Kant and the Problem of Self-Knowledge (Routledge Studies in Eighteenth-Century Philosophy) 9781138385467, 9780429427091, 1138385468

This book addresses the problem of self-knowledge in Kant’s philosophy. As Kant writes in his major works of the critica

149 38 2MB

English Pages 214 [227] Year 2018

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Cover
Half Title
Title
Copyright
Contents
Key to Abbreviations of Kantian Works
Acknowledgements
A Brief Introduction
1 Self-Knowledge and Self-Consciousness
The Problems of Self-Knowledge
The Question of Self-Consciousness
Indexicality
Immunity to Error Through Misidentification
Kant’s Theory of the Mind: Faculties and Representations
The Inner Sense
An Introspective Account of Self-Consciousness
The Inner Sense and the Ideality of Time
The Epistemic Limits of the Inner Sense
The Transcendental Deduction and the Principle of the Unity of Apperception
The Metaphysical Deduction
The Transcendental Deduction
I Think and the Principle of the Necessary Synthetic Unity of Apperception
2 Two Senses of ‘I think’
The Synthetic Unity of Apperception and the I Think
Transcendental Dialectic: The I Think and the Analysis of the Paralogisms
The Analysis of the Paralogisms
I as a Concept
The ‘I Exist Thinking’: The Empirical Apperception
Thinking Is Being
The Concept of a ‘Transcendental Subject’
I think qua Thinking and I think qua Representation
3 The Problem of Self-identification
‘I think’ and the Question of Self-Identification
The Genesis of the Cartesian Self
Two No-Ownership Readings and One Thesis of Exclusion
The I of Pure Apperception and the I as Subject
The I of Empirical Apperception
4 The Simple Representation I and the Transcendental Designation
I Think and the Direct Reference Theory
Kant Between Conceptualism and Non-Conceptualism: Concepts and Intuitions
Intuition as an Indexical Representation
The Logical Form of Singular Judgments
The Dismissal of the Lowest Species
The Conceptualist Form of Singular Terms
The Role of Designation in Natural Kind Terms
Kripke and Putnam on the Theory of Direct Reference
Kant and the A Priori Nature of the Judgment “Gold Is a Yellow Metal”
The Relationship of Term, Concept, and Natural Substance
The Semantic Reflection in the Logical Corpus
The Role of Designation
The Simple Representation ‘I’ and the Role of Transcendental Designation
5 On de se and de re
Knowledge by Acquaintance and Knowledge by Description: The Weak Conceptualist View
The Theory of the Transcendental Object
Non-Conceptualism Versus Weak Conceptualism
“The Object Must Be Thought of Only as Something in General = x”
De se Thoughts Between Descriptivism and Singularism
Kant on de se
I Think and de se Thoughts
I as Form of Thinking
The Paradox of Self-Consciousness and the Non-Conceptual Forms of Self-Awareness
The Problem of Self-Knowledge
The Dualism of the I of Apperception and the I as Human Being
Longuenesse and Ginsborg on the I Considered as Human Being
Capozzi and the Role of Attention in the Inner Sense
A Brief Conclusion
References
Index
Recommend Papers

Kant and the Problem of Self-Knowledge (Routledge Studies in Eighteenth-Century Philosophy)
 9781138385467, 9780429427091, 1138385468

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Kant and the Problem of Self-Knowledge

This book addresses the problem of self-knowledge in Kant’s philosophy. As Kant writes in his major works of the critical period, it is due to the simple and empty representation ‘I think’ that the subject’s capacity for selfconsciousness enables the subject to represent its own mental dimension. This book articulates Kant’s theory of self-knowledge on the basis of the following three philosophical problems: (1) a semantic problem regarding the type of reference of the representation ‘I’; (2) an epistemic problem regarding the type of knowledge relative to the thinking subject produced by the representation ‘I think’; and (3) a strictly metaphysical problem regarding the features assigned to the thinking subject’s nature. The author connects the relevant scholarly literature on Kant with contemporary debates on the huge philosophical field of self-knowledge. He develops a formal reading according to which the unity of self-consciousness does not presuppose the identity of a real subject, but a formal identity based on the representation ‘I think’. Luca Forgione is Associate Professor in Philosophy of Language and in Philosophy of Mind at the University of Basilicata, Italy.

Routledge Studies in Eighteenth-Century Philosophy

Kant and the Cultivation of Virtue Chris W. Surprenant The Post-Critical Kant Understanding the Critical Philosophy through the Opus Postumum Bryan Wesley Hall Kant’s Inferentialism The Case Against Hume David Landy Rousseau’s Ethics of Truth A Sublime Science of Simple Souls Jason Neidleman Kant and the Scottish Enlightenment Edited by Elizabeth Robinson and Chris W. Surprenant Kant and the Reorientation of Aesthetics Finding the World Joseph J. Tinguely Hume’s Science of Human Nature Scientific Realism, Reason, and Substantial Explanation David Landy Hume’s Moral Philosophy and Contemporary Psychology Edited by Philip A. Reed and Rico Vitz Kant and the Problem of Self-Knowledge Luca Forgione For more information about this series, please visit: www.routledge.com/ Routledge-Studies-in-Eighteenth-Century-Philosophy/book-series/SE0391

Kant and the Problem of Self-Knowledge Luca Forgione

First published 2019 by Routledge 52 Vanderbilt Avenue, New York, NY 10017 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2019 Taylor & Francis The right of Luca Forgione to be identified as author of this work has been asserted by him in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data Names: Forgione, Luca, 1973– author. Title: Kant and the problem of self-knowledge / by Luca Forgione. Description: 1 [edition]. | New York : Taylor & Francis, 2018. | Series: Routledge studies in eighteenth-century philosophy ; 17 | Includes bibliographical references and index. Identifiers: LCCN 2018039559 | ISBN 9781138385467 (hardback) Subjects: LCSH: Kant, Immanuel, 1724–1804. | Self-knowledge, Theory of. Classification: LCC B2798 .F665 2018 | DDC 126.092—dc23 LC record available at https://lccn.loc.gov/2018039559 ISBN: 978-1-138-38546-7 (hbk) ISBN: 978-0-429-42709-1 (ebk) Typeset in Sabon by Apex CoVantage, LLC

Contents

Key to Abbreviations of Kantian Works Acknowledgements A Brief Introduction

viii xi 1

1 Self-Knowledge and Self-Consciousness 8 The Problems of Self-Knowledge 8 The Question of Self-Consciousness 12 Indexicality14 Immunity to Error Through Misidentification 17 Kant’s Theory of the Mind: Faculties and Representations 20 The Inner Sense 25 An Introspective Account of Self-Consciousness 25 The Inner Sense and the Ideality of Time 27 The Epistemic Limits of the Inner Sense 29 The Transcendental Deduction and the Principle of the Unity of Apperception32 The Metaphysical Deduction 32 The Transcendental Deduction 34 I Think and the Principle of the Necessary Synthetic Unity   of Apperception 37 2 Two Senses of ‘I think’ 49 The Synthetic Unity of Apperception and the I Think 52 Transcendental Dialectic: The I Think and the Analysis of the Paralogisms54 The Analysis of the Paralogisms 54 I as a Concept 55 The ‘I Exist Thinking’: The Empirical Apperception 60

vi Contents Thinking Is Being The Concept of a ‘Transcendental Subject’ I think qua Thinking and I think qua Representation

62 65 68

3 The Problem of Self-identification 75 ‘I think’ and the Question of Self-Identification 76 The Genesis of the Cartesian Self 79 Two No-Ownership Readings and One Thesis of Exclusion 83 The I of Pure Apperception and the I as Subject88 The I of Empirical Apperception 96 4 The Simple Representation I and the Transcendental Designation 103 I Think and the Direct Reference Theory 104 Kant Between Conceptualism and Non-Conceptualism: Concepts and Intuitions 109 Intuition as an Indexical Representation 113 The Logical Form of Singular Judgments 114 The Dismissal of the Lowest Species 115 The Conceptualist Form of Singular Terms 116 The Role of Designation in Natural Kind Terms 118 Kripke and Putnam on the Theory of Direct Reference 118 Kant and the A Priori Nature of the Judgment “Gold Is a   Yellow Metal” 120 The Relationship of Term, Concept, and Natural Substance 122 The Semantic Reflection in the Logical Corpus 125 The Role of Designation 127 The Simple Representation ‘I’ and the Role of Transcendental Designation130 5 On de se and de re143 Knowledge by Acquaintance and Knowledge by Description: The Weak Conceptualist View 143 The Theory of the Transcendental Object 145 Non-Conceptualism Versus Weak Conceptualism 148 “The Object Must Be Thought of Only as Something in  ­General = x” 152 De se Thoughts Between Descriptivism and Singularism 154 Kant on de se160 I Think and de se Thoughts 160 I as Form of Thinking 164

Contents  vii The Paradox of Self-Consciousness and the Non-Conceptual   Forms of Self-Awareness The Problem of Self-Knowledge The Dualism of the I of Apperception and the I as Human  Being Longuenesse and Ginsborg on the I Considered as Human  Being Capozzi and the Role of Attention in the Inner Sense A Brief Conclusion

170 177 177 179 183 195

References201 Index212

Key to Abbreviations of Kantian Works

The majority of English language quotations from Kant’s works in this book are from The Cambridge Edition of the Works of Immanuel Kant, edited by P. Guyer and A. Wood (Cambridge University Press, 1992ff.). When a translation was not available, I provided my own. Apart from the references to the Critique of Pure Reason, all references to Kant are to the volume and pages of Kant’s gesammelte Schriften (KGS), herausgegeben von der Deutschen (formerly Koniglichen Preussischen) Akademie der Wissenschaften, 29 vols. (Berlin: Walter de Gruyter, 1902). References to the Critique of Pure Reason are to the standard A and B pagination of the first and second editions, respectively. Specific works cited are referred to by means of the abbreviations listed below. The translations used are also listed below and, with the exception of the Critique of Pure Reason, are referred to immediately following the reference to the volume and page of the German text. A/B  Kritik der reinen Vernunft [KrV] (KGS 3–4) Critique of Pure Reason, ed. and trans. P. Guyer and A. Wood (Cambridge: Cambridge University Press, 1997) Anth  Anthropologie in pragmatischer Hinsicht (KGS 7) Anthropology From a Pragmatic Point of View, ed. and trans. R. B. Louden (Cambridge: Cambridge University Press, 2006) Anth-B  Anthropology, History, and Education, ed. R. B. Louden and G. Zöller, trans. M. Gregor, P. Guyer et al. (Cambridge: Cambridge University Press, 2007) Br  Kant’s Briefwechsel (KGS 10–13) Correspondence, ed. and trans. A. Zweig (Cambridge: Cambridge University Press, 1999) FM  Welches sind die wirklichen Fortschritte, die die Metaphysik seit Leibnitzens und Wolf’s Zeiten in Deutschland gemacht hat? (KGS 20) ‘What Real Progress Has Metaphysics Made in Germany Since the Time of Leibniz and Wolff?’, in Theoretical Philosophy After 1781, ed. H.

Key to Abbreviations of Kantian Works  ix Allison and P. Heath, trans. G. Hatfield, M. Friedman et al. (Cambridge: Cambridge University Press, 2002) KGS  Kant’s Gesammelte Schriften, ed. Königlich Preußischen, später Deutschen Akademie der Wissenschaften zu Berlin (Berlin: de Gruyter, 1900—) KU  Kritik der Urteilskraft (KGS 5) Critique of the Power of Judgment, ed. P. Guyer, trans. P. Guyer and E. Matthews (Cambridge: Cambridge University Press, 2000) LBl B 12  Lose Blätter B 12 Critique of Pure Reason (KGS 23) Notes and Fragments, ed. P. Guyer, trans. C. Bowman, P. Guyer, and F. Rauscher. (Cambridge: Cambridge University Press, 2005) LM  Lectures on Metaphysics, ed. and trans. K. Ameriks and S. Naragon (Cambridge: Cambridge University Press, 1997) Log  Logik (KGS 9) ‘The Jäsche Logic’, in Lectures on Logic, ed. and trans. M. Young, pp. 521– 640. (Cambridge: Cambridge University Press, 1992) MAN  Metaphysische Anfangsgründe der Naturwissenschaften (KGS 4) Metaphysical Foundations of Natural Science, ed. and trans. M. Friedman (Cambridge: Cambridge University Press, 2004) OP  Opus postumum (KGS 21 u. 22) Opus Postumum, ed. and trans. E. Forster and M. Rosen (Cambridge: Cambridge University Press, 1993) PhilEnz  Philosophische Enzyklopädie (KGS 29) Prol  Prolegomena zu einer jeden künftigen Metaphysik (KGS 4) Prolegomena to Any Future Metaphysics That Will Be Able to Come Forward as Science, ed. and trans. G. Hatfield (Cambridge: Cambridge University Press, 2004) Refl  Reflexionen (KGS 14–19) Notes and Fragments, ed. P. Guyer, trans. C. Bowman, P. Guyer, and F. Rauscher. (Cambridge: Cambridge University Press, 2005) V-Lo/Blomberg  Logik Blomberg (KGS 24) ‘The Blomberg Logic’, in Lectures on Logic, ed. and trans. M. Young, pp. 5–246 (Cambridge: Cambridge University Press, 1992). V-Lo/Dohna  Logik Dohna-Wundlacken (KGS 24) ‘The Dohna-Wundlacken Logic’, in Lectures on Logic, ed. and trans. M. Young, pp. 438–516 (Cambridge: Cambridge University Press, 1992) V-Lo/Pölitz  Logik Pölitz (KGS 24)

x  Key to Abbreviations of Kantian Works V-Lo/Wiener  Wiener Logik (KGS 24) ‘The Vienna Logic’, in Lectures on Logic, ed. and trans. M. Young, pp. 251– 377 (Cambridge: Cambridge University Press, 1992) V-Met-L1/Pölitz  Kant Metaphysik L 1 (Pölitz) (KGS 28) ‘Metaphysik L1, mid-1770s’, in Lectures on Metaphysics, ed. and trans. K. Ameriks and S. Naragon, pp. 17–106 (Cambridge: Cambridge University Press, 1997, 2001) V-Met/Mron  Metaphysik Mrongovius (KGS 29) ‘Metaphysik Mrongovius, 1782–1783’, in Lectures on Metaphysics, ed. and trans. K. Ameriks and S. Naragon, pp. 109–288 (Cambridge: Cambridge University Press, 1997).

Acknowledgements

In the last years I dwelled on the topics of self-knowledge, philosophy of language, and philosophy of mind in Kantian philosophy through several papers and talks in international journals and congresses, for instance: Kant on de se (forthcoming in Akten des XII Internationaler KantKongresses—Kant-Gesellschaft); Kant, the Transcendental Designation of I, and the Direct Reference Theory (forthcoming in Theoria. An International Journal for Theory. History and Foundations of Science); The Structure of I-Thoughts. Kant and Wittgenstein on the Genesis of Cartesian Self (forthcoming in Paradigmi); Kant and the Simple Representation I (International Philosophical Quarterly); Kant and Natural Kind Terms (Theoria. An International Journal for Theory. History and Foundations of Science); Kant on de re: Some Aspects of the Kantian Non-Conceptualism Debate (Kant Studies Online); Kant and I as Subject (Akten des XI Internationaler Kant-Kongresses—Kant-Gesellschaft); Kant and the Problem of Self-identification (Organon F). This book stems from the ideas contained in these works and also from discussions with several scholars and reviewers. It’s impossible to list all them here, but I want to express all my gratitude to my editor Andrew Weckenmann for his support and to the two scholars who have reviewed this book for their comments and suggestions. In publishing this book, my first and warm greeting goes to Stefano Gensini; once again on this occasion he encouraged me by restating his teaching on the inseparable bond between theoretical reflection and historical–interpretive investigation. If all of Mirella Capozzi’s work has always represented for me an indispensable reference point in my research on Kant, her 2007 essay on Kant and self-knowledge was the compass that allowed me to orient myself in this fascinating Kantian territory, starting from the Kantian quotation that serves as an exergue to this book. Finally, an affectionate thought to Pasquale Frascolla, from whom I have drawn so many teachings every day for almost 15 years; I hope there will be many others to come. I dedicate this book to Emilia de Lucia, for her affection that has accompanied me ever since I was born.

A Brief Introduction

Gott erkennt alles, indem er sich selbst erkennt. Der Mensch erkennt sich selbst, indem er andere Dinge erkennt. God knows all knowing itself. The man knows himself knowing the other things. (R. 3826, KGS 17: 304)

In the famous first paragraph of the Anthropology, by drawing attention to both concepts of ‘person’ and ‘unity of consciousness’, i.e., the transcendental apperception expressed by the representation ‘I’, Kant links the moral and theoretical aspects of his philosophical approach in order to consider the human being infinitely above all other living beings on Earth: The fact that the human being can have the “I” in his representations raises him infinitely above all other living beings on earth. Because of this he is a person, and by virtue of the unity of consciousness through all changes that happen to him, one and the same person—i.e., through rank and dignity an entirely different being from things, such as irrational animals, with which one can do as one likes. (Anth 7: 127, 15) The aim of this book is to enquire about the theoretical aspects of Kant’s philosophy that are connected to the representation ‘I’, whereas the moral dimension will not be considered.1 As Kant said in his major works of the critical period, it is due to the ‘simple’ and ‘empty’ representation ‘I think’ that the subject’s capacity for self-consciousness enables the subject to represent its own mental dimension, as well as itself as one and the same subject through all changes. More specifically, the subjective capacity to represent itself through the representation I will be articulated on the basis of the following three questions, which cover different philosophical areas: (1) a semantic question regarding the type of reference of the representation ‘I’,

2  A Brief Introduction (2) an epistemic question regarding the type of knowledge relative to the thinking subject produced by the representation ‘I think’, and (3) a strictly metaphysical question regarding the features assigned to the thinking subject’s nature. These three different questions obviously touch on the huge philosophical field of self-knowledge, which is concerned with the knowledge of one’s own mental states, e.g., the knowledge of one’s current experiences, thoughts, beliefs, or desires. A classic problem, for instance, involves the possibility of determining what a subject is feeling or thinking at a given moment, and yet there is significant disagreement about the nature of this knowledge among scholars. The problem of the knowledge of one’s mental states involves the self-conscious subjective dimension. The fact that a subject acquires knowledge of her belief that Naples is a lovely city implies that the state is registered as her own; this is related to the question of ‘self-consciousness’ or ‘self-awareness’ proper (the terms are interchangeable in this context), one of the major topics in the philosophical arena. Since expressions of selfknowledge employ terms such as “I”, as in “I feel an itch”, the problem of self-consciousness also concerns how the determination of the reference of I and the identification of those mental states as one’s own may be achieved. In fact, as Gertler (2017) points out, “self-knowledge” can also be used to refer to knowledge of the self and its nature, which are connected to selfconsciousness and a few related issues, for instance: how it is that one distinguishes oneself from others, as the object of a self-attribution; whether self-awareness yields a grasp of the material or non-material nature of the self; whether self-awareness yields a grasp of one’s personal identity over time. In his approach to self-consciousness, Rödl (2007) takes a step further by linking self-consciousness and self-knowledge in an extremely strong way; firstly, he says, “self-consciousness is the nature of a subject that manifests itself in her thinking thoughts whose linguistic expression requires the use of the first person pronoun, ‘I’ ” (2007, VII). Secondly, he adopts the spirit of Evans’ approach, according to which forms of reference have to be understood through corresponding forms of predication, and the theme at issue is a manner of thinking about an object; in other words, a form of reference: As aspects of thinking a predicative thought, referring to an object and predicating a concept of it bear a unity, which suggests that formal distinctions in the one are linked to formal distinctions in the other. Since, fundamentally, reference is to something real, the relevant forms of predication are forms of knowledge, forms of knowing how things stand with the object. (2007, VIII)

A Brief Introduction  3 In this way, an inquiry into self-consciousness corresponds to an inquiry into a form of knowledge, which is knowledge of oneself as oneself; that is, self-knowledge. Thus, since the form of knowledge connected to the first-person component is a form of knowing acts of thinking, Rödl distinguishes two kinds of thinking, namely practical and theoretical thinking, or action and belief: The former concerns the way in which “I know that I am doing something when my knowing it is an act of self-consciousness”, while the latter concerns “the way in which I know that I believe something when, again, I know it in such a way as to know that I believe it”. If belief and action can only be known by the subject via a first-person perspective, acts of thought are essentially self-conscious: “Therefore, a theory of self-consciousness is a theory of action, belief, and knowledge” (2007, VIII). Rödl stresses that the German Idealist tradition (specifically that of Kant and Hegel) considers the philosophical study of action and knowledge as part of an inquiry into self-consciousness. The attempt of this book is precisely to shed light on those central aspects of Kant’s philosophy regarding self-consciousness that are essential in order to understand the thinking activity itself, as Kant considered self-consciousness and thinking to be two sides of the same coin. In Chapter 1, the notion of self-consciousness that will be considered can be referred to as basic self-consciousness. This consists of two specific, correlated features that do not pertain to the consciousness of things other than oneself: Based on the first feature, in fact, self-consciousness can be said to be grounded in a first-person perspective, whereas due to the second feature, self-consciousness must be regarded as a consciousness of the self as subject rather than a consciousness of the self as object. Both peculiarities are grounded in the possibility of using the term/concept I, which presents a few specific epistemic and semantic features: Essential indexicality and immunity to error thorough misidentification. The former is relative to the meaning of the term/concept I, any expression of self-consciousness being based on indexical terms such as “I” or “me”; the latter, on the other hand, refers to the fact that certain singular judgments involving the self-ascription of mental (and physical, as will be seen later) properties are immune to error through misidentification relative to the first-person pronoun (IEM). The subject formulating such judgments in given epistemic contexts cannot be mistaken as to whether it is she herself who is attributing a particular mental property to her own self. Briefly stated, the capacity for self-consciousness depends on the possibility of producing I-thoughts, which, as such, employ an indexical self-reference immune to error through misidentification relative to the concept I. The general point that will be developed in this book is that Kant’s approach to self-consciousness seems to succeed in explaining these features as it does consider the above-mentioned features of the concept I in the terms of transcendentalism in some way. Kant points to two forms of self-consciousness: The inner sense, or empirical apperception, based on a sensory form of self-awareness, and transcendental apperception.

4  A Brief Introduction Through the notion of inner sense, Kant also allows for an introspective account of self-awareness; nonetheless, the point and purpose of this book is to show that an utterly sophisticated notion of basic self-consciousness is, in fact, provided for by the notion of transcendental apperception. As we will see, the doctrine of apperception is not to be confused with an introspective psychological approach: in reality, it is a formal model for the thinking activity itself. Chapter 2 aims to address certain characterisations of ‘I think’ connected to the problems of self-consciousness and self-knowledge according to the so-called Formal Ownership Reading so as to set the Kantian metaphysical assumptions about the thinking subject against this interpreting background that is so central to the Kantian debate. Most importantly, two distinct meanings of ‘I think’ need be identified: while in its first meaning, mainly found in the Transcendental Deduction, the ‘I think’ is the act of apperception, in the second meaning, found in Transcendental Deduction and in the section of Paralogisms in particular, I think (in italics in these pages) is assumed in its representational nature. The notion of the ‘transcendental subject’ will be interpreted in formal terms as a specific concept that, mutatis mutandis, has the same function as the concept of the ‘transcendental object’. With regard to the three philosophical key questions—the semantic, the epistemic, and the metaphysical problems—the ‘I think’ and the so-termed transcendental subject will be analysed in an attempt to establish certain specific metaphysical characterisations of the thinking subject introduced by Kant in the critical period. In particular, starting from the general interpretative scenario of the formal reading, I will attempt to present and develop three specific theses: (1) a metaphysical thesis: Thinking, the being itself, is spontaneity, (2) an epistemic thesis: Thinking is not the object of knowledge. If Thinking is the being itself, and if nothing is thereby given to the subject for thinking, then the subject can think of itself only as an object in general through the concept of a transcendental subject, i.e., the thought of a thinking being that has an existence in itself, and (3) a semantic thesis: The I think, the bare or empty representation I, is the representational vehicle for the concept of the transcendental subject; as such, it is a simple representation. The awareness of oneself as thinking is only expressed by the I—the intellectual representation of the spontaneity of a thinking subject. While the epistemic bone of contention concerns how the subject thinks of itself, and the knowledge it can form through the I-thought based on the representation I think, the semantic question involves the nature of the representation I and the type of designation (if any) that is involved in the reference of the thinking subject. In the transcendental system, the epistemic and the semantic theses seem to stem directly from the metaphysical thesis. The

A Brief Introduction  5 epistemic thesis and some points of the semantic thesis will be discussed in Chapter 3, with special attention to the problems of self-identification and the reference of I. The semantic thesis and the problem of transcendental designation will be addressed in depth in Chapter 4. In Chapter 3, the notion of transcendental designation and the relative question of self-identification will be explored in depth because Kant seems to anticipate some of the self-reference without identification features. As mentioned previously, due to the absence of identification components, certain singular judgments involving the self-ascriptions of mental and physical properties are immune to error through misidentification relative to the first-person concept. The subject formulating such judgments in given epistemic contexts cannot be mistaken as to whether it is she who is attributing a particular property to herself. The issue is introduced by Wittgenstein in his philosophico-linguistic analysis of the grammatical rule governing the term I, used as subject as well as object. Shoemaker specifies the question of self-reference without identification, and Evans of identification-free self-ascription. As will be discussed later, the issue is slightly more complex when expounding on the Kantian approach. The transcendental unity of apperception is the foundation of representational synthesis, through which an objective determination of representations for possible cognition arises. In this picture, the ‘I think’ resides in a metaphysical frame, which necessarily involves any thinking activity because it does identify with such an activity. At least at this level of investigation, and with reference to the passages that will be considered, this represents the highest level of abstraction in the transcendental reflection. In this scenario, several Kantian interpretative readings in the debate will be considered and discussed in order to specify the appropriate framework for the I think feature associated with the question of self-identification so as to highlight the difference from Wittgenstein’s approach and the contemporary debate. In Chapter 4, the issue of the nature of the representation I and of the transcendental designation will be analysed in order to answer the following questions: What exactly does Kant mean when he states that I is a simple and empty representation? Can the features of the representation I and the correlative ‘transcendental designation’ explain the indexical nature of the I? Do the Kantian considerations on indexicality anticipate any of the semantic elements or—if nothing else—the spirit of the direct reference theory? With regard to the last question, some sort of contiguity between the Kantian approach to the I think and the contemporary direct reference theory concerning the semantic function of I has been suggested in the Kantian debate. In addition, the direct reference theory has also been applied to the Kantian approach to the semantics of natural kind terms. In order to rule out any proximity to the direct reference theory in these specific semantic issues, Chapter 4 will focus on how Kant treats indexicality. Furthermore, non-conceptual content theorists have taken Kant as a reference point in recent years due to his notion of intuition, and some Kantian scholars in

6  A Brief Introduction the current debate regard sensible intuition as an indexical representation. It is necessary to explore a number of complementary issues intertwined with the notion of non-conceptual content in order to understand Kant’s treatment of indexicality. Of these, the first is solely concerned with the role of the intuition as an indexical representation, whereas the second pertains to the presence of some epistemic features that will be discussed in the next chapter based on the distinction between knowledge by acquaintance and knowledge by description. Following this, the designation involved in the relationship of words, concepts, and intuition will be discussed with particular regard to the kind of designation involved in natural kind terms. In so doing, the features that Kant assigns to the different representations and to the correlative designation will be compared to the representational features of I and its correlative transcendental designation: Undoubtedly, since transcendental designation displays utterly unique designation features, a simple or empty representation is a representational unicum among the kinds of representations examined by Kant. Accordingly, it will be possible to pinpoint the peculiarities of the representation I in order to comprehend the role of transcendental designation. In the last chapter, other complementary questions will be addressed. The first of these revolves around the question of de re thoughts, whereas the second refers to the articulation of de se thoughts in the transcendental system. As mentioned previously, shifting from the semantic considerations concerning the referential mechanism of the intuitive representations expounded in the preceding chapter to a more strictly epistemic perspective, with regard to the distinction between conceptual and non-conceptual content, the Kantian difference between concepts and intuitions has been partly associated with the distinction between knowledge by acquaintance and knowledge by description, which specifies two basic types of knowledge. Firstly, this chapter intends to dismiss the possibility that intuition has an autonomous function of de re knowledge in support of an interpretative reading that could be labelled “weak conceptualism”. To this end, the exploration will be conducted from a strictly transcendental perspective; that is, by referring to the so-called theory of the concept of a transcendental object. The interpretative reading features a number of results that are in contrast to the main points stirring the debate on Kantian non-conceptualism. At the same time, with reference to de se thoughts, when beginning with the faculty analysis, one may certainly reconstruct Kant’s stance regarding the different types of de se thoughts. There is no doubt that Perry and Recanati’s de se-thoughts perspective discussed in these pages is particularly attractive: Certain features of transcendental apperception and I think seem to anticipate certain points of this approach, with particular reference to a basic typology of implicit de se thoughts. Finally, the problem of self-knowledge in the empirical and transcendental dimension will also be analysed. Prima facie, there seems to be some incompatibility between self-knowledge as a human being, that is, as an object embedded in a spatio-temporal causal order governed by

A Brief Introduction  7 natural laws, and the knowledge that the subject possesses as a thinking subject through the spontaneity of the I of apperception: The representation I think contains no intuition that can connect it to the subject considered to be an empirical object. The final part of the book is dedicated to overcoming this kind of dualism between the I of apperception and the I as human being on the basis of the formal reading adopted here.

Note 1. Cf. Ware (2009), La Rocca (2013), for a first introduction to the Kantian issue of self-knowledge in the moral sphere, starting from the famous § 14 from the “Doctrine of Virtue” in the Metaphysics of Morals, where Kant introduces “to know (scrutinize, fathom) [Erkenne (erforsche, ergründe)] yourself” as “the first command” among all duties. Bagnoli’s works on Kant’s account of practical reason in metaethical debates touch on different issues pertaining to the philosophical area of self-knowledge.

1 Self-Knowledge and Self-Consciousness

The Problems of Self-Knowledge In general terms, the philosophical area of self-knowledge is concerned with the knowledge of one’s own mental states; for example, the knowledge of one’s current experiences, thoughts, beliefs, or desires. A classic problem, for instance, revolves around the possibility of determining what a subject is feeling or thinking at a given moment: Although statements such as “I feel an itch” or “I’m thinking about summertime” can sometimes express knowledge, there is significant disagreement regarding the nature of this knowledge among scholars. As a rule, a subject can know what she is thinking or feeling, what she believes or desires: If asked, she can form a description of her own mental dimension and reach knowledge of her mental states under normal circumstances—due to the very nature of the subjective mental dimension, each subject is obviously in a better position than anybody else to identify her own mental states. Subjects seem to be authoritative about what they are thinking or feeling because the method they employ is different from that available to others; in other words, subjects seem to be provided with a first-person authority about their own mental states. The nature of the first-person authority is directly at the heart of a fierce philosophical dispute. While some philosophers regard first-person authority to be epistemic, contending that subjects enjoy privileged access to their own mental states, others believe that subjects have no such access whatsoever. Even if many empirical and conceptual considerations have highlighted how restricted the sphere of first-person authority is—and there is a vast philosophical debate about the types of mental states involved therein— the main theories of self-knowledge aim to explain the sense of firstperson authority and the reasons for the introduction of a special method: An exclusively first-personal method that enables the subject to grasp her own mental states in a completely different way from those that are adopted to form knowledge about other spheres. In particular, compared to the knowledge about other domains, Gertler (2011) regards self-knowledge to be special in three respects: (1)

Self-Knowledge and Self-Consciousness  9 self-knowledge can be considered particularly secure from an epistemic angle, (2) self-knowledge can be grounded in a unique epistemic method, and (3) self-knowledge can be considered epistemically special. However, what is special about self-knowledge is not epistemic. Statement (1) points out the unique feature of self-knowledge as a result of its epistemic status; for instance, by supporting the strongest epistemic claims, such as the subject’s infallibility or omniscience about her own mental states. Statement (2) emphasises the special method employed to achieve self-knowledge; for example, the introspective method adopted by the acquaintance theory. Lastly, with regard to statement (3), what is fundamental in self-knowledge is non-epistemic, as its epistemic specialness depends on its non-epistemic specialness—as contended, for instance, by Rationalism. For my purposes, I will mainly focus on the methods employed to access one’s own mental dimension; nonetheless, this will be done by following the leading theories of self-knowledge that can be singled out in the debate on the basis of the three statements mentioned above: the acquaintance theory, the inner-sense theory, and the rationalist theory. Acquaintance theorists regard first-person authority as an epistemic phenomenon based on introspection and as an exclusively first-personal method of grasping our own mental states. The method of introspection is grounded in a type of inward reflection concerning the contents of one’s own mind, so that the subject can be directly aware of or “acquainted” with her mental dimension. It follows that the first-personal method differs from perceptual awareness. While the latter depends on certain causal relations—for instance, the fact that there is a book causally produces a visual experience—the introspective awareness of a mental state is unmediated. For example, with regard to the awareness of a sensation such as pain, the pain will be considered as part of the introspective experience because it goes beyond shaping the subjective introspective experience. Hence, introspective awareness is not based on causal processes but is metaphysically direct, with no mediation between one’s awareness of the mental state and the mental state itself taking place. The result is an epistemic asymmetry. If the perceptual belief that there is a book relies on a causal process, then the knowledge that can be achieved by a subject will depend on whether its perception is appropriately caused by the presence of a book: If it is caused by a hallucinogenic drug, a subject will not know that there is a book in front of her. According to the acquaintance theory, the introspective awareness of a mental state does not depend on this type of causal matter. The point here is not that judgments based on introspection are infallible: Some sources of error may affect introspective beliefs—such as inattention—and errors of this type can also affect perceptual beliefs. The point is, rather, that self-attributions based on introspection are not liable to many sources of error affecting perceptual beliefs; for this reason, introspective beliefs can enable a high level of epistemic security or certainty according to acquaintance theorists.1

10  Self-Knowledge and Self-Consciousness In contrast to the acquaintance theory, the inner-sense theory points out the similarities between self-knowledge and perceptual knowledge. While outer sense relies on the model of perception or sensing, inner sense is construed through the process of introspection, which is considered to be on par with outer sense based on a causal process broadly similar to that involved in perceptual awareness. The theory’s classic frame of reference is Locke (1690, II.1.IV), who claims that This Source of Ideas, every Man has wholly in himself. . . . And though it be not Sense, as having nothing to do with external Objects; yet it is very like it, and might properly enough be call’d internal Sense. Assuming that inner sense is a casual process of introspection somewhat similar to that of perception, while vision and the other “outer” senses address the external world, inner sense is directed towards one’s own mind. For inner-sense theorists, the acquaintance theory’s tenet that introspective awareness involves a metaphysically direct and, hence, a non-causal relation to one’s own mental dimension is to be rejected. Since inner-sense theorists regard the knowledge of one’s mental states to be grounded in a perception-like process, and, consequently, introspective justification is considered tantamount to perceptual justification, there is no epistemic difference between this type of knowledge and knowledge about the external world. Aside from the metaphysical difference between the acquaintance and the inner-sense theories—while the former considers the relationship between introspection and the state introspected to be a direct process, the latter believes it to be causal—they are both empiricist theories of self-knowledge involving a non-inferential, introspective process.2 The rationalist theory of self-knowledge embraces the perspective that the subject is authoritative about her own states but rejects the empiricist orientation because it denies that first-person authority relies on the subject’s observation of her own mental states. While the empiricist perspective considers the subject to be passive and sees her as a mere observer of her own mental dimension, the rationalist theory of self-knowledge firmly rejects an interpretation of self-knowledge as a mere matter of observation; on the contrary, it acknowledges the subject to be authoritative because she is an agent who shapes her own states. Furthermore, since the subject uses this agency to keep her attitudes in line with her reasons, she has to be considered a mainly rational agent. Kant himself inspired the rationalist theory of self-knowledge.3 Unlike Descartes and Locke who, mutatis mutandis, employ an observational account of self-knowledge—the cogito argument depending on the subject observing that she is thinking, and Locke’s perceptual model of inner sense relying on introspection as an observational process—Kant points out the difference

Self-Knowledge and Self-Consciousness  11 between the observational account of one’s thoughts, achieved through inner sense, and the transcendental apperception device that accounts for one’s grasp of the activity of thinking. While in an observational account of self-knowledge the subject is seen as entirely passive regarding the thoughts she introspects, the rationalist theory of self-knowledge proposes an idea of agential self-knowledge whereby the subject is conscious of her agency, as well as of her power of thinking. As Kant states, “I exist as an intelligence which is conscious solely of its power of combination” (B158–9), of “the activity of the self” (B68). Following Brook (2008), “When one is conscious of oneself by doing cognitive and perceptual acts, one is conscious of oneself as spontaneous, rational, self-legislating, free—as the doer of deeds, not just as a passive receptacle for the contents of representations”. Accordingly, for Kant, the capacity for agential self-knowledge is a prerequisite for observational self-knowledge: The subject cannot know her own mental dimension by simply observing it unless she can grasp that she is a rational agent, i.e., the author of the thoughts. In this sense, the I is to be regarded as a rational agent. With regard to the role of reason in selfknowledge, the rationalist theory deals exclusively with the types of mental states that can be supported (or undermined) by reasons—namely, propositional attitudes such as beliefs, desires, intentions, etc.—since propositional attitudes can be justified by reasons: For example, the belief that there is a book can be justified by seeing a book. As Gertler remarks: The rationalist invokes this connection between attitudes and reasons in explaining how we know our attitudes. The basic idea is that, because propositional attitudes are tied to justifying reasons, we can determine our own attitudes by directly considering our reasons. If asked what I believe about the weather, I can consider the fact that I’m being drenched, and report that I believe that it’s raining. If asked whether I intend to continue exercising, I can consider the benefits of exercise and report that I intend to continue. Crucially, rationalism implies that one can know one’s own attitudes without reflecting on those attitudes: I need not reflect on my belief or intention, but only on the reasons that bear on them. (Gertler 2011, 19) The main distinctive features of self-knowledge are not epistemic but normative: The subject does not need to occupy any special position in order to observe her own attitudes; instead, being the only one who shapes her attitudes through rational agency, she is the only one responsible for them. Since the subject has to ensure that her attitudes combine well with her reasons, self-knowledge chiefly concerns a normative dimension that involves rights and responsibilities. As will soon be seen, such theories can explain the different approaches to self-consciousness.

12  Self-Knowledge and Self-Consciousness

The Question of Self-Consciousness The problem of the knowledge of one’s mental states centres on the involvement of the subjective self-conscious dimension. The fact that a subject acquires knowledge of her belief that Naples is a lovely city implies that this state is registered as her own; this is related to the question of selfconsciousness proper, one of the major topics in the philosophical arena. One can be self-aware in several ways, each of which corresponds to one of the several interpretations of the term “self” (i.e., the embodied self, the ecological self, the narrative self, etc.); nevertheless, at first glance, the selfconsciousness at issue is based on the consciousness of ourselves and of our personal mental dimensions, of thoughts taking place and of feelings being experienced. More specifically, the notion of self-consciousness that will be considered here can be referred to as basic self-consciousness. This involves two specific, correlated features that are not owned by the consciousness of things other than oneself: The first is related to the fact that self-consciousness is grounded in a first-person perspective, whereas the second concerns the fact that it should be considered a consciousness of the self as subject rather than a consciousness of the self as object. Both peculiarities are grounded in the possibility of using the term or concept ‘I’, as will be seen in the following paragraphs, in which a few epistemic and semantic features of the term or concept ‘I’ will be analysed. The first point can be explained by Baker’s approach. According to the author (cf. Baker 2000, 2013), two types of first-person perspectives can be distinguished: a rudimentary first-person perspective, manifested by many mammals and human infants, and a robust first-person perspective, manifested by language users who master first-personal language. The latter is the conceptual capacity to not only recognise oneself as distinct from things other than oneself, but also to conceive of oneself as oneself. A robust firstperson perspective is exactly what marks the difference between a creature with a rudimentary first-person perspective that can only be conscious of the environment and a fully self-conscious subject. In fact, a mature human subject with a robust first-person perspective can attribute a first-person reference on the basis of a self-concept to herself; not only can she refer to herself in the first person, she can also attribute a first-person reference to herself. In this way, it is possible to distinguish between consciousness in terms of a rudimentary first-person perspective and self-consciousness in terms of a robust first-person perspective. Mutatis mutandis, Peacocke (2014) distinguishes three degrees of self-representation: While the subject does not have mental states with I-contents in what he calls Degree 0, subjects with Degree 1 of self-representation have states with objective I-contents that represent the subject as standing in a spatial relationship to other objects and events in the spatial world. However, these contents are non-conceptual (cf. infra, Chapter 5); in other words, they are not based on the use of the concept I,

Self-Knowledge and Self-Consciousness  13 contrary to the Degree 2 of subject involvement in the representation of the objective world, in which the subject employs the first-person concept I. According to Baker (2013) and Matthews (1992), “I*” pronouns4 are used reflexively to pick out the subject from her own point of view: I*sentences are sentences containing “I*”, whereas I*-thoughts are thoughts expressible by I*-sentences. Via an I*-thought, a subject conceives of herself as herself*, and needs no third-person referential device, such as a name, a description, or a demonstrative to identify herself. As we will see in the following two paragraphs, certain semantic and epistemic features of the term/ concept ‘I’5 can be identified in this subject’s capacity for self-identification: Essential indexicality and immunity to error through misidentification. The former is relative to the meaning of the term/concept ‘I’ as any expression of self-consciousness is based on indexical terms such as “I” or “me”; the latter refers to the fact that some singular judgments involving the self-ascription of mental (and physical, as will be seen) properties are immune to error through misidentification relative to the first-person pronoun (IEM). As Wittgenstein (1958) and Shoemaker (1968) have pointed out, if I feel hungry, it is nonsensical to ask whether I am sure that I am the one who feels hungry. At the same time, the basic self-consciousness at issue here is also to be regarded as the consciousness of self-as-subject, or subject selfconsciousness, rather than as the consciousness of the self-as-object. Following Kriegel (2003, 2007), it is possible to make a distinction between the consciousness of oneself qua object and the consciousness of oneself qua subject. For instance, Mario can be conscious of Naples—Mario is the subject of the thought, and Naples its object. Mario, however, can also be conscious of himself*; in this case, Mario is both the subject and the object of the thought. Even though there is one single entity, and the subject and the object of the thought are the same, it is possible to draw a conceptual distinction between Mario’s ability as object of thought and Mario’s ability as subject of thought, namely the concepts of self-as-subject and selfas-object. James (1890) distinguishes between I and me; thus, we can tell the difference between “I am self-conscious that I think that p” and “I am self-conscious that me thinks that p”. While the former refers to the self-assubject, “me” refers to the self-as-object: Corresponding to these two concepts, or conceptions, of self, there would presumably be two distinct modes of presentation under which a person may be conscious of herself. She may be conscious of herself under the “I” description or under the “me” description. Thus, my state of self-consciousness may employ either the “I” mode of presentation or the “me” mode of presentation. . . . In the latter case, there is a sort of “conceptual distance” between the thing that does the thinking and the thing being thought about. Although I am thinking of myself, I am not thinking of myself as the thing that does the thinking. By contrast,

14  Self-Knowledge and Self-Consciousness in the former case, I am thinking of myself precisely as the thing that is therewith doing the thinking. (Kriegel 2007) This being the case, the self-as-subject may be interpreted as the thing that does the thinking, whereas the consciousness of oneself as a subject is the consciousness of oneself as the thing doing the thinking. It seems clear that the first-person perspective and the consciousness of the self-as-subject are two interdependent features, one being the condition of the other and vice versa. The subject’s manifestation of strong firstperson phenomena—one’s ability to attribute a first-personal reference to one’s self—is based on the subject’s possibility of being consciousness of herself as the thing doing the thinking, and the subject’s consciousness of self-as-subject cannot be attained unless the subject exhibits a manifestation of strong first-person phenomena. As has just been mentioned, these two features defining the notion of basic self-consciousness are grounded in a few epistemic and semantic peculiarities related to the ability to use the term or concept ‘I’ in I- or de se thoughts. These will be discussed in the following two paragraphs. Indexicality As mentioned previously, an I*-thought allows an individual to refer to herself as herself* without the need for third-person referential devices, such as names, descriptions, or demonstratives to identify herself: The term/concept ‘I’ employed in a self-conscious or I*-thought is essentiality indexical;6 as such, it necessarily involves information indexed to the context and—more specifically—to the thinker who has produced the thought. In detail, (1) ‘I’ is a singular term/concept; that is, a term with a single individual as its reference, (2) this term is governed by the token-reflexive rule, whereby every token of ‘I’ refers to the subject who has produced or used it, either mentally or linguistically, and (3) with the information available in context, and once the circumstances for evaluation are established, this rule is prima facie sufficient to determine its reference.7 More importantly, the indexical information about oneself based on the use of the term/concept ‘I’ cannot be reduced to non-indexical information; for this reason, indexicality is essential. Two well-known examples by Perry describe this issue. The first example concerns a fictional character named Rudolf Lingens: An amnesiac, Rudolf Lingens, is lost in the Stanford library. He reads a number of things in the library, including a biography of himself, and

Self-Knowledge and Self-Consciousness  15 a detailed account of the library in which he is lost. . . . He still won’t know who he is, and where he is, no matter how much knowledge he piles up, until that moment when he is ready to say, This place is aisle five, floor six, of Main Library, Stanford. I am Rudolf Lingens. (Perry 1977, 479) Amnesiac Rudolf Lingens can gather all sorts of information about himself by reading the books in the library, and yet no such information can provide him with the missing conceptual tool he needs to link the information to himself. In other words, there is no logical connection between third-person descriptive information, no matter how detailed, and a first-person grasp of oneself through the use of ‘I’. The second example concerns indexical judgments (beliefs and desires), which are crucial to explain and predict the motivating action: I once followed a trail of sugar on a supermarket floor, pushing my cart down the aisle on one side of a tall counter and back the aisle on the other, seeking the shopper with the torn sack to tell him he was making a mess. With each trip around the counter, the trail became thicker. But I seemed unable to catch up. Finally it dawned on me. I was the shopper I was trying to catch. (Perry 1979, 3) In the example, Perry’s thoughts (1) “the shopper with the torn sack is making a mess” and (2) “I am making a mess” refer to the same individual; thus, two intrinsically different kinds of self-reference are at play here. In the former, self-reference is external and available in the third person only: Babis can refer to an object by using a name, a definite description, or a demonstrative, and the object to which he is referring might be himself; there is no difference between this kind of self-reference and the reference made to an object that is different from oneself. An external self-reference can occur without the subject realising he is referring to himself, as in the first thought formulated by Perry, when he does not realise that he is the very shopper with the torn sack who is making a mess. Instead, the internal self-reference expressed by the second thought (2) “I am making a mess” produces an authentic I-thought that is only accessible from a first-person perspective because it is based on the use of ‘I’. As soon as he realises that he is the shopper in question, Perry produces a new thought, which he terms locating belief, based on the use of the essential mental indexical ‘I’. This entails that the ascription of an authentic I-thought to oneself cannot be achieved without the concept ‘I’, as there is no way of thinking an I-thought other than via indexical reference; hence, I-thoughts are irreducible to the other kind of non-indexical thoughts. Accordingly, by articulating an I-thought in a propositional way, an I-thought will contain a content of which the subjective essential indexical reference is expressible in the natural language by the personal pronoun

16  Self-Knowledge and Self-Consciousness ‘I’. This thought can be reported in either direct or indirect form: While the former is in oratio recta and reports the above-mentioned example as in (1) “I am making a mess” (thought by Perry), in the indirect form, the thought can be expressed in oratio obliqua and the report will be in the third person: (2) “Perry thinks that he himself is making a mess”. In turn, this sentence can be interpreted as the report of still yet another thought: (3) “Perry thinks that Perry is making a mess”. Obviously, it is possible to employ a definite description, “the φ” that picks out Perry uniquely, as in this example: (4) “Perry thinks that the author of The Essential Indexical is making a mess”. The thought expressed in (1) is not equivalent to (3) or to (4). In cases (3) and (4), Perry might be amnesiac and not remember his name or being the author of The Essential Indexical. It is only in (1) that an authentic I-thought is present: The subject who thinks the thought “I am making a mess”, provided that she knows the rule associated with ‘I’, cannot use it without realising that she is referring to nobody but herself. Although (1) and (3) or (4) are not equivalent, prima facie (2) seems to be a report of both. To capture the difference, Castañeda employs two different uses of the third person pronouns in oratio obliqua sentences. In the first case, to make (2) equivalent to (1), the pronoun is to be used in an indirect reflexive modality (Anscombe 1975) or as a quasi-indicator—the above-mentioned artificial pronoun (‘she*’, ‘he*’, ‘it*’) introduced by Castañeda (1966, 1967, 1968) to attribute a first-person essential indexical use from a third-person angle: (2.1) “Perry thinks that he* himself is making a mess”. The quasiindicator “he*” in the example is used as an anaphora, and its reference is not determined directly but only through the propositional attitude subject. In the other case, assuming that Perry is amnesiac, to make (2) equivalent to (3), “he” will not be employed as a quasi-indicator but as a simple indexical; thus, (2) is the report of Perry’s belief that someone else in the context (named Perry, too) is making a mess: Perry has not realised that it is he* himself who is doing that. What matters here is the fact that token-reflexive expressions such as firstperson pronouns and quasi-indicators are essential indexicals: They cannot be eliminated or replaced by a name, description, or demonstrative without losing the content expressed by the sentences/thoughts that contain them. To refer to (to think of) oneself qua oneself, the subject has to use the essential indexical ‘I’. As Shoemaker (1968, 560) summarised, there is no description at all which is free of token-reflexive expressions and which can be substituted for ‘I’; no matter how detailed a tokenreflexive-free description of a person is, . . . it cannot possibly entail that I am that person. In other words, essentiality lies in the fact that the use of ‘I’ is not based on cognitive mediation but is indispensable or essential in order to form

Self-Knowledge and Self-Consciousness  17 I-thoughts. This means that an identifying description is not a necessary or a sufficient condition for the self-reference of the self-conscious subject. In one case, an amnesiac subject locked up in a completely dark room was able to use I in order to refer to herself even though she could not apply thirdperson descriptions to herself (cf. Castañeda 1999, 268–9). In the other case, as has already been stated by Shoemaker and suggested by Perry’s example of the amnesiac Rudolf Lingens, no matter how detailed a thirdperson description about a certain subject is, the latter can grasp it without realising she is the subject to whom the description refers. The self-reference expressed in linguistic or mental self-ascriptions such as (2.1) is referred to as de se (i.e., of oneself), an expression used in Lewis (1979) and Castañeda’s works, and is semantically unique: The linguistic or mental self-reference is irreducible to either a de dicto or a de re linguistic or mental reference (for the above considerations, the truth conditions of the de se “Perry thinks that he* himself is making a mess” may be different from both de dicto “Perry thinks that Perry is making a mess” and de re “Perry thinks, of Perry, that he is making a mess”). If de se thoughts are mental states reported in de se reports, they are irreducible to mental states reported in de dicto and de re reports. In fact, first-person and third-person self-ascriptions may present a different cognitive significance for the subject, as self-ascriptions involving the first-person pronoun have immediate implications for action. If de se thoughts can only be entertained by a subject who is aware that the thoughts are self-referring, this means that de se thoughts are non-accidentally selfreferring: Perry thinks “I am making a mess!” only when he realises that it is he himself who is carrying a leaking sugar bag because, in order to produce a self-ascription involving the first-person pronoun, he has to know that “I” necessarily refers to himself when he is entertaining this type of self-ascription. This self-referring thought is non-accidental and will prompt him to stop the cart and fix the leak. For this reason, in contrast to de re or de dicto thoughts, de se thoughts have immediate relevance for action. As we will also see later, I-thoughts constitute an irreducible class of mental phenomena, some of which have been articulated in transcendental terms by Kant. Immunity to Error Through Misidentification In a well-known passage, Wittgenstein (1958, 66–7) introduces his philosophico-linguistic analysis of the grammatical rule of the term ‘I’, where he identifies two types of uses, i.e., ‘I’ used as object (“I have grown six inches”) and ‘I’ used as subject (“I have toothache”): One can point to the difference between these two categories by saying: The cases of the first category involve the recognition of a particular person, and there is in these cases the possibility of an error. . . . On the other hand there is no question of recognizing a person when I say

18  Self-Knowledge and Self-Consciousness I have toothache. To ask ‘are you sure it’s you who have pains?’ would be nonsensical. According to some theses (some of which will be examined in Chapter 3), this passage should be seen as part of the philosophical framework articulated by Wittgenstein since the 1930s and should be regarded as the background to the analyses of the two uses of ‘I’. While the I used as object performs a referential function relative to one’s body and physical features in general, the I used as subject apparently regards mental states and processes involving no subject identification. Particular judgments displaying first-person reference (e.g., “I have pain”) display what Shoemaker (1968, 565) defines as self-reference without identification: My use of the word ‘I’ as the subject of my statement is not due to my having identified as myself something of which I know, or believe, or wish to say, that the predicate of my statement applies to it. The self-ascription of the thoughts on which the self-consciousness is based considers the consciousness of oneself qua subject—that is, as the subject of every thought or mental state—rather than as the object based on the previous identification. Due to the absence of any identification component, particular singular judgments involving the self-ascription of mental (and physical, as will be seen) properties are immune to error through misidentification relative to the first-person pronoun (IEM). For example, if a subject feels pain and judges “I am in pain”, the subject formulating such judgments cannot be mistaken as to the person who happens to be in pain. To say that a statement ‘a is Φ’ is subject to error through misidentification relative to the term ‘a’ means that the following is possible: the speaker knows some particular thing to be Φ, but makes the mistake of asserting ‘a is Φ’ because, and only because, he mistakenly thinks that the thing he knows to be Φ is what ‘a’ refers to. The statement ‘I feel pain’ is not subject to error through misidentification relative to ‘I’: it cannot happen that I am mistaken in saying ‘I feel pain’ because, although I do know of someone that feels pain, I am mistaken in thinking that person to be myself. (Shoemaker 1968, 557) To make this passage clearer, we might consider an example of identification-dependent thought, e.g., Laura’s thought that her neighbour is a nice person. Following Evans (1982) and Kriegel (2007), the structure of this thought consists of an identification component and a predication component, which can be explained by Laura’s first-person perspective as follows: “My neighbour [identification component] is a person who smiles

Self-Knowledge and Self-Consciousness  19 at me every day and the person who smiles at me every day is a nice person [predication component]”. Here, two types of errors are possible. Laura can be mistaken as to the predicational component, i.e., that her neighbour is a nice person: For example, Laura later discovers that her neighbour’s tendency to smile is nothing more than a cynical strategy to have her consent to cut the trees in the garden. Laura can also be mistaken as far as the identificational component is concerned—i.e., with respect to the person who is her neighbour—and, for example, become confused and mistake the postman for her neighbour. On the other hand, there is a class of self-ascriptions that do not involve identification-dependent thoughts; as such, this class is not subject to error through misidentification. Shoemaker (1968, 565) examines the kinds of psychological predicates involved in such self-ascriptions: There is an important and central class of psychological predicates, let us call them P* predicates, each of which can be known to be instantiated in such a way that knowing it to be instantiated in that way is equivalent to knowing it to be instantiated in oneself. For instance, the judgment “I have pain” is IEM because the way in which the predicate is expressed (“there is pain”), that is, based on our own subjective experience, will suffice to realise that it is ascribed to ourselves (“I have pain”). It is in this particular sense that “there is pain” is tantamount to “I have pain”. Evans, in turn, goes beyond the terms of the matter as suggested by Wittgenstein and, to some extent, by Shoemaker. In particular selfascriptions, the self-reference is direct and unmediated: As Evans notes, this is identification-free self-reference. More to the point, and moving from the self-ascription of properties that are not only mental but also physical, the author discloses his approach: Judgments are IEM when they result from the connection between the information acquired in the first person and its justification, in contrast to the identification-dependent judgments involved in the ordinary perception of external objects. The IEM feature does not depend on the kind of predicate involved in the self-ascription, but on the epistemic and justification ground on which the subject produces such judgments8 in a context in which—from Strawson’s lesson onward—the subject is conceived of as a spatio-temporally located object. Evans (1982, 220) contends that a judgment such as “I am F” is identification-free unless it corresponds to the inferential conclusion drawn from the two premises, i.e., “a is F” (predication component) and “I am a” (identification component). Such a judgment is based on the unmediated self-ascription of properties through the introspective consciousness (as is the case with mental properties) or proprioception (as with physical properties). For example, according to our general capacity to perceive bodies, and to our sense of proprioception, of balance, of heat and cold, and

20  Self-Knowledge and Self-Consciousness of pressure, the kind of information generated by each of these modes of perception seems to give rise to judgments with immunity to error through misidentification: “None of the following utterances appears to make sense when the first component expresses knowledge gained in the appropriate way: ‘Someone’s legs are crossed, but is it my legs that are crossed?’ ”. Peacocke’s (2008) strategy, in turn, consists of associating IEM properties with more fundamental characterisations. By specifying the manner and circumstances in and under which an IEM judgment relative to the occurrence of a particular concept is formed, Peacocke pushes the characterisations of the functional rule of first-person concepts highlighted by Evans into the background. In addition, in his last book, Peacocke (2014, 107) contends that the best explanation for the phenomenon of judgments formed according to the self-attributions of mental and physical properties relies on the basic reference rule for which “what makes someone the reference of I in a thinking is that he or she is the producer of that thinking”. This is Peacocke’s (2014, 109) criterion for immunity to error through misidentification of a judgment ‘I’m F’ when reached in a certain way: [A] thinker can rationally come to make the judgement in that way without relying on some identity I = m, for some individual mode of presentation m distinct from the first person. . . . In paradigm cases of first person immunity to error through misidentification, you can know I am F without reliance on any substantive empirical identity involving yourself.9 Therefore, the capacity for self-consciousness depends on the possibility of producing I-thoughts that employ indexical self-references immune to error through misidentification relative to the term/concept I. As Musholt (2015, 14) points out, if not all self-conscious thoughts are immune to misidentification, as those that involve the use of “I” “as object” are based on an identification component, there must be some self-conscious thoughts that are immune in order to explain the possibility of self-conscious thoughts occurring at all. Self-consciousness seems to succeed in explaining these features in Kant, since he does consider the above-mentioned feature of the concept I in some way, as we will see later.

Kant’s Theory of the Mind: Faculties and Representations In the previous paragraph, the notion of basic self-consciousness was defined as the subject’s capacity to have a first-person perspective and a consciousness of herself as subject, which are both grounded in the particular epistemic and semantic features of the term or concept ‘I’ used in I- or de se thoughts. Kant points to two forms of self-consciousness: The inner sense, or empirical apperception, grounded in a sensory form of self-awareness, and

Self-Knowledge and Self-Consciousness  21 transcendental apperception. Through the notion of inner sense, Kant also allows for an introspective account of self-awareness; nonetheless, the point and purpose of this book are to show that an utterly sophisticated notion of basic self-consciousness is, in fact, provided for by the notion of transcendental apperception. As we will see later, the doctrine of apperception is not to be confused with an introspective psychological approach: In reality, it is a formal model for the thinking activity. In his approach to self-consciousness, Rödl (2007) points out this distinction very well by focusing on the link connecting subjectivity, self-consciousness, and thinking: self-consciousness is a form of subjectivity of a thinking subject, and this is really a Kantian point: An investigation of self-consciousness cannot start with sensation because sensation does not constitute the kind of subjectivity we call self-consciousness. A sentient creature bears a special nexus to her sensations, which we may call consciousness. Its description would be an account of the subjectivity of the sentient living being, the animal. This subjectivity is not self-consciousness. Self-consciousness is a relation a subject bears to herself by virtue of being a subject of thought. This means that it is in the first instance a character of the nexus a subject bears to her acts of thinking, of the way in which she represents herself as thinking. Acts of sensibility of a thinking subject are part of the content of her self-consciousness, which shows that, in her, the nexus of subject and sensation has a different form from the one it has in an animal without thought. An inquiry into the nexus that a subject bears to her sensations in virtue of being able to think them inquires into the sentience of a thinking subject and presupposes an understanding of the subjectivity of such a subject. Hence, thoughts about sensations are not the place to begin when investigating self-consciousness. The primary topic of a theory of self-consciousness is thought about thought. (Rödl 2007, 12–13) Kant’s views on the mind and self-consciousness are mainly developed in the Critique of Pure Reason. Other important points are expounded in the late Anthropology From a Pragmatic Point of View, first published in 1798, worked up from notes for popular lectures, and in several other passages dating to the critical period. In KrV, the mind and self-consciousness are only discussed in relation to major projects—epistemic and metaphysical in particular—but never in their own right. Therefore, the arguments appear to be somewhat dispersed. The two parts of KrV containing Kant’s main remarks about the mind and self-consciousness are the chapters on Transcendental Deduction and Paralogisms; as is well known, Kant completely revised both chapters for the second edition. The faculties and the different kinds of representations will be singled out in this paragraph in an attempt to explain the conditions for ‘cognition’

22  Self-Knowledge and Self-Consciousness (Erkenntnis), namely for the mind’s relationship to a possible object of experience. The main features of empirical apperception will be examined in the “The Inner Sense” section, while transcendental apperception will be discussed in relation to Transcendental Deduction in the “The Transcendental Deduction and the Principle of the Unity of Apperception” section in order for the property Kant attributes to the subject’s self-consciousness and its difference from the object’s self-consciousness of the inner sense can be developed in the following chapters. According to Kant, the mind consists of two fundamental capacities (Fähigkeiten), namely ‘receptivity’ (Receptivität) and ‘spontaneity’ (Spontaneität): If we will call the receptivity of our mind to receive representations insofar as it is affected in some way sensibility, then on the contrary the faculty for bringing forth representations itself, or the spontaneity of cognition, is the understanding. It comes along with our nature that intuition can never be other than sensible, i.e., that it contains only the way in which we are affected by objects. The faculty for thinking of objects of sensible intuition, on the contrary, is the understanding. (B75/A51) While the former is the mind’s capacity to be affected by something, be this itself, or something outside of itself, the power of spontaneity can initiate its activity by itself without external triggers.10 From both the affection of receptivity and the activity of spontaneity, ‘representations’ (Vorstellungen) can be regarded as the discrete mental events or states, as well as the vehicles and results of all mental activities. Moving from these two very general aspects of the mind, Kant introduces three specifications of receptivity and spontaneity, three faculties or ‘powers’ (Vermögen): ‘sensibility’ (Sinnlichkeit), ‘understanding’ (Verstand), and ‘reason’ (Vernunft). As will be seen, the power of judgment (Urteilskraft) is, in turn, a central cognitive faculty of the human mind; it is a spontaneous and innate cognitive “capacity” (Fähigkeit). As a “faculty of judging” (Vermögen zu urteilen) (A69/B94), Kant also considered it to be the “faculty of thinking” (Vermögen zu denken) (A81/B106). These characterise specific cognitive powers that have a particular cognitive task: While sensibility is the faculty through which human beings—as well as other animals—are receptive, the faculties of the understanding and reason characterise spontaneity as a form unique to human and rational beings. Such faculties deal with and produce different representations: one particular passage, generally referred to as the Stufenleiter (“stepladder”) passage from the Transcendental Dialectic of KrV, explains and classifies the varieties of representation: The genus is representation in general (repraesentatio). Under it stands the representation with consciousness (perceptio). A perception that

Self-Knowledge and Self-Consciousness  23 refers to the subject as a modification of its state is a sensation (sensatio); an objective perception is a cognition (cognitio). The latter is either an intuition or a concept (intuitus vel conceptus). The former is immediately related to the object and is singular; the latter is mediate, by means of a mark, which can be common to several things. (A320/B376) ‘Sensations’ (Empfindungen), ‘intuitions’ (Anschauungen), and ‘concepts’ (Begriffe) are the main kinds of representation. While sensibility is the faculty of sensory representations, sensations, and intuitions, because they are affected either by entities distinct from the subject or by the subject itself, the faculty of understanding deals with conceptual representations in a spontaneous way. Reason is also a spontaneous faculty that is concerned with concepts, called ‘ideas’, the objects of which can never be met in ‘experience’: In fact, some of these ideas include God and the soul. Concepts and intuitions11 can be pure as well as empirical; in particular, an empirical intuition is the product of both ‘sensation’ (Empfindung)—the effect produced by an external object on the capacity of sensible representation, i.e., the matter of sensibility—and pure intuitions, namely time and space, i.e., the a priori pure forms in which the sensible matter is organised. At the empirical level, it seems possible to contend that an object is only given through sensibility, which produces intuitions by virtue of a relationship with the object. Such a relationship depends on the sensations forming the matter of the intuition; they are articulated before the intervention of thought through the pure forms of sensibility—time and space—in order to produce a singular representation that is immediately linked to the object. Due to the pure forms of sensibility, the formal element of an empirical intuition is either spatial-temporal or temporal. Kant connects the two forms of intuition to two distinct spheres or senses: The ‘outer sense’ and the ‘inner sense’. The outer sense—the form of which is space—involves the spatial world of material objects, whereas the inner sense—the form of which is time—involves the inner intuition that orders the states of mind (A22/B37; cf. Anth 7:154) temporally: Sensibility in the cognitive faculty (the faculty of intuitive representations) contains two parts: sense and the imagination. . . . But the senses, on the other hand, are divided into outer and inner sense (sensus internus); the first is where the human body is affected by physical things, the second is where the human body is affected by the mind. (Anth 7:153; 45) If concepts and intuitions are two distinct types of objective representations, then they are both necessarily involved in the judgment for the determination of objective knowledge from an epistemic perspective: According to the well-known adage, “Thoughts without content are empty, intuitions

24  Self-Knowledge and Self-Consciousness without concepts are blind”. The togetherness principle, whereby knowledge can only be produced through the joint intervention of concepts and intuitions, was supported by McDowell’s Kant-inspired conceptualist position: Our cognition arises from two fundamental sources in the mind, the first of which is the reception of representations (the receptivity of impressions), the second the faculty for cognizing an object by means of these representations (spontaneity of concepts); through the former an object is given to us, through the latter it is thought in relation to that representation (as a mere determination of the mind). Intuition and concepts therefore constitute the elements of all our cognition, so that neither concepts without intuition corresponding to them in some way nor intuition without concepts can yield a cognition. (A50/B74)12 At this point, it is necessary to recall the two theoretical dimensions under discussion and the relative distinction between the transcendental and the empirical dimensions. While the latter concerns the formation of empirical concepts and intuitions, the former accounts for both the a priori principles of sensibility—the pure forms of sensibility, i.e., time and space—and the a priori principles of the intellect. These are the twelve categories discussed in the analytic of concepts and listed in the metaphysical deduction under four titles (‘quantity’, ‘quality’, ‘relation’, ‘modality’). Their use is also legitimised in the Transcendental Deduction through the ‘transcendental apperception’ and is associated with the logical function of judging, whereby a cognition can arise. The categories, that is the pure concepts of the understanding, are employed by the faculty of judgment (Urteilskraft), seen as “the capacity to subsume under rules, that is, to distinguish whether something falls under a given rule” (A132/B171). The question here concerns the application of the categories to the intuitions of the sensibility, marking a difference between the faculty of understanding and the faculty of judgment. While the former is the faculty of the rules (das Vermögen der Regeln), the latter is the capacity to subsume under rules. This shift takes place through the synthesis of the imagination: As it mediates between the faculties of sensibility and understanding, the synthesis of the imagination plays a key role in the application of the concept. Kant calls this mediating role the “transcendental function” of the imagination (A124). Even though the faculty of judgment is a non-basic faculty, is nevertheless the central cognitive faculty of the human mind. In fact, judging combines all the acts and contents of intuition, conceptualisation, imagination, and reason via transcendental apperception in order to generate a single cognitive product. As Brook (2008) points out, for Kant, rational humans are judging animals.

Self-Knowledge and Self-Consciousness  25

The Inner Sense An Introspective Account of Self-Consciousness Similar to Hume and Locke’s view,13 according to which internal sense is reflection, i.e., “the notice which the Mind takes of its own Operations” (Locke, II,1, § IV),14 Kant’s theory of inner sense uses an introspective account of self-consciousness, and in part, this approach to self-knowledge is subject to the same transcendental conditions as the knowledge of an object given in the outer sense. As we have just seen, Kant associates the two pure forms of intuition with two distinct spheres or senses, the ‘outer’ and the ‘inner’ sense. While the former is grounded in the pure form of space and regards the spatial world of material objects, the latter is grounded in the pure form of time and concerns the inner intuition, which orders the mental states temporally (A22/B37). With regard to this specific point, Kant distinguishes the different functions of inner and outer sense. Due to the pure forms of sensibility, appearance—the object of the empirical intuition—will be either spatial-temporal or temporal: Time is the a priori formal condition of all appearances in general. Space, as the pure form of all outer intuitions, is limited as an a priori condition merely to outer intuitions. But since, on the contrary, all representations, whether or not they have outer things as their object, nevertheless as determinations of the mind themselves belong to the inner state, while this inner state belongs under the formal condition of inner intuition, and thus of time, so time is an a priori condition of all appearance in general, and indeed the immediate condition of the inner intuition (of our souls), and thereby also the mediate condition of outer appearances. (A34/B50) In the a priori formal condition of time, all possible representations are revealed; in space, however, there is only one specific sub-class of representations—those referring to sheer external objects. Due to space, the form of the outer sense, objects are represented as something real and different from the subject. It is only through a spatio-temporal allocation that the object is knowable in the strict sense, and the relative representations become Erkenntnis through the application of the conceptual–discursive dimension. As a result of time—the form of the inner sense—the subject can become aware of the representations as its own: As with the condition of all appearances in general, time is the immediate condition of the inner intuition—“of our souls”—and the mediate condition of outer appearances. Let us now examine the first important passage about the self-awareness function of the inner sense: By means of outer sense (a property of our mind) we represent to ourselves objects as outside us, and all as in space. In space their form,

26  Self-Knowledge and Self-Consciousness magnitude, and relation to one another is determined, or determinable. Inner sense, by means of which the mind intuits itself, or its inner states, gives, to be sure, no intuition of the soul itself, as an object; yet it is still a determinate form, under which the intuition of its inner state is alone possible, so that everything that belongs to the inner determinations is represented in relations of time. (A22–3/B37) Here, Kant seems to be assigning two distinct and yet deeply connected roles to the inner sense, as also summarised in passage A33/B50, in which time is described as a form of the inner sense: “Time is nothing other than the form of inner sense, i.e., of the intuition of our self and our inner state” (cf. also B55, where Kant states that the objects of “our inner sense” are “myself and my state”). With regard to the first role, the subject intuits itself through the inner sense, whereas with the second, it intuits its representations or states; in other words, due to the capacity to be affected by the determinations of inner states, the subject can recognise them as states of itself. These two points are extremely important and have generated much controversy among scholars.15 With regard to the former, Kant specifies a Humean approach by introducing a few different and, at first sight, contrasting features of the inner sense. Even if Kant asserts that the subject can intuit itself due to the inner sense in more than one passage, in the above-mentioned passage A22–3/ B37, it is just as evident that the subject cannot intuit its soul or mind— its essential nature—as an object, since no object encountered in the inner experience can be singled out. The point is made again in one of the most famous Kantian passages about the inner sense, which brings to mind Hume’s approach:16 The consciousness of oneself in accordance with the determinations of our state in internal perception is merely empirical, forever variable; it can provide no standing or abiding self in this stream of inner appearances [es kann kein stehendes oder bleibendes Selbst in diesem Flusse innrer Erscheinungen geben], and is customarily called inner sense or empirical apperception. (A107) This particular feature of what is referred to as empirical subject or self in the debate will be considered several times in later chapters, in which the contrasting features of the notion of ‘transcendental subject’ will be examined. As we will see, Kant introduces two Is: The ‘I as subject of thinking’ relating to pure apperception, and the ‘I as object of perception’, or empirical subject, relating to the inner sense, which contains various determinations that make inner experience possible. If apperception is the intellectual

Self-Knowledge and Self-Consciousness  27 consciousness of thinking or spontaneity, the inner sense is the sensory consciousness of the self and its thoughts. Even though the empirical subject or ‘I’ considered by the inner sense in the empirical dimension is not represented as a standing or abiding object because the subject does not represent itself to itself as a spatio-temporal object, Kant often asserts that, due to the inner sense, the subject intuits itself as an appearance nonetheless. However, how can we reconcile these contrasting features? The Inner Sense and the Ideality of Time To shed light on this point, it is necessary to move from the doctrine of the ideality of time, which we will also investigate later in the so-called selfaffection argument. In the following passage, Kant compares the external affection of objects in space to the self-affection of the inner sense despite their different features, as we will see later. The point at issue is that the subject deals with an appearance through the inner sense; hence, it cognises itself as appearance in the same way in which it deals with appearances referring to external objects. In both cases, however, the subject does not deal with the thing itself: We must order the determinations of inner sense as appearances in time in just the same way as we order those of outer sense in space; hence, if we admit about the latter that we cognize objects by their means only insofar as we are externally affected, then we must also concede that through inner sense we intuit ourselves only as we are internally affected by our selves, i.e., as far as inner intuition is concerned we cognize our own subject only as appearance but not in accordance with what it is in itself. (B156) When comparing this excerpt to the Humean passage A107, the contrast is apparent. On one hand, there is no standing or abiding self in the inner sense, since no object can be distinguished as a self; on the other hand, Kant claims that the subject only cognises itself as appearance, not as it is in itself. More specifically, the point of contrast is that if the subject cognises or intuits itself as appearance, then something must have produced it. If there is no self-object producing such an appearance in the same way that the external objects affect the outer sense, then something else has to play this role. So, what is this something that produces the appearance through which the subject cognises itself as appearance in the inner sense? To answer this question, and eventually reconcile the various features, it is necessary to recollect the second role assigned by Kant to the inner sense: The subject’s intuition of its own representations or states.

28  Self-Knowledge and Self-Consciousness With regard to this particular role, a few important distinctions require our attention. In general, the inner sense has no separate manifold of its own. The matter of inner intuitions consists of no sensory impressions produced by or grounded in the self, much in the same way that the matter of outer intuitions consists of impressions produced by the affection from outer objects; as just mentioned, there is no empirical intuition of the self as an external object, since the self cannot be encountered as a spatio-temporal object. For this reason, according to Kant, the manifold of the inner sense properly consists of outer sense representations: “the representations of outer sense make up the proper material with which we occupy our mind” (B67). As a result, the inner sense is to be regarded as the limiting condition through which the empirical manifold given in outer sense may be attained: “I exist as an intelligence . . . which, in regard to the manifold that it is to combine, is subject to a limiting condition that it calls inner sense” (B158). It is now possible to connect the two different functions of the inner sense: The intuition of the self and the intuition of the representations. If the inner sense is the means through which the subject intuits itself—although only in these terms—but there is no self-object originating or causing the sensory impression, then only the empirical manifold given in the inner sense can produce the appearance of the self. However, as the manifold of the inner sense properly consists of the representations of the outer sense, in order to realise the cognition of the self as appearance—as held by Kant in several passages—the only representations that can achieve the goal are the introspectively available ones, and these are nothing but the representations derived from the outer sense. This point is related to the important self-affection argument which, in turn, is linked to the doctrine of the ideality of time and, in particular, with the fact that, in the inner sense, the subject cognises itself only as appearance. This argument is pivotal from an exegetical perspective; in this paragraph and in the following chapters, only the specific consequences of the theory of self-consciousness and self-knowledge will be explored in order to examine the difference between the inner sense and transcendental apperception in greater depth. It is through the self-affection argument that the distinction between an active I (of transcendental apperception) and a passive I (of the inner sense) is singled out for the first time. The double I, both active and passive, is introduced by Kant as a kind of paradox: Here is now the place to make intelligible the paradox that must have struck everyone in the exposition of the form of inner sense (§ 6): namely how this presents even ourselves to consciousness only as we appear to ourselves, not as we are in ourselves, since we intuit ourselves only as we are internally affected, which seems to be contradictory, since we would have to relate to ourselves passively; for this reason it is customary in the systems of psychology to treat inner sense as the same as the faculty of apperception (which we carefully distinguish). (B153)

Self-Knowledge and Self-Consciousness  29 Here, Kant characterises self-affection, or affection of the inner sense, as the determination of the inner sense through the understanding under the designation of a transcendental synthesis of the imagination (B153), or figurative synthesis. The understanding exerts its action, the inner sense is thereby affected, and, accordingly, the subject becomes passive to itself. In this way, the active subject coincides with apperception and its synthetic unity: Apperception . . . is so far from being the same as the inner sense that the former, rather, as the source of all combination, applies to all sensible intuition of objects in general, to the manifold of intuitions in general, under the name of the categories; inner sense, on the contrary, contains the mere form of intuition, but without combination of the manifold in it, and thus it does not yet contain any determinate intuition at all, which is possible only through the consciousness of the determination of the manifold through the transcendental action of the imagination (synthetic influence of the understanding on the inner sense), which I have named the figurative synthesis. (B154) Kant thus distinguishes the ‘I’ as ‘I think’, or ‘intelligence’ or ‘thinking subject’, given in transcendental apperception, and the ‘I that intuits itself’ as two different dimensions of the same mind or subject.17 The ‘I’ as intelligence cognises itself as an object being thought, insofar as the subject is also given to itself in intuition, as with other phenomena, “not as I am for the understanding but rather as I appear to myself”. Nonetheless, as has just been mentioned, the subject given to itself as appearance in intuition must be regarded as the result of the combination or determination of the inner sense by understanding. In the following chapters, the self-affection argument will be reconsidered in a broader sense with reference to the question of empirical self-knowledge and the embodied subject. The Epistemic Limits of the Inner Sense As mentioned previously, all the representations that can be ascribed to the subjective dimension are related to the inner sense and conform to its temporal form. Although the inner sense is the intuition of one’s own self and inner state—in a way that has demanded and will demand further explanation— unlike Descartes and Hume, Kant rejects the role of the inner sense from an epistemological perspective because he does not view it as the main or basic self-consciousness. Firstly, Kant makes several remarks about the empirical study of the mind: In The Metaphysical Foundations of Natural Science in particular, he harshly criticises introspection as an empirical method for engaging in psychology. In debating the epistemic status of natural science, Kant asserts that “in any special doctrine of nature there can be only as much proper science as there is mathematics therein” (MAN 4: 470, 6): Proper science, and above all proper natural science, requires a pure part

30  Self-Knowledge and Self-Consciousness as the basis of the empirical part. Here, Kant attacks the epistemic status of chemistry (in his time, the development of a single, unified theory of chemical reactions was yet to be accomplished) and asserts that “the empirical doctrine of the soul must remain even further from the rank of a properly so-called natural science than chemistry” (MAN 4: 471, 7). There are several reasons for this conclusion, the first being that the appearances of pure inner intuition are only constructed within the dimension of time; since the contents of the inner sense cannot be quantified, mathematics cannot be applied to the phenomena of the inner sense and its laws. Accordingly, the empirical doctrine of the soul can also never approach chemistry even as a systematic art of or experimental doctrine, for in it the manifold of inner observation can be separated only by mere division in thought, and cannot then be held separate and recombined at will (but still less does another thinking subject suffer himself to be experimented upon to suit our purpose), and even observation by itself already changes and displaces the state of the observed object. Therefore, the empirical doctrine of the soul can never become anything more than an historical doctrine of nature, and, as such, a natural doctrine of inner sense which is as systematic as possible, that is, a natural description of the soul, but never a science of the soul, nor even, indeed, an experimental psychological doctrine. (MAN 4: 471, 7) If the only thinking subject whose inner sense one can investigate is oneself, and even if observation alters and distorts the state of the object observed, then the contents of introspection within the inner sense cannot be analysed scientifically. The main point is that empirical psychology cannot achieve any necessary truth about the mind, nor be assigned any role in the transcendental task. As just seen in B153, Kant warns psychologists to avoid confusing inner sense and apperception; the latter is an active faculty, whereas inner sense is a passive faculty. Kant introduces two different accounts on self-consciousness, contrasting empirical and transcendental apperception. A passage from Anthropology is particularly self-evident: If we consciously represent two acts: inner activity (spontaneity), by means of which a concept (a thought) becomes possible, or reflection; and receptiveness (receptivity), by means of which a perception (perception), i.e., empirical intuition, becomes possible, or apprehension; then, consciousness of one’s self (apperception) can be divided into that of reflection and that of apprehension. The first is a consciousness of the understanding, pure apperception; the second a consciousness of the inner sense, empirical apperception. In this case, the former is falsely named inner sense.—In psychology we investigate ourselves according

Self-Knowledge and Self-Consciousness  31 to our ideas of the inner sense; in logic, according to what intellectual consciousness suggest. Now here the “I” appears to us as to be double (which would be contradictory): 1) the “I” as subject of thinking (in logic), which means pure apperception (the merely reflecting “I”), and of which there is nothing more to say except that it is a very simple idea; 2) The “I” as object of the perception, therefore of the inner sense, which contains a manifold of determinations that make an inner experience possible. (Anth 7: 135, 23) The double I—as subject and as object—will be one of the issues discussed in the following chapters because it also mirrors the distinction between subject and object self-consciousness. For instance, as McLear (2011) highlights, due to inner sense, an animal can be aware of being in pain, and yet this cannot be regarded as subject self-consciousness. A subject can be conscious of herself in different inner mental changes, but only on condition that she represents herself as one and the same subject in those changes, and this is only possible due to transcendental apperception. At the same time, Kant restricts the use of the empirical kind of selfconsciousness, particularly in his main philosophical concerns: Even though a subject can be aware of its own states via the inner sense as a temporally ordered series of mental states, it cannot be aware of the series as a whole. In a Kantian passage, the inner sense is said to contain “the mere form of intuition, but without combination of the manifold in it, and thus it does not contain any determinate intuition at all” (B154). This entails that the mere form of intuition does not establish what representations represent; in other words, what they are representations of. The point is stated in a letter by Kant to Marcus Herz in 1789: [Representations] could still (if I imagine myself to be an animal) carry on their play in an orderly fashion, as representations connected according to empirical laws of association, and thus even have an influence on my feeling and desire, without my being aware of them (assuming that I am even conscious of each individual representation, but not of their relation to the unity of representation of their object, by means of the synthetic unity of their apperception). This might be so without my knowing the slightest thing thereby, not even what my own condition is. (Br 11:52, 314) The synthetic unity of apperception must be regarded not only as the supreme principle in the whole of human cognition (B135) and the objective condition of all cognition (B138), but also as the only principle whereby an exhaustive account of basic self-consciousness can possibly be developed. The double nature of apperception—regarded as the principle of both human cognition and basic self-consciousness—is summarised by

32  Self-Knowledge and Self-Consciousness Rosenberg (2005), who asserts that, at a particularly high level of generality, every perception or experience has the same form as an encounter between subject and object: To think of something present in intuition as an object of representation is, however schematically, to commit oneself to the in-principle possibility of filling in both its history as an item in nature, i.e., in a lawful system of spatio-temporal items, and one’s own history as an experiencer of this nature here and now in perceptual encounter with it. . . . The form of experience . . . is: an identical self persisting through time in relation to a systematic world of spatio-temporal items that affect it. (Rosenberg 2005, 121–2) In the following paragraphs, the unity of apperception will be examined in order for its role in the transcendental system to be clarified. In particular, the main features of apperception will be articulated in relation to the Transcendental Deduction so as to develop the features Kant attributes to subject self-consciousness and its difference from the object self-consciousness of empirical apperception in the following chapters.

The Transcendental Deduction and the Principle of the Unity of Apperception The first two chapters of the Transcendental Analytic articulate the condition and use of pure concepts: The first is “On the Clue to the Discovery of All Pure Concepts of the Understanding” (A66/B91), retroactively called the Metaphysical Deduction in the second edition (B159), whereas the second contains “The Deduction of the Pure Concepts of the Understanding”, i.e., the Transcendental Deduction proper (A84/B116). The next two paragraphs address these two fundamental topics whereby the unity of apperception is developed for the first time in the critical approach. The Metaphysical Deduction While the aim of the Metaphysical Deduction is to determine the full set of categories, or pure concepts, from the table of logical judgments, the Transcendental Deduction legitimates the categories’ application in order to have cognition and experience. In particular, the term “Deduktion” has the semantic value given by jurists, who distinguish what is lawful (quid juris) from what concerns the fact (quid facti); furthermore, such scholars regard it as a demonstration of the legitimacy of a property claim. Within the transcendental approach, it is necessary to demonstrate the claims that the categories have objective validity; that is, that they are the conditions for the possibility of experience.

Self-Knowledge and Self-Consciousness  33 In order to establish the origin of the categories in general through their complete coincidence with the universal logical forms of judgment, in the Metaphysical Deduction Kant assigns cognition18 a few general features. Firstly, cognition is based on the form of the judgment. In this regard, and more specifically, it is necessary to recall that intuitions, the immediate and singular representations of an object (A19/B33, A320/B377), are subsumed under the higher-order representation of concepts, and that understanding, which can be regarded as a faculty for judging (A69/B94), uses these concepts to produce judgments in compliance with them: All intuitions, as sensible, rest on affections, concepts therefore on functions. By a function, however, I understand the unity of the action of ordering different representations under a common one. Concepts are therefore grounded on the spontaneity of thinking, as sensible intuitions are grounded on the receptivity of impressions. Now the understanding can make no other use of these concepts than that of judging by means of them. Since no representation pertains to the object immediately except intuition alone, a concept is thus never immediately related to an object, but is always related to some other representation of it (whether that be an intuition or itself already a concept). Judgment is therefore the mediate cognition of an object, hence the representation of a representation of it. (A68/B93) Judgments can be articulated in a given variety of logical forms linked to the ‘logical functions’ of quantity, quality, relation, and modality (each judgment being either singular or plural, affirmative or negative, and so forth). Nonetheless, judgments featuring such forms can only refer to objects if the concepts of objects employed are structured according to such forms—i.e., judgments based on a subject-predicate structure can only refer to objects if these are conceived of as substances with accidents. Thus, if the forms of judgments are identified through the ‘logical functions’ of judgments, and if they are articulated according to the general headings of quantity, quality, relation, and modality, then the table of pure concepts of the understanding, or ‘categories’, is derived from the logical functions of judgments and the categories are regarded as the general ways in which the particular concepts of objects are structured in the form of judgments. The issue can be explained in more detail by recalling the issue of synthesis and the fact that transcendental logic arises from general logic and makes the thinking about objects possible. Kant regards synthesis as “the act of putting different representations together, and grasping what is manifold in them in one cognition” (B103). It “gathers the elements for cognition, and unites them to form a certain content” (B103), and, accordingly, “it is the first thing to which we have to

34  Self-Knowledge and Self-Consciousness attend if we wish to judge about the first origin of our cognition”. ‘Pure synthesis’, which relies on the a priori synthetic unity ground, yields the pure concepts of the understanding. In the so-called guiding thread (Leitfaden) passage, Kant holds: The same function that gives unity to the different representations in a judgment also gives unity to the mere synthesis of different representations in an intuition, which, expressed generally, is called the pure concept of understanding. 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 analytical unity, also brings a transcendental content into its representations by means of the synthetic unity of the manifold in intuition in general, on account of which they are called pure concepts of the understanding that pertain to objects a priori; this can never be accomplished by universal logic. (A79/B104–5) The Transcendental Deduction While the Metaphysical Deduction aims to demonstrate that the pure concepts of the understanding are discovered through reflection on the logical structure of judgments, the aim of the Transcendental Deduction—the heart of the transcendental approach—is to show how such concepts necessarily apply to any and all experiences that a subject might have; in other words, its aim is to legitimate the objective validity of concepts and the way in which they can be related to objects. If there are any such things as concepts destined for pure use a priori, that is completely independently of all experience, then these always require a deduction of their entitlement, since proofs from experience are not sufficient for the lawfulness of such a use, and yet one must know how these concepts can be related to objects that they do not derive from any experience. I therefore call the explanation of the way in which concepts can relate to objects a priori their transcendental deduction. (A85/B117) More specifically, the Transcendental Deduction is concerned with the connection between the intellectual and the sensible conditions of human cognition in that it pertains to what Allison (2004, 160) called a “cognitive fit”. For Descartes, the problem is epistemological because it concerns the way epistemology accounts for the problem of scepticism, which is synthesised by the well-known action of a malignant genius who can systematically deceive us and our most evident cognitions. Here, the problem revolves around the correspondence between the subjective realm of cognitions and the external world; with Kant’s theses of transcendental idealism, the

Self-Knowledge and Self-Consciousness  35 problem relates to the correspondence between the contents of sensibility and the a priori rules of thought. The task of the Transcendental Deduction is precisely to tackle the justification of this connection so as to characterise the necessity of the categories as the condition of possibility for experience. As is known, Henrich’s (1969) important work has influenced the debate about the argumentative structure of the Transcendental Deduction. Firstly, he disagrees with Adickes (1889, 139–4) and Paton (1936, 501), who assert that, while that which precedes § 21 is an ‘objective deduction’, § 24 and § 26 contain a ‘subjective deduction’. Secondly, Henrich highlights the twostep nature of the proof of the B-Deduction, namely the division of the argument into two distinct parts—the first presented in §§ 15–21, and the second in §§ 22–7—both establishing the necessity for the categories:19 We can now formulate a criterion for a successful interpretation of the whole text of the deduction in this way: the interpretation must show that, contrary to the initial impression that the two conclusions merely define the same proposition, on the contrary, sections 20 and 26 offer two arguments with significantly different results, and that these together yield a single proof of the transcendental deduction. We shall call this task the problem of the two-steps-in-one-proof. (Henrich 1969, 642) This argument has provoked much debate about the nature of two such steps. For the purposes of this work, I will refer to Allison’s account: While the first part asserts the necessity of the categories with regard to the object of sensible intuition in general in order to have the thought of it, the second states the necessity of the categories with regard to human sensibility and its objects in order to link the categories to the perception of objects:20 [The function of categories] in the first part is to serve as rules for the thought of an object of sensible intuition in general, that is, as discursive rules for judgment. That is why the argument abstracts from the particular nature of human sensibility and refers to sensible intuition in general. It shows that any representation that is brought to the “objective unity of apperception” is also thereby related to an object in a judgment and, as such, necessarily stands under the categories. By contrast, the aim of the second part of the Deduction is to establish the applicability of the categories to whatever is given under the conditions of human sensibility. It attempts to do this by demonstrating (through their connection with the imagination) that the categories also have a non-discursive function as conditions under which whatever is given (in accordance with the forms of sensibility) can enter empirical consciousness. In short, it attempts to link the categories (albeit indirectly) to the perception rather than merely the thought of objects. (Allison 2004, 162)

36  Self-Knowledge and Self-Consciousness Therefore, in the first part, the function of the categories is regarded as a rule whereby the thought of an object of sensible intuition in general is possible: The categories are strictly interpreted as discursive rules employed in the judgments. The argument unfolds in five steps: (1) the manifold given in a sensible intuition necessarily pertains to the synthetic unity of apperception, whereby the unity of intuition is possible, (2) the action of unifying the representations and bringing them under an apperception is based on the understanding—regarded as the faculty of the rules—and is characterised as a logical function of judgment; it is important to highlight that this very act enables the manifold to be brought to the synthetic unity of apperception and to be unified in a judgment, (3) if the manifold is unified in a judgment, then it is necessarily determined in accordance with the logical functions of the judgment through which the manifold is brought under apperception, (4) the categories are regarded as these very functions for judging: The manifold of a given intuition is determined according to them, and (5) hence, “the manifold in a given intuition also necessarily stands under the categories” (B143). The “second step” goes even further than the first step’s claim concerning the mere condition of possibility of the thought of an object: Here, Kant refers to the possibility of cognizing a priori through categories whatever objects may come before our senses, not as far as the form of their intuition but rather as far as the laws of their combination are concerned, thus the possibility of as it were prescribing the law to nature and even making the latter possible. (B159–60) Thus, the second step establishes a link between the categories and what is given through the forms of sensibility in perception and experience. In more detail, Kant connects the categories to time—the form of the inner sense through the transcendental synthesis of the imagination—in order to relate the categories to the content of the inner sense, i.e., perception and empirical intuition. If the transcendental, a priori synthesis of the imagination, regarded as an expression of the spontaneity of thought, determines the form of the inner sense, and if its determination is based on the condition of the synthetic unity of apperception, then the function in question must be subject to the categories.21 What is particularly important for our purposes is the first part of the Transcendental Deduction, in which it is stated that sensible intuitions in general must be subject to the categories if they are to be brought to the

Self-Knowledge and Self-Consciousness  37 unity of consciousness. It is exactly here that Kant specifies this particular role and assigns the unity of apperception the main characterisations thereof. I Think and the Principle of the Necessary Synthetic Unity of Apperception Transcendental apperception is defined in the first edition of the Transcendental Deduction at A107: Now no cognitions can occur in us, no connection and unity among them, without that unity of consciousness that precedes all data of the intuitions, and in relation to which all representation of objects is alone possible. This pure, original, unchanging consciousness I will now name transcendental apperception. (A107) It is in the famous passage (B131–2) in the B-Deduction that Kant introduces I think in the context of the transcendental unity of apperception. For Van Cleve (1999, 79), it is important to underline that it indicates both a property and a principle; namely, a principle attributing the property to certain collections of representations of being apprehended by the act of apperception: The I think must be able to accompany all my representations; for otherwise something would be represented in me that could not be thought at all, which is as much as to say that the representation would either be impossible or else at least would be nothing for me. That representation that can be given prior to all thinking is called intuition. Thus all manifold of intuition has a necessary relation to the I think in the same subject in which this manifold is to be encountered. (B131–2) Soon after, Kant associates I think with spontaneity and the pure or original apperception or self-consciousness: But this representation [I think] is an act of spontaneity, i.e., it cannot be regarded as belonging to sensibility. I call it the pure apperception, in order to distinguish it from the empirical one, or also the original apperception, since it is that self-consciousness which, because it produces the representation I think, which must be able to accompany all others and which in all consciousness is one and the same, cannot be accompanied by any further representation. (B132)

38  Self-Knowledge and Self-Consciousness Later, Kant calls it transcendental unity of self-consciousness to mark it as the condition for the possibility of cognition: I also call its unity the transcendental unity of self-consciousness in order to designate the possibility of a priori cognition from it. For the manifold representations that are given in a certain intuition would not all together be my representations if they did not all together belong to a self-consciousness; i.e., as my representations (even if I am not conscious of them as such) they must yet necessarily be in accord with the condition under which alone they can stand together in a universal selfconsciousness, because otherwise they would not throughout belong to me. From this original combination much may be inferred. (B132–3) In particular, passage B131–2 marks the incipit of the B-Deduction, condensing two crucial points that have been widely discussed by leading commentators. The first point has a general scope and concerns single representations: For a representation to represent and to be anything to the subject, the subject must be able to think of it as its own (cf. Dyck 2017). Obviously, the point here is not that de facto representations are accompanied by the ‘I think’, but concerns the necessity of the possibility of representation of self-ascriptions (Allison 2004, 163). In this regard, Ameriks (1997, 58) refers to a ‘personal quality’ assigned to individual representations in order for them to display the form (E): I think that x, I think that y, I think that z. Similarly, Carl (1997, 153) refers to one’s ability to make judgments from a first-person perspective, holding that “the very first sentence or ‘§ 16’ gives an account of what is involved by the notion ‘my representations’ in terms of the notion of ‘I’ and the ability to make judgments from the first-person point of view”. All these points will be extremely important to consider when the notion of de se thought is examined. The second point concerns the set of representations that account for a complex thought based on synthetic unity: The different representations merge into a single consciousness as a thought ascribable to a thinking subject, i.e., (T) I think that (I think that x, I think that y, I think that z, etc.) (Ameriks 1997, 58). Not only must the uses of ‘I’ be co-referential—the ‘I’ thinking x must be identical to the ‘I’ thinking y, and so on—but the identity of the ‘I think’ must also concern the highest-ranking ‘I think’ (outside of parentheses here) upon which is based the synthesis of various representations into a single, complex thought. Taken together, for Allison (2004, 164), these two points determine the principle of the necessary synthetic unity of apperception, establishing a connection between the possibility of ascribing the representations to a thinking subject and the possibility of a complex thought through the ‘I think’: All my representations in any given intuition must stand under the condition under which alone I can ascribe them to the identical self as my

Self-Knowledge and Self-Consciousness  39 representations, and thus can grasp them together, as synthetically combined in an apperception, through the general expression I think. (B138) As will be seen below, in passages B137 and B138 Kant holds this principle to be analytic. The single constituents of a complex thought must be combined in a synthetic unity to allow their ascription to a single thinking subject. In particular, a single complex thought requires a unified thinking act through which the single representations x, y, z are brought into synthetic unity. Even if the I that thinks x is the same I that thinks y, this does not imply an I thinking a single, complex thought; for the subject to think them together, a single act must be involved, and through this, the thinking subject brings the components of a complex thought into a synthetic unity and ascribes them to its identical self. One is the condition for the other and vice versa: The condition of possibility for the self-ascription of single representation relies on the fact that these can be brought into a synthetic unity, and the condition of possibility for such a synthetic unity is determined by the fact that the representations composing it can be ascribable to a single, thinking subject. If a single complex thought logically involves a single thinking subject, then every single component of this complex thought must be ascribable to an identical thinking subject; in turn, the thinking subject must be aware of its identity in the synthetic unity of such a complex thought. In another famous excerpt, all this can be reformulated by stating that the analytical unity of apperception presupposes a synthetic unity: This thoroughgoing identity of the apperception of a manifold given in intuition contains a synthesis of the representations, and is possible only through the consciousness of this synthesis. For the empirical consciousness that accompanies different representations is by itself dispersed and without relation to the identity of the subject. The latter relation therefore does not yet come about by my accompanying each representation with consciousness, but rather by my adding one representation to the other and being conscious of their synthesis. Therefore it is only because I can combine a manifold of given representations in one consciousness that it is possible for me to represent the identity of the consciousness in these representations itself, i.e., the analytical unity of apperception is only possible under the presupposition of some synthetic one. (B134) Two points are at issue here. The consciousness of the identity of I contains the synthesis of the representations, i.e., the ‘I’ that thinks X = the ‘I’ that thinks Y presupposes the synthetic unity of apperception, namely ‘I think’ (X + Y) (Rosenberg 1986, 511). Nonetheless, the consciousness of the ‘I think’ identity requires not only a synthesis, but also a consciousness of such a synthesis.

40  Self-Knowledge and Self-Consciousness More specifically, the thought of the identity of ‘I’ in the manifold—as Kant states at the beginning of the preceding passage, “the thoroughgoing identity of the apperception of a manifold given in intuition”—implies a synthesis. This means that the combination (conjunctio) of the representations through the action of understanding—referred to as ‘synthesis’ by Kant—represents the condition of possibility for the thought of an identical ‘I think’. However, the condition for the thought of an identical ‘I think’ also requires a consciousness of such a synthesis, as Kant also states in the A-Deduction: This unity of consciousness would be impossible if in the cognition of the manifold the mind could not become conscious of the identity of the function by means of which this manifold is synthetically combined into one cognition. (A108) In other words, the consciousness of the identity of the ‘I’ that thinks x with the ‘I’ that thinks y is nothing but the consciousness of the identity of the action of the ‘I’ when the unified act of thinking produces a combination of representations to bring them into a complex thought, i.e., a synthetic unity. For this reason, in passage A108, Kant also specifies a correlation between the original, necessary consciousness of the identity of oneself and the consciousness of the necessary unity of the synthesis of all appearances in accordance with concepts: Without the latter, the subject cannot think of itself as the same ‘I’: The mind could not possibly think of the identity of itself in the manifoldness of its representations, and indeed think this a priori, if it did not have before its eyes the identity of its action, which subjects all synthesis of apprehension (which is empirical) to a transcendental unity, and first makes possible their connection in accordance with a priori rules. (A108) Within the B-Deduction, all these points should be related to the associations of synthesis, representations, and object on one hand, and of apperception, intellect, and judgment on the other, specifying the categories as the conditions of possibility for knowledge. With reference to the first dimension, Understanding is, generally speaking, the faculty of cognitions [Erkenntnisse]. These consist [bestehen] in the determinate relation of given representations to an object. An object, however, is that in the concept of which the manifold of a given intuition is united. Now, however, all unification of representations requires unity of consciousness in the synthesis of them. Consequently the unity of consciousness is that which

Self-Knowledge and Self-Consciousness  41 alone constitutes the relation of representations to an object, thus their objective validity, and consequently is that which makes them into cognitions and on which even the possibility of the understanding rests. (B137) While the intellect is the faculty of knowledge through concepts, the object is that which is given through the synthesis of the manifold in intuition on a conceptual basis. As seen previously, Kant defines synthesis as the act of putting different representations together and grasping what is manifold in them in one cognition. Moreover, it gathers the elements for cognition and unites them to form a certain content (B103). In this way, the multiple representations—namely, the manifold—are connected with one another to form a single, further representation. The union of the representations via conceptual synthesis could not exist without the unity of apperception, which refers them to a given object so as to establish their objective validity (B137). At this point, it is important to highlight a fundamental reciprocity thesis between the unity of apperception and the objective unity of the representations relating to an object (cf. passages B137 and B138, and as early as A105): The unity of self-consciousness is the condition for the cognition or experience of an object understood as the notion of ‘experience’ described in Transcendental Deduction (§ 22 and § 26) and in B218, where it is identified with an empirical cognition, i.e., a cognition that determines an object through perceptions based on the unity of apperception. Thus, while the synthetic unity of the representations relating to an object—i.e., the representations’ objective validity—is impossible without the unity of apperception, the latter cannot affirm the identity of ‘I’ without synthetic unity. The unity of apperception can be viewed as not only a necessary but also a sufficient condition for the representation of an object: In essence, the objective and the synthetic unities coincide.22 The unification of representations should also be linked to the faculty of judgment—regarded as the faculty for applying rules—and articulated in accordance with the well-known distinction between the objective and the subjective unity of consciousness: The transcendental unity of apperception is that unity through which all of the manifold given in an intuition is united in a concept of the object. It is called objective on that account, and must be distinguished from the subjective unity of consciousness, which is a determination of inner sense, through which that manifold of intuition is empirically given for such a combination. Whether I can become empirically conscious of the manifold as simultaneous or successive depends on the circumstances, or empirical conditions. Hence the empirical unity of consciousness, through association of the representations, itself concerns an appearance, and is entirely contingent. (B139)

42  Self-Knowledge and Self-Consciousness While the objective unity of consciousness presupposes the use of the categories and enables the manifold provided by intuition to be unified in the concept of an object, the subjective unity of consciousness is merely a product of the reproductive imagination. More specifically, the empirical unity of consciousness is non-universal, contingent, and only subjectively valid, since representations are ordered by the empirical laws of association. On the other hand, being universal and necessary, the transcendental unity of apperception can order the representations produced by the synthesis to assign objective validity to them:23 The pure form of intuition in time, on the contrary, merely as intuition in general, which contains a given manifold, stands under the original unity of consciousness, solely by means of the necessary relation of the manifold of intuition to the one I think, thus through the pure synthesis of the understanding, which grounds a priori the empirical synthesis. That unity alone is objectivity valid; the empirical unity of apperception, which we are not assessing here, and which is also derived only from the former, under given conditions in concreto, has merely subjective validity. One person combines the representation of a certain word with one thing, another with something else; and the unity of consciousness in that which is empirical is not, with regard to that which is given, necessarily and universally valid. (B140) In the subsequent paragraph, Kant connects the synthesis to the judgment to show that the categories are employed in the synthesis of experience. The judgment is the vehicle through which each representation is unified within the subject to make synthesis possible: “I find that a judgment is nothing other than the way to bring given cognitions to the objective unity of apperception” (B141). Ultimately, it is judgment that exhibits the universality and necessity found at the base of objective validity (B142). As far as this point is concerned, Kant rejects the logicians’ definition of judgment as the representation of a relationship between two concepts: In his view, this does not specify what such a relationship entails. It is the copula that characterises the relationship of judgment as that which links representations to the objective unity of apperception so as to convert them into objective knowledge: The act of judgment, seen as the spontaneity of the subject, is the very condition for synthesis. In this regard, the connection between the representational manifold and the unity of apperception enacted by the intellect is given by the logical function of judgments determining the manifold in empirical intuition. The categories are functions of judgment whereby the manifold is determined: As long as it is linked to the synthetic unity within one single consciousness, the manifold is subject to the categories. If features are to be ascribed to an object as such through judgment, the scheme ‘x is F’—and the unity that follows—must be related to more than

Self-Knowledge and Self-Consciousness  43 a simple ‘I think that x is F’. The original synthetic unity of apperception stands for the necessary possibility of an all-inclusive ‘I think’, i.e., the highest level of ‘I think’, which is used to link all ‘first-order’ acts of thought based on the scheme ‘I think that: I think that x is F, I think that y is G, I think that z is H, and so forth’ (Ameriks 2004, 57).

Notes 1. In Russell’s epistemology (1912), acquaintance is central to the metaphysically direct relationship between awareness and sense data: According to his perspective, acquaintance—regarded as a non-causal relationship of metaphysical immediacy—is the epistemic foundation of not only self-knowledge but also all knowledge. In brief, Russell’s notion of self-knowledge is based on the subject’s introspective knowledge, interpreted as a metaphysically direct relationship of acquaintance with her mental object. For this reason, it is strongly justified. Contemporary acquaintance accounts are based on Russell’s thesis that the subject enjoys metaphysically direct access to her mental states: Although this is held to support strongly justified non-inferential judgments about these states, one of the major questions in the current debate is regarding the types of mental states as objects of acquaintance (sensations such as pain being the best candidates). Cf. BonJour (2003), Chalmers (2003), Feldman (2004), Fumerton (1995), Gertler (2001), Levine (2007), and Pitt (2004). 2. Contemporary versions of the inner-sense theory have been developed by Armstrong (1968), Lycan (1996), Goldman (2006), and Nichols and Stich (2003). According to these versions, introspection is a “self-scanning” process with mental states as inputs and introspective states or self-attributing beliefs based on a reliable causal process as outputs. From these perspectives, introspection can also explain the notion of consciousness or awareness. For instance, following Armstrong and Lycan’s HOP (high-order perception) theory, a state is conscious only if the subject is aware of that state through inner sense: “As I would put it, consciousness is the functioning of internal attention mechanisms directed at lower-order psychological states and events. I would also add a soupçon of teleology: attention mechanisms are devices that have the job of relaying and/ or coordinating information about ongoing psychological events and processes” (Lycan 1996, 14). Other theories of consciousness are called HOT (higher-order thought) and contend that what makes a state conscious is the fact that a subject has a higher-order thought directed onto the state (Rosenthal 1993; Carruthers 2005). The HOT theories construe awareness as conceptualised and belief-like, whereas HOP theories construe it according to the perceptual model. 3. Given Kant’s historical opposition to rationalist metaphysics (cf. infra, Chapter 2), it is worth pointing out that the term “rationalist” is employed here to refer to the main feature of self-knowledge singled out by the contemporary rationalist theory of self-knowledge, namely its contribution to rational, critical thought (for instance, cf. the views of Burge and Moran). In this way, according to the Rationalists, the normative dimension, which regards what the subject ought to do or how she ought to think, represents what distinguishes the selfknowledge from other-knowledge. 4. Castañeda (1966, 1967) employs an asterisk, or star, next to a pronoun (“he*”) to attribute first-person reference to someone else, as in “Alessandro believes that he* is tall”. This sentence is not true unless Alessandro expresses his belief in the first person: “I am tall”. Matthews (1992) introduces the “I*” for sentences with first-person subjects in order to analyse the phenomena expressed by “I think that I* am F”.

44  Self-Knowledge and Self-Consciousness 5. There is a close relationship between the linguistic and the mental dimensions, between the semantic features that will be explained later with regard to the first-person pronoun (linguistic self-reference, the way in which “I” refers) and the concept ‘I’ (the mental self-reference) expressed by the linguistic device (cf. Bermúdez 2017). Given the topic of this book, my interest is related to the concept of ‘I’ and the relative I-thoughts; when no other specification is present, the ‘I’ is to be considered in its conceptual nature. 6. Following Kaplan (1989) and Perry’s (1997, 594) classic approaches, “a defining feature of indexicals is that the meanings of these words fix the designation of specific utterances of them in terms of facts about these specific utterances. The facts that the meaning of a particular indexical deems relevant are the contextual facts for particular uses of it”. Kaplan and Perry detected several difficulties in the Fregean approach to indexical terms, in particular with the notion of sense as articulated in two components. Kaplan (1989) distinguishes between character, namely the linguistic meaning fixed by linguistic convention, and the content, which is determined according to the context associated with at least one agent, time, location, and possible world. On the other hand, Kaplan distinguishes between pure indexicals (for example, “I”, “now”, and “here”) and impure indexicals (also called true demonstratives or deictic terms; e.g., “this”, “that”). In the latter case, the determination of reference is based not only on the associated rule but also on the mediation of the subject’s intention. 7. This approach to indexicality has been acknowledged by major scholars, cf. Shoemaker (1968, 91), Peacocke (1983, 133–9), Rovane (1987, 147), Campbell (1994, 73), and Kaplan (1989, 493). Nonetheless, their positions are not entirely uniform—cf. Castañeda (1983, 1987, 1989, 1990), Kapitan (2001, 2008), and de Gaynesford (2006) for an analysis of the debate. 8. As Wright (1998, 19) points out, “the ground has to be such that in the event that the statement in question is somehow defeated, it cannot survive as a ground for the corresponding existential generalization”. Returning to the example examined previously, “there is pain” is tantamount to “I have pain”; if the second judgment is defeated—for instance, if it is not true that “I have pain”—the kind of access to the information through the subject’s first-person mental dimension on which the judgment has been made does not allow us to maintain the first judgment, “there is pain”, i.e., the corresponding existential generalisation: If I am not the one who has the pain, then there cannot be anybody else with the pain. Instead, in the judgments displaying uses of I as object, the error through misidentification is always possible. In such cases, the corresponding existential generalisation survives: If Elena is mistaken regarding who is the owner of the hand she is seeing in the mirror, and she realises she is not the owner of the hand, the corresponding existential generalisation “someone has the dirty hand” survives. 9. If a subject makes the judgment I feel hungry in normal circumstances, and on the rational basis of an experience of hunger, the following hold: (a) The judgment is true if the thinker of the thought feels hungry (by the fundamental reference rule for the first person), and (b) the concept feels hungry is such that an experience of hunger makes the self-ascription of the concept by a subject having such an experience reasonable. Peacocke points out that “In being a concept whose instantiation makes reasonable a self-ascription by the subject, the concept feels hungry is one of several kinds that are in a more general sense anchored in the subject. The displayed explanation of the rationality of the judgement does not at any point appeal to the rationality of a judgement on the part of the thinker I = m, for some mode of presentation m distinct from the first person concept” (2014, 111). The explanation of IEM concerns the fact that

Self-Knowledge and Self-Consciousness  45 there is good reason for making the judgement I feel hungry in this way; in other terms, there is no vulnerability to misidentification when the judgement is made in such a way that the explanation of the correctness of the judgement made in this way does not involve any empirical identities. 10. In the first Critique, Kant defines the faculty of understanding, in other words the faculty for theoretical knowledge, to be spontaneous as it brings forth representations from itself (KrV A51/B75)—thus, spontaneity is connected to acts of synthesis (B129–30) or to functions (A68/B93). In the second Critique, transcendental freedom is considered “absolute spontaneity” and is based on the moral law. In this way, acts of theoretical knowledge (from a Vermögen der Erkenntnisse) and practical knowledge (from Wille) are considered to be spontaneous acts. There is a debate about the kind of relationship between theoretical and practical capacity—between theoretical freedom (the freedom of the understanding) and practical freedom (the freedom of the will)—in order to explain the nature of spontaneity (see Ellis 2017). A few scholars consider spontaneity to be a kind of freedom: McDowell (1994, 5) considered spontaneity to be a conceptual activity that originates from the “realm of freedom”; Allison’s interpretation is that understanding is “absolutely spontaneous”. Thus, the spontaneity of the understanding is a notion of freedom in judgment, since the spontaneity of the cognitive faculty consists of the ability to produce the complex representations that constitute the content of judgments. 11. Moving from the mathematical–philosophical debate initiated by Hintikka (1967, 1969, 1972) and Parsons (1969, 1984, 2012), there are certain conditions that must be met for a representation to be an intuition or concept: Kant contrasts the ‘immediacy’ (Unmittelbarkeit) and ‘singularity’ (Einzelheit) (cf. A19/B33, A68/B93; JL 9:91) of the intuition with the mediacy and ‘generality’ (Allgemeinheit) of the conceptual representation (A68/B93; Log 9:91); cf. infra, Chapter 4. 12. A further, crucial notion is ‘experience’ (Erfahrung), which Kant links with ‘empirical cognition’ (B166, A176/B218, A189/B234). In this sense, Kant identifies ‘perception’ (Wahrnehmung), regarded as the conscious apprehension of the content of an intuition (Prol 4:300; A99, A119–20, B162, B202–3) (since the perception is objective, however, it is also the consciousness of the content of whichever representation. More specifically, an objective perception is a cognition, which is either an intuition or a concept—A320/B376). It follows that Kant regards ‘experience’ as a set of interconnected perceptions by means of fundamental concepts, namely the ‘categories’: “Experience is cognition through connected perceptions [durchverknüpfte Wahrnehmungen]” (B161; cf. B218; Prol 4:300). 13. Hohenegger (2013) points out that Kant particularly supports Baumgarten (1739), who establishes the distinction between “outer” (externus) and “inner” (internus) sense. The sense can represent the inner state of the soul, or the outer state of the body, so the sensation is internal, or through the internus sensus (the consciousness in the strict sense), or external, through the externus sensus. 14. In this regard, Locke (1690, II, 1, § IV) employs an additional, ‘internal’ sense: “The other Fountain, from which Experience furnisheth the Understanding with Ideas, is the Perception of the Operations of our own Minds within us, as it is employ’d about the Ideas it has got; which Operations, when the soul comes to reflect on, and consider, do furnish the Understanding with another set of Ideas, which could not be had from things without: and such are Perception, Thinking, Doubting, Believing, Reasoning, Knowing, Willing, and all the different actings of our own Minds; which we being conscious of, and observing in our selves, do from these receive into our Understanding, as distinct Ideas as

46  Self-Knowledge and Self-Consciousness we do from bodies affecting our Senses. This source . . . though it be not Sense, as having nothing to do with external Objects; yet it is very like it, and might properly enough be call’d internal Sense”. 15. Cf. Allison (2004), Valaris (2008), and Schmitz (2013). It is important to elucidate three features stemming from the metaphysical and transcendental expositions of the concept of time (Allison 2004, 276). Firstly, “Time is not something that would subsist for itself or attach to things as an objective determination, and thus remain if one abstracted from all subjective conditions of the intuition of them” (A32/B49). The second feature pertains to the previously mentioned roles of the inner sense, through which “Time is nothing other than the form of inner sense, i.e., of the intuition of our self and our inner state” (A33/B49–50). The third feature draws attention to the fact that time is the a priori formal condition for all appearances in general, whereas space is nothing but the pure form of all outer intuitions (A34/B50). Allison argues that if there is a restriction of time to the inner sense when Kant remarks that “time cannot be a determination of outer appearances; it belongs neither to a shape or a position, etc., but on the contrary determines the relations of representations in our inner state” (A33/ B49–50), then this claim seems to contradict his other general thesis whereby time is the a priori formal condition of all appearances. The author solves this problem by focusing on the passage in which Kant states that, although time is the immediate condition of the inner intuition and “of our souls”, it is “also the mediate condition of outer appearances” (A34/B51). Accordingly, if all representations or modifications of the mind are given in the inner sense, and if time is the form of the inner sense, then Kant can conclude that “all appearances in general, i.e., all objects of the senses, are in time, and necessarily stand in relations of time” (A34/B51). 16. In the first book of his Treatise of Human Nature, Hume (1739–40, 364) dismisses the idea that the subject can perceive a persistent self through the introspective consciousness: “For my part, when I enter most intimately into what I call myself, I always stumble on some particular perception or other, of heat or cold, light or shade, love or hatred, pain or pleasure. I never can catch myself at any time without a perception, and never can observe any thing but the perception”. We may reach a metaphysical thesis of exclusion: The thinking subject is not a substantial object in the world. This thesis is obtained as a conclusion of the previous elusive epistemic thesis: Given that, in introspection, the thinking subject cannot manifest itself to itself; nor can it be known as an object (epistemic argument); thus, the subject is not a substantial object (metaphysical thesis). 17. In a passage in Metaphysik Mrongovius (V-Met/Mron 29: 882, 550–1) Kant states: “Inner sense is the consciousness of our representations themselves. (Apperception is the ground of inner sense.) It has the soul as its object. If the soul is conscious of itself to itself, without being conscious of its state, that is apperception. If it is also conscious of its state, then it is sensation or perception”. 18. On this subject, cf. Guyer (2010), Reich (1992), and Brandt (1995). 19. The A-Deduction has a different structure and presents a few problems, as shown by Guyer (2010). Firstly, the A-Deduction does not affirm that which is stated in the Metaphysical Deduction concerning the “clue” that cognition necessarily employs the form of the judgment: The necessity for the categories is established on the basis of the transcendental unity of apperception. Secondly, the A-Deduction does not present a single argument in two steps, but rather a single line of arguments in two forms: An argument “from above” and an argument “from below” (cf. Pereboom 2006). While the latter demonstrates

Self-Knowledge and Self-Consciousness  47 that synthesis and its correlated categories are necessary to explain how objects are represented, the former focuses on apperception. Representations must be interconnected through a concept in order to produce the representation of an object; this “carries something of necessity” because it is determined a priori, neither at leisure nor arbitrarily (A104–5). Such a necessity is attained through the transcendental apperception, “that unity of consciousness that precedes all data of the intuitions, and in relation to which all representations of objects is alone possible” (A106–7). In attributing the unity of apperception to the faculty of understanding, Kant maintains that apperception requires the categories: It involves the application of the categories, which are regarded as rules for understanding. In this way, the categories are consequently regarded as the conditions for thinking in a possible experience, just as space and time contain the conditions for intuition. Thus, they are the fundamental concepts for thinking objects in general, and have a priori objective validity: “The necessity of these categories rests on the relation that the entire sensibility, and with it also all possible appearances, have to the original apperception, in which everything is necessarily in agreement with the conditions of the thoroughgoing unity of selfconsciousness, i.e. must stand under universal functions of synthesis, namely of the synthesis in accordance with concepts, as that in which alone apperception can demonstrate a priori its thoroughgoing and necessary identity” (A111–2). Some passages from the A-Deduction will also be employed in connection with the similar argument contained in the above-mentioned B-Deduction § 16. 20. On one hand, Henrich (1969) contends that the first step reaches this result in order to justify the connection between the contents of sensibility and the a priori rules of thought for a range of sensible intuitions that already possess unity; on the other hand, the second step eliminates this restriction since it is based on the principle shown in the Transcendental Aesthetic, whereby space and time are unities and all intuitions are spatio-temporal, thus removing the possibility of non-conformity for all human sensible intuition. To examine this difficult issue in depth, cf. Wagner (1980) and Evans (1990). On the other, Guyer criticises Allison’s approach. By focusing on the interpretation of § 21 and § 22, Guyer (2010, 143) emphasises how Kant moves from the objective validity of the categories—the necessary condition for the transcendental unity of apperception, regardless of the specificity of sensibility—to the objective validity of the categories regarded as the necessary condition for apperception relating to the specifically spatio-temporal, empirical intuition. It follows that there is a shift from an abstract to a concrete account, despite the fact that that ‘second step’ involves an independent argument pertaining to the application of the categories to more specific, sensible, spatio-temporal objects. 21. This second part and the experiential function of the categories pose a few interpretative problems; on the role and success of the arguments in the Deduction, cf. Prauss (1971), Keller (2001), and Allison (2004). 22. For this interpretation, cf. Allison (1983, 144ff; 2004, 173–5), Howell (1992, 227), and Aquila (1989, 159); for a recent defence, see Schulting (2012b). For a presentation and analysis of the issue, cf. Carl (1998, 197ff.) and Wunderlich (2017). Pereboom (2006) attacks Allison and Howell’s interpretations of Kant’s assertion as a statement of the sufficiency claim in B137: “The unity of consciousness is that which alone constitutes the relation of representations to an object, thus their objective validity”. Although it is true that, in Kant, the cognitions of objects consist of the relationship between representation and objects produced by a synthesis that involves the unity of consciousness, on the other hand, this does not mean that the synthesis involving the unity of consciousness unfailingly results in such a relationship (cf. also Pereboom 2001). Guyer (1992,

48  Self-Knowledge and Self-Consciousness 151) is also highly critical of the issue; in his view, the thesis that transcendental apperception is not only necessary but also sufficient for the representation of objects is an “excessive assumption”. Furthermore, Guyer (1987, 118) stresses that Kant fails to equate the transcendental unity of apperception with the knowledge of objects by fiat, “instead of demonstrating a synthetic connection between them (in either direction). That is, Kant does not develop an argument holding that self-consciousness as such has special a priori conditions also applying to objects, whatever the source of the latter” (for another important adverse criticism of this issue, cf. Carl 1989a and 1989b). Fortunately, for my purposes, I can refrain from taking a stance concerning this issue. 23. A representation is objectively valid if it is a representation of an objective feature of reality in which existence and nature are represented independently from the way they are perceived (cf. Guyer 1987, 11–24).

2 Two Senses of ‘I think’

Three key issues have emerged from the discussion of ‘I think’: (1) a semantic problem connected to the type of reference of the representation ‘I’, (2) an epistemic problem regarding the type of knowledge relative to the thinking subject produced by the representation ‘I think’, and (3) a strictly metaphysical problem1 associated with the features assigned to the thinking subject’s nature. In this chapter, ‘I think’ and the so-termed transcendental subject will be analysed in an attempt to establish a connection to these theoretical points and, in particular, to certain specific metaphysical characterisations of the thinking subject introduced by Kant in the critical period: Thinking as spontaneity is the being itself. Other semantic and epistemic issues will also be examined in the next two chapters. Inasmuch as it accompanies every single representation, ‘I think’ refers to the thinking subject, no matter its metaphysical nature. At the level of the metaphysics of the self, Kant speaks of the ‘I’ (das Ich), of the mind (das Gemüt), of the thinking subject (das denkende Subjekt), and of the soul (die Seele). According to this perspective, several interpretations may be singled out within the debate, among which an anti-metaphysical account is deserving of particular attention (cf. Marshall 2010): Noticeably, although some of its arguments rest on Kant’s claims concerning the limits of cognition and objective significance, others are based on his rejection of rational psychology—at heart, they all aim at ruling out the possibility of a doctrine concerning the nature of the self as such. As will be seen later, even if no Erkenntnis (in Kantian technical terms) of the self is possible because no relationship between a given representation and a spatial-temporal object can be regarded as a self, Kant maintains a number of metaphysical claims regarding the nature of the thinking subject. On the other hand, the empty form of the referential apparatus in transcendental apperception has been appraised in intrinsically different ways. At first glance, Kant presents a metaphysical thesis of exclusion whereby the ‘I’ of the ‘I think’ as an intellectual representation produces no knowledge about the nature of the thinking subject or self: rather, it refers to something that, in no respect, can be said to be an object—the transcendental subject

50  Two Senses of ‘I think’ (Cassam 1997). This approach seems to anticipate Wittgenstein’s Tractatus’ claims of the subject as something not belonging to, but limiting, the world (Bennett 1974). An elusive interpretation of the ‘I’ in the ‘I think’ as having no reference has been argued more than once by the so-termed Noownership Reading (cf. Strawson 1966, 166; Chisholm 1976, 42; Rosenberg 1986, 68; Ameriks 2000, 54). The affinity between Wittgenstein’s Blue Book’s argument concerning the ‘I as subject’ and Kant’s ‘I think’, as alleged by a few commentators,2 lies precisely within this framework: A few details of this approach will be discussed in Chapter 3. Two further antithetical views on the pivotal notions of ‘unity of apperception’ and ‘identity of self-consciousness’ are central to this framework: The Substantial Ownership Reading, discussed in Henrich and Guyer’s influential works, and the Formal Ownership Reading, specifically upheld by Allison (cf. Bermúdez 1994). According to the Substantial Ownership Reading, there is no way to determine what the identity of self-consciousness consists of across different representations unless it is regarded as something involving a single conscious self; in fact, this remains the same despite its different representations (cf. Henrich 1976).3 Several Kantian passages point out the difficulty of interpreting the knowledge of a persisting, enduring self in these substantial or Cartesian terms, as if some identity about the subject could be disclosed. According to the Paralogism section—the Third Paralogism in particular—it is fallacious to infer the persistence of a numerically identical self in the self-ascription of experiences. The formal condition under which the subject can ascribe the experience to itself through the mere logical identity of the representation ‘I’ of the ‘I think’ is one thing, but eliciting knowledge of a subject’s substantial identity from these formal constraints is quite another, as we will see later: “Kant’s point seems to be that constancy in the sense of the expression ‘I’ can be no guarantee of constancy in its reference” (Bermúdez 1994, 222). The fact that the unity of apperception produces the a priori knowledge of something that is persisting and numerically identical certainly does not require the rational psychologists’ claim that this a priori knowledge is the knowledge of a simple substance.4 While the Substantial Ownership Reading maintains that, in the absence of something identical or unified in different states, neither identity nor unity of consciousness is possible, according to the Formal Ownership Reading the identity of self-consciousness cannot be seen as if there existed an identical or unified self. Moving from the Paralogisms chapter of the Dialectic, Allison (2004, 340) highlights Kant’s disclosing the illusion of regarding the unity of apperception as the unity of a thing. Even if the representation ‘I think’ produces no knowledge of a particular self whatsoever, the ‘I’ can still play a representational role. For instance, Horstmann (1993) identifies two theses that bring the two accounts to mind in more than one way: In the A-edition Paralogism, the representation ‘I’ refers to what is termed “substrate”, while the ‘I’ refers to an object even though this is not knowable. In the B-edition Paralogism, by contrast, ‘I think’ should simply be regarded as

Two Senses of ‘I think’  51 an act of spontaneity: No entity is picked out by the ‘I’ of apperception, as the ‘I’ solely represents the spontaneous activity of thought. Similarly, Melnick (2009, 4) maintains that the self should be considered as an activity simply because this avoids understanding it as an entity of any sort: “the simple thinking subject or the ‘I think’ is literally an action, not an entity”; in particular, for Horstmann (2007) as well as for Melnick (2009), the self should be regarded as an act or activity in a Fichtian fashion, marking a sharp break with the previous metaphysical views. Even though there is a fundamental connection between a mental activity of representational combination and the nature of the self, and even though the ‘I think’ is considered an act of spontaneity (cf. infra), as Marshall (2010, 11) reminds us, in several passages Kant “consistently ascribes activity to the self, without suggestion of identification”. According to Rosefeldt (2000, 128), if the ‘I think’ only concerns the logical identity of the I in the Deduction passages, then any assumption of a particular metaphysical identity should be excluded. The distinction at issue is precisely between logical and real features of the self, or between features of the self as a real entity and features of the self as a logical entity, i.e., the so-called logical I (Rosefeldt 2003, 142). The self is not a “real subject of inherence” (A350), but the “permanent logical subject of thinking”: In other words, the “logical identity of the I” has to be considered in opposition to the numerical identity of a real thinking being (A363). For this reason, a “logically simple subject” does not entail a simple substance (B407–8). The aim of this chapter is to highlight certain characterisations of ‘I think’ according to the Formal Ownership Reading: Two distinct meanings of ‘I think’ will be identified in the following two paragraphs. In its first meaning, mainly found in the Transcendental Deduction, ‘I think’ is the act of apperception; in the second, found in Transcendental Deduction and in the section on Paralogisms in particular—I think (in italics in these pages) is supposed to have a representational nature. The notion of ‘transcendental subject’ will be interpreted in formal terms as a specific concept that, mutatis mutandis, has the same function as the concept of the ‘transcendental object’. Kant maintains a sharp distinction between the metaphysical and the epistemic dimensions, and yet he introduces two meanings of ‘I think’ and the concept of ‘transcendental subject’ to create a bridge between the two. In the last three paragraphs (§§ 3–5), I will attempt to show that (1) in line with the Formal Ownership Reading, the identity or unity of selfconsciousness does not presuppose the identity of a real subject, but a formal identity based on the representation I think, (2) Kant establishes general, metaphysical characterisations concerning the nature of the thinking subject: The thinking as spontaneity is the being itself, and (3) to some extent, such characterisations are captured by the concept of the ‘transcendental subject’: The thinking being represents itself as a

52  Two Senses of ‘I think’ transcendental subject = x through the simple representation I. Accordingly, the simple representation I will be explained from both the semantic and the epistemic perspectives.

The Synthetic Unity of Apperception and the I Think As seen in Chapter 1, the principle of the necessary synthetic unity of apperception establishes a connection between the possibility of ascribing the representations to a thinking subject and the possibility of a complex thought based on the ‘I think’; consequently, it is to be regarded as the supreme principle in the whole of human cognition (B135), as well as the objective condition of all cognition (B138). In the first place, the principle seems to equate the transcendental apperception with ‘I think’ as asserted in B137, in which Kant explicitly states that ‘I think’ is the act of apperception itself: The supreme principle of all intuition in relation to the understanding is that all the manifold of intuition stand under conditions of the original synthetic unity of apperception. All the manifold representations of intuition stand under the first principle insofar as they are given to us, and under the second insofar as they must be capable of being combined in one consciousness; for without that nothing could be thought or cognized through them, since the given representations would not have in common the act of apperception, I think, and thereby would not be grasped together in a self-consciousness. (B136–7) As seen above, in B138 Kant also uses ‘I think’ as an expression of the act of apperception: All representations in any given intuition must stand under the condition of apperception, through which the subject can ascribe such representations as its own to its identical self and combine them synthetically through the general expression ‘I think’. The same point is expressed in B140, where Kant distinguishes the necessary, universally valid ‘objective unity of consciousness’ from the ‘subjective unity of apperception’, a determination of the inner sense through which the manifold of intuition is given empirically: The pure form of intuition in time, on the contrary, merely as intuition in general, which contains a given manifold, stands under the original unity of consciousness, solely by means of the necessary relation of the manifold of intuition to the one I think, thus through the pure synthesis of the understanding, which grounds a priori the empirical synthesis. As mentioned above, the principle of the unity of apperception expresses an analytic proposition even though this is based on the introduction of the

Two Senses of ‘I think’  53 synthetic unity of the manifold in the intuition (B138): The statement “the synthetic unity of consciousness is an objective condition of all cognition” expresses an analytical principle asserting that all representations must stand under such a principle so that the thinking subject can ascribe them to its identical self as its own. Only in this way can the subject grasp such representations as a synthetically combined whole in apperception, and this act of apperception of conjunction is precisely expressed by the universal expression ‘I think’. The point is that I think (in italics in these pages) is also seen as a concept or representation, although of a special kind. Moreover, it is precisely at this stage in the Transcendental Deduction that Kant recognises the double meaning of ‘I think’—as an act and as a representation—which seems to trigger paradoxical consequences in a famous passage. In B132, Kant states that the representation I think is an act of spontaneity because—to distinguish it from the empirical apperception—it is not concerned with sensibility but with pure or original apperception; it corresponds to the self-consciousness that produces the representation I think, which accompanies all representations even though it cannot be accompanied by any further representation in turn. Howell (2000, 141) remarks that the ensuing circularity—the representation I think is the act of apperception producing the representation I think—is “a hasty way of stressing the spontaneity and underivedness of the I think”: The capacity or faculty for apperception is original (A117) and produces “the underived and not-further-explicable I think”. In other words, based on the transcendental apperception understood as the act of conjunction of the representations in synthetic unity, the thinking subject is characterised as a self-conscious subject through the first-person representational component I think. Since it is equated to the transcendental apperception, ‘I think’ is regarded as an act of conjunction of the representations establishing the supreme principle of all cognition. On the other hand, at the very heart of the Transcendental Deduction, I think is described as a representation that accompanies and makes any other representation possible by representing the identical self of the thinking subject in the synthetic unit. This last point—the I as a representation of the identical self—is restated at the very heart of the Transcendental Deduction, namely in B135, much more straightforwardly than it is in B132. After the passage in which the analytic nature of the principle of the necessary unity of apperception is stated, Kant asserts that, “without [a synthesis of the manifold given in an intuition] [the] thoroughgoing identity of self-consciousness could not be thought”. In the following paragraph, this constant identity is related to the representation I. Kant goes even further to assert that I is a representation, and a simple one, too: “through the I, as a simple representation, nothing manifold is given; it can only be given in the intuition, which is distinct from it, and thought through combination in a consciousness” (B135). Notwithstanding Kant’s efforts to keep the two dimensions apart, these passages are paradigmatic examples of the close connection between the

54  Two Senses of ‘I think’ metaphysical and the representational planes, as well as of the fact that two different meanings of ‘I think’ are taken into consideration here: As will be discussed later, they are two facets of a single principle. The suggestion outlined previously should be discussed and argued in light of the ontological question raised by Kant while building the equation between thinking and being. Before that, however, a more detailed discussion of I think in its representational sense is needed.

Transcendental Dialectic: The I Think and the Analysis of the Paralogisms The Analysis of the Paralogisms It is in the Transcendental Dialectic, specifically in the section on the Paralogisms of Pure Reason, that Kant focuses more explicitly on I think as a concept or representation and assigns certain features to it. In this section, Kant denies that we can have any knowledge of the ontological nature of the thinking subject by dismissing the possibility of the pure or empirical apperception to be the object of knowledge regardless of the conditions of possible experience. No psychological idea can be associated with a knowable content: In Kantian terms, transcendental ideas are necessarily produced by reason; hence, unlike the categories of the intellect, they have no objective reference in experience. With regard to the first transcendental idea—the psychological idea proper—by introducing the idea of an ultimate, unconditional subject, the reason attempts to find a suitable object in “I myself, considered merely as thinking nature (soul)” (A682/B710). At the beginning of both editions of the chapter on Paralogisms (B399), Kant defines the I think as a concept: Even if it has no place in the general list of transcendental concepts, and despite the fact that it has no special title, I think is a transcendental concept in its own right—it is the vehicle for all concepts, including the transcendental ones, and serves the purpose of introducing the whole of thinking to the sphere of consciousness. Kant attacks all metaphysical approaches that attempt to find a ‘rational doctrine of the soul’ (rationale Seelenlehre) based on the inferences that employ the concept I independently of experience. By means of such inferences, and without the conditions for the applicability of the categories to the intuitions, the rationale Seelenlehre claims to know the nature of the subject as a soul or as a thinking substance by the a priori ascription of the categories, considered as transcendental predicates, to the ‘I’, or soul, taken as noumenal object. In particular, due to a systematic connection (Zusammenhang), Kant outlines the dialectical propositions to reject Rational Psychology’s claims and all related topics produced following the Leitfaden of categories. As a result, a thinking entity is assigned these intrinsic features: (1) the soul is substance, (2) in its quality, simple, (3) in the different times in which it exists, numerically identical, i.e., unity (not plurality), and (4) in relation

Two Senses of ‘I think’  55 to possible objects in space (A345/B403). The categorical titles do not follow the standard method in that they begin with the category of relation and proceed to quality, quantity, and mode—Kant’s chief aim is to analyse Rational Psychology’s propositions without the intrusion of experience or the intervention of a perceived existence. In the A-version, Kant highlights the paralogistic nature of the four propositions of topics: In his view, the application of these characteristics can only result in delusive knowledge, which may eventually take the form of a fallacious syllogism—a paralogism proper. With regard to the B-version, Capozzi (2007) points out two distinct arguments: In the first argument, Kant examines the soul based on pure apperception and discloses that (1) analytically, the topics of the rational doctrine of the soul are not attainable, and (2) synthetically, the first proposition of the topics referring to the substantiality of the thinking being as such, from which the others stem, is not attainable. With regard to the second argument, Capozzi highlights how, in fact, Kant proceeds from a different assumption, i.e., a rationale Seelenlehre constructed via an analytical method, the object of which is the perception of an existence—the Cartesian cogito ergo sum—is groundless. In this case, Kant refers to the empirical apperception, namely the inner sense, and applies his analysis to ‘I think’: Since this is assumed to be real, it is perceived as something existing in the stream of consciousness. The two arguments are based on the distinction (that will be discussed later) between the proposition ‘I think’ from a Cartesian perspective, which entails the perception of an existence, and the proposition ‘I think’ assumed problematically, that is, in “its mere possibility, in order to see which properties might flow from so simple a proposition as this for its subject (whether or not such a thing might now exist)” (A347; cf. B406). This is associated with a further distinction between transcendental and empirical apperception, establishing “how the I that I think is to differ from the I that intuits itself . . . and yet be identical with the latter as the same subject”, and therefore I as intelligence and thinking subject cognize my self as an object that is thought, insofar as I am also given to myself in intuition, only, like other phenomena, not as I am for the understanding but rather as I appear to myself. (B155) I as a Concept As mentioned above, the B-version’s first argument consists of two parts: An analytic one, developed in B407–9, and a synthetic one, to be found in B410–413. Both sections show that the propositions for topics of rationale Seelenlehre are not valid: No knowledge of the thinking subject’s metaphysical nature is attainable.

56  Two Senses of ‘I think’ The analytic approach dwells on the analysis of the consciousness of self and thought in general, which cannot lead to a determination of the metaphysical thinking subject as object. Concerning this specific point, Kant is forthright: All modi of self-consciousness in thinking are therefore not yet themselves concepts of the understanding of objects (categories), but mere functions, which provide thought with no object at all, and hence also do not present my self as an object to be cognized. (B407) In fact, it is not the determining self but rather that which pertains to the determinable self, namely the inner intuition, that can be an object to cognise. At the same time, in the four categorical titles, Kant introduces a clearcut distinction between a metaphysical and a representational dimension, between the ‘thinking subject’, or ‘thinking I’ considered as a determining subject, and the logical subject, or ‘I’, in apperception. The thinking or determining subject is what determines the relationship constituting every judgment, as Kant clearly expounds in the first title on substantiality: “In every judgment I am always the determining subject of that relation that constitutes the judgment” (B407). With regard to the arguments in the Transcendental Deduction, the subject that determines the relationship constructing the judgment is clearly determined by virtue of the features assigned to the unity of transcendental apperception. In the following passages taken from B407, Kant shifts the focus from the metaphysical plane of the determining subject to a representational level, i.e., the representation of ‘I’: The subject that thinks is one thing, but what is represented in thought as a subject is another. This step is of the utmost importance when introducing the wellknown epistemic closure: It is not possible to assert that the thinking self is a substance through the representation of a subject. The same argumentative structure recurs in three more titles. For example, under the title of simplicity, the I of apperception is a representation designating a logically simple subject; it follows that I features in every single thought. If the concept of ‘thought’ is analysed, the representation I is implicit; hence, an analytic proposition is produced. Similarly, the proposition that the I of apperception is identical in each manifold and that its existence as a thinking being is distinct from external things are both analytic. Nonetheless, the features assigned to the representation I allow no epistemic conclusion; in particular, the representation of a subject as an essentially simple, identical substance detached from matter does not entail that the thinking self is a simple, identical substance detached from the body. These are synthetic propositions requiring not only the involvement of the categories but also an intuition. In the transcendental system, however, intuition can only be sensible: It plays no part in the intellect nor

Two Senses of ‘I think’  57 in the field of thought; thus, the simple representation I—referred to as the poorest representation of all—lies solely in that field (B408). The same conclusion can be drawn when it comes to the synthetic moment. The synthetic method is based on syllogism, regarded as a synthetic, rational, and necessary procedure used as early as 1781 to criticise Rational Psychology’s attempts to obtain the four syllogisms of the topics. Although the introduction to the four paralogisms remains unaltered, in the B-version, Kant illustrates the paralogistic character of the syllogism on the substantial nature of the thinking subject:5 What cannot be thought otherwise than as subject does not exist otherwise than as subject, and is therefore substance. Now a thinking being, considered merely as such, cannot be thought otherwise than as subject. Therefore it also exists only as such a thing, i.e., as substance. (B410–1) Kant explains why the syllogism6 is a paralogism by unmasking a sophisma figurae dictionis based on the ambiguous nature of the middle term: The predicate of the minor premise does not coincide with the subject of the major premise. Even in the synthetic moment, Kant analyses the representational nature of I, beginning with the different meanings indicated by the middle term in the two premises. In the major premise, the thought or concept’s reference is a being that can be thought of in every respect and that, as such, can be given via intuition. For the minor premise, instead, the concept’s reference is the self-consciousness; here, if what is represented is the relation to oneself as subject (as the form of thinking), then, in the concept of ‘Thinking’, “the I always serves as subject of consciousness” (B412). The only possible conclusion that can be drawn from this argument is not the paralogistic interpretation of the metaphysical nature of the thinking entity as something existing as a substantial subject,7 but the recognition of a tautological proposition that simply specifies the concept of the thought in analytic terms. For this reason, no property can be assigned to the manner of existence of the thinking subject: “In thinking my existence I can use myself only as the subject of judgment, which is an identical proposition, that discloses absolutely nothing about the manner of my existence” (B412). All this mirrors Kant’s ‘Copernican turn’: In the transcendental approach, the first-order discussion of objects is replaced by a second-order discussion of the concept of an object and its conditions of representation. What I am referring to is the thinking ‘I’; thus, the first-order discussion of the determining subject, or apperception, is replaced by the second-order discussion of the concept of thinking or the self-consciousness through which the subject represents its relation with itself as subject. In Anthropology (Anth 7: 135, 23) and Fortschritte (KGS 20: 270), Kant describes the I of apperception as being linked strongly to the logical I.

58  Two Senses of ‘I think’ This connection is described in Opus Postumum, where the first act of the faculty of representation is held to be the consciousness of oneself, which is a merely logical act, as we will see in the next paragraph. The I of apperception is the logical subject, which specifies the simple representational nature of the I in I think. The logical I is strictly connected to the role of the ‘absolute subject’ of judgments. As is known, according to Kant, “the specific nature of our understanding consists in thinking everything discursively, that is by concepts, and so by mere predicates, to which therefore the absolute subject must always be missing” (Prol 4: 333, 86). A standard categorical judgment is made of the combination of two concepts (the subject-concept and the predicate-concept of the judgment); it follows that the ‘subject of a judgment’ is a thing represented by the concept that is consequently used as the subject-concept in this judgment. The point is that an absolute subject of judgments cannot be represented by means of general concepts: An absolute subject of judgments is a specific thing represented by a concept that can only be used as a subject-concept, and not as a predicate of a judgment. However, it is well known that, in Kant’s view, all concepts must offer the possibility of being employed as predicates of possible judgments (A69/B94). The principle that there are no concepts that can only be used as subject-concepts in the judgments but never as the determination of another thing is on par with the claim that there are no singular concepts: No concept has a singular representative function determining a particular, complete concept of a Leibnizian individuum to be used as the absolute subject of a possible judgment. This issue will be discussed in Chapter 4; at this point, we can only state that, since all concepts are general concepts, to make a judgment about a certain particular object it is necessary to make a ‘singular use’ of a general concept. As the representation of a thinking being, the I is regarded by Kant as an absolute subject of judgments, a representation that cannot be employed as the predicate of any other thing because the thinking subject cannot apply the I to any being other than itself: Now it does appear as if we have something substantial in the consciousness of ourselves (the thinking subject) . . . for all the predicates of inner sense are referred to the I as subject, and this I cannot again be thought as the predicate of some other subject. It therefore appears that in this case completeness in referring the given concepts to a subject as predicates is not a mere idea, but that the object, namely the absolute subject itself, is given in experience. (Prol 4: 334, 86) The fact that “I, as a thinking being, am the absolute subject of all my possible judgments, and this representation of myself cannot be employed as predicate of any other thing” (A348) should lead to the conclusion that the

Two Senses of ‘I think’  59 representation I is a singular concept; however, the above-mentioned simplicity of I established by Kant is precisely antipodal to the definition of a ‘singular concept’, with its rich content that is only applicable to one object (the particular representational nature of I will be addressed in Chapter 4). Nevertheless, the concept I is used as an absolute subject of judgment, since the thinking subject can only apply it to one possible object, i.e., itself. Consequently, in the judgments “I saw a dog yesterday” and “I see a cat today”, the thinking subject ascribes the properties of having seen a dog yesterday and of seeing a cat today; thus, it ascribes these two properties to one and the same object, i.e., itself. In Metaphysics L1, Kant restates that I is a nonpredicable representation: I am a substance. The I means the subject, so far as it is no predicate of another thing. What is no predicate of another thing is a substance. The I is the general subject of all predicates, of all thinking, of all actions, of all possible judgments that we can pass of ourselves as a thinking being. I can only say: I am, I think, I act. Thus it is not at all feasible that the I would be a predicate of something else. I cannot be a predicate of another being; predicates do belong to me; but I cannot predicate the I of another, I cannot say: another being is the I. (LM 28: 266, 79) At the same time, the logical feature of the concept I introduces the distinction between the logical and real features of the self, or between features of the self as a real entity and features of the self as a logical entity.8 Following Rosefeldt (2003, 147), “a logical subject is everything of which we can predicate something, i.e., everything the concept of which we can use as the subject-concept in a judgement”; instead, a real subject is a substance, something in which properties inhere. This distinction emphasises the fact that not everything of which we can predicate something in a judgement is a thing in which properties inhere. This seems like a rather sensible thing to say, for we can even make judgements about properties, such as “Red is my favourite colour”, although properties like the colour red are by definition not substances. (Rosefeldt 2003, 146) As mentioned previously, at the very heart of the Transcendental Deduction, Kant states that the I is a representation of the identical self; however, this identity presents nothing but a logical nature, as “through the I, as a simple representation, nothing manifold is given” (B135). Apart from the simple or empty representation I, the subject has no acquaintance with itself, since this representation I cannot be identified with a spatio-temporal object. It is precisely for this reason that Kant states that the thinking subject can only know that its self is the “permanent logical subject of thinking”,

60  Two Senses of ‘I think’ and not that it is a “real subject of inherence” (A350). Rosefeldt explains the point very clearly: A ‘permanent logical subject, of thinking’ [is] something that is represented in mere thought as something that cannot be a predicate of something else because it is represented by a representation that can only occur in the subject positions of judgments. Hence, the predicate ‘logical’ would not signify a certain way in which the subject exists but rather the way in which it is represented, namely merely by conceptual means and not by intuition. (Rosefeldt 2017, 231) If the ‘logical identity of the I’ is in opposition to the numerical identity of a real thinking being (A363), there are no criteria for the identity of noncorporeal objects. For this reason, the thing to which the thinking subject ascribes its thoughts—i.e., the subject itself—can only be given through a representational–logical identity. In brief, the thinking subject always represents itself as an ‘absolute logical subject of judgments’: It is absolute because the subject can apply the concept I to itself and not to any other being, and logical because the identity in question can only be representational, as the ‘I’ cannot be identified with a real subject of inherence. The ‘I Exist Thinking’: The Empirical Apperception In Anthropology, Kant (Anth 7: 135, 23) introduces the ‘I as subject’ and the ‘I as object’, based on the distinction between transcendental and empirical apperception: If we consciously represent two acts: inner activity (spontaneity), by means of which a concept (a thought) becomes possible, or reflection; and receptiveness (receptivity), by means of which a perception (perception), i.e., empirical intuition, becomes possible, or apprehension; then, consciousness of one’s self (apperception) can be divided into that of reflection and that of apprehension. The first is a consciousness of the understanding, pure apperception; the second a consciousness of the inner sense, empirical apperception. In this case, the former is falsely named inner sense.—In psychology we investigate ourselves according to our ideas of the inner sense; in logic, according to what intellectual consciousness suggest. Now here the ‘I’ appears to us as to be double (which would be contradictory): 1) the ‘I’ as subject of thinking (in logic), which means pure apperception (the merely reflecting ‘I’), and of which there is nothing more to say except that it is a very simple idea; 2) The ‘I’ as object of the perception, therefore of the inner sense, which contains a manifold of determinations that make an inner experience possible.

Two Senses of ‘I think’  61 The distinction enables the subject to (re)present itself in two ways: Through the ‘I’ that thinks and through the ‘I’ that intuits itself. In a different passage, Kant (V-Met-L1/Pölitz 28: 224, 44–5) states that This I can be taken in a twofold sense: I as human being, and I as intelligence. I, as a human being, am an object of inner and outer sense. I as intelligence am an object of inner sense only. Obviously, this does not imply two Is: “I as a thinking being am one and the same subject with myself as a sensing being” (Anth 7: 142, 33). Unlike the proposition ‘I think’ being regarded as the a priori equivalent of ‘I am thinking’, from the empirical angle the proposition ‘I think’ is equivalent to ‘I exist thinking’; there is no longer a logical function, but only the determination of the subject at the level of existence—the Cartesian cogito ergo sum. The object of intuition must involve the inner sense—the ‘I’ as object of perception is revealed by the empirical apperception as a phenomenon unfolding through the form of time. Consequently, the nature of the justification of the existence is empirical, i.e., it is based on the perception of thinking expressed by the proposition ‘I think’ (cf. Anderson 2015; Longuenesse 2017). In the following passage, Kant attributes a syllogistic interpretation of the cogito to Descartes—‘I exist’ is derived from ‘I think’ through a syllogistic inference, the major premise of which is ‘Everything that thinks, exists’—and he argues with such a view: The ‘I think’ is, as has already been said, an empirical proposition, and contains within itself the proposition ‘I exist’. But I cannot say ‘everything that thinks, exists’; for then the property of thinking would make all beings possessing it necessary beings. Hence my existence also cannot be regarded as inferred from the proposition ‘I think’, as Descartes held (for otherwise the major premise, ‘everything that thinks, exists’ would have to precede it) but rather it is identical with it. (B422 n.) In fact, there are several problems with a syllogistic interpretation of the Cartesian cogito, and it should not be taken for granted that Descartes (1641) assessed the cogito argument in this way.9 Whatever the case may be, what matters at this point is the type of connection between ‘I think’ and ‘I exist’: For Kant, the subject knows itself to exist by perceiving itself while it is thinking; thus, the perception becomes the justification or the knowledge of existence. In the next paragraph, this ontological question will be specified from the pure apperception point of view. The type of existence in question is not the category of existence, since the use of this category is in connection with the object of experience. As Longuenesse points out, in order to use the category of existence,

62  Two Senses of ‘I think’ we need either to have a perception of it or to represent its existence as connected, according to known empirical laws, to the existence of objects we do perceive. But this is not what happens with ‘I’ in ‘I think’. The perception of an act of thinking just is the perception of that act as being my own, and thus as being the ground of my use of ‘I’. So this is not a case of the application of the category of existence, but rather an immediate, pre-categorial perception of existence that is a component in any individual perception of thinking. (Longuenesse 2017, 90) Nevertheless, from the perspective of receptivity, the consciousness of the self as the object of perception appears variable at first glance, as “it can provide no standing or abiding self in this stream of inner appearances” (A107) (cf. supra, Chapter 1). In primis, with regard to the empirical apperception, Kant also rejects the possibility of moving from the inner perception of something existing as thinking (referred to as eine unbestimmte empirische Anschauung, “an indeterminate empirical intuition”, B423 n.) to the determination of this particular something as an existing substance in time and space, the forms of inner and outer sense through which all phenomena are given; otherwise, this would not be thought, but matter. The consciousness of the self as contemplated by the empirical apperception is the inner perception of something that is not object of the outer sense. In other words, from an empirical point of view, the representation I designates an object of the inner sense, and as such, it should not be regarded as a concept for the determination of the subject: For the I is not a concept at all, but only a designation of the object of inner sense insofar as we do not further cognize it through any predicate; hence although it cannot itself be the predicate of any other thing, just as little can it be a determinate concept of an absolute subject, but as in all the other cases it can only be the referring of inner appearances to their unknown subject. (Prol 4: 334, 86) For this reason, from the empirical perspective, ‘I think’ “expresses an indeterminate empirical intuition, i.e., a perception” (B423 n.); this is an empirical intuition because the empirical–existential proposition ‘I think/I exist’ relies on a sensation pertaining to sensibility, which is only revealed in time. Moreover, it is indeterminate due to the absence of space, the form in and through which objects manifest themselves and are determined.10

Thinking Is Being Thus far, two distinct meanings of ‘I think’ have been distinguished: The first, mainly found in the Transcendental Deduction, is ‘I think’ as the act

Two Senses of ‘I think’  63 of apperception; the second, found in Transcendental Deduction and particularly in the section on Paralogisms, considers I think (in italics in these pages) in its representational nature. For Ameriks (2004, 61), the latter can be regarded as the thought of the I as an epistemic subject, the general representation of the I as the subject of apperception standing for something present in the general thought of the transcendental apperception. The former, on the other hand, is the thought of the I as an existing subject: “Wherever there really is an epistemic subject, there is a confrontation with one’s own ‘being itself’, an I that at least has ‘existence’ ”. In Ameriks’ view, it is no mystery that the same being can be both kinds of subjects, and yet these characterisations are very different. In what follows, Kant spells out a precise ontological issue: I as an existing subject is the thinking, so the thinking is the being itself. In the next section, it will be argued that the thinking as being can be thought through the concept of the transcendental subject. As seen in KrV, Kant asserts that pure apperception is the original consciousness. As a fundamental, transcendental principle, in Opus Postumum Kant (OP 22: 77, 186) states that “the first act of the faculty of representation is the consciousness of myself which is a merely logical act underlying all further representation, through which the subject makes itself into an object”. If the faculty of representation proceeds from apperceptio, understood as a merely logical act or act of the thought, then self-consciousness is the first act of knowledge: The first act of knowledge, rather, is: I am an object of thought (cogitabile) and intuition (dabile) for myself, initially as pure (not empirical) representation, which knowledge is called a priori. This act contains as the formal element of this unity a principle of the connection of the manifold of these representations, independent of all perception. (OP 22: 79, 187) Pure apperception can be expressed by sum. According to Capozzi (2007), an ontological question arises at this point: Sum is nothing but activity, which has nothing receptive about it. It will not mingle with any element in the sensible dimension; it is a thinking activity to the extent that sum and cogito are on par, as ‘I am thinking’ is a tautology in the first act of knowledge: “In the proposition: I am thinking, because it is completely identical, no progress, no synthetic judgment is given to me; for it is tautological and the alleged inference: I think, therefore I am, is no inference” (OP 22: 79, 187). The ontological question is specified when Kant asserts that the subject bound to the first act of knowledge—that is, to former apperception—is both the first subject and the first Wesen being thought. With the first act of knowledge, expressed by the verb ‘I am’, the subject is the being itself: The first act of knowledge is the verb: I am,—self-consciousness, for I, [as] subject, am an object to myself. In this, however, there lies a

64  Two Senses of ‘I think’ relation which precedes all determination of the subject, namely, the relation of intuition to the concept, in which the I is taken doubly (that is, in a double meaning) insofar as I posit myself: that is, on the one hand, as thing in itself (ens per se), and, secondly, as object of intuition; to be precise, either objectively as appearance, or as constituting myself a priori into a thing (that is, as thing [Sache] in itself). (OP 22: 413, 179–80) It follows that the ‘I’ as subject—the self-consciousness—is the being itself. B429 highlights this point: “In the consciousness of myself in mere thinking I am the being itself”. Sum/cogito refers to a subject/being displaying no property because it is something in general (A355); what Kant is asserting here is that the subject has an intellectual consciousness of itself, as the existence of the being itself (Das Wesen selbst) is regarded as something (ein Etwas) unknown. This self-consciousness is specified in representational terms, “as intellectual consciousness of my existence, in the representation I am, which accompanies all my judgments and actions of my understanding” (BXL). In B277 Kant states: “The representation I am, which expresses the consciousness that can accompany all thinking, is that which immediately includes the existence of a subject in itself”. From this metaphysical perspective emerge a few characterisations of the thinking subject that may explain the absence of an epistemic identification in the representational synthesis. On one hand, I think/I am is the I as an epistemic subject, the formal condition of all thinking: “The I think must be able to accompany all representations” (B132), and “the representation I am accompanies all my judgments and actions of my understanding” (BXL). However, the self-consciousness is the consciousness of a subject/being regarded as something in general—something unidentifiable from an epistemic perspective. I as an existing subject is a form of intellectual awareness of self-existence summarised as those I am or I think representations that accompany every other representation; as such, they display no property whatsoever. In fact, due to the absence of intuition, it is not possible to determine if this something exists as a persistent substance in order to produce knowledge. This basic kind of consciousness discloses one’s own self and existence—what has been referred to as a kind of special self-familiarity—in a way that cannot be reduced in the terms of a description.11 Ameriks (2004, 59) refers to this as the fundamental thesis: “In the consciousness of myself in mere thinking I am the being itself, about which, however, nothing yet is thereby given to me for thinking” (B429). What is being assumed according to the representation I is merely an existence devoid of any property. The subject is able to know that it exists as a thinking activity: “I exist as an intelligence that is merely conscious of its faculty for combination” (B159). However, it is not able to know what it is: “I cannot determine my existence as that of a self-active being; rather

Two Senses of ‘I think’  65 I merely represent the spontaneity of my thought. . . . Yet this spontaneity is the reason I call myself an intelligence” (B158).12 The subject’s being is inaccessible from an epistemic view: What is given is nothing but thoughts that are regarded as its predicates, which cannot enable us to grasp the thinking subject’s nature. If the subject that is bound to the first act of knowledge is the first object to be thought, and if the relative representation is a representation of an indeterminate aliquid, then nothing more than a logical identity can be presented via the I taken as the representational subject that accompanies all thoughts.

The Concept of a ‘Transcendental Subject’ In light of these metaphysical considerations, how many selves are there? This issue also touches on the question of the nature of transcendental idealism and the relevant distinction between appearances and things in themselves (cf. infra). As is known, there is much debate about this topic (cf. Schulting 2011b for a first overview): Even though there is a continuum of interpretive views, Kantian scholars generally account for this distinction in two ways—the first being known as ‘one-world’ or ‘two-aspect’ interpretations (appearances and things in themselves are the same things), and the second being known as ‘two-world’ or ‘two-object’ interpretations (appearances and things in themselves are not, in any important sense, the same things). As Marshall (2013) argues, “understanding Kant’s metaphysics of the self is more directly relevant to interpreting his transcendental idealism than many of his readers have assumed”, above all in order to corroborate the ‘one-world’ or ‘two-aspect’ interpretation. For instance, with reference to Allison’s two-aspect view (1983, 1996), the distinction between appearances and things in themselves has an epistemic rather than an ontological import—it should be regarded as a difference between two types of concepts of an object rather than as two kinds of objects. While the concept of the object as appearance implies a reference to the necessary conditions for the cognition of an object, the concept of the object as a thing in itself includes no such reference—it is abstracted from the cognitive standpoint. Not only are both phenomenal and noumenal aspects ascribed to the self or subject but, in several passages (cf. infra), Kant suggests that, if the subject appears as to be double, then this does not imply an ontological commitment to two Is: “I as a thinking being am one and the same subject with myself as a sensing being” (Anth 7: 142, 33). Consequently, Kant does not seem to introduce any different self or entity but only different types of self-representation. Ameriks (2004) distinguishes two different theoretical dimensions. The first concerns the just-mentioned distinction between the I as an existing subject and the I as an epistemic subject, which is internal to the transcendental apperception apparatus. The second dimension regards the manifold characterisations attributed to the different “selves”, identified at any given

66  Two Senses of ‘I think’ time according to the transcendental and empirical distinctive features of apperception and to the two subsequent ways the subject can (re)present itself (cf. infra).13 In this regard, Van Cleve (1999) pointed out at least three ways: (1) the empirical self or subject, namely the self as we encounter it in introspection based on empirical apperception, (2) the transcendental self or subject, namely the thinker of our thoughts, the haver of our experiences, and (3) “finally, the noumenal self is the ‘self in itself’—the real self or the self as it really is” (1999, 182). However, as Van Cleve (1999, 182) believes the three selves to be too many, he reduces their number by establishing an equation between the transcendental and the noumenal self: “The transcendental self must exist in its own right, I mean that it must exist independent of being represented”.14 One cannot but assume the existence of the noumenal subject (1999, 184): “If thinkers or representers were not noumenal beings, we would have the absurdity of something that exists only as the content of representations, yet is itself the subject of representations”. Nevertheless, this does not imply that the transcendental subject can be equated with the noumenal subject; for instance, according to Schulting (2011a), the transcendental subject, regarded as the logical characteristics of the thinking subject, is not—or, at least, need not be—equivalent to the noumenal self. For Rosenberg (1986, 514), the transcendental subject is a ‘formal’ notion related to modes of representation. In my view, the notion of ‘transcendental subject’ is just as ‘formal’ as that of ‘transcendental object’ in terms of its representational nature as a transcendental pointer, as will be seen shortly. The thinking subject is endowed with an intellectual consciousness of itself and of the existence of the being itself (Das Wesen selbst); this is regarded as an indeterminate aliquid (ein Etwas), an object that cannot be acquired as cognition in the same way that phenomena can. Nor can an object be known as a noumenon; otherwise, the thinking subject might intuit himself through apperception. For these reasons, Kant surmises that, “in the synthetic original unity of apperception, I am conscious of myself not as I appear to myself, nor as I am in myself, but only that I am. This representation is a thinking, not an intuiting” (B157). “The consciousness of myself in the representation I is no intuition at all, but a merely intellectual representation of the self-activity of a thinking subject” (B278). To think of this something as an indeterminate object (which is not known in the same manner as phenomena), Kant introduces the concept of the ‘transcendental subject’ that must be related to the representation I to play the same role as the concept of the ‘transcendental object’. As will be seen, despite this functional equivalence, the two notions display remarkably different characterisations—I think is a representational unicum. According to the well-known distinction between noumenon and phenomenon,15 the object has an existence in itself and is presented as a phenomenon through intuition. Provided that it is not possible to know an object in itself nor to assign it certain properties beyond the representational order, for an

Two Senses of ‘I think’  67 object to be thought in itself, one must employ the indeterminate thought that something in general (= x) exists in itself and appears via the intuition (A104). The concept of an object in general is nothing but the concept of a ‘transcendental object’. Intuitions are sensible, singular representations that refer directly to objects, whereas appearances—the outcomes of this relationship—are representations to be kept distinct from what is referred to as the ‘transcendental object’ (= x). Considered in terms of its function, the concept of a general or transcendental object (= x) is the indeterminate thought of a single object with an existence in itself, i.e., the condition of having the possibility to think about the singular object—spatio-temporally determined by the intuition as Erscheinung—and to provide it with objective reality through the subsequent unification of empirical concepts and the relative attribution of the properties presented by the intuition. On the other hand, when Kant discusses “Thinking, taken in itself”, he states that ‘Thinking’ is the logical function as well as sheer spontaneity, although no subject of consciousness is presented. The subject thinks of itself in the same way that it thinks of an object in general: “In this way I represent myself to myself neither as I am nor as I appear to myself, but rather I think myself only as I do every object in general from whose kind of intuition I abstract” (B429). The concept of transcendental subject (= x) is the indeterminate thought of a thinking being having an existence in itself. The famous passage A346/ B404 introduces the concept of ‘transcendental subject’: When he dismisses the possibility of a science of pure reason concerning the nature of the thinking being, Kant argues that: We can place nothing but the simple and in content for itself wholly empty representation I, of which one cannot even say that it is a concept, but a mere consciousness that accompanies every concept. Through this I, or He, or It (the thing), which thinks, nothing further is represented than a transcendental subject of thoughts = x which is recognized only through the thoughts that are its predicates, and about which, in abstraction, we can never have even the least concept. (A346/B404) The thinking being represents itself as a transcendental subject = x through the completely empty representation I. Nonetheless, the concept of a transcendental subject is different from the concept of a transcendental object in a few respects. The transcendental object is introduced to address the problem of objectivity, that is, the issue of the ‘immanentization’ of cognition (Allison 2004, 60). Since it is impossible to stand outside of the representations and compare them to transcendentally real entities, the concept of a transcendental object should be regarded as a sort of transcendental pointer: “It serves to define the philosophical task by indicating that the commonsensical and transcendentally realistic concern with the ‘real’

68  Two Senses of ‘I think’ nature of objects must be replaced by a critical analysis of the conditions of the representation of an object”. The concept of the transcendental subject can also be seen as a kind of transcendental pointer; however, since no intuition is given here, the concept of a transcendental subject does not enable the determination of a spatiotemporally singular object through the unification of the conceptual dimension: Even though the I can be employed as a pointer to conceive of the difference between the representation and that which is represented, it cannot determine any experiential object. Nothing else is given, nor can it be thought of, except the completely empty representation I. The point is also made in A350, where Kant critically remarks that the first syllogism of transcendental psychology “imposes on us an only allegedly new insight” as it mistakes the constant logical subject of thinking for “the cognition of a real subject of inherence, with which we do not and cannot have the least acquaintance”. According to Kant, “consciousness is the one single thing that makes all representations into thoughts, and in which, therefore, as in the transcendental subject, our perceptions must be encountered”. Shortly after, Kant states that “apart from this logical significance of the I, we have no acquaintance with the subject in itself that grounds this I as a substratum, just as it grounds all thoughts” (italics added). The concept of a transcendental subject is tantamount to the representation I, or the logical I: The I is the simple, entirely empty representation; the transcendental subject is what is represented, i.e., an x. In B399, Kant states that I think is a transcendental, hence simple, concept: It is the vehicle for all transcendental concepts and “serves to introduce all thinking as belonging to consciousness”. Through this representation, the subject can think of itself as a thinking being and refer to itself based on a ‘transcendental designation’. In the notion of transcendental designation, Kant anticipates some of the features of self-reference without identification, as will be seen in the next chapter.

I think qua Thinking and I think qua Representation Passage B429 can be employed to summarise the topics developed in these pages. These might be grouped according to five ‘steps’, two metaphysical (the first and the fourth) and three epistemic (the others): (1)  [Thinking, taken in itself, is merely the logical function and hence the sheer spontaneity of combining the manifold of a merely possible intuition]; (2) [and in no way does it present the subject of consciousness as appearance, merely because it takes no account at all of the kind of intuition, whether it is sensible or intellectual. In this way I represent myself to myself neither as I am nor as I appear to myself], (5) [but rather I think myself only as I do every object in general from whose kind of intuition I abstract]. (3) [If here I represent myself as

Two Senses of ‘I think’  69 subject of a thought or even as ground of thinking, then these ways of representing do not signify the categories of substance or cause, for these categories are those functions of thinking (of judging) applied to our sensible intuition, which would obviously be demanded if I wanted to cognize myself]. (5) [But now I want to become conscious of myself only as thinking]; (2) [I put to one side how my proper self is given in intuition, and then it could be a mere appearance that I think, but not insofar as I think]; (4) [in the consciousness of myself in mere thinking I am the being itself, about which, however, nothing yet is thereby given to me for thinking]. (1) ‘I think’ (qua Thinking) is spontaneity: “Thinking, taken in itself, is merely the logical function and hence the sheer spontaneity of combining the manifold of a merely possible intuition”. Understanding is spontaneous (A51/B75) and “can be represented as a faculty for judging” (A69/B94); it follows that the work of the judgment can be regarded as the sheer spontaneity of the understanding in action (Allison 2004, 36). If to think is to unify the manifold conceptually, then every thought expressed by a judgment must be based on the principle of transcendental apperception; as a result, apperception is the consciousness of the act of thinking or the consciousness of spontaneity. Accordingly, Kant identifies ‘I think’ with apperception taken in itself as an act of spontaneity (B132); thus, ‘I think’ (qua Thinking) is regarded as an act of connection of the representations, the supreme principle of all cognition—it is the bearer or producer of all judgments, the ground of all mental self-ascriptions, and the act of judging one’s given representations. This entails that (2) ‘I think’ (qua Thinking) is not determined representationally. As spontaneity, ‘I think’ (qua Thinking) represents no thinking subject: “In no way does it present the subject of consciousness as appearance, merely because it takes no account at all of the kind of intuition, whether it is sensible or intellectual. In this way I represent myself to myself neither as I am nor as I appear to myself”. “I put to one side how my proper self is given in intuition, and then it could be a mere appearance that I think, but not insofar as I think”. A fortiori, this entails that (3) ‘I think’ (qua Thinking) is not the object of knowledge. “If here I represent myself as subject of a thought or even as ground of thinking, then these ways of representing do not signify the categories of substance or cause, for these categories are those functions of thinking (of judging) applied to

70  Two Senses of ‘I think’ our sensible intuition, which would obviously be demanded if I wanted to cognize myself”. (4) ‘I think’ (qua Thinking) is the being itself. “In the consciousness of myself in mere thinking I am the being itself, about which, however, nothing yet is thereby given to me for thinking”. This entails that (5) I think (qua Representation) is how the subject represents or thinks of itself to itself. If the ‘I think’ (qua Thinking) is the being itself and nothing is thereby given to the subject for thinking, then the subject can only think itself as an object in general: “I think myself only as I do every object in general from whose kind of intuition I abstract”. This entails that, to become conscious of oneself only as thinking or to think of oneself as the subject of a thought, or even as the ground of thinking, the subject uses the concept of the transcendental subject, i.e., the indeterminate thought of a thinking being with an existence in itself. However, apart from this indeterminate thought, which has its representational vehicle in the simple or empty representation I, the subject has no acquaintance with itself. For this reason, the consciousness of oneself as thinking is only expressed by the I; that is, by the intellectual representation of the spontaneity of a thinking subject. (B278) In short, if the I is the representational correlate of the apperception or the representation of a logical subject, then it should be understood as the representation of something in general, as self-consciousness is an intellectual awareness of self-existence, unidentifiable from an epistemic angle. What is represented is a transcendental subject of thoughts = x, only recognised through the thoughts serving as its predicates. The fact that the transcendental subject of thoughts = x is only recognised through the thoughts that are its predicates simply amounts to saying that the representation I is contained analytically in every thought (A350); to put it in different words, the analytical unity of apperception is only possible under the presupposition of some synthetic unity (B133–4). To think is to unify the manifold conceptually; every thought expressed by a judgment must be based on the principle of transcendental apperception. For this reason, Kant holds that the I is inherent in the very concept of thought (B132), determining the form of every judgment in general terms (B406). While Kant states that the mere representation I in relation to all others is the transcendental consciousness in A117, in the A-edition Paralogism, Kant addresses the relationship between all thoughts and the ‘I’ taken as the common subject in which they inhere to affirm that the representation I features in all thoughts (A350). The I of apperception seen as a ‘logically simple subject’ is analytically implied in the concept of ‘thinking’ (B407–8);

Two Senses of ‘I think’  71 if the I is the subject of thinking (KGS 7: 134), and if it is not represented except in the form of the judgment established by the synthetic unity of apperception, then the I expresses the analytical unity of apperception, which is only possible by assuming some synthetic unity. In conclusion, the ‘I think’ (qua Thinking) is the synthetic unity of apperception determining the I think (qua Representation) as an analytical unity of apperception, i.e., the I think (qua Representation) is the representation through which the spontaneity of ‘I think’ (qua Thinking) is given to one’s self. In the absence of epistemic mediations of identification, regarded as simple representations, the I think (qua Representation) merely designates the activity of thinking transcendentally; that is, as the nexus established in the judgment by the copula linking the representational synthesis to the synthetic unity of apperception on a conceptual basis. The possibility of thinking of oneself as subject is properly and solely given by the simple and empty representation I.

Notes 1. I will use the term “metaphysics” in its contemporary philosophical sense; that is, a metaphysics of the self that accounts for the particular types of entities known as ‘selves’. While Kant asserts that “this name (metaphysics) can also be given to all of pure philosophy including this critique” (A841/B869), in the Transcendental Dialectic, he uses the term “metaphysical” to describe and label any dogmatic enterprise aimed at ascertaining the reality of God, freedom, and immortality. 2. Cf. Becker (1984), Sturma (1985), Powell (1990), and McDowell (1994). Kitcher (1990, 1998) has suggested a different No-ownership Reading: The transcendental deduction is a reaction to Hume’s challenge to the notion of ‘personal identity’; i.e., it is a more complex version of Hume’s perspective: The mind is not something distinct from the representations that it has. In fact, the self should be regarded as an interconnected system of mental states: “the ‘unity of apperception’ refers to the fact that cognitive states are connected to each other through syntheses required for cognition. ‘Apperception’ does not indicate any awareness of a separate thing, a ‘self’, or even that different cognitive states belong to a separate thing, a ‘self’. Rather, they belong to the unity of apperception in being connected by syntheses to each other” (Kitcher 1990, 105). Similarly, Brook (1994, 44) identifies the self with the notion of global representation: Kant “did not just think of the mind as having a system of representations; he also thought of it as being a representation, namely, the global representation within which many of the usual denizens of a system of representations are all contained”. The dualism in the thesis of exclusion, which is between a person and a transcendental (metaphysical or philosophical) subject expelling any subjective reference from the world, is remarkably different from the No-ownership Reading, which does not contemplate any subjective reference at all. It is worth noting that these two interpretative views are sometimes considered to constitute an anti-metaphysical approach (Marshall 2010). 3. Henrich (1989, 278) points out the formal implications of an ‘I think-instance’: Every self-conscious thought involves an implicit reference to any other selfconscious thought the same thinking subject can think of; in other words, “the subject is not just the subject of actual self-consciousness. Rather it is only the subject of consciousness in so far as it is actually thought, in the thought being

72  Two Senses of ‘I think’ thought on a particular occasion, as the subject of indefinitely many other ‘I think’-thoughts” (Henrich 1989, 270). The numerical identity of the self consists of the fact that the same subjective dimension is the reference for all these ‘I think’ instances. Henrich contends that all relevant instances of the ‘I think’ consciousness involve the consciousness of the same subject, so much so that all actual and potential self-conscious thoughts constitute a whole in which “the same reference from them to the system as a whole, and to every other individual ‘I-think’ instance in it, is thought as possible” (1989, 273). In this way, his insistence is that the Transcendental Deduction is concerned with a Cartesian certainty about the thinker (1976, 58–9, 69–70). Moving from A116—“we are conscious a priori of the complete identity of the self in respect of all representations which can ever belong to our knowledge”—Guyer (1987) points out the main features of the Transcendental Deduction’s argument: The unity of apperception is the source of a priori knowledge concerning the existence of a persisting, enduring self (as a result, the Transcendental Deduction is a doomed project). In particular, the transcendental argument unfolds as follows: The subject has a priori knowledge of the fact that its experiences concern a single, persisting self; considering that this presupposes knowing that experiences are connected independently of any particular empirical content, the subject is sure that its experiences are the experiences of a single self. If experiences can only be connected based on the assumption of the acts of synthesis, in order to have experiences, the subject knows a priori that what it performs are nothing but acts of synthesis; since such acts cannot depend upon any particular empirical content, it follows that they cannot be anything but a priori. In this scenario, the a priori knowledge of the unity of apperception is only possible by assuming a priori synthesis. Guyer articulates his substantial interpretation to demonstrate that the Transcendental Deduction is a doomed project: As a transcendental argument, it should involve a condition or feature of experience that a sceptic would regard as either absolutely not contentious or as sufficiently acceptable. Nevertheless, if the unity of apperception is seen as a source of a priori knowledge of the existence of a persisting, enduring self, this condition can be hardly accepted. 4. Similarly, for Marshall (2010), that “there must be something that’s identical doesn’t itself presuppose any of the views Kant attacks in the Paralogisms chapter, namely, that we know that thing is a substance (First Paralogism), or simple (Second Paralogism), or distinct from material things (Fourth Paralogism)”. Kant’s metaphysical assertions concerning the thinking being that will be considered later merely presuppose that there is something, and that this something is the being itself. 5. In 1787 Kant attacks the system of Rational Psychology—understood as a rational science resting on rational necessary principles—through his criticism of the first syllogism of topics; in his view, you only have to denounce its paralogistic character and the whole system will collapse. 6. Kannisto (2018) points out two principal lines of interpretation on whether the paralogisms really commit any formal error: According to Allison (2004), the transcendental paralogisms are deductively invalid syllogisms, whereas Bennett (1974), Ameriks (2000), and Van Cleve (1999) deny that they are formal fallacies. The author introduces a clear definition which reconciles the two extremes by focusing on Kant’s distinction between logical and transcendental paralogism; the former “consists in the falsity of a syllogism as to form”, while the latter has a specifically “transcendental ground for inferring falsely as to form” (A341/B399, translation amended by the author). Kannisto develops another view: Transcendental paralogisms are (deductively) valid and sound syllogisms in general logic, but according to a transcendental logic they are considered formal fallacies.

Two Senses of ‘I think’  73 7. To become a cognition, the concept of ‘substance’ must indicate an object; hence, it must be necessarily grounded on a persisting intuition as the indispensable condition of a concept’s objective reality. However, nothing persists in the inner intuition, for the “I is only the consciousness of my thinking; thus if we stay merely with thinking, we also lack the necessary condition for applying the concept of substance, i.e., of a subject subsisting for itself, to itself as a thinking being” (B413). Cf. Waibel (2017) for a more general approach on this topic. 8. The distinction between the logical and the real subject recalls the distinction between the logical and the real predicate as used by Kant in the ontological proof of the existence of God. Existence is not a real but a logical predicate; it can be used as a predicate-concept in a judgment to predicate a thing, e.g., in a judgment such as “God exists”. Given that it is not the property of a thing, it is not a real predicate: A logical predicate only concerns the relationship between a concept and an object, i.e., there is an object to which the concept ‘God’ corresponds. In fact, over and above the best-known distinction between logical and real predicates, Kant considers several notions in their ‘logical’ and ‘real’ aspects (cf. Chapter 4). 9. Going through the steps that outline the Cartesian distinction between deduction, i.e., the inference of something that necessarily follows from other propositions that are known with certainty, and intuition, i.e., the faculty by which it is possible to capture the initial certainty of a proposition in an immediate and self-evident way to make deduction possible, Markie (1992, 145) asserts that, in the cogito argument, Descartes’ use of intuition is based on two different types of intuition: “The set of intuited propositions includes both self-evident propositions not inferred from any others and propositions immediately inferred from self-evident premises”. Propositions about mental states are self-evident and not inferred from others—i.e., they are intuitive in a narrow sense—while propositions about one’s existence are immediately inferred from propositions about mental states, i.e., they are intuitive in a broad sense. In support of this argument, the author presents several Cartesian passages (e.g., his replies to Mersenne’s objections) in which he refers to lumen naturale, ratio naturalis, and distinct perceptions as criteria to explain his certainty about the thinking and its existence. There are, however, several problems related to The Self-Evident Intuition/Immediate Inference Interpretation: Markie himself reformulates his argument to reconcile the contrast between the passages in which Descartes refers explicitly to the two senses of intuition as the grounds for the cogito argument and for those claiming the need for prior knowledge of the general proposition for which it is impossible for that which thinks not to exist (for example, cf. the tenth article of the first part of the Principles of Philosophy). A syllogistic structure of the cogito argument is thus maintained (cf. Williams 1978). Other commentators bring into question the very possibility of inferring the existence of the subject from the consciousness of thought; cf. Kenny (1968, 169), Wilson (1978, 55), and, obviously, Hintikka (1962). For an effective critique of Hintikka’s non-inferential reading of the cogito argument, cf. Bonomi (1991, 19): his reading is similar to Markie’s (1986, 1992) in some respects, particularly when he focuses on the plain inferences highlighted by Descartes, also with reference to the cogito, which can be grasped intuitively to affirm the compatibility between the inferential nature of the cogito and its intuitive character. Moreover, on several occasions Descartes himself argues that the cogito is not based on a syllogism: Obviously, as Markie and Bonomi’s interpretations suggest, the non-syllogistic nature of the cogito does not imply its non-inferentiality. 10. In the same passage, Kant adds: “An indeterminate perception here signifies only something real, which was given, and indeed only to thinking in general, thus not as appearance, and also not as a thing in itself (a noumenon), but rather

74  Two Senses of ‘I think’ as something that in fact exists and is indicated as an existing thing in the proposition ‘I think’ ” (B423 n.). Here, Kant employs different boundaries—certainly intertwined, although kept separate—and operates between the epistemological and the ontological dimensions, the empirical and the transcendental levels, and the empirical and the transcendental apperceptions. In the same passage, he also points out that if ‘I think’ is an empirical proposition, the I is a purely intellectual representation of thought in general; nonetheless, the ‘I think’ act occurs only by means of an empirical representation providing material for thought. 11. This entails holding the anti-reflexive character of self-consciousness, which Sturma and Ameriks (although under different circumstances) have argued against Henrich’s Fichtian interpretation of apperception based on the Reflexion Theory, as will be seen later on. 12. If the ‘I think’ expresses the act of determining the transcendental subject’s existence, then such an existence, though given, is not determined based on manner, that is, to the manifold which belongs to it. A similar determination would require not only an intuition—more precisely, a self-intuition, which is grounded in time as the a priori sensible form of the receptivity of the determinable—but also another self-intuition solely relative to what is determining, namely to mere spontaneity, in much the same way as time provides the condition of the determinable. However, there is only an intellectual consciousness of spontaneity, and since Kant allows for no intellectual intuition, the subject cannot determine its existence as that of a self-active being (cf. B158, n.). 13. These two theoretical dimensions must be kept distinct, although it is possible to find a point of convergence. The self-reference involved in apperception considers the consciousness of oneself as that which carries out the action or activity of synthesising; the relative representation I in apperception is exactly one of the two ‘Is’ contemplated in Anth.: The ‘I’ as the subject of thinking (in logic) means pure apperception (the merely reflecting ‘I’) and should be distinguished from the way the subject appears to itself in the inner sense, i.e., the ‘I’ as the object of the perception. On the other hand, the I as a representation of apperception is not merely a function of thought but has a more substantial underlying nature compared to that which is manifest at the phenomenal level. Schulting (2011a, 168) highlights this when he states that the ‘I’ of the transcendental apperception refers to someone in particular who is doing the synthesising, “viz., the ‘I’ as a substantial thing in itself, so the noumenal and not the empirical self, rather than to a mere activity”. 14. The author refers to several passages, particularly to B409–10 and A360, and especially to A492/B520, in which Kant seems to equate the transcendental subject with the self proper, as this exists in itself. The issue is also referred to as ‘Kant’s third self’; cf. Kitcher (1984), Schulting (2011a), and Marshall (2013). 15. As is well known, with Prauss (1974, 2015) and Allison (2004)—to mention two classics within the huge Kantian debate—some important distinctions need be made. On one hand, Kant introduces Ding an sich (and its variants, i.e., Sache, Gegenstand, and Object an sich) as a shortened form of Ding an sich selbst (and its variants, i.e., Sache, Gegenstand, and Object an sich selbst), particularly for Ding an sich selbst betrachtet. On first approximation, if it is possible to distinguish between Erscheinung (appearance), the indeterminate object of sensible intuition, and Phänomenon (phenomena), the sensible object falling under the categories, then it should be also possible to recognise a further distinction between Ding an sich (thing-in-itself) and Noumenon: “The former is conceptually undetermined, since our thought of it is empty of real content, while the latter, as putative object of an intellectual intuition, is ‘conceptually’ determined, though not for our discursive intellect” (Allison 2004, 58).

3 The Problem of Self-identification

The preceding chapter examined how the ‘I think’ (qua Thinking) can be regarded as the synthetic unity of apperception determining the I think (qua Representation) as an analytical unity of apperception. Accordingly, the I think (qua Representation) can be said to be the representation through which the spontaneity of ‘I think’ (qua Thinking) is given to one’s self. Due to the absence of any epistemic mediation of identification, the I think (qua Representation) is regarded as a simple representation, designating the activity of thinking transcendentally: It serves as the nexus established by the copula in the judgment that links the representational synthesis and the synthetic unity of apperception on a conceptual basis. The possibility of thinking of oneself as a subject is solely ensured by the simple and empty representation I. This interpretative reading raises the three questions: (1) a semantic problem concerning the nature of the representation ‘I’, (2) an epistemic problem concerning the type of knowledge relative to the thinking subject that is produced by the representation I think, and (3) a metaphysical problem posed by those features assigned to the nature of the thinking subject. Thus far, we have singled out three theses in order to answer these questions: (1) a metaphysical thesis: Thinking, the being itself, is spontaneity, (2) an epistemic thesis: Thinking is not the object of knowledge. If Thinking is the being itself, and nothing is thereby given to the subject for thinking, then the subject can think of itself only as an object in general through the concept of the transcendental subject, i.e., the thought of a thinking being that has an existence in itself, and (3) a semantic thesis: The I think, the bare or empty representation I, is the representational vehicle of the concept of the transcendental subject; as such, it is a simple representation. The awareness of oneself as thinking is only expressed by the I—the intellectual representation of the spontaneity of a thinking subject. As seen in the first chapter, self-consciousness is characterised by a first-person perspective and consciousness of the self-as-subject as two

76  The Problem of Self-identification interdependent features, one being the condition of the other and vice versa. The subject’s manifestation of strong, first-person phenomena—one’s ability to attribute first-personal reference to oneself—depends on the subject’s capacity to be the consciousness of herself as the thing doing the thinking, and the consciousness of self-as-subject cannot be gained unless the subject displays strong first-person phenomena. Two such features defining the notion of basic self-consciousness are grounded in a few epistemic and semantic peculiarities in terms of the ability to use the term or concept I in de se or I-thoughts: Essential indexicality and immunity to error through misidentification. In the transcendental system, the epistemic (2) and the semantic theses (3) seem to stem directly from the metaphysical thesis (1). In this chapter, the epistemic thesis and some points of the semantic thesis will be described with special attention to the problems of self-identification and reference of I; on the other hand, the semantic thesis and the problem of transcendental designation will be addressed in depth in the ensuing chapter. Although Kant regards the epistemic and the semantic issue as strictly interconnected and, as such, treats them in the same contexts, when it comes to the problem of selfidentification the two theoretical dimensions should be kept apart. While the epistemic bone of contention consists in how the subject thinks of itself and the knowledge it can form through the I-thought based on the representation I think (cf. the following two paragraphs), the semantic question is concerned with the nature of the representation I and the type of designation (if any) involved in the reference of the thinking subject. It has already been hinted that there are several interpretative readings on these issues; in this chapter, both an elusive reading—the so-termed No-ownership Reading—and the thesis of exclusion will be rejected (cf. the third paragraph), whereas the nature of transcendental designation will be discussed in the next chapter. The affinity between Wittgenstein and Kant, alleged by some commentators within this framework, will be examined in a new light in order to assess the appropriate theoretical bond between the Kantian I think and the contemporary perspectives on the question of self-identification (cf. the last two paragraphs).

‘I think’ and the Question of Self-Identification The question of self-identification in the transcendental apparatus can be discussed via the three theoretical dimensions touched upon above: The metaphysical, the epistemic, and the semantic angles. Firstly, the metaphysical reflection on Thinking must be taken into consideration in order to detect those characterisations assigned to the transcendental apperception that can be traced in several passages from the Transcendental Deduction and the Paralogisms sections. As has been shown previously, from the perspective of the spontaneity of understanding, the transcendental constraints consist of the conditions for the possibility of

The Problem of Self-identification  77 experience and knowledge and are ultimately based upon transcendental apperception, i.e., self-consciousness, which, via the I think, Kant sees as the highest point of Transcendental Philosophy itself. Regarded as a synthetic unity of apperception, apperception is the foundation of representational synthesis in order for knowledge to occur; this calls the claim that categories have objective validity into question, and that they are predicates for objects in general; it follows that judgments can be formed wherever knowledge arises. On the other hand, regarded as an analytical unity of apperception, the representation I produced by apperception is a feature of every representation precisely because I think must be able to accompany every representation. With regard to the epistemic thesis, this can be said to consist of two parts. The first part is negative and asserts that Thinking is not the object of knowledge; the second part is positive and maintains that the subject can think of itself only as an object in general through the concept of the transcendental subject, the representational vehicle of which is I. Secondly, within the transcendental constraints characterising the designation of I in I think, the empirical dimension and the epistemic conditions in and under which the subject reveals itself in the temporal sphere of receptivity will be object of discussion. In order to specify the negative part of the thesis, it is necessary to refer to the analysis of the Paralogisms, in which Kant establishes an epistemic closure regarding the ontological nature of the thinking subject and ultimately denies the possibility of pure or empirical apperception to be the object of knowledge, regardless of the conditions of possible experience. Firstly, Thinking is not determined representationally; according to the points highlighted in the previous chapter, ‘I think’ as spontaneity (qua Thinking) represents no thinking subject—it represents nothing with regard to the properties or the nature of ‘I’. For instance, in B157, Kant states that, in the original synthetic unity of apperception, the subject is not conscious of itself as it appears to itself, nor as it is in itself, whereas in B429, he reiterates that in no way does Thinking present the subject of consciousness as appearance because it takes no account of the kind of intuition involved, whether sensible or intellectual: “In this way I represent myself to myself neither as I am nor as I appear to myself”. Secondly, Thinking is not the object of knowledge in the sense that, whatever its entity, it cannot be defined as one particular object that one can distinguish from others in experience. Thinking does not present the subject as appearance (B428–9), nor is there any representation that can identify an object (A346/B404): The representation I, which is in every thought, is in no way bound up with an intuition, which would distinguish it from other objects of intuition (A350). If Thinking is not the object of knowledge, then it is not possible to acquire cognition about its nature or its identity conditions through I think. As a result, Thinking cannot be explained in terms of, or be derived from, the occurrence of further acts, representations or other

78  The Problem of Self-identification mental phenomena. No further explanation of ‘I think’ qua Thinking can be given other than it being pure and original spontaneity: All manifold of intuition has a necessary relation to the I think in the same subject in which this manifold is to be encountered. But this representation is an act of spontaneity, i.e., it cannot be regarded as belonging to sensibility. I call it the pure apperception, in order to distinguish it from the empirical one, or also the original apperception, since it is that selfconsciousness which, because it produces the representation I think, which must be able to accompany all others and which in all consciousness is one and the same, cannot be accompanied by any further representation. (B132) From this metaphysical perspective emerge a few transcendental characterisations that form the small positive part of the epistemic thesis, and which may also explain the lack of identification in representational synthesis. On one hand, the Thinking Being is something in general, unidentifiable from an epistemic viewpoint; on the other hand, the Thinking Being is an intellectual, self-existence awareness summarised by the I am or I think representations that accompany every other representation and which, as such, present no property: I think/I am is a merely formal condition of all thinking. In fact, due to the absence of intuition, it is not possible to determine whether that something exists as a persistent substance in order to produce knowledge: “The consciousness of myself in the representation I is no intuition at all, but a merely intellectual representation of the self-activity of a thinking subject” (B278); “in the synthetic original unity of apperception, I am conscious of myself not as I appear to myself, nor as I am in myself, but only that I am. This representation is a thinking, not an intuiting” (B157). What is being assumed on the basis of the representation I is merely an existence devoid of properties. The subject is able to know that it exists as a thinking activity, but it is not able to know what it is: Its being is inaccessible from an epistemic angle, and what is given is nothing more than thoughts that are regarded as its predicates, which do not allow us to grasp the thinking subject’s nature. In the well-known passage from Paralogisms concerning a transcendental doctrine of the soul, Kant states: At the ground of this doctrine we can place nothing but the simple and in content for itself wholly empty representation I, of which one cannot even say that it is a concept, but a mere consciousness that accompanies every concept. Through this I, or He, or It (the thing), which thinks, nothing further is represented than a transcendental subject of thoughts = x, which is recognized only through the thoughts that are its predicates, and about which, in abstraction, we can never have even the least concept. (A346/B404)

The Problem of Self-identification  79 Accordingly (we are finally confronted with the semantic thesis), a few peculiarities concerning the self-referential apparatus involved in transcendental apperception appear: “The subject of inherence is designated only transcendentally through the I that is appended to thoughts, without noting the least property of it, or cognizing or knowing anything at all about it” (A355). The act of reference performed by the subject to refer to its self does not entail a mediation of knowing—it does not involve identification by means of properties ascribable to the subject itself. With the notion of transcendental designation, Kant anticipates some of the self-reference without identification features (cf. Brook 2001). The condition of possibility for all judgments relies on the act ‘I think’; at this level, the intellectual representation I only designates transcendentally because no conceptual mediation is involved: It is a simple representation that has no content and merely refers to something in general, i.e., the concept of the transcendental subject—“Its properties [of subject] are entirely abstracted from if it is designated merely through the expression ‘I’, wholly empty of content (which I can apply to every thinking subject)” (A355). As an empty or bare form (A443/B471), I designates but does not represent: For in that which we call the soul, everything is in continual flux, and it has nothing abiding, except perhaps (if one insists) the I, which is simple only because this representation has no content, and hence no manifold, on account of which it seems to represent a simple object, or better put, it seems to designate one. (A381) As mentioned previously, the next chapter will examine the semantic features of transcendental designation in depth; before addressing the nature of designation, however, it is necessary to ascertain whether the I of I think designates at all. In Kant, this issue is closely connected with the epistemic features involved in the question of self-identification: Even though the above-mentioned passages seem to be quite clear on this point—“the subject of inherence is designated through the I”—the lack of identification has led some commentators to detect a close proximity between Kant and Wittgenstein on the elusive No-ownership Reading, whereby the I does not refer at all.

The Genesis of the Cartesian Self In general, it should be highlighted that the absence of an identification component does not imply that the I performs no referential function, nor that it necessarily involves a specific metaphysical thesis pertaining to the nature of the self-conscious subject. In fact, the I-thoughts’ self-reference features have been supported by both a materialist conception regarding the self-conscious subject as a bodily object—by Strawson and Evans, for

80  The Problem of Self-identification example—and a different metaphysical framework, as in Wittgenstein’s eliminativist thesis. The point here consists of the possible contiguity between the analysis of the form of I think and the contemporary reflections on the question of selfidentification: Although it is possible to find different elements of affinity, the theoretical contexts are still deeply distant. With regard to the transcendental designation—based on the principle that “the subject is designated only transcendentally without knowing anything at all about it”—Kant certainly does not seem to be anachronistic with regard to the considerations raised by the contemporary debate. On the contrary, at least as far as the genesis of the Cartesian illusion concerning the thinking subject’s immaterial nature is concerned—one of the issues addressed in the Blue Book and in the first Critique—Kant and Wittgenstein seem to share the same philosophical concerns, as they both focus, although not exclusively, on the type of reference involved in the I, obviously via extremely different philosophical paths and with what may appear antipodal, metaphysical assumptions at first sight. As mentioned in the first chapter, in a well-known passage, Wittgenstein (1958, 66–7) introduces his philosophical–linguistic analysis of the grammatical rule of the term I, where he distinguishes two types of uses, the use as object (“I have grown six inches”) and the use as subject (“I have toothache”): One can point to the difference between these two categories by saying: The cases of the first category involve the recognition of a particular person, and there is in these cases the possibility of an error. . . . On the other hand, there is no question of recognizing a person when I say I have toothache. To ask ‘are you sure it’s you who have pains?’ would be nonsensical. This passage should be considered as part of the philosophical framework articulated by Wittgenstein since the 1930s on the basis of some theses that might be regarded as the background for the analyses of the two uses of I. For example: (1) the irreducibility of the manifold ‘games’ that build up language, the rules of which are to be made explicit in order to solve any type of philosophical problem, (2) anti-referentialism, which relies on the recognition of the manifold functions performed by language, as well as on the necessity of avoiding the erroneous search for the use of a sign on the basis of the object–sign relationship, and (3) anti-mentalism, according to which suggesting that thinking is a mental activity is misleading. While the I used as object performs a referential function relative to the body and to physical features in general, the I used as subject apparently regards mental states, as well as processes, but no subject identification is taken into account. From a Wittgensteinian angle, the I used as subject has no referential function: According to this thesis—supported by Geach (1957), Hacker (1972), and Anscombe (1975)—it is our inclination to assume that a linguistic term only

The Problem of Self-identification  81 has a meaning if it stands for an object that induces us to believe that the I used as subject denotes the thinking subject, mind, soul, etc. (cf. Sluga 1996; Wright 1998). The above-mentioned theses are directed towards anti-objectivism, i.e., the refusal to conceive of meaning in terms of the object and the designation model, as Wittgenstein (1953, § 293) would later use to demolish the model based on his well-known private language argument. The use of I as subject has no object (whether material or immaterial) to which to refer: The mental and the physical belong to two different language-games, and reducing their relationship to a simple formula is indefensible. Instead, it is necessary to show the differences between these games related to mental and physical properties so as to examine the grammars of first- and third-person statements. In this way, Wittgenstein (1958, 43) begins with the analysis of language and the use of the I as subject to dissolve any issue concerning the nature of the ego in an anti-metaphysical key. Philosophical inquiry must only investigate the grammars of the mentalistic terms used—no metaphysical distinction between the mental and the physical should follow from the distinction between propositions describing facts of the world and propositions describing psychological experiences. It is necessary to analyse the uses of and related grammars for terms such as thinking, meaning, and wishing because the investigation “rids us of the temptation to look for a peculiar act of thinking, independent of the act of expressing our thoughts, and stowed away in some peculiar medium”. Thinking is using signs according to rules, and philosophical difficulties may only arise from those misleading uses of language that can lead us to look for something that might correspond to a noun. This may be the case with the use of the I as subject: The referential thesis, according to which the use of a sign is based on its relationship to the object (strongly criticised when taken as the sole basis to explain the semantics of the language), along with the proper consideration that some uses of the I do not denote physical properties, leads to false Cartesian metaphysical conclusions: We feel then that in the cases in which ‘I’ is used as subject, we don’t use it because we recognize a particular person by his bodily characteristics; and this creates the illusion that we use this word to refer to something bodiless, which however, has its seat in our body. In fact, this seem to be the real ego, the one of which it was said, ‘Cogito ergo sum’. (Wittgenstein 1958, 69) In no way is the absence of identification in the use of the I omitted by Kant, as seen previously with the transcendental designation of the I in I think. Given that there is no empirical intuition, the I in I think cannot be based on public employment through the identifying mediation of properties attributable to the thinking subject. However, and in contrast

82  The Problem of Self-identification to Wittgenstein, Kant moves from a metaphysical reflection in the sense of transcendental idealism concerning the conditions of possibility of experience and knowledge and from the transcendental assertion that I think is the centre of such conditions. Philosophical inquiry can only analyse the formal constraints of knowledge: In revealing the genesis of the illusion of a Cartesian, immaterial ego, mainly addressed in the analysis of paralogisms, Kant argues that nothing about the metaphysical order and the ontological nature of the thinking subject can be elicited from the conscious form of the unity of apperception or from the representational order of I think, precisely because the I in I think is not the concept of an object, but an empty representation that refers to “the concept of a mere something”. Needless to say, it was Strawson himself who insisted on the characterisations of the form of the I think. Based on some arguments recalling Wittgenstein’s theses in more than one way, Strawson claims that Kant has revealed the source of the Cartesian error from a purely internal, referential use of the I, which severs all ties with the ordinary, empirical criteria of self-identity: If we try to abstract this use, to shake off the connexion with ordinary criteria of personal identity, to arrive at a kind of subject-reference which is wholly and adequately based on nothing but inner experience, what we really do is simply to deprive our use of “I” of any referential force whatever. It will simply express, as Kant would say, “consciousness in general”. If we nevertheless continue to think of the “I” as having referential force, as referring to a subject, then, just because we have really nothing left but the bare form of reference, it will appear that the object of this reference must be an object of singular purity and simplicity—a pure, individual, immaterial substance. Kant sees clearly how the key fact, that immediate self-ascription of thoughts and experiences involves no application of criteria of subject-identity, simultaneously explains three things: it explains the temptation to permit ourselves the use of the notion of the subject of experience (“I”) while thinking exclusively in terms of the inner contents of consciousness (the contents of “inner sense”); it explains why that notion, so used, is really quite empty of content; and it explains why it seems, therefore, to be the notion of an absolutely simple, identical, immaterial individual. (Strawson 1966, 166) At the same time, Strawson criticises Kant for not explicitly considering the requirement that the subject is recognisable as an object of intuition as a condition of possibility of experience; this thesis has been further developed by Evans (1982) and Cassam (1997), and challenged by some Kantian commentators in turn (cf. infra). As said, even though the empty form of the referential apparatus in the transcendental apperception has been appraised in intrinsically different ways (from the Substantial Ownership Reading to the Formal Ownership

The Problem of Self-identification  83 Reading; cf. supra, Chapter 2), an elusive reading suggesting a referenceless I in I think has been argued more than once—the so-termed No-ownership Reading. The close affinity between Wittgenstein and Kant as alleged by some commentators lies precisely within this framework. For instance, Sturma (1985) and Becker (1984) dwell upon the apperception structure involving a criterionless identification of the I in I think. Powell (1990) brings up Anscombe’s Wittegensteinian lesson to define the I think absence of reference. To this purpose, McDowell (1994) offers a Kantian reading which is totally in favour of the I as subject; by doing so, he contributes to what Cassam (1994) terms—in the self-knowledge context of discussion— Elusiveness Thesis, which pertains to both Kant and Hume. By contrast, Kitcher’s (2000) proposal is quite the reverse of the one just discussed. In putting forward her Kantian account of the Transcendental Psychology, Kitcher (1990) rebuts Strawson’s judgment of “imaginary discipline” and upholds her redescription in psychologistic terms, refuting every Wittgensteinian reading of the I think and making a critical attack upon McDowell’s Kantian interpretation. Although antithetical, both readings will be considered in the next paragraph as they account for two key examples of Kantian No-ownership Reading. Besides, Peacocke’s metaphysical thesis of exclusion will be tackled, according to which the I in I think refers to something which in no respect can be said to be an object of the world.

Two No-Ownership Readings and One Thesis of Exclusion Moving from Strawson’s classical interpretation and the assumptions of the introspective model, McDowell criticises the elusive consequences of the I think and the merely formal nature of the continuity of consciousness of the subject through time—a continuity which seems to imply no criterion of identity. The idea of a persisting referent of the I in I think has been acknowledged by Kant as Descartes’ move to generate a substantial subject. However, it is necessary to introduce a conception of merely formal persistence of the representation I think in the flow of consciousness to avoid whichever Cartesian substantialism. As a consequence, the use of I is identification-free, since no track-keeping of an object (of I) in the flow of consciousness is required. There is nothing, as McDowell claims in drawing on one of Locke’s ideas, except the flow of consciousness itself. If, however, there is no track of a persisting object in the flow of consciousness, it does not follow that the subject of the flow of consciousness must be substantially simple, nor that a merely formal, elusive conception of identity of the I involved in the acts of self-reference must be reached. It is here that McDowell recognises a contiguity between Wittgenstein and Kant, between the use of the I as subject—which has no reference—and a merely formal conception of transcendental apperception. This latter is conceived as a geometrical presence in the world at the most, something unlikely to integrate any exhaustive, complete conception of a subject having a body

84  The Problem of Self-identification and a substantial continuity in the objective world—a perspective that McDowell borrows from Strawson and Evans. What is missing in Kant is the idea of second nature and the corresponding notion of Bildung; hence, Kant’s mistake would consist in challenging Descartes’ mistake and substantialism by replacing it with an empty formalism of I, which is not capable of integrating the idea of an ordinary subject situated in the world as a substantially persistent object: Kant’s aim is to exorcize Cartesian temptations about the self, and he gets to the very brink of success. He wants to acknowledge the peculiarities of self-awareness that encourage Cartesian philosophy, but without letting them seem to show that the object of self-awareness is a Cartesian ego. But he thinks the only alternative is a transcendental self-awareness, something that has no object substantially present in the world. If we insist on supplying this self-awareness with an object, we can locate the object in the world only geometrically, as a point of view. This avoids the familiar Cartesian problems about the relation between a peculiar substance and the rest of reality. But it leaves us with what look like descendants of those problems. If we start from a putative sense of self as at most geometrically in the world, how can we work up from there to the sense of self we actually have, as a bodily presence in the world? (McDowell 1994, 104) Kitcher’s proposal, on the other hand, is quite the reverse. She criticises the different commentators resorting to Wittgenstein to read Kant’s I think. Within the cognitive science framework reassessed in the transcendental psychology terms, the doctrine of apperception is purged of any self-reflective component and relation with a subjective reference for the attribution of experience and turns into the assumption (in fully cognitivist terms) of a necessary connection of the mind-system. The referential apparatus of the I think will still have a function only because I refers to neither substantial subject nor formal apparatus, but only to an interconnected system of mental states. Both Kitcher’s criticism of McDowell and her different Kantian Noownership Reading are made clear. The scholar poses the issue of criterionless self-ascription as a key factor in the philosophical issue of the other minds problem, that is, the alleged asymmetry between first- and third-person ascriptions of mental states. While the inference that can be elicited from the existence of the other minds and the corresponding mental states is always dubious, from a first-person perspective there is no doubt as to who is the subject of the self-ascribed thoughts; for this reason, the ascription of mental states is immune to error through misidentification. Kitcher draws on different Kantian passages on the inadequate role played by psychology and introspection in the analysis of one’s own as well as others’ psychological dimensions to prove that Kant has not posed any asymmetry issue

The Problem of Self-identification  85 whatsoever. The subject manifests itself as Erscheinung in the same way that others do: There is not “any special knowledge of our own selves and our states” (2000, 50). The self-ascription of thoughts by the I in I think is an obviously analytic, necessary truth, and yet it depends on the synthetic unity and has no relationship with the feature of the irreducibility of self-conscious thoughts. Indeed, the scholar remarks that Kant’s alleged acknowledgement of the IEM has not brought him to prove “any special knowledge of a special subject of thoughts”; precisely, according to Kitcher “our knowledge of our own selves, as appearances, is on a par with our knowledge of all else, including the minds of other”: Kant stressed the mediacy of self-knowledge in principle as well as practice. Central tenets of transcendental idealism implied that subjects had no direct or privileged access to their own mental states. Probably in response to the accusation of Berkeleyan idealism, Kant took some pains to explain in the second edition that his position was that, since we can have knowledge of our inner states only through the form of inner sense and the categories of the understanding, we know “also even our own selves, only as we appear” (B152–3, see also B409). Because Kant denied—for reasons of principle—any special knowledge of our own selves and our states, the asymmetry between self-knowledge and knowledge of ‘other minds’ that is essential for taking the problems of ‘error-proof’ self-ascription and other minds seriously is absent from his philosophy. I do not see how he can be credited with the insight that, although self-ascription of mental states is immune to error through misidentification, that does not imply any special knowledge of a special subject of thought, when his own position was that our knowledge of our own selves, as appearances, is on a par with our knowledge of all else, including the minds of others. (Kitcher 2000, 50) Although diametrically different, Kitcher and McDowell’s readings are based on the perceptual model of self-knowledge to stress the elusive, nonreferential character of the I think as both a premise and a conclusion. McDowell takes it as a premise that there is no track of a self in the flow of consciousness, which eventually leads us to assume, with Kant, a nonexhaustive conclusion of subjectivity. On the other hand, it is a conclusion in the view taken by Kitcher, who, in reading apperception only as a presupposition of the synthetic mind, holds the Dennettian virtuality of the I think. This manifests itself in the mind as Erscheinung like the others—it has no referential function if not in the transcendentally interconnected mindsystem. Therefore, as there is no thinking subject whatsoever, the scholar dismisses any feature of self-ascription of self-conscious thoughts. Mutatis mutandis, both readings apparently assume a perceptual model of self-consciousness in simplistic terms; however, no use of an introspective

86  The Problem of Self-identification model of self-conscious forms is being refuted here—Kant had been awakened from his dogmatic slumbers by reading Hume. What is rather necessary to highlight here is that the empirical model has to be connected in a more strict way to the transcendental framework in order to avoid an elusive conception of the thinking subject. In introspection, as well as in the empirical apperception, no fixed, abiding self can be found, yet this simply entails that we cannot draw a knowledge of the thinking subject, not that there is no thinking subject in the least. On the other hand, at a transcendental level the hypothesis is that the I in I think does not manifest itself as Erscheinung like the others (in opposition to Kitcher’s reading), but it has a referential function (unlike McDowell’s reading), adding together the I-thoughts features in the peculiar terms of transcendentalism. Let us begin with the passage in which Kant states that the inner sense offers consciousness the self only as it appears phenomenically (B153). Kitcher quotes this passage to equate the alleged epistemic status of one’s own mind with that of the others; in reality, it is necessary to set this argument against the general background of the transcendental framework, and, first and foremost, of the distinction between inner sense and transcendental apperception (cf. supra, Chapter 1). As said previously, pure apperception is the original consciousness which can be expressed by sum (cf. supra, Chapter 2). As emphasised by Capozzi (2007, 287), an ontological question arises here: Sum is nothing but activity, which has nothing receptive about it in the sense that it does not mingle with any element of the sensible dimension— it is a thinking activity in such a way that sum and cogito are on a par. The ontological question is specified in the assertion that “the subject bound to the first act of knowledge, to primitive apperception, is the first subject as well as the first Wesen being thought: with the first act of knowledge, the subject is the being itself” (2007, 287). This being, however, is something in general, unidentifiable from an epistemic point of view; it is an intellectual consciousness of self-existence, summarised by either the I am or the I think representations which accompany every other representation. Only with empirical apperception does the I as object reveal itself as an appearance in the intuition (and phenomenalises itself): The empirical apperception indirectly obtains a persistence and identity, as well as the kind of existence an appearance has (cf. infra, the last paragraph). Thus, the I as subject is not an appearance on a par with others: The I in I think is a purely intellectual representation: From its first-person perspective the subject can know “I am”, and this representation is a thinking (B153). At any rate, in the flow of consciousness the taking on of an I is not missing, for the I think has a designative function referring to something which really exists as an act of thinking (and phenomenalises itself as empirical apperception). As said previously, if the Thinking is the synthetic unity of apperception, I think is an analytical unity of apperception, i.e., the representation through which the spontaneity is given to one’s own self. In the absence of epistemic mediations of identification, considered as simple

The Problem of Self-identification  87 representations, the I think (qua Representation) merely designates the activity of thinking transcendentally. As for the exclusion thesis, although Peacocke does not explicitly mention Strawson, he uses the same type of argument used by Strawson against Kant in order to affirm a metaphysical thesis of exclusion whereby the I in I think as an intellectual representation produces no knowledge about the nature of the thinking subject or self because it refers to something which in no respect can be said to be an object. As said in the previous chapter, the dualism in the thesis of exclusion, that is, between a person and a transcendental or metaphysical subject expelling any subjective reference from the world, is remarkably different from the No-ownership Reading’s, which contemplates no subjective reference at all. Peacocke contends that the German philosopher mistook an epistemological phenomenon for a metaphysical one, thus establishing a wrong slide “from truths about the level of uses of the first-person concept to very questionable claims about the level of reference” (1999, 285). The author refers, among others, to the above-mentioned B404 to specify the nature of the transcendental subject; unlike the formal reading adopted in these pages, Peacocke conceives the “transcendental subject” as the thinking subject, whose metaphysical nature cannot be object of knowledge for the simple reason that it is not an object of the world: Accordingly, this is not an object empirically determinable through the application of the categories, and it can only be known through the thoughts regarded as its predicates. This metaphysical conclusion can only stem from a failure to recognise the representationally independent use of the I in I think, which is the form of every judgment; as such, however, it neither presents the subject in the content of the judgment itself, nor can determine or identify it a fortiori: The reason why the transcendental subject can be known only through its thoughts is that consciousness in itself is not a representation distinguishing a particular object, but a form of representation. Precisely what distinguishes a representationally independent use of ‘I’ when a thinker knowledgeably thinks ‘I am F’ is that the thinker’s reason for self-applying the predicate F is not that one of his conscious states has the content ‘I am F’. His reason is not given by a ‘representation distinguishing a particular object’ as F. It is rather the occurrence of a certain kind of conscious state itself which is his reason for making the judgement. . . . It obviously does not follow that what a representationally independent use of ‘I’ refers to is something which is not subject to the categories. That would be a slide from truths about the level of uses of the first-person concept to very questionable claims about the level of reference. That slide does seem to be present in some passages in Kant. . . . From the fact that in certain self-ascriptions the reference of ‘I’ is not then given as an object, it does not follow that it is not subject to the categories. (Peacocke 1999, 284)

88  The Problem of Self-identification For those who harbour sympathies for the hypothesis for which Kant asserts a metaphysical thesis of exclusion, Peacocke’s argument might be reversed by contending that Kant may have held a metaphysical position on the exclusion of the subject from the metaphysical order of reality to generate the epistemic phenomenon of the representationally independent use of I, not vice versa. For those wishing to recall the Formal Ownership Reading adopted in these pages, and the above-mentioned Kantian arguments contained in the analysis of paralogisms, it is not possible to elicit any metaphysical conclusion about the nature of the thinking subject from the formal and representational order of the I think. Indeed, when Kant focuses on the logical exposition of apperception, he describes the subject’s function through a completely abstract locution such as Das Denken, pointing up that the subject can think itself only as abstraction from any kind of intuition thanks to the thought which is normally and typically used to think an object in general (cf. supra, Chapter 2): Thinking, taken in itself, is merely the logical function and hence the sheer spontaneity of combining the manifold of a merely possible intuition; and in no way does it present the subject of consciousness as appearance, merely because it takes no account at all of the kind of intuition, whether it is sensible or intellectual. In this way I represent myself to myself neither as I am nor as I appear to myself, but rather I think myself only as I do every object in general from whose kind of intuition I abstract. (B428–9) At this transcendental level, the thinking thing is the being itself for the simple thought; it shows no properties at all, to the extent that Kant leaves pronouns (‘I’, ‘he’, ‘it’) indeterminate. As a result, the subject’s thought employed to think itself as an object in general is the concept of something in general, i.e., the concept of transcendental subject. Once again, it is worth noting that, according to the Formal Ownership Reading, the transcendental subject is a concept, so the first-order talk about the nature of the determining subject can be replaced by the second-order talk about the concept of transcendental subject through which the subject represents its relation with itself as subject. The I think is the very representation which means the concept of transcendental subject; it displays a designative function referring to something which really exists as (an act of) Thinking precisely because Thinking, as spontaneity, is the synthetic unity of apperception which produces the representation I think as an analytical unity of apperception.

The I of Pure Apperception and the I as Subject All things considered—and we are finally confronted with the conclusions that mark the distance between Kant and contemporary perspectives—the

The Problem of Self-identification  89 considerations concerning the I in I think reside at a remarkably different level of investigation from the one assumed by contemporary approaches. The transcendental unity of apperception is the foundation of the representational synthesis, through which there arises an objective determination of representations for possible cognition: Each empirical manifold given in the intuitions of sensibility is determined by the functions of the power of judgment based on the application of the categories of understanding that bring it back to consciousness. In this sense, every manifold has a necessary relationship to ‘I think’; that is, the foundation of the necessary unity of the objectively valid connection of all representations expressed by the judgment. In this view, ‘I think’ resides in a metaphysical frame that necessarily involves any thinking activity because it identifies with such an activity. At this level of investigation at least, and with regard to the passages considered, this represents the highest level of abstraction in the transcendental reflection. At the same time, in transcendental self-consciousness, the selfattributions of any thought (and of a transcendental category, e.g., ‘I think’ substance, cause, etc.) are not based on any identification component relative to the representation I underlying the determination of those thoughts. As mentioned previously, although the representation I designates transcendentally in a way that requires a more detailed explanation (cf. infra, Chapter 4), it cannot be determined to identify the thinking entity as an empirical object: “The consciousness in itself is not even a representation distinguishing a particular object but rather a form of representation in general, insofar as it is to be called a cognition” (A346/B404). However, if the act of spontaneity expressed by I think is necessarily involved in the making of any judgment, the lack of a subjective identification component entailed in the transcendental designation of I appears to be totally empty of meaning. The Kantian reflections cannot articulate the different types of singular judgments expressing self-ascriptions of mental and physical properties: The ‘I think’ concerns the form and condition of the possibility of any kind of judgment, regardless of the particular use of I (as subject or as object, in Wittgensteinian terms) involved in the singular judgments being produced. In other words, and more practically, the transcendental designation mechanism of I think cannot account for the presence or absence of a subjective identification component relative to the first person in judgments such as “I have grown six inches” or “I have toothache” because such is the condition of possibility of both. In this regard, Longuenesse (2012) distinguishes two different uses of I as subject. The first use is associated with Strawson–Evans and Cassam (cf. Longuenesse 2006), who refer to the consciousness of the self as a spatiotemporally located object and to the channels through which the subject determines information about himself to produce possible IEM judgments relative to the first person. In Chapter 1, an I-thought was seen to consist of a thought in which, to put it in Evans’ terms (1982, 207), a subject of knowledge and action has thoughts about himself—i.e., about a subject of

90  The Problem of Self-identification thought and action. Thoughts regarding the very subject of these thoughts rely on a special form of access to the information about the subject herself, and it is due to this information that the knowledge expressed in the judgment takes shape. Evans singles out several kinds of access: According to our general capacity for perceiving bodies, and to our sense of proprioception, of balance, of heat and cold, and of pressure, the kind of information generated by each of these modes of perception seems to give rise to judgments with immunity to error through misidentification.1 On the other hand, the second I as subject use highlighted by Longuenesse (2012, 88) is linked to Kant’s view as well as to the subject’s awareness of mental unity: There is no longer a particular spatio-temporal entity at issue, as “the subject of thought is to herself part of the content of her thought in no other capacity than as the agent of a process of thinking that generates a specific kind of unity of mental contents”. As a result, “we should now add to consciousness of oneself via self-location and, in some cases, proprioception/kinaesthesia, the mere consciousness of oneself as the agent of the unity of one’s thought, a unity one is accountable for” (2012, 91). Longuenesse rejects Evans’s criticism of Kant—in fact, this is an attack that was previously put forth by Strawson (1966) and hence emphasised by McDowell (1994)—whereby the I think displays a purely formal characterisation that cannot account for the self-referential capacity of the selfconscious subject. In some instances, the Kantian self-consciousness may suffice to account for the use of ‘I’ (Longuenesse 2012, 91). In her last book, the author (2017, 27) clarified the point and the two different uses of the I as subject in more detail, distinguishing between information about the subject that justifies the subject’s self-ascription of a given predicate and the concept of the subject due to which the subject can use the representation I in order to refer to itself. In her example, the subject can ascribe to itself the predicate “think this is a tree” on the basis of information concerning its own location and physical properties, but—and this is the point in question highlighted by the author—“this does not mean that a conception of oneself as spatially located and embodied is a condition for the very use of ‘I’ in the argument place of one’s judgment”. Longuenesse (2012, 92) accurately remarks that there is no need to know what it is that entity ultimately consists of as a means to use the I in I think— I refers to the subject of thinking in its logical rule concerning the connectedness of the thoughts. The Kantian reasons advocated to explain the immunity to the misidentification of I think are quite different from the cases addressed by Evans, which are based on the sources of information of the modes of perception that give rise to the epistemic and justification grounds upon which the subject produces such judgments. In the case of I think, the subject’s capacity to think of itself via the representation I, as well as the process of connecting its own thoughts and representations, are mutually conditioning. It is worth pointing out that, for the author, the capacity to refer to oneself is based on the representation ‘I’, which is governed by the token-reflexive

The Problem of Self-identification  91 rule, whereby every token of ‘I’ refers to the subject that has produced or used it, either mentally or linguistically (cf. supra, Chapter 1 and infra, Chapter 4). Thus, having this kind of concept available, ‘I’ is the condition of being engaged in the kind of unifying activity and vice versa. For this reason, as argued by Longuenesse (2017, 29), it is nonsensical to suppose that I could wonder, “Someone is thinking [that this proof is valid, that this is a tree, that it will rain] but is it me?” In this case, and in this case alone, the reason there is no place for this question is not that the judgment is justified by a particular kind of experience. Rather, the reason is the very nature of the process of thinking. In this way, in I think, Longuenesse distinguishes a thin notion of immunity to error through misidentification compared to the cases examined by Evans, which refer to a self-locating subject and an experiencing subject: The subject to which ‘I’ is taken to refer is individuated by no other feature than the fact that she is the agent of the connection under scrutiny as well as the agent of the scrutinizing itself, where the agency is individuated only by its result: the connection of particular mental contents to a totality of experience and knowledge that can be accessed whenever needed. Nevertheless, the notion of immunity to misidentification is not thereby made empty. Saying that attributing ‘think’ to ‘I’ in the circumstances described is immune to error through misidentification relative to the first-person pronoun makes sense. It cannot be the case that ‘think’ is true of someone but false of me—the subject that goes through the proof and checks its steps, the subject that reviews the reasons for thinking this is a tree. (Longuenesse 2012, 93–4) Even though several theoretical differences have been pinpointed in Kant and Evans’ approaches, it seems difficult to compare such two different levels of investigation, as highlighted, mutatis mutandis, by Perry and Recanati within a different theoretical framework (as will be seen in Chapter 5). In addition, if Kant’s ‘I think’ is not a case of the use of I as subject in Evans’ sense, it is not a case of the use of I as subject in Wittgenstein’s sense (amended by Shoemaker), either (cf. Longuenesse 2017, 31), but it is the condition of possibility of both. As mentioned previously, since the act of spontaneity expressed by I think is necessarily involved in the making of each judgment, it resides on a much more abstract level than does the reflection on the distinction between I as subject and I as object: Contrary to Longuenesse’s interpretation, the lack of identification components entailed in the transcendental designation of I think appears to be totally empty of meaning.

92  The Problem of Self-identification The contrast between Longuenesse’s view and my approach is due to the different reading of I think: As seen previously, if Thinking is the synthetic unity of apperception determining I think (qua Representation) as an analytical unity of apperception, the I in I think is the representation produced by the spontaneity of Thinking. Thus, it is not possible to state that ‘think’ is attributed to ‘I’; consequently, it is not possible to state that the judgment “I think” is immune to error through misidentification relative to the concept I. In other words, the lack of an identification component relative to the subject in the forming of the judgment I think is no longer an issue: I is the bare and empty representation generated by, and, as such, analytically contained in, the concept of Thinking. For this reason, as will become apparent in Chapter 4 when the transcendental designation will be taken into account, it is not possible to employ the fundamental reference rule, according to which ‘I’ refers, in any instance of its use, to the author of the thought or the speaker of the sentence in which ‘I’ is a component. If Kant’s I think is an empty form generated by Thinking as an analytical unity of apperception, this thesis is an epistemological conclusion gained from a metaphysical reflection on the features of Thinking and rests on a different theoretical level from that involved in Wittgenstein and contemporary scholars’ approaches (cf. Carl 1997). It is necessary to change perspective; in other words, it is necessary to move to the sensitive dimension of receptivity and the plane of the empirical apperception articulated by Kant. Like Wittgenstein, in Anthropology Kant also introduces ‘I as subject’ and ‘I as object’ on the basis of the distinction between transcendental and empirical apperception; as mentioned, however, their theoretical uses are completely different: If we consciously represent two acts: inner activity (spontaneity), by means of which a concept (a thought) becomes possible, or reflection; and receptiveness (receptivity), by means of which a perception (perception), i.e., empirical intuition, becomes possible, or apprehension; then, consciousness of one’s self (apperception) can be divided into that of reflection and that of apprehension. The first is a consciousness of the understanding, pure apperception; the second a consciousness of the inner sense, empirical apperception. In this case, the former is falsely named inner sense.—In psychology we investigate ourselves according to our ideas of the inner sense; in logic, according to what intellectual consciousness suggest. Now here the ‘I’ appears to us as to be double (which would be contradictory): 1) the ‘I’ as subject of thinking (in logic), which means pure apperception (the merely reflecting ‘I’), and of which there is nothing more to say except that it is a very simple idea; 2) The ‘I’ as object of the perception, therefore of the inner sense, which contains a manifold of determinations that make an inner experience possible. (Anth 7: 135, 23)

The Problem of Self-identification  93 Such a distinction determines that the subject can (re)present itself in two ways: Through the I that thinks and through the I that intuits itself. In another passage, Kant (V-Met-L1/Pölitz: 224, 44–5) states that This I can be taken in a twofold sense: I as human being, and I as intelligence. I, as a human being, am an object of inner and outer sense. I as intelligence am an object of inner sense only. Obviously, this does not imply that there are two Is; on the contrary, the “I as a thinking being am one and the same subject with myself as a sensing being” (Anth 7: 142, 33). This point should also be related to passage B415, in which Kant maintains that the thinking being considered as a human being is also an object of outer sense. From the transcendental perspective, and by abstracting from any modality of intuition, through the pure representation of I, transcendental apperception does not render the thinking subject as a noumenon or as an appearance: “I think myself only as I do every object in general from whose kind of intuition I abstract” (B429). From the empirical perspective, considering I think as an empirical proposition equivalent to I exist thinking, there is no longer a logical function, but only the determination of the subject at the level of existence, the object of intuition that necessarily involves inner sense: I as the object of perception is revealed by empirical apperception as an appearance that unfolds through the form of time. As noted above, if, from the transcendental angle, I think cannot account for the presence or absence of a subjective identification component (in judgments such as “I have grown six inches” or “I have toothache”) because such is the condition of possibility for both (the Kantian I as the subject of thinking, namely, transcendental apperception, is the condition of possibility for all judgments), then certain characterisations of the identification mechanisms of the egological dimension can only be added to the empirical plane of investigation. Empirical apperception involves the pure forms of sensibility, which articulate the Wittgensteinian uses of I as subject and I as object in the specific terms of transcendentalism in judgments expressing self-ascriptions of mental and physical properties. The argument can be constructed as follows: (1) In primis, as well as with regard to empirical apperception, Kant rejects the possibility of moving from the inner perception of something existing as thinking (which he calls “eine unbestimmte empirische Anschauung”) to the determination of this something as substance existing in time and space, the forms of inner and outer sense through which all appearances are given. In any other way, this would not be thought but matter. Recalling the considerations presented in Chapter 1, the consciousness of the self as contemplated by empirical apperception is

94  The Problem of Self-identification the inner perception of something that is not an object of outer sense. In other words, from an empirical perspective, the representation I designates an object of inner sense and, as such, it should not be regarded as a concept for the determination of the subject: For the I is not a concept at all, but only a designation of the object of inner sense insofar as we do not further cognize it through any predicate; hence although it cannot itself be the predicate of any other thing, just as little can it be a determinate concept of an absolute subject, but as in all the other cases it can only be the referring of inner appearances to their unknown subject. (Prol 4: 334, 86) In this way, Allison (2004) rejects any attempt to consider inner sense as a source of empirical self-knowledge, starting from the Kantian account of inner experience described in the so-called Leningrad Reflexion on Inner Sense: Through the intellectual consciousness we represent ourselves but we do not cognize ourselves either as we appear nor as we are, and the proposition ‘I am’ is not an experiential proposition, rather I lay it at the ground of every perception in order to constitute experience. (It is also not a cognitive proposition.) In the case of inner experience, however, which I order, I affect myself insofar as I bring the representations of outer sense into an empirical consciousness of my condition. (Refl 18: 623, 364–5) According to Allison (2004, 278–9), “that inner experience involves a kind of reflective reappropriation of the contents of outer experience”; thus, “its content consists of the very representations through which we cognize external objects”. This point—which reasserts what is stated in B67, as remarked in Chapter 2, although from the perspective of inner experience2—draws a crucial asymmetry between outer and inner experience: As mentioned previously, inner sense does not present a manifold of its own; thus, the sensible representations involved in judgments of inner sense through which the subject represents itself to itself in the empirical dimension do not represent itself as an object. As a result, these empirical representations of the self cannot be conceived of in the same way in which outer intuitions are conceived of as representations of outer objects. Allison’s conclusion is that inner sense “conceives of these representations as belonging to itself as its own ‘subjective objects’. Correlatively, the subject regards itself merely as the substratum or subject in which these representations inhere” (Allison 2004, 279). This reference to the substratum recalls the ‘I of apperception’ involved in connection with inner experience in another Reflexion pointed out by Allison.3

The Problem of Self-identification  95 Therefore, it is due to the absence of impressions of the self that the only role of inner sense that can be taken into account is its enabling the subject to recognise the representations belonging to it as its own. In other words, if the empirical predicates referring to the subject in a judgment of empirical experience can be conceived of as a mere reflective reappropriation of the contents of outer experience, then the subject cannot cognise itself in the same way that it cognises outer objects. The latter are objects of empirical intuitions with spatio-temporal forms, which can therefore be cognised due to the application of intellectual forms of categories based on the judgment; instead, the ‘I’ or self is not an empirical, spatio-temporal object. While Allison’s conclusion is that inner sense does not produce empirical self-knowledge, a different conclusion will be reached in the last chapter following Capozzi’s analysis of the mechanism of attention. (2) Nonetheless, from the empirical perspective, namely by considering I think as an empirical proposition equivalent to I exist thinking, there is still the determination of the subject at the level of existence as long as I as object of the perception, i.e., the passive I of inner sense, is revealed by empirical apperception as an appearance that unfolds through the form of time. As mentioned above, and recalling the self-affection argument and the arguments drawn from the Refutation of Idealism (cf. supra, Chapter 1 and infra, Chapter 5), external sense plays a crucial role: In order for the subject to determine its existence in time, it is necessary to assume the existence of objects perceived by the outer sense, the subject’s body itself, in the first place.4 Following Capozzi (2007), in the consciousness of the self based on the empirical determination of inner perception—i.e., on the subject’s ability to perceive itself, principally in the paradigmatic instance of the psychological mechanism of attention as something that thinks while apprehending representations in the psychological flow towards the outside—the subject reveals itself as appearance and phenomenalises; de facto, by doing so, it indirectly obtains a persistence that exceeds the afore-mentioned variable nature of empirical apperception. This specific point will be examined in the last chapter, when the issue of empirical self-knowledge will be discussed. In this way, according to the distinction between the I that thinks and the I that intuits itself as two different dimensions of the same mind, the subject is given to itself as an appearance in intuition as long as the combination or determination of inner sense by the spontaneity of Thinking is also given. The I as intelligence cognises itself as an object being thought insofar as the subject is also given to itself in intuition, just as with any other appearance.5 (3) This does not imply that the subject phenomenalising in empirical apperception, which requires outer sense in order to be determined in

96  The Problem of Self-identification time as inner appearance, can account for an instance of identification. Just as the I as subject in transcendental apperception poses no question of identification—as Kant contends when he states that “I represent myself to myself neither as I am nor as I appear to myself”—the I as object of empirical apperception cannot account for any instance of identification: It is an appearance that only unfolds in time.

The I of Empirical Apperception Kant states that cognition consists of “the determinate relation of given representations to an object” (B137, cf. also B146–74), whereas in the BIntroduction, the production of cognition is described as a process involving the comparison of different sensory representations. As a result, we can only refer to Erkenntnis, which is always discursive, with regard to the result of the application of conceptual forms to the forms of sensibility, and only with reference to both the pure forms of sensibility, i.e., time and space— two sources of cognition (zwei Erkenntnißquellen), which, taken together, make up the pure forms of all sensible intuitions (A38–9/B55–6). While all possible representations are revealed in the first pure intuition, there is only a specific sub-class of representations in space, i.e., those referring to sheer external objects. Thanks to space itself—the form of outer sense—objects are represented as real and different from the subject (cf. Bencivenga 1987). It is only through a spatio-temporal allocation that the object is knowable in the strict sense and the relative representations become Erkenntnis. As mentioned above, as far as the empirical apperception features are concerned, the activity of thinking (I exist thinking) only manifests itself in time, not space: It follows that the representation I, i.e., the designation of the object of inner sense, cannot turn into knowledge because, paradoxically, this would require an intuition in space; that is, in the form of sensibility wherein the only representations refer to that which is represented as something different from the subject. Thus, the subject cannot determine itself as an object within inner sense: By means of outer sense (a property of our mind) we represent to ourselves objects as outside us, and all as in space. In space their form, magnitude, and relation to one another is determined, or determinable. Inner sense, by means of which the mind intuits itself, or its inner states, gives, to be sure, no intuition of the soul itself, as an object; yet it is still a determinate form, under which the intuition of its inner state is alone possible, so that everything that belongs to the inner determinations is represented in relations of time. (A22–3/B37) For this reason, from the empirical dimension, I think “expresses an indeterminate empirical intuition, i.e., a perception” (B423 n.). As such, it is an

The Problem of Self-identification  97 empirical intuition because the empirical–existential proposition I think/I exist relies on a sensation that does in fact belong to sensibility and reveals itself only in time.6 In addition, it is indeterminate due to the lack of space, i.e., the form in and through which objects manifest themselves and can be determined (cf. supra, Chapter 2). In this context, it is clear—although unexpected—that the I of empirical apperception presents a Wittgensteinian use of the I as subject in some self-attributions. Firstly, the question concerns the judgment, and the Prolegomena contains the famous distinction between judgments of perception and judgments of experience: 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. But the former always demand, in addition to the representations of sensory intuition, special concepts originally generated in the understanding, which are precisely what make the judgment of experience objectively valid. (Prol 4: 298, 50) Judgments of perception only have subjective validity, while judgments of experience involve the principles of understanding that make empirical judgments objectively valid: A judgment of perception is only based on the relationship between a perception and a subject, whereas a judgment of experience indicates an object’s property (Prol 4: 298, 50). In other words, a judgment of perception “is merely a connection of perceptions within my mental state, without relation to the object” (Prol 4: 300, 52); it “is merely subjectively valid and . . . contains in itself no basis for necessary universal validity and, thereby, for a relation to an object” (Prol 4: 299 n., 51). Nevertheless, a judgment of perception—for instance, “in touching the stone I sense warmth”—can become a judgment of experience—“the stone is warm”— due to the involvement of the object and, consequently, of the categories: All of our judgments are at first mere judgments of perception; they hold only for us, i.e., for our subject, and only afterwards do we give them a new relation, namely to an object, and intend that the judgment should also be valid at all times for us and for everyone else; for if a judgment agrees with an object, then all judgments of the same object must also agree with one another, and hence the objective validity of a judgment of experience signifies nothing other than its necessary universal validity. (Prol 4: 298, 50) The point, however, is that some judgments of perception can never become judgments of experience because they are based solely on subjective

98  The Problem of Self-identification sensations or feelings (to readapt Kant’s examples: “I’m hot in this room”, “I am disgusted by wormwood”). A dual reference to the subject’s experience as well as consciousness is required in order for representations to fall within the knowable (cf. A 320/B376; Log 9: 33). Instead, the internal or subjective sensation (cf. Anth 7: 156), thus equated to the Gefühl (cf. KU 5: 206), lacks a potential reference to an object of reality—such as intuitions (in an immediate way) or concepts (in a mediate way). Consequently, it is a representation that is connected exclusively to the subject, and for that reason, it only manifests in time. The distinction between judgments of experience and judgments of perception should be intended as a counterpart to the Transcendental Deduction (Allison 2004, 179), and two accounts of judgment have been distinguished (cf. Allison 2004; Longuenesse 1998). The first type of judgment considers the act of judgment as the unification of distinct representation in a concept that is correlated with a unity in the consciousness. Since “a concept is never immediately related to an object, but is always related to some other representation of it (whether that be an intuition or itself already a concept)”, the judgment is the mediate cognition of an object, or the representation of a representation of it (A68/B93). Therefore, the judgment includes two concepts related both to each other and to the object being judged: In every judgment there is a concept that holds of many, and that among this many also comprehends a given representation, which is then related immediately to the object. So in the judgment, e.g., “All bodies are divisible”, the concept of the divisible is related to various other concepts; among these, however, it is here particularly related to the concept of body, and this in turn is related to certain appearances that come before us. These objects are therefore mediately represented by the concept of divisibility. All judgments are accordingly functions of unity among our representations, since instead of an immediate representation a higher one, which comprehends this and other representations under itself, is used for the cognition of the object, and many possible cognitions are thereby drawn together into one. 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. (A68–9/B93–4) The second account focuses on objectivity:7 In some passages (B § 18–19), Kant points out the distinction between an objective unity of self-consciousness, which presupposes the use of categories, and subjective unity, which is nothing but the product of reproductive imagination. For this reason, he criticises the logicians’ definition of judgment as “the representation of a relation between two concepts” because these do not specify what the relation amounts to: I have never been able to satisfy myself with the definition [Erklärung] that the logicians give of a judgement in general: it is, they say, the

The Problem of Self-identification  99 representation of a relation between two concepts. Without quarrelling here about what is mistaken in this explanation, that in any case it fits only categorical but not hypothetical and disjunctive judgements (which latter two do not contain a relation of concepts but of judgements themselves) . . ., I remark only that it is not here determined wherein this relation consists. If, however, I investigate more closely the relation of given cognitions in every judgment, and distinguish that relation, as something belonging to the understanding, from the relation in accordance with laws of the reproductive imagination (which has only subjective validity), then I find that a judgment is nothing other than the way to bring given cognitions to the objective unity of apperception. That is the aim of the copula is in them: to distinguish the objective unity of given representations from the subjective. (B142) Kant distinguishes a relationship of given cognitions to the objective unity of apperception from a relationship between representations that only present subjective validity. Only in the former case are there judgments that can be considered objectively valid—since the latter are only subjectively valid, they cannot be regarded as judgments at all. To understand these claims, it is important to specify the notion of objective validity with reference to several passages (A125; B137; A202/B247), according to which judgments are objectively valid if and only if they are about objects, i.e., ‘phenomena’ (Phänomena), ‘appearances’ (Erscheinungen), and objects of possible experience, regardless of whether they are true or false (cf. Prauss 1971; Vanzo 2012). In this way, prima facie only objectively valid judgments are truth-apt.8 From a Kantian perspective, and with regard to the forms of sensibility and understanding, a judgment such as “I’m hot” does not involve an identification component because it consists of perceptions only revealed in time, the nexus of which cannot involve any pure concept of understanding and, therefore, any objective unity of apperception—there is no object directed to the outer sense to be determined and hence identified. The heart of the matter changes in the self-ascriptions of properties manifesting themselves in time and space and that, as a result, refer to a spatio-temporally located subject/ body. If, as a human being, the thinking being is also an object of the outer sense (cf. B415), from a Kantian perspective, a judgment such as “I have a dirty hand” is a judgment of experience because it involves an outer object— the body—and the principles of pure understanding, which are based on the categories applied to the formal conditions of a possible intuition: The nexus of the representations is necessarily produced by the presence of the object that affects sensibility and which is objectively valid through the intervention of the intellectual dimension, thus determining the object. This is obviously related to the conditions that make a judgment of experience possible based on the relationships between forms of thoughts and forms of sensibility assumed as rules; in turn, according to the Copernican revolution, these are the universal laws without which nature in general,

100  The Problem of Self-identification as an object of sense, could not be thought (cf. KU 5: 183). It goes without saying that the transcendental account cannot tell whether the judgment “I have a dirty hand” is true: It provides the conditions for possibility required to produce an assertive Erkenntnissatz, which may be confirmed or denied with reference to the attribution of the predicate—the hand of the person who made the judgment may not be dirty—and the subject’s identification. If the judgment is produced according to the reflected image of a tangle of hands in a mirror, then the dirty hand might not belong to the person who has produced the judgment but to someone else. In this context, a judgment of experience expressing the self-attribution of a subject/body’s physical property involves a subject’s identification component—the subject can be mistaken as to whether she is the one who is attributing this particular property to itself. Prima facie, it is only through the objective unity of apperception that one can make a truth-evaluable judgment (i.e., a judgment of experience) that refers to objects and contains an identification component, making Erkenntnis possible; conversely, a judgment that lacks any reference to a possible object of experience is a judgment that involves no objective unity of apperception (i.e., a judgment of perception); hence, it cannot be said to be truthevaluable. As a result, such a judgment includes no identification component.9 In conclusion, following Kitcher (2000, 34), it is true that Kant and Wittgenstein start from what might appear to be different theses at first glance; thus, “it is an interpretive and philosophical mistake to try to force an alliance between what are, in fact, deeply opposed camps”. In fact, in order to point out the difference between Kant and Wittgenstein, we can refer to Carl’s approach (1997, 149): Kant’s distinction between the ‘I as subject’ and the ‘I as object’ is not concerned with Wittgenstein’s project of distinguishing different kinds of predicates to be ascribed to oneself, but with an epistemological perspective focusing on the distinction between spontaneity and receptivity, seen as conditions of all possible knowledge, in order to provide an account that incorporates the ‘I as subject’ of apperception into the foundation of the formal conditions of knowledge. However, it is also true that, although originating from different philosophical outlooks, they both arrive at similar conclusions: The I used as subject concerns the self-attributions of mental properties and involves no instance of identification, while the I used as object concerns the selfattributions of physical properties, and the subject’s identification will be provided. As seen previously, the question is further described from the transcendental and the empirical angles, and can be connected to Carl’s (1997, 157) distinction between two classes of self-ascriptions: self-ascriptions that regard the ‘I as passive’, as related to receptivity, and self-ascriptions that regard the ‘I as active’, as related to spontaneity: (1) Being necessarily involved in the making of any judgment, the ‘I as the subject’ of thinking, which means pure apperception, is the condition of

The Problem of Self-identification  101 possibility for the Wittgensteinian uses of the I as subject and the I as object in judgments expressing self-ascriptions of mental and physical properties; the representation I, which means the concept of the transcendental subject, cannot be used to identify or misidentify the thinking entity because I think (qua Representation) is the analytical unity of apperception, the simple and empty representation produced by the spontaneity of Thinking. It follows that the lack of an identification component regarding the subject in the forming of the judgment I think cannot be in question. (2) With regard to the I as passive in empirical apperception, there are two possibilities: (2.1) if the subject reveals itself only in time, selfattributions only concern mental properties; thus, there is no question of identification—the I is used as the subject in a judgment of perception, and (2.2) if the subject reveals itself in time and space, the question of the identification does arise as there is an explicit self-attribution of bodily, physical properties relative to I, used as the object in a judgment of experience. This Kantian reading is also obtained through a reflection on the sources of epistemic sensibility and mediation of the forms of time and space; also, for this reason, it should be included in a very different theoretical framework from that of Wittgenstein’s approach. Furthermore, such an articulation between two planes as described by Kant will also be proposed in the terms of de se thoughts in the final chapter.

Notes 1. At the same time, Evans claims that self-locating beliefs play a very important role not only in the representation of objects as independent entities in the world, but also in terms of one’s representation of oneself and with regard to one’s selfattribution of any physical or mental property. One’s capacity to locate oneself in space in relation to other individuals so as to be distinguished from them is nothing more than the instantiation of the Generality Constraint, which Evans regards as a necessary condition to form thoughts about individuals: “We cannot avoid thinking of a thought about an individual object x, to the effect that it is F, as the exercise of two separable capacities; one being the capacity to think of x, which could be equally exercised in thoughts about x to the effect that it is G or H; and the other being a conception of what it is to be F, which could be equally exercised in thoughts about other individuals, to the effect that they are F” (1982, 75). The possibility of forming I-thoughts and even I-thoughts based on the self-ascription of mental predicates (e.g., beliefs or perceptions) is grounded in the capacity to think of oneself as something located in space. 2. Kant maintains that the manifold of inner sense consists precisely of the representations of outer sense: “The representations of outer sense make up the proper material with which we occupy our mind” (B67). 3. “All inner experience is (has) a judgment in which the predicate is empirical and the subject is I. Independently of experience, therefore, there remains merely the I for rational psychology; for the I is substratum of all empirical judgments” (Refl. n. 5453, 18: 186).

102  The Problem of Self-identification 4. Taking the cue from Allison (2004, 298), who maintains that “one’s body functions as the enduring object, with reference to which one’s existence is determined in time” in the framework of the Refutation of idealism, several scholars, such as Cassam (1993) and Hanna (2000), have questioned the notion of “embodied subject” in Kant’s reflection. Capozzi’s (2007) historical–critical reconstruction clarifies a few aspects of the issue, as will be addressed in Chapter 5. 5. According to some scholars (Washburn 1970; Allison 2004), two distinct senses of self-affection are to be specified. The first sense is related to the transcendental synthesis, the condition of all experience, whereas the other is related to the empirical synthesis of apprehension and is only regarded as the condition for inner experience, the topic of the previous discussion on inner sense. As a result, there arise two applications, the first being the transcendental synthesis of the imagination through which the pure forms of understanding synthesise the objects of the intuition, and the second being linked to the phenomenon of attention—introduced in a footnote in § 25 and touched upon previously— whereby self-affection is required for the constitution of inner experience. In the latter case, “the point is that in attending to its representations, the mind makes them into objects represented. . . . As a second-order, reflective act, this presupposes a prior outer experience. . . . as a ‘second application’, this act involves an active seeking out by the mind of the representations it endeavours to make into objects of inner sense” (Allison 2004, 284). In this way, Allison focuses on the epistemic change based on a reconceptualisation; due to the initial conceptualisation of transcendental synthesis, the given representations can be referred to an object and, due to the reconceptualisation of the second application, they can become objects of inner sense themselves: “Kant’s claim that self-knowledge requires self-affection boils down to the claim that the mind must reconceptualize its representations in order to grasp them as objects” (Allison 2004, 284). Accordingly, in this second sense of self-affection linked to inner experience, the given contents can be represented as objects and constitute inner experience. 6. As mentioned previously, in another passage, Kant (Prol 4: 334 n., 86) reaffirms that I, as a representation of apperception, is not a concept: “It is nothing more than a feeling of an existence without the least concept, and is only a representation of that to which all thinking stands in relation (relatione accidentis)” (cf. infra, Chapter 5). 7. For Allison (2004, 84), it is possible to see the two accounts “as concerned with different aspects of a single conception that is considered from two points of view. To judge is both to unify representations by combining them in a concept (producing an analytic unity) and to relate these same representations to an object in a manner that purports to be valid with respect to the object”. 8. The expressions “objective validity” and “objectively valid” not only refer to judgments but also to concepts (Bxxvi n., A239/B298) and intuitions of space and time (A28/B44; A156/B195). In fact, some judgments with objective validity or that are objectively valid are not intended by Kant as truth-apt, but as being ‘objective truth’ or as ‘objectively true’. In this narrow sense, a judgment is objectively valid if it concerns objects and it is true (A125; Prol 4:298, 300). Thus, only true judgments about objects can properly be called judgments. 9. Needless to say, there is a huge debate among scholars concerning judgments lacking a truth value: Analytic judgments, judgments about items of which humans cannot have experience, judgments of perception and non-assertoric judgments are the main candidates, even though Kant seems to ascribe truth values to those kinds of judgments lacking either objective validity or any reference to any possible object of experience in some passages. Nevertheless, there is some agreement on the fact that judgments of perception are not truth-apt (Prauss 1969; Savile 1974; Wettstein 1980; Heckmann 1981; Schulz 1993; Hanna 2000; Vanzo 2012).

4 The Simple Representation I and the Transcendental Designation

At this point, three theses regarding the three questions at hand (a semantic, an epistemic, and a metaphysical one) have been developed: Thinking, the being itself, is spontaneity, which is not the object of knowledge. The subject can think of itself only as an object in general through the concept of the transcendental subject, i.e., the indeterminate thought of a thinking being with an existence in itself. The representation I think is the representational vehicle for the concept of the ‘transcendental subject’ and, as such, is a simple or empty representation. In the previous chapter, the epistemic thesis and a few points relating to a semantic thesis were addressed in order to examine the problems of selfidentification and of the reference of I: The Kantian reflections on I think and empirical apperception have been explored to explain the different uses of I as subject and I as object, and to investigate the problem of the identification of the subjective dimension in greater depth. On the other hand, the so-termed No-ownership Reading, whereby the I in I think has no reference whatsoever has to be rejected: I think (qua Representation) plainly designates the activity of thinking transcendentally. It is now time to address the issue of the nature of I and of the transcendental designation: To begin with, what exactly does Kant mean when he states that I is a simple and empty representation? Secondly, and from a strictly semantic viewpoint, can the features of the representation I and the correlative ‘transcendental designation’ explain the indexical nature of the I? Thirdly, do the Kantian reflections on indexicality anticipate any of the semantic elements or the spirit of the direct reference theory, if nothing else? Some sort of contiguity between the Kantian approach to I think and the contemporary direct reference theory concerning the semantic device of I has been suggested in the debate; this issue will be introduced in the next paragraph. In fact, the direct reference theory has also been applied to the Kantian approach to the semantics of natural kind terms; in order to rule out any proximity to the direct reference theory concerning these specific semantic issues, the present chapter will focus on how Kant treats indexicality. In recent years, non-conceptual content theorists have taken Kant as a reference point on account of his notion of intuition, and some Kantian

104  The Simple Representation I scholars in the current debate regard sensible intuition as an indexical representation (cf. the second paragraph). This chapter aims to explore a number of complementary issues intertwined with the notion of non-conceptual content in order to understand Kant’s treatment of indexicality. Of these, the first is solely concerned with the role of intuition as an indexical representation (cf. the third paragraph), whereas the second pertains to the presence of some epistemic features that will be discussed in the next chapter based on the distinction between knowledge by acquaintance and knowledge by description. The designation involved in the relationship between word, concept, and intuition will then be discussed, with particular regard to the kind of designation involved in natural kind terms (cf. the fourth paragraph). In so doing, the features that Kant assigns to the different representations and to the correlative designation will be compared to the representational features of I and its correlative transcendental designation: Undoubtedly, as transcendental designation displays utterly unique designation features, a simple or empty representation is a representational unicum among the kinds of representations considered by Kant. In this way, it will be possible to identify the peculiarities of the representation I in order to grasp the role of transcendental designation (cf. the last paragraph). With the designation of I, just as with the designation of natural kind terms, any application of the direct reference theory aimed at explaining the Kantian approach to indexicality will be rejected. Furthermore, as will be presented in the last paragraph, I is not regarded as an indexical representation at all.

I Think and the Direct Reference Theory Some scholars have developed Kant’s views on the semantic device of I think so as to appraise their analogies to contemporary claims about the direct reference of I. According to the features of the direct reference theory highlighted in Chapter 1, we are reminded of two theses concerning the use of the term/concept I: essential indexicality and lack of identification. (1) The term/concept I employed in a self-conscious or I*-thought is essentially indexical. More specifically, I is a singular term/concept governed by the token-reflexive rule, whereby every token of “I” refers to the subject that has produced or used it, either mentally or linguistically; with the information available in context, and once all evaluation circumstances are established, this rule is prima facie sufficient to determine its reference. Thus, moving from the contextual factors, the reference relationship between I and the subject who has produced it is determined in a direct way. By virtue of the token-reflexive rule, the fact that a subject produces the relevant token “I” makes her the very referent of that token; nonetheless, the process of indexical production does not occur within the propositional or thought content. For instance, when Nicola thinks the thought “I am happy”, the fact that he is the subject

The Simple Representation I  105 who has produced the token I determines that the I refers to him; however, it is the subject, i.e., Nicola, not the fact that he has produced the token, which enters the content of that thought. As a result, it can be argued that reference is direct: The singular term I is directly referential. (2) The reference of the I is determined without the subject observing or identifying herself as the subject who possesses a specific property: In other words, without the possibility of misidentifying or making thoughts liable to error through misidentification relative to the concept I, as discussed in the first chapter according to Shoemaker, and as explained from a Kantian perspective in the preceding chapter. It is a fact that a subject produces an occurrence of “I” that guarantees that this “I” will refer to that subject, not to the subject’s observation and identification of herself. Indexical information about oneself based on the use of I cannot be reduced to non-indexical information; for this reason, indexicality is essential. In other words, token-reflexive expressions such as the term/concept “I” are essential indexicals, as they cannot be eliminated or replaced by names, descriptions, or demonstratives without losing the content expressed by the sentences/thoughts that contain them: To refer to (to think of) oneself qua oneself, the subject has to use the essential indexical I. ‘Essentiality’ lies in the fact that, although the use of I is not based on any cognitive mediation, it is indispensable in order to form I-thoughts; this entails that an identifying description is a neither a necessary nor sufficient condition for the subject to refer to herself (cf. supra, Chapter 1). In the previous chapter, a few self-references without identification features were developed in Kantian terms; as mentioned, a few peculiarities of the self-referential apparatus involved in transcendental apperception due to the fact that the subject is designated only transcendentally through the I attached to thoughts emerged, without noting or cognising its least property. The act of reference performed by the subject to refer to itself entails no mediation of knowing, as it involves no identification through properties ascribable to the subject herself. The crucial point is to ascertain whether there is any clue about the idea of the direct reference of I in the Kantian approach. It is noteworthy that, according to some contemporary perspectives, the phenomenon of the lack of identification in self-reference entails that indexical reference to the self is both essential and grounded in direct reference because the act of self-reference is independent of non-indexical identification. As mentioned previously, Kant introduces the notion of ‘transcendental designation’ by pointing out that I is a simple representation: No manifold is given (B135). In A355 of the recently mentioned first-edition attack on the second Paralogism, Kant crucially claims that, by attaching I to its thoughts, the subject is only designated transcendentally, without noting any quality in it: The entirely empty of content I may be applied to every thinking

106  The Simple Representation I subject without knowing or identifying anything about it, either directly or by inference. The act of reference that produces the awareness of oneself as a subject is non-ascriptive—it displays no properties of oneself. The nonascriptive feature of the self-conscious subject’s self-reference is reasserted in B155 in the second edition, where Kant contends that, from a representational perspective, even if the subject is the same, the ‘I that thinks’ is distinct from ‘the I that intuits’. The subject is given to itself beyond what is given in intuition. In B422, the subject of the categories is considered to be unable to acquire a concept of itself as an object via the categories: In order to think of them, pure self-consciousness itself—namely, what must be explained—must be assumed. Moving on from this, Brook (2001) suggests that, in some respects, Kant is aware of the fact that the use of indexicals such as ‘I’ is essential for certain purposes because awareness of oneself as a ‘subject of the categories’ relies on an act of ascription-free transcendental designation: In order to apply the categories to oneself, i.e., in order to make ‘any judgment upon’ oneself or know oneself as an object, one must already and independently be aware of oneself as subject, i.e., as oneself. But this is nothing less than the core of the idea of the essential indexical. . . . Kant seems to have been aware of two features of reference to self that Shoemaker views as distinctive: 1. Kant was clearly aware of what Shoemaker calls reference to self without identification; in his jargon, we designate the subject “transcendentally, without noting in it any properties whatsoever” (A355); and, 2. There are indications that Kant was also aware of the idea of the essential indexical. In his terms, awareness of properties as properties of oneself presupposes awareness of oneself as subject, as oneself. (Brook 2001, 15) It follows that, if identification-free self-reference appears to imply the essentiality of those first-person indexicals involved in judgments, the pure self-consciousness resulting from it captures the subject’s ability to be selfconscious as a subject (Brook 2001, 1994). Howell (2000) is a scholar who has possibly explored this specific issue in the closest detail in recent years. Several authors—from Henrich (1976) and Frank (2004) to Ameriks (1997) and Sturma (1985)—have dealt expressly with the issue of the lack of identification, as has been touched upon previously; however, Howell (2000, 145–6) contends that they have failed to recognise that much of Kant’s account of ‘I think’ is related to his semantic account of the I as addressed by contemporary direct reference theorists, whose perspective provides the best approach to the Kantian approach to self-awareness. Howell has discussed the issue at greater length than has Brook, maintaining that, at the level of representational thought, the Kantian approach to ‘I think’ is, in some respects, similar to the contemporary

The Simple Representation I  107 direct reference views on the linguistic reference of I: “Given its own intellectual virtues, the direct-reference framework thus offers the best current way of coming to grips with the distinctive (and I think correct) points in Kant’s account”. Rosefeldt (2017, 226) restates this same point: “Because of its lack of descriptive content and its ties to intuition, Kant’s characterisation of the representation ‘I’ seems very similar to our contemporary notion of a directly referential expression and a demonstrative in Kaplan’s sense”. Similarly, Longuenesse (2017, 31) and Peacocke (2014, 170) draw on the use of the mental token-reflexive rule in order to specify how the reference of “I” in “I think” is determined. According to the theory of direct reference, the major point highlighted by Howell concerns the way in which I refers to the subject: The subject’s self-reference via I establishes a directly referential relationship between the subject and herself by importing the entity she is—the subject herself—into the thought’s content, without involving any of the subject’s properties. In other words, the relationship of designation between I and the subject is direct, and the propertyless nature of this relationship introduces the thing designated—the self—into the content of the ‘I-thought’. In the same manner, the scholar (2000, 134) highlights that Kant introduces a special representation of the self, the I in I think, which is supposed to be simple precisely because the designation that is called “transcendental” ensures that the representation I “gets us, by itself, into the designation relation with a specific being, our self, although it does not thereby enable us to distinguish that being from other beings”. In this scenario, and in keeping with the mental token-reflexive rule (whereby every token of “I” refers to the subject who has produced it), as well as to reflexivity (a feature contained in the Kantian approach1), the I in I think is regarded as a mental indexical that represents the subject in a direct fashion; what is introduced into the content of the thoughts is none of the subject’s properties but its very self. Beyond that sheer entity, i.e., “the being itself” (B429), nothing else can be introduced or known about the subject given to itself. It follows that, in producing the act of thought, the subject produces the representation I, and since it does so by virtue of the token-reflexive rule, the subject can refer to itself in a directly referential manner, i.e., by establishing a propertyless designation relationship between the I in I think and the subject-producer. This account presents two specific consequences. Firstly, Howell does not expressly agree with the conceptualist view presented in a particularly unequivocal way in the Paralogisms, which is also supported in the previous chapters of this work, whereby the acts of the ‘I-thought’ cannot grasp the subject as an individuated object: As has mentioned previously, in these passages, I think is regarded as the completely undetermined concept of a thinking being in general. Expanding on this concept, Howell reinterprets I think as a thought that determines the unique being that thinks—whatever it turns out to be—in a de dicto fashion; given that, in this de dicto manner,

108  The Simple Representation I the concept of a subject of thoughts does not connect the subject to itself in any way, this conceptualist construal is ultimately rejected by Howell. Howell does not support a de re interpretation, either. Certainly, when a direct reference theory is used to explain the designation of I think, a de re reading should be examined in depth with regard to the issue of the relationship of (self-)acquaintance. For example, in his paper The First Person, Kripke (2011, 305) states that “it is a rule of the common language that each of us fixes the reference of ‘I’ by the description ‘the subject’ ”. As Peacocke (2014, 84) points out, it is clear that, for Kripke, the grasp of the first-person concept does not rely on a thinker’s use of some definite description (‘the thinker’); what matters is the thinker’s position in a relationship of acquaintance to herself produced by the fact that she is the subject of certain mental events and states: “Each one of us can fix the reference of the word ‘I’ by means of acquaintance with oneself, self-acquaintance” (Kripke 2011, 302). In this way, “self-acquaintance is more fundamental than anything purely linguistic, and is the basis of our use of first person locutions” (2011, 320). The point is to explain what kind of relationship the acquaintance is in order to make the de se content available in the thought. Lewis (1979) considers identity to be a relationship of acquaintance par excellence; thus, de se beliefs are assessed as de re beliefs based on a causal dependence. Peacocke offers quite a different perspective concerning the first-person concept: It seems to be the case that the de se way of thinking is always distinct from any way of thinking grounded in acquaintance relations that are a matter of causal dependence. As is famously the case, the thought I am identical with that F, where that F is either a perceptual demonstrative, or a memory demonstrative, or a recognitional notion or concept, seems always to be potentially informative, provided that the demonstrative or recognitional notion or concept does not embed the first person itself. The relations in virtue of which the first person has the reference it does are distinct from the causal-dependence relations that ground these other demonstrative and recognitional notions. (Peacocke 2014, 102) Even though Howell does not develop the issue in this way, and a comprehensive evaluation of Peacocke’s perspective is not possible in this context, it is worth pointing out the possibility of considering a constitutive variety of self-acquaintance made available by consciousness, which is different from the de re modes of acquaintance explicated in terms of causal dependence on the model of perception, establishing a genuinely indexical way of thinking of a subject. It seems that this is Howell’s view, as he does not sustain a de re account explicated in terms of causal dependence on the model of perception: According to this, and based on a relationship of acquaintance, I is a mental,

The Simple Representation I  109 indexical representation that represents its referent through a relationship between the occurrence of a thought and the subject. However, the representation I is not an intuition, i.e., the kind of representation that presents a semantic device which, under certain perceptual conditions, establishes an indexical relationship between the representation and the object designated, as will be mentioned in the following paragraphs. It is Howell himself (2000, 138) who stresses this point, maintaining that the ‘I think act’ cannot possibly give the subject “the de re thought-awareness of a self that the producer grasps in some mysterious fashion as a definite but wholly propertyless entity, a bare particular”. In the last chapter, the distinction between de re and de se thoughts will be thoroughly analysed in order to provide an outline of the results about I-thoughts achieved thus far in the Kantian approach. This interpretation should be extended to counterbalance the different and, at first glance, incompatible views concerning this issue. In any event, it seems impossible to follow a Kantian reading of I think according to the direct reference theory for two reasons—one general and the other particular—which will be developed in this chapter. Firstly, as has been mentioned in previous chapters and will also be seen later when recalling and analysing the passages from the Paralogisms and others (cf. infra, the last paragraph), Kant firmly rejects any possible relationship between the I and the self in a direct way, enabling the subject to get into the content of its thought. Secondly, and with regard to the general considerations on indexicality that will be expounded in the following paragraphs, not only is the representation I not an intuition (the main representation connected to an indexical component), but the phenomenon of indexicality addressed by Kant cannot be investigated in terms of the direct reference theory because Kant’s theory is essentially bound to a conceptual dimension; this will also become apparent when the Kantian approach to the semantics of natural kind terms is discussed. In the next two paragraphs, Kant’s investigation of indexicality in general will be examined in depth, while the role of designation and indexicality in the semantics of natural kind terms will be addressed in the fourth paragraph. In both cases, the direct reference theory will be dismissed on the grounds of textual evidence. Finally, in the last paragraph, it will be possible to understand the peculiarities of the simple and empty representation I and the role played by transcendental designation in order to verify whether these are compatible and consistent with the frame lying in the three (metaphysical, epistemic, and semantic) theses developed previously.

Kant Between Conceptualism and Non-Conceptualism: Concepts and Intuitions As is well known, Kant is regarded as the philosopher who adheres to the conceptualist position—the perspective maintaining the impossibility of

110  The Simple Representation I attaining knowledge, experience, or the perception of reality in the absence of conceptual capacities. In the last decade, Kant has become a key frame of reference even for the theorists of non-conceptual content through a comprehensive reflection on the sensible dimension and the relevant notion of intuition in particular. One of the most influential contemporary theorists of conceptualism is McDowell. In Mind and World, McDowell addresses several parts of the Kantian approach to assert a Kantian conceptualist theory of experience and to attack any other approach based on non-conceptual content. According to recent views within the Kantian debate on non-conceptual content, Sellars and McDowell have not acknowledged Kant’s fundamental contribution to the non-conceptualist theory, nor has the contemporary debate adequately emphasised its own debt to Kant, to the extent that Hanna (2006, 90–1) has claimed that “Kant’s theory of intuition is the hidden historical origin of both sides of the debate between conceptualists and non-conceptualists”. The debate on non-conceptual content is divided among several, not entirely consistent positions. Although it is unquestionably difficult to find an agreed-upon definition in the current debate, Bermúdez’s general considerations (2012) can be taken as a starting point. Bermúdez argues that, if the content of a mental state of a (human or non-human) creature is what the mental state actually represents, according to the theory of nonconceptual mental content certain mental states represent reality even if their subject does not possess the necessary concepts to articulate their contents. More precisely, while conceptualism holds that the mental states of nonhuman creatures lack conceptual capacities, non-conceptualism regards the representational content as being determined not only (or not entirely) by conceptual abilities but also by non-conceptual capacities shared by infants and non-human creatures (cf. Evans 1982; Bermúdez 2003; Gunther 2003; Hanna 2008, 2011).2 Within the non-conceptualist approach, Speaks (2005, 360) distinguishes two theses. According to the first—holding an absolutely non-conceptual content, and endorsed by Evans (1982), Martin (1992), Peacocke (1992), and Heck (2000)—a state of mind has an absolutely non-conceptual content if and only if the type of content of the mental state is different from that of beliefs and thoughts. The second argument concerns the relationship between subject and content and asserts a relatively non-conceptual content: A subject’s mental state at time t has a relatively non-conceptual content if, and only if, the content of the mental state in question includes contents not conceptually grasped or held by the subject at time t.3 Hanna (2008) introduces several Kantian non-conceptualist arguments, among which is the well-known Two Hands Argument. Expounded as early as in the pre-critical period, the argument states that the incongruent counterparts do not have any descriptive or conceptual difference and can be distinguished only from a perceptual standpoint. Therefore, while Hanna supports an absolutely non-conceptual content, Kantian non-conceptualists

The Simple Representation I  111 have articulated several argumentative strategies in the transcendental system to identify a comparatively non-conceptual content in the intuition (see Allais 2009). Within this general framework, the difference between non-conceptual cognition and its content and conceptual cognition and its content seems to reflect the Kantian distinction between concepts and intuitions: The intuitive representations are assumed to possess certain semantic features pertaining to the indexical dimension, together with some epistemic features articulated according to the distinction between knowledge by acquaintance and knowledge by description (see Hanna 2008; de Sá Pereira 2013). The latter will be the focus of the next chapter, after the following framing of the types of representations in the transcendental system. In the three classic passages presented below, Kant describes the differences in the main types of representations: All cognitions, that is, all representations related with consciousness to an object, are either intuitions or concepts. An intuition is a singular representation (repraesentatio singularis), a concept a universal (repraesentatio per notas communes) or reflected representation (repraesentatio discursiva). (Log 9: 91, 589) In whatever way and through whatever means a cognition may relate [beziehen] to objects, that through which it relates immediately [unmittelbar bezieht] to them, and at which all thought as a means is directed as an end [als Mittel abzweckt], is intuition. This, however, takes place only insofar as the object is given to us; but this in turn, is possible only if it affects [affiziere] the mind in a certain way. The capacity (receptivity) to acquire representations through the way in which we are affected by objects is called sensibility. Objects are therefore given [gegeben] to us by means of sensibility, and it alone affords us intuitions; but they are thought [gedacht] through the understanding, and from it arise concepts [Begriffe]. But all thought, whether straightaway (directe) or through a detour (indirecte), must ultimately be related to intuitions, thus, in our case, to sensibility, since there is no other way in which objects can be given to us. (A19/B33) The genus is representation in general (repraesentatio). Under it stands the representation with consciousness (perceptio). A perception that refers to the subject as a modification of its state is a sensation (sensatio); an objective perception is a cognition (cognitio). The latter is either an intuition or a concept (intuitus vel conceptus). The former is immediately related to the object and is singular; the latter is mediate, by means of a mark, which can be common to several things. (A320/B376)

112  The Simple Representation I Starting with the mathematical–philosophical debate instigated by Hintikka (1967, 1969, 1972) and Parsons (1969, 1984, 2012), certain conditions to be fulfilled in order for a representation to be an intuition have been identified. In the selected passages above, Kant argues that intuition is a singular representation: In other words, he introduces (1) the singularity condition based on the type of denotation involved, where the intuition is a singular representation denoting an individual object, as opposed to the concept, which relates to different objects falling under it in view of the presence of the very property the concept represents. In the second and third passages, in which Kant adds that intuition is “immediately related to the object”, he virtually spells out (2) the immediacy condition, which concerns the type of relationship between the representation and its denotation, as opposed to concepts referring to the object through the mediation of the conceptual features, or marks, composing the very concept’s intention. Although the latter condition has been at the centre of a harsh dispute,4 several commentators are inclined to link the immediacy condition to the referential directness of intuition. These scholars contend that intuitive representations not only bring about the immediate cognition of objects, but also do not identify them with mediation of conceptual or descriptive contents. In the second passage, Kant (Prol 4: 282, 33) also introduces (3) the object dependence condition, i.e., the condition of immediacy implying a criterion of object-dependency whereby intuitions are produced provided that the objects are given: “An intuition is a representation of the sort which would immediately depend on the presence of an object”. Intuition is assumed to be a relational form of cognition, its objective validity relying on the existence of the object. This essentially demonstrative relationship implies (4) the relatedness to sensibility condition: The intuitive representation has a direct reference to the sensible object because this can only be given through the faculty of sensibility, i.e., the faculty to receive sensible representations based on the way in which the subject is affected by the objects: “It comes along with our nature that intuition can never be other than sensible, i.e., that it contains only the way in which we are affected by objects” (A51/B75) (cf. Howell 1973, 209). This amounts to saying that, for Kant, it is logically possible for a non-human mind—e.g., the divine mind—to have intellectual intuition (cf. B72). Hanna (2006, 102) points to a further (5) priority-to-thought condition, whereby the intuition is a representation that can be given before thinking. The priority of intuition is both cognitive and semantic, its independence and objective validity being marked regardless of the conceptual dimension. Concepts and intuitions can be both pure and empirical; in particular, an empirical intuition is the product of both ‘sensation’ (Empfindung)—the effect produced by an object on the capacity of sensible representation, i.e., the matter of sensibility—and pure intuitions, namely time and space, i.e., the a priori pure forms according to which the sensible matter is organised.

The Simple Representation I  113 At the empirical level, it seems possible to summarise these conditions by arguing that an object is given through sensibility only, producing intuitions by virtue of a relationship with the object. This relationship depends on the sensations composing the matter of intuition; they are articulated before the intervention of thought through the pure forms of sensibility—time and space—producing a singular representation that is linked immediately with the object. If concepts and intuitions are two distinct types of objective representations, from an epistemic point of view they are both necessarily involved in the judgment for the determination of objective knowledge: According to the well-known adage, “Thoughts without content are empty, intuitions without concepts are blind”. The togetherness principle, whereby knowledge is produced only through the joint intervention of concepts and intuitions (A50/B74), has been supported by McDowell’s Kant-inspired conceptualist position (cf. supra, Chapter 1). In contrast to this view, Hanna argues that concepts and intuitions are cognitively and semantically interdependent only in terms of the constitution of objectively valid judgments. Beyond this specific epistemic dimension consisting of empirically meaningful judgments, empty concepts or blind intuitions are certainly possible. Hanna distinguishes a direct relationship with the perceptual dimension in the Kantian approach: Intuitions are representations cognitively and semantically independent of the concepts with non-conceptual cognitive contents (Hanna 2006, 100). Against this background, the togetherness principle is consistent with Kantian nonconceptualism due to the epistemic and metaphysical independence of the intuitive representation.5 Based on these considerations, the two abovementioned theoretical issues, which are closely linked to the notion of nonconceptual content, emerge: The first issue is semantic and will be addressed in the next paragraph, whereas the second is more closely connected to epistemic considerations, which will be analysed in the following chapter.

Intuition as an Indexical Representation Some of the scholars involved in the current debate on Kantian nonconceptual content regard the sensible intuition as an indexical representation; for this reason, they are prone to link it to directly referential singular terms (Hanna 2000, 21). This approach has already been supported within the Kantian philosophical–mathematical debate: Regardless of the different approaches, both Hintikka and Parsons correlate the distinction between concepts and intuitions to that between singular and general terms. Hanna (2001, 195–6) focuses on the features of the immediacy criterion to highlight that only intuitions can refer immediately to an object; concepts, on the other hand, are provided with a mediated reference based on the conceptual marks composing their intention and determining the set of objects that fall under them. Hanna redefines the nature of the referential

114  The Simple Representation I mechanism of the intuition in contemporary terms by referring explicitly to the theory of direct reference as developed by Kaplan, Kripke, and Putnam: Just as particular singular terms (including demonstratives and indexicals proper) are directly referential, denoting their object directly—i.e., without a mediation or satisfaction of descriptive conditions—sensible intuitions are indexical, singular representations relating directly to their objects, that is, without the mediation of a conceptual content: So the Kantian distinction between conceptual (mediate) reference and intuitive (immediate) reference is most accurately construed as the difference between, on the one hand, indirect or description-determined reference to an object, and, on the other, direct or non-description-determined reference to an object. More plainly put, intuitional reference is direct reference. (Hanna 2001, 197) While it is possible to identify the same referential device, it does not seem as possible to maintain the existence of a link between the role of the intuition and that of singular terms (Thompson 1972; Capozzi forthcoming). In this regard, at least two fundamental theoretical assumptions are to be isolated in the transcendental system: The logical form of singular judgments and the dismissal of the lowest species. The Logical Form of Singular Judgments Firstly, the logical form of the judgment does not involve a singular representation—according to Kant, judgment constituents are always concepts, namely general representations. More specifically, these consist of cases of singular judgments using (1) proper names as subjects, as in Kant’s examples “Caius is mortal” (Log 9: 103, 599; A322/B378) and “Adam was fallible” (Refl 3080, 16: 647), (2) demonstratives with the form “This F is G”; “this house is plastered in this way or that” (V-Lo/Wiener 24: 909, 352) and “This world is the best” (Refl 3173, 16: 695), and (3) definite descriptions expressing a concept’s singular use have to be contextualised within the traditional logical frame accepted by Kant. From the perspective of logical form, singular judgments are to be treated as universal judgments: The logicians rightly say that in the use of judgments in syllogisms singular judgments can be treated like universal ones. For just because they have no domain at all, their predicate is not merely related to some of what is contained under the concept of the subject while being excluded from another part of it. The predicate therefore holds of that concept without exception, just as if the latter were a generally valid concept with a domain with the predicate applying to the whole of what is signified. (A71/B96)

The Simple Representation I  115 This entails that the subject of a singular judgment can be used in an equivalently universal judgment, as in Kant’s example “God is without error; everything which is God is without error” (Refl 3080, 16: 647). Certainly, from an epistemic point of view—but not from that of logical validity—if a singular judgment (judicium singular) is compared to a universal one (judicia communia) with regard to quantity, the two will appear quite different. In fact, Kant states that the former relates to the latter as unity relates to infinity, implying a singular use of a concept, as will be discussed shortly. The Dismissal of the Lowest Species The second assumption concerns the anti-Aristotelian and anti-Leibnizian dismissal of the lowest species and singular concepts (cf. Gensini 2000). Concepts are considered predicates of possible judgments via their application to other concepts; no concept has a singular representative function that thoroughly determines an individual or particular (cf. A656/B684 for explicit support). Following Capozzi, it is important to remember that, in Kant’s view, every concept has an intention (Inhalt) and an extension (Umfang).6 The inverse connection between intention and extension (“Quantum cognitio ab una parte lucri facit, tantum ab altera mulctatur”—Refl 2893, 16: 564) sets forth a relation of subordination specifying the hierarchy among the concepts based on genus and species. The possibility to articulate the hierarchy of concepts depends on two complementary operations: Abstraction and determination. Abstraction is a bottom-up process from lower to higher concepts; as marks are subtracted, the complexity of intension decreases. The inverse operation is determination, which is similar to addition in that it practically adds marks and concepts to higher concepts so as to move down the hierarchy. The more marks abstraction subtracts from the concept’s intension, the higher we climb up the summum genus, i.e., the concept of something (Etwas) (cf. Log 9: 95, 593; V-Lo/Pölitz 24: 570), of a being (Wesen), or of a thing (Ding) (cf. V-Lo/Dohna 24: 754, 488). On the contrary, and according to the law of specification, there is no concept or species infima under which no other conceptual content can possibly stand. Determination is virtually unlimited: Every genus requires different species, and these subspecies, and since none of the latter once again is ever without a sphere, (a domain as a conceptus communis), reason demands in its entire extension that no species be regarded as in itself the lowest; for since each species is always a concept that contains within itself only what is common to different things, this concept cannot be thoroughly determined, hence it cannot be related to an individual, consequently, it must at every time contain other concepts, i.e., subspecies, under itself. This law of specification could be expressed thus: entium varietates non temere esse minuendas. (A655–6/B683–4)

116  The Simple Representation I Although each concept is a species contained in some higher genus, and this all the way up to the top of the summum genus hierarchy, which expresses maximum generality, every concept can always be further articulated by a more specific concept still. It follows that no concept can be not common, free of extension, and not subject to further determination: In the series of species and genera there is no lowest concept (conceptus infimus) or lowest species, under which no other would be contained, because such a one cannot possibly be determined. For even if we have a concept that we apply immediately to individuals, there can still be specific differences in regard to it, which we either do not note, or which we disregard. (Log 9: 97, 595) All this results in an anti-Leibnizian rejection of the principle of indiscernibles, which relies on the assumption that, in the absence of haecceity— i.e., in compliance with the dictates of the Discourse on Metaphysics, the complete notion of an individual that defines a set of attributes determining the metaphysical identity of the individual substance (monad)—one cannot state that two individuals with the same concept are the same. The Conceptualist Form of Singular Terms It is because the possibility of singular concepts is rejected that intuition—a singular representation by definition—seems the best possible candidate to occupy the subject role in singular judgments. The principle whereby intuitions cannot occupy the subject role in singular judgments holds, together with the denial of an infima species and singular or individual concepts. In fact, Kant allows for the possibility of a singular use of a concept in the subject role within singular judgments. While Kant (Log 9: 91, 589) states explicitly that concepts are representations in which the logical form is always general by definition, he also allows for three different uses of concepts, namely general, singular, and particular: “It is a mere tautology to speak of universal or common concepts—a mistake that is grounded in an incorrect division of concepts into universal, particular, and singular. Concepts themselves cannot be so divided, but only their use [Gebrauch]”. In this regard, one should refer to the Kantian view that no thought is nonlinguistic, as every name refers to the concept’s logical essence. This consists of marks being (1) necessary or constitutive in order to define the concept itself (V-Lo/Wiener 24: 838, 293); (2) primitive (cf. Letter to Reinhold, May 12, 1789, Br 11: 37); (3) inseparable and immutable, no other concepts being possible without them (V-Lo/Dohna 24: 727, 463); (4) numerically limited (A728/B756); (5) arbitrarily associated with the name attributed to the concept (Anth 7: 191, 84); and (6) subject to logical–linguistic analysis (V-Lo/Dohna 24: 728, 464). If the word “home” designates a common

The Simple Representation I  117 concept, then this can be used in three different ways based on the quantitative determination (universal, particular, or singular) of its extension class, i.e., the set of things that can be articulated in toto, in parte, or in individuo (V-Lo/Wiener 24: 908). The Wiener Logik explains the issue as follows: I can make use of a concept insofar as it is applied to many objects[;] then the concept is used as a repraesentatio communis, i.e., is used in abstracto, e.g., house. If I say of all houses, now, that they must have a roof, then this is the usus universalis. It is always the same concept, however, and is here used wholly universally. For having a roof holds for all houses. This use of the concept is concerned universally with all, then. But a particular use is concerned only with many. E.g., some houses must have a gate. Or I use the concept only for an individual thing. E.g., this house is plastered in this way or that. (V-Lo/Wiener 24: 908–9, 352) It follows that it is judgments, not concepts, which are divided into universal, particular, and singular, as only judgments can specify the use of concepts based on their quantitative determinations. In an epistemic context, adopting the singular judgments used by Kant (employed singularly as the subject of a judgment to represent a given object and to ascribe it a certain property), a concept must fulfil two conditions: The condition of existence, according to which the concept must represent an existing object in a space–time dimension that can only be met through the intervention of intuition, and the condition of uniqueness, whereby the concept must represent an object through the specific features that only that particular object possesses. These conditions further articulate the issue of identification, whereby the same intuition—that is, the same phenomenal object—can be identified on several occasions only with regard to a conceptual dimension under which the intuition falls (see Thompson 1972). At first glance, the emerging picture is descriptivistic. In view of the fact that it relates to its object in an immediate way, i.e., without the mediation of any conceptual content, intuition is a demonstrative, singular representation; nonetheless, while the intuitive dimension can only attest to the presence of something spatio-temporally located, it cannot identify or reidentify it as the same particular object. To take a trivial example, if proper names are the linguistic labels of intuitions, then we cannot apply or reapply the same name to a specific individual: “Names can be applied, reapplied, and misapplied; so can concepts, but not intuitions” (Thompson 1972, 91). Mutatis mutandis, the same argument can be used with regard to demonstratives; as Kant probably concurred with the Port Royalists’ claim that the pronoun “this” is equivalent to “this thing”, the demonstrative simply makes a concept (even the most general of all) singular. Despite the remarkable differences between names and demonstratives highlighted by Capozzi (forthcoming),7 Kant’s use of singular terms seems essentially bound to the

118  The Simple Representation I conceptual component—the intuitive dimension alone cannot even identify something spatio-temporally located as an object.

The Role of Designation in Natural Kind Terms After discussing the main kinds of representations (i.e., concept and intuition) and how Kant accounts for indexicality, it is now necessary to focus on the particular designation involved in the association of word, empirical concept, and intuition, with special attention to designation in natural kind terms. In this way, it will be possible to capture the difference between this type of designation and the designation of I that Kant calls ‘transcendental’ in the next paragraph. The direct reference theory has also been linked to empirical concepts and the relative designation involved in natural kind terms. Kripke and Putnam on the Theory of Direct Reference Scientific essentialism accounts for the philosophical framework wherein the theory of direct reference was articulated by Kripke and Putnam in the early 1970s in several papers and essays.8 I will take Kripke’s doctrine as the paradigm of scientific essentialism:9 Scientific investigation generally discovers characteristics of gold which are far better than the original set. For example, it turns out that a material object is (pure) gold if and only if the only element contained therein is that with atomic number 79. Here the ‘if and only if’ can be taken to be strict (necessary). In general, science attempts, by investigating basic structural traits, to find the nature, and thus the essence (in the philosophical sense) of the kind. (Kripke 1980, 138) Kripke and Putnam object to Frege’s semantic descriptivist approach, asserting that the reference of natural kind terms, as well as that of proper names, is determined on the basis of a descriptive content because the individual/substance satisfies the properties expressed by the descriptive contents associated with the terms. In particular, Putnam states that a natural kind term can be used notwithstanding the fact that the descriptive content of the term may not be satisfied by the substance in question: If a speaker does not know the difference between an elm and a beech, and yet she employs the corresponding terms, these will refer to two different trees even if the content ascribed to ‘elm’ and ‘beech’ is the same, making it impossible to distinguish them. Similarly, in Putnam’s well-known Twin-Earth thought experiment, in which he hypothesises a planet that is identical to Earth in every way except for water—the microstructure of which is different from H2O and is represented by the formula XYZ—an earthling and his or her twin will refer

The Simple Representation I  119 to two different substances every time he or she uses the term “water”, although the conceptual content or stereotype associated with the term is the same—they both perceive the same macro-properties (a clear, odourless, tasteless liquid, etc.). It is not the content/meaning in the two speakers’ minds that determines the reference: Natural kind terms, just like proper names, are directly referential—or have an indexical component, to employ Putnam’s terminology. Reference is fixed by a baptism, an act of intentional labelling that establishes a causal, punctual relationship between a single term and a substance, a relationship that remains unchanged in both name and substance based on the assumption of the indexical component fixing the paradigm of the substance’s extension. The relationship between name and substance—the causal chains of transmission in a linguistic community— must be socially inherited; in other words, the intention to refer to that which possesses the same essential properties as the substance of the initial baptism must remain unaltered in order for the two most extreme rings of the chain to be assumed. Once the introductory event is created by a baptism, it is nature, i.e., the metaphysical order, which determines a term’s reference. However, Putnam maintains that the changeover fixing the reference will occur through an intentional device that introduces and establishes the chain between a name and a substance, its final ring being the indexical act ‘this’, conventionally determined and socially inherited. In this framework, we can distinguish no fewer than three different theoretical dimensions. In particular, the first dimension is metaphysical: Natural kinds are identified by virtue of their essence or nature based on micro-properties or microstructures analysed by science, on which depend those macro-properties that are the objects of empirical observation (of course, the philosophical debate on this dependency relationship offers several metaphysical options with a crucial impact on the essentialist thesis). Secondly, there is an epistemic dimension: Structures and micro-properties that make up the substances of natural kinds are investigated by science. The resulting knowledge is just as a posteriori as is the knowledge of macroscopic properties, the objects of empirical observation; by contrast, knowledge of the macrophysical properties of natural kinds belongs to conceptual stereotypes of common sense, and a priori knowledge is specified by the conceptual or logical/linguistic analysis of the dictionary meanings of the terms that make up the natural kind stereotypes employed in a given society (Hanna 2006, 156). Lastly, we can also distinguish a semantic dimension: Natural kind terms are rigid designators, i.e., terms referring to the same referent in all possible worlds in which the referent exists, and are based on a direct reference mechanism.10 The semantic content of these terms does not consist of descriptions or conceptual contents because it is identified by its referents, which in turn are determined through a baptism that establishes a causal chain, socially inherited, between terms and substances.

120  The Simple Representation I If true, a statement of essential identity involving two co-referential, rigid designators—e.g., one relative to the natural kind and the other relative to what are regarded as its micro-properties, as in the statement “water is H2O”—expresses a necessary truth from a metaphysical point of view which is valid in all possible worlds, however epistemically a posteriori. On the other hand, to retrieve an example from Kant (“gold is a yellow metal”) that I will attempt to analyse in the next section, statements of identity involving a natural kind rigid designator and a term referring to one of the macro-properties of the substance in question express a truth that is metaphysically contingent and epistemically a posteriori: The macroproperties of the substance, such as the yellow colour of gold, do not express the essence of the natural kind and are used empirically to categorise substances from a linguistic/cognitive point of view. Kant and the A Priori Nature of the Judgment “Gold Is a Yellow Metal” On the basis of the identification of metaphysically necessary truths—a posteriori though they are, from an epistemic angle—the essentialist approach overtly polemicises with Kant. An example can be found in Putnam: Since Kant there has been a big split between philosophers who thought that all necessary truths are analytic and philosophers who thought that some necessary truths were synthetic a priori. But none of these philosophers thought that a (metaphysically) necessary truth could fail to be a priori: the Kantian tradition was as guilty as the empiricist tradition of equating metaphysical and epistemic necessity. In this sense Kripke’s challenge to received doctrine goes far beyond the usual empiricism/ Kantianism oscillation. (Putnam 1975, 233) Several commentators, however, have attempted to articulate Kant’s arguments against an essentialist perspective—cf. Kroon and Nola (1987), Anderson (1994), and Hanna (1998) (2006)—in an effort to frame Kant’s semantic approach within the theory of direct reference. As I will try to argue, although a number of theoretical points connected to this framework can certainly be discovered, Kant’s approach to natural kind terms is still bound to a nominalist/conceptualist position spelled out according to the dictates of transcendental idealism. Kant articulates a sharp division between epistemic and metaphysical order. In agreement with the principles of transcendental idealism and the separation between noumena and phenomena, the ultimate nature of substances— i.e., their internal foundation—is completely inaccessible from an epistemic angle. On one hand, the thesis of transcendental idealism is even more radical than Locke’s caesura: While Kant explicitly rejects Berkeley’s anti-realism in

The Simple Representation I  121 compliance with the noumena/phenomena distinction, he agrees with the Irish bishop’s dismissal of a subdivision between primary and secondary qualities, as also supported by Locke, eventually to reject a representational paradigm based on resemblances even with regard to primary qualities (Prol 4: 289). On the other hand, based on his transcendental idealism, Kant defends his own version of empirical realism (A370)—what Hanna (2006, 142) has re-labelled ‘manifest realism’—based on his adoption of the Newtonian principles of physics and his subsequent rebuttal of the corpuscular theory adopted by Locke (A265/B321). The metaphysically assumed existence of material reality conjugates with Newton’s principles, which state that reality amounts to a series of movements of material bodies containing perceptible properties and interacting with one another in compliance with the laws of physics in a macroscopic space–time dimension apprehended as a phenomenon. If, in Kant’s view, empirical concepts associated with nature only apply to objects of possible experience (Log, 9: 143–4), then the foundation of natural substances cannot be an object of knowledge because it lies outside the forms of possible experience established by transcendental investigation. In a letter to Reinhold, dated May 1789, Kant is particularly explicit on this matter: The real essence (the nature) of any object, that is, the primary inner ground of all that necessarily belongs to a given thing, this is impossible for man to discover in regard to any object. For example, extension and impenetrability constitute the whole logical essence of the concept of matter, that is, they are all that is necessarily and primitively contained in my, and every man’s, concept of matter. But to know the real essence of matter, the primary, inner, sufficient ground of all that necessarily belongs to matter, this far exceeds the capacity of human powers. We cannot discover the essence of water, of earth, or the essence of any other empirical objects. (Br 11: 36–7, 299–300) In this context, we should also add that Kant’s famous statement “gold is a yellow metal” is an analytic, a priori, necessary judgment: All analytic judgments rest entirely on the principle of contradiction and are by their nature a priori cognitions, whether the concepts that serve for their material be empirical or not. For since the predicate of an affirmative analytic judgment is already thought beforehand in the concept of the subject, it cannot be denied of that subject without contradiction; . . . For that reason all analytic propositions are still a priori judgments even if their concepts are empirical, as in: Gold is a yellow metal; for in order to know this, I need no further experience outside my concept of gold, which includes that this body is yellow and a metal; for this constitutes my very concept, and I did not have to do anything except analyze it, without looking beyond it to something else. (Prol 4: 267, 17)

122  The Simple Representation I Kripke (1980, 39) overtly rejects the analytical, a priori character attributed to the judgment. According to the above-mentioned dimensions composing the essentialist approach of the direct reference theory, and considering the term “gold” as a subject and, for brevity, only the term “yellow” (the reference of which is a phenomenal quality) as predicate, for a pure essentialist perspective such as Kripke’s, the statement can only be synthetic, contingent, and a posteriori. ‘Being yellow’ is a macro-property of the substance gold—which, to a certain extent, depends on one’s perceptual system—but it does not concern its essence. As opposed to the tenets of the essentialist approach, Kant’s articulation of the epistemic dimension precisely addresses those macro-properties of the substances because there can only be knowledge of the possible objects of experience based on the principles of transcendental idealism, any possibility of determining the inner nature of the substances themselves being precluded. In order to understand the reasons that led Kant to state the a priori nature of the judgment “gold is a yellow metal”, it might be worthwhile to recall the indissoluble link between thought and language and, more specifically, the characteristics of the empirical/conceptual and linguistic dimensions. Accordingly, these are employed to determine the semantics of the terms referring to the substances that surround us in nature by means of the plexus that binds concept, logical essence, and linguistic terms. However, this cannot be attained without specifying that it is possible—within a conceptualist frame—to identify two different dimensions in Kant’s semantic approach to natural kind terms, i.e., a logical/formal dimension, as it were, and a material dimension.11 The Relationship of Term, Concept, and Natural Substance With the formal/logical dimension as a starting point, every concept has a logical essence, namely a set of essential, necessary, immutable, and limited conceptual notes. Considering that the concept is associated with a linguistic term that expresses it, its nominal definition clarifies the link between logical essence and term. These considerations must be comprised within the context of general logic and its difference from transcendental logic, which regards the former as something that is not concerned with the content or matter of thought, but only with the form of representations and the way they become concepts. The point is well summarised in the section of the KrV that addresses the difference between the mathematical and the philosophical methods, starting from the well-known distinction between analytic and synthetic propositions: I could analyze my empirical concept of gold without thereby gaining anything more than being able to enumerate what I actually think by means of this word, which would certainly produce a logical improvement in my cognition, but no augmentation or supplementation of it.

The Simple Representation I  123 But I can take the matter that goes by this name and initiate perceptions of it, which will provide me with various synthetic though empirical propositions. (A721/B749) Conceptual analysis cannot increase knowledge from a material point of view because it merely expounds the notes contained in logical essence, which, in turn, are associated with the word-concept in question. It is precisely within the articulation of conceptual form that the statement “gold is a yellow metal” (seen as an analytical judgment expressing a priori knowledge) should be framed. It is sufficient to apply a linguistic and conceptual analysis to the bond linking word and concept to state that the conceptual note ‘yellow’ must be contained in the concept ‘gold’. Once the conceptual plot and its relative linguistic/conceptual links are established, they can be traced back a priori in order to find the notes that make up the logical essence of a concept; in other words, it is possible to identify the specific connection between the subject-concept and the concept-predicate contained in it that make up an analytic judgment. Empirical concepts can only contain conceptual notes relative to macrostructural properties that are objects of possible experience. Therefore, if the concept ‘gold’ contains the conceptual note ‘yellow’ to form an analytic judgment, then, in a similar way—and regardless of the fact that chemical knowledge was extremely limited in Kant’s time—the concept of gold does not contain the conceptual note ‘element having atomic number 79’ analytically. For these reasons, Kant considers the judgment “gold is the element having atomic number 79” to be synthetic and known a posteriori (Anderson 1994, 357; Hanna 1998, 501). If this is the conceptualist soul of Kant’s approach to the semantic theory of natural kind terms, a material dimension can be specified by investigating how conceptual plots are formed and used. The formal dimension of the conceptual nexus whereby we can determine the analytical nature of judgments is one thing, but the material dimension of transcendental logic, which can be referred to the concepts’ formation and application to intuitions according to synthetic judgments, e.g., in denomination performances, is quite another. In the passage from the KrV (A721/B749) on the synthetic judgments concerning matter, we can detect an early mention of both the perceptual relationships that increase knowledge and an unavoidably indexical link between term and matter. This indexical link is more explicitly restated in the quoted passage from Prolegomena (Prol 4: 267, 17), and it is for this reason that Hanna (2006, 212) adds: “[Kant] explicitly describes my concept of gold, which had as its content that this body is yellow and metal”. Although I will return to this point to mitigate the strength of certain conclusions, it is now necessary to point out the interdependence between the conceptual and the intuitive dimension in reference determination, which

124  The Simple Representation I depends not only on conceptualist criteria but also on the introduction of empirical intuition. Both the conceptualist and the Kripkian/Putnamian essentialist perspectives operate a clear-cut distinction between metaphysical and epistemic order. Of course, this takes place from two different directions, i.e., from the epistemic side of nominal essence in one case, and from the essentialist side of the substance nature in the other. The causal chain created by an introductory event binding name and substance returns the issue of linguistic reference to the essential nature of substances rather than to contingent, nominalist criteria; as a result, the reference between term and substance is secured, regardless of whether the latter is captured epistemically. Kant’s caesura concerns the epistemic perspective, implying the antiLeibnizian objection to the fixing of a species infima or a complete notion of an individual or substance from a metaphysical point of view (A655–6/ B683–4; cf. supra). At the same time, Kant (Log 9: 97) acknowledges the determination of a lowest species by convention, ignoring the conceptual differences which—according to the regulative principle of reason given by the transcendental law of specification (A656/B684)—can be articulated within any concept at all times. In addition, the contingent condition of uniqueness of the lowest species, in exactly the same way as the denotation of definite descriptions, requires an empirical intuition: A substance is specifically identified by a concept as long as an intuition incompatible with those conceptual assumptions is given (Thompson 1972, 331). In other words, the introduction of species infima or of criteria for the determination of a substance cannot provide identification conditions that are always valid, its conventional and provisional nature unfailingly implying the possibility for further intuitions to disavow those criteria. The following Kantian passage specifies this aspect by introducing new elements of reflection on the relationship of term, concept, and natural substance: In the concept of gold one person might think, besides its weight, color, and ductility, its property of not rusting, while another might know nothing about this. One makes use of certain marks only as long as they are sufficient for making distinctions; new observations, however, take some away and add some, and therefore the concept never remains within secure boundaries. And in any case what would be the point of defining such a concept?—since when, e.g., water and its properties are under discussion, one will not stop at what is intended by the word “water” but rather advance to experiments, and the word, with the few marks that are attached to it, is to constitute only a designation and not a concept of the thing; thus the putative definition is nothing other than the determination of the word. (A728/B756) This passage offers several key issues concerning the topics that are being touched on, and it should be connected to the distinction between the

The Simple Representation I  125 notion of definition, typical of mathematical concepts, and that of exposition, belonging to the sphere of empirical concepts (cf. Capozzi 1980). If we analyse the material dimension of concepts, the use of the terms associated with them, and the experimental scientific research on the making of the substances that are the very terms’ referents, from a transcendental point of view, it is not possible to establish the conceptual characteristics that define the contents of a term once and for all and, along with them, the properties that make up the essence of a natural substance. As it is not possible to establish a definition of the concept, we can only proceed with an exposition or description—provisional for its own nature—of the notes composing it. The definition is an exclusive privilege of the a priori synthesis of mathematical concepts: As this is arbitrary, originating in pure intuition without the material intervention of an object, it coincides with the construction, i.e., with the a priori representation of an object in intuition that contains nothing but the elements that compose the concept itself in a complete, precise Darstellung of the concept in intuition. The Semantic Reflection in the Logical Corpus Although the issue is somewhat articulated more extensively in the logical corpus, it reaches the same conclusion. Kant (Log 9:139) states that the doctrine of method, understood as the second part of the logic that follows the doctrine of the elements, analyses the form of science in general—the way to connect the manifold of knowledge to make it a science—and expounds on how we reach the perfection of knowledge. One of the essential, logical perfections of knowledge is expressed via the distinctness, thoroughness, and systematic ordering required by science; in turn, the distinction of concepts regarding what is contained within and below them is part of the distinction of knowledge. In this context, we find the notions of exposition and definition—where the latter is meant as a precise, sufficiently distinct concept—which are mainly analysed in Wiener Logik. In order to frame these concepts, it is necessary to follow Kant in his definition of the different types of concepts. The first distinction concerns (1) conceptus dati and (2) conceptus factitii: while the former do not depend on any choice whatsoever, the latter are produced on a voluntary basis. Conceptus dati can be: (1.1) a priori, if the concept is given in the intellect, e.g., the concept of cause, or (1.2) a posteriori, if the concept is given through experience, e.g., the concept of water. In turn, conceptus factitii can be (2.1) a priori, if the concept is formed through reflection without the intervention of any element of experience, e.g., the conceptus fictitius that supports the representation of a thousand-sided figure, or (2.2) a posteriori, if the concept is formed on the basis of an object given in experience, e.g., if we wish to form a distinct concept regarding a piece of metal we must use to carry out a series of experiments or tests, in order to record all the properties that might emerge from experience in order to extend the concept through the addition of subsequent notes. At

126  The Simple Representation I first glance, all conceptus dati (1)—be they a priori (1.1) or a posteriori (1.2)—can only be defined through the analysis: “For because [the concept] is given, I cannot make it distinct except by making clear the marks that lie in it, and that is just analysis” (V-Lo/Wiener 24: 915, 357). On the other hand, conceptus factitii (2)—be they a priori (2.1) or a posteriori (2.2)—can only receive a per synthesin definition (V-Lo/Wiener 24: 915). The synthetic definitions of concepts given a posteriori cannot be determined, as there is no knowledge of every possible, undefined conceptual mark experience can offer. On the whole, the analytic definitions of concepts given a posteriori themselves are incomplete: “When the inquirer into nature defines water, e.g., as a fluid body without taste or colour, one readily sees how precarious the definition is. He who is not already acquainted with water will not thereby become acquainted with it”. It follows that “the definition . . . is completely unsuitable for acquainting others with water. For in the concept water there lies so little that I immediately go outside the concept and have to collect new marks through experience” (V-Lo/Wiener 24: 918, 360). Furthermore, even the analytic definitions of concepts given a priori present several limitations—it is not easy to expound a concept’s marks in a precise way. In this regard, it is necessary to introduce the distinction between nominal and real definition. As mentioned previously, while the former corresponds to the meaning associated with a given name to designate logical essence, the latter concerns the knowledge of the object according to its internal determinations and corresponds to the foundation of the possibility of the thing (Capozzi 1980, 426–9). In particular, real definitions are only given for the conceptus factitii, e.g., the arbitrary, mathematical concepts, concepts of reason and arbitrary inventions. Apart from being real, they are always complete and are never wrong.12 Kant then addresses the possibility of employing the notion of ‘definition’ with regard to empirical concepts. To begin, the corresponding object cannot be defined by a real definition—we can only resort to the concept through which the object is thought of along with the corresponding marks involved in the nominal definition. Nonetheless, while these marks are always variable, they can never make up the complete concept of an object; once again, by using an argument similar to those also introduced in the KrV, Kant says that, in naming and defining empirical concepts, “we do not all have the same opinions”. As each speaker may have different experiences of the same thing, we can only acquire a provisional description of empirical concepts with reference to specific purposes. In the logical corpus, Kant also argues that a real definition—here synthetic and real—is possible, coinciding with the accomplished perfection of knowledge only as far as conceptus factitii are concerned, particularly for arbitrary concepts exemplified paradigmatically by mathematical concepts. Instead, since the synthesis of empirical concepts is endless, it is always possible to find new notes in experience that modify the concept. For this

The Simple Representation I  127 reason, it is only with mathematical, arbitrary concepts that we can associate a definition that Kant (Log 9: 142, 633) also terms declarations, “insofar as through them one declares his thoughts or gives account of what one understands by a word. This is the case among mathematicians”. This does not mean that, in a logical context, a logical/semantic reflection cannot serve the specific purpose of linking the use of language to a reliable semantic side (at least in a provisional way) due to the concept’s nominal definition and the corresponding logical essence governing the relationship between word and concept. Kant states this principle in several passages: By mere definitions of names, or nominal definitions, are to be understood those that contain the meaning that one wanted arbitrarily to give to a certain name, and which therefore signify only the logical essence of their object, or which serve merely for distinguishing it from other objects. (Log 9: 143, 634) It (Nominal definition) means almost nothing more than what the expression nominal definition says, a certain attestation to the name of the thing, in order to make the name of the thing distinct, but not to have better insight into the thing itself. (V-Lo/Wiener 24: 919, 361) In other words, at one pole, and with regard to the concepts given a priori and a posteriori, Kant seems to provide an analytic definition that, in the case of a material or real context for a posteriori concepts such as natural kinds, is judged to be “precarious” because such concepts can be modified at all times. In the case of a logical/formal context, an analytic definition will be needed in order to capture the logical essence of a concept and to assign a stable meaning. At the other pole there are the synthetic definitions of conceptus factitii: As with the above-mentioned example of the artefacts of the works of art, these can be named by an act of baptism and without the mediation of a specific, conceptual core so that their designation may become the connecting core for subsequent, synthetic definitions. The Role of Designation This creates a breach in Kant’s conceptualist approach to natural kind terms. In addition to the considerations from the corpus, a number of passages have also been quoted up to this point: (1) the passage from Prolegomena (Prol, 4: 267) concerning the content of the concept of gold as involving an indexical dimension; (2) the passage from the KrV (A721/B749) in which, once the content associated with the term gold has been explained, it is stated that “I can take the matter that goes by this name and initiate perceptions of it”; (3) the passage from the KrV (A728/B756) in which it is

128  The Simple Representation I asserted that we can never be sure that the same number of notes can be thought with the word that designates the same object; and, finally, (4) once again from the passage from the KrV (A728/B756), whenever we speak of water and its properties with specific reference to its substance and properties, we cannot limit ourselves to what is contained in the thought and concept given a posteriori: When, e.g., water and its properties are under discussion, one will not stop at what is intended by the word ‘water’ but rather advance to experiments, and the word, with the few marks that are attached to it, is to constitute only a designation and not a concept of the thing. If, in the material dimension of transcendental logic, the word, with the few marks that are attached to it, only constitutes a designation and not a concept of the thing, then logical essence is no longer the centre of mediation between substance and term. At the same time, as has been mentioned, the synthetic definitions of concepts given a posteriori cannot be determined because new observations can modify the conceptual content associated with the term;13 as a result, we cannot know all the possible marks provided by experience. In Logik Dohna-Wundlacken, Kant (V-Lo/Dohna 24: 728, 464) also points out that, for a natural substance such as water, “We can never have complete insight into the real essence, e.g., we can never experience all the marks of water no matter how far physics advances”. It is precisely for this reason that, in a material dimension, the designation (Bezeichnung)— i.e., the link between term and substance, even if accompanied by the notes that are associated with the term—becomes the fulcrum of the semantic device of natural kind terms. Since synthetic judgments are what come into play in this dimension, we need to abandon the logical/formal dimension and move to the indexical mechanism of the empirical intuition, which will eventually become the bridge linking word-concept and substance, the hook that enables us to speak of the same substance even if the same notes that constitute the related concept are not shared. Undoubtedly, these considerations link Kant to the spirit of the direct reference theory or, at the very least, they contradict a pure conceptualist approach à la Locke. At the same time, it does not seem possible to posit a direct relationship between terms and substances along the lines of Putnam’s indexical component through the attribution of a twofold component (both demonstrative and descriptive) to natural kind concepts.14 This is due to the presence in the designation, even so specified, of certain notes that constantly accompany the word-concept in the first place. More importantly, a natural kind concept is not a special empirical concept containing a demonstration component—as with any other empirical concept, it may have a singular use15—and be employed in indexical, cognitive performances as it involves empirical intuition within the form of a singular judgment. Kant’s

The Simple Representation I  129 examples, among which is the previously quoted passage from Prolegomena, should be contextualised in terms of this bigger picture. In brief, even when Kant focuses on the Bezeichnung in the KrV (A728/ B756)—the word that designates the thing, rather than its concept—this involves some of the notes that accompany the word and the intuition serving as the substance designative terminal. The demonstrative component is irrelevant to the concept and, in a material dimension—both in the formation of an empirical concept and in its application—only involves the intuition in an act of judgment that, in the applied case, tries to bring it under the umbrella of the abstract concept through the schematic process of imagination. Probably it is starting with a semantic reflection on proper names that the tension between a conceptualist approach and one based on a direct, semantic device becomes tangible, to the extent that, in certain passages, Kant seems to be attributing more importance to the latter than to the former. Following Capozzi’s (2009, forthcoming) contributions, proper names are associated with conceptus singulares, “as disconcerting may be the presence of this term in the Kantian logical lexicon” because Kant explicitly rejects the existence of singular concepts; ‘being something’ and ‘being one’ constitute the only necessary notes articulating their contents. This minimalist/conceptualist approach to the semantics of proper names allows considerations that seem to parallel the theory of direct reference if we consider the problem of epistemic identification and its relative application to proper names, and the fact that singular concepts designated by proper names cannot differ in content— which is the same for everyone—but only in number: “differentia numerica (Caius, Titius)” (Refl 2901, 16, 566). In Logik Dohna-Wundlacken (V-Lo/ Dohna 24: 756, 489), the numerical difference is explained as “the distinction of the conceptus singulares, insofar as they are not common to several. Among men we indicate them by nomina propria”. Accordingly, it is the names themselves—and their bearers’ designations—rather than the mediation of conceptus singulars—that determine the differences among individuals. Once again, in the example of Logik Dohna-Wundlacken, if two learned men are not distinguishable through the conceptual notes attributed to the singular concepts representing them, then it must be their numerical difference that makes a difference; this will be signified by the relative designation of proper names or their initials: “Learned men are specifically the same and generically, too, and nonetheless numerically different[,] as C. and J” (V-Lo/ Dohna 24: 764, 496). In conclusion, due to the presence of certain notes in the designation that constantly accompany natural kind terms (as well as proper names), Kant does not introduce a direct referential mechanism; in fact, he appears to continue to maintain a nominalist/conceptualist position spelled out according to the principles of transcendental idealism, even though he seems to be well aware of a few issues that are key to the theorists of direct reference.16

130  The Simple Representation I

The Simple Representation ‘I’ and the Role of Transcendental Designation Even though the representation I depicts quite a different scenario, as mentioned in the first paragraph, some scholars have also turned to the direct reference theory to account for the designation of I. While the use of a similar approach to I think has been quickly touched upon previously, any semantic involvement of the representation I in the direct reference theory will now be more firmly rejected, and the transcendental designation will be integrated into the metaphysical and the epistemic theses articulated previously. As mentioned, ‘thinking’ is the synthetic unity of apperception determining I think (qua Representation) as an analytic unity of apperception: I think is the representation through which spontaneity or thinking is given to oneself. Regarded as a simple representation (cf. B135, B138, A345/B404, A355, and many others), the representation I think simply designates the activity of thinking transcendentally. The concept of the transcendental subject (= x) has also been taken into account and regarded as the indeterminate thought of a thinking being with an existence in itself; on the other hand, the I in I think has been seen as the representation through which this thought is expressed. The well-known passages A346/B404, in which Kant introduces the concept of ‘a transcendental subject’ and specifies the I as a completely empty representation, can be appreciated once more: We can place nothing but the simple and in content for itself wholly empty representation I, of which one cannot even say that it is a concept, but a mere consciousness that accompanies every concept. Through this I, or He, or It (the thing), which thinks, nothing further is represented than a transcendental subject of thoughts = x which is recognized only through the thoughts that are its predicates, and about which, in abstraction, we can never have even the least concept. (A346/B404) The thinking being represents itself as a transcendental subject (= x) through the completely empty representation I. Hence, the concept of the transcendental subject amounts to the representation I, or the logical I— while I is the simple and entirely empty representation, the transcendental subject (= x) is what is represented. The I appended to thoughts is simple as it does not represent any property, signifying only ‘a something in general’ or ‘a transcendental subject’: “Nothing can be represented as more simple than that which is represented through the concept of a mere Something” (A355). As has already been pointed out with reference to the Copernican turn, in the transcendental approach a first-order talk about objects is replaced by a second-order discussion of the concept of an object and the conditions for the representation of an object. However, with regard to ‘thinking’, Kant often moves from the first-order discussion of its nature— for instance, when he maintains that the thinking is the being itself—to the

The Simple Representation I  131 second-order talk. By virtue of the concept of ‘a transcendental subject’ (or of ‘a something in general’) and the representation I, here he employs the second-order discussion, through which the thinking being represents its relationship with itself as a subject of inherence and, in turn, the latter is designated transcendentally through the simple representation I: But it is obvious that the subject of inherence is designated only transcendentally through the I that is appended to thoughts, without noting the least property of it, or cognizing or knowing anything at all about it. It signifies only a Something in general (a transcendental subject), the representation of which must of course be simple, just because one determines nothing at all about it; for certainly nothing can be represented as more simple than that which is represented through the concept of a mere Something. But the simplicity of the representation of a subject is not therefore a cognition of the simplicity of the subject itself, since its properties are entirely abstracted from if it is designated merely through the expression “I”, wholly empty of content (which I can apply to every thinking subject). (A355) Kant restates this point in his Metaphysical Foundations of Natural Science: The I, the general correlate of apperception, and itself merely a thought, designates, as a mere prefix [ein bloßes Vorwort],17 a thing of undetermined meaning—namely, the subject of all predicates—without any condition at all that would distinguish this representation of the subject from that of a something in general: a substance, therefore, of which, by this term, one has no concept of what it may be. (MAN 4: 542–3, 82) Assuming the relationship between the synthetic unity of apperception and I think (qua Representation) as an analytic unity of apperception, the I stands for the thinking being as follows: Due to this representation, the thinking being can think of itself as a thinking subject and refer to itself through the transcendental designation, i.e., the self-referential apparatus involved in transcendental apperception. The I is a simple representation, which does not convey any content and simply signifies ‘a something in general’ or ‘a transcendental subject’: As the properties of a thinking being are entirely abstracted, this is designated through the completely empty of content expression “I”. In this scenario, if the I simply signifies ‘a something in general’, namely ‘a transcendental subject’, then the I is a representational unicum, as it is neither intuition nor concept: For in that which we call the soul, everything is in continual flux, and it has nothing abiding, except perhaps (if one insists) the I, which is simple

132  The Simple Representation I only because this representation has no content, and hence no manifold, on account of which it seems to represent a simple object, or better put, it seems to designate one. This I would have to be an intuition, which, since it would be presupposed in all thinking in general (prior to all experience), would, as an intuition, supply a priori synthetic propositions if it were to be possible to bring about a pure rational cognition of the nature of a thinking being in general. Yet this I is no more an intuition than it is a concept of any object; rather, it is the mere form of consciousness, which accompanies both sorts of representations and which can elevate them to cognitions only insofar as something else is given in intuition, which provides the material for the representation of an object. (A381–2) It is interesting that Kant ponders the possibility of regarding I as an intuition, with the ultimate aim of demonstrating quite the reverse. Were it possible to construe a pure, rational cognition, I should be an intuition—and a pure one, too—coming into play to form a priori, synthetic propositions about the nature of the thinking being; in spite of that, this is not possible according to the Paralogisms section. It is worth noting that the I could also be regarded as an intuition at first glance, as it displays some of the features summed up as the intuition conditions mentioned above: A singular representation denoting an individual object (the singularity condition), as opposed to the concept which relates to different objects falling under it and grounded in an immediate relationship with its denotation (the immediacy condition), as opposed to concepts referring to the object through the mediation of the conceptual features, or marks, composing the concept’s very intention. Prima facie, it might seem possible, for instance, to account for the features of simplicity or emptiness of I’s representational form by the immediacy condition. If one also recalls the Kantian consideration that the expression “I”, which is completely devoid of content, can be applied to every thinking subject (A355), then Kant does seem to anticipate a contemporary proto-approach to indexicality, even based on the direct reference theory, as argued by Howell (2000) (cf. the first paragraph). Accordingly, and based on indexical features, I is a singular term/concept with a single individual as its reference and is governed by the token-reflexive rule, whereby every token of “I” refers to the subject who has produced or used it (cf. supra, Chapter 1). To rule out any Kantian interpretation based on the direct reference theory, some points need be discussed. Firstly, the I is an intellectual representation, not an intuition corresponding to whichever indexical condition is associated with sensibility: In no respect can the thinking being be regarded as a spatio-temporal object, as it is nothing more than that through which an object can be determined and known. As a result, the properties of a subject are entirely and necessarily abstracted, precisely because they cannot be determined at all:

The Simple Representation I  133 The consciousness in itself is not even a representation distinguishing a particular object, but rather a form of representation in general, insofar as it is to be called a cognition; for of it alone can I say that through it I think anything. (B404/A346) The ‘I’ exists in all thoughts, but not even the least intuition is involved in this representation, which would distinguish it from other objects of intuition (A350): Thinking, taken in itself, is merely the logical function and hence the sheer spontaneity of combining the manifold of a merely possible intuition; and in no way does it present the subject of consciousness as appearance, merely because it takes no account at all of the kind of intuition, whether it is sensible or intellectual. (B428–9) Moreover, if I cannot be said to be an intuitional representation due to the absence of any type of relationship to the sensible spatio-temporal forms, it is not conceptual either. The I does not represent any thinking subject conceptually, because it employs neither content mediation nor a prior instance of identification articulated in conceptual marks: An empty or bare form (A443/B471) is quite the reverse of the definition of the concept as a common representation, which is not free from extension and is subject to further determination at all times. In A381, Kant remarks that, at first glance, the I seems to designate rather than to represent: “[the I] is simple only because this representation has no content, and hence no manifold, on account of which it seems to represent a simple object, or better put, it seems to designate one”. The difference is crucial, and it calls into question the notion of Bezeichnung explained previously, as well as the difference from transcendental designation. As has been mentioned, when Kant addresses the designation of empirical concepts, he focuses on the word designating the thing rather than its concept, even though the latter involves some of the conceptual notes that accompany the word each time, as well as the intuition that serves as the substance’s designative terminal, namely the indexical device. With regard to the transcendental designation of I, the question is different and is also much more complicated. It is better to state that I seems to designate rather than to represent, as it seems to refer to a simple object (the thinking being) that has no properties to be articulated into conceptual or representational marks in turn. However, by recalling the formal reading used in these pages, and by highlighting passage A355, it is not possible to insist that there is any material or immaterial substance to be referred to by I: I is a simple representation not because it designates a simple object, but because it has no content at all; in other words, no manifold.

134  The Simple Representation I More specifically, the I appended to thoughts designates the ‘subject of inherence’ transcendentally: The I stands for the ‘subject of inherence’ or the ‘subject of all predicates’, which is not the real thinking subject—the thinking being—but concerns a mere logical role which is filled by the concept of ‘a transcendental subject’ or ‘a something in general’ instead. As said in Chapter 2, there is a difference between the logical and real features of the subject, and thanks to this difference it is possible to state without fear of contradiction that the I is both not a (general) conceptual representation and a simple representation which stands for the concept of ‘a transcendental subject’ or ‘a something in general’. If the former is simple—nothing is simpler than the concept of ‘a mere something’ (A355)—then the I also has to be a ‘simple’ or ‘empty’ representation. In this way, the thinking subject always represents itself as an ‘absolute logical subject of judgments’: As said previously in Chapter 2, it is absolute because the subject applies the concept I only to itself, and logical because the I cannot stand for a real subject of inherence. The point is also stated in A350, in which Kant critically points out that the first syllogism of transcendental psychology “imposes on us an only allegedly new insight”, as it mistakes “the constant logical subject of thinking” for “the cognition of a real subject of inherence with which we do not and cannot have the least acquaintance”. In Kant’s view, “consciousness is the one single thing that makes all representations into thoughts, and in which, therefore, as in the transcendental subject, our perceptions must be encountered”. Later, Kant claims that, “apart from this logical significance of the I, we have no acquaintance with the subject in itself that grounds this I as a substratum, just as it grounds all thoughts” (italics added). For this reason, Kant states that the thinking subject can only know that its self is the “permanent logical subject of thinking”, and not that it is a “real subject of inherence” (A350). With reference to the questions at issue here, two main results emerge from this analysis: The first considers the designative role of I, and the second the way in which the I designates based on the fact that it is associated with the logical role of a thinking subject. With regard to the first point, this interpretation does not entail an elusive account suggesting an I in I-think that is devoid of reference (the so-termed No-ownership Reading; cf. supra, Chapter 3). For Howell (2000, 121), it is true that “Kant’s talk of the I think or I as designating an entity, I, is quite full-blooded” because Kant is committed to there being a single selfidentical entity that is referred to by the occurrences of “I”. However, there are no identity conditions through the I think: As remarked previously, what is being assumed based on the representation I is nothing but an existence devoid of properties; the subject is able to know that it exists as a thinking being (B157, B159; cf. supra, Chapter 2) but is not able to know what it is. It follows that, to the extent that the subjective grasp of this entity through the I think extends, its identity is purely formal. In this manner, although the

The Simple Representation I  135 I does not designate any specific material or immaterial entity, it does stand for an existence devoid of properties represented. On the other hand, as just mentioned, the I does hold a logical role that is also filled by the concept of ‘a transcendental subject’. Two such roles will be balanced and settled together in the following pages. Secondly, with regard to the way I designates, besides anything else, the question can be clarified by recalling the general results of the analysis conducted in these pages. Both a general conceptualist/descriptivist strategy and strict indexical approach based on the direct reference theory have to be dismissed. (1) As has been mentioned in the first paragraph, by availing himself of the mental token-reflexive rule (R), Howell regards the I in I think as a mental indexical that represents the subject in a direct fashion: No properties of the subject are introduced into the contents of the thoughts other than the subject’s self. In particular, Howell’s analysis displays several specific features: 1) a rejection of the conceptualist view, whereby I think is regarded as the completely undetermined concept of a thinking being in general, 2) the use of the direct reference theory, according to which the subject’s self-reference via “I” establishes a directly referential relation to herself, importing the entity she is into the thought’s content, and 3) the ensuing use of the mental token-reflexive rule (R) in order to specify how the reference of “I” is determined. The critical point of this approach is the following: Due to the conceptualist view of I think expressed by Kant, it does not seem that the relationship between I in I think and the thinking being is a relation of designation which introduces the entity designated—the subject—into the content of the ‘I-thought’ in a directly referential way. It has just been seen that a conceptualist view of I think is expressed in a clear-cut way, especially in the passages from the Paralogisms that were analysed previously: I think must be regarded as the concept of a thinking being in general. (2) Even though a conceptualist view of I think rejects the use of the direct reference theory, this does not entail that I think should be regarded as a thought that, in a de dicto fashion, determines the unique thinking being—whichever it turns out to be—that corresponds to the concept: The I does not designate the subject-entity through the concept of ‘a transcendental subject’ or ‘a something in general’, as this would entail that the relation of designation is not of the propertyless type that Kant attributes to the I. As said previously, according to the difference between logical and real features, the representation I only designates a logical subject, i.e., the subject of all predicates, which cannot be distinguished from the concept of ‘a something in general’ and, above all, has to be precisely considered only as a mere logical role. Since this concept does not express a real feature of thinking subject, the I cannot designate it in a de dicto fashion.

136  The Simple Representation I In this way, even if the I stands for an existence devoid of properties, it does have a logical role that is fulfilled by the concept of ‘a transcendental subject’: As the representational correlate of transcendental apperception, the I designates the thinking being in self-consciousness because it is analytically contained in the synthetic unity of apperception as the representation of (the concept of) ‘a something in general’ or ‘a transcendental subject’. In a nutshell, the I does not designate the thinking being because it is the representation of ‘a transcendental subject’ that designates the thinking being in a de dicto fashion; neither does the I designate the thinking being because it is an indexical representation that designates its referent through a self-acquaintance relationship or a relationship linking the occurrence of a thought with the thinking being in a de re fashion. Kant firmly states that the thinking being is not an object of self-acquaintance (“apart from this logical significance of the I, we have no acquaintance with the subject in itself” A350) or perceptual relationship (B428–9). Instead, the I designates the thinking being inasmuch as this is analytically contained as a representation in the synthetic unity of apperception: For this reason, its origin is not indexical at all. It is necessary to recall the theses posited previously in order to develop this semantic conclusion concerning the transcendental designation of I. Since the transcendental apperception is an intellectual awareness of selfexistence, the thinking being is an existence devoid of properties represented to such an extent that Kant leaves the pronoun indeterminate (‘I’, ‘he’, ‘it’): “In the synthetic original unity of apperception, I am conscious of myself not as I appear to myself, nor as I am in myself, but only that I am. This representation is a thinking” (B157); “Through this I, or He, or It (the thing), which thinks, nothing further is represented than a transcendental subject of thoughts = x” (A346/B404); “in the consciousness of myself in mere thinking I am the being itself, about which, however, nothing yet is thereby given to me for thinking” (B429). On one hand, it seems possible to derive an approach on the nature of the determining thinking being: As has been mentioned above, the principle of the unity of apperception expresses the statement “the synthetic unity of consciousness is an objective condition of all cognition”, i.e., an analytical principle asserting that all representations must stand under such a principle in order to ascribe them to an identical self as its representations; only in this way can they be grasped as a synthetically combined whole in this act of apperception of conjunction. At the same time, the subjective component comes into play to affirm a first-person perspective: A single complex thought logically involves a single subjective dimension; every single component of such a complex thought is ascribable to an identical thinking subject—thus, the thinking being becomes aware of its identity as a thinking subject in the synthetic unity of such a complex thought. As mentioned in Chapter 1, Kant elaborates on this point in a famous passage in which

The Simple Representation I  137 he states that the analytical unity of apperception presupposes a synthetic unity: This thoroughgoing identity of the apperception of a manifold given in intuition contains a synthesis of the representations, and is possible only through the consciousness of this synthesis. For the empirical consciousness that accompanies different representations is by itself dispersed and without relation to the identity of the subject. The latter relation therefore does not yet come about by my accompanying each representation with consciousness, but rather by my adding one representation to the other and being conscious of their synthesis. Therefore it is only because I can combine a manifold of given representations in one consciousness that it is possible for me to represent the identity of the consciousness in these representations itself, i.e., the analytical unity of apperception is only possible under the presupposition of some synthetic one. (B134) On the other hand, when the first-order discussion about the ‘thinking being’ or the synthetic unity of consciousness is replaced by the secondorder discussion, the representation I is introduced, and the determining subject is therefore enabled to represent its relationship with itself from its first-person perspective. As the representational correlate of apperception, the I is the representation of a logical subject, and yet it is to be understood as the representation of ‘a something in general’—just as self-consciousness is an intellectual awareness of self-existence—that is, something unidentifiable from an epistemic angle. What is represented is a transcendental subject of thoughts (= x) that can only be recognised through the thoughts that are its predicates. As mentioned previously, the fact that the transcendental subject of thoughts = x is only recognised through the thoughts that are its predicates amounts to saying that the analytical unity of apperception is only possible under the presupposition of some synthetic unity (B133–4). The former is a second-order approach that introduces the first-person perspective through I think (qua Representation); the latter, on the other hand, consists of a first-order approach that takes the ‘thinking’ as a synthetic unity of apperception in a completely impersonal way into account. For this reason, it is not possible to explain transcendental designation on the basis of the direct reference theory. If to think is to unify the manifold conceptually, and hence every single thought expressed by a judgment is grounded in the principle of transcendental apperception, then the I simply expresses the analytical unity of apperception, which is only possible under the presupposition of a synthetic unity. In other words, and from a second-order perspective, the I is the representation that is inherent in the very concept of thought (B132); in the A-edition Paralogism, Kant tackles

138  The Simple Representation I the relationship of all thoughts and the ‘I’ taken as the common subject to which they inhere to affirm that the representation ‘I’ features in all thoughts (A350). Kant sees the I as the nexus established by the copula in the judgment by the representational synthesis of the unity of apperception. In other and more concise words, the I is solely given as the analytical form of thought. In a way, not only in the inner sense but also at the heart of transcendental apperception proper, Kant is as Humean as Hume himself. If the I of apperception is not originated by the mental token-reflexive rule but is seen, instead, as the representation of a logically simple subject analytically contained in the concept of ‘thinking’ (B407–8), then it must determine (i.e., it is inherent in) the form of every judgment in general terms (B406): For this reason, Kant can state, in a quite indexical fashion, that the wholly empty of content I can be applied to every thinking subject (A355). This implies that the subject’s self-reference via I does not import any entity into the thought’s content: The I is the representation of the subject of thinking (KGS 7: 134), but the thinking being is represented in nothing other than the form of the judgment. Thus, the possibility of thinking of oneself as a subject is not given by importing the entity of which the subject consists into the thought’s content, but properly and solely by the form of the thought that, analytically, contains the simple and empty representation I. In brief, it is not an entity in the thought’s content, which could justify the indexical nature of I, or the thought of the ‘thinking subject’, which could determine the reference of the subject through the mediation of properties in a de dicto fashion—as there is no property to associate with the ‘thinking being’, except for the fact (but, again, not the property!) that it exists—but only the form of ‘spontaneity’ or ‘thinking’ that contains the representation I. Kant’s view and naming of the designation of apperception of I “transcendental” is further evidence of the fact that the origin of I as the representational correlate of the synthetic unity of transcendental apperception is not indexical. It is known that Kant calls all cognition that is concerned not with objects but with their a priori concepts “transcendental” (hence, Transcendental Philosophy is the general system of such concepts; A12); furthermore, Kant specifies that not all a priori cognition must be called “transcendental”, “but only that by means of which we cognize that and how certain representations (intuitions or concepts) are applied entirely a priori, or are possible (i.e., the possibility of cognition or its use a priori)” (A56/B81). Regarded as a simple representation, I think designates transcendentally the activity of thinking given by the synthetic unity of apperception, namely as the nexus established in the judgment by the copula serving as the link between the representational synthesis and the synthetic unity of apperception on a conceptual basis. It follows that the transcendental designation is concerned with the conditions of possibility for and how the representation I think is used a priori—given that the designation of I is transcendentally inherent in the ‘thinking’—and not by picking out some entity a posteriori. In conclusion, I think consists of the representation that

The Simple Representation I  139 means the concept of the ‘transcendental subject’ and presents a designative function referring to something that really exists as (an act of) Thinking (Being) for no other reason than that ‘thinking’ as spontaneity is the synthetic unity of apperception that contains the representation I think as an analytical unity of apperception.

Notes 1. Howell (2000, 137): “The reflexivity of the I or I think act of thought means, roughly, that that act of thought is not just the thinker’s act of awareness of the content that the thinker thinks through that act. (I think the content that: it is raining, etc.) Instead, it is an act of thought-awareness by the thinker, that not only grasps that content but also takes that very act itself to have the thinker as its subject”. In this way, by turning on the self-referential aspect of the I think act and involving the idea of “an act centered on (experienced as referring back to) a certain subject” (2000, 139), Howell puts forward a Kantian way of understanding how the representation I—considered as a device of direct reference—acquires its peculiarly first-person character. Thanks to the selfreflexive I think act, this thought-experience becomes the experience of that act itself as having its origin in the very subject. 2. Along the lines of Gunther (2003)—cf. also Speaks (2005) and Bermúdez (2003)—Hanna (2008) outlines seven different arguments used in the debate for non-conceptualism: (1) if animals and children do not possess concepts, then their perceptual capacity must depend on non-conceptual cognition with non-conceptual content; (2) if the perceptual experience has a qualitative character based on the so-called phenomenological fineness of grain, which cannot be captured via a conceptual articulation, then a part of the human perceptual experience is non-conceptual; (3) if a subject can discriminate an item from a perceptual point of view, without the strictly conceptual ability to re-identify the item in question, then the subject is capable of non-conceptual cognition with non-conceptual content; (4) if it is possible to have a perception (or experience) of something without a judgment, then the cognition that is not based on judgments is non-conceptual with non-conceptual content; (5) if one assumes the distinction between knowing-how and knowing-that (or knowing-what), then the capacity to know how to do something without knowing that or knowing what you are doing presupposes that the knowing-how is non-conceptual cognition with non-conceptual content; (6) if one of the assumptions of the theories in concept-acquisition is based on the non-conceptual perception of the objects falling under such concepts, then part of the perceptual capacity is nonconceptual with non-conceptual content; and (7) if the theory of demonstratives includes the thesis that demonstrative reference is essentially fixed indexically— and, on that account, non-descriptively—then perception is non-conceptual with non-conceptual content. 3. Within what he terms Absolutist Non-Conceptualism, Hanna (2008) distinguishes a weak Absolutist Non-Conceptualism from a strong Absolutist Non-Conceptualism. According to the first theory, the structure and function of perceptual mental contents are contingently distinct from the structure and function of conceptual content (contingently absolutely non-conceptual content). According to the second thesis, the structure of perceptual mental contents is essentially distinct (essentially absolutely non-conceptual content). What Tye (2006) terms the robustly non-conceptual content of a perceptual state—in his view, a Russellian rather than a Fregean proposition—according to Hanna might be related to a contingently absolutely non-conceptual content: As Tye

140  The Simple Representation I himself (2006) remarks, it is not excluded that this can be articulated through the conceptual dimension. 4. According to Hintikka, singularity and immediacy are not to be regarded as distinct criteria in the definition of an intuition: While the former is the only distinctive feature needed to define an intuitive representation, the condition of immediacy is a secondary trait arising from the criterion of singularity itself. Expressly, singularity is the necessary and sufficient condition to define the notion of intuition, the criterion of immediacy—understood as the absence of mediation of conceptual marks—being just a mere corollary of singularity. This concerns the intuitions’ reference modality based on a specifically immediate or direct relationship with their respective objects; on the other hand, the concepts’ characteristics or notes are involved in the determination of the reference (Hintikka 1972, 342). It follows that the immediacy criterion does not establish a necessary link between intuition and sensibility—Hintikka (1967) points out that in a phase of Kant’s approach the intuitive representation was not connected to sensibility—nor, a fortiori, can the criterion serve as an additional specification for its definition: Understood as a mere corollary of singularity, it does not presuppose the presence of any object (Hintikka 1972, 341). On account of his different approach to the Kantian philosophy of mathematics, Parsons rejects the primacy Hintikka assigns to the singularity criterion in the definition of the intuition: The condition of immediacy—considered by the author as an essentially phenomenological–perceptual epistemic criterion—is not so much a dark formula nor a corollary of the singularity condition but rather the criterion to identify intuitive representations, which also implies that the object to which the intuition refers is somehow directly present to the mind and perception. To Parsons, singularity is a necessary but not sufficient condition, too large a criterion to encompass the notion of intuition: In fact, there can be singular representations which are not immediate, such as those linguistically expressed by definite descriptions; for this reason, he introduces the immediacy criterion as a separate condition from singularity. Ultimately, while to Hintikka the intuition is immediate for the very reason that it is singular, to Parsons it is singular and immediately linked to its object. 5. The different interpretations present in the debate between conceptualism and non-conceptualism will not be addressed here; cf. Schulting (2016) for an introduction to the Kantian debate. 6. Both intention and extension fall under the notion of ‘containment’. While the former consists of the set of marks contained by the concept (V-Lo/Pölitz 24: 569)—a generic content to be kept distinct from the logical essence (cf. infra)— “the extension of a concept is a sphaera, and it is concerned with the multitude of things that are contained under the concept” (V-Lo/Wiener 24: 911, 354). In particular, “the more the things that stand under a concept and can be thought through it, the greater is its extension or sphere” (Log 9: 96, 593). It is necessary to specify that the extension can also consist of a set of concepts (cf. PhilEnz 29: 17). Following the Port-Royal tradition (see Capozzi and Roncaglia 2009, 99–100; Capozzi 2009, 127), there are two definitions of ‘extension’ in the literature: (1) in the logical doctrine of judgments and inferences, the extension consists of the things contained by the concept, and is also referred to as the “extension-class” by Capozzi, and (2) in the logical doctrine of concepts, Kant speaks of extension logic (Log 9: 98, 596), whereby the extension consists of the concepts in the intentions of which the concept is contained (Capozzi 2009, 128; Capozzi forthcoming). 7. Capozzi (forthcoming): “There is a difference between the singularization of ‘house’ and the singular ‘Julius Caesar’, a difference that is disregarded in

The Simple Representation I  141 Thompson’s interpretation. As we have seen, ‘this house’ preserves the logical essence of the general concept ‘house’ and indicates a point in the class-extension of the concept ‘house’ and of no other concept. ‘Julius Caesar’ conveys a conceptual content, but this content differs from the logical essence of general concepts. A singular concept does not contain a complex of few, primitive, immutable, necessary and unchangeable marks that are associated to the name and are available to anyone. A singular concept contains a single obligatory mark—the thought of something singular—which can be complemented with variable aggregates of conceptual marks with a freedom impossible for ‘house’ ”. 8. Cf. essays in Putnam (1975), particularly The Meaning of ‘Meaning’, and Kripke (1980). Needless to say, while a number of authors judge metaphysical essentialism to be a consequence of the theory of direct reference, others consider the two positions to be independent; Kripke and Putnam’s approaches to scientific essentialism, which will not be taken into account here, differ because Putnam has reworked fundamental metaphysical and epistemic aspects of his philosophical reflection. 9. Hanna (2006, 152–3): “Scientific essentialism is held to follow directly from: (1) A general thesis concerning the correct analysis of necessary or strongly modal statements (namely, that they are logically or strictly metaphysically necessary and ‘Leibnizian’, or true in all logically possible worlds accessible from a designated world, our actual world). (2) A general theory of the semantics of natural kind terms (namely, that every natural kind term is a ‘rigid designator’: it holds its actual-world reference fixed across all possible worlds in which its referent exists, and never picks out anything else otherwise). (3) A specific thesis about the modality of identity-statements involving rigid designators (namely, that if they are true, then they are necessarily true). (4) A specific linguistic claim to the effect that natural kind terms are rigid designators, based on some proposals concerning the nature of the linguistic mechanisms of reference-fixing at work in society at large and in the natural sciences in particular (namely, the ‘causal theory of names’, and the socio-linguistic hypothesis of ‘linguistic division of labor’)”. 10. Obviously, not all rigid designators are directly referential (for a presentation of the theory of direct reference, cf. Kaplan 1989). Needless to say, there has been much debate concerning whether and (if so) how natural kind terms are to be understood as rigid, and there is no consensus—cf. Bird, Tobin (2012). Since the main aim of this chapter is also to verify the presence of a direct referential mechanism in the Kantian approach to the semantics of natural kind terms, I will not consider these aspects of the debate. 11. These cannot overlay the two components of the semantic device included in Anderson (1994), one ‘stable’ and the other ‘unstable’, wherein Kant’s natural kind terms are quasi-indexical terms, the meaning and reference of which can be determined through a core-concept and, in a broad sense, a pragmatic dimension based on an indexical device. However, I will return to this point later when discussing Hanna’s (2006) somewhat similar proposal. 12. In Logik Blomberg, which is based on lectures given early in the 1770S and then before the introduction of the distinction between transcendental and general logic, Kant argues: “it lies solely with me to make up the concept and to establish it as it pleases me, and the whole concept has thus no other reality than merely what my fabrication wants[;] consequently I can always put all the parts that I name into a thing[,] and these must then constitute the complete, possible concept of the thing, for the whole thing is actual only by means of my will” (V-Lo/Blomberg 24: 268, 216). It is exactly because concepts or artefacts are arbitrary that we can choose how to call them through an act of baptism:

142  The Simple Representation I “since . . . I will that this shall be called thus, it is called thus” and this is the reason why “the names of things that arise in art are also always true[;] their inventor is always right, for they are arbitrary concepts” (V-Lo/Blomberg 24: 268, 216). 13. In Logik Blomberg (V-Lo/Blomberg 24: 112, 87): “Gold differs from all other metals through its weight, but a metal has also been discovered, named platina, that is white in colour but has the same weight as gold. Consequently this mark, this distinguishing ground of gold from all remaining metals, is not sufficient enough” Cf. Capozzi (2002, 497). 14. Hanna (2006, 212–13): “Kant regards natural kind concepts as a special case of his theory of empirical concepts. He explicitly describes ‘my concept of gold, which had as its content that this body is yellow and metal’; hence gold partially decomposes to this conceptual microstructure: . It can be immediately seen that for Kant gold contains two distinct components: (a) a referential component, this body; and (b) an attributive or descriptive component that reflects some (but obviously, not all) of the manifest identifying properties of gold, namely, yellow and metal. What I want to argue is that for Kant this fusion of distinct components yields the very feature of natural kind concepts that sets them radically apart from all other sorts of empirical concepts”. Kroon and Nola (1987) have already argued that Kant implicitly—as well as inconsistently—identifies two different natural kind concepts, a rigidly designating concept and a concept-stereotype. 15. As mentioned previously in the third paragraph, Kant (Log 9: 91) explicitly states that there is no singular concept whatsoever: By definition, concepts are representations, the logical form of which is always general. However, he also admits the possibility of three different uses of concepts: one general, one singular, and one particular. 16. Due to the designation of a substance, we can speak of the same substance even if two speakers have different knowledge of it and even if they do not share exactly the same conceptual notes, or when science generates new knowledge and further observations change the concept associated with the substance term. As has been mentioned with regard to the difference between definition and exposition, “an empirical concept cannot be defined at all but only explicated. For since we have in it only some marks of a certain kind of objects of the senses, it is never certain whether by means of the word that designates the same object one does not sometimes think more of these marks but another time fewer of them” (A728/B756). 17. Rosefeldt (2017, 228): “Interpreters have been rather creative in explaining why Kant called the representation ‘I’ ‘ein bloßes Vorwort’ here, and Friedman . . . seems to think that he has to use the rather vague and extravagant expression ‘prefix’ for ‘Vorwort’ in his translation in order to capture what Kant meant. In fact, the exegetical situation is much less complicated. ‘Vorwort’ is just the 18th-century German translation for the Latin grammatical term ‘pronomen’, a translation that has survived in contemporary German where a pronoun is sometimes called a ‘Fürwort’ ”.

5 On de se and de re

As pointed out in Chapter 1, since the classic works by Castañeda (1966), Perry (1979), and Lewis (1979), de se thoughts have been described as thoughts about oneself “as oneself”. In recent years, several theoretical perspectives have gained ground, and even though the transcendental system does not seem to include any explicit articulation of de se thoughts, a few features of transcendental apperception and I think do seem to anticipate some of the points made by Perry and Recanati concerning the so-called implicit de se thoughts in the specific terms of Transcendentalism. In this chapter, the distinction between de re and de se thoughts will be addressed in order to provide an outline of the results relating to I-thoughts achieved thus far in the Kantian approach.

Knowledge by Acquaintance and Knowledge by Description: The Weak Conceptualist View From the semantic considerations concerning the referential mechanism of intuitive representations expounded in the preceding chapter to a more strictly epistemic perspective, with regard to the distinction between conceptual and non-conceptual content, the Kantian difference between concepts and intuitions has been partly associated with the distinction between knowledge by acquaintance and knowledge by description, which specifies two basic types of knowledge (Hanna 2008, 52). Since Russell onward to the current debate between descriptivism and singularism—which involves both philosophy of mind and language (see Jeshion 2010)—the intuitive difference between descriptive thoughts about a particular object or individual (e.g., “The strongest man in the world can lift 150 kg”) and the so-called non-descriptive or de re thoughts (e.g., “that man is drunk”), based on a relationship of acquaintance, can be grasped immediately on account of the different nature and role of the respective modes of presentation in play (see Burge 1977, 1979; Bach 1987; Recanati 2009). Based on a reconsideration of Russell’s notion of acquaintance through Evans’s neo-Fregean lesson, which explicitly takes non-descriptive modes of presentation into account, in de re thoughts the individual or

144 On de se and de re object to which the thought refers is determined by a demonstrative mode of presentation specified through a relationship of information-perception linking the object to the occurrence of the thought (Recanati 2009, 252). Descriptive representations represent their referents through the properties they instantiate, their reference being determined by the existence of whatever may satisfy such properties. Instead, non-descriptive representations represent their referents through a contextual relationship linking the occurrence of a thought with the object in question. In this context, and due to their indexical nature, non-descriptive representations are token-reflexive and, as such, display two semantic levels.1 In the so-called de re thoughts, the individual or the object to which the thought refers is determined by a demonstrative mode of presentation specified by a relationship of information-perception between the object and the occurrence of the thought. In brief, non-descriptive representations represent their referents through a contextual relationship between the occurrence of a thought and the object itself. This implies that the determination of the reference does not depend on some inherent representational criterion—in other words, it is not based on the fact that the representation conforms to the object to which it refers— but on an external criterion—a perceptual relationship between the representation and its object, which, as such, is not represented in the content of the representation itself. Expressed in a different way, what makes an object the referent of the thought in de re thoughts is not the fact that the object satisfies any of the properties expressed by a concept; obviously, properties may also come into the picture, as the property of being drunk in the de re thought exemplified by “that man is drunk”, and yet the referent— i.e., the man associated with the property of being drunk—is not so much determined by the properties instantiated by the direct mode of presentation through a relationship of acquaintance that is not represented as such. In de re thoughts, properties play no role in determining to what a non-descriptive representation refers, to the extent that I can correctly refer to the man to whom I attribute the property of being drunk even if I actually see that the man is not drunk at all: What matters is the particular perceptual relation established between the object and the mental occurrence. In Kant, we can find an articulation of the different types of cognition that calls to mind a specific difference between propositional knowledge (erkennen dass) and knowledge by acquaintance (kennen) (cf. de Sá Pereira 2013). Following the Jäsche Logik (cf. also Blomberg Logic, § 139): The first degree of cognition is: to represent something; The second: to represent something with consciousness, or to perceive (percipere); The third: to be acquainted with [kennen] something (noscere), or to represent something in comparison with other things, both as to sameness [Einerleiheit] and as to difference;

On de se and de re  145 The fourth: to be acquainted with [kennen] something with consciousness, i.e., to cognize it [erkennen]; (cognoscere). Animals are acquainted with [kennen] objects too, but they do not cognize [erkennen] them. (Log 9: 33, 569–70) The classic Kantian example of the savage used by Hanna (2006, 104) to explain one of the Kantian non-conceptual content types seems to fall under the types of de re thoughts: If a savage [Wilder] sees a house from a distance for example, with whose use he is not acquainted, he admittedly has before him in his representation the very same object as someone else who is acquainted with it determinately as a dwelling established for humans. But as to form, this cognition of one and the same object is different in the two. With one it is mere intuition, with the other it is intuition and concept at the same time. (Log 9: 65, 544–5) As opposed to a more “urbanised” subject, the savage does not possess the concept of ‘house’, nor does he recognise that particular object as such; be that as it may, he does possess an intuitive presentation that establishes an immediate, singular representational link with the object in question.2 The Theory of the Transcendental Object With reference to the issues summarised above, which are at the heart of the debate between conceptualists and non-conceptualists, this paragraph will address certain aspects of the so-called theory of the transcendental object. The theory has important repercussions because it appears to deny that intuition is provided with an autonomous function of de re presentation. The condition of possibility for intuition to apprehend a phenomenal object is based on an act of thinking of the intellect, which involves the concept of the transcendental object (Allison 1968). According to the well-known distinction between noumenon and phenomenon (cf. supra, Chapter 2), the object o has an existence in itself and is presented as a phenomenon through intuition i. Since it is not possible to know an object in itself outside the representational order, to think of an object in itself, it is necessary to employ the indeterminate thought that something in general (= x) exists in itself (A104). As seen in the Chapter 2, the concept of an object in general is nothing but the concept of a transcendental object: Intuitions are singular representations that immediately refer to objects, whereas appearances are representations to be kept distinct from that which is referred to as “transcendental object” (= x). Thus, the concept of a general or transcendental object (= x) is the indeterminate thought of

146 On de se and de re a single object having an existence in itself, the condition of possibility to think of the singular object—which is spatio-temporally determined by the intuition as Erscheinung—and to provide it with objective reality through the resulting unification of empirical concepts and the relative attribution of the properties presented by the intuition: We are now also able to determine our concepts of an object in general more correctly. All representations, as representations, have their object, and can themselves be objects of other representations in turn. Appearances are the only objects that can be given to us immediately, and that in them which is immediately related to the object is called intuition. However, these appearances are not things in themselves, but themselves only representations, which in turn have their object, which therefore cannot be further intuited by us, and that may therefore be called the non-empirical, i.e., transcendental object = X. The pure concept of this transcendental object (which in all of our cognitions is really always one and the same = X) is that which in all of our empirical concepts in general can provide relation to an object, i.e., objective reality. Now this concept cannot contain any determinate intuition at all, and therefore concerns nothing but that unity which must be encountered in a manifold of cognition insofar as it stands in relation to an object. (A108–9) Leaving aside the many exegetical and theoretical problems posed by the relationship of the transcendental object, the noumenon, and the thing in itself,3 and based on the terminological change introduced in the second edition of the KrV with reference to the distinction between a positive and a negative sense of noumenon—which is not supposed to affect the gist of the theory, despite Kant’s dismissal of the expression “transcendental object”— for Allison (2004, 61), the concept of the transcendental object should be regarded as a sort of transcendental pointer: “It serves to define the philosophical task by indicating that the commonsensical and transcendentally realistic concern with the ‘real’ nature of objects must be replaced by a critical analysis of the conditions of the representation of an object”. At this stage, it might be useful to refer to the intensional logic apparatus employed by Howell to articulate the theory of the transcendental object, starting with a slight adaptation of one of his examples. Suppose that a subject H has a cognitive relationship with a triangular wooden house: The empirical intuition i displays the different properties contained in the manifold of representations, e.g., i1 displays property P1 of being made of wood, i2 displays property P2 of being a house, i3 displays a spatial part s1, corresponding to the upper part of the house positioned in l1, and, lastly, i4 displays a spatial part s2, corresponding to the lower part of the house positioned in l2. The manifold is synthesised in the intuition by the imagination and the application of concepts so that the intuitive, singular

On de se and de re  147 representation falling under the empirical concept “triangular wooden house” displays properties and spatial parts jointly (C being the concept of “wooden house”, D being the concept of “being triangular”). Given that the conditions of possibility for (1) H knows that (object o is made of wood) presuppose (2) H thinks that (object o has the property of being made of wood), the example can be articulated as follows: (3) i1 displays P1 to H & then i2 displays P2 to H & then i3 displays s1 (as occurring at l1) to H & then i4 displays s2 (as occurring at l2) to H & H thinks that [some single object X is such that (X has P1 & X has P2 & X has s1 & X has s2 & s1 occurs at l1 & s2 occurs at l2 & X occurs at l1 + l2 & P1 and P2 jointly constitute concept C & s1 and s2 jointly specify concept D)]. In summary, what is involved here is an intensional context that does not allow the substitution of co-referential singular terms. As is well known, for Quine (1956), who speaks expressly of the relational and notional sense of those attitude verbs such as “believe”, a sentence containing an intensional operator can be ambiguous with regard to a de re or a de dicto interpretation. However, this can be grasped immediately as soon as the scope of the existential quantifier is ascertained: Using the same example in a synthetic way, (4) H thinks that (some single object X is such and such) presents a de re and a de dicto interpretation—respectively, (4.1) (∃x)[x has existence in itself & x is an object & H thinks that (x is an object & x is such and such)], and (4.2) H thinks that (∃x) (x is an object & x is such and such). The de dicto reading captures the peculiar features assigned by Kant to the concept of a general or transcendental object (= x): Specifically, this takes on the role of an indeterminate thought, which, although referring to a single object (= x), does not relate to any object in particular: [H’s de dicto thought] is in fact an indeterminate thought, a thought about an object in general, and a thought nevertheless not about a special type or kind of object, just in the following sense: namely, just in the sense that in this thought H thinks that there is an object x, in the most general sense of ‘object’, which is such and such; but in this thought H does not think, de re fashion, of any particular object (or type or kind of object), that that object is such and such. It does seem to be in just this sense that Kant means us to understand the sort of indeterminacy that, according to the transcendental-object theory, is supposed to attach to H’s act of thought. (Howell 1981, 102) The de dicto reading of (3) produces (5): (5) i1 displays P1 to H & then i2 displays P2 to H & then i3 displays s1 (as occurring at l1) to H & then i4 displays s2 (as occurring at l2) to H & H

148 On de se and de re thinks that (∃ x) (X is an object & X has P1 & X has P2 & X has s1 & X has s2 & s1 occurs at l1 & s2 occurs at l2 & X occurs at l1 + l2 & P1 and P2 jointly constitute concept C & s1 and s2 jointly specify concept D). In this Kantian scenario, the typically de re knowledge of a single particular object—e.g., of a subject perceiving a triangular wooden house before her—is a de dicto act of thought. On the basis of a de re mechanism, a phenomenal object is identified by the presentation function performed by the indexical empirical intuition locating the object in a given position within the spatio-temporal forms of sensibility. Through a de dicto mechanism, the relative concept of a transcendental object is involved as a condition of possibility in order to think of an object; therefore, the attestation of the singularity of a spatio-temporal something by intuition can be represented as an object to which is attributed certain properties due to the synthesis of the empirical concepts. The concept of the transcendental object is a pointer, although a conceptual rather than a relational one by definition. With regard to the above remarks concerning the difference between descriptive and non-descriptive representations, if the intuition presents a non-descriptive relational mode of presentation, and if such a relationship is not displayed in the representation’s content, in order to represent the object presented by the intuition as an object, it is then necessary to employ the thought of such a relationship and to represent it a priori as its own condition of possibility proper: “We find, however, that our thought of the relation of all cognition to its object carries something of necessity with it” (A104). Non-Conceptualism Versus Weak Conceptualism The interpretative reading based on the theory of the transcendental object that might be referred to as weak conceptualism features a number of contrasting results, with the main points stirring the debate on Kantian non-conceptualism. In addition to what was said about Kantian conceptualism and nonconceptualism in Chapter 4 according to Hanna (2016), conceptualism can be divided according to two theses: (1) all rational humans’ objectively valid representational content is strictly determined by conceptual capacities alone, and (2) non-rational humans or non-human animals are not capable of objectively valid representations. Kant’s conceptualism follows directly from the togetherness principle and can be divided into (1) strong Kantian conceptualism: The innate conceptual human capacities not only strictly determine all conscious objective representational content, but also strictly determine the faculty of sensibility itself and all the intuitions (see Sellars 1963, 1968; McDowell 1994; Abela 2002), and

On de se and de re  149 (2) weak Kantian conceptualism: The innate conceptual human capacities strictly determine all conscious objective representational content, although the faculty of sensibility independently provides a necessary condition for conscious objective representation (see Ginsborg 2006, 2008; Grüne 2009; Land 2011; Griffith 2012; Williams 2012). According to Hanna (2014), the weak variety at least minimally preserves Kant’s cognitive dualism of faculties, and also some sort of semi-independent cognitive role for intuitions (even though it still rejects the thesis that intuitions have an essentially different kind of representational content from concepts), the strong variety does not countenance any of these concessions to non-conceptualism, and thereby, in effect, strong Kantian conceptualism explanatorily reduces the faculty of sensibility to the faculty of understanding. Instead, non-conceptualism can be defined according to three theses: (1) not all rational humans’ objectively valid representational content is determined by conceptual capacities alone, (2) at least some rational humans’ objectively valid representational contents are both autonomous from and independent of conceptual content, and are also strictly determined by non-conceptual capacities alone (see infra), and (3) at least some and perhaps most non-rational humans or non-human animals are capable of objectively valid representation. The Kantian non-conceptualists’ aim is to delimit the conceptualists’ claims behind the argumentative structure of the Transcendental Deduction, including some of the arguments discussed in the previous chapter, through a deflationary strategy aimed at weakening the togetherness principle. As observed previously, the involvement of the faculty of judgment is alleged to be necessary with reference to objective knowledge only, not to perception, as perception is assumed to rely on the use of intuitive representations without the intervention of concepts understood as general rules constituting judgments (cf. Hanna 2006; Allais 2009). Hanna distinguishes two theses, the independence of intuitions from concepts thesis and the autonomy of intuitions from concepts thesis: Now focusing on the case of intuitional content, empirical or pure concepts are neither sufficient nor necessary for empirically meaningful, objectively valid intuitions. Sufficiency would mean that fixing the objectively valid representational contents of all relevant concepts would necessarily fix the objectively valid representational content of any intuition, and necessity would mean that fixing the objectively valid

150 On de se and de re representational content of at least some concepts is required for the determination of the objectively valid representational content of every intuition. I call the thesis that empirical or pure concepts are not sufficient for the determination of the objectively valid representational content of intuitions, the independence of intuitions from concepts, and I call the thesis that intuitional, or essentially non-conceptual, cognitions and contents can also exist and be objectively valid in the total absence of concepts and conceptual capacities alike, hence the thesis that that empirical or pure concepts are not necessary for the determination of the objectively valid representational content of intuitions, the autonomy of intuitions from concepts. (Hanna 2016, 47–8) As seen, Hanna argues that the togetherness principle is consistent with Kantian non-conceptualism on account of the epistemic and metaphysical independence of the intuitive representation. Hanna (2016) develops this point very clearly; it is worth quoting the following excerpt: What Kant is actually saying in the famous texts at A50–1/B74–6 is that intuitions and concepts are cognitively complementary and semantically interdependent for the specific purpose of constituting objectively valid judgements. This in turn corresponds directly to a special, narrower sense of ‘cognition’ (Erkenntnis) that Kant highlights in the B edition of the First Critique, which means the same as objectively valid judgement (Bxxvi, Bxxvi n.), as opposed to the wider definition of ‘cognition’ (Erkenntnis) that he had used in the A- or 1781 edition, which means the same as conscious objective representation (A320/ B376–7). But from this it does not follow that there cannot be ‘empty’ concepts or ‘blind’ intuitions outside the special context of objectively valid judgements. ‘Empty concept’ for Kant does not mean either bogus concept or wholly meaningless concept: rather it means concept that is not empirically meaningful or objectively valid, and for Kant there can be very different sorts of concepts that are not objectively valid, including rationally intelligible concepts of noumenal objects or noumenal subjects, which are meaningful, or informative, only in a thin sense that implies at least bare logically self-consistent conceivability, although not in a thick sense that also implies empirical meaningfulness. (Hanna 2016, 48) The Kantian non-conceptualist strategy has several further ramifications: In particular, (1) it refers to those passages from the Transcendental Deduction which claim that no intervention of the intellect is needed to ensure that phenomenal objects are given in the intuition,4 (2) it distinguishes between figurative and intellectual synthesis, and (3) it asserts that the intuition can be based on a non-conceptual activity of synthesis.

On de se and de re  151 For example, Allais (2009) draws attention to the passages focusing on the threefold synthesis in the first edition of the Transcendental Deduction— unlike apprehension and reproduction, only recognition involves the conceptual dimension. Secondly, Allais refers to A78/B103, in which a distinction between the synthesis of the imagination and the functions of the intellect is established. Accordingly, Allais (2009) argues for the attribution of the relative non-conceptual content to Kant by rejecting a strong conceptualist approach in line with McDowell—for whom the intuition does not even make a notionally separable contribution to cognition.5 For Allais, intuition can be assumed to provide a separate, perceptual presentation of spatiotemporally located, mind-independent entities—be they objects or empirical details—making at least a notionally separable representational contribution. It follows that a subject can have a perceptual representation with a content without possessing any concept to describe that content: While . . . what perceptual states a subject might be in does not depend on what concepts she possesses, this need not mean denying that our perceptual states are brought under concepts, and that experience, for us, typically is an actualization of conceptual capacities in sensory consciousness itself. Rather, what is denied is that experience is representational only to the extent that it is brought under concepts. Applying this to Kant, the idea is not that we need deny that our intuitions are brought under concepts (and must be, if we are to cognize an objective world), but that they need to be brought under concepts in order to present us with particulars. (Allais 2009, 386) As indicated above, this interpretative scenario is essentially based on a distinction between the synthesis of the imagination and the functions of the intellect; the matter of the different interpretations of the passages suggested in the debate will not be touched on here if not for a point. Several commentators (see, e.g., Ginsborg 2008; Schulting 2012a) have rejected this non-conceptualist reading by highlighting the passages in which the synthetic activity is attributed to the intellect (e.g., B129) and the spontaneity of the imagination is parallel to that of the intellect (e.g., B162n). For instance, according to Schulting (2017, 221) the view that “synthesis per se” is not “governed” by the categories “must be considered mistaken”. The author points out Kant’s assertion in one of his Reflexionen from the 1780s; in this passage, transcendental synthesis of the imagination is said to involve “a concept of the object in general”: The transcendental synthesis of the imagination pertains solely to the unity of apperception in the synthesis of the manifold in general through the imagination. Through that a concept of the object in general is conceived in accordance with the different kinds of transcendental synthesis. (LBl B 12, 23:18, 258)

152 On de se and de re Schulting (2017, 221) holds that synthesis and categories cannot come separate for their intimate relation; “the set of categories just is the set of rules for synthesis”, a point that is further made clear by the so-called guiding thread (Leitfaden) passage (B104–5/A79) (cf. supra, § 1.5.1) and in these Reflexionen: Now the categories are nothing other than the representations of something (appearance) in general so far as it is represented through transcendental synthesis of imagination. . . . The manifold, however, cannot thoroughly belong to one apperception except by means of a thoroughgoing synthesis of imagination and its functions in one consciousness. This transcendental unity in the synthesis of imagination is thus an a priori unity under which all appearances must stand. Those [i.e., den Functionen derselben = the functions of the synthesis of imagination] however are the categories, thus the categories express the necessary unity of apperception under which all appearances belong insofar as they belong to one cognition a priori and necessarily. (LBl B 12, 23:19, 259) The passages discussed above, which make up the so-called transcendental object theory, certainly suggest new points to be addressed by those scholars who aim to investigate Kantian non-conceptualism. As seen in the previous chapter, it is one thing to establish that the intuitive representation makes a peculiar, autonomous (as to conceptual forms) contribution to content, and quite another to maintain that this contribution enables an object to be perceived regardless of any conceptual articulation through an autonomous, epistemic function of de re presentation assigned to the intuitions. For this reason, my perspective might be referred to as weak conceptualism: In contrast to McDowell’s strong conceptualism, here intuition makes a notionally separable representational contribution to cognition and does so by making reference possible in the first place (cf. supra, Chapter 4). In contrast to Allais’s non-conceptualism, this epistemic contribution cannot be realised without at least the concept of the transcendental object. “The Object Must Be Thought of Only as Something in General = x” To delve even deeper into this view, it is necessary to consider the expression “an object of representations” (cf. A 104). Appearances are nothing more than sensible representations—they are not to be regarded as objects in themselves, outside of the power of representation. Hence, the object has to be considered as distinct from cognition: “This object must be thought of only as something in general = X, since outside of our cognition we have nothing that we could set over against this cognition as corresponding to it”.

On de se and de re  153 The subject cannot be representationally (or perceptually) presented with a particular/object if it does not employ the concept of the transcendental object understood as a thought concerning the relationship between object and cognition. Since “our thought of the relation of all cognition to its object carries something of necessity with it”, this concept is a necessary condition to determine a de re link with an object in the spatio-temporal, indexical, intuitive device. It is worthwhile to quote the full passage A104–5: What does one mean, then, if one speaks of an object corresponding to and therefore also distinct from the cognition? It is easy to see that this object must be thought of only as something in general = X, since outside of our cognition we have nothing that we could set over against this cognition as corresponding to it. We find, however, that our thought of the relation of all cognition to its object carries something of necessity with it, since namely the latter is regarded as that which is opposed to our cognitions being determined at pleasure or arbitrarily rather than being determined a priori, since 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. (A104–5) If so, perception seems to be no mere intuitive event. For Allais (2009, 386), a subject can perceptually represent those particular aspects of things that can be included in concepts even if she does not possess the relevant concepts, e.g., a subject can perceive a round, red particular without the concepts of roundness, redness, and particularity. On the contrary, in my view, the non-conceptual, intuitive content is not assumed to exist without at least one transcendental conceptual dimension through a de dicto act and its relative concept of a transcendental object. The subject can perceive a round, red particular without the concepts of roundness, redness and particularity, but cannot ascribe them to an object regardless of the employment of the concept of the transcendental object: The pure concept of this transcendental object (which in all of our cognitions is really always one and the same = X) is that which in all of our empirical concepts in general can provide relation to an object, i.e., objective reality. (A109) In A250, Kant seems to be even more explicit about such a descriptivist result. On one hand, all representations are related to some object through the understanding; since appearances are nothing but representations, the understanding relates them to the object of sensible intuition, i.e., the transcendental object: This signifies, however a something = X, of which we know nothing at all nor can know anything in general (in accordance with the current

154 On de se and de re constitution of our understanding), but is rather something that can serve only as a correlate of the unity of apperception for the unity of the manifold in sensible intuition, by means of which the understanding unifies that in the concept of an object. On the other hand, and this is the point at issue, “this transcendental object cannot even be separated from the sensible data, for then nothing would remain through which it would be thought”. However, as it is not an object of cognition in itself, the transcendental object is “only the representation of appearances under the concept of an object in general, which is determinable through the manifold of those appearances”. It follows that the very appearances cannot subsist without the employment of the representation or concept of an object in general, to the extent that the categories do not represent any special object given to the understanding, “but rather serve only to determine the transcendental object (the concept of something in general) through that which is given in sensibility, in order thereby to cognize appearances empirically under concepts of objects”. It seems that the subject can be perceptually presented with a particular/ object regardless of the application of empirical concepts – and, probably, of pure concepts, too – provided that the non-conceptual content theorists’ arguments about the Transcendental Deduction are accurate (cf. supra); in any event, this cannot possibly be the case without at least the concept of the transcendental object. To return to the example of the savage perceiving a house, the savage and an urbanised subject will share the same type of intuitive representation, but not the same conceptual dimension from a strictly empirical standpoint: Both of them must employ the concept of the transcendental object—i.e., the concept of an object in general—as a condition of possibility for “the representation of appearances”. Following this, the object is likely to be attributed a number of specific properties in the form of a judgment through the synthesis of the manifold, provided that the relevant empirical concepts are available. In conclusion, the concept of a transcendental object allows the thought of the phenomenal object presented by the intuition in a non-descriptive mode and at a spatio-temporal level. Hence, the Kantian de re proposal appears to be essentially descriptivist.

De se Thoughts Between Descriptivism and Singularism The following is a summary of the key arguments that have emerged thus far: The transcendental subject is regarded to be just as ‘formal’ as that of ‘transcendental object’. The analysis in the previous pages has shown that the thinking subject has intellectual consciousness of itself, just as it has consciousness of the existence of the being itself (Das Wesen selbst); this is regarded as an indeterminate aliquid (ein Etwas), an object that cannot be acquired as cognition in the same way as phenomena. Obviously, it cannot be acquired as a noumenon, either; otherwise, the thinking subject might

On de se and de re  155 intuit itself through apperception. For these reasons, Kant concludes that, “in the synthetic original unity of apperception, I am conscious of myself not as I appear to myself, nor as I am in myself, but only that I am. This representation is a thinking, not an intuiting” (B157). “The consciousness of myself in the representation I is no intuition at all, but a merely intellectual representation of the self-activity of a thinking subject” (B278). To think of this something as an indeterminate object (which is not known in the same way as phenomena), Kant introduces the concept of a transcendental subject that must be related to the representation I to play the same role as the concept of the transcendental object. As seen in Chapter 2, despite this functional equivalence, the two notions display remarkably different characterisations: The I in I think remains a representational unicum. Based on the distinction between noumenon and phenomenon, the object has an existence in itself, and is presented as a phenomenon through intuition. As mentioned previously, in order to think of an object in itself, one must employ the indeterminate thought that something in general (= x) exists in itself and appears through the intuition (A104). Considered in terms of its function, the concept of a general or transcendental object (= x) is the indeterminate thought of a single object having an existence in itself. On the other hand, when Kant addresses the “Thinking, taken in itself”, he states that ‘Thinking’ is the logical function as well as sheer spontaneity, even if no subject of consciousness is presented. The subject thinks itself in the same way that it thinks an object in general: “In this way I represent myself to myself neither as I am nor as I appear to myself, but rather I think myself only as I do every object in general from whose kind of intuition I abstract” (B429). The concept of a transcendental subject (= x) is the indeterminate thought of a thinking being having an existence in itself. Once again, the famous passage A346/B404 introduces the concept of ‘a transcendental subject’: When he rejects the possibility of a science of pure reason concerning the nature of the thinking being, Kant states: We can place nothing but the simple and in content for itself wholly empty representation I, of which one cannot even say that it is a concept, but a mere consciousness that accompanies every concept. Through this I, or He, or It (the thing), which thinks, nothing further is represented than a transcendental subject of thoughts = x which is recognized only through the thoughts that are its predicates, and about which, in abstraction, we can never have even the least concept. (A346/B404) The thinking being represents itself as a transcendental subject = x through the completely empty representation I. Nonetheless, the concept of a transcendental subject is different from the concept of a transcendental object in more than one respect. The transcendental object is introduced in order to address the problem of objectivity, i.e., the issue of the ‘immanentization’ of cognition (Allison 2004, 60). It is impossible to stand outside representations

156 On de se and de re and compare them with transcendentally real entities; consequently, the concept of a transcendental object, regarded as a sort of transcendental pointer, “serves to define the philosophical task by indicating that the commonsensical and transcendentally realistic concern with the ‘real’ nature of objects must be replaced by a critical analysis of the conditions of the representation of an object”. As indicated previously, the subject cannot establish a representational relation with a particular/object without the use of the concept of the transcendental object understood as a thought of something in general = x (A104–5). If “our thought of the relation of all cognition to its object carries something of necessity with it”, then this concept is a necessary condition in order to determine a de re link with an object. As seen in Chapter 2, the concept of the transcendental subject can also be regarded as a sort of transcendental pointer; however, since no intuition is given here, the concept of a transcendental subject enables no determination of a spatio-temporally singular object through the unification of the conceptual dimension. Even though the I can be employed as a pointer to conceive of the difference between the representation and that which is represented, it cannot determine any experiential object: Nothing is given, nor can it be thought of, but the entirely empty representation I. In other words, while the concept of a general or transcendental object (= x) is the indeterminate thought of a single object having an existence in itself, the concept of a transcendental subject (= x) is the indeterminate thought not of a single object, but of an existence in itself. In this way, the I and the concept of transcendental subject can only be used to think of the difference between the representation and that which is represented—the thinking being—but cannot be employed to determine the subject as an object: The subject cannot be representationally or perceptually presented with itself as a particular/object. As contended in the preceding chapters, the I of apperception is not a representation that functions in a de re way, but one that singles out the most fundamental kind of de se thought. Thus, now it is necessary to articulate the different kinds of de se thoughts. Taking up Frege’s (1918–19, 25–6) idea that “every one is presented to himself in a particular and primitive way, in which he is presented to no one else”, Perry (2000) and Recanati (2009, 255) describe de se thoughts as special cases of de re thoughts. A de se thought is nothing more than a de re thought, the object of which is the same thinking subject that has produced the thought; this type of thought rests on a special relationship of acquaintance between the subject and itself through a first-person mode of presentation, referred to as the EGO mode of presentation by Recanati and as self files—the mental particulars whereby the subject acquires firstperson information—by Perry. It goes without saying that one may also formulate accidental de se thoughts, namely de re thoughts about oneself, without resorting to a first-person mode of presentation and without realising that the subject to whom the thought is referring is nobody but oneself. Castañeda, Kaplan, and Perry’s examples are well known, particularly the

On de se and de re  157 forgetful painter who judges the painter of a particular picture very accomplished without realising that he painted it himself (Castañeda 1999, 256). In this way, if de re thoughts require the res being thought about and the way—the mode of presentation—in which the res is thought about, then de re thoughts can generate ‘Frege cases’ (Fodor 1995), that is, those instances in which the subject does not realise that two distinct modes of presentation concern the same object and determine it as his or her own referent. To recall Kaplan’s example of a subject who thinks the thought “His pants are on fire” while looking at himself in the mirror, the subject entertains a de re thought about himself under the mode of presentation ‘that man (in the mirror)’, but he does not realise that that particular thought pertains to nobody but himself. In other words, he does not recognise ‘That man = myself’; therefore, and without irrationality, the subject can simultaneously assent to ‘I am F’ and to ‘That man is not F’. Through an example offered by Wittgenstein (1958, 67), the opposite case also exists: “It is possible that, say in an accident, I should feel a pain in my arm, see a broken arm at my side, and think it is mine, while in fact it is my neighbour’s”. In this example, the subject wrongly believes ‘that man = myself’; hence, he wrongly self-ascribes property F (to have the broken arm) based on the connection between two thoughts. The first, “My arm hurts”, is a judgment based on the subject’s feeling of pain through proprioception; the second thought, “that arm is broken”, is a demonstrative judgment based upon the subject’s visual perception. Through these two thoughts, and through the wrong identification of ‘that arm = my arm’, which is grounded in the visual demonstrative ‘that arm’ used in order to determine the subject’s ‘arm that hurts’ as the referent, the subject reaches the false judgment “my arm is broken”. According to Recanati, (1) de re thoughts may concern (1.1) someone other than oneself or (1.2) oneself. In the latter case, i.e., with de se thoughts, these can be either (1.2.1) accidental or (1.2.2) genuine, also widely referred to as I-thoughts (cf. schema 1). (Schema 1) (1)  De re Thought (1.1) About something other than oneself

(1.2) About oneself (1.2.1) Accidentally de se

(1.2.2) First-person thoughts

On the other hand, (1.2.2) genuine de se thoughts can be explicit and grounded in an identification component, once again explicitly represented in thought, as in Kaplan’s reversed example of a subject who identifies himself in a mirror with the man whose pants are on fire. De se thoughts may also be implicit and based on identification-free self-reference. As such, they

158 On de se and de re are immune to error through misidentification relative to the first person, as opposed to thoughts involving some explicit self-identification. Several authors from Shoemaker onward have observed that, in this particular type of de se thought, the subject is not represented as a constituent in the content of thought but rather serves as a circumstance of evaluation for the judgment in question. In other words, a de se thought does not express a complete proposition ascribing a property to the thinking subject, but simply arises from a representational content expressing the instantiation of that property. This is to say that thinking of such a content also implies the self-ascription of the property expressed by the content of the de se thought (Lewis 1979; Chisholm 1979). As shown by Wright (1998, 19), it is the peculiar epistemic ground of the first-person mode of presentation that assigns these features: For example, a judgment such as “I have pain” expresses an implicit de se thought because the way the property expressed by the predicate is instantiated (“there is pain”)—based on one’s own subjective experience focused on introspective consciousness—will be sufficient to realise that the predicate is ascribed to oneself even if the subject is not represented as a constituent in the content of the judgment. It is in this sense that “there is pain” is tantamount to “I have pain”: Identification-freedom, in the strong sense, characterizes thoughts that are ‘implicitly’ de se as opposed to thoughts that involve an explicit self-identification. Thoughts that are implicitly de se involve no reference to the self at the level of content (Perry 1986b): what makes them de se is simply the fact that the content of the thought is evaluated with respect to the thinking subject. The subject serves as ‘circumstance of evaluation’ for the judgment, rather than being a constituent of content. Or, to put it in slightly different terms, in such cases the content of the thought is not a complete proposition ascribing a certain property to an object (viz. the subject himself/herself): the content is the property, but to think the thought—or to think it in the relevant mode—is, for the subject, to self-ascribe that property (Loar 1976, 358; Lewis 1979; Chisholm 1979, 1981). In this framework the content of the subject’s self-ascription of pain is something like ‘pain’ or ‘there is pain’ or ‘pain is being experienced’. Being gained from inside (through introspective awareness), the information that pain is being experienced necessarily concerns pain that the subject himself or herself is experiencing. The judgment is about the subject not because the subject is represented in the content of the judgement, but because the experiential basis of the judgment determines its domain of application. (Recanati 2009, 259) When Recanati mentions the “experiential basis of the judgment”, what he is actually referring to is the mode of presentation of the thought’s content. For instance, recalling Evans’ example, when a subject knows that her legs are crossed through proprioception, she gains this information “from inside”: It is

On de se and de re  159 the proprioceptive mode that enables such a state to represent the position of the limbs as the subject’s limbs and no one else’s. Nobody can gain information about the position of other people’s limbs through his or her own proprioceptive mode. That being so, a subject can self-ascribe a bodily property such as one’s legs being crossed and produce that thought via this particular mode of presentation, which guarantees the thought’s self-ascription to the subject without explicitly representing her in the content of the thought. Instead, when the subject looks in the mirror and acknowledges that her legs are crossed, she explicitly thinks of herself as the person whose legs are crossed. Accordingly, not all de se are also de re thoughts, as this only applies to explicit de se thoughts; on the contrary, implicit de se thoughts are not included in the class of de re thoughts because the subject is not a res that can be made the object of a representational relation. The subject is not represented in the de se thought because this only applies to the properties instantiated in the experiential dimension:6 Implicit de se thoughts are not [de re thoughts]. Their content is thetic, while the content of de re thoughts is categoric. Admittedly, explicit de se thoughts are a sort of de re thought: they are de re thoughts that involve a special mode of presentation of the res thought about, namely the ego mode of presentation. But implicit de se thoughts are not: their content corresponds to that of a predicate, and the subject of which it is predicated remains implicit. (Recanati 2009, 259) By way of summary, three types of de se thoughts are the issue at the moment (cf. schema 2): (1) accidental de se thoughts, which are de re thoughts about a particular subject who happens to be herself. Among the ‘essential’ or ‘genuine’ de se thoughts through which the subject thinks of herself as herself by using the first person, one can distinguish: (2) explicit de se thoughts, in which the subject is represented in the thought’s content in the same way as in de re thoughts,7 and (3) implicit de se thoughts, the content of which does not contain any representational correlate of the subject, making them immune to error through misidentification. As will be seen below, a more fundamental typology of implicit de se should also be taken into account. (Schema 2) De re Thoughts About something other than oneself

First-person thoughts

About oneself Accidentally de se

Explicitly de se Vulnerable to EM

Implicitly de se Immune to EM (error through misidentification)

160 On de se and de re

Kant on de se Bearing in mind the different philosophical approaches, one does not seem to be able to find any explicit articulation of de se thoughts in the transcendental system; starting with the faculty analysis, however, one may certainly reconstruct Kant’s stance on the different types of de se thoughts. There is no doubt that the de se thoughts perspective discussed thus far is particularly attractive: certain features of transcendental apperception and I think seem to anticipate certain points of this approach, with particular reference to a basic typology of implicit de se thoughts that has not yet been introduced in the present discussion, as will be explained shortly. I Think and de se Thoughts Recalling the excerpts discussed in Chapter 1, Kant introduces the principle of the transcendental unity of apperception in the famous passage B131–2: The I think must be able to accompany all my representations; for otherwise something would be represented in me that could not be thought at all, which is as much as to say that the representation would either be impossible or else at least would be nothing for me. That representation that can be given prior to all thinking is called intuition. Thus all manifold of intuition has a necessary relation to the I think in the same subject in which this manifold is to be encountered. (B131–2) This fragment has already been seen to condense three important points: (1) in order to represent something to a subject, the subject must be able to think that every representation is its own. At first glance, the point is not that de facto representations are accompanied by I think, but the necessity of the possibility of the representations’ self-ascriptions, (2) the second point consists of the complex thought based on synthetic unity: The different representations merge into one single consciousness as a thought ascribable to a subject, i.e., (T) I think that (I think that x, I think that y, I think that z, etc.). The several uses of I are co-referential—the I thinking x is identical to the I thinking y, and so forth—and the identity of the I think also concerns the higher-ranking I think upon which depends the synthesis of various representations in a single complex thought, (3) taken together, these two points set up the necessary synthetic unity of apperception, whereby the representations of a complex thought are connected in such a way that they are linked to a single thinking subject, which ultimately entails that they make up a synthetic unity. As highlighted several times throughout this work, Kant asserts that the

On de se and de re  161 analytical unity of apperception presupposes synthetic unity: The consciousness of the I think identity requires not only a synthesis, but also a consciousness of the synthesis identity. With regard to deduction, there are two relationships. The first is of synthesis, representations, and object, whereas the second concerns apperception, judgment, and intellect, specifying the categories as conditions for the possibility of knowledge. If the intellect is the faculty of knowledge through concepts, then the object is given through the synthesis of the manifold in intuition. The merging of the representations is based on a conceptual synthesis, and it cannot exist without the unity of apperception, which refers them to a given object in order to establish their objective validity. The synthesis of the representations is also linked to the faculty of judgment and is articulated according to the distinction between the objective and the subjective unity of consciousness. The objective unity of consciousness presupposes the use of categories; consequently, the manifold given in intuition is unified in the concept of an object. In particular, it is the copula that specifies the relationship of judgment as that which links representations to the objective unity of apperception so as to convert them into objective knowledge. The act of judgment is the very condition of synthesis. Thus, the function of the intellect that links the representational manifold to the unity of apperception is specified by the logical function of judgments determining the manifold in empirical intuition. Categories are functions of judgments determining the manifold; thus, the manifold is linked to the synthetic unity within one single consciousness. As the above pages have attempted to explain, I think has mainly been regarded as the formal condition for all thinking: I is the representational correlate of the thinking being in the self-consciousness; as such, it designates an existence devoid of properties because it is analytically contained in the synthetic unity of apperception as the representation of (the concept of) ‘a something in general’ or ‘a transcendental subject’. If the awareness of intellectual self-existence summarised by I think does not display any property, then it is not possible to determine whether that something exists as a persistent substance to produce knowledge due to the absence of intuition: The thinking being is merely represented as a something general, unidentifiable from an epistemic perspective. All the subject knows is that it exists as a thinking activity. A few peculiarities concerning the self-referential apparatus involved in the transcendental apperception have been revealed. The act of reference performed by the subject to refer to itself entails no mediation of knowing; in other words, the notion of transcendental designation involves no identification through the properties ascribable to the subject. At this level, the intellectual representation I is a simple representation bearing no content, and it only designates transcendentally; that is, without any conceptual mediation.

162 On de se and de re In summary, the general correlate of apperception I cannot be used to constitute a de re thought: The thinking being, seen as ‘this I, or He, or It (the thing), which thinks’ is not a res that can be made the object of a representational relationship. While I think is the condition for every thought, I is not an explicit representational reference within the content of a judgment: It only designates transcendentally and, as such, involves neither conceptual nor intuitional mediation. Two related issues arise with regard to this initial result. The first concerns the relationship between I think and the articulation of de se thoughts. If the act of spontaneity expressed by I think is necessarily involved in the making of a judgment, then it must reside in a much more abstract level than the reflection on the distinction between implicit and explicit de se thoughts. As remarked in Chapter 3, the Kantian arguments concerning I think cannot articulate the different types of de se thoughts expressing self-ascriptions of mental and physical properties: These consider the form and condition of the possibility for any kind of judgment, regardless of the particular uses of I (as subject or object, in Wittgensteinian terms) involved in the single judgments produced. In other words, the I think mechanism of transcendental designation cannot account for the presence or absence of the representational reference of the subject in judgments such as “I have grown six inches”—an explicit de se thought—or “I have a toothache”—an implicit de se thought—since such is the condition of possibility for both. Things are somewhat different with a more fundamental, implicit de se thought. As we have seen, Recanati (2007, 147) distinguishes between implicit and explicit self-ascriptions according to the distinction between the mode of representation and the content of representation. The former concerns all the information gained through the proprioceptive/kinaesthetic mode: On the basis of his proprioceptive/kinaesthetic experience, if the subject judges his legs to be crossed, then he cannot be mistaken as to whom he ascribes the property of having his or her legs crossed. If the judgment is not made ‘from the inside’ and relies on the perception of the subject’s body in the mirror, the subject will see that his legs are crossed and may misidentify the person whose legs are crossed. In this latter case, the subject is explicitly represented in the content of the representation, whereas in the former he is not represented but is implicitly determined by the mode; for this reason, the statement is ‘identification-free’. Despite this difference, the author claims that the self is always an unarticulated constituent, while every (implicit or explicit) self-ascription presupposes, in turn, a specific or more basic implicit self-ascription:8 [In the mirror case], the perception is (explicitly) about myself because it is my legs which, on the basis of my visual experience, I judge to be crossed. But the perception is (implicitly) about myself also in the sense that I am the one who sees the mirror and what it shows. That the perception is mine and concerns my surroundings rather than someone

On de se and de re  163 else’s, is something which is guaranteed by the architecture of the system, hence gives rise to immunity. So, in the mirror case, the perceptual representation I entertain is such that one property (the property of seeing in the mirror that my legs are crossed) is ascribed to myself implicitly, in virtue of the perceptual mode of representation, while another property (the property of having one’s legs crossed) is ascribed to myself explicitly, in virtue of the content of the perceptual judgment. In the internal case, I self-ascribe both properties (or their counterparts) implicitly, in virtue of the internal mode of representation. What is represented is only the property of having one’s legs crossed: its ascription to myself is not an aspect of the (explicit) content of the representation, but something that follows from the mode of representation. (Recanati 2007, 147) Perry (1986a, 360) expresses the same idea when he claims that the subject of every judgment is regarded as an unarticulated constituent:9 The information that we get at a certain spot in the world is information about objects in the neighborhood of that spot in a form suitable for the person in that spot. As long as this is the only source of information we have about ourselves, we need no way of designating ourselves, indexical or insensitive. Our entire perceptual and doxastic structure provides us with a way of believing about ourselves, without any expression for ourselves. (Schema 3) Basic implicit self-ascription De re Thoughts About something other than oneself

First-person thoughts

About oneself Accidentally de se

Explicitly de se Vulnerable to EM

Implicitly de se Immune to EM (error through misidentification)

Schema 3 summarises the articulation of the thoughts. To recall the above-mentioned example offered by Recanati, if a subject is looking in the mirror and recognises her legs as being crossed, then the thought is explicit de se because the subject features as a constituent in the content. On the other hand, the perceptual mental state involved should also be considered to be an implicit de se thought regarding its basic, implicit self-ascription, as the subject in question is the one who sees the mirror and what it shows. This is not represented in the thought’s content; instead, it is the architecture

164 On de se and de re of the mind’s system, i.e., the mode of representation, which ensures that its perception belongs to the subject in question and not to someone else. In the case of the mirror, two properties in the same perceptual representation are self-ascribed by the subject, one being the property of ‘seeing in the mirror that her legs are crossed’, which is implicitly ascribed to the subject due to the perceptual mode of representation, and the other being the property of ‘having one’s legs crossed’, which is explicitly ascribed to the subject due to the fact that she is represented in the content of her own perceptual thought. One should distinguish between a first level, in which the subjective dimension producing a thought is not represented as the producer of that thought, and a second level, in which the subject can be implicitly or explicitly represented in the content of the thought as the subject to whom a given property is attributed. Perry and Recanati’s relativist perspective articulates such two levels in compliance with specific cognitive constraints and discards any explicit representation of the self to eventually reach Perry’s paradox, whereby some de se attitudes should be regarded as selfless. I as Form of Thinking On the contrary, and yet with the same result, for Kant, the role of the I, seen as a correlative of apperception, is preserved as the basis of synthesis. Kant’s perspective describes the two levels according to a distinction between two classes of self-ascriptions, i.e., those taking ‘I as passive’ and those taking ‘I as active’ (Refl 208, 20: 270). Following Carl’s (1998, 157) epistemological perspective, which maps this distinction onto one between spontaneity and receptivity, while the former self-ascriptions are determined by representations given in sensibility in independent ways, the ‘I as active’ is ‘the logical I’ (KGS: 20, 270), the bearer and ground of all judgments, the referent of all mental self-ascriptions expressing the act of judging one’s given representations. This point was already discussed in Chapter 3, when Kant’s distinction between ‘I as subject’ and ‘I as object’ was investigated with reference to the problem of the lack of identification of the subject; this was done starting from the epistemological perspective focusing on the distinction between spontaneity and receptivity, seen as the conditions for all possible knowledge, to provide an account that incorporates the ‘I as subject’ of apperception into the foundation of the formal conditions of knowledge. As a result, from the Kantian perspective, the scenario is the following (cf. schema 4). (1) The first level of basic, implicit self-attribution concerns the act of spontaneity—the Thinking—for it is the synthetic unity of apperception that is the origin of the representational synthesis and hence the source of the self-attribution of all thoughts. As has been pointed out, the act of spontaneity expressed by I think is necessarily involved in the making of any judgment: At this first level, the lack of an identification component

On de se and de re  165 entailed in the transcendental designation of I think and the resulting implicitness feature of the self-ascription of all thoughts are grounded in the Thinking features regarded as the synthetic unity of apperception that determines the I think (qua Representation) as an analytical unity of apperception. In other words, the subjective component is not represented explicitly: it is not a res that can enter the thought’s content as a constituent, and it is implicitly represented because the I is only given as the analytical form of thought. Kant accounts for the implicitness of self-ascription by pointing out that the I is the bare and empty form that indicates the nexus established by the copula in the judgment through the representational synthesis of the unity of apperception. If the I of thinking, which means pure apperception, is the condition for the possibility of all thoughts (de re and de se), then the first level of spontaneity also regards the problem of I-thoughts and the condition of possibility of the different Wittgensteinian uses of the I as subject and I as object in judgments expressing the self-ascriptions of mental and physical properties. (2) With regard to the second level concerning the empirical apperception, Recanati’s distinction between the mode of representation and the content of representation can be mapped onto the articulation between inner and outer sense with two possible outcomes: (2.1) if the subject only reveals itself in time, i.e., if the mode of representation of inner sense is all that is involved, then its selfattributions only concern mental properties. There is no problem with identification, and the I is used as the subject in judgments of perception that can be regarded as implicit de se thoughts, and (2.2) if the subject reveals itself in time and space, i.e., if both inner and outer senses are involved, then the identification problem will arise due to the presence of the explicit self-attribution of a bodily physical property relative to I, used as the object in judgments of experience, which can consequently be regarded as explicit de se thoughts. (Schema 4) Thinking as spontaneity De re Thoughts About something other than oneself

First-person thoughts

About oneself Accidentally de se

Explicitly de se

Inner and Outer Sense Receptivity

Implicitly de se Inner Sense

166 On de se and de re At this point, a second issue concerning the specific and, to some extent, controversial nature of the representational reference of I in judgments arises. While I think must be able to accompany every representation, I only designates ‘this I, or He, or It (the thing) that thinks’ transcendentally, without involving any explicit egological representational reference in the judgment. This entails holding the anti-reflexive character of self-consciousness, which Sturma (1985) and Ameriks (1995, 1997, 2000)—although under different circumstances—have used to argue against Henrich’s (1966) Fichtian interpretation of apperception based on the Reflexion Theory (cf. Schulting 2017b). In his conception of self-consciousness, Kant addresses the issue of the reflexive and representational model—already traceable in his predecessors, and particularly in Locke10—which follows the structure of perceptual consciousness and, more generally, of a model of knowledge built on the subject–object scheme: The awareness of an object is always mediated by the object’s representation, implying a representing subject and a represented object. In the shift from consciousness to self-consciousness, the model structure remains unaltered, whereas the subject and the object of consciousness coincide: The subject of the representation represents itself, which entails its inward withdrawal. The subject is objectified and, in the process, splits into two poles. However, by saying that the subject comes into contact with itself by taking a step back towards itself (so to speak) in order to be objectified, we enter a vicious circle in which we are merely assuming what we actually aim to explain: For we can only speak of an ‘I’ if a subject knows itself, that is to say if the self addresses itself as the self. In this case the subjective self, the self that was meant to be our starting point, is not the only self we have in mind. We are already taking it into account in the sense that it is also its own object of investigation. So if one wants to explain the phenomenon of self-consciousness in its entirety using, as a starting point, the self as subject and its reflection upon itself, then one presupposes a priori what one wants to explicate: one is thematizing the reflecting subjective self as identical with the reflected-upon self. One does not just presuppose the self but the self that conceives of itself as identical with itself. (Henrich 1982, 63; cit. in Freundlieb 2003, 43) Difficulties arise because the subject of reflection cannot identify the object of its self-conscious reflection act as itself unless it is assumed that it already possesses some form of contact and self-knowledge before the reflexive act. In other words, the reflexive model assumes that the subject acquires knowledge of itself by connecting with itself—in a way, by retreating into itself—but in some way the subject must already distinguish its object of reflection in order to establish this connection and assume what one would like to explain. In short, this concept of self-consciousness does

On de se and de re  167 not stand up if its assumption is that the subject is in a connection with itself as if it were standing in front of the mirror: To relate to itself, it should already be able to recognise itself in the reflected image and should already be in some form of contact with itself before, and independently of, any act of reflection. If the reflexive model is held, then any result is equally undesirable, both if what one wants to explicate is assumed and if the very possibility of self-consciousness is dismissed:11 If their relation is interpreted via reflection and thus as an achievement through which the act of reflection becomes conscious of itself, then the subject of the act must either already be the Self, or the equation “I = I” will never hold. If the Subject-Self is not the Self, then neither can the Self, of which we come to have knowledge, that is, the Object-Self, ever be identical with it. Thus, the reflection theory of self-consciousness either presupposes the phenomenon of Self without clarifying it, or totally invalidates it. (Henrich 1966, 21) Actually, Kant does not encounter the paradoxical consequences triggered by the self-consciousness reflexive model: In fact, not only does Kant not seem to address any reflexive model at all but, as Henrich himself (1976) put it in different contexts—as it is also argued here, at least to some extent— the I think self-reference at the basis of the self-conscious capacity does not involve any kind of identification component. While it cannot constitute a de re thought, it assumes the implicit form of every thought. The theoretical issue can be summarised via what Cassam describes as the ‘elusiveness’ of the self and, as far as Kant is concerned, via the passage in which Horstmann (2010, 449) points out that “the self-conscious I can . . . never be an item of which I can be conscious as an object”. The subject is always conscious of the self qua subject, and yet it cannot catch and objectify itself qua conscious subject. McGinn probably offers one of the best accounts of the issue: When I think of myself that which thinks occurs as subject; thus I never become merely an object of my own apprehension. The self always, and systematically, steps out of cognitive reach. Even if the reflecting self and the self reflected upon are numerically identical, I can never stand back and apprehend this identity, since I shall always occur as a subject in my reflections, as well as an object. Qua subject I can never become an intentional object to myself. Yet it is qua subject that I have my essence. (McGinn 1993, 48) Qua subject, a subject of thought cannot think of itself as such in an objective way. If it tries to think of itself and of what is going on in its

168 On de se and de re own head, then the subject becomes the object of that thought; however, as Castañeda (1966, 64) remarks, “there is no object of experience that one could perceive as the self that is doing the perceiving”. In this way, the peculiar subjective perspective of the consciousness of the self qua subject escapes the objective take by definition. To employ an image offered by Ryle, thinking of oneself as subject is similar to the shadow of one’s own head: The I will not wait to be jumped upon. Emundts (2006, 306) condenses the issue in Kantian terms by remarking that the representation of the subject as something spontaneous cannot be seen as object-like: “If we wanted to determine the I as being present . . . the I would be determined as appearance. But then it is represented not as determining [das Bestimmende] but as determined [das Bestimmte]”. This specific point is made clear when Kant introduces two poles, namely the reflecting subject and the self that is reflected upon, as two ways of representing the subjective dimension. These will remain distinct, even though the subject is numerically identical. The issue has been developed previously, particularly in Chapters 1 and 3, through the self-affection argument, which, in the Transcendental Aesthetic B67–8 and in § 24 of the Transcendental Deduction, introduces the distinction between an active ‘I’ (of transcendental apperception) and a passive ‘I’ (of inner sense). Kant describes this topic as a paradox that must have struck everyone in the exposition of the form of inner sense (§ 6): namely how this presents even ourselves to consciousness only as we appear to ourselves, not as we are in ourselves, since we intuit ourselves only as we are internally affected, which seems to be contradictory, since we would have to relate to ourselves passively. (B152–3) As has been discussed, the paradox is resolved when the I that thinks and the I that intuits itself are simply regarded as two different representational dimensions of the same mind—the subject is given to itself as appearance in intuition as long as the combination or determination of inner sense through the spontaneity of Thinking is also given. The ‘I’ as intelligence cognises itself as an object being thought insofar as the subject is also given to itself in intuition, just as with any other appearance. Since this is nothing but two different ways of representing the subjective dimension, it follows that any attempt to identify them will prove to be a mistake: Qua subject, the subject is conscious of itself by means of the representation I of apperception, as “a merely intellectual representation of the self-activity of a thinking subject” (B278); “in the synthetic original unity of apperception, I am conscious of myself not as I appear to myself, nor as I am in myself, but only that I am. This representation is a thinking, not an intuiting” (B157). Instead, inner sense only presents the subject as it appears, “since we intuit ourselves only as we are internally affected”. As a result, considering that

On de se and de re  169 the representation I in I think is not an intuition, it will be impossible to either connect or identify the representational way of an active ‘I’ of transcendental apperception with the way the subject is given and represented in the intuition as a passive ‘I’ of inner sense. For Kant, this should not be regarded as a problematic point, as it only concerns the ways in which the subject represents itself to itself: But how the I that I think is to differ from the I that intuits itself (for I can represent other kinds of intuition as at least possible) and yet be identical with the latter as the same subject, how therefore I can say that I as intelligence and thinking subject cognize my self as an object that is thought, insofar as I am also given to myself in intuition, only, like other phenomena, not as I am for the understanding but rather as I appear to myself, this is no more and no less difficult than how I can be an object for myself in general and indeed one of intuition and inner perceptions. (B155) As a result, the reflexive model of self-consciousness is rejected and the relative theoretical problems avoided: The I of apperception can never be the representational vehicle employed to form an intentional object for oneself because the thinking being cannot determine any de re thought about itself; the I in I think does not involve any identification component—hence, the bare and empty features of the representation I, which explain the transcendental self-consciousness as an intellectual representation of the self-activity of a thinking subject, can be connected to the feature of implicitness introduced previously, i.e., the implicit form given by I, which is present in every thought. Ameriks (1997, 63) identifies this feature in the above-mentioned first level of ‘one’s own’, ‘not nothing for me’ representations: “Without any objective reflection on the self having taken place, the state [of awareness] is structured by the form ‘I think that x’, and therefore is already in a personal, even if implicit, sense an instance of ‘our’ consciousness”. Every thought is an instance of self-awareness in itself, even when no subjective dimension is involved. As for the upper level concerning the function of synthesis, I think—seen as the representation of an act of spontaneity—is “the bearer and ground of all judgments” which makes any thought possible. Moreover, I think is considered to be a simple representation: In the absence of any identifying epistemic mediation, it merely designates the activity of thinking transcendentally, namely as the nexus established by the copula in the judgment linking the representational synthesis on a conceptual basis to the synthetic unity of apperception. In conclusion, to think is to unify the manifold conceptually—every thought expressed by a judgment is necessarily based on the principle of transcendental apperception. For this reason, Kant holds that I is implicit in the concept of thought itself (B132, B134 n.), determining the form of every judgment in general terms (B406). As indicated previously, the I of

170 On de se and de re apperception—seen as a ‘logically simple subject’—is analytically contained in the concept of ‘thinking’ (B407–8). If I is the subject of thinking, and if it is not represented except in the form of the judgment established by the synthetic unity of apperception, in the specific terms of Transcendentalism every thought is an implicit de se thought. This feature of implicitness is connected to the limits of self-knowledge articulated between spontaneity and receptivity. Before addressing this question in the last paragraph, it is worth pointing out in the next paragraph how this feature of implicitness can avoid the problems of the so-called paradox of self-consciousness. The Paradox of Self-Consciousness and the Non-Conceptual Forms of Self-Awareness As seen previously, the capacity for self-consciousness depends on the possibility of producing I-thoughts that employ indexical self-references that are immune to error through misidentification relative to the term/concept I,12 and which have immediate implications for action. In this way, selfconscious thoughts (de se or I-thoughts) are thoughts that are typically expressed through the linguistic property of the first-person pronoun according to which “I” always refers to the person producing it; hence, it is natural to think that I-thoughts depend on the capacity for self-reference using “I”. In The Paradox of Self-Consciousness, Bermúdez (1998) points out that this guaranteed reference of “I” is connected to the immunity property in order to explain the nature of I-thoughts and, above all, to introduce a paradox that originates when self-consciousness and self-reference are considered to be interdependent. Bermúdez introduces the terms of the question whereby the paradox arises: (A) Once we have an account of what it is to be capable of thinking “I”thoughts we will have explained everything that is distinctive about selfconsciousness. (B) Once we have an account of what it is to be capable of thinking thoughts that are immune to error through misidentification we will have explained everything that is distinctive about the capacity to think “I”-thoughts. (C) Once we have explained what it is to master the semantics of the first-person pronoun (e.g., via mastery of some version of the token-reflexive rule that a given utterance of “I” always refers to the person uttering it), we will have explained everything that is distinctive about the capacity to think thoughts that are immune to error through misidentification. (Bermúdez 2018, 44) The problem with this deflationary account of self-consciousness (Bermúdez 1998, 2018) is that to master the semantics of the first-person pronoun correctly presupposes the capacity to think I-thoughts in a way that creates two forms of vicious circularity which, when combined, Bermúdez terms the paradox of self-consciousness.

On de se and de re  171 The first type of circularity is called explanatory circularity, and it arises because the capacity for self-conscious thought must be presupposed in any satisfactory account of mastery of the first-person pronoun: I cannot refer to myself as the producer of a given token of ‘I’ without, for example, knowing that I intend to refer to myself—which is itself a self-conscious thought of the type that we are trying to explain. (Bermúdez 2018, 43) In other words, if the semantic role of the first-person pronoun holds that “I” necessarily refers to the subject who has produced the sentence/thought containing it, the subject has to understand that “I” refers to herself. As Musholt (2015, 21) explains, “this understanding—‘When I use the firstperson pronoun, it always refers to myself’—is itself an ‘I’-thought”. Consequently, it’s not possible to explain the mastery of the first-person pronoun without taking the capacity to think I-thoughts into account. The second type of circularity is called capacity circularity: Due to the interdependence between self-consciousness and the semantics of the firstperson pronoun, it is not possible to explain how the capacity for selfconscious thought or for the linguistic mastery of the first-person pronoun arises in the normal course of human development, either ontogenetically or phylogenetically, since the ability to use the first-person pronoun is not present in new-borns or in non-human animals. For Bermúdez (2018, 43), it is not possible to meet the Acquisition Constraint, according to which, “if a given psychological capacity is psychologically real, then there must be an explanation of how it is possible for an individual in the normal course of human development to acquire that capacity”. In order to avoid the paradox of self-consciousness, several authors have carefully distinguished the forms of fully fledged self-consciousness, which depend on the mastery of the first-person concept and the linguistic mastery of the first-person pronoun, from the forms of primitive or non-conceptual self-consciousness that do not rely on any such linguistic or conceptual mastery. It is these non-conceptual forms of self-consciousness that allow the avoidance of both types of circularity. In other words, the ability to produce de se or I-thoughts must be independent of the mastery of the first-person pronoun and of the concept of possession in general; therefore, it is necessary to address certain kinds of non-conceptual representation in order to clarify the specific features of I-thoughts: These are thoughts that, although first-personal in the sense that their content is to be specified directly by means of the first-person pronoun and indirectly by means of the indirect reflexive pronoun ‘he*’, can nonetheless be correctly ascribed to creatures who have not mastered the first-person concept (as evinced in mastery of the first-person pronoun). (Bermúdez 1998, 45)

172 On de se and de re At this point, it is possible to distinguish two different type of accounts, namely self-representationalist and non-self-representationalist accounts of non-conceptual self-awareness (Musholt 2015, 65). According to selfrepresentationalist accounts – for example, cf. Bermúdez (1998), we can demonstrate the existence of non-conceptual forms of self-representation that (1) are self-referring, (2) have immediate implications for action, and (3) are immune to error through misidentification. These are primitive, nonconceptual forms of self-representation; in other words, the self is part of the representational content of experience. By taking Gibson’s ecological theory into account, Bermúdez (1998) develops one of the most important self-representationalist accounts of non-conceptual self-awareness. In particular, Bermúdez singles out several domains in which genuinely non-conceptual forms of self-consciousness can be found, thus providing the basis for more complex forms of “I”-thoughts. The most important are perceptual experience and somatic proprioception (bodily self-awareness). For the first, one of J. J. Gibson’s great insights in the study of visual perception was that the very structure of visual perception contains propriospecific information about the self, as well as exterospecific information about the distal environment. Visual perception incorporates a firstperson perspective in the information it picks up about the world. This is the basis of nonconceptual self-awareness, the foundation on which all forms of self-awareness are built. In his book The Senses Considered as Perceptual Systems, Gibson identifies what he calls the “fallacy of ascribing proprioception to proprioceptors”. This is the fallacy (he claims) of making a sharp distinction between outwardly directed exteroceptors (i.e., the five senses of sight, touch, smell, taste, and hearing) and inwardly directed proprioceptors (i.e., receptors in muscles, joints, and the inner ear providing information about bodily position and movement) and interoceptors (receptors in the visceral organs providing information about homeostatic states such as hunger and thirst). On the view that Gibson is attacking, information about the self can come only from proprioceptors and interoceptors. The job of the exteroceptors is solely to provide information about the external world. On Gibson’s view, each of the exteroceptive senses brings with it a distinctive type of proprioception. . . . What is particularly significant in Gibson’s analysis of visual proprioception (as compared to cutaneous or auditory proprioception) is his bold suggestion that visual perception involves direct perception of the self. (Bermúdez 2018, 49–50) Although visual perception is directed outward, the information received via perception necessarily also pertains to the perceiving subject itself, as the self is part of the content of visual perception due to the form of

On de se and de re  173 self-specifying information (among self-specifying structural invariants, there are the boundedness of the visual field and the occlusion of parts of the visual field by various bodily parts). The second domain is somatic proprioception, which concerns bodily self-awareness: Somatic proprioceptive information provides a way of registering the boundary between self and nonself. To appreciate this we need to note that there is an important variation among the information systems that provide information about the body. Some provide information solely about the body (e.g., the systems providing information about general fatigue and nutrition). The vestibular system, in contrast, is concerned with bodily balance and hence with the relation between the body and the environment. Other systems can be deployed to yield information either about the body or about the environment. Receptors in the hand sensitive to skin stretch, for example, can provide information about the hand’s shape and disposition at a time, or about the shape of small objects. Similarly, receptors in joints and muscles can yield information about how the relevant limbs are distributed in space, or, through haptic exploration, about the contours and shape of large objects. These latter information systems, underpinning the sense of touch, yield a direct sense of the limits of the body—and hence of the limits of the self. (Bermúdez 2018, 58) According to self-representationalist theorists of non-conceptual selfconsciousness, ecological perception and bodily experience provide selfreferring information to the subject; that is, information about itself, thus constituting primitive forms of self-consciousness. In fact, they deliver selfspecifying information that is relevant for action to the organism and are simultaneously immune to error through misidentification inasmuch as this self-specifying information is necessarily about the subject itself. As mentioned previously, the debate offers another, antipodal kind of account, which is the non-self-representationalist account of non-conceptual self-awareness; Musholt (2015) develops this by employing a few counterarguments to the self-representationalist approach. She points out two main reasons that self-representationalist theories are mistaken and suggests that non-self-representationalist, or “no-self”, theories are more suitable to address the problem of self-consciousness. Firstly, while self-representationalist theories hold that the self is explicitly represented in experience, for this perspective, the self has to be considered an “unarticulated constituent” of the content of experience: “positing that the self is explicitly represented in perception and bodily awareness puts an unnecessary and implausible cognitive burden on organisms capable of basic interactions with the environment” (Musholt 2015, 55). Moreover, if the self is part of the explicit content of perception and bodily awareness, not only is the phenomenology of perceptual and bodily experience misrepresented, but it creates an implicit

174 On de se and de re commitment to the subject–object model of self-consciousness which, as seen previously, seems to be too deeply flawed to account for the nature of self-consciousness. Instead, as an explicitly self-referring component seems to lack, the non-conceptual forms of self-consciousness do not represent the self explicitly. Consequently, ecological perception and bodily experience cannot be considered forms of genuine self-consciousness, as they do not constitute self-representation but contain only implicitly self-related information: As Perry has convincingly argued, facts that are provided by the context do not figure as part of the explicit representational content of an utterance or of an intentional state (see also Recanati). Moreover, it would put an unnecessary cognitive burden on the organism to represent itself explicitly if such explicit representation is not required for successful interaction with the environment. However, self-consciousness requires explicit self-representation. Hence, while perception and bodily experience are instances of conscious experience, they do not constitute forms of self-consciousness. To put it differently: all sentient beings are subjects of experience, and they experience the world from their own egocentric perspective. Because of this, perceptual content necessarily contains self-related information, which the organism must use to interact with the environment. However, not all subjects of conscious experience also have explicit self-representations or think of themselves as themselves. (Musholt 2015, 56) This point is also well developed by Campbell (1994, 119) in his discussion of egocentric spatial representation: The egocentric frame used in vision employs monadic spatial notions, such as ‘to the right’, ‘to the left’, ‘above’, ‘in front’, and so on, rather than relational notions, such as ‘to my right’, ‘above me’, ‘in front of me’, and so on. The content of visual experience itself does not need to employ an explicit first-person component but merely provides the organism with implicitly self-related information that, once conceptualised, becomes the condition for producing first-person thought. In other words, this kind of experience can represent properties and states of affairs without the necessity of representing the subject of these states. Self-related information is implicit in perception and bodily experience; consequently, it necessarily concerns the subject in question, but it does not require a representation of the perceiving and proprioceiving subject. This stresses the precise difference between self-related information and self-representation: The former does not represent the self explicitly in perception and bodily experience, unlike

On de se and de re  175 self-representation, in which the self is represented explicitly in the content of thought. These considerations are based on Recanati and Perry’s approaches, which were introduced previously. As we saw earlier, Recanati and Perry distinguish between implicit and explicit self-ascriptions according to the distinction between the mode of representation and the content of representation. As seen in the previous paragraphs, the former concerns all the information gained from inside; for example, the subject is not represented through the proprioceptive/kinaesthetic mode but is determined implicitly by the mode while, for judgments that are not made ‘from the inside’, the subject is represented explicitly in the content of the representation. Despite this difference, as seen previously, the authors claim that the self is always an unarticulated constituent; every (implicit or explicit) self-ascription presupposes, in turn, a specific or more basic implicit self-ascription. This further passage from Perry explains the issue in question once again: What each of us gets from perception may be regarded as information concerning ourselves, to explain connections between perception and action. There is no need for a self-referring component of our belief, no need for an idea or representation of ourselves. . . . The eyes that see and the torso or legs that move are parts of the same more or less integrated body. And this fact, external to the belief, supplies the needed coordination. The belief need only have the burden of registering differences in my environment, and not the burden of identifying the person about whose relation to the environment perception gives information with the person whose action it guides. (Perry 2000, 182–3) Thus, two levels have been distinguished: On the first level, the subjective dimension producing a thought is not represented as the subject of that thought; on the second level, the subject can be represented implicitly or explicitly in the content of the thought as the subject to whom a given property is attributed. Perry and Recanati’s relativist perspective articulates such two levels in compliance with specific cognitive constraints. As seen previously, and bearing the obvious differences in mind, Kant achieves the same result in the specific terms of transcendentalism. The two levels in question are preserved due to the distinction between two classes of self-ascriptions, namely those taking ‘I as passive’, and those taking ‘I as active’: While the former self-ascriptions are determined by representations given in sensibility in independent ways, the ‘I as active’ is the basis of all judgments and the referent of all mental self-ascriptions expressing the act of judging one’s given representations. The first level of basic, implicit self-attribution concerns the act of spontaneity—thinking—because it is the synthetic unity of apperception that is the origin of the representational synthesis; hence, it the source of the self-attribution of all thoughts: The act

176 On de se and de re of spontaneity expressed by I think is necessarily involved in the making of any judgment. A few further points can be articulated. (1) The lack of an identification component in the transcendental designation of I think and the resulting implicitness feature of the selfascription of all thoughts are based on the thinking features that are regarded as the synthetic unity of apperception that determines the I think (qua Representation) as an analytical unity of apperception. The I designates the thinking being inasmuch as this is analytically contained as a representation in the synthetic unity of apperception. For precisely this reason, as seen in Chapter 4, a semantic conclusion concerning the transcendental designation of I has been reached: I think, qua Representation, is not indexical at all. (2) At this first level, precisely because the representation I in I think is not indexical, the problems concerning the paradox of self-consciousness introduced by Bermúdez do not seem to touch the Kantian approach to self-consciousness. As mentioned, the terms of paradox consist in the fact that the ability to use the first-person component does not explain but presupposes the ability to think I-thoughts: If the semantic role of the first-person pronoun holds that “I” necessarily refers to the subject who has produced the thought containing it, the subject must already understand that “I” refers to itself in a self-conscious way. Instead, since it is not indexical at all, the use of the representation I in I think does not presuppose any self-conscious ability to think I-thoughts. It is the form contained analytically as representation in thinking: The act of spontaneity expressed by I think is necessarily involved in the making of a judgment, then it takes the form of a much more abstract level. (3) In this way, the role of the I is preserved as the basis of synthesis, since the I is considered the representational correlative of apperception. However, the subjective component is not represented explicitly: There is no res that can enter the thought’s content as a constituent; instead, it is represented implicitly, as the I is given as the analytical form of thought. The lack of an identification component and the resulting implicitness feature that Kant assigns to the self-ascription of all thoughts avoids all of the problems pointed out by Musholt when she defends her nonself-representationalist account in contrast to the self-representationalist account of non-conceptual self-awareness. As mentioned previously, Kant rejects the reflexive model of self-consciousness, as the I of apperception can never be the representational vehicle employed to form an intentional object. It is worth noting that it almost seems that Kant accounts for the implicitness of self-ascription by introducing the nonconceptual content of I. In fact, when he states that the I is the bare and empty form that indicates the nexus established by the copula in the judgment through the representational synthesis of the unity of apperception, establishing the I as the analytical form of thought, he

On de se and de re  177 introduces a representational form that is not conceptual at all. As seen in Chapter 4, the I does not represent any thinking subject conceptually because it employs neither content mediation nor a prior instance of identification articulated in conceptual marks. An empty form is quite the reverse of the definition of the concept as a common representation, which is not free from extension, and is subject to further determination at all times. A representational unicum of the kinds of representations is the basis of self-conscious ability.

The Problem of Self-Knowledge The Dualism of the I of Apperception and the I as Human Being As seen in Chapter 3, Allison (2004) does not regard inner sense as a possible source of empirical self-knowledge due to the radical asymmetry between outer and inner experience accounted for by the absence of impressions of the self. Kant certainly offers different clues as to the specific topic of self-knowledge and, as has already been noted, seems quite explicit on a few occasions, as in § 25 of the Transcendental Deduction, for instance. Here, he introduces two levels of self-knowledge based on the distinction between the transcendental and the empirical plane, or between spontaneity and receptivity. As seen in Chapter 2, what can be assumed according to the representation I of apperception is only a being devoid of properties. With regard to the self-knowledge expressed by the ‘I’ of apperception, Kant seems unambiguous and contends that there is no cognition of the self as such: Now since for the cognition of ourselves, in addition to the action of thinking that brings the manifold of every possible intuition to the unity of apperception, a determinate sort of intuition, through which this manifold is given, is also required, my own existence is not indeed appearance (let alone mere illusion), but the determination of my existence can only occur in correspondence with the form of inner sense, according to the particular way in which the manifold that I combine is given in inner intuition, and I therefore have no cognition of myself as I am, but only as I appear to myself. (B157–8) In this way, the subject cannot determine its existence as that of a selfactive being, since its existence can only be determined via the form of inner sense; for this reason, Kant holds that “I therefore have no cognition of myself as I am, but only as I appear to myself”. Even though the subject cannot determine its existence as that of a self-active being, it is through the I think that the spontaneity of the thinking being is represented: “I cannot determine my existence as that of a self-active being; rather I merely

178 On de se and de re represent the spontaneity of my thought. . . . Yet this spontaneity is the reason I call myself an intelligence” (B158). The reason that there is no determination of the existence as a self-active being is expounded in further detail: If the ‘I think’ as ‘Thinking’ expresses the act of determining the subject’s existence, then such an existence, although given, is not determined based on manner; that is, not ­according to the manifold that belongs to it. Such a determination would require not only an intuition—­more precisely, a self-intuition, which is grounded in time as the a priori sensible form of the receptivity of the determinable—but also another self-intuition specifically relative to that which it is determining, namely mere spontaneity, in much the same way as time provides the condition for the determinable. However, there is only an intellectual consciousness of spontaneity, and since Kant does not allow for any intellectual intuition, the subject cannot determine its existence as that of a self-active being. It is worth quoting the entire passage: The I think expresses the act of determining my existence. The existence is thereby already given, but the way in which I am to determine it, i.e., the manifold that I am to posit in myself as belonging to it, is not yet thereby given. For that self-intuition is required, which is grounded in an a priori given form, i.e., time, which is sensible and belongs to the receptivity of the determinable. Now I do not have yet another self-intuition, which would give the determining in me, of the spontaneity of which alone I am conscious, even before the act of determination, in the same way as time gives that which is to be determined, thus I cannot determine my existence as that of a self-active being, rather I merely represent the spontaneity of my thought, i.e., of the determining, and my existence always remains only sensibly determinable, i.e., determinable as the existence of an appearance. Yet this spontaneity is the reason I call myself an intelligence. (B158 n.) If existence is the only sensibly determinable, that is, if it is determinable as the existence of an appearance, then, in contrast to Allison’s view, a dimension of empirical self-knowledge seems still possible (cf. infra). In several passages, it is Kant himself who points out that, even though the subject has no cognition of itself as it is, it still has knowledge of how it appears to itself. On this specific topic, some interpretations show the possibility of empirical self-knowledge, to the extent that the problem ultimately consists of how to reconcile the subject, or I of apperception—the transcendental self-consciousness expressed by the formula ‘I think’, whose existence is not determinable—and the self-knowledge expressed by the empirical apperception, which allows the subject to cognise itself as part of the spatio-temporal world. This topic is partly connected to the metaphysical theses introduced previously: One of the main points in question essentially concerns whether the

On de se and de re  179 I of apperception refers to an object and, if so, how it connects to the ‘I’ of empirical apperception. For example, Rosefeldt (2006) states that, in both editions, the ‘I’ should be regarded as a representation that picks out an object, even if this is a Gedankending, that is, a kind of ‘non-real’ object: Although intuition is absent, it still qualifies as an object. For Wolff (2006), if the I refers to an object, then this object is the human being proper 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. For Ginsborg (2015), the I of apperception introduced in § 16 of the B-Deduction must be identified with a particular human being in space and time, i.e., with something that can be the object of empirical self-knowledge. By employing the thought ‘I think’ in its relevant sense, every subject refers to a particular human being, a human being of which the subject can acquire empirical knowledge through both outer and inner senses. Horstmann (1993) identifies two arguments in the two editions: In the A-edition Paralogism, the representation I refers to what is termed “substrate”, i.e., the I refers to an object even though this is not knowable. Conversely, in the B-edition Paralogism, the ‘I think’ should simply be regarded as an act of spontaneity: No entity is singled out by the I of apperception, the I solely representing the spontaneous activity of thought. As a result, according to the B-edition, the transcendental self-consciousness articulated in the Deduction and Paralogisms cannot be related to the consciousness of oneself as object. Prima facie, there seems to be some incompatibility between self-knowledge as a human being, that is, as an object embedded in a spatio-temporal causal order governed by natural laws, and the knowledge the subject possesses as a thinking subject through the spontaneity of the ‘I’ of apperception: The representation I think contains no intuition that can connect it to the subject considered as an empirical object. Longuenesse and Ginsborg on the I Considered as Human Being Several scholars have acknowledged this topic, and some strategies can be isolated in order to overcome this type of dualism involving the ‘I’ of apperception and the ‘I’ as human being. According to Ginsborg (2015, 358), if one does not identify these two dimensions in the Kantian approach, it is hard to see how Kant’s account of the conditions of cognition in the first 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.

180 On de se and de re The problem mainly concerns the nature of transcendental idealism and the fact that, as seen in the main parts of the first Critique, transcendental apperception refers to a thinking subject, which can be regarded as ‘outside of’ space and time (in fact, the authors also refer to the ‘transcendental subject’, even though this has been regarded as a formal notion in these pages). Although the activity of unifying representations through apperception is the source of the constitution of appearances as spatio-temporal objects, the thinking being, which is the origin of this unification activity, resides outside of space and time. For these reasons, it is not easy to reconcile the ‘I’ of transcendental apperception and the ‘I’ considered as human being, i.e., as a spatio-temporal object. In other words, if, in the first instance, the ‘I’ refers to that which is regarded as the thinking subject, then the problem consists of whether it also picks out a human being in order to reconcile the spatio-temporal human being with the self-aware thinking subject as that which generates the representation of a world that includes objects such as human beings. Longuenesse tries to weaken certain possible consequences of Kant’s idealism related to this dualistic result. The thinking activity addresses representations, gives them objectivity, and enables them to become representations directed towards an objective spatio-temporal world. The task of this activity, however, is only to confer intentionality on the representations, not to generate a spatio-temporal world as such. According to this reading, the activity of apperception makes spatio-temporal objects, understood simply as ‘represented objects’, possible: The ‘I’ is not responsible for their existence. The activity of apperception has to be regarded exclusively in this epistemic task. It follows that it is possible to consider the event that the ‘I’ of apperception can be identified with a spatio-temporal object, specifically with one of the human beings in the world, since the existence of the human being would not depend on the activity of the ‘I’ of apperception, but on the possibility of being represented to itself as a spatio-temporal object. In this way, the thinking activity makes the representation of the objects—and of human beings themselves—possible, and the subject can recognise one of them as itself. The author moves from the analysis of the Analogies of Experience, in particular the third analogy and the forms of judgments under the heading of relation, pointing out that the ‘I’ of apperception generates the representation of a unified space and time in which the subject can cognise all empirical objects and their relationships of interaction. By cognising these causal interactions, “we also situate ourselves, as empirical unities of consciousness associated to a body we represent as our own, in the unified empirical space and time whose representation we thereby generate” (Longuenesse 1998, 378). This argument is connected to Longuenesse’s own reading of the Anticipations of Perception, according to which our awareness of the universal ‘community’ or interaction of the objects we perceive in space . . . includes an awareness of their interaction with

On de se and de re  181 our own body, and thus of their causal determination of the ‘matter of our perceptions’, sensations. (Longuenesse 1998, 322) According to Longuenesse, the subject synthesises a sensory manifold to represent a spatio-temporal world in which interacting substances are cognised; among these substances, the subject represents its own body, which it recognises as interacting causally with other bodies, and, in fact, as a body in a causal relationship with its mental states, in particular with the sensations that constitute the material of the synthesis: 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; or the regularly repeated succession of rays of light shining on raindrops and sensations of rainbows of color. Thus the subjective manifold of our empirical representations is itself temporally determined in causal relation to the universal temporal order of the objects of experience, although we do not need to suppose any other community of reciprocal interaction than that of corporeal substances. For since only corporeal substances are given in space, only they can be cognized as simultaneous and thus in relations of universal reciprocal determination. And this is how the astonishing edifice of Kant’s Analogies of Experience comes to completion: by the location of “us” in the empirically given world. An “us”, that is, 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. (Longuenesse 1998, 392–3) Even though Longuenesse does not explicitly refer to any identification of the ‘I’ of apperception with a particular human being as an object located in space and time, this identification is quite plausible (Ginsborg 2015): Due to the spontaneous activity, the ‘I’ as a transcendental subject locates itself as a unity of empirical consciousness associated with a phenomenal body. In this way, the representation I in I think picks out a human being who is regarded as both a transcendental subject—hence as the generating origin of the representation of the spatio-temporal world—and as an object located in the spatio-temporal world. Ginsborg (2015) is highly critical of this view: Longuenesse’s attempt to explain the question of how a transcendental subject, initially conceived of

182 On de se and de re as independent of the spatio-temporal world, can be ‘located’ or ‘situated’ within that world proves unsuccessful. The author argues that the dualism of spontaneity and receptivity remains unsettled because these are two very different dimensions that cannot connect. Ginsborg recalls McDowell’s view: If the ‘I’ in ‘I think’ is spontaneous, then the ‘I’ cannot pick out a human being because this is part of the natural causal order. The point is the following: When Longuenesse describes the ‘us’ we locate in the empirical world as unities of consciousness that are both passive (receptive, capable of conscious sensation and associative imagination) and active (spontaneous, intellectual, capable of judgment and synthesis speciosa), she affirms that being spontaneous is compatible with being part of the natural causal order; the human being is not only regarded as passively affected by the sensations but can also be identified as a spontaneous activity. Hannah Ginsborg summarises Longuenesse’s point as follows: The suggestion is that I—the I that thinks—simply ascribe to HG the same spontaneous understanding which (on transcendental reflection) I ascribe to myself. That is to say, I ascribe to HG a capacity not only of being affected by things so as to undergo sensations, but also of spontaneously synthesizing those sensations, in accordance with the categories and forms of judgement, so that they come to form a unified 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 lifting, by generating, out of the material provided by the cup’s affecting 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. . . . At first 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. (Ginsborg 2015, 215) According to Ginsborg, Longuenesse’s solution, whereby the ‘I’ that thinks can identify itself with a particular human being in the world by conceiving of the human being as endowed with spontaneity, is somewhat misleading. It is not possible to regard the candidate human being as simply standing in the same spontaneous relation to her sensory manifold in the same way that the correlative ‘I’ of the transcendental subject stands in relation to its own; instead, 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 spatiotemporal appearance but

On de se and de re  183 which at the same time allows her to have cognition of, and not merely to be affected by, other spatiotemporal appearances in her environment. In short, what HG needs, in order to be me, is the appearance of spontaneity. (Ginsborg 2015, 216) In order to fill the role of the phenomenal correlate of the spontaneity activity of the ‘I’, Ginsborg suggests taking the faculty of judgment characterised by Kant in the third Critique into consideration: 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. (Ginsborg 2015, 216) The spontaneous activity of ‘I’ is nothing but the human beings’ exercise of judgment, with neither determination at the empirical level nor—more generally—spatio-temporal character. Spontaneity is nothing more than the faculty of judgment from a transcendental, rather than from an empirical, perspective. Following Ginsborg’s approach, my perspective contrasts with Longuenesse’s view: The dimensions of spontaneity and receptivity can be connected and identified only from a representational perspective. In particular, regardless of whether Ginsborg’s view is a correct interpretation of the role of judgment, a similar result can be achieved by considering the articulation between the ‘I’ of apperception and inner sense, in which the thinking being is manifest as appearance, although from an epistemic rather than from a metaphysical perspective. Turning to Capozzi’s interpretation of Kantian empirical self-knowledge, it is possible to fill the role of the phenomenal correlate of the spontaneity activity through the task of attention in the inner sense. Capozzi and the Role of Attention in the Inner Sense At this point, and in moving towards the conclusion of this inquiry, the general interpretative reading employed here will be recapped in order to present a final overview of the self-knowledge topics touched upon thus far. When taking the empirical plane into consideration, i.e., the ‘I’ of empirical apperception and the way in which the ‘I’ as appearance is connected to the transcendental dimension, any immediate identification of different dimensions that does not seem justified in the frame of transcendental idealism must be avoided. Unlike the previous readings, including Ginsborg’s appeal to the faculty of judgment to enable the metaphysical identification of the ‘I’

184 On de se and de re of apperception with a particular human being, the interpretative approach employed in these pages follows the general principles of Transcendentalism accounted for previously in a Formal Ownership Reading (cf. supra, Chapter 2)—the representation I does not allow for a metaphysical conclusion. In particular, it is not possible to reach the metaphysical conclusion that the activity of thinking can be ascribed to a spatio-temporal object or human being; certainly, this does not rule out the epistemic conclusion that the activity of thinking reveals itself in the empirical dimension as an appearance, and that it is also cognisable through empirical means in order for the subject to regard itself as a human being who can be the object of consciousness through the inner and outer senses. To point out an important textual example, in B415, Kant maintains that the thinking being, regarded as a human being, is also an object of the outer sense. It is also worth pointing out a Handschrift crossed out in Anthropology, a remark in which Kant claims that empirical self-cognition presents “the human being as he appears to it, not as he is in himself” to inner sense, specifying that “every cognition explains merely the affectability of the subject, not the inner characteristic of the subject as object”. Recalling the self-affection argument, which will be reconsidered shortly, Kant states that, even though the subject has a right to attribute its representation to itself considered as a person, its self-representation merely depends on the form through which the subject is affected in the inner sense: In the self-cognition of the human being through inner experience he does not make what he has perceived in himself, for this depends on impressions (the subject matter of representations) that he receives. Therefore he is so far enduring, that is, he has a representation of himself as he is affected by himself, which according to its form depends merely on the subjective property of his nature, which should not be interpreted as belonging to the object, even though he still also has the right to attribute it to the object (here his own person), but with the qualification that he can only recognize himself as an object through his representation in experience as he appears to himself, not as he, the observed, is in himself. (Anth-B 7:141, 252) Later, when explaining the intrinsic difference between inner sense (empirical self-consciousness) and apperception (intellectual selfconsciousness), “which are usually taken to be one and the same”, Kant reasserts the epistemic limit of empirical self-knowledge as related to the concept of a human being: The I in every judgment is neither an intuition nor a concept, and not at all a determination of an object, but an act of understanding by the determining subject as such, and the consciousness of oneself; pure

On de se and de re  185 apperception itself therefore belongs merely to logic (without any matter and content). On the other hand, the I of inner sense, that is, of the perception and observation of oneself, is not the subject of judgment, but an object. Consciousness of the one who observes himself is an entirely simple representation of the subject in judgment as such, of which one knows everything if one merely thinks it. But the I which has been observed by itself is a sum total of so many objects of inner perception that psychology has plenty to do in tracing everything that lies hidden in it. And psychology may not ever hope to complete this task and answer satisfactorily the question: “What is the human being?” (Anth-B 7:141, 253) This point is connected to many passages in which Kant indicates that the subject can (re)present itself in two ways: through the I that thinks and through the I that intuits itself. As noted previously, in another passage, Kant (V-Met-L1/Pölitz 28: 224, 44–5) also states that the I can be taken in a twofold manner: I as human being and I as intelligence. I in the first sense means: I am an object of the inner and the outer sense. I in the second sense means that I am the object of the inner sense only. Obviously, this does not imply two Is; on the contrary, and in line with a formal reading, Kant claims that the “I as a thinking being am one and the same subject with myself as a sensing being” (Anth 7: 142, 33). When Kant considers the self of animals in Metaphysik Mrongovius, he does not consider two selves, but three: The self underlies consciousness and is what is peculiar to spirit. But we can consider this self in three ways: I think as intelligence, i.e., the subject of thinking is intelligence. I think as subject which has sensibility, and am soul. I think as intelligence and soul, and am a human being. A body which is animated only by a soul, and not by intelligence, is an animal. (V-Met/Mron 29: 878, 247–8) In this way, if the problem is examined from an epistemic rather than from a metaphysical angle, i.e., from the representational views through which the thinking being is given to itself as subject (as Kant remarked, the self can be considered in three ways), it can be maintained that, in the first place, the representation I should be regarded as the representational correlate of transcendental apperception, which provides the condition of possibility for the thinking subject to employ several representational means in order to cognise and refer to itself as a human being. In the absence of this I as a conditio sine qua non, the thinking subject cannot cognise itself

186 On de se and de re in any other way. In the same passage from Metaphysik Mrongovius, Kant states that the faculty for grasping the thought: I am, belongs solely to the intelligence. This I remains [even] when everything has changed, when bodies and principles have changed. Now what the identity of its self consists in is difficult to know; everything is related to this, everything can change, only consciousness and apperception, or the faculty for referring representations to one’s self, remain. (V-Met/Mron 29: 878, 248) Consequently, the dualism of the transcendental and the empirical dimensions can be overcome without paying the price of a metaphysical commitment to the identification of the ‘I’ of transcendental apperception and the ‘I’ of empirical apperception. These are not regarded as two different entities, but only as two different ways of representing the thinking being, in compliance with the two-aspect view of transcendental idealism (cf. supra, Chapter 2). Another passage also seems to move in this direction: Kant points out the subject’s ability to use I to refer to itself, both as a soul—as an object of inner sense—and as a body (and, as just mentioned, both depend on apperception, namely the faculty of grasping the thought: I am or I think): Here we consider the soul merely as the object of inner sense, and that rests on our own experience which no one can deny. Furthermore, we are not worrying here about the question whether the soul [is] material or not, because we cannot prove this from experience, and thus it belongs in rational psychology . The soul is merely our I, not the body, but body and soul together, as human consciousness, are also called I. In empirical psychology we consider our I as soul and as human being. But we consider the body, on the one hand, as an organ of the soul which depends on the soul, but on the other hand as a lodging, since the soul also often depends on it. (V-Met/Mron 29: 876–7, 246) Furthermore, as mentioned previously, Capozzi’s reading of Kantian empirical self-knowledge will be employed to explain the role of the phenomenal correlate of the spontaneity activity in order to connect the dimensions of spontaneity and receptivity based on a representational perspective. The question was partially introduced in Chapter 3 and can be resumed in three steps by highlighting the usual problems associated with the possibility of forming stable, empirical self-knowledge. (A) To recall the considerations expounded in the previous chapters, in the empirical apperception the inner perception of something existing as thinking cannot be determined as an existing substance in time and

On de se and de re  187 space—the forms of inner and outer sense through which all appearances are given—otherwise, this would not be thought, but matter. The consciousness of the self as contemplated by empirical apperception is the inner perception of something that is not the object of outer sense; thus, as with the I of pure apperception, the representation I of empirical apperception is no concept at all and represents no thinking subject conceptually: I is neither a repraesentatio communis (common representation), which is predicable, nor is it subjected to further determinations that can be articulated via conceptual marks. Clearly, I is a designation of the object of inner sense: Quoting the passage from Prolegomena, For the I is not a concept at all, but only a designation of the object of the inner sense insofar as we do not further cognize it through any predicate; hence although it cannot itself be the predicate of any other thing, just as little can it be a determinate concept of an absolute subject, but as in all the other cases it can only be the referring of inner appearances to their unknown subject. (Prol 4: 334, 86) In another passage from Prolegomena, not only does Kant reaffirm that the I, as a representation of empirical apperception, is not a concept, but he also specifies that the I is the feeling of an existence: If the representation of apperception, the I, were a concept through which anything might be thought, it could then be used as a predicate for other things, or contain such predicates in itself. But it is nothing more than a feeling of an existence without the least concept, and is only a representation of that to which all thinking stands in relation (relatione accidentis). (Prol 4: 334 n., 86) As observed in Chapter 3 with reference to the distinction between judgments of experience and judgments of perception, some judgments of perception can never become judgments of experience because they are based exclusively on subjective sensations or feelings. A dual reference to the subject’s experience and consciousness is required in order for representations to become knowledge (cf. A 320/B376; Log 9: 33); instead, the internal or subjective sensation (cf. Anth 7: 156), thus equated to the Gefühl (cf. KU 5: 206), which lacks a potential reference to an object of reality, is a representation that is connected exclusively to the subject and, as such, only manifests in time. On this specific topic, we can consider the distinction between the I of pure apperception and the I of empirical apperception once again. While the latter is regarded as the feeling of an existence, with regard to the former Kant refuses to consider I think/I am as a feeling. A passage from Opus

188 On de se and de re Postumum13 makes this point straightforwardly: I is a logical representation of consciousness (cf. supra, Chapter 2), a facultas motrix, i.e., a faculty that acts while in motion. Since the consequent action is an inner action, it does not move from a physical or spatial position; however, as it expresses an activity, it cannot be seen as a feeling connected to receptivity, either. If the empirical apperception only expresses the feeling of an existence, which is exclusively connected to an occasional internal sensation, then the subjective dimension is only represented in a varying, unstable way, allowing no experience of the subject as a permanent appearance. This recalls passage A107, in which Kant affirms that the empirical subject considered by the inner sense is not represented as a standing or an abiding object. The corresponding I think is nothing more than a judgment of perception, which can never become a judgment of experience due to it being based exclusively on subjective sensations or feelings. In brief, the issue of the possibility of empirical self-knowledge can be summarised as follows: (1) From the empirical perspective, I think is an empirical proposition equivalent to I exist thinking; thus, there is a determination of the subject at the level of existence insomuch as the I as object of the perception is revealed as appearance by empirical apperception. (2) The empirical apperception, however, only expresses the feeling of an existence, which is exclusively connected to an occasional internal or subjective sensation; as a result, it only expresses a continuously interrupted existence. (3) It follows that two elements can be taken into account to identify an empirical self-knowledge approach in the Kantian perspective and to overcome the fact that any consideration of the psychological subjective dimension cannot be construed on the mere grounds of the feeling of an existence. In order for the subject to know itself as appearance, its determination at the level of existence is based on affection or on an activity deriving from the subject itself; in fact, the activity or affection is a self-affection. The second element, on the other hand, concerns the external sense, which is necessarily involved in order to obtain the permanence of the self in the empirical apperception. (B) As pointed out in Chapter 1, it is through the self-affection argument that the distinction between an active ‘I’ (of transcendental apperception) and a passive ‘I’ (of inner sense) is identified for the first time. The argument can be found in two different places in the first Critique—in the Transcendental Aesthetic B67–8 and in § 24 of the Transcendental Deduction—and concerns mind and inner sense, as well as how the mind affects itself while apprehending its own representations in time, i.e., the form of inner sense in which all appearances appear. It is exactly for this reason that we can reach the

On de se and de re  189 conclusion, as emphasised previously and more than once, that the subject only cognises itself “as it appears to itself, not as it is” (B69). Paton and Allison ponder the analogy between self-affection and the affection by external objects, i.e., the matter of empirical intuition. In the latter case, the mind, which receives this material and is considered passive in this task, is affected by the external material through its spatio-temporal form of receptivity; as a result, the objects are cognised as appearances. In the former, the mind affects itself when it apprehends its representational contents in time and, because of this, it also cognises itself in a sensible way, i.e., just as it appears to itself. As has been discussed previously (cf. supra, Chapter 1), in B67, Kant states that the representations of outer sense constitute the material proper of the inner sense; it follows that time, as a formal condition containing the relationships of sequencing and simultaneity in which these representations are allocated, precedes the consciousness of them in experience and establishes how the mind is affected by its own activity: Now that which, as representation, can precede any act of thinking something is intuition and, if it contains nothing but relations, it is the form of intuition, which, since it does not represent anything except insofar as something is posited in the mind, can be nothing other than the way in which the mind is affected by its own activity, namely this positing of its representation, thus the way it is affected through itself, i.e., it is an inner sense as far as regards its form. (B67–8) If the subject affects itself by its own activity as it apprehends its representational contents, in the passage B68 quoted below, Kant specifies the difference between an active task and a passive side of the mind: Everything that is represented through a sense is always appearance, and it is due to that which is represented in inner sense that the subject becomes the object of this sense and can be represented as appearance at the same time, “not as it would judge of itself if its intuition were mere self-activity, i.e., intellectual”. The difference from apperception is restated in the following passage, in which Kant speculates about how a subject can intuit itself internally: Consciousness of itself (apperception) is the simple representation of the I, and if all of the manifold in the subject were given self-actively through that alone, then the inner intuition would be intellectual. In human beings this consciousness requires inner perception of the manifold that is antecedently given in the subject, and the manner in which this is given in the mind without spontaneity must be called sensibility on account of this difference. (B68)

190 On de se and de re Later, Kant specifies what it means to become conscious of oneself through sensibility: “It is to seek out (apprehend) that which lies in the mind”. Thus, the subject becomes conscious of itself in this way because it affects itself, and this determination produces an intuition of itself. This intuition is grounded in the pure form of time, which determines the way in which the manifold is united in the mind. For this reason, the subject, here considered as the faculty of consciousness, intuits itself “not as it would immediately self-actively represent itself”, but passively, “in accordance with the way in which it is affected from within”, i.e., in the way the manifold is united in inner sense, which consequently presents the subject as appearance to itself (passively) but not as it is (self-actively). In the Transcendental Deduction, in which the link between the united manifold and the transcendental synthesis is established, the disanalogy between outer affection and self-affection is obvious. In § 24, Kant considers self-affection or affection of the inner sense as the determination of the inner sense via understanding under the designation of a transcendental synthesis of the imagination. Therefore, while the external objects influence outer sense in the former case, in the latter it is the understanding, i.e., the transcendental synthesis, which influences and determines the inner sense. The understanding exerts its action; inner sense is thereby affected and, accordingly, the subject becomes passive to itself. Thus, the active subject coincides with apperception and its synthetic unity: Apperception . . . is so far from being the same as the inner sense that the former, rather, as the source of all combination, applies to all sensible intuition of objects in general, to the manifold of intuitions in general, under the name of the categories; inner sense, on the contrary, contains the mere form of intuition, but without combination of the manifold in it, and thus it does not yet contain any determinate intuition at all, which is possible only through the consciousness of the determination of the manifold through the transcendental action of the imagination (synthetic influence of the understanding on the inner sense). (B154) The Kantian account of inner experience described in the so-called Leningrad Reflexion on Inner Sense—in the inner experience “I affect myself insofar as I bring the representations of outer sense into an empirical consciousness of my condition” (Refl 18: 623, 364–5) (cf. supra, Chapter 1)— restates the issue, maintaining that inner experience involves a reflective reappropriation of the contents of the outer experience. This point merely ratifies passage B67—“the representations of outer sense make up the proper material with which we occupy our mind”—although this conclusion is reached based on inner experience: Since there is no empirical intuition of the self as an external object, the matter of inner intuitions consists of no sensory impressions produced by or grounded in the self; as a result,

On de se and de re  191 the manifold of the inner sense can only consist of outer sense representations. The result reaffirms an important asymmetry between outer and inner experiences: If the inner sense presents no manifold of its own, the sensible representations involved in the judgments of inner sense through which the subject represents itself to itself in the empirical dimension do not represent the subject as object. As partly indicated previously, according to Allison and due to the absence of impressions of the self, the only role of inner sense that can be taken into account is that it can enable the subject to recognise the representations that belong to it as its own. In other words—as described in Chapter 3, in which the problem of self-identification was addressed—if the empirical predicates referring to the subject in judgments of empirical experience can be conceived of as mere reflective reappropriations of the contents of outer experience, then the subject cannot cognise itself in the same way in which it cognises the outer objects. The latter are spatio-temporal objects of empirical intuitions and, as such, can be cognised due to the application of sensible and intellectual forms of categories based on judgment; instead, the ‘I’ or self is not an empirical spatio-temporal object. Allison’s conclusion is that inner sense does not produces any empirical self-knowledge. Several passages have been introduced to refute this conclusion; Capozzi’s interpretative reading moves from these arguments to reach a very different conclusion through his analysis of the mechanism of attention involved in inner sense. (C) Capozzi’s hypothesis is to consider the external sense and the permanence of the object (matter or body) to obtain the permanence of the self in the empirical apperception. In this way, the external sense is assumed to play a crucial role: In order to determine the subject’s existence in time, it is necessary to involve the existence of objects perceived by the outer sense, starting with the subject’s body itself. With the involvement of the mechanism of attention in the inner sense, this role appears to be even more crucial to account for empirical self-knowledge. Two Reflexionen can be considered to introduce this specific issue. The first is the Reflexion 3826 (KGS 17: 304): “Gott erkennt alles, indem er sich selbst erkennt. Der Mensch erkennt sich selbst, indem er andere Dinge erkennt”; (“God knows all knowing itself. The man knows himself knowing the other things”). The second is the Reflexion 5654: One can well set time in oneself, but one cannot set oneself in time and determine oneself in it, and yet in that consists empirical selfconsciousness. Thus in order to determine one’s existence in time, one must intuit something else in an outer relation, which for that very reason must be considered as persisting. Because time cannot be externally perceived in things, since it is only a determination of the inner sense, we can only

192 On de se and de re [crossed out: represent] determine ourselves in time insofar as we stand in relation to things outside us and consider ourselves therein, and that which is outside us introduces an existence insofar as it is not subject to alteration, i.e., that is persistent. (Refl 18: 313, 285) This passage highlights Kant’s effort to develop his transcendental idealism as a “critical” or “formal” idealism in order to distinguish it from both Descartes’ problematic idealism and Berkeley’s dogmatic idealism, an issue developed in a few passages from Prolegomena and, above all, in the second edition of first Critique, in the section on the Refutation of Idealism (B274–9). Here, Kant maintains that the possibility of self-consciousness presupposes the existence of an external world of objects that are also conceived of as existing independently of whichever representation relates to them. Hence, the subject’s empirically determined consciousness of its own existence proves the existence of objects in space: I am conscious of my existence as determined in time. All timedetermination presupposes something persistent in perception. This persistent thing, however, cannot be something in me, since my own existence in time can first be determined only through this persistent thing. Thus the perception of this persistent thing is possible only through a thing outside me and not through the mere representation of a thing outside me. Consequently, the determination of my existence in time is possible only by means of the existence of actual things that I perceive outside myself. Now consciousness in time is necessarily combined with the consciousness of the possibility of this time-determination: Therefore it is also necessarily combined with the existence of the things outside me, as the condition of time-determination; i.e., the consciousness of my own existence is at the same time an immediate consciousness of the existence of other things outside me. (B275–6) If “we do not even have anything persistent on which we could base the concept of a substance, as intuition, except merely matter, and even this persistence is not drawn from outer experience, but rather presupposed a priori as the necessary condition of all time-determination”, then this persistence also serves as the determination of inner sense with regard to the subject’s own existence. Once again, what is being stated is that the representation I, which expresses the subject’s consciousness of itself, is no intuition at all, but “a merely intellectual representation of the selfactivity of a thinking subject”. On the other hand, the reference to the intellectual activity serves to link these considerations to the self-affection argument: In order for the subject to know itself as appearance, the activity has to be regarded as selfaffection, i.e., the understanding produces the combination of the manifold

On de se and de re  193 affecting the inner sense, and the permanence of the objects in the external sense is required to obtain a sort of permanence of the subject in the empirical apperception. This last point is highlighted by Capozzi through an analysis of the mechanism of attention: In the consciousness of the self relying on the empirical determination of the inner perception, i.e., on the subject’s ability to perceive itself in the paradigmatic instance of the psychological mechanism of attention as something that thinks while apprehending representations in the psychological flow towards the outside, the subject reveals itself as appearance and phenomenises. De facto, in so doing, it indirectly obtains a persistence that exceeds the afore-mentioned variable nature of empirical apperception. Kant addresses the psychological mechanism of attention in a note to B156: I do not see how one can find so many difficulties in the fact that inner sense is affected [afficirt] by ourselves. Every act of attention can give us an example of this. In such acts the understanding always determines the inner sense, in accordance with the combination that it thinks, to the inner intuition that corresponds to the manifold in the synthesis of the understanding. How much the mind [Gemüth] is commonly affected by this means, everyone will be able to perceive [wahrnehmen] in himself. In the lexicon of the time, attention is generally regarded as an activity in which one part of perception is clearer than another. In the Fortschritte, attention is called “the inner psychological observation”: Every inner psychological observation that we undertake can serve us as proof and example; since for this we are required to affect the inner sense, in part also doubtless to the point of fatigue, by means of attention (for thoughts, as factual determinations of the power of representation, also belong to the empirical representation of our state), in order to have first of all in the intuition of ourself a knowledge of what inner sense is presenting to us; which then merely makes us aware of ourself as we appear to ourselves; whereas the logical self does indeed point to the subject as it is in itself, in pure consciousness, not as receptivity but as pure spontaneity, but beyond that is also incapable of knowing anything of its nature. (FM 20:270, 362) The inner sense is subjected to self-affection by intellectual activity; if, however, there is an involvement of the act of attention and the subsequent choice or selection of the thoughts due to the inner psychological observation, then the inner sense expresses not only the synthesis of the thoughts in a passive way but also the self-knowledge of the subject’s identity as an

194 On de se and de re activity that unfolds through appearance. Capozzi develops this point very clearly by referring to a few crucial Kantian passages: The empirical apperception or inner sense—the consciousness of an existent thinking I—is the inner perception of one’s own state as being thinking linked to the occasions when one is aware of rhapsodic external representations by which he or she is affected. Hence, it “is by itself dispersed and without relation to the identity of the subject” (B133). However, in the precise moment its external representations are known as related to existing, identifiable, and reidentifiable objects through the order and regularity introduced “by ourselves” (A125), this empirical apperception can also know itself as an appearance, the existence of which does not reside in the instant, but interrupts and reappears in ever changing guises every time it accompanies what is subjected to external affection. The perceptual consciousness that I have of my being thinking becomes the consciousness of the identity of an I when I no longer limit myself “by my accompanying each representation with consciousness, but rather by my adding one representation to the other and being conscious of their synthesis” (B133). So I can say that I was the person who identified the existence of a substance in the manifold, and that I am the same person when the same substance returns, although changed. As a result, the empirical apperception obtains permanence and identity, as well as the kind of existence that an appearance has . . . because it now asserts “that I was, I am, and I will be, i.e., I am a thing of past, present, and future time” (Refl 18: 623, 365). (Capozzi 2007, 48–9) This empirical self-knowledge does not only regard the passive way in which the intellectual synthesis affects the inner sense, but also the activity of thinking itself; as such, it proves to be the phenomenal correlate of the spontaneity activity invoked previously by Ginsborg to avoid the dualism of the transcendental and the empirical dimensions. The point is that the dualism conceived of in this interpretive scenario is epistemic, not metaphysical. The identity of this activity lasts as long as it acts for knowing: Obviously, with regard to the transcendental considerations introduced previously, there is no spatio-temporal object in the world that can be identified with empirical apperception or its features of non-corporeal simplicity unfolding in the inner form of time; nonetheless, the empirical self-knowledge concerning the thinking activity expresses the phenomenal identity of this activity until this maintains a relationship with the intuitions of the outer sense. In other words, by adopting a formal rather than a metaphysical reading, as long as the thinking activity assigns permanence and identity to the object of the external world through the spatio-temporal forms of sensibility, a type of permanence and identity unfolding in time is assigned to the

On de se and de re  195 same empirical apperception, i.e., one’s consciousness of perceiving oneself as thinking persists in the inner sense as long as the existence of a permanent object is established in the outer sense. For this reason, and recalling the Leningrad Fragment, while the transcendental apperception only asserts “I am”, the empirical apperception asserts that “I was, I am, and I will be”. However, the latter, and the possibility for the subject to assert “I was, I am, and I will be”, depends on the former, i.e., on the possibility of the subject grasping itself through the empty and simple transcendental representation I think/I am produced as an analytical unity by the synthetic unity of transcendental apperception.

A Brief Conclusion Two kinds of self-knowledge can be identified, one concerning the justmentioned I of empirical apperception and the other the I of transcendental apperception; they are both grounded in the two ways of representing the subjective dimension. The following passage from Fortschritte summarises the question, recalling the distinctions introduced in this book. Firstly, the subject, here referred to as person by Kant, presents two ways or senses of representing itself; as in the Anthropology, Kant points out that, if the subject appears to us as double, this does not imply that there are two Is. On the contrary, “I as a thinking being am one and the same subject with myself as a sensing being”. That I am conscious of myself is a thought that already contains a twofold self, the self as subject and the self as object. How it should be possible that I, who think, can be an object (of intuition) to myself, and thus distinguish myself from myself, is absolutely impossible to explain, although it is an undoubted fact; it demonstrates, however, a power so far superior to all sensory intuition, that as ground of the possibility of an understanding it has as its consequence a total separation from the beasts, to whom we have no reason to attribute the power to say ‘I’ to oneself, and looks out upon an infinity of self-made representations and concepts. We are not, however, referring thereby to a dual personality; only the self that thinks and intuits is the person, whereas the self of the object that is intuited by me is, like other objects outside me, the thing. The second passage concerns the subject of apperception: Of the self in the first sense (the subject of apperception), the logical self as a priori representation, it is absolutely impossible to know anything further as to what sort of being it is, or what its natural constitution may be; it is like the substantial, which remains behind after I have taken away all the accidents that inhere in it, but absolutely cannot be known

196 On de se and de re any further at all, since the accidents were precisely that whereby I was able to know its nature. Finally, the third passage concerns the subject of perception, which confirms that which has been explained thus far with reference to empirical self-knowledge: But the self in the second sense (as subject of perception), the psychological self as empirical consciousness, is capable of being known in many ways, among which time, the form of inner intuition, is that which underlies a priori all perceptions and their combination whose apprehension (apprehensio) conforms to the manner in which the subject is thereby affected, i.e., to the condition of time, in that the sensory self is determined by the intellectual to take up this condition into consciousness. (FM 20:270, 362) Even though the thinking being is one, the reference to the many ways in which the subject can be known highlights the different theoretical dimensions involved. On one hand, Kant’s conception of the inner sense is influenced by Descartes and Locke’s views. Moving from the famous Meditations on First Philosophy and its first-person methodology, Descartes’ use of the acquaintance theory calls attention to the fact that self-knowledge is epistemically special because it also serves to ground other types of knowledge. Instead, the inner sense theory points out the similarities between self-knowledge and perceptual knowledge: Inner sense is construed through the process of introspection, considered to be on par with the outer sense based on a causal process that is broadly similar to that involved in perceptual awareness. However, both Descartes and Locke employ, mutatis mutandis, a model of introspection regarded as observational; in both cases, the introspecting subject can be passive to herself due to the thoughts she introspects. As just seen, Kant points out that some kinds of self-knowledge can be achieved observationally through the inner sense or empirical apperception, but he acknowledges that the observational model depends on the apparatus of the transcendental apperception. Kant contrasts the observational model based on inner sense with a grasp of the activity of thinking based on pure apperception through a form of agential model in which, by means of the thinking activity, the thinking subject realises that it is the agent that has the intellectual power that forms the thoughts. It is the consciousness of the faculty of representational combination that introduces an agential self-knowledge, which is different from the observational model of inner sense. As explained in the first chapter, Kant inspired the rationalist theory of self-knowledge: Agential self-knowledge is related to the subject’s knowledge of her own rational agency considered as the thinking subject’s activity

On de se and de re  197 since the subject is conscious of her existence as an intelligence based on her power of combination. Therefore, the subject is aware of her own agency through agency, i.e., by performing thinking acts. As remarked previously, Thinking is the synthetic unity of apperception determining I think (qua Representation) as an analytical unity of apperception, i.e., I think is the representation through which spontaneity is given to one’s self, the consciousness through which the thinking activity is given. At this transcendental level, the self-knowledge expressed by I think concerns the fact that there is only a being devoid of properties. According to the representation I of apperception, the subject is able to know that it exists as a thinking being; through this consciousness alone, however, it is neither possible to know that there is anything on which this activity is based, nor— provided that there is anything at all—to know what it is. According to the reading embraced in these pages, the features assigned to the representation I do not allow for an epistemic conclusion; in particular, the representation of a subject as an essentially simple, identical substance detached from matter does not entail that the thinking self is a simple, identical substance detached from the body. The subjective identity only pertains to the subject as a formal identity of thought. In particular, the selfconsciousness expressed by ‘I think’ qua transcendental apperception is nothing but the consciousness of a thinking activity. In the absence of epistemic mediations of identification regarded as simple representations, I think (qua Representation) merely designates an activity—the activity of thinking transcendentally—precisely because this activity produces the simple and empty representation I. This is an empty form that makes all the difference, as Kant acknowledges in the above-mentioned famous passage that opens the Anthropology: The fact that the human being can have the “I” in his representations raises him infinitely above all other living beings on earth. Because of this he is a person, and by virtue of the unity of consciousness through all changes that happen to him, one and the same person- i.e., through rank and dignity an entirely different being from things, such as irrational animals, with which one can do as one likes. This holds even when he cannot yet say “I”, because he still has it in thoughts, just as all languages must think it when they speak in the first person, even if they do not have a special word to express this concept of “I”. For this faculty (namely to think) is understanding. But it is noteworthy that the child who can already speak fairly fluently nevertheless first begins to talk by means of “I” fairly late (perhaps a year later); in the meantime speaking of himself in the third person (Karl wants to eat, to walk, etc.). When he starts to speak by means of “I” a light seems to dawn on him, as it were, and from that day on he never again returns to his former way of speaking.—Before he merely felt himself, now he thinks himself. (Anth 7: 127, 15)

198 On de se and de re

Notes 1. In the mental case, the character—what, in the linguistic case, is the conventional meaning of the expression (type) based on the token-reflexive rule determining its reference—is the equivalent of the cognitive role of the representation (type) based on the token-reflexive rule that specifies its reference. Its content, on the other hand, is what the representation (token) represents, namely its reference. 2. This particular case involves two different types of knowledge. Animals, as well as the savage in the example, have the kennen—i.e., they represent what they experience—and can distinguish an object from another due to the different types of sensations behind the material dimension of the intuition. On the other hand, they are not able to gain knowledge and awareness, i.e., they do not possess the erkennen, as no concept is involved. 3. In several passages, Kant sharply distinguishes the noumenon from the transcendental object; for example, A 253: “The object to which I relate appearance in general is the transcendental object, i.e., the entirely undetermined thought of something in general. This cannot be called the noumenon; for I do not know anything about what it is in itself, and have no concept of it except merely that of the object of a sensible intuition in general, which is therefore the same for all appearances”. In A288–9/B344–5, Kant suggests that the transcendental object might also be considered as a noumenon only if it is correctly understood. Even if there are certain KrV passages in which Kant relates the transcendental object to the thing as it is in itself, for Allison (2004, 60), the concept of the transcendental object is introduced in order to address the issue of the “immanentization” of cognition: “The basic problem is that we cannot, as it were, stand outside our representations in order to compare them with some transcendentally real entity. Accordingly, such an object ‘must be thought only as something in general = X’ (A104), which is later identified with the transcendental object”. 4. “Objects can indeed appear to us without necessarily having to be related to functions of the understanding” (A89/B122); “Appearances can certainly be given in intuition without functions of the understanding” (A90/B122); “The manifold for intuition must already be given prior to the synthesis of the understanding and independently from it” (B145). Cf. Hanna (2001, 199; 2005, 259), Allais (2009, 396), Schulting (2012a, 84); however, cf. the antithetical conceptualist interpretations of A89/B122 by Allison (2001, 38) and Grüne (2011). 5. McDowell (1994, 9): “The original Kantian thought was that empirical knowledge results from a co-operation between receptivity and spontaneity. . . . We can dismount from the seesaw if we can achieve a firm grip on this thought: receptivity does not make an even notionally separable contribution to the cooperation. The relevant conceptual capacities are drawn on in receptivity. . . . It is not that they are exercised on an extra-conceptual deliverance of receptivity. We should understand what Kant calls ‘intuition’—experiential intake—not as a bare getting of an extra-conceptual Given, but as a kind of occurrence or state that already has conceptual content”. As pointed out by Allais, even if McDowell’s intention in Mind and World is theoretical rather than historical–interpretative, his reading has largely influenced the Kantian debate. 6. It goes without saying that this is not the only option within the debate. Lewis’ (1979) perspective reverses Perry and Recanati’s terms of the issue: All types of property ascriptions are mere variations of a single kind, namely de se ascriptions. It follows that de re thoughts should be reduced to de se thoughts. Within this line of thought, which internalises the contextual relations of acquaintance in the content of thought, one should also include Searle and Higginbotham’s

On de se and de re  199 (2003) reflexivity perspective. Among the many studies, see García-Carpintero (2013) for an overview of the different positions. 7. The position in Schema 2 suggests that explicit de se thoughts can be also immune to error through misidentification: such is Recanati’s (2009) view when explicit de se thoughts present the same way of grounding experience as an implicit de se thought. 8. In rejecting Shoemaker’s distinction between absolute and circumstantial immunity, Recanati distinguishes two types of implicit self-ascription, that is, two kinds of immunity to error through misidentification. These are called ‘grade1’ and ‘grade-2’ immunity. Although they are not usually distinguished, they should be shown in fantastic situations—rather than through a conceptual distinction—in which grade 1 IEM is legitimate and grade 2 is not: “In perception, one attributes to oneself the conscious state one is in; but there is another form of implicit self-ascription. . . . The perceiving subject does not merely ascribe to himself or herself the property of being in a certain inner state when she, for example, sees a flower in front of her. She also ascribes to herself the world-involving property of standing in a certain relation to the flower—a relation which makes it possible for the flower to cause the current conscious experience which the subject implicitly self-ascribes. The two types of implicit self-ascription are hard to distinguish in the perception case, because the inner state we are aware of reveals the world to us and our situation in it. The state is, as it were, ‘transparent’, and our awareness of it is also, simultaneously, awareness of what it reveals to us. In the case of memory the distinction is made easier due to the temporal distance between the inner memory state and what it reveals. What the inner state reveals is how the world was and how we experienced it in the past” (Recanati 2007, 151–52). 9. On this particular point, also see Perry (1986b). Perry (2002, 208) expounds the “Tractarian or Carnapian way of making this point” by asserting that “the world as we perceive it does not include ourselves, but has ourselves as sort of a point of origin”. As is well known, this is also, and principally, a Kantian issue, resumed—mutatis mutandis—by Schopenhauer and, through Schopenhauer, by Wittgenstein (cf. Frascolla 2007, 204–9). 10. Locke offers one of the plainest formulations of the reflexive model of selfconsciousness. Firstly, thinking takes on a broad sense, which ultimately encompasses all sorts of mental activities, and must be accompanied with the consciousness of thinking. Not unlike the Cartesian perspective, according to Locke, thinking consists of one’s being conscious of the fact of performing the action of thinking. Nonetheless, Locke sees in sensation and reflection the only two sources for the formation of ideas: In one case through the different perceptions of external objects and, in the other, based on the perception of the operations carried out by the mind concerning the ideas received. Reflection is defined as ‘internal sense’ by analogy with the faculty of sensation, i.e., the sense that receives the perceptions of external objects, and it consists of the capacity of the mind to direct its gaze towards itself and its own mental acts. With regard to this theory, consciousness is thus regarded as a perception-like second-order representation of psychological states; cf. Armstrong (1968) and Lycan (1996). 11. Thus, the first difficulty, which calls into question the petitio principi, is how to explain the identity of the two relata without presupposing what one would like to explain. The second difficulty is establishing how an act of consciousness on the part of an object can be the explicative basis for the self-consciousness of a subject. The second horn of the dilemma, which affirms the impossibility of self-consciousness, can be articulated on the basis of Zahavi’s argument (1999, 16): (1) awareness is a relationship between a subject (qua experiencing) and an

200 On de se and de re object (qua experienced); (2) if the subject is to be aware of itself, it must take itself as an object; (3) if the subject is aware of an object, it is not aware of itself; thus, (4) real self-awareness is impossible. Therefore, for Zahavi, if awareness is a relationship between a subject and an object, self-awareness is impossible: “The subject of experience can never truly be its own object, insofar as this would imply both a negation of its subjectivity, as well as a violation of the principle of identity”. 12. Cf. supra, Chapter 1, for the close relationship between the linguistic and the mental dimensions. Bermúdez points out the thought-language principle according to which, in order to understand the capacity to think I-thoughts, it is necessary to analyse the linguistic expression of the first-person pronoun. 13. “Wenn ich meinen Hebel an einem Punkt in mir selbst ansetze so kann ich mich selbst nicht von der Stelle bewegen. Es ist zwar facultas motrix aber nicht locomotiva die daraus entspringe. Daß das Bewußtsein ein Gefühl sei ist falsch denn [das] die Vorstellung meiner selbst ist bloß logisch um ein Objekt (an mir selbst) zu haben. Das Wort sum” (KGS 21: 484).

References

Abela, P. (2002), Kant’s Empirical Realism. Oxford: Clarendon Press/Oxford University Press. Adickes, E. (1889), Immanuel Kants Kritik der reinen Vernunft. Berlin: Mayer & Müller. Allais, L. (2009), “Kant, Non-Conceptual Content and the Representation of Space”. In Journal of the History of Philosophy 47 (3): 383–413. Allison, H. E. (1968), “Kant’s Concept of the Transcendental Object”. In KantStudien 59 (1–4): 165–86. Allison, H. E. (1983), Kant’s Transcendental Idealism: An Interpretation and Defense. New Haven: Yale University Press. Allison, H. E. (1996), Idealism and Freedom. Cambridge: Cambridge University Press. Allison, H. E. (2001), Kant’s Theory of Taste: A Reading of the Critique of Aesthetic Judgment. Cambridge: Cambridge University Press. Allison, H. E. (2004), Kant’s Transcendental Idealism: An Interpretation and Defense. Revised and Enlarged Edition. New Haven/London: Yale University Press. Ameriks, K. (1995), “From Kant to Frank: The Ineliminable Subject”, in K. Ameriks, D. Sturma (eds.), The Modern Subject: Conceptions of the Self in Classical German Philosophy. Albany: SUNY Press, pp. 217–30. Ameriks, K. (1997), “Kant and the Self: A Retrospective”, in D. Klemm, G. Zöller (eds.), Figuring the Self: Subject, Absolute, and Others in Classical German Philosophy. Albany: SUNY Press, pp. 55–72. Ameriks, K. (2000), Kant’s Theory of Mind. Second Edition. Oxford: Oxford University Press. Ameriks, K. (2004), “Apperzeption und Subjekt: Kants Lehre vom Ich”, in D. H. Heidemann, K. Engelhard (eds.), Warum Kant heute? Systematische Bedeutung und Rezeption seiner Philosophie in der Gegenwart. Berlin: de Gruyter, pp. 76–99. Repr. in K. Ameriks, Kant and the Historical Turn: Philosophy as Critical Interpretation. Oxford: Clarendon Press, 2006, pp. 51–66. Anderson, E. (1994), “Kant, Natural Kind Terms, and Scientific Essentialism”. In History of Philosophy Quarterly 11 (4): 355–73. Anderson, R. L. (2015), The Poverty of Conceptual Truth: Kant’s Analytic/Synthetic Distinction and the Limits of Metaphysics. Oxford: Oxford University Press.

202 References Anscombe, G. E. M. (1975), “The First Person”, in S. D. Guttenplan (ed.), Mind and Language. Oxford: Clarendon Press. Aquila, R. (1989), Matter and Mind. Bloomington: Indiana University Press. Armstrong, D. M. (1968), A Materialist Theory of the Mind. London: Routledge. Bach, K. (1987), Thought and Reference. Oxford: Oxford University Press. Baker, L. R. (2000), Persons and Bodies: A Constitution View. Cambridge: Cambridge University Press. Baker, L. R. (2013), Naturalism and the First-Person Perspective. New York: Oxford University Press. Baumgarten, A. (1739), Metaphysica. Hildesheim: Olms (1982). Becker, W. (1984), Selbstbewusstsein und Erfahrung: zu Kants transzendentaler Deduktion und ihrer argumentativen Rekonstruktion. Freiburg: K. Alber. Bencivenga, E. (1987), Kant’s Copernican Revolution. Oxford: Oxford University Press. Bennett, J. F. (1974), Kant’s Dialectic. New York: Cambridge University Press. Bermúdez, J. L. (1994), “The Unity of Apperception in the Critique of Pure Reason”. In European Journal of Philosophy 2: 213–40. Bermúdez, J. L. (1998), The Paradox of Self-Consciousness, Cambridge, MA: MIT Press. Bermúdez, J. L. (2003), “Nonconceptual Content: From Perceptual Experience to Subpersonal Computational States”, in Y. H. Gunther (ed.), Essays in Nonconceptual Content. Cambridge, MA: MIT Press. Bermúdez, J. L. (2012), “Nonconceptual Mental Content”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy. Available at http://plato.stanford.edu/ archives/spr2012/entries/content-nonconceptual/. Bermúdez, J. L. (2017), Understanding ‘I’: Language and Thought. Oxford: Oxford University Press. Bermúdez, J. L. (2018), The Bodily Self, Selected Essays. Cambridge, MA: MIT Press. Bird, A., Tobin, E. (2012), “Natural Kinds”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy. Available at http://plato.stanford.edu/archives/win2012/ entries/natural-kinds. BonJour, L. (2003), “Back to Foundationalism”, in E. Sosa, L. Bonjour (eds.), Epistemic Justification: Internalism vs. Externalism, Foundations vs. Virtues. Malden, MA: Blackwell. Bonomi, A. (1991), “Sul cogito cartesiano: natura inferenziale e criteri di giustificabilità”, in G. Usberti (a c. di), Problemi fondazionali nella teoria del significato. Firenze: Olschki Editore, pp. 7–33. Brandt, R. (1995), The Table of Judgments: Critique of Pure Reason A67–76/B92– 101. Trans. and ed. E. Watkins. North American Kant Society Studies in Philosophy, Vol 4. Atascadero: Ridgeview Publishing Company. Brook, A. (1994), Kant and the Mind. Cambridge: Cambridge University Press. Brook, A. (2001), “Kant, Self-Awareness and Self-Reference”, in A. Brook, R. C. DeVidi (eds.), Self-Reference and Self-Awareness. Philadelphia: John Benjamins. Brook, A. (2008), “Kant’s View of the Mind and Consciousness of Self”, in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy. Winter 2008 Edition. Available at http://plato.stanford.edu/archives/win2008/entries/kant-mind/. Burge, T. (1977), “Belief de re”. In Journal of Philosophy 74 (6): 338–62. Burge, T. (1979), “Sinning Against Frege”. In Philosophical Review 88 (3): 398–432.

References  203 Campbell, J. (1994), Past, Space and Self. Cambridge, MA: MIT Press. Capozzi, M. (1980), “Kant on Mathematical Definition”, in M. L. Dalla Chiara (a c. di), Italian Studies in the Philosophy of Science. Dordrecht: Reidel Publishing Company. Capozzi, M. (2002), Kant e la logica. Napoli: Bibliopolis. Capozzi, M. (2007), “L’io e la conoscenza di sé in Kant”, in E. Canone (a c.di.), Per una storia del concetto di mente. Firenze: Olschki. Capozzi, M. (2009), “La teoria kantiana dei concetti e il problema dei nomi propri”. In Dianoia 14: 119–46. Capozzi, M. (forthcoming), “Singular Term and Intuitions in Kant: A Reappraisal”, in C. Posy, O. Rechter (eds.), Kant’s Philosophy of Mathematics. Cambridge: Cambridge University Press. Capozzi, M., Roncaglia, G. (2009), “Logic and Philosophy of Logic from Humanism to Kant”, in L. Haaparanta (ed.), The Development of Modern Logic. Oxford: Oxford University Press. Carl, W. (1989a), Der schweigende Kant. Göttingen: Vandenhoeck & Ruprecht. Carl, W. (1989b), “Kant’s First Drafts of the Deduction of the Categories”, in E. Förster (ed.), Kant’s Transcendental Deductions. The Three “Critiques” and the “Opus postumum”. Stanford: Stanford University Press, pp. 3–20. Carl, W. (1997), “Apperception and Spontaneity”. In International Journal of Philosophical Studies 5 (2): 147–63. Carl, W. (1998), “Die transzendentale Deduktion in der zweiten Auflage (B129– 169)”, in G. Mohr, M. Willaschek (eds.), Klassiker Auslegen: Immanuel Kant, Kritik der reinen Vernunft. Berlin: Akademie Verlag, pp. 189–216. Carruthers, P. (2005), Consciousness: Essays from a Higher-Order Perspective. Oxford: Oxford University Press. Cassam, Q. (1993), “Inner Sense, Body Sense, and Kant’s Refutation of Idealism”. In European Journal of Philosophy 1: 111–27. Cassam, Q. (ed.) (1994), Self-Knowledge. Oxford: Oxford University Press. Cassam, Q. (1997), Self and World. Oxford: Oxford University Press. Castañeda, H-N. (1966), “ ‘He’: A Study in the Logic of Self-Consciousness”, in H-N. Castañeda (1999), pp. 35–60. Castañeda, H-N. (1967), “Indicators and Quasi-Indicators”, in H-N. Castañeda (1999), pp. 61–88. Castañeda, H-N. (1968), “On the Phenomeno-Logic of the I”, in H-N. Castañeda (1999), pp. 89–95. Castañeda, H-N. (1983), “Reply to John Perry: Meaning, Belief, and Reference”, in J. Tomberlin (ed.), Agent, Language, and the Structure of the World. Indianapolis: Hackett Publishing Company, pp. 313–28. Castañeda, H-N. (1987), “Self-Consciousness, Demonstrative Reference, and the Self-Ascription View of Believing”, in H-N. Castañeda (1999), pp. 143–79. Castañeda, H-N. (1989), Thinking, Language, and Experience. Minneapolis: University of Minnesota Press. Castañeda, H-N. (1990), “Indexicality: The Transparent Subjective Mechanism for Encountering a World”. In Noûs 5 (24): 735–49. Castañeda, H-N. (1999), The Phenomeno-Logic of the I: Essay on Self-Consciousness. Bloomington: Indiana University Press. Chalmers, D. J. (2003), “The Content and Epistemology of Phenomenal Belief”, in Q. Smith, A. Jokic (eds.), Consciousness: New Philosophical Perspectives. Oxford: Oxford University Press.

204 References Chisholm, R. (1976), Person and Object: A Metaphysical Study. Chicago/La Salle: Open Court. Chisholm, R. (1979), “Objects and Persons: Revisions and Replies”, in E. Sosa (ed.), Essays on the Philosophy of Roderick M. Chisholm. Amsterdam: Rodopi, pp. 317–88. Chisholm, R. (1981), The First Person. Minneapolis: University of Minnesota Press. de Gaynesford, M. (2006), I: The Meaning of the First Person Term. Oxford: Oxford University Press. De Sá Pereira, R. H. (2013), “What Is Nonconceptualism in Kant’s Philosophy?” In Philosophical Studies 164 (1): 233–54. Descartes, R. (1641), Meditations on First Philosophy: With Selections from the Objections and Replies. Trans. with an Introduction and Notes by Michael Moriarty. Oxford: Oxford University Press (2008). Dyck, C. W. (2017), “The Principles of Apperception”, in Udo Thiel, Giuseppe Motta (eds.), Immanuel Kant: Die Einheit des Bewusstseins (Kant-Studien Ergänzungshefte). Berlin: de Gruyter, pp. 32–46. Ellis, A. (2017), “The Case for Absolute Spontaneity in Kant’s Critique of Pure Reason”. In Con-Textos Kantianos 6: 138–64. Emundts, D. (2006), “Die Paralogismen und die Widerlegung des Idealismus in Kants ‘Kritik der reinen Vernunft’ ”. In Deutsche Zeitschrift für Philosophie 54 (2): 295–309. Evans, G. (1982), The Varieties of Reference. Oxford: Oxford University Press. Evans, J. C. (1990), “Two-Steps-in-One-Proof: The Structure of the Transcendental Deduction of the Categories”. In Journal of the History of Philosophy 28 (4): 553–70. Feldman, R. (2004), “Foundational Justification”, in J. Greco (ed.), Ernest Sosa and His Critics. Malden, MA: Blackwell. Fodor, J. (1995), The Elm and the Expert. Cambridge, MA: MIT Press/Bradford Books. Frank, M. (2004), “Fragments of a History of the Theory of Self-Consciousness from Kant to Kierkegaard”. In Critical Horizons 5 (1): 53–136. Frascolla, P. (2007), Understanding Wittgenstein’s Tractatus. London: Routledge. Frege, G. (1918–19), “The Thought: A Logical Enquiry”, in P. F. Strawson (ed.), Philosophical Logic. Oxford: Oxford University Press (1967). Freundlieb, D. (2003), Dieter Henrich and Contemporary Philosophy: The Return to Subjectivity. Aldershot: Ashgate. Fumerton, R. (1995), Metaepistemology and Skepticism. Lanham, MD: Rowman & Littlefield. García-Carpintero, M. (2013), “The Self File and Immunity to Error Through Misidentification”. In Disputatio 5 (36): 191–206. Geach, P. (1957), Mental Acts: Their Content and Their Objects. London: Routledge. Gensini, S. (2000), ‘De linguis in universum’. On Leibniz’s Ideas on Languages. Five Essays. Münster: Nodus Publikationen. Gertler, B. (2001), “Introspecting Phenomenal States”. In Philosophy and Phenomenological Research 63: 305–28. Gertler, B. (2011), Self-Knowledge. London: Routledge. Gertler, B. (2017), “Self-Knowledge”, in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy. Fall 2017 Edition. Available at https://plato.stanford.edu/ archives/fall2017/entries/self-knowledge/.

References  205 Ginsborg, H. (2006), “Empirical Concepts and the Content of Experience”. In European Journal of Philosophy 14: 349–72. Ginsborg, H. (2008), “Was Kant a Nonconceptualist?” In Philosophical Studies 137 (1): 65–77. Ginsborg, H. (2015), The Normativity of Nature: Essays on Kant’s Critique of Judgement. Oxford: Oxford University Press. Goldman, A. (2006), Simulating Minds. New York: Oxford University Press. Griffith, A. (2012), “Perception and the Categories: A Conceptualist Reading of Kant’s Critique of Pure Reason”. In European Journal of Philosophy 20: 193–222. Grüne, S. (2009), Blinde Anschauung: Die Rolle von Begriffen in Kants Theorie sinnlicher Synthesis. Frankfurt am Main: Klostermann. Grüne, S. (2011), “Is There a Gap in Kant’s B-Deduction?” In International Journal of Philosophical Studies 19 (3): 465–90. Gunther, Y. H. (2003), Essays on Nonconceptual Content. Cambridge, MA: MIT Press. Guyer, P. (1987), Kant and the Claims of Knowledge. Cambridge: Cambridge University Press. Guyer, P. (1992), “The Transcendental Deduction of the Categories”, in P. Guyer (ed.), The Cambridge Companion to Kant. Cambridge: Cambridge University Press, pp. 123–60. Guyer, P. (2010), “The Deduction of Categories: The Metaphysical and Transcendental Deductions”, in P. Guyer (ed.), The Cambridge Companion to Kant’s Critique of Pure Reason. Cambridge: Cambridge University Press, pp. 118–50. Hacker, P. M. S. (1972), Insight and Illusion. Oxford: Clarendon Press. Hanna, R. (1998), “A Kantian Critique of Scientific Essentialism”. In Philosophy and Phenomenological Research 58 (3): 497–528. Hanna, R. (2000), “The Inner and the Outer: Kant’s Refutation Reconstructed”. In Ratio 2: 146–74. Hanna, R. (2001), Kant and the Foundations of Analytic Philosophy. Oxford: Oxford University Press. Hanna, R. (2005), “Kant and Nonconceptual Content”. In European Journal of Philosophy 13 (2): 247–90. Hanna, R. (2006), Kant, Science, and Human Nature. Oxford: Clarendon Press. Hanna, R. (2008), “Kantian Non-Conceptualism”. In Philosophical Studies 137 (1): 41–64. Hanna, R. (2011), “Kant’s Non-Conceptualism, Rogue Objects, and the Gap in the B-Deduction”. In International Journal of Philosophical Studies 19 (3): 399–415. Hanna, R. (2014), “The Togetherness Principle, Kant’s Conceptualism, and Kant’s Non-Conceptualism: Supplement to Kant’s Theory of Judgment”, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy. Summer 2014 Edition. Available at http://plato.stanford.edu/archives/sum2014/entries/kant-judgment/. Hanna, R. (2016), “Kantian Madness: Blind Intuitions, Essentially Rogue Objects, Nomological Deviance, and Categorial Anarchy”. In Contemporary Studies in Kantian Philosophy 1: 44–64. Heck, R. G. (2000), “Nonconceptual Content and the ‘Space of Reasons’ ”. In Philosophical Review 109 (4): 483–523. Heckmann, H-D. (1981), Was ist Wahrheit? Eine systematisch-kritische Untersuchung philosophischer Wahrheitsmodelle. Heidelberg: Winter. Henrich, D. (1966), “Fichtes ursprüngliche Einsicht”, in D. Henrich, H. Wagner (hrsg. v.), Subjektivität und Metaphysik. Festschrift für Wolfgang Cramer.

206 References Frankfurt am Main: Klostermann; English Trans. “Fichte’s Original Insight”, in D. E. Christensen (ed.), Contemporary German Philosophy. Vol I. University Park: The Pennsylvania State University Press (1982), pp. 15–53. Henrich, D. (1969), “The Proof-Structure of the Transcendental Deduction”. In Review of Metaphysics 22: 640–59. Henrich, D. (1976), Identität und Objektivität: Eine Untersuchung über Kants transzendentale Deduktion. Heidelberg: Winter. Henrich, D. (1982), Selbstverhältnisse: Gedanken und Auslegungen zu den Grundlagen der klassischen deutschen Philosophie. Stuttgart: Reclam. Henrich, D. (1989), “The Identity of the Subject in the Transcendental Deduction”, in E. Schaper, W. Vossenkuhl (eds.), Reading Kant. Oxford: Blackwell. Higginbotham, J. (2003), “Remembering, Imagining, and the First Person”, in A. Barber (ed.), Epistemology of Language. Oxford: Oxford University Press, pp. 496–533. Hintikka, J. (1962), “Cogito, Ergo Sum: Inference or Performance?” In Philosophical Review LXXI: 3–32. Hintikka, J. (1967), “Kant on the Mathematical Method”. In The Monist 51: 352–75. Hintikka, J. (1969), “On Kant’s Notion of Intuition (anschauung)”, in T. Penelhum, J. J. MacIntosh (eds.), The First Critique: Reflections on Kant’s Critique of Pure Reason. Belmont: Wadsworth Publishing Company. Hintikka, J. (1972), “III. Kantian Intuitions”. In Inquiry: An Interdisciplinary Journal of Philosophy 15 (1–4): 341–5. Hohenegger, H. (2013), “La terminologia della coscienza in Kant: pars destruens”, in R. Palaia (a cura di.), Coscienza nella filosofia della prima modernità (LEI). Vol 119. Firenze: Olschki, pp. 135–68. Horstmann, R. P. (1993), “Kants Paralogismen”. In Kant-Studien 83: 408–25. Horstmann, R. P. (2007), “Kant und Carl über Apperzeption”, in J. Stolzenberg (hrsg.), Kant in der Gegenwart. Berlin: de Gruyter, pp. 131–48. Horstmann, R. P. (2010), “The Limited Significance of Self-Consciousness”. In Revue de Métaphysique et de Morale 68 (4): 435–54. Howell, R. (1973), “Intuition, Synthesis, and Individuation in the Critique of Pure Reason”. In Noûs 7 (3): 207–32. Howell, R. (1981), “Kant’s First Critique Theory of the Transcendental Object”. In Dialectica 35 (1): 85–125. Howell, R. (1992), Kant’s Transcendental Deduction. Boston: Kluwer Academic Publishers. Howell, R. (2000), “Kant, the I Think, and Self-Awareness”, in P. Cicovacki (ed.), Kant’s Legacy: Essays in Honor of Lewis White Beck. Rochester: University of Rochester Press. Hume, D. (1738-40), A Treatise of Human Nature, ed. L. A. Selby-Bigge and P. H. Nidditch. Oxford: Clarendon Press, 1993. James, W. (1890), The Principles of Psychology. New York: Cosimo (2007). Jeshion, R. (2010), “Singular Thought: Acquaintance, Semantic Instrumentalism, and Cognitivism”, in R. Jeshion (ed.), New Essays on Singular Thought. Oxford: Oxford University Press. Kannisto, T. (2018), “Transcendental Paralogisms as Formal Fallacies—Kant’s Refutation of Pure Rational Psychology”. In Kant-Studien 109 (2): 195–227. Kapitan, T. (2001), “Indexical Identification: A Perspectival Account”. In Philosophical Psychology 14 (3): 293–312.

References  207 Kapitan, T. (2008), “Perry, Castañeda, and ‘I’ ”, in The XVIIIth Edition of the InterUniversity Workshop on Philosophy and Cognitive Science. Available at http:// fs-morente.filos.ucm.es/actividades/2008/conference/papers/Kapitan.pdf. Kaplan, D. (1989), “Demonstratives and Afterthoughts”, in J. Almog, J. Perry, H. Wettstein (eds.), Themes from Kaplan. Oxford: Oxford University Press, pp. 481–614. Keller, P. (2001), Kant and the Demands of Self Consciousness. Cambridge: Cambridge University Press. Kenny, A. (1968), Descartes: A Study of His Philosophy. New York: Random House. Kitcher, P. (1984), “Kant’s Real Self”, in A. W. Wood (ed.), Self and Nature in Kant’s Philosophy. Ithaca, NY: Cornell University Press, pp. 113–47. Kitcher, P. (1990), Kant’s Transcendental Psychology. New York: Oxford University Press. Kitcher, P. (1998), “Kant’s Cognitive Self”, in P. Kitcher (ed.), Critique of Pure Reason: Critical Essays. Lanham, MD: Rowman & Littlefield. Kitcher, P. (2000), “On Interpreting Kant’s Thinker as Wittgenstein’s ‘I’ ”. In Philosophy and Phenomenological Research 61 (1): 33–63. Kriegel, U. (2003), “Consciousness as Intransitive Self-Consciousness: Two Views and an Argument”. In Canadian Journal of Philosophy 33: 103–32. Kriegel, U. (2007), “Self-Consciousness”, in J. Fieser, D. Bradley (eds.), Internet Encyclopedia of Philosophy. Available at www.iep.utm.edu/self-con/. Kripke, S. (1980), Naming and Necessity. Cambridge, MA: Harvard University Press. Kripke, S. (2011), “The First Person”, in S. Kripke (ed.), Philosophical Troubles, Volume 1. New York: Oxford University Press. Kroon, F., Nola, E. (1987), “Kant, Kripke, and Gold”. In Kant-Studien 78: 442–58. La Rocca, C. (2013), “Kant on Self-Knowledge and Conscience”, in Dieter Hüning, S. Klingner, C. Olk (hrsg. v.), Das Leben der Vernunft. Beiträge zur Philosophie Kants. Berlin-Boston: de Gruyter, pp. 364–85. Land, T. (2011), “Kantian Conceptualism”, in Abel et al. (eds.), Rethinking Epistemology. Berlin: de Gruyter. Levine, J. (2007), “Phenomenal Concepts and the Materialist Constraint”, in T. Alter, S. Walter (eds.), Phenomenal Concepts and Phenomenal Knowledge. New York: Oxford University Press. Lewis, D. (1979), “Attitudes De Dicto and De Se”, in D. Lewis (ed.), Philosophical Papers. Oxford: Oxford University Press, pp. 133–59. Loar, B. (1976), “The Semantics of Singular Terms”. In Philosophical Studies 30: 353–77. Locke, J. (1690), Essay Concerning Human Understanding, ed. P. H. Nidditch. Oxford: Clarendon Press. Longuenesse, B. (1998), Kant and the Capacity to Judge: Sensibility and Discursivity in the Transcendental Analytic of the Critique of Pure Reason. Princeton: Princeton University Press. Longuenesse, B. (2006), “Self-Consciousness and Consciousness of One’s Own Body: Variations on a Kantian Theme”. In Philosophical Topics 34: 283–309. Longuenesse, B. (2012), “Two Uses of ‘I’ as Subject?” in S. Prosser, F. Recanati (eds.), Immunity to Error Through Misidentification. Cambridge: Cambridge University Press. Longuenesse, B. (2017), I, Me, Mine: Back to Kant, and Back Again. Oxford: Oxford University Press. Lycan, W. (1996), Consciousness and Experience. Cambridge, MA: MIT Press. Markie, P. (1986), Descartes’s Gambit. Ithaca, NY: Cornell University Press.

208 References Markie, P. (1992), “The Cogito and Its Importance”, in J. Cottingham (ed.), The Cambridge Companion to Descartes. Cambridge: Cambridge University Press, pp. 144–73. Marshall, C. (2010), “Kant’s Metaphysics of the Self”. In Philosophers’ Imprint 10 (8): 1–21. Marshall, C. (2013), “Kant’s One Self and the Appearance/Thing-in-Itself Distinction”. In Kant-Studien 104 (4): 421–41. Martin, M. (1992), “Perception, Concepts, and Memory”. In The Philosophical Review 101 (4): 745–63. Matthews, G. B. (1992), Thought’s Ego in Augustine and Descartes. Ithaca, NY: Cornell University Press. McDowell, J. (1994), Mind and World. Cambridge: Cambridge University Press. McGinn, C. (1993), The Problem of Consciousness: Essays Towards a Resolution. Oxford: Blackwell. McLear, C. (2011), “Kant on Animal Consciousness”. In Philosophers’ Imprint 11 (15): 1–16. Melnick, A. (2009), Kant’s Theory of the Self. London: Routledge. Musholt, K. (2015), Thinking About Oneself: From Nonconceptual Content to the Concept of Self. Cambridge, MA: MIT Press. Nichols, S., Stich, S. (2003), Mindreading. Oxford: Oxford University Press. Parsons, C. (1969), “Kant’s Philosophy of Arithmetic”, in S. Morgenbesser, P. Suppes, M. White (eds.), Philosophy, Science and Method: Essays in Honor of Ernest Nagel. New York: St. Martin’s Press. Parsons, C. (1984), “Arithmetic and the Categories”. In Topoi 3 (2): 109–21. Parsons, C. (2012), From Kant to Husserl: Selected Essays. Cambridge, MA: Harvard University Press. Paton, H. J. (1936), Kant’s Metaphysics of Experience: A Commentary on the First Half of the Kritik der reinen Vernunft. 2 vols. London: George Allen & Unwin. Peacocke, C. (1983), “Castañeda on He and I”, in J. Tomberlin (ed.), Agent, Language, and the Structure of the World. Indianapolis: Hackett Publishing, pp. 15–42. Peacocke, C. (1992), A Study of Concepts. Cambridge, MA: MIT Press. Peacocke, C. (1999), Being Know. New York: Oxford University Press. Peacocke, C. (2008), Truly Understood. Oxford: Oxford University Press. Peacocke, C. (2014), The Mirror of the World. Oxford: Oxford University Press. Pereboom, D. (2001), “Assessing Kant’s Master Argument”. In Kantian Review 5: 90–102. Pereboom, D. (2006), “Kant’s Metaphysical and Transcendental Deductions”, in G. Bird (ed.), A Companion to Kant. Oxford: Blackwell, pp. 154–68. Perry, J. (1977), “Frege on Demonstratives”. In Philosophical Review 86 (4): 474–97. Perry, J. (1979), “The Problem of the Essential Indexical”. In Noûs 13 (1): 3–21. Perry, J. (1986a), “Perception, Action, and the Structure of Believing”, in R. Grandy, R. Warner (eds.), Philosophical Grounds of Rationality. Oxford: Oxford University Press. Perry, J. (1986b), “Thought Without Representation”. In Proceedings of the Aristotelian Society, Supplementary 60: 137–51. Perry, J. (1997), “Indexicals and Demonstratives”, in B. Hale, C. Wright (eds.), A Companion to the Philosophy of Language. Oxford: Blackwell, pp. 586–612. Perry, J. (2000), The Problem of the Essential Indexical and Other Essays. Expanded Edition. Stanford: CSLI Publications.

References  209 Perry, J. (2002), Identity, Personal Identity, and the Self. Indianapolis: Hackett Publishing. Pitt, D. (2004), “The Phenomenology of Cognition, or, What Is It Like to Think That P?” In Philosophy and Phenomenological Research 69: 1–36. Powell, C. T. (1990), Kant’s Theory of Self-Consciousness. Oxford: Oxford University Press. Prauss, G. (1969), “Zum Wahrheitsproblem bei Kant”. In Kant-Studien 60: 166–82. Prauss, G. (1971), Erscheinung bei Kant. Berlin: de Gruyter. Prauss, G. (1974), Kant und das problem der dinge an sich. Bonn: Bouvier. Prauss, G. (2015), Die Einheit von Subjekt und Objekt. Kants Probleme mit den Sachen selbst. Freiburg: Verlag Karl Alber. Putnam, H. (1975), “The Meaning of ‘Meaning’ ”, in H. Putnam (ed.), Philosophical Papers, Vol II. Cambridge: Cambridge University Press. Quine, W. V. (1956), “Quantifiers and Propositional Attitudes”. In The Journal of Philosophy 53: 177–87. Recanati, F. (2007), Perspectival Thought: A Plea for (Moderate) Relativism. New York: Oxford University Press. Recanati, F. (2009), “De re and De se”. In Dialectica 63: 249–69. Reich, K. (1992), The Completeness of Kant’s Table of Categories. Stanford: Stanford University Press. Rödl, S. (2007), Self-Consciousness. Cambridge, MA: Harvard University Press. Rosefeldt, T. (2000), Das Logische Ich: Kant über den Gehalt des Begriffes von sich selbst. Berlin: Philo. Rosefeldt, T. (2003), “Kant’s Self: Real Entity and Logical Identity”, in H. J. Glock (ed.), Strawson and Kant. Oxford: Oxford University Press. Rosefeldt, T. (2006), “Kants Ich als Gegenstand”. In Deutsche Zeitschrift für Philosophie 54 (2): 277–93. Rosefeldt, T. (2017), “Subjects of Kant’s First Paralogism”, in A. Gomes, A. Stephenson (eds.), Kant and the Philosophy of Mind: Perception, Reason, and the Self. Oxford: Oxford University Press. Rosenberg, J. F. (1986), “ ‘I Thinks’: Some Reflections on Kant’s Paralogisms”. In Midwest Studies in Philosophy 10 (1): 503–30. Rosenberg, J. F. (2005), Accessing Kant: A Relaxed Introduction to the Critique of Pure Reason. Oxford: Oxford University Press. Rosenthal, D. (1993), “Thinking That One Thinks”, in M. Davies, G. W. Humphreys (eds.), Consciousness: Psychology e Philosophical Essay. Oxford: Blackwell, pp. 198–223. Rovane, C. (1987), “The Epistemology of First-Person Reference”. In Journal of Philosophy 84: 147–67. Russell, B. (1912), Problems of Philosophy. New York: Henry Holt & Co. Savile, A. (1974), “Kant, Truth and Affinity”, in G. Funke (ed.), Akten des 4. Internationalen Kant-Kongresses. Vol 2. Berlin: de Gruyter, pp. 336–43. Schmitz, F. (2013), “On Kant’s Conception of Inner Sense: Self-Affection by the Understanding”. In European Journal of Philosophy 23 (4): 1044–63. Schulting, D. (2011a), “Limitation and Idealism: Kant’s ‘Long’ Argument from the Categories”, in D. Schulting, J. Verburgt (2011). Schulting, D. (2011b), “Kant’s Idealism: The Current Debate”, in D. Schulting, J. Verburgt (2011).

210 References Schulting, D. (2012a), “Kant, Non-Conceptual Content, and the ‘Second Step’ of the B-Deduction”. In Kant Studies Online 51–92. Schulting, D. (2012b), Kant’s Deduction from Apperception: On Explaining the Categories. Basingstoke/New York: Palgrave Macmillan. Schulting, D. (ed.) (2016), Kantian Nonconceptualism. New York: Palgrave Macmillan. Schulting, D. (2017a), “Apperception, Self-Consciousness, and Self-Knowledge in Kant”, in Matthew Altman (ed.), The Palgrave Kant Handbook. New York: Palgrave Macmillan, pp. 139–61. Schulting, D. (2017b), Kant’s Radical Subjectivism Perspectives on the Transcendental Deduction. Basingstoke/New York: Palgrave Macmillan. Schulting, D., Verburgt, J. (eds.) (2011), Kant’s Idealism: New Interpretations of a Controversial Doctrine. Dordrecht/Heidelberg: Springer Science. Schulz, G. (1993), Veritas est adaequatio intellectus et rei: Untersuchungen zur Wahrheitslehre des Thomas von Aquin und zur Kritik Kants an einem überlieferten Wahrheitsbegriff. Leiden: Brill. Sellars, W. (1963), “Empiricism and the Philosophy of Mind”, in W. Sellars (ed.), Science, Perception, and Reality. New York: Humanities Press. Sellars, W. (1968), Science and Metaphysics: Variations on Kantian Themes. London: Routledge & Kegan Paul. Shoemaker, S. (1968), “Self-Reference and Self-Awareness”. In Journal of Philosophy 65: 555–67. Sluga, H. (1996), “Whose House Is That? Wittgenstein on the Self”, in H. Sluga, D. G. Stern (eds.), The Cambridge Companion to Wittgenstein. Cambridge: Cambridge University Press. Speaks, J. (2005), “Is There a Problem About Nonconceptual Content?” In The Philosophical Review 114: 359–98. Strawson, P. F. (1966), The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason. London: Methuen. Sturma, D. (1985), Kant über Selbstbewusstsein. New York: Georg Olms Verlag. Thompson, M. (1972), “Singular Terms and Intuitions in Kant’s Epistemology”. In Review of Metaphysics 26 (2): 314–43. Tye, M. (2006), “Nonconceptual Content, Richness, and Fineness of Grain”, in T. Gendler, J. Hawthorne (eds.), Perceptual Experience. Oxford: Oxford University Press. Valaris, M. (2008), “Inner Sense, Self-Affection, and Temporal Consciousness in Kant’s Critique of Pure Reason”. In Philosophers’ Imprint 8 (4): 1–18. Van Cleve, J. (1999), Problems from Kant. Oxford: Oxford University Press. Vanzo, A. (2012), “Kant on Truth-Aptness”. In History and Philosophy of Logic 33 (2): 109–26. Wagner, H. (1980), “Der Argumentationsgang in Kants Deduktion der Kategorien”. In Kant-studien 71 (3): 352–66. Waibel, V. L. (2017), “Das reine Selbst, die Kausalität des Begriffs und die Zeit”, in Udo Thiel, Giuseppe Motta (eds.), Immanuel Kant: Die Einheit des Bewusstseins (Kant-Studien Ergänzungshefte). Berlin: de Gruyter, pp. 236–57. Ware, O. (2009), “The Duty of Self-Knowledge”. In Philosophy and Phenomenological Research 79 (3): 671–98.

References  211 Washburn, M. C. (1970), The Problem of Self-Knowledge and the Evolution of the Critical Epistemology: 1781 to 1787. Unpublished Ph.D. dissertation, University of California. Wettstein, R. H. (1980), Eine Gegenstandstheorie der Wahrheit: Argumentativ rekonstruierender Aktualisierungs- und Erweiterungsversuch von Kants kritischer Theorie. Königstein: Forum Academicum. Williams, B. (1978), Descartes: The Project of Pure Inquiry. Abingdon: Routledge. Williams, J. (2012), “How Conceptually-Guided Are Kantian Intuitions?” In History of Philosophy Quarterly 29: 57–78. Wilson, M. D. (1978), Descartes. London: Routledge. Wittgenstein, L. (1953), Philosophical Investigations. Oxford: Blackwell. Wittgenstein, L. (1958), The Blue and the Brown Books. Oxford: Blackwell. Wolff, M. (2006), “Empirischer und Transzendentaler Dualismus”. In Deutsche Zeitschrift für Philosophie 54 (2): 265–75. Wright, C. (1998), “Self-Knowledge: The Wittgensteinian Legacy”, in C. Wright, B. Smith, C. Macdonald (eds.), Knowing Our Own Minds. Oxford: Oxford University Press. Wunderlich, F. (2017), “Kant on Consciousness of Objects and Consciousness of the Self”, in Udo Thiel, Giuseppe Motta (eds.), Immanuel Kant: Die Einheit des Bewusstseins (Kant-Studien Ergänzungshefte). Berlin: de Gruyter, pp. 164–80. Zahavi, D. (1999), Self-Awareness and Alterity: A Phenomenological Investigation. Evanston: Northwestern University Press.

Index

Allison, H. E. 34 – 8, 45n10, 46n15, 50, 69, 72n6, 74n15, 98, 102n4, 102n5, 102n7, 145 – 6, 155, 198n3, 198n4; on inner sense 94 – 5, 177 – 8, 189, 191; on transcendental deduction 35 – 6, 47n20, 47n21, 47n22; on two-aspect view 65 – 7, 186; see also formal ownership reading; self-affection Ameriks, K. 38, 43, 50, 63 – 5, 72n6, 74n11, 106, 166, 169 Anscombe, G. E. M. 16, 80, 83 appearance 25 – 30, 41, 46n15, 62, 65 – 7, 74n15, 77, 85 – 6, 93 – 6, 99, 145 – 6, 152 – 4, 168, 178, 182 – 4, 187 – 9, 192 – 4, 198n3 apperception, transcendental unity of 11, 21, 24, 28 – 31, 31, 36, 37 – 43, 46n19, 47n19, 48n22, 49, 52 – 4, 56, 63, 65 – 6, 69 – 71, 71n2, 72n3, 74n10, 74n13, 75 – 7, 79, 82 – 3, 86, 88, 93, 96, 105, 131, 136 – 9, 143, 160 – 1, 164 – 5, 168 – 70, 175 – 6, 180, 185 – 6, 188, 195, 197; empirical 60 – 2, 66, 77, 86, 92 – 3, 96 – 101, 165, 178 – 9, 183 – 95 attention 43n2, 95, 102n5, 183, 191 – 3 Baker, L. R. 12 – 13 Berkeley, G. 85, 120, 192 Bermúdez, J. L. 44n5, 50, 110, 139n2, 176, 200n12; on the paradox of self-consciousness 170 – 3 Brook, A. 11, 24, 71n2, 79, 106 Burge, T. 43n3, 143 Capozzi, M. 55, 63, 86, 95, 102n4, 114 – 17, 125 – 6, 129, 140n6, 140n7, 142n13, 183 – 95 Carl, W. 38, 47n22, 48n22, 92, 100, 164

Cassam, Q. 50, 82 – 3, 89, 102, 167 Castañeda, H-N. 16 – 17, 43n4, 44n7, 143, 156 – 7, 168 Chisholm, R. 50, 158 concept 23 – 5, 32 – 4, 40 – 2, 45n11, 45n12, 47n19, 58, 73n7, 73n8, 74n15, 98 – 9, 102n7, 102n8, 109 – 15, 117, 121 – 9, 140n4, 140n6, 141n12, 142n14, 142n15, 142n16, 145 – 6; singular concept 58 – 9, 115 – 17, 129, 141n7 conceptualism vs. non-conceptualism 109 – 13, 139n2, 139n3; weak conceptualism vs. non-conceptualism 148 – 52 Descartes, R. 10, 29, 34, 83 – 4, 192, 196; on the cogito 61, 73n9, 79 – 83 direct reference theory 103 – 9, 118 – 20, 127 – 39 Evans, G. 18 – 19, 79, 82, 84, 89 – 91, 101n1, 110, 143, 158 first-person perspective 12, 14 – 15, 18, 38, 84, 86, 136 – 7, 168, 172, 174 formal ownership reading 50 – 1, 88, 133, 184 Frege, G. 44n6, 118, 139n3, 143, 156 – 7 Ginsborg, H. 149, 151, 179 – 83, 194 Guyer, P. 46n18, 46n19, 47n20, 47n22, 48n22, 50, 72n3 Hanna, R. 102n4, 102n9, 110 – 14, 119 – 21, 123, 139n2, 139n3, 141n9, 141n11, 142n14, 143, 145, 148 – 50, 198n4 Henrich, D. 35, 47n20, 50, 71n3, 72n3, 74n11, 106, 166 – 7

Index  213 Hintikka, J. 45n11, 73, 112 – 13, 140n4 Howell, R. 47n22, 53, 106 – 9, 112, 132 – 5, 139n1, 146 – 7 human being 22, 61, 93, 99, 177 – 86, 189, 197 Hume, D. 25 – 7, 29, 46n16, 71n2, 83, 86, 138 immunity to error through misidentification 13, 17 – 20, 76, 90 – 1, 163, 170, 199n8 indexicality 13 – 17, 20, 44n6, 44n7, 76, 103 – 9, 111 – 14, 118 – 23, 127 – 9, 130 – 8, 139n2, 141n11, 144, 148, 153, 163, 170, 176 introspection 9 – 10, 29 – 30, 43n2, 46n16, 66, 84, 86, 196; see also sense, inner intuition 22 – 37, 45n11, 45n12, 46n15, 47n19, 47n20, 52 – 3, 55 – 6, 60 – 70, 73n7, 74n12, 74n15, 77 – 8, 81 – 2, 86 – 9, 92 – 9, 103 – 4, 109 – 17, 124 – 5, 128 – 9, 132 – 3, 140n4, 143, 145 – 55, 168 – 9, 177 – 9, 189 – 94 I think 29, 37 – 43, 49 – 60, 61 – 4, 66 – 8, 71n3, 72n3, 74n10, 74n12, 176 – 82, 185 – 8, 195; and de se thoughts 160 – 70; and the direct reference theory 104 – 9, 130 – 9; I think qua Thinking vs. I think qua Representation 68 – 71, 75, 87, 92, 101, 103, 165, 176; and the question of self-identification 75 – 83, 88 – 101 James, W. 13 judgment 22 – 4, 32 – 43, 45n10, 46n19, 56 – 60, 69 – 71, 73n8, 77, 87, 101n3, 102n8, 102n9, 113, 120 – 3, 134, 137 – 8, 140n6, 162 – 6, 169 – 70, 175 – 6, 180 – 5, 191; faculty of 22, 24, 41, 89, 149, 161, 183; logical form of singular judgments 114 – 17; of perception vs. of experience 97 – 101, 187 – 8 Kitcher, P. 71n2, 74n14, 83 – 6, 100 knowledge by acquaintance vs. knowledge by description 143 – 5 Kripke, S. 108, 114, 118, 120, 122, 141n8 Leibniz, G.W. 58, 115 – 16, 124 Locke, J. 10, 25, 45n14, 83, 120 – 1, 128, 166, 196, 199n10

Longuenesse, B. 61, 62, 89 – 92, 98, 107, 179 – 83 McDowell, J. 24, 45n10, 71n2, 83 – 6, 90, 110, 113, 148, 151 – 2, 182, 198n5 natural kind terms 118 – 25, 127 – 9, 141n9, 141n10, 141n11 no-ownership reading 71n2, 79, 83 – 7, 103, 134 – 5; and the thesis of exclusion 87 – 8 paralogisms of pure reason 50, 54 – 8, 70, 72n4, 72n6, 77 – 8, 82, 88, 105, 107, 109, 132, 137 – 8, 179 Parsons, C. 45n11, 112 – 13, 140n4 Peacocke, C. 12, 20, 44n7, 44n9, 83, 87 – 8, 107 – 8, 110 Perry, J. 14 – 17, 44n6, 91, 143, 156, 158, 163 – 4, 174 – 5, 198n6, 199n9 phenomenon/noumenon 54, 65 – 6, 73n10, 74n13, 74n15, 93, 120 – 1, 145 – 6, 150, 154 – 5, 198n3 Prauss, G. 47n21, 74n15, 99, 102n9 Putnam, H. 114, 118 – 20, 124, 128, 141n8 Quine, W. V. O. 147 Recanati, F. 91, 143 – 4, 156 – 9, 162 – 5, 174 – 5, 198n6, 199n7, 199n8 Rosefeldt, T. 51, 59 – 60, 107, 142n17, 179 Rosenberg, J. F. 32, 39, 50, 66 Russell, B. 43n1, 139n3, 143 self-affection 27 – 9, 95, 102n5, 168, 184, 188 – 93 self-consciousness 12 – 14, 18, 20 – 2; 28 – 32, 37 – 8, 50 – 3, 56 – 7, 63 – 4, 70, 71n3, 74n11, 75 – 7, 85, 89 – 90, 98, 106, 136 – 7, 161, 166 – 77, 184, 192, 199n11; an introspective account of 25 – 7; see also sense, inner Sellars, W. 110, 148 sense, inner 9 – 10, 20 – 3, 25 – 7, 36, 41, 43n2, 46n15, 46n17, 52, 55, 58, 60 – 2, 74n13, 82, 85 – 6, 92 – 6, 102n5, 138, 165, 168 – 9, 177 – 9, 183 – 95; the epistemic limits of 29 – 32; and the ideality of time 27 – 9; outer 10, 23, 25, 27 – 8, 61 – 2, 93 – 6, 99, 101n2, 165, 179, 184 – 5, 187, 189 – 91, 194 – 6

214 Index Shoemaker, S. 13, 16 – 19, 44n7, 91, 105 – 6, 158, 199n8 Strawson, P. F. 19, 50, 79, 82 – 4, 87, 89 – 90 subject, thinking 21, 29 – 30, 38 – 9, 46n16, 49, 51 – 60, 64 – 6, 69 – 70, 71n3, 75 – 82, 85 – 8, 93, 97, 131 – 6, 138, 154 – 6, 158, 160, 168 – 9, 177, 179 – 80, 185, 187, 192, 196; logical 51, 56, 58 – 60, 68, 70, 134 – 5, 137; transcendental 26, 49, 51 – 2, 63, 65 – 8, 70, 74n12, 74n14, 75, 77 – 9, 87 – 8, 101, 103, 130 – 1, 134 – 7, 139, 154 – 6, 161, 180 – 3 substantial ownership reading 50, 82, 71n3, 72n3 thing in itself 64 – 5, 73, 74n15, 146 thoughts, de dicto 17, 107, 135 – 6, 138, 147 – 8, 153; de re 17, 108 – 9, 136,

143 – 8, 152 – 9, 162 – 3, 165, 169, 198n6; de se 14, 17, 20, 38, 76, 101, 108 – 9, 143, 156 – 60, 162 – 5, 170 – 1, 198n6, 199n7 transcendental deduction 34 – 7, 41, 51 – 3, 59, 62 – 3, 71n2, 72n2, 98, 149 – 51, 168, 177, 188, 190 transcendental designation 68, 79 – 81, 89, 91 – 2, 103 – 6, 130 – 9, 161 – 2, 165, 176 transcendental object 51, 66 – 7, 145 – 8, 152 – 6, 198n3 Van Cleve, J. 37, 66, 72n6 Wittgenstein, L. 13, 17 – 19, 50, 76, 79 – 84, 89, 91 – 3, 97, 100 – 1, 157, 162, 165, 199n9 Zahavi, D. 199n11, 200n11