents nature, as in a landscape painting, what it is really doing is expressing human “feeling and insight” about nature, which nature itself can never do. Here “spirit” or “spiritual value” seems to be equivalent to “human interest.” But Hegel concludes this passage with the statement that “Besides, no natural being is able, as art is, to present the divine Ideal,” which suggests that representing spirit in its human form, the form of human insight and action, is only the rst step toward representing spirit in all its forms, which include a divine form not reducible to human form. And then he continues with an argument that seems to make the latter point. He begins by rejecting the view that natural beauty is superior to artistic beauty because it is a product of God rather than mere man,¹³ but does not advocate that we should instead think of art simply as a product of man and more important than natural beauty for that reason. Rather, he argues that we must think of artistic beauty as the product of God working through man and as more important than mere natural beauty because although that too is a product of God, it is not a product of God working through man, and thus does not reveal that human spirit is only part of a spirit larger than itself: For nature and its products, it is said, are a work of God, created by his goodness and wisdom, whereas the art-product is a purely human work, made by human hands according to human insight. In this contrast between natural production as a divine creation and human activity as something merely nite there lies directly the misunderstanding that God does not work in and through men at all, but restricts the sphere of his activity to nature alone. This false opinion must be completely rejected if we are to penetrate to the true nature of art […] .For not only is there something divine in man, but it is active in him in a form appropriate to the being of God in a totally di erent and higher manner than it is in nature. God is spirit, and in man alone does the medium, through which the Divine passes, have the form of conscious and actively self-productive spirit […] Now in art-production God is just as operative as he is in the phenomena of nature; but the Divine, as it discloses itself in the work of art, has been generated out of the spirit, and thus has won a suitable thoroughfare for its existence, whereas just being there in the unconscious sensuous of nature is not a mode of appearance appropriate to the Divine.¹⁴

12 Hegel, Aesthetics, vol. I, p. 29. 13 Hegel names no name here, but we nd such a view in Moses Mendelssohn, “On the Main Principles of the Fine Arts and Sciences” (1757), in Mendelssohn, Philosophical Writings, ed. by D. O. Dahlstrom (Cambridge: Cambridge University Press, 1997), pp. 174–5. 14 Hegel, Aesthetics, vol. I, pp. 29–30.

288

Guyer

This argues that human art is more important than natural beauty because it is a product of spirit working through a sensuous medium rather than a property of the sensuous medium alone, and that art thereby reveals that human spirit is itself the product of the divine rather than a product of something merely natural (for example, the human body alone). The importance of art rather than natural beauty is that it reveals that the human is not merely human but also divine. The irony of the passage, however, is that art reveals this only to or through philosophical re ection on art, and that the nature of the divine and its relation to human spirit can thus not be fully expressed by art alone but only by philosophical reection on art. The foundation for Hegel’s theory that art must be superseded by philosophy is already laid here. But we would hardly expect the defenders of the humanistic approach to Geist to concede to the transcendent interpretation of Geist as soon as I have invoked Hegel’s thesis of the end of art and pointed to art’s need but ultimately incoherent e ort to represent the divine by sensuous means in order to explain it. And indeed they do not. Rather, they defend the humanistic approach by arguing that art is incapable of fully representing the character of human freedom. To prepare the way for my own argument, then, which is that Hegel conceives of art as capable of adequately representing everything about human being except its relation to the divine, or our own divinity, I must rst look at the humanists’ argument that what art is incapable of adequately representing is nothing other than human freedom itself. My conclusion will be that their argument depends on a conception of human freedom that is itself best considered transcendent, that is, itself interprets human freedom as something divine.

2 The Artistic Representation of Human Freedom Terry Pinkard’s view is that Hegel’s overall project is to show how humans fail or succeed in understanding their own agency, that art is one of the ways in which they attempt to do so, but that it must ultimately fail to be a successful way of understanding our own agency because it is essential to art that it retain a sense of mystery about our agency but essential to our self-understanding that we eliminate mystery. Thus, he says rst that “Hegel aims […] at providing a comprehensive account of the conditions under which forms of life fail and succeed by virtue of their having committed themselves to certain determinate views of what it means to be an agent.” In particular, the goal is to understand the collective agency of human beings: “The goal of Hegel’s theory is a comprehensive account of how certain collective expressions of normativity dissolve and fail because of

The End of Art and the Interpretation of Geist

289

their irrationality,”¹⁵ he says rst, but this is incomplete, because humans can ultimately succeed in becoming agents together through “reciprocal recognition.”¹⁶ Art is one of the forms of the “Idea’s self-realization,” that is, the “collective attempt by humanity to articulate for itself just what the ‘whole’ is within which it orients itself and interprets itself,” ultimately, to understand its collective agency and its potential. Thus “Art as Idea […] is one way of re ecting and assuring ourselves about the meaning of what it is to be human […] a beautiful sensuous presentation of what it would be like to be free.”¹⁷ The three stages of art history, or as Hegel to our ears misleadingly names them the three “forms” of art, are then interpreted as three stages in our attempt to present our own freedom to ourselves by sensuous means, all of which, however, are ultimately inadequate to this task. Symbolic art, epitomized by archaic architecture, is only capable of hinting at the “mythically conceived origin of human freedom, the divine ‘whole’ that brings about the existence and sustains the continued lives of the ‘minded’ creatures we are,” and moreover only through works that “do not link in any clear way to the human realities they are trying to portray.”¹⁸ Classical art, epitomized by the representation of the Olympian gods in human form in sculpture or other media, shows “what it would be like to be an agent with the capacity for free sense-making activities in such a way that the individual would be a law unto himself,”¹⁹ and thus fails to display the essentially collective character of human agency, that is, its dependence on reciprocal recognition. The Olympic Gods are constantly represented as competing with one another rather than as cooperating, and even when humans are represented, such as in the Antigone, such works “do not resolve the deeper con ict contained in forms of life in which agents take themselves to be free individuals who nonetheless must keep faith with laws […] and […] see these laws as expressing who they most deeply are.”²⁰ Finally, “Romantic art begins with the conviction that what we mean by our actions is not completely disclosed by what we do and that there is therefore an ‘inwardness’ which must be discovered or uncovered if we are to nd out who we really are”²¹ and interprets this inwardness as “in nite subjectivity,” but then progresses toward “increasing re ection, incorporation of theory into itself, and, in short, to greater abstraction” and “then there is no longer a

15 Pinkard, “Symbolic, Classical, and Romantic Art,” p. 5. 16 Pinkard, “Symbolic, Classical, and Romantic Art,” p. 7. 17 Pinkard, “Symbolic, Classical, and Romantic Art,” p. 9. 18 Pinkard, “Symbolic, Classical, and Romantic Art,” pp. 12–13. 19 Pinkard, “Symbolic, Classical, and Romantic Art,” p. 14. 20 Pinkard, “Symbolic, Classical, and Romantic Art,” p. 16. 21 Pinkard, “Symbolic, Classical, and Romantic Art,” pp. 18–19.

290

Guyer

metaphysical secret to freedom or to the world, no basic enigma, that only art can hint at or intimate.” Thus “art […] loses its power to say the unsayable or represent the unrepresentable, since what was supposedly ‘unrepresentable’ has turned out to be conceptually comprehensible, namely the ‘Kantian paradox’ as involving the sociality and historicity of agency and reason.”²² Pinkard concludes: Romantic art as modern thus loses its “vocation” as art for us, for although it remains crucial and irreplaceable in human experience, it cannot satisfy us on its own since it seeks something – namely, a comprehensive aesthetic exhibition of the meaning of freedom – that it cannot in principle provide.²³

But why does art fail to satisfy in this way: because we in the modern age have fully succeeded in understanding our own freedom in purely conceptual terms, and therefore no longer need art, with its inescapable residue of the sensuous even in poetry, or because we in the modern age by understanding our own freedom in purely conceptual terms have failed to understand our own freedom, which contains something mysterious that can never be completely captured in our concepts, and are therefore incapable of employing art’s mysterious capacity to intimate the mysterious? Pinkard quotes Hegel as saying that modern art is “a withdrawal of man into himself, a descent into his own breast,” which “makes Humanus its new holy of holies,”²⁴ but then says that “it is clear that Hegel is shi ing the meaning of ‘divine’ from its normal usage and redescribing it in terms of his view of the absolute as Geist, as self-determination, as mankind’s ‘true’ or ‘highest interests.”²⁵ But it is by no means clear to me that this is what Hegel is doing here. His argument seems rather to be that modern art is perfectly adequate to represent everything about human freedom except its mysterious and divine aspect, the fact that human freedom itself cannot be fully grasped through abstract and conceptual means. Hegel’s argument seems to be not that art is unnecessary because freedom has now been rendered unmysterious, but rather that modern art is a failure because it no longer recognizes the mysterious nature of freedom. Pippin seems somewhat more con icted about whether for Hegel modern art has been le behind by modernity’s successful attempt to eliminate mystery from the understanding of human freedom or whether it is a failure precisely because of its own attempt to eliminate mystery from freedom. In “The Absence of Aes-

22 23 24 25

Pinkard, “Symbolic, Classical, and Romantic Art,” p. 21. Pinkard, “Symbolic, Classical, and Romantic Art,” p. 21. Cited from Hegel, Aesthetics, vol. I, p. 607. Pinkard, “Symbolic, Classical, and Romantic Art,” p. 21.

The End of Art and the Interpretation of Geist

291

thetics,” he suggests that Romantic or modern art successfully exempli es modernity’s clear self-understanding of the real nature of human freedom. In general, Pippin suggests, Hegel takes there to be a rm correspondence between a period’s self-understanding and its art: Hegel’s philosophy of art is dependent rst of all on a theory of spirit, on some account of collective, norm-governed human mindedness and an account of the kind of nitude or lack of which explains the production of art works and the legislation of norms for their production and evaluation. We simply need to know how social norms work in order to know how artistic norms work.²⁶

Thus he states that “Art is an achieved form of self-knowledge,”²⁷ romantic art included or even above all, and thus suggests that romantic art adequately reects modernity’s self-understanding of “collective, norm-governed human mindedness.” He writes, [t]o use Hegel’s narrative metaphors, having discovered that human beings do not have a xed, purposive “place” in nature, no natural home (that nature is disenchanted), spirit abandons its attempt to “see itself” or “ nd itself” in nature or in corporeal externality at all […] and begins the attempt to see itself in its own products, to nd a way to see its culture, work-world, politics, laws, and religion as “its own” […] Romantic art is then both psychologically sensuous and re ective, expressive of how an experience, another person, a world, seems, or feels, “for the subject” as the most important and privileged dimension of experience […] ²⁸

On this account, it seems as if modernity has come to a fully adequate understanding of human agency and that modern, romantic art fully expresses this understanding. Art may have lost its sense of mystery, but that is only because there is really no mystery to human agency. And Hegel’s thesis of the end of art has fallen by the wayside. Pippin’s previous paper “What Was Abstract Art?” seems to suggest a more complex picture, however, in which there is still something divine about human freedom and modern art has two choices, either to nd ways to intimate this divinity and thus avoid the end Hegel foresaw for it, or succumb to this end. Here, following his claim that by the divine Hegel just means the highest interests of mankind, Pippin continues that Art, in other words, is treated as a vehicle for the self-education of human being about itself, ultimately about what it means to be a free, self-determining being, and when Hegel calls

26 Pippin, “Absence of the Aesthetic,” p. 409. 27 Pippin, “Absence of Aesthetics,” p. 411. 28 Pippin, “Absence of Aesthetic,” p. 414.

292

Guyer

that dimension of aesthetic meaning divine, he seems to be rather attering the seriousness and nality of the enterprise (its independence from sensual need, utilitarian interest, and so on; its “absolute” importance) than in any sense worrying about the God of revealed religion. Another way to put Hegel’s quite heretical view would be to say that for Hegel artistic activity is not about representing divinity, but expressing divinity and even becoming divine.²⁹

This passage suggests that Pippin realizes that the humanistic interpretation of Geist can be attributed to Hegel only by also attributing to him a conception of the human being, speci cally of human freedom and agency, that goes beyond any normal understanding of human agency and straddles the boundary between the human and divine. He tries to resist this implication by saying that for Hegel “Nature will not be lost or rendered a mere object” in the process of achieving rather than recovering “an original harmony between our corporeality or natural fate and our agency, spontaneity, and freedom,” but “transformed, remade into a ‘second nature’,” and that “none of this means that we become, or realize we always were, supernatural beings, or that we can ignore our corporeality. We remain nite, constrained in all the obvious ways by natural limitations.”³⁰ But while on his account of Hegel our corporeality and its embeddedness in the rest of nature may establish the circumstances within which we can exercise our agency, our norms are entirely self-determined, “self-legislated” and “self-authorized,” and in that regard humans are not determined by nature but are more like something divine. Hegel’s end-of-art thesis is then based on the assumption that we are incapable of representing this supernatural capacity in ourselves by natural, that is, sensuous means: “representational art cannot adequately express the full subjectivity of experience, the wholly self-legislating, self-authorizing status of the norms that constitute such subjectivity, or thus cannot adequately express who we (now) are.”³¹ It is because there is something essentially non-natural about human agency that it cannot be represented by natural means. Pippin’s ultimate suggestion in “What Was Abstract Art?” is that Hegel was right about human agency but wrong about the limits on any artistic expression of our understanding of it, because Hegel, in this regard still a child of the eighteenth century, assumed that art had to be representational, and could not foresee the possibility of a form of art that depends upon “abstraction from dependence on sensual immediacy, and so a kind of enactment of the modernist take on normativity since Kant.” His argument here involves two steps: rst, that paintings by the likes of Jackson Pollock and Mark Rothko are not just “presentations of

29 Pippin, “What Was Abstract Art?”, p. 250. 30 Pippin, “What Was Abstract Art?”, p. 253. 31 Pippin, “What Was Abstract Art?”, p. 259.

The End of Art and the Interpretation of Geist

293

paint drips and color elds and at canvas,” but “thematize and so render selfconscious components of sensible meaning” or “present the materiality of such components in their conceptual signi cance”; and second, that this conceptualization or thematization of the material components of painting can in turn lead to an understanding of the self-legislating character of human norms in general, a “generalized idealism” that “has come to be part of the intellectual habits of mind of modern self-understanding.”³² One might question both steps of this argument: rst, one might raise a question about this idealist interpretation of painting like the one Eduard Hanslick raised about idealist theories of absolute music, namely what entitles us to assign to a work in a non-representational medium, or any aspect of it, any de nite content beyond simple feelings that we can associate directly with such directly perceptible features as “audible changes of strength, motion, and ratio”? – what entitles us to say that music expresses “love, wrath, or fear” rather than just being “gentle” or “violent”?;³³ and second, one might ask why we should suppose that even if abstract art can express de nite ideas about its own materials it can also express de nite ideas about the general nature of human agency? As I have already suggested, it might seem as if only the philosophy of art, not art itself, could make the move from a claim about the meaning of a certain kind of artistic production to a general claim about the nature of human agency, which would not undercut Hegel’s end-of-art thesis but con rm it. But arguing whether abstract art could save us from the end-of-art thesis or only con rms it is not my purpose; my point is rather that the humanistic interpretation of the meaning of Geist in Hegel’s aesthetics can be defended only by means of an interpretation of human agency that undercuts the point of the distinction between the humanistic and the more traditional and transcendent interpretations of Geist in the rst place. Once human agency has been understood as “absolute,” having perhaps to apply itself within the constraints of nature but remaining entirely free from constraint by nature in the determination of its norms or principles, human agency is being understood as to this extent like divine agency, the agency of a self-caused cause. That the humanistic interpreters of Geist end up attributing a view of human agency like this to Hegel does not, to be sure, impugn the accuracy of their reading of Hegel, only the description of their interpretation as humanistic. The ironic accuracy of their reading will be con rmed by a closer look at Hegel’s own text, as we will see in the next section.

32 Pippin, “What Was Abstract Art?”, p. 262. 33 Eduard Hanslick, The Beautiful in Music, transl. by G. Cohen (Indianapolis: Bobbs-Merrill, 1957), pp. 23–4.

294

Guyer

To be sure, imputing to Hegel a non-naturalistic and in that sense transcendent theory of human agency would only place him in the mainstream of continental philosophy from the beginning of modernity to his own time. In British philosophy, in Hobbes, Locke, and Hume, we may nd a thoroughly naturalistic account of human agency, where freedom is reduced to the “liberty” to do what one chooses independent of interference by other agents, but where one’s choice itself is determined by entirely natural mechanisms. But in the mainstream of European thought from Descartes to Kant, there is always something supernatural or divine about human agency. Descartes had asserted, not for purposes of moral philosophy but for the epistemological purpose of demonstrating our ability to withhold assent from anything that is not completely clear and distinct, that “the will, or freedom of choice, which I experience within me [is] so great that the idea of any greater faculty is beyond my grasp; so much so that it is above all in virtue of the will that I understand myself to bear in some way the image and likeness of God.”³⁴ And Kant had argued that our consciousness of the moral law as a “pure practical law” proves the existence in us of a “pure understanding” and a “pure will” that can neither explain nor be explained by anything “in appearances,”³⁵ thereby giving us “practico-dogmatic knowledge” of our own freedom as “supersensible,” like God,³⁶ a knowledge that can be “forti ed” by experience of “the physical purposiveness that is to be met with in the world”³⁷ but that is not dependent on such experience. In his nal work, unknown to Hegel, Kant went so far as to suggest that we can understand human agency and human normativity only by understanding human being as identical to God, although that in turn means understanding God as identical to human being: “There is a God, not as a world-soul in nature, but as a personal principle of human reason […] which, as the idea of a holy being, combines complete freedom with the law of duty.”³⁸ For the categorical imperative, “which commands for nature freedom under laws and through which freedom itself demonstrates the principle of its own possibil-

34 René Descartes, “Meditations on First Philosophy,” Fourth Meditation; in The Philosophical Writings of Descartes, transl. by J. Cottingham, R. Stootho , and D. Murdoch (Cambridge: Cambridge University Press, 1984), vol. II, p. 40. 35 Immanuel Kant, Critique of Practical Reason, 5:30, in I. Kant, Practical Philosophy, ed. and transl. by M. J. Gregor (Cambridge: Cambridge University Press, 1996), p. 163. 36 Immanuel Kant, What Real Progress Has Metaphysics Made in Germany since the Times of Leibniz and Wol ?, 20:296, in I. Kant, Theoretical Philosophy a er 1781, ed. by H. E. Allison and P. Heath (Cambridge: Cambridge University Press, 2002), p. 385. 37 Kant, Real Progress, 20:300, p. 388. 38 Immanuel Kant, Opus postumum, 21:19, in I. Kant, Opus postumum, ed. by E. Förster, transl. by E. Förster and M. Rosen (Cambridge: Cambridge University Press, 1993), p. 225.

The End of Art and the Interpretation of Geist

295

ity[,] the commanding subject is God,” although “This commanding being is not outside man as a substance di erent from man.”³⁹ Thus for Kant human normativity and human agency can be understood only as something divine, although the divine can be understood as something existing only within the human. This is not humanism of any normal sort, and this is the sort of thought to which Hegel’s conception of human normativity and agency is indebted. To be sure, Hegel has a story to tell about how humankind’s understanding of its own divine character emerges that Descartes did not have at all and that Kant only sketched, but Hegel’s theory of human agency and of the inherent limitations of artistic representation of it can only be understood in these decidedly non-naturalistic terms. So let us now return to Hegel’s own account of art and the crucial moments in his characterization of its ultimate limitations.

3 The Limits of Romantic Art The heart of Hegel’s argument for the end of the signi cance of art is that art is charged with sensuously representing a content that is ultimately incapable of sensuous representation. The humanistic interpretation of Geist interprets this content as human freedom (and Pippin then argues that forms of art – to t them into Hegel’s historiography, call them post-romantic – are capable of representing human freedom a er all). But Hegel himself most typically describes the intended content of art as God or the divine, and argues that an adequate understanding of God is incapable of sensuous representation. When he does suggest that it is human freedom that is the intended content of art, it is only on an understanding of human freedom that makes it something divine rather than reducing the divine to the level of something human. The key to his argument is that while romantic art is the art of inwardness, either the divine or human freedom itself must ultimately be understand as absolute, in nite, or pure inwardness, and that inwardness so understood cannot be sensuously represented. Indeed, Hegel’s account of the individual arts ends not with tragedy but with comedy, not the ultimate kind of art but rather the symbol of the ultimate fate of art. The seed for the eventual conclusion of Hegel’s argument is planted in his “Division of the Subject” in the Introduction even though the denouement of the argument is not stated here. The premise of the Division is that “the content of art is the Idea, while its form is the con guration of sensuous material. Now art has to

39 Kant, Opus postumum, 21:21; pp. 227–8.

296

Guyer

harmonize these two sides and bring them into a free reconciled totality.”⁴⁰ The ensuing outline of the division may allow it to appear that this reconciliation is possible, but Hegel’s larger argument is that it is not; and the reason why appears at the end of his sketch of the division. From his premise Hegel derives as his “ rst point […] the demand that the content which is to come into artistic representation should be in itself quali ed for such representation,” and the “second demand, derived from the rst, requires of the content of art that it be not anything abstract in itself, but concrete, though not concrete in the sense in which the sensuous is concrete when it is contrasted with everything spiritual and intellectual and these are taken to be simple and abstract.”⁴¹ The inference to be drawn is that, working as it always does with some sensuous medium, where the sensuous is always concrete (a supposition that clearly derives from Kant’s conception of intuition as individual rather than general), art must represent something other than its own medium, thus something intellectual, but still, because of its medium, something concrete, thus something intellectual but not abstract; and this can be nothing other than God. Thus, “thirdly, if a sensuous form and shape is to correspond with a genuine and therefore concrete content,” that content “must likewise be something individual, in itself completely concrete and single,”⁴² but that can be nothing other than God, indeed, “God as conceived by Christian ideas” rather than the “Greek gods,” which, for all their concrete human form, from an intellectual point of view are abstractions, each representing a general property such as wisdom or love rather than a genuine individual. But while the “Christian God […] is indeed a concrete personality,” it is “pure spirituality and is to be known as spirit and in spirit,”⁴³ and although Hegel does not yet say this, pure spirituality is precisely what ultimately turns out to be incapable of sensuous representation a er all. This point becomes clear in Hegel’s ensuing outline of the three historical stages of art, as does his commitment to a conception of human freedom that makes it incapable of sensuous representation because of its absoluteness even if we take the content of art to be human freedom rather than God. Thus, on Hegel’s account the rst stage of art, symbolic art, is inadequate to the task of art because neither its conception of spirit nor its use of the form of sensuous media is adequate. It is “a mere search for portrayal” rather “than a capacity for true presentation” because its conception of spirit falls far short of comprehending its absoluteness or in nitude but also because it just “exagger-

40 Hegel, Aesthetics, vol. I, p. 70. 41 Hegel, Aesthetics, vol. I, p. 70. 42 Hegel, Aesthetics, vol. I, p. 71. 43 Hegel, Aesthetics, vol. I, p. 72.

The End of Art and the Interpretation of Geist

297

ates natural shapes and the phenomena of reality itself into inde niteness and extravagance”⁴⁴ rather than truly transforming them into something ideal. “The symbolic shape is imperfect because (i) in it the Idea is presented to consciousness only as indeterminate or determined abstractly, and, (ii) for this reason the correspondence of meaning and shape is always defective and must itself remain purely abstract.”⁴⁵ Hegel then says that “The classical art-form clears up this double defect; it is the free and adequate embodiment of the Idea in the shape peculiarly appropriate to the Idea itself in its essential nature,”⁴⁶ but this is a misleading representation of his ensuing argument, which is that classical art must also give way because its more determinate use of its media, for example its idealization of the human form in sculpture rather than the exaggeration of animal forms, is necessarily inadequate to an adequate comprehension of spirit. Thus Hegel continues that “if the correspondence of meaning and shape is to be perfect,” as it is in classical art, then “the spirituality, which is the content, must be of such a kind that it can express itself completely in the natural human form,” thus “here the spirit is at once determined as particular and human, not as purely absolute and eternal, since in this latter sense it can proclaim and express itself only as spirituality” – but “This last point in its turn is the defect which brings about the dissolution of the classical art-form and demands a transition to a higher form, the third, namely the romantic.”⁴⁷ But although the transition to the romantic art-form is motivated by the recognition that spirit is “purely absolute and eternal” and thus cannot “express itself completely in the natural human form,” this is also what dooms the romantic art-form and with it art itself, because no sensuous medium, not even the most rare ed, such as poetry, can adequately represent what is “purely absolute.” Ultimately only philosophy can do that. The ultimate fate of art is immediately pre gured in Hegel’s opening characterization of romantic art, which has to be quoted at length: The romantic form of art cancels again the completed uni cation of the Idea and its reality, and reverts, even if in a higher way, to that di erence and opposition of the two sides which in symbolic art remained unconquered. The classical form of art has attained the pinnacle of what illustration by art could achieve, and if there is something defective in it, the defect is just art itself and the restrictedness of the sphere of art. This restrictedness lies in the fact that art in general takes as its subject-matter the spirit (i.e., the universal, in nite and concrete in its nature) in a sensuously concrete form, and classical art presents the complete uni cation of spiritual and sensuous existence as the correspondence of the two. But in this blending of

44 45 46 47

Hegel, Aesthetics, vol. I, p. 76. Hegel, Aesthetics, vol. I, p. 77. Hegel, Aesthetics, vol. I, p. 77. Hegel, Aesthetics, vol. I, pp. 78–9.

298

Guyer

the two, spirit is not in fact represented in its true nature. For spirit is the in nite subjectivity of the Idea, which as absolute inwardness cannot freely and truly shape itself outwardly on condition of remaining moulded into a bodily existence as the one appropriate to it.⁴⁸

The problem for art is that spirit is not just concrete but also, however this is supposed to be combined with concreteness, absolute or in nite. While subjectivity or inwardness may be expressed in sensuous media, and the paradigmatically romantic arts of painting, music, and poetry succeed in doing precisely that, absolute or in nite subjectivity or inwardness cannot be so expressed. Thus when “Christianity brings God before our imagination as spirit, not as an individual, particular spirit, but as absolute in spirit and in truth,” then “For this reason it” necessarily “retreats from the sensuousness of imagination into spiritual inwardness.” “Thus the unity of divine and human nature is a known unity, one to be realized only by spiritual knowing and in spirit,” and “In this way romantic art is the self-transcendence of art.”⁴⁹ To be sure, the last sentence continues “but within its own sphere and in the form of art itself,” but what Hegel means by this is that romantic art in its paradigmatic form expresses discontent at art itself, which is why romantic art culminates not in sublime poetry such as tragedy but in comedy. Even when Hegel shi s from talking about God or the unity of divine and human nature as the intended but unreachable subject-matter of art to putting human freedom in this place, he does so only with a conception of human freedom that sees it not as working within the bodily conditions of human existence as we ordinarily understand them but rather as completely transcending the limits of bodily existence, thus as something supernatural or more divine than human. This is evident in Hegel’s next paragraph, which begins with the statement that “at this stage the subject-matter of art is free concrete spirituality, which is to be manifested as spirituality to the spiritually inward.” At rst it sounds as if romantic art is simply in the business of representing the internal side of human action, the domain of the feelings that may inspire our actions, the intentions that may be so inspired, and our feelings about those intentions, all of which the romantic arts might seem perfectly well capable of representing: In conformity with this subject-matter, art cannot work for sensuous intuition. Instead it must, on the one hand, work for the inwardness which coalesces with its object simply as if with itself, for subjective inner depth, for re ective emotion, for feeling which, as spiritual,

48 Hegel, Aesthetics, vol. I, p. 79; Hegel’s (that is, Hotho’s) emphases throughout. 49 Hegel, Aesthetics, vol. I, p. 80.

The End of Art and the Interpretation of Geist

299

strives for freedom in itself and seeks and nds its reconciliation only in the inner spirit. This inner world constitutes the content of the romantic sphere and must therefore be represented as this inwardness and in the pure appearance of this depth of feeling.

Romantic arts such as music and poetry are surely capable of representing inwardness and depth of feeling. But then Hegel concludes that “Inwardness celebrates its triumph over the external and manifests its victory in and on the external itself, whereby what is apparent to the senses alone sinks into worthlessness.”⁵⁰ Hegel’s idea is not that human freedom consists in the control of the sensuous and the bodily, but ultimately in the complete transcendence of the sensuous and the bodily. “The aspect of external existence is assigned to contingency” and [t]hereby the separation of Idea and shape, their indi erence and inadequacy, come to the fore again, as in symbolic art, but with this essential di erence, that, in romantic art, the Idea, the de ciency of which in the symbol brought with it de ciency of shape, now has to appear perfected in itself as spirit and heart. Because of this higher perfection, it is not susceptible of an adequate union with the external, since its true reality and manifestation it can seek and achieve only within itself.⁵¹

With its essential employment of sensuous media, art would nevertheless be capable of representing human freedom if that itself consisted in the achievement of harmony between human feeling and thought on the one hand and human embodiment and physical action on the other (as Kant, for example, conceived of it, thereby concluding that it was necessary to bridge the gulf between the legislation of nature and the legislation of freedom to have an adequate theory of human freedom).⁵² But Hegel conceives of freedom as the transcendence of the bodily into a realm of pure or in nite spirituality, and therefore nds art as incapable of adequately representing freedom as it is incapable of adequately representing the divine. The divine can be equated with freedom as the highest interest of mankind only because freedom is itself understood in terms ordinarily reserved for the divine – Hegel’s conception of freedom is nothing other than Kant’s conception of a holy will, which Kant rejected as a model for human virtue but Hegel does not. The distinctively non-natural character of Hegel’s conception of spirit and its freedom is further on display in the opening pages of the rst of the three main sections of the Aesthetics, the exposition of “The Idea of Artistic Beauty, or the Ideal,” which precedes the sections on the historical art-forms and then on the individual arts. The tone of this section is set by the statement that “If we wish to

50 Hegel, Aesthetics, vol. I, p. 81. 51 Hegel, Aesthetics, vol. I, p. 81. 52 See Kant, Critique of the Power of Judgment, Introduction, section II, 5:175–6.

300

Guyer

indicate brie y what the absolute Idea is in its genuine actuality, we must say that it is spirit, not, as may be supposed, spirit in its restrictedness and involvement with the nite, but the universal in nite and absolute spirit which out of itself determines what is genuinely the true.” Hegel rejects any idea that spirit grows out of nature, as Schelling had argued, or even that it has to work with nature and its limitations; rather, spirit itself posits nature and overcomes any limitations in it: “nature does not stand over against spirit, either as possessing the same value or as spirit’s limitation; on the contrary, it acquires the standing of having been posited by spirit, and this makes it a product, deprived of the power of limiting and restricting.”⁵³ Speci cally, Hegel interprets spirit’s activity as one of freeing itself from all limitations of nature, and genuine freedom only as spirit’s transcendence of all such limitations. It is precisely this conception of spirit and of spirit’s freedom as the content of art that places art within absolute knowing but also guarantees that art itself must be transcended: This is the point at which we have to begin in the philosophy of art. For the beauty of art is neither the Idea as conceived in Logic, i.e., absolute thought as it is developed in the pure element of thinking, nor yet, on the other hand, the Idea as it appears in Nature; on the contrary, it belongs to the sphere of spirit, though without stopping at the knowledge and deeds of the nite spirit. The realm of art is the realm of the absolute spirit. […] the Idea in logic has, in accordance with its own Concept, to transpose itself into natural existence and then, out of this externality, into spirit; and nally to free itself from the nitude of spirit again to become spirit in its eternity and truth.⁵⁴

It is essential to spirit that it free itself from all nitude, thus all constraint by nature. If spirit is identical to humanity, then it is essential to humanity that it free itself from all nitude, thus all constraint by nature. The task of art is to represent this transcendence of the nite and external, but since art is essentially connected to nite and external media, it is the task of art to transcend itself. Art can represent spirit only by transcending itself, and if human freedom is spirit then art can represent human freedom only by transcending itself. “From this point of view, which pertains to art in its highest and true dignity, it is at once clear that art belongs to the same province as religion and philosophy,”⁵⁵ but insofar as it belongs to this province it must also be superseded by religion and philosophy – and religion too, insofar as it remains tied to representations if not necessarily sensuous representations, must itself be transcended by philosophy.

53 Hegel, Aesthetics, vol. I, p. 92. 54 Hegel, Aesthetics, vol. I, p. 94. 55 Hegel, Aesthetics, vol. I, p. 94.

The End of Art and the Interpretation of Geist

301

The remarkable character of Hegel’s conception of freedom is made even more explicit a few pages further when he states that Now the highest content which the subject can comprise in himself is what we can pointblank call freedom. Freedom is the highest destiny of the spirit. In the rst place, on its purely formal side, it consists in this, that in what confronts the subject there is nothing alien, and it is not a limitation or a barrier; on the contrary, the subject nds himself in it. Even under this formal de nition of freedom, all distress and every misfortune has vanished, the subject is reconciled with the world, satis ed in it, and every opposition and contradiction is resolved.⁵⁶

The subject does not nd freedom, as the Hellenistic conception of ataraxia or the Kantian conception of virtue recommended, by learning to recognize and accept the limitations of nature, but only by overcoming them – distress and misfortune are not to be accepted as part of the fabric of nature within which human freedom must be exercised, but are to vanish, or be vanquished. What man seeks […] is the region of a higher, more substantial, truth, in which all oppositions and contradictions in the nite can nd their nal resolution, and freedom its full satisfaction. This is the region of absolute, not nite, truth. The highest truth, truth as such, is the resolution of the highest opposition and contradiction. In it validity and power are swept away from the opposition between freedom and necessity, between spirit and nature, between knowledge and its object, between law and impulse […] To grasp this Concept of truth is the task of philosophy,⁵⁷

but also the task of art, even though it is a task that art cannot discharge while remaining art. Hegel’s string of oppositions that are to be overcome shows how radical his approach to spirit and freedom is: freedom consists not just in learning to control impulse, but in eliminating all tension between law and impulse, just as knowledge consists not in learning the limits of our own representational schema but in overcoming all opposition between knowledge and its object, thus all distinction between the phenomenal and the noumenal. This is a superhuman conception of human knowledge and human freedom, and thus art can represent human spirit only by representing human spirit as superhuman – transcending itself into the bargain. This conception of spirit as the content of art is at work in Hegel’s account of art’s self-transcendence in its nal stage, romantic art, and in his argument that even the most universal art, poetry, can only end in comedy. The concluding section of Hegel’s discussion of the romantic form of art is entitled “The End of the

56 Hegel, Aesthetics, vol. I, p. 97. 57 Hegel, Aesthetics, vol. I, pp. 99–100.

302

Guyer

Romantic Form of Art.”⁵⁸ Hegel begins this section by reminding us that “Art, as it has been under our consideration hitherto, has as its basis the unity of meaning and shape and so the unity of the artist’s subjective activity with his topic and work.” He then states that “it is the vocation of art to nd for the spirit of a people the artistic expression corresponding to it,”⁵⁹ which might seem to imply, rst, that each people or culture has a distinctive spirit and that “Spirit” in general is nothing but the collectivity of these more particular cultures or national spirits, which would be consistent with the humanistic approach to spirit, and, second, that every such spirit can nd adequate expression in some artistic medium (or several), thus even that the spirit of post-classical modern Europe should be able to nd adequate artistic expression. Hence romantic art should be an adequate form for the expression of spirit in general in its modern European form. However, whereas in pre-romantic art there has been an unbreakable bond between the content of art and its medium, such that “for the artist the content [has] constitute[d] the substance, the inmost truth, of his consciousness, and [made] his chosen mode of presentation necessary,” in the development of romantic art “the whole situation has altogether altered.”⁶⁰ Now the artist must express his own freedom but also the very nature of spirit as such by liberating himself from both all “content which on every occasion was determinate for a particular people, a particular age,” and also all “speci c consecrated forms and con gurations”; the artist thus “moves freely on his own account, independent of the subject-matter and mode of conception in which the holy and eternal was previously made visible to human apprehension.”⁶¹ This might make it seem as if the artist is to free himself precisely from all conceptions of the holy and eternal in order to express a purely human conception of (his own) freedom; the artist could express any aspect of his own inwardness or subjectivity and in any form he likes as long as that form passes the purely human test of beauty: “every material may be indi erent to him if only it does not contract the formal law of being simply beautiful and capable of artistic treatment.”⁶² To this extent the “self-transcendence” of art would be “a withdrawal of man into himself, a descent into his own breast, whereby art strips away from itself all xed restriction to a speci c range of content and treatment, and makes Humanus,” something human rather than divine, “its new holy of holies.”⁶³ It would be “the appearance and activity of imperishable humanity in

58 Hegel, Aesthetics, vol. I, p. 602. 59 Hegel, Aesthetics, vol. I, p. 603. 60 Hegel, Aesthetics, vol. I, p. 604. 61 Hegel, Aesthetics, vol. I, pp. 604–5. 62 Hegel, Aesthetics, vol. I, p. 605. 63 Hegel, Aesthetics, vol. I, p. 607.

The End of Art and the Interpretation of Geist

303

its many-sided signi cance and endless all-round development” that would “now constitute the absolute content of our art.”⁶⁴ Art would seem to be a completely adequate medium of cognition because there would seem to be nothing about purely human experience or inwardness that it would be incapable of expressing. However, the cost of this focus on purely human experience is precisely that art loses its seriousness and becomes a form of “harmless play.” Hegel’s concluding example of romantic art is Goethe’s poem cycle West-östlicher Divan,⁶⁵ which can fuse a modern European sensibility with the forms and imagery of much older, symbolic Persian art precisely because it is merely a form of play. In a paradigmatic poem such as Wieder nden, Hegel argues, “love is transferred wholly into the imagination, its movement, happiness, and bliss”; here we nd an inexhaustible self-yielding of imagination, a harmless play, a freedom in toying alike with rhyme and ingenious metres – and, with all this, a depth of feeling and a cheerfulness of the inwardly self-moving heart which through the serenity of the outward shape li the soul high above all painful entanglements in the restrictions of the real world.”⁶⁶

This last remark suggests that the ultimate limitation of romantic art is that it elevates the soul above all entanglement with the “restrictions of the real world” by sheer imagination, rather than recognizing the seriousness of human freedom: that it must elevate itself above the restrictions of the real world, but only by recognizing its own divinity rather than by “liberating” itself from all conceptions of the divine and eternal. Romantic art is not, in the end, a genuine recognition of the true nature of human freedom, its genuine divinity, but liberation from the real world in mere imagination. Art is not a serious way to represent human freedom because human freedom can only be understood as something divine, not a mere form of play.

64 Hegel, Aesthetics, vol. I, p. 608. 65 A poem of six stanzas, ending with the remarkably evocative lines: So, mit morgenrothen Flügeln, Riß es mich an deinen Mund, Und die Nacht mit tausend Siegeln Krä igt sternenhell den Bund. Beyde sind wir auf der Erde Musterha in Freud und Quaal Und ein zweytes Wort: Es werde! Trennt uns nicht zum zweytenmal.

From the 1819 edition, from Johann Wolfgang Goethe, West-östlicher Divan, ed. by H. Birgus, 2 vols. (Insel Verlag: Berlin, 2010), vol. I, pp. 96–7. For variants, see also pp. 399–400 and 570–1. 66 Hegel, Aesthetics, vol. I, pp. 610–11.

304

Guyer

This is why Hegel’s lengthy description of poetry as the paradigmatic medium of romantic art ends rst with a description of modern tragedy but then with a brief account of comedy. Like most other Germans of his time, Hegel regards Shakespeare as the high point of modern tragedy, who gives us “the nest example of rm and consistent characters who come to ruin simply because of [their] decisive adherence to themselves and their aims.”⁶⁷ Precisely because of their identi cation with their own inwardness rather than with anything larger than themselves, such characters act “Without ethical justi cation, but upheld solely by the formal inevitability of their personality.”⁶⁸ In such cases “the sole spectacle o ered to us is that the modern individual with the non-universal nature of his character […] is necessarily surrendered to all that is mundane and must endure the fate of nitude.”⁶⁹ In other words, the only way to avoid tragedy is by means of a selfconception of human inwardness as in nite rather than nite, a self-conception that sees the human being as part of the divine rather than reducing the divine and our highest interest to ordinary human subjectivity or any account of freedom as merely the liberty to pursue individual preferences. Then what becomes laughable in modern comedy, such as that of Molière, following the New Comedy of antiquity, is “folly and one-sidedness”⁷⁰ that is recognized as such by the audience but not by the characters themselves, who in their self-seriousness, their lack of self-knowledge, are like the self-destructive protagonists of modern tragedy re ected in fun-house mirrors. The characters of modern comedy are “deadly serious in their aims” and pursue them “with all the fervour of this seriousness,” like “Molière’s miser whose absolutely serious involvement in his narrow passion inhibits any liberation of his mind from this restriction.” When at the end such characters “are deceived or have their aim frustrated by themselves, they cannot join in the laughter freely and with satisfaction but, duped, are the butt of the laughter of others, o en mixed as it is with malice.”⁷¹ Thus, whether we pity a king or somewhat maliciously laugh at a miser, the real object of our response is the spectacle of self-restriction, self-involvement, or self-satisfaction: all forms of nitude. Hegel’s conclusion is that because of its sensuous medium, even in the case of poetry, even modern art can reveal to us the limits of a nite conception of our own subjectivity but cannot positively display the genuinely in nite or divine character

67 Hegel, Aesthetics, vol. II, p. 1229. 68 Hegel, Aesthetics, vol. II, p. 1230. 69 Hegel, Aesthetics vol. II, p. 1231. 70 Hegel, Aesthetics, vol. II, p. 1233. 71 Hegel, Aesthetics, vol. II, p. 1234.

The End of Art and the Interpretation of Geist

305

of our subjectivity. Modern art “brings the negative side of this dissolution into consciousness in the humour of comedy,” but in it “the presence and agency of the Absolute no longer appears positively uni ed with the characters and aims of the real world”; in it “subjective personality alone shows itself self-con dent and self-assured […] in this dissolution.” Taking a swipe at any theory of art as o ering us a possibility for a form of play the value of which does not depend upon cognition of some great truth, Hegel says that “in art we have to do, not with any agreeable or useful child’s play, but with the liberation of the spirit from the content and forms of nitude.”⁷² But this liberation, the truly in nite or absolute character of spirit, or the participation of human spirit in something more than merely human, is precisely what art cannot in the end adequately represent. It might be objected to this conclusion that I have cut Hegel o in mid-sentence, for he continues that in art we also have to do with “the presence and reconciliation of the Absolute in what is apparent and visible, with an unfolding of the truth which is not exhausted in natural history but revealed in world-history.”⁷³ This suggests that the presence of the absolute, thus of the ultimately in nite and divine character of human freedom, can be made visible in appearances and thus in art a er all, perhaps in some form of art that transcends the tragedy and comedy in which romantic art culminates. But this might be a point on which the posthumous Hotho edition of Hegel’s lectures on which I have been relying is misleading. For Hotho’s own transcription of Hegel’s 1823 lectures, presumably closer to the actual words of the master than the edition Hotho compiled four years a er Hegel’s death, ends on a less reassuring tone. Here Hegel is reported as saying that “the subjectivity” of comedy, which “is satis ed with and consoled in itself, and only plays with the objective,” is the “opposite” of the representation of “the individual as divine, as standing beyond particular subjectivity,” thus that in comedy “Subjectivity annihilates objectivity.” He then concludes with the statement that [w]ith this we have completed the circle of art. Art in its seriousness is for us something past. For us other forms are necessary to make the divine into an object. We need thought. But art is an essential means for the presentation of the divine, and we must understand this form. It does not have as its object the agreeable, not subjective skillfulness. Philosophy has to consider what is true in art.⁷⁴

72 Hegel, Aesthetics, vol. II, p. 1236. 73 Hegel, Aesthetics, vol. II, p. 1236. 74 G. W. F. Hegel, Vorlesungen über die Philosophie der Kunst, ed. by A. Gethmann-Siefert (Hamburg: Felix Meiner, 2003), pp. 311–12 (my translation).

306

Guyer

This passage is perhaps ambiguous, or ambivalent, but it seems best to read it as saying that art has been an essential means for the presentation of the divine, and for the understanding of ourselves as part of the divine, but its seriousness is for us something past because it is not really up to the task that has been imposed on it. In considering what is true in art, what philosophy considers is that art has striven to represent something, humanity or human agency, which it is ultimately incapable of representing because that can be understood only by elevating the human to the divine rather than reducing the divine to the human, and art, which is con ned to sensuous and therefore nite media, is ultimately incapable of adequately representing the in nitude of the divine and of ourselves as part of it. Art can adequately represent the tragedy or comedy that results from any nite self-conception of humanity, but not the properly in nite conception of humanity. Hegel’s thesis of the end of art thus can be understood only on an understanding of humanity that is not humanistic.

Anton Friedrich Koch

Metaphysik bei Hegel oder analytische, synthetische und hermeneutische Philosophie 1 Das Ideal der Transparenz Unter dem Ideal der Transparenz verstehe ich eine Erwartung, die so tief in unserem Denken verwurzelt ist, daß wir sie quer durch unterschiedlichste philosophische und weltanschauliche Lager antre en, nämlich die Erwartung, daß im Prinzip alles erkannt werden kann. Platon lehrte, wahrha real seien die Ideen jenseits der Erfahrungswelt, und zugleich seien sie das wahrha Erkennbare. Sein schließt demnach Erkanntwerdenkönnen ein und umgekehrt. Wo es De zite in der Erkennbarkeit gibt, nach Platon im Bereich des sinnlich Erfahrbaren, da haben wir es nicht mit gediegenem Sein, sondern mit einer Mischung von Sein und Nichtsein, mit bloßem Werden zu tun. In kritischem Anschluß an Platon lehrte Aristoteles, daß die Ideen nicht transzendent, sondern der Erfahrungswelt immanent sind, so daß also auch diese sich dem Erkennen darbietet. Auf solche Vorarbeiten konnte sich die Philosophie des Mittelalters stützen, um die religiöse Überzeugung, daß Gott alles weiß und sieht, in metaphysische Lehren zu übersetzen. Statt vom Ideal der Transparenz kann man insofern auch vom Ideal der Allwissenheit sprechen. Dieses Ideal hat in der postreligiösen Theoriebildung unserer Tage keineswegs ausgedient; es ist nur säkularisiert worden. Die Aufgabe der modernen Physik ist es, die Naturgesetze zu entdecken und sie in einem Vokabular zu formulieren, das im Prinzip ausreicht, um die Welt vollständig zu beschreiben. Transparenz, Allwissenheit, vollständige Beschreibung – das sind Varianten ein und desselben Ideals. Präzise fassen als Ideal einer vollständigen Weltbeschreibung läßt es sich im Rahmen der analytischen Philosophie: Man nehme (oder denke sich) eine Wissenscha ssprache, die für jeden Gegenstand einen Namen und für jede fundamentale Eigenscha ein Prädikat hat. Das kommt der Sprache Gottes nahe, der dem Kinderlied zufolge Sternlein und Wolken gezählt hat, daß ihm auch nicht eines fehlet an der ganzen großen Zahl, und der alle Mücklein und alle Fischlein mit Namen rief, daß sie all’ ins Leben kamen und nun so fröhlich sind. Wie das Kinderlied, so Rudolf Carnap, dem wir den Begri der Zustandsbeschreibung verdanken. Darunter versteht Carnap eine vollständige Beschreibung einer möglichen Welt mittels der Namen und der Grundprädikate einer physikalischen Ide-

308

Koch

alsprache. Die zutre ende Zustandsbeschreibung beschreibt die wirkliche Welt; nichtzutre ende Zustandsbeschreibungen beschreiben alternative mögliche Welten. Statt von Zustandsbeschreibungen kann man dann aber auch direkt von möglichen Welten reden, wie einst Leibniz. Dieser lehrte, daß Gott in seiner Allwissenheit alle möglichen Welten kennt und in seiner Allgüte und Allmacht die beste von ihnen wirklich gemacht, d.h. gescha en hat. Da haben wir das Ideal der Transparenz und Allwissenheit in Vollendung. Aber am Werke ist es auch dort, wo man von Gott abstrahiert und die vielen Welten rein für sich betrachtet, wie David Lewis es getan hat. Lewis war Realist bezüglich der Welten und begri sie als große, konkrete Einzeldinge. Alle Welten sind real, sind alternative Maximalbestände des Realen, und jeder Weltbewohner nennt seine eigene Welt die wirkliche. Realität und Wirklichkeit, realitas und actualitas, fallen insofern auseinander. – So wird in der metaphysischen Tradition von Platon und Aristoteles über die scholastische Philosophie und Leibniz bis hin zu Carnap und Lewis und zum Naturalismus, Szientismus und Materialismus unserer Tage das Ideal der Transparenz unhinterfragt aufrechterhalten. Was Hegel angeht, hege ich die Überzeugung, daß seine Philosophie den erwähnten metaphysischen Ansätzen überlegen ist, daß aber auch sie noch dem Ideal der Transparenz untersteht und dadurch in ihrem Wahrheitsgehalt beeinträchtigt wird.

2 Thetisches zu Hegels Gesamtkonzeption Hegels Wissenscha der Logik (WdL) besticht durch ihre monumentale Grundspielregel, der zufolge sie strikt voraussetzungslos sein soll. Eine voraussetzungslose Theorie könnte freilich leicht in der Trivialität des unartikulierten Ausrufs „Sein!“ enden, weil alles Weitere voraussetzungsvoll zu sein scheint. Die WdL beginnt tatsächlich mit diesem Ausruf („Sein, reines Sein, – ohne alle weitere Bestimmung“), aber sie endet nicht damit. Für ihre Nichttrivialität und ihren geregelten Fortgang sorgt die Negativität, die von Anfang an mit von der Partie ist. Näher darauf einzugehen, würde im gegenwärtigen Kontext zu weit führen; es muß hier genügen, Hegels Logik, da sie ja kritisiert werden soll, zunächst möglichst stark zu machen, und dabei muß ich kurz und thetisch zu Werke gehen. Die Metaphysik kann aufgefaßt werden als die Theorie des logischen Raumes und der logische Raum als die Gesamtheit dessen, was der Fall sein und was gedacht werden kann. Platon denkt den logischen Raum als den Kosmos der Ideen und den physikalischen Raum als einen Vorhof des logischen Raumes im umgebenden Dunkel des logischen Chaos. Am anderen Ende des Theoriespektrums be-

Metaphysik bei Hegel

309

grei Davis Lewis den logischen Raum als die Menge aller möglichen Welten. Beide, Platon wie Lewis, betrachten den logischen Raum als etwas Fixes und Ewiges, der eine als die Gesamtheit der Ideen, der andere als die Gesamtheit der Welten. Hegel hingegen, und das ist Teil seiner Ausnahmestellung, bietet eine Evolutionstheorie des logischen Raumes an. Die Dimension der logischen Evolution ist freilich nicht die Zeit, die es im Bereich des rein Logischen noch gar nicht gibt, sondern eine ursprünglichere, zeitanaloge, prätemporale Ordnung, eine rein logische Sukzession, vergleichbar der prätemporalen Sukzession der natürlichen Zahlen. Zuerst, in diesem rein logischen Sinn von „zuerst“, ist der logische Raum unmittelbares, unbestimmtes, homogenes Sein, also ungefähr die Seinskugel des Parmenides. Dann bricht am Sein das widerspruchsvolle Werden auf in einer Art logischem Urknall; oder vielmehr ist es immer schon aufgebrochen. Und weil das Werden in sich widerspruchsvoll ist, ist es auch immer schon sofort wieder in sich zusammengesunken. Das heißt, der logische Urknall hat nur einen logischen Augenblick lang gedauert und hat sodann dem ersten halbwegs stabilen Zustand des logischen Raumes Platz gemacht, den Hegel das Dasein nennt. Das Dasein gibt sich als ruhiges Sein, ist aber insgeheim von Negativität in ziert. Denn es ist ja das Negative – der Negativabzug – des Werdens. Das ist sein Pro l, und weiter ist über es zunächst nichts zu sagen. Weil es in seinem Pro l, seiner Bestimmtheit ganz aufgeht, nennt Hegel die Bestimmtheit seine Qualität. Mit dieser fällt das Dasein vollständig zusammen. In der analytischen Philosophie des Geistes ist heute viel von sogenannten Qualia die Rede, von einfachen Sinneseindrücken, z.B. Farbemp ndungen, die keine Qualitäten haben, sondern Qualitäten sind. Das Hegelsche Dasein ist ebenfalls seine Qualität und insofern das singuläre logische Quale. Mit ihm beginnt der logische Raum die geregelte Karriere seiner Evolution. Triebfeder der Evolution ist die Negativität in ihrem Selbstwiderspruch, die am Dasein zunächst verborgen war, dann aber an ihm hervortritt und mit ihm synthetisiert werden muß und dann für Veränderung sorgt. In der systematischen Veränderung löst ein Zustand des logischen Raumes den anderen ab, und die Wahrheit über den logischen Raum ist die ganze Geschichte seiner Evolution. Deren drei Hauptepochen sind die des Seins, die des Wesens und die des Begri s, und Hegels Logik ist demgemäß unterteilt in eine Lehre vom Sein, eine Lehre vom Wesen und eine Lehre vom Begri . In der Seinslogik werden die Kategorien des Seins (Dasein, Qualität, Etwas, Grenze, Endliches, Unendliches, Fürsichsein, Quantität, Maß usw.) als sukzessive Zustände des logischen Raumes hergeleitet. In der Wesenslogik sieht man dann, wie die sogenannten Re exionsbestimmungen (Identität und Unterschied usw.) wechselweise ineinander scheinen und jede sich auf Kosten ihres Pendants als das ganze Wesen präsentiert. Es folgt die Lehre vom Begri , den man auch den

310

Koch

Logos nennen könnte und der sich über Rückfälle ins Sein und ins Wesen hinweg zur Idee entwickelt, die den Ziel- und Fixpunkt der ganzen logischen Entwicklung bildet, von dem aus dann der Übergang in die sogenannte Realphilosophie erfolgt, die als Philosophie der Natur mit der Betrachtung des Raumes und der Zeit anhebt. Im Lauf der langen logischen Entwicklung vom reinen Sein des Anfangs bis zur absoluten Idee werden alle metaphysischen Grundkonzeptionen in nuce abgeleitet und als vorläu g und überboten durch eine Nachfolgerkonzeption dargestellt. Der letzte Nachfolger, die absolute Idee, ist dabei keine inhaltlich neue Konzeption, sozusagen Hegels eigene positive Metaphysik, sondern bringt den Weg durch alle vorhergehenden Stationen, also die Methode der Logik, zusammenfassend zur Darstellung. Der Weg ist hier einmal wirklich das Ziel, der Inhalt die Methode, das Wahre das Ganze – das heißt, das Wahre über den logischen Raum ist die ganze Geschichte seiner Evolution. Hegel tritt also nicht in Konkurrenz zu den vorherigen metaphysischen Systemen, indem er eine neue Konzeption als Alternative entwickelt, sondern er schließt die Metaphysik ab, indem er ihre konkurrierenden Systeme als Glieder eines logischen Prozesses herleitet, der die Evolution des logischen Raumes bildet. Die nachhegelsche Metaphysik und namentlich die metaphysischen Theorien, die in der analytischen Philosophie entwickelt worden sind, können daher vermutlich nichts vollkommen Neues in den Schatz der metaphysischen Grundgedanken mehr einbringen.

3 Analytische, synthetische und hermeneutische Philosophie Um nun die Hegelsche Philosophie dem Ideal der Transparenz zuordnen und das Ideal kritisieren zu können, sei zunächst der Begri der theoretischen Wissenscha eingeführt und gegen den Begri der hermeneutischen Wissenscha pro liert. Wir werden dabei auf vier philosophische Grundoptionen stoßen, die ich als analytische Philosophie (z.B. Leibniz und Lewis), synthetische Philosophie (Hegel und vor ihm Fichte), hermeneutische Philosophie (der frühe Heidegger) und nachwissenscha liche Philosophie (der späte Heidegger) unterscheiden werde. Die beiden erstgenannten, also die analytische und die synthetische Philosophie, verstehen sich als theoretische Wissenscha ; der vierten begegne ich mit Skepsis; der dritten aber – der hermeneutischen Philosophie – traue ich zu, daß sie die zweite und die erste (und das Haltbare an der vierten) einbetten kann. Aristoteles unterschied drei theoretische Wissenscha en: Metaphysik, Physik und Mathematik. Die Metaphysik betrachtet das Reale in seinem Sein, die Physik betrachtet es in seinem Werden, und die Mathematik untersucht rein formale, nichtsubstantielle Aspekte des Realen. Alle sind auf allgemeingültige Wahrheiten aus und thematisieren Einzelnes nicht als Einzelnes, sondern als Fall allgemei-

Metaphysik bei Hegel

311

ner Bestimmungen und Gesetze. Sie machen infolgedessen nur unwesentlichen Gebrauch von indexikalischen Ausdrücken wie „dies“, „ich“, „hier“, „jetzt“ und dem Tempus verbi. Zweifellos hat es die Mathematik im Verzicht auf solche Indikatoren am weitesten gebracht, und mit gutem Grund streben daher die theoretischen Wissenscha en unserer Tage nach mathematischer Darstellung, während die hermeneutischen Wissenscha en, etwa die Geschichtsschreibung, auf Indikatoren angewiesen bleiben (letztere zum Beispiel auf die Indikatoren „vor unserer Zeitrechnung“ und „nach unserer Zeitrechnung“). Die neuzeitliche Verbrüderung der Physik mit der Mathematik hat die Metaphysik ins Hintertre en gebracht; verlegen um Methode und Darstellungsform steht sie seither am Rande der Wissenscha en und in keinem guten Ruf mehr. Als Grundlagendisziplin muß sie hinter gegebene Begri ichkeiten, auch die mathematische, zurückgehen und zahlt dafür den Preis, daß ihre Lehren nicht präzise xierbar und nicht streng beweisbar sind, sondern umstritten bleiben. Nach 2500 Jahren Geschichte der Metaphysik überwiegt schon lange der Eindruck, das kunstvolle Hin und Her ihrer Begri e, Thesen und Argumente sei ausgereizt. Dennoch wird es – auch nach Hegel – fortgesetzt auf höchstem wissenscha lichem und scholastischem Niveau in der analytischen Philosophie, allerdings nun doch mit leichten Anleihen bei der Mathematik und um den fälligen Preis einer Abkehr von Grundproblemen, was zur Folge hat, daß die analytische Philosophie, auch und gerade da, wo sie sich metaphysikkritisch gibt, dogmatische Metaphysik bleibt. Wer analysiert, will vom Zusammengesetzten zum Einfachen, vom Molekularen zum Atomaren kommen. Die Analyse als philosophisches Programm lebt insofern von einem logischen und metaphysischen Atomismus. Eine markante Form des letzteren ist der Materialismus, d.h. die Lehre, daß die physikalischen Tatsachen alles Weitere festlegen bzw. daß alles Weitere über den physikalischen Tatsachen „superveniert“. Diese Behauptung kann mittels des Lewisschen Weltenpluralismus wie folgt präzisiert werden: In jeder Welt, in der dieselben physikalischen Tatsachen bestehen wie in unserer Welt, bestehen auch alle weiteren (biologischen, psychologischen, soziologischen usw.) Tatsachen wie in unserer Welt. Dies nur als kleines Beispiel für die Art und Weise, wie logische Analyse und metaphysische Analyse sich in der analytischen Philosophie wechselseitig stützen. Die Präzision und Strenge der analytischen Philosophie begünstigt indes die Verdrängung von Grundproblemen zugunsten des Traktablen. Zu den Grundproblemen gehört, daß das Denken mit einem Grundwiderspruch beha et ist, wie die unscheinbare Antinomie des Lügnersatzes bezeugt, der dann und nur dann wahr ist, wenn er nicht wahr ist: „Dieser Satz ist nicht wahr“. Da die Welten als widerspruchsfrei konzipiert sind, wird mit ihnen der Grundwiderspruch verdeckt, nicht behandelt. Die Behauptung, daß er dennoch da und daß unser Denken widerspruchsvoll ist – obwohl es unter der Norm der Widerspruchsfreiheit steht –,

312

Koch

nenne ich die Antinomiethese. Sie zwingt uns dazu, die analytische Philosophie hinter uns zu lassen und sie in die synthetische Philosophie einzubetten. Die Antinomiethese zeigt nämlich, daß sogar die Logik, das Festeste in unserem Wissen, zuletzt auf schwankendem Boden ruht. Die Vernun setzt sich und zugleich die Logik in Kra durch eine Art Machtspruch, dessen Umsetzung eigentlich über ihre Krä e geht. Die Grundgesetze der klassischen Logik – das Nichtwiderspruchsprinzip und das Tertium non datur – können daher nur als Normen, also nur im Konjunktiv, nicht im Indikativ gelten: „tertium non detur“, nicht „datur“: Es soll kein Drittes geben zwischen einer Aussage und ihrer Verneinung. Und es soll kein Widerspruch sein. Der Widerspruch ist aber da, wie die Lügnerantinomie beweist. Im Anfang ist – oder war –logisches Chaos, Tohuwabohu. It’s a jungle out there. Und im Dschungel einen logischen Raum zu errichten – das ist die große, die übermäßige Aufgabe der Vernun . Fichte untersucht, wie die Vernun ihre Aufgabe löst: Sie setzt sich selbst als absolutes Ich, und sie setzt, den Widerspruch als gegeben anerkennend, sich ihr kontradiktorisches Gegenteil, ein Nicht-Ich, entgegen. Und dann versucht sie eine Synthese dieser beiden widersprüchlichen Seiten. Die Synthese gelingt auch, jedenfalls vorläu g; aber der Widerspruch tritt erneut an ihr auf und erfordert eine weitere Synthese und so fort. Wir kennen dieses Muster aus Hegels Logik; es ist das Grundmuster der synthetischen oder dialektischen oder spekulativen Philosophie. Auch diese aber versteht sich weiterhin, wie die analytische Philosophie, als theoretische Wissenscha . An der Antinomiethese wurde ein Grundproblem sichtbar, das in der analytischen Philosophie verdrängt wird und das zur synthetischen Philosophie motiviert. Es gibt ein weiteres Grundproblem, eines, das uns dazu bringt, auch die synthetische Philosophie noch hinter uns zu lassen. Dieses Problem hat mit der Natur des Raumes und der Zeit als schieren Stellenmannigfaltigkeiten zu tun. Die Stellen in Raum und Zeit unterscheiden sich nämlich voneinander nur dadurch, daß sie eben verschieden sind, was jedoch einer logischen Wahrheit, dem Prinzip der Identität des Ununterscheidbaren, zuwiderläu . Das Prinzip besagt: Wenn x dieselben Eigenscha en hat wie y, sind x und y ein und dasselbe. Nehmen wir an, x und y seien zwei qualitativ identische Billardkugeln. Wenn dann auch noch beide in denselben Beziehungen zu anderen Dingen stehen, beide etwa jeweils mitten auf einem Billardtisch liegen, der jeweils mitten in einem Raum steht, der sich jeweils im Erdgeschoß eines Hochhauses be ndet, das … usw., dann muß es sich am Ende um ein und dieselbe Billardkugel handeln. Es lassen sich aber dank der Natur des Raumes und der Zeit leicht Welten denken, in denen das Prinzip der Identität des Ununterscheidbaren scheinbar verletzt wird, etwa Welten mit ewiger Wiederkehr des Gleichen, in denen jedes Ding exakte Duplikate in anderen Weltepochen hat, wobei diese Epochen ihrerseits Duplikate

Metaphysik bei Hegel

313

voneinander sind. Bei ewiger Wiederkehr seit einem Anfangszeitpunkt wäre zwar eine Weltepoche die erste, eine andere die zweite, wieder eine andere die dritte usw., und dadurch wären sie unterschieden. Bei ewiger Zweibahnwiederkehr – unendlicher Wiederholung in beiden zeitlichen Richtungen – aber gäbe es keine erste Epoche, von der aus man zählen und die Epochen durch ihren Platz in der Reihe unterscheiden könnte. Also wäre ein Ding von seinen Duplikaten hier vollkommen ununterscheidbar und nach Voraussetzung dennoch von ihnen numerisch verschieden – entgegen dem Prinzip der Identität des Ununterscheidbaren. Will man die Zweibahnwiederkehr nicht ad hoc aus dem Reich des Möglichen ausschließen, so muß man etwas benennen, das für alle Welten die Identität des Ununterscheidbaren garantiert. Das Gesuchte ist auch leicht gefunden, sobald wir an unseren eigenen Fall denken: Wenn die Spree unbekannte Duplikate in anderen Weltepochen hätte, wären diese zwar nicht für einen göttlichen, außerweltlichen Blick von nirgendwo, wohl aber für meinen endlichen, innerweltlichen, perspektivischen Blick unterschieden. Von mir aus gesehen ist die Spree der Fluß, den ich kenne, und von meiner Gegenwart aus kann ich in die Vergangenheit und in die Zukun abzählen: Es gibt das vorige und das nachfolgende, das vorvorige und das nachnachfolgende Spreeduplikat usw. Dann aber gehört der Standpunkt endlicher innerweltlicher Subjekte zu allen Dingen in allen Welten, und in jeder möglichen Welt gibt es verkörperte endliche Subjekte, die sich denkend auf sich und ihre Umwelt beziehen. Die Dinge verlangen es so. Diese Behauptung nenne ich die Subjektivitätsthese. Mein Einwand gegen die Hegelsche Philosophiekonzeption lautet nun, daß Hegel zwar die Antinomiethese, nicht aber die Subjektivitätsthese in ihrem vollen Gewicht anerkannt hat. In der schwachen Fassung, in der ich sie soeben formuliert habe, ist allerdings auch die Subjektivitätsthese mit dem Ideal der Transparenz und dem Programm einer theoretischen Wissenscha des logischen Raumes vereinbar. Die schwache Version besagt (um es zu wiederholen): „In jeder möglichen Welt gibt es endliche Subjekte“, oder kürzer: „Verkörperte Subjektivität ist metaphysisch notwendig“. Dies ist mit Hegels Lehre kompatibel bzw. folgt sogar aus ihr. Die eigentliche Formulierung und starke Version der These sollte aber anders lauten, etwa so: „Die Welt ist wesentlich bezogen auf je mich als körperliches, endliches Subjekt in der Welt.“ Heidegger hat unabhängig von den hier skizzierten Begründungen die Antinomiethese und die Subjektivitätsthese anerkannt, erstere, wenn er in Sein und Zeit davon spricht, daß wir in der Grundstimmung der Angst vor das Nichts gebracht sind, welches selbst „nichtet“ als ein widerspruchsvoller Grundsachverhalt, und letztere in ihrer starken Fassung, wenn er, ebenfalls in Sein und Zeit, den Anspruch preisgibt, sein Philosophieren sei theoretische Wissenscha . Die theoretische Wissenscha kann Indexikalität und Standpunktgebundenheit zwar the-

314

Koch

matisieren, soll es aber von dem Standpunkt der Standpunktfreiheit und in einem quasigöttlichen, außerweltlichen Blick von nirgendwo tun. Heidegger hingegen lehrt, daß erkennende (und erst recht handelnde) Subjektivität mit der Welt untrennbar verwachsen ist, und nennt sie deswegen nicht mehr Subjektivität, sondern „Dasein“. Er lehrt ferner, daß im Gegenzug die Dinge auf das Dasein (leibliche Subjektivität) in ihrer Mitte angewiesen sind und nur dank ihm aus absoluter Verborgenheit – Unsichtbarkeit sogar für ein (per impossibile) als allwissend vorgestelltes Wesen – ins Sein und Sichzeigen hervorgehen können. Während die Metaphysik – die dogmatisch-analytische wie auch die kritischsynthetische – als theoretische Wissenscha au ritt, wird in Sein und Zeit ein Schritt der Emanzipation der Philosophie von der Metaphysik vollzogen. Die Philosophie reiht sich ein unter die hermeneutischen Disziplinen in Anerkennung ihrer eigenen Perspektivität, stellt sich aber zugleich über die theoretischen Wissenscha en durch ihre These, das Reale sei wesentlich auf je meine endliche Perspektive inmitten des Realen bezogen. Wird es aperspektivisch betrachtet, so fallen wesentliche seiner Aspekte aus der Betrachtung heraus. Die mathematisch formulierten Theorien der Physik abstrahieren also nicht nur vom Geistigen, sondern sogar von Zügen der Natur selber, etwa vom phänomenalen, qualitativen Grün der Wiese oder Blau des Himmels. In der Relativitätstheorie lassen sich die Ergebnisse von Messungen, die ein Beobachter macht, mittels Transformationsgleichungen auf die Ergebnisse von Messungen abbilden, die Beobachter in anderen Bezugssystemen machen; und die Lehrsätze der Theorie (die Naturgesetze) gelten ohnehin in allen Bezugssystemen. Ein Analogon von Transformationsgleichungen gibt es nach der Subjektivitätsthese für unser Grundverhältnis zum Realen nicht. Die Jemeinigkeit verkörperter Subjektivität und damit auch die Alterität der je anderen sind unhintergehbar. Wir haben keine exakten Gleichungen, sondern nur Faustregeln für die Abbildung der verschiedenen Standpunkte aufeinander. Dieses Abbilden ist kein Berechnen, sondern ein Übersetzen, kein Erklären, sondern ein Verstehen, keine Theorie, sondern Hermeneutik. Und die Hermeneutik bettet die Theorie ein. Denn der umfassende Blick von nirgendwo geht nicht nur über unser Vermögen als endlicher Subjekte, sondern ist eine inkohärente Fiktion, wie die Subjektivitätsthese belegt. Die Physik wird also mit ihrer Aufgabe der O enlegung der fundamentalen Naturgesetze nie ans Ende kommen können, sondern einen potentiell unendlichen Progreß von eine die andere überbietenden Nachfolgertheorien bilden müssen. In seiner Spätphilosophie ist Heidegger noch einen Schritt weiter gegangen und hat sogar den Anspruch preisgegeben, sein Denken sei überhaupt noch Wissenscha . Weder Forschen noch Dichten, weder Kunst noch Wissenscha , sondern etwas Neues soll es nun sein, für das freilich nicht weniger strenge Maßstäbe

Metaphysik bei Hegel

315

zu gelten hätten als für große Kunst und gediegene Wissenscha . Dies ist die vierte der erwähnten philosophischen Grundoptionen. Aber vermutlich gibt es diese Option gar nicht; vermutlich folgt Heidegger hier einem Phantom: Er will nach 2500 Jahren Metaphysikgeschichte einen „anderen Anfang“ des Denkens vorbereiten. Der erste Anfang bei Parmenides und Heraklit hat uns nach seiner tre enden Diagnose weggeführt von der vollen Erfahrung der Dinge über die Erfahrung von Gegenständen hin zur Erfahrung von nur noch bloßen Beständen des Realen. Ding, Gegenstand, Bestand – was hat es mit dieser Abfolge des näheren auf sich? Dinge „dingen“, das heißt, sie versammeln. Germanisch „thing“ ist die Versammlung, abgeleiteter Weise dann der Tagesordnungspunkt. („Das ist ein Ding!“: ein TOP für die Ratsversammlung.) Die Dinge versammeln, mit Heidegger zu reden, den Himmel und die Erde, die Göttlichen und die Sterblichen in das „Geviert“. Man nehme beispielsweise (oder denke an) einen Krug, einen Krug Wein.¹ Im Wein sind die Sonne und der Regen des Himmels mit den Mineralien der Erde für den Genuß durch die Sterblichen und, sofern der Wein zu sakralen Zwecken ausgeschenkt wird, mit den Göttlichen versammelt. – Ein Gegenstand ist gegenüber dieser Dingfülle nur noch das, was als Inhalt der Vorstellung einem vorstellenden Subjekt gegenübersteht. Immerhin wird im Begri des Gegenstandes noch ein wesentlicher Bezug des Realen zur verkörperten Subjektivität, dem menschlichen Dasein, mitgedacht. – Vollends zum bloßen Bestand werden die Gegenstände dann aber unter der Herrscha der Technik. Wir sahen schon, wie unter einer Carnapschen Zustandsbeschreibung die Welt als ein Maximalbestand von Realem katalogisiert und gleichsam bestellbar gemacht wird. So triumphiert im modernen Szientismus die vollständige Trivialisierung des Realen. Darin liegt die höchste Gefahr, aus der Rettung not täte. Wenn eine Programmformel gesucht wird für Heideggers Entfaltung des Standpunktes der Endlichkeit, so wüßte ich nichts Passenderes zu nennen als das Heraklitische physis kryptesthai philei, die Physis liebt es, sich zu verbergen.² Sie verbirgt sich in ihrem Sichzeigen und zeigt sich in ihrem Sichverbergen. Es gibt keinen vollständigen Überblick über sie, weil sie an ihr selber so verfaßt ist, daß sie sich nur für endliche Standpunkte innerhalb ihrer ö net, und dies dann nie vollständig, sondern jeweils um den Preis eines Entzugs. Zum Sein gehört zwar Seinsverständnis, nicht aber die Möglichkeit, sondern vielmehr die Unmöglichkeit eines totalen Überblicks über das Seiende.

1 Martin Heidegger Gesamtausgabe, hg. v. F. W. Herrmann, Frankfurt am Main 1977 . (GA). Bd. 7, 168 . 2 Heraklit. Fragmente, hg. v. B. Snell, München/Zürich 1989, (Fr.). B 123.

316

Koch

Sein und Erkennbarkeit gehören demnach zwar wesentlich zusammen, aber sie können grundsätzlich nicht zur Deckung kommen. Sie stehen in einem Wechselverhältnis, d.h. in einem Verhältnis wechselseitiger wesentlicher Abhängigkeit von wohlunterschiedenen Gliedern. Keines der Glieder ist, was es ist, wenn nicht die anderen Glieder sind, was sie sind; dennoch kollabieren sie nicht unterschiedslos in eines, kommen auch nicht miteinander zur Deckung, sondern bleiben spannungsvoll als Unterschiedene gegeneinander stehen. Heraklit ist der Philosoph der Wechselverhältnisse; sie sind, so lesen wir bei ihm: „Ganzes und Nichtganzes, Einträchtig-Zwieträchtiges, Einstimmend-Mißstimmendes, und aus Allem Eins und aus Einem Alles“ (Fr. B 10). In ihnen waltet eine unsichtbare Harmonie (harmoniê aphanês),die stärker ist als alle sichtbare Harmonie (Fr. B 54). Der Kern eines Wechselverhältnisses, das eigentlich Versammelnde, bleibt dunkel, opak. Der aktive Kern des tiefsten aller Wechselverhältnisse, das alles zu einem versammelt, aber ist der Logos (Fr. B 2, B 50). Wenn Heraklit der Denker opaker Wechselverhältnisse und der sich entziehenden Physis ist, so könnte man demgegenüber in Parmenides den Denker der transparenten Identität erblicken, von dem sich die Tradition des Ideals der Transparenz und der Metaphysik als einer theoretischen Wissenscha herleitet. Heidegger allerdings vereinnahmt auch Parmenides für seinen eigenen Heraklitismus und liest das bekannte Fragment 3 des Parmenideischen Lehrgedichts: to gar auto noein estin te kai einai („das nämlich Selbe ist Denken sowohl als auch Sein“), ganz in diesem Sinn: „Denken und Sein gehören in das Selbe und aus diesem Selben zusammen“ – ohne jedoch restlos zur Deckung zu kommen (GA, 11, 36). Dies ist die Weise der Wechselverhältnisse, deren Selbes und Versammelndes dunkel bleibt. Für den allwissenden Gott der theologischen und philosophischen Tradition lassen sich dann kein Raum und keine Funktion mehr nden. Nehmen wir zum Beispiel die Leibnizsche Monadologie: Jede gescha ene Monade spiegelt das gesamte Universum aus einer individuellen Perspektive und damit zu guten Teilen dunkel oder verworren. Gott hingegen sieht das Universum in vollkommener Klarheit, Deutlichkeit und Adäquatheit, ohne Bindung an Perspektive oder Standpunkt. So aber kann das Universum nicht gesehen werden, wenn das Reale gar kein Korrelat möglicher Allwissenheit, sondern nur endlichen epistemischen Standpunkten zugänglich ist, als ein System von Wechselverhältnissen, dessen Grund sich wesentlich entzieht und das keine letzte Basis für reduktive wissenscha liche Erklärungen bereitstellen kann. Zwar könnte man Gott seinerseits als ein trinitarisches Wechselverhältnis denken. Aber so gedacht wäre Gott nicht nur uns, sondern sich selbst ein Deus absconditus; das Versammelnde seines Wesens bliebe ihm selber dunkel. Andererseits hält die Konzeption der Wechselverhältnisse und der sich entziehenden physis vielleicht eine erklärungskrä ige Antwort auf die Frage nach dem Grund der

Metaphysik bei Hegel

317

Inkarnation: „Cur Deus homo?“ bereit – Gott mußte Mensch werden, weil er sich nur von innerhalb der Schöpfung, von einem endlichen Standpunkt aus, auf diese beziehen konnte. Und nur von einem innerweltlichen Standpunkt aus war die Scha ung der Welt überhaupt möglich. Deswegen mußte die Schöpfung durch dieselbe göttliche Person erfolgen, als welche Gott in sie eintreten mußte: durch den Sohn. Aber ich bezwei e, daß die Theologie ein derartiges Theorieangebot von der Philosophie gerne annimmt.

4 Hegel oder Heidegger? Die Hegelsche Logik tritt als theoretische Wissenscha , d.h. vom Standpunkt der Standpunktlosigkeit auf. Beeindruckend an ihr ist, wie sie (angemessener als die Monadologie) das Endliche und Perspektivische anzuerkennen und einzubetten vermag, ohne sich je mit ihm gemein zu machen. Nehmen wir etwa die Lehre vom Wechselverhältnis. Zu Beginn der Begri slogik, im Abschnitt über den allgemeinen Begri , lesen wir über „die eigene Vermittlung des Begri es mit sich selbst, seine eigene immanente Re exion“ das Folgende: Diese Vermittlung, welche das Zufällige zunächst [in der Wesenslogik] zur Notwendigkeit erhebt, ist aber [in der Begri slogik] die manifestierte [zugänglich und transparent gewordene] Beziehung; der Begri ist nicht […] die Notwendigkeit als die innre Identität voneinander verschiedener und sich beschränkender Dinge oder Zustände, sondern als absolute Negativität das Formierende und Erscha ende, und […] so ist der Schein die Erscheinung als des Identischen.³

Hier wird präzise das Wechselverhältnis gefaßt als Notwendigkeit, und zwar näher als innere, verborgene, nicht manifestierte, nicht transparente Identität von endlichen, einander beschränkenden Gliedern. Diese dunkle Identität ist, als nur innerliche, zugleich eine nur äußerliche (Enz. § 140), nur ein notwendiger äußerer Zusammenhang von Verschiedenen. Mit dem Übergang vom Wesen zum Begri aber ist nach Hegel eine neue logische Situation eingetreten: Die innere Identität und ihre äußerliche Daseinsweise als notwendige Verknüpfung endlicher Glieder ist umgeformt und verklärt worden zur Freiheit des Begri s, zur manifestierten Identität; das Identische ist nicht mehr verborgen, die dunkle Innerlichkeit des Wesens ist avanciert zur Erscheinung des Wesens als des Identischen.

3 G. W. F. Hegel. Gesammelte Werke, in Verbindung mit der Deutschen Forschungsgemeinscha hg. von der Rheinisch-Westfälischen Akademie der Wissenscha en. Hamburg 1968 . Wissenscha der Logik II (WdL II), Bd. 12, 242.

318

Koch

Hegel betrachtet demnach den Übergang von Heraklit zu (Heidegger möge verzeihen) Parmenides, vom Wechselverhältnis, in welchem die Physis als „das Lichtscheue“ (WdL II 183) sich zu verbergen liebt, zum transparenten Sichdurchdringen des Begri s, als einen logischen Fortschritt. Das verleiht seiner Theorie scheinbar einen überlegenen Standpunkt; scheinbar kann sie dem Wechselverhältnis und der inneren, verborgenen bzw. äußeren, als Notwendigkeit au retenden Identität vollkommen Rechnung tragen. Da ferner der logische Fortschritt nicht starr auf zeitlichen Fortschritt abbildbar ist, werden wir zu allen Zeiten Fälle der logisch überholten inneren, dunklen Identität nden. Deswegen könnte Hegel sich großzügig gegenüber Heidegger geben und das, was dieser das Dingen des Dinges nennt, als logisch überholte, aber faktisch noch vorkommende symbolische Versammlung ins vierfältige Wechselverhältnis von Himmel und Erde, Göttlichen und Sterblichen gelten lassen: Es dingt eine Mühle im Schwarzwälder Tal. Doch von Heideggers Warte ist der theoretische Triumph bloßer Schein. Es gibt keinen logischen Standpunkt, der den des Wechselverhältnisses überbieten, einbetten, im Hegelschen Sinn au eben könnte. Und darin, glaube ich, hat Heidegger recht. Wenn es einen höheren Standpunkt als den der Endlichkeit und des Wechselverhältnisses gibt, so ist es kein umfassender, einbettender oder au ebender Standpunkt, sondern allenfalls eine radikale und überraschende, philosophisch nicht ableitbare Alternative zum Standpunkt der Endlichkeit. Die Argumente sind nach meiner Überzeugung auf Heideggers Seite, auch wenn er persönlich vornehm gegen philosophische Argumente tut. Zu denken ist besonders an Argumente für die Subjektivitätsthese, deren starker Version zufolge, wie schon gesagt, alles Reale in einem Wechselverhältnis zu je meiner unhintergehbaren, endlichen Subjektivität steht. Wenn es eine Überwindung von deren Endlichkeit gibt, so kann sie nicht a priori als notwendig abgeleitet, sondern allenfalls als möglich beschrieben werden, und muß sich faktisch einstellen als eine unvorhersehbare freudige Überraschung. So könnte das konsequente sich Stellen auf den Standpunkt der Endlichkeit die Voraussetzung dafür scha en, daß er ex improviso überwunden werden kann, etwa in Zuständen von der Art jener Kaskade von Wieder ndungserlebnissen, die Proust im letzten Band der „Recherche“, „Le temps retrouvé“, beschreibt. Ganz trägt die Analogie nicht, denn die Proustsche Erleuchtung ist so konzipiert, daß sie jeden tre en könnte, gleichviel auf welchem Standpunkt er (oder sie) steht. Doch auf dem Hegelschen Standpunkt, dem der Unendlichkeit in der Theoriebildung, wird man das Desiderat einer noch tieferen, auch praktischen und damit erst wirklich beglückenden Einsicht bestreiten müssen. Auf dem Standpunkt der Endlichkeit kann man ihre Möglichkeit immerhin theoretisch entwerfen und als

Metaphysik bei Hegel

319

Lebensdesiderat anerkennen, wenn auch als eines, das auf dem Wege der wissenscha lichen Theoriebildung, des „begründenden Vorstellens“, nicht zu erfüllen ist. Aber in einem anderen Punkte behält umgekehrt Hegel recht gegen Heidegger. Seit Jahrzehnten nimmt die Globalisierung desjenigen Denkens, mit dem die Griechen einst dem Sein entsprachen, einen mühelosen, ungebremsten Lauf, auch wenn (oder vielleicht gerade weil) jenes Denken längst umgeformt ist in ein Rechnen mit bestellbaren Beständen. Die moderne Technik und die mathematischen Naturwissenscha en, deren sie bedarf, gehen ungehindert durch alle politischen, kulturellen und religiösen Absperrungen hindurch. Wo, wenn nicht hier, am unau altsam sich verbreitenden Bestandsdenken, haben wir es mit einem kulturinvarianten menschlichen Universale zu tun? Dann aber war schon das frühe Seinsdenken, das sich heute in der Technik vollendet, ein kulturinvariantes menschliches Universale, in dessen Bereich die Griechen nur zufällig als erste vorstießen. Dann wird es keinen anderen Anfang des Denkens und kein nachwissenscha liches Philosophieren geben können. Es gibt nur einen einzigen, seinem logischen Inhalt nach alternativlosen Anfang, ganz wie Hegel es lehrt, einen Anfang, der faktisch vor 2500 Jahren an den entgegengesetzten Enden der griechischen Welt, in Ionien und in Unteritalien, von Heraklit und Parmenides, vollzogen wurde und in dessen Folge wir heute quer durch alle Kontinente herausgefordert sind, das Seiende als Bestand zu entbergen. Wenn dies – das sich ausbreitende Bestandsdenken – die Gefahr ist, wo wächst dann das Rettende? Heidegger wollte es darin erblicken, daß gerade die äußerste Trivialisierung des Realen uns vor die Frage nach dem Sinn des Seins bringt und so einem anderen Anfang den Weg bereitet. Wenn aber der erste Anfang alternativlos ist, was dann? Es müßte dann dieser alternativlose Anfang des Denkens neu auf seine emanzipatorischen Potentiale hin durchdacht werden in einem Philosophieren, das sich als Wissenscha versteht, aber seine wesentliche Indexikalität anerkennt, das demnach die analytische und die synthetisch-spekulative Philosophie einbettet und auf der anderen Seite Heideggers späte Entdeckungen in die hermeneutische Wissenscha zurückholt. Die Antinomiethese und die Subjektivitätsthese sind dazu ein geeigneter Leitfaden.

Gary Hat eld

Russell’s Progress Spatial Dimensions, the From-Which, and the At-Which Kant’s critical philosophy brought space (and time) to the forefront of philosophical attention. Regarding space as the form of outer sense, his critical doctrines purported to establish that synthetic a priori knowledge is possible, knowledge that describes and constrains any possible human experience of space. The implication was that the very same form of intuition governs all appearances of spatial structure, including all possible sensory experience and the pure intuition of a priori imagination. During the nineteenth century, philosophers, sensory physiologists, and sensory psychologists responded to Kant’s doctrine of space in various ways.¹ Some took his doctrine to be in fact a psychological proposal about the innateness of spatial perception and sought to support or refute it on those grounds. Others believed that the doctrine was intended to explain the possibility of geometrical knowledge. Some philosophers, including Bertrand Russell, held that Kant was seeking to establish a necessary relation between geometry and physical space, founded upon space as a form of intuition. Russell held that Kant’s doctrine was refuted by the existence of non-Euclidean geometries, which made the question of the actual geometrical structure of physical space into an empirical matter.² There is a variety of opinion on the relation between Kant’s arguments that space is an a priori form of intuition and his theory of geometrical knowledge, which draws on that doctrine. Some interpreters contend that Kant’s arguments for the apriority and the intuitional nature of space are separate from his theory of geometry,³ while others nd them to be more closely related.⁴ In this literature, Strawson emphasized the “phenomenal interpretation” of Kant’s theory of geometry: that Kant was addressing, in the rst instance, the structure of space as experienced phenomenally and claiming that this space is of necessity described by Euclid’s geometry. Citing Russell, he reports that the progress of pure mathe-

1 Gary Hat eld, The Natural and the Normative: Theories of Spatial Perception from Kant to Helmholtz (Cambridge: MIT Press, 1990). 2 Bertrand Russell, The Principles of Mathematics, 2nd ed. (New York: Norton, 1938), 372; the rst edition was published in 1903. 3 Rolf-Peter Horstmann, “Space as Intuition and Geometry: Observations on Kant’s Transcendental Aesthetic,” in Ratio 18 (1976), 18–30. 4 P. F. Strawson, The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason (London: Methuen, 1966), 60, 66–7, 277.

322

Hat eld

matics and physics in the twentieth century rendered the phenomenal structure of space less important, because: on the one hand, rigour in mathematics turns on logic, not on looks – which is why Russell said that the development of mathematics showed Kant’s Anschauung to be super uous; and, on the other hand, the physical applications of geometry turn on physical tests and measurements of many kinds, and not on the mere contemplation of phenomenal appearances.⁵

Strawson accurately reports that Russell himself separated geometry as pure mathematics from the empirical determination of the geometry of physical space and that he believed that the rigorization of mathematics took it away from appeals to spatial experience. But Strawson underestimates the importance of phenomenal spatial appearances in Russell’s theory of knowledge and hence in any knowledge of physical space. I have rehearsed these themes concerning geometry and physical space in part so as to specify my topic more closely. I want to focus on a question in Russell that has to do not with the foundations of geometry or, to begin with, the speci c geometry of physical space, but rather with the relation between the sensory experience of the spatial properties of objects and knowledge of a public spatial order. Thus, although Strawson’s observation about the inadequacy of bare phenomenal looks may apply to speci c problems addressed in Russell’s philosophy of physics (viz., determining the precise structure of physical space), it does not hold with respect to his more general epistemology as applied to the spatial properties of things. Any claims about the spatial properties of objects in physical space, including measuring instruments, are for Russell based on perceptual experience. My topic concerns some aspects of Russell’s epistemological turn in the period a er 1911. In particular, it focuses on two aspects of his philosophy in this period: his attempt to render material objects as constructions out of sense data, and his attitude toward sense data as “hard data.” These aspects are epistemological and metaphysical. They involve commitments to the existence of certain sorts of entities, paired, not with the denial of other entities, but with a decision not to include other entities in the general scheme of being that Russell develops; and they involve a discussion of the sort of metaphysical commitments that would allow, as regards the objects around us, the positing of easily knowable entities as their basis. In considering this topic, I focus on Russell’s “breakthrough” of early 1914, in which he concluded that, viewed from the standpoint of epistemology and an-

5 Ibid., 286.

Russell’s Progress

323

alytic construction, space has six dimensions, not merely three. In this scheme, Russell posits a three-dimensional personal or “perspective” space that is inhabited by sense data. This space then forms the basis for constructing the threedimensional space of physics (and of public things). I am concerned with the speci cs of this construction: with the properties of the private spaces, the relations among those spaces, and their relation to physical space and to constructed “things,” such as pennies or tables. I nd that there are di culties of interpretation with respect to these relations, which stem from the di culty of nding a coherent interpretation of Russell’s claim that objects such as tables and pennies look smaller at a greater distance (or look trapezoidal or elliptical from some points of view). I don’t mean to challenge the phenomenal claim that objects do, in some sense, look small in the distance. Rather, I raise di culties with Russell’s analysis of this fact, in which he appeals to both phenomenal experience and the ndings of sensory psychology. I hold that if he wishes to maintain his phenomenal claim about objects appearing smaller with greater distance, he must alter or redescribe aspects of his construction of ordinary things. However, if his construction of things and physical space is based on a problematic description of the private spaces, then his claim that private or perspective spaces are very well known and provide the hard data for knowledge of the physical world faces a challenge. In the rst section, I review the place of sense data in Russell’s philosophical development a er 1911. I then consider: his “breakthrough”; the ndings of sensory psychology that he invoked in connection with it; some problems that arise, both for Russell and textbook psychology, in connection with the perception of size; and nally some problems for Russell’s use of this psychology in describing the “hard” sense data at the basis of his constructions, along with some directions in which he might seek a solution. I end with a brief consideration of why experientially based knowledge of spatial items is crucial for Russell in a way that it is not for Kant.

1 Russell and Sense Data The part of Russell’s metaphysics that is our focus concerns his e ort to render the material world, including the world of physical science, as a construction out of sense data. Russell worked at this project from sometime in 1912 through the Analysis of Matter in 1927. He did not sustain the same metaphysics throughout: indeed, within the range of years just indicated, he switched from endorsing a representative realism with inferred physical entities in Problems of Philosophy; to adopting a metaphysics of sense data and subjects who apprehend them later in

324

Hat eld

1912, according to which matter is constructed; to accepting the position that there exists only neutral stu , that is, to his accepting late in 1918 the neutral monism of William James.⁶ Russell’s e ort to construct physical objects and theoretical entities from sense data begins with the sense data themselves. From sometime in the rst decade of the century onward, Russell was a realist about sense data. In his autobiographical reconstruction of this period, he emphasizes the initial pluralistic realism that he shared with Moore: the real is not one, but many, and objects are known directly, as in naïve realism.⁷ In his writings on sense data from 1912–14, he stresses a second position that he shared with Moore, the mind-independence of sense data: that they exist independently of our perception of them, even while we perceive them.⁸ Accordingly, Russell distinguished sensation, as the mental act by which we are aware of sense data, from sense data as objects of which we are aware. Considering the question of whether the immediate objects of sense are mental entities or at least are in some sense mind-dependent, in Our Knowledge of the External World, the Lowell lectures in Boston from 1914, Russell held them to be physiologically conditioned but mind-independent: I think it must be admitted as probable that the immediate objects of sense depend for their existence upon physiological conditions in ourselves, and that, for example, the coloured surfaces which we see cease to exist when we shut our eyes. But it would be a mistake to infer that they are dependent on the mind, not real while we see them, or not the sole basis for our knowledge of the external world.⁹

In a paper from the same time, “The Relation of Sense-Data to Physics,” he asserted that sense data are “physical,” not mental, by which he meant that they

6 On Russell’s adoption of neutral monism, see Hat eld, “Sense-Data and the Philosophy of Mind: Russell, James, and Mach,” in Principia 6 (2003), 203–30, and “Sense Data and the Mind–Body Problem,” in Perception and Reality: From Descartes to the Present, ed. R. Schumacher (Paderborn: Mentis Verlag, 2004), 305–31. 7 Russell, My Philosophical Development (New York: Simon and Schuster, 1959), 61–2; G. E. Moore’s well-known “Refutation of Idealism,” in Mind n.s. 12 (1903), 433–53, in e ect endorses a naïve realism, according to which we are directly aware of “the existence of a table in space” in the same way in which we are directly aware of the conscious element that accompanies every mental fact (453). 8 Moore, in “The Nature and Reality of Objects of Perception,” in Proceedings of the Aristotelian Society 6 (1905–6), 68–127, distinguished “sense-contents” as non-mental from the mental acts by which we apprehend them; in “The Subject-Matter of Psychology,” in Proceedings of the Aristotelian Society 10 (1909–10), 36–62, he made the same point, now using the term “sense data.” 9 Russell, Our Knowledge of the External World as a Field for Scienti c Method in Philosophy (Chicago: Open Court, 1914), 64; herea er, OKEW.

Russell’s Progress

325

are mind-independent and did not mean that they are composed of the elementary particles of physics.¹⁰ Why posit sense data as distinct from ordinary external objects? Russell accepted the arguments from perceptual variation. The table appears rectangular from one view, trapezoidal from another; brown from one view, glaring whitish grey from another. The coin appears now elliptical, now circular, now larger, now smaller, depending on viewing angle and distance. We believe that the color and shape of the real table and coin do not change from moment to moment; but the “data” that we see do change. Russell concluded that sense data are not parts of the objects themselves, and he posited them as entities distinct from really existing physical objects.¹¹ But then in 1912–14 he came to believe that he could do away with really existing physical objects as usually conceived and achieve a more economical ontology involving only two types of entity: subjects and data. Ordinary objects and microphysical processes had, in Problems, been things known by inference from sense data. Now they would be regarded not as inferred entities but as logical constructions or indeed as logical ctions.¹² One of Russell’s motivations for retaining sense data and jettisoning physical matter was epistemological: sense data are very well known. In the language of the Lowell lectures, he described them as the “hard data” from which he would construct the “things” of common sense and science. In calling the thing a “mere logical construction” or comparing it to a “ ction,” he meant to say that it doesn’t exist, or rather it needn’t be held to exist.¹³ Accordingly, only these particulars exist: sense data; the subjective acts of their apprehension; and, most likely, the domain of sensibilia – entities very like sense data that exist at places where there isn’t anyone to see them at present. The sensibilia are not physiologically conditioned, which is why they are only “very similar” to sense data. The fact of physiological conditioning is an empirical regularity within our sense data: e.g., that staring at a bright light yields an a erimage that a ects our visual experience for several minutes and about which we can construct explanations in the ctional language of physics and physiology.

10 Russell, “Relation of Sense-Data to Physics” (herea er, RSDP), in Scientia 16 (1914), 1–27, as reprinted in Russell, Mysticism and Logic and Other Essays (London: George Allen and Unwin, 1963), 108–31, on p. 112. 11 Russell, The Problems of Philosophy (New York: Oxford University Press, 1997), 12, 22–5; the rst edition was published in 1912. 12 OKEW, vi; RSDP, 115. 13 Russell, OKEW, 70–2, 89; “Ultimate Constituents of Matter” (herea er, UCM), in Monist 25 (1915), 399–417, as reprinted in Mysticism and Logic, 94–107, on p. 102; RSDP, 115.

326

Hat eld

If sense data are the “hard data” Russell takes them to be, the well-known items that are the basis for all other knowledge of the external world, then he should be clear about their phenomenal properties. For vision, these properties are colors, shapes, distances, and sizes. Russell worked on the structure of sense data and their spatial relations to one another in 1912–14. In January, 1914, he had what he considered a breakthrough, which caused him to emend the manuscript of the Lowell Lectures.¹⁴ This breakthrough led him to the notion of a six-dimensional spatial construction, which related the private spaces of sense data to a constructed public space that could be known by all. I want to explore the spatial structure especially of the private spaces, as these are the (ostensibly) well-known data at the foundation of Russell’s system.

2 Russell’s “Breakthrough” and Six-Dimensional Space In Problems, Russell o ered a familiar description of sense data according to which, from most normal points of view, a table is seen as having a trapezoidal shape; or rather, if we carefully attend to our sense data of the table, we will notice that they are trapezoidal: We are all in the habit of judging as to the “real” shapes of things, and we do this so unreectingly that we come to think we actually see the real shapes. But, in fact, as we all have to learn if we try to draw, a given thing looks di erent in shape from every di erent point of view. If our table is “really” rectangular, it will look, from almost all points of view, as if it had two acute angles and two obtuse angles. If the opposite sides are parallel, they will look as if they converged to a point away from the spectator; if they are of equal length, they will look as if the nearer side were longer. All these things are not commonly noticed in looking at a table, because experience has taught us to construct the “real” shape from the apparent shape, and the “real” shape is what interests us as practical men.¹⁵

There are two things to notice about this passage. First, Russell describes what he otherwise calls the sense data (here: appearances) of the table as if they were perspective projections; he describes the apparent shape as trapezoidal and compares it to the shape of the table as we would draw it. He does not indicate that the sense datum is experienced at a distance, but he does not say it is experienced internally or in the eye, either, and I believe the presumption is that the trapezoidal shape is experienced roughly in the location of the table. Second, he tacitly

14 Kenneth Blackwell, “Our Knowledge of Our Knowledge,” in Russell: The Journal of Bertrand Russell Studies 12 (1973), 11–13. RSDP records the breakthrough. 15 Russell, Problems, 10–11.

Russell’s Progress

327

appeals to psychological theories of shape perception, according to which, as a result of experience, we perceive (or think that we perceive) the table top as a rectangle in its proper location.¹⁶ Russell’s position in Problems was subject to two sorts of objection. One type of objection was more generally abroad concerning sense data, and has become known as Stout’s postulate, according to which, as described recently by Nasim, “one and the same thing cannot have more than one sensible quality at one and the same place” (at the same time, one presumes).¹⁷ If four people view the same rectangular table from positions standing back from each of the four sides, they see four di erent trapezoidal sense data; even if the trapezoids seen from either end should by coincidence be congruent (the viewers have the same height and are in exactly similar positions), they show di erent sides of the table as being longer and shorter (for each, the near side is the longer). If one assumes that they are seeing these data as being at the position of the real table, then the data violate Stout’s postulate. But T. P. Nunn replied that, if the various sense data are in the private spaces of distinct observers, then they are not literally together occupying the same location at the same time. In this regard, the public object must be seen as a construction based on the private elements. We shall see that Russell, in his “breakthrough,” draws from Nunn’s suggestion (and he in fact cites Nunn).¹⁸ The second objection was conveyed to Russell by Whitehead, as part of a fourteen-page critical response to a prepublication manuscript version of Problems. Whitehead wrote to his former pupil: “As to the ‘shape of the table’. Why assume that our perception of space is two-dimensional? Perhaps you don’t. I can’t get a decisive instance just now. But the general impression on my mind is

16 G. E. Stout, Analytical Psychology, 2 vols. (London: Swan Sonnenschein, 1896), 2:21–3, o ers a position similar to Russell’s, asserting that we merely think of the real shape, without forming an image of it, which goes along with Russell’s statement that “we come to think we actually see the real shapes,” which might suggest that we don’t actually see them (form an image of them). W. James, Principles of Psychology, 2 vols. (New York: Holt, 1890), 2:238–40, by contrast, suggests that we habitually replace the sensation of a projective shape with an image of the real shape seen from an ideally revealing point of view (so as to capture symmetry and equality of sides, if these exist). This could be Russell’s position if we think that the “construction” of the real shape is a phenomenal construction. Russell read these works by Stout and James in 1894–96; see Russell, “What Shall I Read?” in Cambridge Essays, 1888–99, ed. K. Blackwell et al. (London: George Allen and Unwin, 1983), 345–70, on pp. 354–5, 357. 17 Omar W. Nasim, Bertrand Russell and the Edwardian Philosophers (Houndmills, England: Palgrave Macmillan, 2008), p. 6. 18 RSDP, 113, citing T. P. Nunn, “Are Secondary Qualities Independent of Perception?” in Proceedings of the Aristotelian Society 10 (1909–10), 191–218.

328

Hat eld

that you do. Surely such an assumption is false psychology.”¹⁹ Whitehead assumes from Russell’s description that the sense data are two-dimensional, but he also correctly points out that there is no passage from which it can be seen that Russell decisively held this position (at least there is none to be found in the published version of Problems). All the same, in the six-dimensional construction of the “breakthrough,” Russell explicitly renders his private spaces as threedimensional. He also may have gleaned an idea from Whitehead’s materials intended for the fourth volume of Principia Mathematica (but never published in the form Russell saw them, and later destroyed),²⁰ involving the distinction between the “from which” and the “at which.”²¹ Russell’s breakthrough from January, 1914, is recorded in the paper “The Relation of Sense-Data to Physics” (written in the rst week of January). Russell responded to the above problems with his six-dimensional account of space. Russell’s six-dimensional space consists of two three-dimensional spaces: a private one and a public one that is constructed from it.²² The private spaces are each experienced by an individual. These spaces have a three-dimensional structure that constitutes a “perspective” or point of view. As Russell explains, “two places of di erent sorts are associated with every ‘sense-datum’, namely the place at which it is and the place from which it is perceived.”²³ The from-which may be thought of as the phenomenal position of the subject or percipient, the place from which the subject perceives the world. The at-which is the location of the datum,

19 Victor Lowe, “Whitehead’s 1911 Criticism of The Problems of Philosophy,” in Russell: Journal of the Bertrand Russell Archives 13 (1974), 3–11, on p. 5. 20 Ibid., p. 3. Some of the ideas contained in these papers were published by Whitehead in 1919 and 1920, on which, see Elizabeth Ramsden Eames, Bertrand Russell’s Dialogue with His Contemporaries (Carbondale: Southern Illinois University Press, 1989), 121. Famously, in the preface to OKEW (p. vi), Russell credited Whitehead with his changed position from Problems, saying: “I owe to him the de nition of points, the suggestion for the treatment of instants and ‘things,’ and the whole conception of the world of physics as a construction rather than an inference.” As is well known (see, e.g., Eames, Russell’s Dialogue, 103, 118–21), Whitehead was not pleased with the use to which Russell put these ideas, and he refused to send Russell any more of his “notes” and asked him not to discuss his ideas (in a letter from 1917). 21 A distinction similar to Russell’s between the from-which and the at-which appears in A. N. Whitehead, An Enquiry Concerning the Principles of Natural Knowledge (Cambridge: Cambridge University Press, 1919), 85, 189, and The Concept of Nature (Cambridge: Cambridge University Press, 1920), ch. 5. The distinction is not worked out in such a way that it adds to Russell’s version, and the relation between these later publications and Whitehead’s earlier “notes” is not clear. 22 RSDP, 119: “The world which we have so far constructed is a world of six dimensions, since it is a three-dimensional series of perspectives, each of which is itself three-dimensional.” 23 RSDP, 117.

Russell’s Progress

329

perceived to be at a certain distance (the space is three-dimensional) and having a shape and size as experienced from the from-which. As Russell elaborates: each person, so far as his sense-data are concerned, lives in a private world. This private world contains its own space […] . The place at which a sense datum is, is a place in private space. This place therefore is di erent from any place in the private space of another percipient.²⁴

These private spaces would seem to speak to both of the problems from Problems. They obviate, in Russell’s words, the problem “of combining what we call di erent appearances of the same thing in the same place,”²⁵ since the place of each at-which occurs only in the private space of the individual. Further, Russell here makes clear that the private spaces are three dimensional; the relation between the from-which and the at-which must be supposed to include a distance at which the datum is seen. Although not crucial for our purposes, Russell augments the actual private spaces of individuals with additional locations, from-whiches, that correspond to the points of view from which a sense datum would be observed if anyone were there; in these cases, the postulated entity located in the at-which is called a “sensibile” and amounts to an unsensed sense datum (occurring minus the e ects of any physiological conditioning, as can accrue with actual perceivers).²⁶ The second three-dimensional space is the public or physical space of “perspectives” or points of view. It is essentially an ordering of the from-which locations, that is, the places from which each perceiver experiences a datum. He introduces this space as follows: In addition to the private spaces belonging to the private worlds of di erent percipients, there is, however, another space, in which one whole private world counts as a point, or at least as a spatial unit. This might be described as the space of points of view, since each private world may be regarded as the appearance which the universe presents from a certain point of view. I prefer, however, to speak of it as the space of perspectives, in order to obviate the suggestion that a private world is only real when someone views it.²⁷

The various from-whiches, or perspective units, can, according to Russell, be ordered so as to produce “the one all-embracing space of physics.”²⁸ To each po-

24 RSDP, 117. Sajahan Miah, Russell’s Theory of Perception, 1905–1919 (London: Continuum, 2006), 158, implausibly makes the “at-which” a location in public space. Rather, locations in public space are constructed from the from-whiches. 25 RSDP, 118. 26 RSDP, 110–11. 27 RSDP, 118. 28 RSDP, 119

330

Hat eld

sition relative to a table, or, more conveniently, to a penny, an entire domain of private worlds can be imagined, each addressing the penny from a location so as to yield an appearance of a speci c size and shape. These sizes and shapes are similar enough to be compared among perceivers through language,²⁹ and they therefore can serve as a basis for ordering the positions from-which into a set of locations that de ne the penny considered as a “public” object. Once these locations are determined, the notion of the penny as a real material thing is no longer of use, and it is reduced to the class of its appearances. For our purposes, the crucial aspect of the construction of public space and material things is that it is accomplished, in Russell’s view, by an ordering of private spaces according to their similarities and relations to one another. Russell explains the process: The arrangement of perspectives in a space is e ected by means of the di erences between the appearances of a given thing in the various perspectives. Suppose, say, that a certain penny appears in a number of di erent perspectives; in some it looks larger and in some smaller, in some it looks circular, in others it presents the appearance of an ellipse of varying eccentricity. We may collect together all those perspectives in which the appearance of the penny is circular. These we will place on one straight line, ordering them in a series by the variations in the apparent size of the penny. […] By such means, all those perspectives in which the penny presents a visual appearance can be arranged in a three-dimensional spatial order.³⁰

It must be remembered that there is no material penny, in the usual sense. The penny is the class of its appearances. In considering a material penny as a ctive construction, we may hold that “the matter of a thing is the limit of its appearances as their distance from the thing diminishes.”³¹ Talk of the penny “appearing” in the perspectives is a way of characterizing the class of sense data that are all “as of” a single material thing. This classi cation must be able to be achieved simply by comparing sense data empirically. Whether this project is workable in general is not my present concern.³² Assuming for the sake of argument that the appearances can be ordered as he says, I want to look into the speci c structure of the private spaces. In doing so, I will raise some questions about the coherence of this description, questions that were available within the conceptual framework in the

29 RSDP, 117. 30 RSDP, 118–19. 31 RSDP, 121. 32 Miah, Russell’s Theory of Perception, ch. 7, collects many objections to Russell’s constructivist project.

Russell’s Progress

331

psychology of perception of his time, and which apply more generally to the notion of the penny looking smaller at a greater distance and looking elliptical from an angle.

3 Russell among the Psychologists In ordering the sense data of the penny, Russell appeals to their “apparent sizes.” The smaller the apparent size (as of the same penny), the farther away the fromwhich from the constructed location of the penny. Russell might also have appealed to their apparent shapes, to the e ect that, the greater the eccentricity of the ellipse, the farther the from-which is (as regards viewing angle) from a line perpendicular to and through the center of the upward facing disc of the penny. Russell also o ers a di erent description of the series of phenomenal sizes. Instead of equating the series with apparent sizes, he relates them to the amount of visual eld taken up: When a number of people are said to see the same object, those who would be said to be near to the object see a particular occupying a larger part of their eld of vision than is occupied by the corresponding particular seen by people who would be said to be farther from the thing. By means of such considerations it is possible, in ways which need not now be further speci ed, to arrange all the di erent spaces in a three-dimensional series.³³

While this may seem equivalent to talk of apparent size, it is not. Occupying a larger part of the visual eld is a matter of the visual angle an object subtends. Visual angle does vary as Russell says: for an object of constant size viewed head on (as with the penny), the object will take up more of the visual eld the closer the vantage point (or from-which) is to the object. But, as discussed in the next section, the perceived size of an object is a function of both visual angle and the perceived distance to the object. Both of Russell’s descriptions connect with the language used by psychologists, and indeed in his discussion of “hard data” in Our Knowledge, he makes clear his dependence on the ndings of psychology. He observes that: Psychologists […] have made us aware that what is actually given is much less than most people would naturally suppose, and that much of what at rst sight seems to be given is really inferred. This applies especially in regard to our space-perceptions. For instance, we instinctively infer the “real” size and shape of a visible object from its apparent size and shape, according to its distance and our point of view.³⁴

33 UCM, 103–4. 34 OKEW, 68.

332

Hat eld

Here again we have talk of apparent size and shape, and now distance is brought in as something that is added to what Russell may well have considered to be a twodimensional sense datum. Russell notes that the “real” size “at rst sight seems to be given,” suggesting that the real size is in some sense phenomenally present; but it remains unclear whether he still holds the position from Problems, according to which we merely “come to think we actually see the real shapes” and presumably sizes, a phrasing which allows that we merely mistakenly believe that we see the real size. The above passage occurs in the portion of the Lowell lectures thought to have been composed before the breakthrough in January, 1914.³⁵ The part inserted a er the breakthrough goes through the six-dimensional space, using much the same language as “Relation of Sense-Data to Physics.” Russell’s description of the appearances of the penny accords with some descriptions of apparent size and shape in a psychological account with which he was familiar. G. E. Stout, whose psychology textbook Russell read and who was one of his teachers at Cambridge, o ers a position that is similar to Russell’s prebreakthrough discussion in Our Knowledge. He reports that in visual perception we actually experience a visual magnitude that accords with visual angle and a visual shape that accords with the projective shape of the object as seen from a particular point of view. We then infer the real size and forget the apparent size. But, interestingly, he asserts that this real size is not present to us phenomenally. Here is what Stout says: If we attend to the visual appearance rather than to the object, we nd that it varies in size according to our distance from the thing seen. At great distances this variation is forced upon our notice. Looking down from a high tower on men walking in the streets of a town, we remark with a kind of surprise how small they look. But for comparatively short distances the variation passes unnoticed unless we expressly measure the visual angle. The reason is that the visual magnitude, as such, is habitually disregarded, serving only as a sign of the real magnitude. But this real magnitude, in its distinction from apparent magnitude, is neither actually seen nor mentally visualized. Nor is it represented by an ideally revived tactual or motor complex. There is no trace of such a complex in consciousness.³⁶

There are three points to note. First, Stout a rms a phenomenology of things looking smaller at a distance, according to visual angle. Second, he remarks that, such appearances are overlooked in favor of the real magnitude (by a process he has characterized as an inference). Third, he reports that this real magnitude is not present in consciousness as a visual content, that is, as a phenomenal spatial extent that corresponds to the real magnitude. These positions accord at least

35 Blackwell, “Our Knowledge”, 13. 36 Stout, Analytical Psychology, 2:22–3.

Russell’s Progress

333

partially with Russell’s pre-breakthrough report of what “the psychologists” have revealed. A potential discrepancy is that Stout suggests that the men are both seen at a distance and look small, while Russell in the passage above writes that, when we take distance into account, we get to real size. As it happens, the relations among visual angle, apparent size, distance, and real size are a problem for both Stout and Russell, and could have been seen to be so on the basis of other psychological literature produced in England, some at Cambridge, during the latter decades of the nineteenth century and the beginning of the twentieth century.

4 Problems with Perceived Size By Russell’s day, the relations among visual angle, apparent size, distance, and real size had long been studied in the optical tradition and were under intense investigation in sensory psychology. From the time of Ibn al-Haytham and Descartes, it was known that, for adult perceivers, perceived size depends on both visual angle and perceived distance.³⁷ Visual angle varies inversely with distance. The perception of size involves both visual angle and perceived distance. Consider an example in which a smaller object at a closer distance takes up the same visual angle as a larger object at a farther distance (Fig. 1a). Assuming that visual angle is accurately recorded by the visual system³⁸ and that distance is accurately perceived, both objects are perceived with their true sizes (a perceptual result later known as “size constancy”). If two objects of di erent size are at the same distance from the eye, then the smaller one subtends a smaller visual angle (Fig. 1b); if the distance is accurately perceived, then the sizes of the two objects are accurately perceived. A further factor arose with Kepler’s discovery of the retinal image. Visual angle can be de ned by relations among directions in the eld of view. Once Kepler accurately characterized the optics of the retinal image, it became known that the visual angle subtended by an object corresponds to the size of the image that the

37 See Gary Hat eld and William Epstein, “The Sensory Core and the Medieval Foundations of Early Modern Perceptual Theory,” in Isis 70 (1979), 363–84. 38 In what follows, I assume that the visual system is basically accurate in recording visual angles, or visual directions. This is a separate matter from whether perceivers are good at judging visual angles. It simply assumes that we see things in the direction that they actually are. When both eyes are open, especially for near distances visual angle is constructed in relation to the “cyclopean eye.” For a discussion, see Maurice Hershenson, Visual Space Perception: A Primer (Cambridge: MIT Press, 1999), ch. 2.

334

Hat eld

(a)

P

(b)

A

P

A

P

B

B

Fig. 1: Part (a) shows how two objects of di erent sizes (A and B) can subtend the same visual angle when placed at di erent distances from the perceiver P. Assuming that visual angle (or visual direction) is accurately registered by the visual system, then, if distance is perceived veridically, A and B are perceived with their true sizes. Part (b) shows how the same two objects relate to the perceiver when they are the same distance away. The smaller object (A) subtends a smaller visual angle. Again, if distance is perceived veridically, then both A and B are perceived with their true sizes.

object projects onto the retina. The larger the retinal projection, the larger the visual angle. This relation includes the previous one: a larger and a smaller object at the same distance have larger and smaller retinal projections, but if distance is accurately perceived their di ering sizes are accurately perceived; further, they can be placed in such a way that they project the same size onto the retina and, again, if distance is accurately perceived, the objects are perceived with their true sizes. Kepler’s nding led to the question of whether perceivers are aware of retinal image size and also of the shapes that objects project onto the retina. In the case of the penny, the retinal image size varies inversely with distance, so that, when farther away, the penny projects a smaller size. Similarly, when viewed straight on, the penny projects a circle, but when viewed from an angle, as this angle diverges

Russell’s Progress

335

from straight on, the penny projects ellipses of increasing eccentricity (until the penny projects an elongated rectangle, when viewed from the side).³⁹ These facts thus far only specify the sizes and shapes of images projected on the retina. Except when the retinal values are combined with a perception of distance, these facts don’t specify a visual experience of an object with a shape and size at a distance. Theorists varied in their interpretation of what sorts of experiences perceivers have of projective values. Some theorists spoke of perceiving or having sensations or visual appearances that correspond to the retinal projections: for a given object, it is smaller in sensation at greater distance, and its shape in sensation varies with viewing angle. But the question of where these sensations appear to be became problematic. Berkeley placed them “in the mind,” Reid had them at “no distance from the eye.” Stout didn’t specify.⁴⁰ Others, such as James Sully, in his widely known textbooks, contended that the notion of a two-dimensional sensation that is experienced in adult life is a ction. He held that adults always localize and experience sensations as being at some distance.⁴¹ This distance may be perceived more or less accurately. Writers on vision at Russell’s Cambridge, including W. H. R. Rivers and C. S. Myers, observed that perceived size (or magnitude) is determined by retinal image size together with perceived distance. Here is how Rivers explained the relation and an instance of its empirical manifestation: The two chief factors on which depends the perception of the size of an object are the size of the retinal image and the estimated distance of the object. The importance of the latter factor is shown by an a er-image experiment. If the a er-image of an object is projected on a moving screen, it will be seen to change in size, becoming smaller as the screen approaches the eyes, larger as it recedes. The retinal image upon which the a er-image depends remains constant in size, and the changes in apparent magnitude depend on the projection. Emmert measured the a er-image at di erent distances, and found the linear size of the image equal to the linear size of the object, multiplied by the distance at which the image was seen.⁴²

39 Although one might carry through this discussion using the examples of both diminishing size with distance and altered shape with viewing angle, henceforth I focus on the former. It is emphasized by Russell in RSDP, and the psychological literature of his time is better developed on this topic. 40 George Berkeley, An Essay Towards a New Theory of Vision (Dublin: Rhames and Papyat, 1709), § 41; Thomas Reid, An Inquiry into the Human Mind, 4th ed. (Edinburgh: Bell and Creech, 1785), ch. 6, sec. 23, p. 416; Stout, Analytic Psychology, 2:22–3. 41 James Sully, Outlines of Psychology (New York: Appleton, 1886), 149, 181, and The Human Mind: A Text-Book of Psychology, 2 vols. (New York: Appleton, 1892), 1:208, 245. 42 W. H. R. Rivers, “Vision,” in Text-Book of Physiology, ed. E. A. Schäfer, 2 vols. (Edinburgh: Pentland, 1898–1900), 2:1026–1148, on pp. 1139–40. An abbreviated account is given by C. S. Myers, A Text-Book of Experimental Psychology, Part I, 2nd ed. (Cambridge: Cambridge University

336

Hat eld

In speaking of “projection,” Rivers means that a subjective phenomenon, such as an a erimage, is perceived as localized at a speci c location – in this case, on the surface of a screen. He notes that in Emmert’s experiment, the size of the a erimage varies directly with the distance at which the screen is placed. This relation, which speci es that perceived size is a function of visual angle and perceived distance, is known as “Emmert’s law.”⁴³ Rivers generalizes the function to apply to the relations among visual angle, perceived distance, and perceived size in normal perception; the function later became known as the size-distance invariance relation.⁴⁴ According to this relation (Fig. 2a), if distance is veridically perceived, size is veridical (size constancy is attained); if distance is underperceived, size is underperceived, etc. For present purposes, the most interesting feature of Rivers’s description is that, as the a erimage is perceived as being at a greater distance, its size increases. This is exactly the opposite of Russell’s account of the penny, in which apparent size decreases with distance. What is going on? Russell, in describing the penny, is describing the perceived size of an object of xed physical size as it recedes (Fig. 2b). It is unclear from his description whether he means to be describing the diminishing retinal-image size of the penny, its diminishing visual angle, or merely to be reporting a fact about how phenomenal size changes with increasing distance. Rivers is describing a situation in which a constant visual angle or retinal size (the physiological process underlying the a erimage, which remains constant on the retina) is paired with di ering distances. He nds that, for a constant visual angle, perceived size increases with perceived distance, in accordance with Emmert’s law. Thus far, there is no direct contradiction between Rivers and Russell. But Rivers’s description entails that, in Russell’s situation, if the distance to the penny

Press, 1911), ch. 22. Russell was at Cambridge when Rivers and Myers published these works. Although he may have met Myers before 1914, that matters not for my point, which is to indicate what knowledge was available whether Russell availed himself of it or not. 43 On Emil Emmert’s law, see Edwin G. Boring, Sensation and Perception in the History of Experimental Psychology (New York: Appleton-Century-Cro s, 1942), 292–3, 310. 44 The connection between the size–distance invariance relation and Emmert’s law has frequently been discussed. I assume here that Emmert’s law relates visual angle (as perceived) with distance (as perceived). For a review of the earlier literature that raises interpretive questions about Emmert’s own intentions, see William Epstein, John Park, and Albert Casey, “The Current Status of the Size-Distance Hypotheses,” in Psychological Bulletin 58 (1961), 491–514. In any event, a recent review of the constancy literature nds that the “vast majority” of studies validate the size–distance invariance relation: Mark Wagner, The Geometries of Visual Space (Mahwah, N. J.: Erlbaum, 2006), 226. Many psychologists use the term “size–distance invariance hypothesis” rather than “relation.”

337

Russell’s Progress

(a)

P

B

A

C

D

(b) P

Fig. 2: Part (a) illustrates the size–distance invariance relation and Emmert’s law. Assume that the a erimage is formed by segment A, which is perceived to be at the distance shown from the perceiver P. Then, if the a erimage is perceived at distances B, C, or D, it is perceived as being correspondingly smaller or larger. These relations work whether the perceiver veridically perceives the distances, or underrepresents them. Part (b) shows the geometrical relation in which the visual angle subtended by a penny decreases with increasing distance from the perceiver.

is perceived accurately, then the penny should appear of constant size as it recedes.⁴⁵ That is, its size should be perceived veridically, so long as its distance is perceived veridically. Again there is no direct contradiction, because Russell has not made explicit whether or not the distance is perceived veridically. But trouble is brewing, since Russell indicates (by calling the private spaces “three-dimensional”) that the penny is seen at a distance. Further, if Russell accepts Stout’s problem as stated above, that the various sense data are supposed to be “in the same place,” then he must assume that the distance is perceived to increase as the viewer’s actual distance from the penny increases; otherwise, the di ering sense data of the penny would not be in danger of being perceived as being at the same

45 For convenience’s sake, I speak of Russell’s penny as if it were a material object at a location in relation to a perceiver. Of course, for Russell, this notion of the penny is constructed from private spaces in which a penny datum is at a distance from the from-which, that is, from the location from which the datum is perceived.

338

Hat eld

place at the same time. In that case, a contradiction with Rivers ensues; for, if the members of the series of sense data of the penny at the at-which are perceived as all being in the same place (that is, at correspondingly appropriate distances in private space so that they would be constructed as being in the same place), then as the from-which recedes from the at-which, the perceived distance between them must increase appropriately.⁴⁶ But if the increasing distance between at-which and from-which is perceived veridically, then the sense data will all be perceived as having the object’s true size (the actual size of a penny), not as being smaller with increasing distance, as Russell reports them to be.

5 Problems and Prospects for Russell’s Ordering The more sophisticated psychological accounts of Sully and Rivers raise problems for understanding the structure of the private spaces in the six-dimensional space of Russell’s “breakthrough.” The problems are, to begin with, not technical but interpretive. They are not technical, for in this case a technical problem can arise only once we have settled on an interpretation of what Russell means to be asserting in the various passages scouted above, when he speaks of “apparent size” and describes the size of the penny as diminishing with increasing distance between its location and the place from which it is perceived. Re ecting on the relations among apparent size, visual angle, and perceived distance as reviewed above, there are several options for interpreting what Russell means. Although it is not clear that any of the more plausible interpretations of what Russell intended to say are in the end coherent, we may be able to repair things on his behalf. Here are the positions that are initially plausible as interpretations of Russell’s intent: 1. The penny data are located in the same place at-which, and they appear in that place to be increasingly smaller as the distance to the from-which increases. 2. The penny data that diminish in size are perceived as being in the eye or at the eye, and these data are then, as result of experience, overlooked in favor of a perception of the penny in its true location with its true size.

46 I am assuming for now that the distances in question in the private spaces can be ascertained by reference to or comparison with that distance as constructed in the public space. This assumption is reasonable, as the public space locations are built upon relations among the spatial structures of the private spaces. Of course, Russell could decide that this was not the case, because assuming it to be so creates a problem for something else he believes. I take up this possibility in the next section.

Russell’s Progress

339

3. The penny is perceived as being at its true distance and with its true size, but the visual angle diminishes as the distance between the at-which and the from-which increases and the perceiver notices this and says that the penny “looks smaller” with increasing distance. The rst position (1) would be Russell’s intent if he accepted Stout’s problem as formulated, according to which the data at the at-which are in danger of all being in the same place, and so in private space they are at appropriate distances from the from-which so as to be related to the same location in public space. The main problem with (1) is that it violates the size-distance invariance relation (and Emmert’s law). If the penny data appear to be at increasing (veridical) distances as the from-which moves back and yet also appear to be smaller, then their visual angles must decrease even more than is called for by geometry. For if the visual angle decreased as geometry requires and if distance were perceived veridically, the penny data would appear with their veridical sizes, not with diminishing sizes. So if Russell sticks with (1), he violates the basic geometry of size and distance perception. Position (2) would have Russell relying on the historical positions of Berkeley and Reid or the notion of overlooked “sensations” employed by Stout. It conforms to texts that seem to hint of particulars or data radiating from the object;⁴⁷ in the case of the penny, the smaller “at-which” sizes would occur at the from-which locations and would diminish regularly with distance. There are two problems with this interpretation of Russell’s intent. First, he doesn’t explicitly say that the ever smaller penny data appear to be in the eye or at the eye. Indeed, in describing his private spaces or perspectives as three-dimensional, he suggests that the at-which is perceived to be some distance from the from-which. Second, and perhaps more decisively, a er the breakthrough he doesn’t describe the decreasing apparent sizes of the penny as something that must be uncovered. His report that penny data appear smaller with increasing distance is o ered as a description of perceptual experience and he doesn’t describe the “real size” as something that is constructed by taking distance into account in relation to penny data (as he did before the breakthrough). Rather, he speaks of the “real size” of a thing as the size that it appears to have as the distance to it approaches zero.⁴⁸ (For an extent near

47 e.g., UCM, 102–3: “At every place between us and the sun, we said, there is to be a particular which is to be a member of the sun as it was a few minutes ago.” However, such passages might also be interpreted as indicating that the various from-whiches of the perspectives are at “every place” between us and the sun. 48 RSDP, 121–2.

340

Hat eld

the surface of the eye, linear size is close to retinal size.) In any case, (2) doesn’t t his post-breakthrough position very well. Position (3) has the advantage that it accords with Emmert’s law and o ers Russell a way to t his description to the textbook perceptual science of his time. It has him suggesting that size constancy is achieved with the penny, but nonetheless visual angle decreases with increasing distance. Accordingly, Russell’s description of the penny data as looking smaller with increasing distance would be a remark, not on the perceived sizes of the penny data, but on the decreasing visual angle. As an interpretation of Russell’s intent, (3) is not very plausible. Although he brie y mentions increasing size in the “ eld of vision” with decreasing distance in “Ultimate Constituents,” he does not integrate this into a discussion that includes a description of size constancy. In any case, in the more extensive discussion in “Relation of Sense-Data,” he regularly speaks of the penny as appearing smaller with greater distance. None of positions (1) to (3) is satisfactory as an interpretation of Russell’s intent, but for di erent reasons: (1) is unsatisfactory on technical grounds, while (2) and (3) do not harmonize well with the relevant texts. We can, however, formulate a fourth position that is technically acceptable and that accords with his basic phenomenal observation that the penny both (a) looks to be at a distance and (b) looks smaller with greater distance. 4. The relation among perceived size, visual angle, and perceived distance in Emmert’s law is preserved, but as the from-which recedes from the location of the penny, the distance to the “real” location of the penny is underperceived in such a way that the penny data are perceived as being closer and smaller than they would be if size constancy obtained. Position (4) is technically sound with respect to Emmert’s law. It has the advantage of preserving Russell’s phenomenal description, in which the penny looks smaller (the penny data appear smaller) with increasing viewing distance. So far, it is better o than the previous positions. However, it introduces phenomenal facts that Russell doesn’t mention, and it makes his constructive project harder to carry out. The unmentioned phenomenal facts concern the underrepresentation of distance, an underrepresentation that increases with viewing distance. The fact that Russell doesn’t mention these facts is consistent with his scattered remarks according to which distance perception isn’t particularly good, especially at farther distances.⁴⁹ It is not so consistent with the implication that the data in the pri-

49 OKEW, 73: “It seems probable that distances, provided they are not too great, are actually

Russell’s Progress

341

vate spaces are very well known, for it entails that Russell overlooked or didn’t feel the need to describe a basic fact about his hard data. Nonetheless, it may be consistent with his use of the private spaces as the basis for the construction of the common physical space. We have seen that, in this construction, Russell must rely on comparisons among private spaces in establishing a serial order of the from-whiches. This serial order then establishes points or spatial units in public space. Leaving aside problems of ambiguity that might arise (e.g., in distinguishing on the basis of phenomenal size a penny at a shorter distance from a quarter at a longer distance), he would be able to establish a series. For according to (4), the penny data do diminish regularly in size with increasing distance. Unlike (1), these data aren’t all referred to the same location in relation to physical space (or to counterpart distance relations in private space). Nor do they track veridical distance in the private spaces, as in (3). As the from-which recedes from the “real” location of the penny, the phenomenal distance to the penny data at the at-which increases but not at the same rate as the increasing distance to the “real” location. This means that the private spaces include a regular separation, with increased viewing distance, between the at-which and the “real” location. That fact puts an extra wrinkle in Russell’s constructive problem. But it may only make the construction more complicated, without (on the grounds we are discussing) making it impossible. Position (4) may not be the best interpretation of Russell’s intent, as it is unclear that he was responsive to the phenomenal relations that it implies. However, (4) might well be an improvement on his actual position. It takes into account some experimental ndings from Russell’s day that suggest just such a contraction in visual space (an underrepresentation of distance that increases with distance). These ndings were made by the German-speaking sensory psychologists Franz Hillebrand and Walter Blumenfeld.⁵⁰ Their work has received only sporadic

given more or less roughly in sight”; he backs o slightly in Philosophy (New York: Norton, 1927), 138: “It is still an open question whether the space of sight has depth, or is merely a surface, as Berkeley contended.” 50 Franz Hillebrand, “Theorie der scheinbaren Grösse beim binokularen Sehen,” in Denkschri der Kaiserlichen Akademie der Wissenscha en Wien, Mathematisch-Naturwissenscha liche Classe 72 (1902), 255–307; Walter Blumenfeld, “Untersuchungen über die scheinbare Grösse im Sehraume,” in Zeitschri für Psychologie und Physiologie der Sinnesorgane 65 (1913), 241–404. For discussions of their work and its application to the problem of constancy and the observation that things look smaller in the distance, see: Boring, Sensation and Perception, 294–6, and “Visual Perception as Invariance,” in Psychological Review 59 (1952), 141–8; and Hat eld, “Representation and Constraints: The Inverse Problem and the Structure of Visual Space,” in Acta Psychologica 114 (2003), 355–78, “On Perceptual Constancy,” in Perception and Cognition: Essays in the Philosophy of Psychology (Oxford: Clarendon Press, 2009), 178–211, and “Phenomenal and Cog-

342

Hat eld

recognition in the century since it was published.⁵¹ But the results remain interesting today. In the context of Russell’s project, they would seem to o er him a better description of his private spaces, one that accords with the phenomenal diminution of the penny with distance, without violating the basic laws of size and distance perception. If distance is increasingly underrepresented, then the penny data will appear at a distance, which is consistent with the three-dimensionality of the private spaces; but they will also appear increasingly smaller with distance. In this way, these results would allow Russell to put psychology in the service of epistemology, a stated aim of Our Knowledge.⁵² The psychology would be more sophisticated than the textbook commonplaces that Russell incorporates into his own discussion.

6 Russell’s Progress If this reconstruction of Russell’s position a er the breakthrough is correct, it suggests that his new insight was not as path-breaking as he might have imagined. The more textually plausible readings of his position a er the breakthrough either have dire technical problems or turn out not to t the texts very well. Position (4), which is technically more acceptable, does not fully t the texts, either, although it allows Russell to retain the decrease in phenomenal size with distance. With respect to Russell’s ordering in terms of phenomenal size, the technically better position would permit him to pursue his ordering of perspectives into series that allow for construction of the all-embracing physical space. From our viewpoint, it permits us to see how Russell could have elaborated his insight, had he delved more deeply into the psychological literature of his time. As it happens, Russell’s subsequent fascination with the project took him rst to an encounter with behaviorism (though by no means a wholesale adoption of its tenets) in the Analysis of Mind, and then, in the Analysis of Matter, on to a deeper

nitive Factors in Spatial Perception,” in Visual Experience: Sensation, Cognition, and Constancy, ed. G. Hat eld and S. Allred (Oxford: Oxford University Press, 2012), 35–62. 51 The contraction found by Hillebrand and Blumenfeld has been used in arguing that visual space is hyperbolic non-Euclidean: Rudolf K. Luneburg, Mathematical Analysis of Binocular Vision (Princeton: Princeton University Press, 1947); also, Patrick A. Heelan, Space-Perception and the Philosophy of Science (Berkeley: University of California Press, 1983). For discussion, see Hateld, “Phenomenal and Cognitive Factors.” 52 On Russell’s conception of how psychology can aid epistemology, see Hat eld, “Psychology, Epistemology, and the Problem of the External World: Russell and Before,” in The Historic Turn in Analytic Philosophy, ed. E. Reck (London: Palgrave Macmillan, in press).

Russell’s Progress

343

analysis of the theory of matter as constructed from sense data now rendered (under the in uence of James’s neutral monism) as “momentary particulars.”⁵³ In these further developments, and especially in the Analysis of Matter, the distinction between the from-which and the at-which that arose with the breakthrough remain important in ways that are crucial to understanding Russell’s ongoing development. But to address the importance of this distinction in that work would take us beyond the breakthrough itself, a moment that yielded some progress in Russell’s analysis of phenomenal space, but not as much as was available to him had he probed more deeply into the psychological science of his time. Finally, we can conclude that, both at the time of the breakthrough and in Matter, empirical knowledge of phenomenal spatial looks is absolutely important to Russell. One cannot arrive at a theory of the spatial structure of physical space without it. In this way, Russell di ers from Kant. In Kant’s critical philosophy, the structure of physical space is determined by the geometrical structures that must unfold in experience in accordance with space as an a priori form of intuition.⁵⁴ Thus, physical bodies and physical space must conform a priori to geometry, that is, in Kant’s thinking, to the geometry of Euclid.⁵⁵ This remains so even if (some of) Kant’s arguments for the a priori intuitional nature of space are separate from this theory of geometry. In contrast, for Russell in the period we have been considering, all knowledge of physical space, which for him means the constructed physical space of everyday experience and of science, arises from sense data. For the purposes of science, these data may be as of measuring instruments and the results of physical tests. Even so, or decidedly so, the scientist must rely on sense data for knowledge of physical space. It is not a matter of “contemplating” phenomenal appearances (as Strawson had it),⁵⁶ but of constructing even ordinary objects from similarities that are perceptually noted in the spatial structures at-which in relation to the place from-which. Exactly how to carry out the scienti c construction of physical space

53 Russell, The Analysis of Mind (London: Allen and Unwin, 1921); The Analysis of Matter (London: Kegan Paul, Trench, Trubner, 1927). On the importance of the from-which for interpreting Russell’s much maligned statement in Matter about the perceptual world being “in his head,” see Hat eld, “Perception and Sense Data,” in Oxford Handbook of the History of Analytical Philosophy, ed. M. Beaney (Oxford: Oxford Univ. Press, in press), footnote 7. 54 Kant, Critique of Pure Reason, trans. P. Guyer and A. Wood (Cambridge: Cambridge University Press, 1998), B 147; A 165/B 206; A 224/B 271; A 239–40/B 298–9. 55 Kant on the mathematics (geometry) of “physical space”: Prolegomena to Any Future Metaphysics, trans. G. Hat eld (Cambridge: Cambridge University Press, 2004), First Part, Note 1 (Akademie edition, 4:287–8). 56 Strawson, Bounds of Sense, 286.

344

Hat eld

on a large scale was becoming an ever more intricate problem during the time in which Russell wrote. But that for Russell this process of construction must begin with comparisons among phenomenal spatial looks is beyond doubt. This is one of the ways in which, in relation to Kant, Russell was an empiricist.⁵⁷

57 Thanks to Bernhard Thöle for helpful criticism of an earlier version given in Berlin in June, 2011, and to Rolf-Peter Horstmann for ongoing discussions on this and related topics, including comments on the penultimate dra .

Index of Names al-Haytham, Ibn 333 Allais, Lucy 159, 243 Allison, Henry 24, 123, 132, 141, 145 f., 170, 173 f., 193, 242, 251 Ameriks, Karl 170, 178 Anscombe, G. E. M. 64 Aquin, Thomas von 82 Armstrong, David M. 68 Bader, Ralf 106 Berkeley, George 216, 227 f., 240 f., 248, 335, 339, 341 Bielefeldt, Heiner 263 Blackwell, Kenneth 326 f., 332 Block, Ned 62, 68 Blumenfeld, Walter 341 f. Boehm, Omri 282 Boring, Edwin G. 336, 341 Brandt, Reinhard 274 Briesen, Jochen 11, 15 Budick, Sanford 269 Burge, Tyler 91 Buroker, J. Vance 210 Campbell, John 202 Cannon, Joseph 264, 282 Carl, Wolfgang 10, 14, 51 ., 68 Casey, Albert 336 Cassam, Quassim 26, 131, 180 Chignell, Andrew 10, 15, 264, 268, 278 Clarke, Samuel 208 f., 211 Cohen, Hermann 263 Crawford, Donald 265 Descartes, René 67, 73, 79, 91 f., 121, 258, 294 f., 333 Eames, Elizabeth Ramsden 328 Emmert, Emil 335 ., 339 f. Emundts, Dina 25 f., 39, 66, 70, 93, 97, 131, 144, 241, 282 f. Epstein, William 333, 336 Euclid 321, 342 f.

Feder, Johann Georg Heinrich 227 f., 240, 247 f. Fichte, J. G. 9 f., 13 f., 19, 207, 219, 222, 238 ., 252, 258, 310, 312 Förster, Eckart 10, 14, 182, 217 f., 274 Frege, Gottlob 87, 177 f., 187, 200, 204 Freud, Sigmund 62 Gareth, Evans 65 Garve, Christian 227, 240, 247 f. George, R. 210 Gerhardt, Volker 177, 192 Ginsborg, Hannah 10, 14, 76, 138, 186 f., 189, 196, 201 f., 265 f., 268 Goethe, Johann Wolfgang von 303 Go on, Tyson 205 Gri oen, Amber 11, 15, 39 Grimm 209 Grüne, Stefanie 10, 14, 144, 168, 171, 196, 198 Guyer, Paul 10, 15, 24, 102, 135, 144, 179, 182 f., 258, 265, 274 f., 282 Haag, Johannes 191, 198 Hanna, Robert 159 Hanslick, Eduard 293 Hartmann, Klaus 283 Hat eld, Gary 11, 15, 321, 324, 333, 341 . Heelan, Patrick A. 342 Hegel, G. W. F. 9 ., 13 ., 19, 24–39, 42, 95 f., 107–117, 207, 221, 283–306, 308–313, 317 . Heidegger, Martin 11, 15, 79, 310, 313–316, 318 f. Henrich, Dieter 52 f., 55, 179 Heraklit 315 f., 318 f. Hershenson, Maurice 333 Hillebrand, Franz 341 f. Hispanus, Petrus 82 Horstmann, Rolf-Peter 9, 11, 13 ., 24, 33, 35, 41–49, 51–55, 58, 61, 72, 74, 76 f., 93, 96 f., 104, 119–124, 131, 142, 144, 192, 207, 221, 245, 250,

346

Index of Names

261, 263, 266, 274 f., 278, 283, 321, 344 Hume, David 48, 138 f., 155, 182, 187, 258, 261 ., 265, 267, 270 f., 294 Husserl, Edmund 17, 79, 178

Nagel, Thomas 68 Nasim, Omar W. 327 Nehamas, Alexander 265 Newton, Isaac 103, 208, 210 ., 214, 216 Nunn, T. Percy 327

Jacobi, Friedrich Heinrich 9, 13, 215, 229–238, 247 f., 251, 256, 258 James, William 324, 327, 343

Ogilby, John 262 ., 271 f., 281 Ostaric, Lara 266 Otter, Martin 71

Kant, Immanuel 9 ., 13 ., 17, 19–30, 32, 34–39, 41–44, 51–77, 79–117, 119–127, 130 ., 134 ., 138, 142 f., 145–194, 196–205, 207–219, 221–259, 261–281, 290, 292, 294 ., 299, 301, 321 ., 343 f. Kepler, Johannes 333 f. Kitcher, Patricia 75, 146, 180 Knappik, Franz 271, 282 Koch, Anton 11, 15 Kripke, Saul 202 f. Kroner, Richard 178 Kusch, Martin 178

Park, John 336 Peirce, C. S. 34 Perry, John 56, 63 Pinkard, Terry 283 f., 288 . Pippin, Robert 145 f., 150, 186, 283 ., 290–293, 295 Prauss, Gerold 177, 242, 251 Proust, Marcel 318

Lailach-Hennrich, Andrea 77 Landig, Anders 11, 15 Leibniz, G. W. 208–212, 216, 308, 310, 316 Locke, John 99, 227, 294 Longuenesse, Béatrice 10, 14, 20, 24, 53, 60, 75, 96 f., 104, 108, 121, 124–129, 131 ., 135, 142, 144, 164, 168, 185, 189–192, 199 Lowe, Victor 328 Ludwig, Bernd 170 Luneburg, Rudolf K. 342 McDowell, John 132 f. Mendelssohn, Moses 287 Miah, Sajahan 329 f. Milton, John 261 f., 269, 271 Mohr, Georg 72 Moore, G. E. 28, 324 Moran, Richard 101 Mühlhölzer, Felix 209 Murray, Bradley 282 Myers, Charles Samuel 335 f.

Quine, Willard V. O. 181, 195 Reich, Klaus 24, 211 ., 313 Reid, Thomas 335, 339 Reimarus, Hermann Samuel 177 Reinhold, Carl Leonhard 24, 229, 231, 236 Rickert, Heinrich John 178 Rivers, W. H. R. 335 f., 338 Rödl, Sebastian 192 . Rogerson, Kenneth 263 Rorty, Richard 79 Rosefeldt, Tobias 10, 14 f., 53, 57, 60, 97, 122, 142, 144, 156, 170, 173, 237, 243, 245, 249 Rosen, Michael 97 Rusnock, P. 210 Russell, Bertrand 11, 15, 321–333, 335–344 Sahnwaldt, Anne Mone 11, 15, 77 Sandkaulen, Birgit 229, 231, 234 Sassen, Brigitte 227 Savile, Anthony 263 Scha arzyk, Wolfgang 11, 15 Schelling, F. W. J. 207, 300 Schi mann, Kelley 106 Schlösser, Ulrich 10, 14

Index of Names

Schulze, Gottlob Ernst 229, 231–238, 247 f., 251, 256, 258 Searle, John 18, 81 Sellars, Wilfrid 150 Shoemaker, Sydney 64, 67 f., 75 f. Smit, Houston 198 Stoecker, Ralf 144 Stout, G. E. 327, 332 f., 335, 337, 339 Strawson, Peter Frederick 75, 321 f., 343 Stroud, Barry 9 f., 14, 26, 51 ., 56, 79, 89–92 Stuhlmann-Laeisz, Rainer 198 Sully, James 335, 338 Taylor, Charles 283 Tetens, Johannes Nikolaus 183

347

Thöle, Bernhard 144, 344 Wagner, Mark 336 Waldorf, David 208 Warren, Daniel 85, 144 Watkins, Eric 155 Waxman, Wayne 167 Whitehead, A. N. 327 f. Willaschek, Marcus 145 f., 150 f., 170 f., 174, 250 f. Windelband, Wilhelm 178 Wittgenstein, Ludwig 64, 75 f., 87, 201 . Wol , Christian 95, 105 f., 209 Wol , Michael 24, 122, 177, 198 Wunderlich, Falk 61 Zuckert, Rachel 266, 282